KRBFull

Kraft recovery boilers -

Principles and practice

Vakkilainen
Esa K.
KRAFT RECOVERY BOILERS
Principles and practice

Esa K. Vakkilainen
Title/Nimi: Kraft recovery boilers - Principles and ptractice
Author/Tekijä: Vakkilainen, Esa K.
Copyright © 2005, Esa K.Vakkilainen
Publisher/Julkaisija: Suomen Soodakattilayhdistys r.y.
Printer/Painopaikka: Valopaino Oy, Helsinki, Finland
Year/Painovuosi: 2005
Pages/Sivuja: 246
ISBN 952-91-8603-7
Preface
This book was created as postgraduate lecture

notes for Lappeenranta University of Technology’s
special course of steam power plants. But as with
anything ever written the ideas shown have nur-
tured for a long time. Parts of these chapters have
appeared elsewhere as individual papers or work
documents. One of the most helpful episodes have
been presentations and discussions during Pohto
Operator training seminars. Input from those ses-
sions can be seen in chapter firing. You who run
recovery boilers, I salute you.
The purpose of this text is to give the reader an

overview of recovery boiler operation. Most parts
of the recovery boiler operation are common to
boilers burning other fuels. The furnace operation
differs significantly from operation of other boiler
furnaces. Oxygen rich atmosphere is needed to
burn fuel efficiently. But the main function of
recovery boiler is to reduce spent cooking chemi-
cals. Reduction reactions happen best in oxygen
deficient atmosphere. This dual, conflicting nature
of recovery furnace makes understanding it so
challenging.
To understand the processes happening in the

recovery furnace one must try to understand the
detailed processes that might occur and their limi-
tations. Therefore chapters on materials, corrosion
and fouling have been added.
Many thanks to Marja and Ari Heinola, who not

only read through the manuscript, but provided
many valuable comments. The book greatly
benefited from drawings, illustrations and other
material given by Esa Vihavainen and Kari Haaga.
As always all the errors, omissions and incompre-
hensible ideas are solely my own fault.
Helsinki, 25.3.2005
Contents i
About the author
Esa Vakkilainen is a graduate of Lappeenranta various technical aspects of kraft recovery boilers.
University. He made his M.Sc., Licentiate and Ph. Main interests have been fouling of heat transfer
D. there. He concluded his graduate studies by surfaces, liquor spraying, air distribution and
enrolling to Institute of Paper Science and Tech- black liquor combustion. He has extensive list of
nology in Atlanta, USA. Esa Vakkilainen started publications in these areas.
his professional career in Lappeenranta University
as an assistant professor (yliassistentti) of Power In 2001 Esa Vakkilainen was employed by Jaakko
plants teaching and researching furnace heat Pöyry Oy, international consultants to pulp and
transfer, optimization of combined cycle proc- paper. He has been involved with majority of
esses, district heating (a Finnish specialty) and recent worldwide recovery boiler purchases.
combustion of biofuels.
Esa Vakkilainen is currently an associate professor
After four years of academics Esa Vakkilainen (dosentti) in Lappenranta University and Helsinki
joined A. Ahlström Corporation and started University of Technology. He has directed over 20
working at their Varkaus Boiler works. He was M. Sc. and 3 Ph. D. theses. Esa Vakkilainen has
responsible for developing a new generation of di- lectured of recovery boilers in technical confer-
mensioning programs for steam generator thermal ences at all major continents. He was the technical
design. It was time of extensive development as chairman for the 2004 International Chemical
circulating fluidized beds were making their way Recovery Conference.
to the mainstream of steam power plants. After
several changes of ownership these boilers are
now part of the Foster Wheeler Corporation.
In 1989 Esa Vakkilainen found himself involved

with kraft recovery boilers, his destiny for the next
15 years. The years with A. Ahlström Corpora-
tion, then Ahlstrom Machinery were spent with
ii Kraft recovery boilers

Contents
Preface i 6 DIMENSIONING OF HEAT
About the author ii TRANSFER SURFACES 6-1
Nomenclature v 6.1 Main heat transfer surfaces 6-2
1 PRINCIPLES OF KRAFT 6.2 Heat load calculation 6-3
RECOVERY 1-1 6.3 Natural circulation 6-5
1.1 Function of recovery boiler 1-1 6.4 Key furnace sizing characteristics 6-7
1.2 Early recovery technology 1-2 6.5 Superheater section dimensioning 6-12
1.3 First recovery boilers 1-4 6.6 Vertical boiler bank design 6-15
1.4 Development of recovery boiler 6.7 Vertical economizers 6-15
technology 1-5 6.8 Heat transfer in boilers 6-16
1.5 High dry solids 1-7 6.9 Example calculation of heat transfer
1.6 Improving air systems 1-7 surface 6-18
1.7 High temperature and pressure 6.10 Effect of high dry solids to recovery
recovery boiler 1-9 boiler dimensioning 6-20
1.8 Gasification 1-10 7 RECOVERY BOILER PROCESSES 7-1
1.9 Alternative recovery 1-11 7.1 Air system 7-2
2 RECOVERY BOILER DESIGN 2-1 7.2 Flue gas system 7-7
2.1 Key recovery boiler design options 2-1 7.3 Water and steam 7-8
2.2 Evolution of recovery boiler design 2-3 7.4 Black liquor and ash 7-14
2.3 Choosing recovery boiler main 7.5 Oil/gas system 7-15
parameters 2-9 7.6 Green liquor 7-16
2.4 Projecting a recovery boiler 2-11 7.7 Auxiliary equipment 7-16
3 MATERIAL AND ENERGY 8 FOULING 8-1
BALANCE 3-1 8.1 Ash deposits on heat transfer
3.1 Material balance 3-1 surfaces 8-1
3.2 Energy balance 3-6 8.2 Formation of ash particles in
3.3 Radiation and convection heat losses 3-8 recovery boiler 8-6
3.4 Steam generation efficiency 3-8 8.3 Deposition of particles and vapors 8-9
3.5 High dry solids black liquor 3-8 8.4 Properties of recovery boiler ash 8-12
4 COMBUSTION OF BLACK 8.5 Deposit behavior 8-16
LIQUOR 4-1 8.6 Deposit removal 8-21
4.1 Drying 4-2 8.7 How to decrease fouling rate 8-22
4.2 Devolatilization 4-3 9 FIRING BLACK LIQUOR 9-1
4.3 Char combustion 4-6 9.1 Liquor gun operation 9-1
4.4 Smelt reactions 4-7 9.2 Liquor gun type 9-6
4.5 Experimental procedures to look at 9.3 Air distribution 9-9
black liquor combustion 4-7 9.4 Reduction control 9-12
4.6 Combustion of black liquor droplet 9.5 NOx control in recovery boilers 9-12
in the furnace 4-8 9.6 SOx and TRS control 9-18
4.7 Combustion properties of high dry 9.7 Firing black liquor – mill experience 9-19
solids black liquor 4-8 9.8 Burning waste streams in a recovery
5 CHEMICAL PROCESSES IN boiler 9-24
FURNACE 5-1 9.9 Auxiliary fuel firing 9-27
5.1 Lower furnace gas phase 5-1 10 MATERIAL SELECTION AND
5.2 Char beds 5-4 CORROSION 10-1
5.3 Smelt 5-8 10.1 Gas side corrosion of heat transfer
5.4 Sodium 5-10 surfaces 10-1
5.5 Potassium 5-13 10.2 Furnace corrosion 10-6
5.6 Sulfur 5-14 10.3 Erosion 10-9
5.7 Chloride 5-18 10.4 Water side corrosion 10-10
5.8 Reactions involving carbon 5-19 10.5 Furnace design and materials 10-13
10.6 Superheater design and materials 10-15
Contents iii
10.7 Screen design and materials 10-18
10.8 Boiler bank design and materials 10-18
10.9 Economizer design and materials 10-19
11 EMISSIONS 11-1
11.1 Typical emissions 11-1
11.2 Reduced sulfur species 11-1
11.3 Carbon monoxide 11-2
11.4 Carbon dioxide 11-3
11.5 NOx 11-3
11.6 VOC 11-8
11.7 Dust emissions 11-8
11.8 Sulphur dioxide 11-10
11.9 HCl 11-10
11.10 Miscellaneous minor emissions 11-13
11.11 Heavy metals 11-14
11.12 Dissolving tank emissions 11-15
12 REFERENCES 1-1
INDEX I-1
Appendices
A EMISSION CONVERSIONS A-1
A.1 Conversion of average emissions
to time not to exceed A-1
A.2 Conversions A-2
iv Kraft recovery boilers

Nomenclature
A surface area, m2 ∆ difference, -
Ad droplet surface area, m2 ∆T temperature difference, K or oC
ad dust emission coefficient, - Δεg overlapping correction for emissivity, -
Bd dust loading, kg/m3 Δαg overlapping correction for absorptivity, -
b correction coefficient, - Φ heat flow, W
C heat capacity, W/K αd absorption coefficient of dust, -
cp specific heat capacity, kJ/kgK αdg absorptivity of the dusty gas, -
Dp particle diameter, m ε ratio of gas temperature drop to total
d diameter, m temperature difference, -,
do outside tube diameter, m emissivity, -
ds inside tube diameter, m εdg emissivity of the dusty gas, -
f correction factor for heat transfer, - εα background emissivity, -
fn form correction εw emissivity of the wall, -
fo overall correction (fouling correction, etc.) ζ loss coefficient, -
G conductance, W/K η viscosity, Pas,
h convective heat transfer coefficient, W/m2K efficiency, -
k heat transfer coefficient, W/m2K θ total temperature difference, oC
kc convective heat transfer coefficient λ heat conductivity, W/moC
ki inside heat transfer coefficient referred to ξ friction factor, -
the outside surface ρ density, kg/m3
ko outside heat transfer coefficient σ Stefan Bolzman coefficient, W/m2K4
kr radiative heat transfer coefficient τ particle relaxation time, s
kex external heat transfer arranged to represent
a heat transfer coefficient
L length, m
l heat of vaporization, kJ/kg
stopping distance, m
n number of, -
mw mass of water in droplet, kg
mo mass of pyrolysable material in droplet, kg
Pr Prandtl number based on gas phase, -
p pressure, bar abs
px partial pressure of substance x, bar
Q heat flow, J
qm mass flow, kg/s
Red Reynolds number based on droplet
diameter and relative speed, -
R ratio of steam side heat capacity to gas
side heat capacity, -
r radius, m
s radiation beam length, m,
tube wall thickness, m
Sl longitudinal pitch, m
St transverse pitch, m
sl dimensionless longitudinal pitch, -
st dimensionless transverse pitch, -
T temperature, K or oC
Tg gas temperature, K
Td droplet temperature, K
U free stream velocity, m/s
w flow velocity, m/s
X fraction, -
x dry solids, -
z number of transfer units, -
Contents v
vi Kraft recovery boilers
1 Principles of kraft recovery
Figure 1-1, Kraft mill unit operations.
Spent cooking chemicals and dissolved organ- non process elements is prevented by disposal of
ics are separated from pulp during washing. This dregs and grits at causticizing. Malodorous non
black, alkaline liquor was at first dumped. Various condensable gases are processed by combustion
chemical recovery systems were then developed at recovery boiler or lime kiln. In some modern
(Niemelä, 2004), but it was in the 1930’s and and closed mills chloride and potassium removal
40’s when modern type of regeneration of spent processes are employed. With additional closure
liquor was widely adopted. New type of equip- new internal chemical manufacturing methods are
ment increased line size and led to more favorable sometimes applied.
economic situation.
Recovery of black liquor has other advantages. 1.1 FUNCTION OF RECOVERY

Concentrated black liquor can, when burnt, pro- BOILER
duce energy for generation of steam and electric-
ity. In the most modern pulp mills, this energy is Concentrated black liquor contains organic
more than sufficient to cover all internal use, dissolved wood residue in addition of cooking
The principal kraft recovery unit operations are, chemicals. Combustion of the organic portion of
Figure 1-1, evaporation of black liquor, combus- chemicals produces heat. In the recovery boiler
tion of black liquor in recovery boiler furnace heat is used to produce high pressure steam,
including of formation of sodium sulfide and
sodium carbonate, causticizing of sodium carbon-
ate to sodium hydroxide and regeneration of lime
mud in lime kiln.
There are other minor operations to ensure con-

tinuous operation of recovery cycle. Soap in the
black liquor can be removed and tall oil produced.
Control of sodium - sulfate balance is done by
addition of makeup chemicals such as sodium
sulfate to mix tank or removal of recovery boiler
ash. Dumping of recovery boiler ash removes
mostly sodium and sulfur, but serves as an impor- Figure 1-2, Net heating value of typical kraft liquor
tant purge for chloride and potassium. Buildup of at various dry solid content.
Principles of kraft recovery 1-1

Figure 1-3, One of the latest recovery boilers, Gruvö from Kvaerner.
which is used to generate electricity in a turbine. green liquor

The turbine exhaust, low pressure steam is used 4. Recovery of inorganic dust from flue gas to save
for process heating. Medium pressure extraction chemicals
steam from between turbine stages is used in 5. Production of sodium fume to capture combus-
cooking, sootblowing and high solids evaporation. tion residue of released sulfur compounds
Combustion in the recovery boiler furnace needs

to be controlled carefully. High level of sulfur 1.2 EARLY RECOVERY
in the black liquor requires optimum process TECHNOLOGY
conditions to avoid production of sulfur dioxide
and reduced sulfur gases emissions. In addition to Early recovery technology concentrated on
environmentally clean combustion, reduction of chemical recovery (Deeley and Kirkby, 1967).
inorganic sulfur must be achieved in the char bed. Chemicals cost money and it was easy to discover
that recycling these chemicals would improve the
The recovery boiler process has several unit proc- profitability of pulp manufacture.
esses
1. Combustion of organic material in black liquor Recovery of pulping chemicals could be based
to generate steam to French chemist Nicholas LeBlanc’s process for
2. Reduction of inorganic sulfur compounds to producing soda at reducing furnace. A flame oven
sodium sulfide was hand filled with black liquor, Figure 1-4. Then
3. Production of molten inorganic flow of mainly the black liquor was dried with flue gases from
sodium carbonate and sodium sulfide and dissolu- burning wood. The dried black liquor was then
tion of said flow to weak white liquor to produce scraped to floor, collected and sent to separate
Figure 1-4, Early flame oven from late 1800 Figure 1-5, Early smelt pot from late 1800
(Edling, 1981). (Edling, 1981).
1-2 Kraft recovery boilers

ds. The evaporation took place in a four stage
evaporator. The heat was recovered with vertical
tube boiler.
Tampella was among the first manufacturers to

build S-S type furnaces, Figure 1-7. Preventing
unnecessary air flow through sealing arrangement
between rotary drum and fixed parts was one of
Figure 1-6, Rotating oven from 1890, liquor in at the major operating problems. The combustion
20 % ds (Edling, 1981). was often conducted at very high air ratio leading
to inefficient energy use. One could generate 3000
… 4000 kg of 3.0 MPa steam for each ton of pulp
(Roschier, 1952).
The main recovery equipment itself remained

unchanged, but details were improved on. Smelt
dissolving tank was introduced, final smelting was
improved on and capacities grew, Figure 1-8. The
use of refractory and rotative oven tended to limit
the recovery capacity to 70 … 75 tds/d (Sebbas et
al., 1983). Rotary part lengths were 7 … 10 m and
Figure 1-7, Early Tampella rotary furnace from diameter about 1.5 m (Swatrz and MacDonald,
about 1925 (Tampella). 1962).
The boilers parts were improved on. In 1930s even

LaMont type forced circulation units were built,
Figure 1-9. The use of rotary furnaces pinnacled
in Murray-Waern type units which were success-
fully built around the world. In these the rotary
precombustion was combined with totally water-
cooled furnace with lower part refractory lined.
The Murray-Waern recovery units were popular
until the fifties.
An epitome of inventivness of that age was the
Godell recovery unit at Stevens Point Wise in
Figure 1-8, Early S-S rotary furnace from about 1940’s. There the liquor was completely dried in
1945 (Vannerus et al., 1948). a chamber after the boiler (Edling 1983). But that
smelt pot, Figure 1-5, for reduction and burning
the remaining organics (Rydholm, 1965). Re-
covery of chemicals with this type of system was
inefficient. Chemicals recovery hardly exceeded 60
% (Whitney, 1968).
Hand operated recovery grew more complicated

with additional heat recovery surfaces. Pre-
evaporation and scrubbing in a rotary device
was invented by Adolph W. Waern (Combustion
Engineering, 1949). The direct contact evapora-
tor improved the heat economy of the recovery
system. The hand feeding operation was soon
replaced by rotative oven, Figure 1-6.
Use of rotary oven improved the heat economy.

Then it was a small step to introduce heat recovery
equipment as was done with other types of boilers
at that time. In 1912 the S-S system (Sundblad-
Sandberg) was taken online at Skutskär. In it Figure 1-9, LaMont type construction used in
liquor was sprayed into rotary furnace at 50 % Kotka, Moss and Frantsach ~1930 (Edling, 1981).

Figure 1-10, First Tomlinson kraft recovery boiler with water cooled furnace from Babcock & Wilcox in 1934
(Steam, 1992). Note spray tower using weak black liquor before the ID fan.
unit and most of those other alternative systems Recovery boiler also improved the smelt removal.
were hard to operate and did not achieve high It is removed directly from the furnace through
availability. smelt spouts into a dissolving tank. Some of the
first recovery units employed the use of Cottrell’s
electrostatic precipitator for dust recovery.
1.3 FIRST RECOVERY BOILERS
Babcock & Wilcox was founded in 1867 and
The modern recovery boiler has a few strong gained early fame with its water tube boilers. It
ideas that have remained unchanged until today. built and put into service the first black liquor re-
It was the first recovery equipment type where all covery boiler in the world in 1929 (Steam, 1992).
processes occurred in a single vessel. The drying, This was soon followed by a unit with completely
combustion and subsequent reactions of black water cooled furnace at Windsor Mills in 1934.
liquor all occur inside a cooled furnace. This is the After reverberatory and rotating furnaces the
main idea in Tomlinson’s work. recovery boiler was on its way (Jones, 2004).
Secondly the combustion is aided by spraying The second early pioneer, Combustion Engi-
the black liquor into small droplets. Controlling neering based its recovery boiler design on the
process by directing spray proved easy. Spraying pioneering work of William M. Cary, who in 1926
was used in early rotary furnaces and with some designed three furnaces to operate with direct
success adapted to stationary furnace by H. K. liquor spraying and on work by Adolph W. Waern
Moore. Thirdly one can control the char bed by and his recovery units.
having primary air level at char bed surface and
more levels above. Multiple level air system was The first CE recovery unit, Figure 1-11, looks a lot
introduced by C. L. Wagner. like a modern recovery boiler. Note direct contact
evaporator on the left, cooled floor tubes and
three drum construction.
Recovery boiler were soon licensed and produced

in Scandinavia and Japan. These boilers were built
by local manufacturers from drawings and with
instructions from licensors. One of the early Scan-
dinavian Tomlinson units employed a 8.0 m high
furnace that had 2,8*4,1 m furnace bottom which
expanded to 4,0*4,1 m at superheater entrance
(Pettersson, 1983). This unit stopped production
for every weekend. In the beginning economizers
had to be water washed twice every day, but after
installation of shot sootblowing in the late 1940s
the economizers could be cleaned at the regular
weekend stop.
The construction utilized was very successful.

One of the early Scandinavian boilers 160 t/day
at Korsnäs, Figure 1-12, operated still almost 50 Figure 1-11, The first CE recovery boiler 1938
years later (Sanquist, 1987). Edling states in 1937 (Combustion Engineering, 1949).
that more than 20 units had already been built of
which 10 in Scandinavia.
1.4 DEVELOPMENT
OF RECOVERY BOILER
TECHNOLOGY
Spread of kraft recovery boilers was fast as
functioning chemical recovery gave kraft pulping
an economic edge over sulfite pulping (Boni-
face, 1985). They had about 20 % better energy
efficiency as more than 5000 kg of 3.0 MPa steam
for each ton of pulp could be generated (Roschier,
Figure 1-12, Korsnäs recovery boiler started opera-
1952, Alava, 1955). The first recovery boilers
tion in 1943 (Götaverken).
had horizontal evaporator surface followed with
superheaters and more evaporation, Figure 1-13.
These boilers resembled the state-of-the-art
boilers of some 30 years earlier. This trend has
continued until today. It is easy to understand that
when any stop will cost a lot of money the adopted
technology tends to be conservative. Conserva-
tism meant that e.g. the new Oulu Oy, 100 000
t/a sulphate mill installed four Tomlinson boilers
when it started operating in 1937 (Oulu, 1937).
The first recovery boilers had severe problems

with fouling (Deeley and Kirkby, 1967, Roschier,
1952). Tube spacing wide enough for normal
operation of a coal fired boiler had to be wider
for recovery boilers. This gave satisfactory per-
formance of about a week before a water wash.
Mechanical steam operated sootblowers were also
quickly adopted, Figure 1-13. To control chemical
losses and lower the cost of purchased chemicals
electrostatic precipitators were added. Lowering
dust losses in flue gases has more than 60 years of Figure 1-13, Early Tomlinson recovery boiler (Van-
practice. nérus et al., 1948).

One should also note square headers in the 1940 Single drum design
recovery boiler (Vannérus et al., 1948), Figure
1-13. There are some early examples of single drum re-
covery boilers. Both B&W and Ahlstrom delivered
The air levels in recovery boilers soon standard- a single drum boiler in the late 1950’s. The first
ized to two. The primary air level was placed at the modern single drum recovery boiler was delivered
char bed level and the secondary above the liquor in 1984 by Götaverken to Leaf River at Hatties-
guns. burg, Mississippi. The boiler size was 1966 tds/d.
By 1990 all manufacturers started providing single
In the first tens of years the furnace lining was drum boilers. Excluding very small boilers, all
often of refractory brick or refractory on cast modern boilers are now of single drum design.
blocks. The flow of smelt on the walls causes
extensive replacement and soon designs that There are several advantages in a single drum
eliminated the use of refractory were developed. boiler. Single drum construction eliminates the
The standard then became the tangent furnace possibility of water leakage to furnace as it is
wall. Membrane wall use became widespread in placed outside the furnace. There are significantly
the 60’s. less holes in a drum wall. Therefore it can be built
thinner. Thinner wall of drum allows faster start
B&W-design up and stop-down. The gas flow to the boiler bank
is smoother and heating surface arrangement is
B&W favored use typically a single black liquor simple. The erection period is shorter because of
gun at front wall. In larger units additional gun large block construction. There is no rolled tube
was placed on back wall (Tomlinson and Richter, work. Enhanced and steady water circulation by
1969). They preferred a significant part of the separated and unheated downcomers.
liquor to be sprayed to walls for drying. Boiler
bottom was in angle causing smelt to flow quickly The largest advantage is that single drum boilers
out. Hardly any space was reserved for smelt layer can be made larger. Tube stiffness limits cross flow
in the furnace. Thus this kind of furnace is named two drum arrangement to about 2300 tds/d size
sloping bottom type. (Steam, 1992). Vertical flow two drum construc-
tions have suffered from plugging because of
Final black liquor evaporation was often carried vibration stiffeners.
in a direct contact evaporator of venturi scrubber
or cyclone evaporator type. The highest practical Furnace protection
black liquor solids was 60 … 65 % depending on
black liquor properties. Use of wall spraying was First recovery units had brick lined lower furnace
promoted by B&W and its licensees Götaverken with straight tubes forming cooling section be-
and Babcock Hitachi. B&W adopted three level air hind bricks. This design persisted until the 1960s.
in the late 1960s. Some of these units are still operating today.
Another design provided corrosion protection
CE-design of furnace tubes with studs and refractory. Some
manufacturers use studs even today, but need of
Early on the CE design stressed use of multiple stud replacement has led to decline of stud use.
guns in all walls (Tomlinson and Richter, 1969). From late fifties onwards the membrane wall
Boiler bottom was flat with space for smelt layer design took over, first with carbon steel walls.
on top of the whole floor. Thus this kind of fur- Tangent tube design was replaced with membrane
nace is named decanting floor type. Final black design. The drawbacks of tangent design were the
liquor evaporation was carried in a direct contact difficulties in inspecting welds and doing mainte-
evaporator of cascade evaporator type. nance work
The basic aims of recovery boiler design could First furnace walls were of carbon steel. With
soon be summarized as; highest possible recovery increasing design pressure there were several cor-
of chemicals, high efficiency, high utilization of rosion problems in lower furnace. The advantage
the calorific values in black liquor and highest of chrome containing alloys as wall corrosion
safety of operation (Hochmuth, 1953). inhibitor was discovered as an answer to high
pressure boiler sulfidation corrosion (Moberg,
CE sticked for a long time with a two level air 1974). In 1972 Tampella delivered first totally
system that had corned fired secondary. They used compound tube recovery boiler furnace to ASSI
similar system in PCF-boilers. Lövholmen mill in Piteå, Sweden. By 1982 there
were 30 recovery boilers with 304 compound tube
bottoms in Scandinavia (Westerberg, 1983). Use
of composite tubing in United States started only
in 1981.
Sanicro 38 is a widespread material that offers

improved corrosion protection for lower furnace.
The first lower furnace made from Sanicro 38 was
delivered by Kvaerner. They used Sanicro 38 in
the lower furnace up to primary ports in 1994 for
their Rauma recovery boiler.
Economizer
Earlier the recovery boilers had horizontal tube
economizers. They plugged fast and had to be
water washed at intervals of 1- 4 weeks (Rissanen,
1965). It was not until the early 1960 that install-
ing vertical economizers started. In economizers
of vertical flow design the gas flows downwards
and water counter currently upwards (Hyöty,
1994). In a period of few years the current long Figure 1-14, One of the first long flow economizers,
flow economizer design emerged, Figure 1-14 Sunila (Moberg, 1967).
(Moberg, 1967). Vertical economizer design loading (Vegeby, 1961). One advantage noted was
spread fast in Scandinavia where by mid 1970’s that partial load capability improved with higher
more than half of the recovery boilers had long dry solids. Increasing black liquor dry solids from
flow economizers without direct contact evapora- 60 % to 68 % enabled running recovery boiler
tor (Environmental, 1976). without auxiliary fuel firing at 65 % of rated MCR
(Rissanen, 1965). At 60 % dry solids hardly any
In competition to purely vertical, the three partial load could be run.
pass design featured gas flow which was forced
crosswise the economizer tubes to improve heat In 1980’s the first high dry solids units started
transfer. coming on line. Extensive tests of effect of
increasing dry solids from 72 % to 84 % were
There have been several rounds of economizer run at Metsä-Botnia Kemi and Rosenlew, Pori,
header designs. In a typical old design each Finland recovery boilers (Hyöty and Ojala, 1987).
economizer tube is connected to a common large They noticed that above 75 % dry solids the SO2
header. As maximum number of tube rows that and H2S emissions were practically zero. Also
fit to this type header is about 8 … 10. The larger reduction increased more than one percentage
economizers must have front and back head- point. Other benefits listed were steam generation
ers. This design has the disadvantage of having increase and boiler controllability increase. High
a header in the gas flow. The header can corrode dry solids require that ESP ash is mixed to the
and the welded joints tend to receive thermal black liquor with 62 … 65 % liquor. Higher reten-
stress. Modern economizers have flat horizontal tion time also improves the stability of resulting
headers. black liquor.
1.5 HIGH DRY SOLIDS 1.6 IMPROVING AIR SYSTEMS

Dry solids at as fired black liquor was between Air system development continues and has been
60 and 65 % in Sweden at the beginning of the continuing as long as recovery boilers existed
1960’s (Jönsson, 1961). In the beginning of 1950’s (Vakkilainen, 1996). As soon as the target set for
the typical as fired black liquor concentration was the air system has been met other new targets
50 % (Vegeby, 1961). The final concentration was are given. Currently the new air systems have
often done with cascade or cyclone evaporator. In achieved low NOx, but are still working on with
practice the as fired dry solids could remain dan- lowering fouling.
gerously low before refractometers started to be
applied in late 1960s and early 1970s (Hellström, The first generation air system in the 1940’s and
1970). The only reasons seen for higher dry solids 1950’s consisted of a two level arrangement; pri-
were the energy economy and increase of bottom
Table 1-1, Development of air systems (Vakkilainen, 1996).
Air system Main target But also should

1st generation Stable burning of black liquor
2nd generation high reduction Burn liquor
3rd generation decrease sulfur emissions Burn black liquor, high reduction
4th generation low NOx Burn black liquor, high reduction and low
sulfur emission
5th generation decrease superheater and boiler Burn black liquor, high reduction, low emis-
bank fouling sions
mary air for maintaining reduction zone and sec- system the air nozzles are in the furnace corners.
ondary air below the liquor guns for final oxida- The preferred method is to create a swirl of almost
tion (Llinares and Chapman, 1989). The recovery the total furnace width. In large units the swirl
boiler size was 100 … 300 tds/d and black liquor caused left and right imbalance. This kind of air
concentration 45 … 55 %. Frequently to sustain system with increased dry solids managed to
combustion auxiliary fuel needed to be fired. Pri- increase lower furnace temperatures and achieve
mary air was 60 … 70 % of total air with second- reasonable reduction. B&W had already adopted
ary the rest. In all levels openings were small and the three level air by then, Figure 1-15.
design velocities were 40 … 45 m/s. Both air levels
were operated at 150 oC. Liquor gun or guns were At first the air port openings were made by bend-
oscillating. Main problems were high carryover, ing one tube away from the opening sideways and
plugging and low reduction. But the main target, making room for this by bending another tube
burning of black liquor could be done. back, Figure 1-16. Airport width was about tube
spacing and large plate areas were needed to make
The second generation air system targeted high airport gas tight. In 1978 CE began experiment-
reduction. In 1954 CE moved their secondary air ing with two level primary air. Upper primary was
from about 1 m below the liquor guns to about designed to about 20 % of total air with velocity
2 m above them (Llinares and Chapman, 1989). up to 60 m/s. Total air split remained the same.
The air ratios and temperatures remained the The aim was to increase hearth temperatures.
same, but to increase mixing 50 m/s secondary air
velocities were used. Third generation air system was the three level air.
In Europe the use of three level air with primary
CE changed their frontwall/backwall secondary and secondary below the liquor guns started about
to tangential firing at that time. In tangential air 1980. At the same time stationary firing gained
Figure 1-15, Typical two level air CE, left and three level air BW, right from early 1960 (Roos, 1963).

Figure 1-16, Typical air port from 1960’s
(Soodakattiloiden, 1968).
ground. Use of about 50 % secondary seemed to

give hot and stable lower furnace (Westerberg,
1983). Higher black liquor solids 65 … 70 %
started to be in use. Hotter lower furnace and
improved reduction were reported. With three
level air and higher dry solids the sulfur emissions
could be kept in place.
Fourth generation air systems are the multilevel

air and the vertical air. As black liquor dry solids
to the recovery boiler have increased, achieving Figure 1-17, Principle of vertical air (Kaila and
low sulfur emissions is not anymore the target of Saviharju, 2003.
the air system. Instead low NOx and low carryover
are the new targets. traditional vertical mixing to horizontal mix-
ing. Closely spaced jets will form a flat plane. In
traditional boilers this plane has been formed by
Multilevel air secondary air. By placing the planes to 2/3 or 3/4
The three level air system was a significant im- arrangement improved mixing results. Vertical air
provement, but better results were required. Use has a potential to reduce NOx as staging air helps
of CFD models offered a new insight of air system in decreasing emissions (Forssén et al., 2000b).
workings. The first to develop a new air system
was Kvaerner (Tampella) with their 1990 multi-
level secondary air in Kemi, Finland, which was
1.7 HIGH TEMPERATURE AND
later adapted to a string of large recovery boilers PRESSURE RECOVERY BOILER
(Mannola and Burel, 1995).
Development of recovery boiler main steam pres-
Kvaerner also patented the four level air system, sure and temperature was rapid at the beginning,
where additional air level is added above the Figure 1-18. By 1955, not even 20 years from
tertiary air level. This enables significant NOx birth of recovery boiler highest steam pressures
reduction. were 10.0 MPa and 480 oC (Vakkilainen et al.,
2004). The typical pressures and temperatures
then backed downward somewhat due to safety
Vertical air (McCarthy, 1968). By 1980 there were about 700
recovery boilers in the world (Westerbeg, 1983).
In vertical air primary is arranged conventionally.
In Japan, because of high electricity prices, more
Rest of the air ports are placed on interlacing 2/3
than ten high temperature and pressure recovery
or 3/4 arrangement, Figure 1-17. Vertical air was
boilers are in use (Tsuchiya et al., 2002). The big-
invented by Erik Uppstu (1995). His idea is to turn
gest one is the 2700 tds/d, 10.3 MPa and 505 oC

Figure 1-18, Development of recovery boiler pressure, temperature and capacity.
recovery boiler at Iwakuni mill (Ohtomo, 2000).

tional targets and reduces production of fossil CO2
New large recovery boilers seem to favor high (Raymond, 2003). More efficient process must be
main steam temperatures and pressures (Vakki- pursued also to offset higher unit cost of gasifica-
lainen, 2004). These increase the amount of back tion (per ton of mass production). One way to
pressure electricity. increase electricity production is to perform gasi-
fication under pressure. Another is to use oxygen
gasification instead of air gasification (Donovan
1.8 GASIFICATION and Brown, 2003). An overview of potential proc-
ess improvements has been presented by McK-
Gasification provides a way to convert solid fuel eough (2003).
into gaseous. Energy conversion always involves
exergy and energy losses. Gaseous fuel enables use Production of extra electricity and high unit cost
of combined cycle power generation, where both of recovery boiler are the main driving forces for
gas turbine and steam process produce electricity. the development of gasification. From energy ef-
This increases energy efficiency (Demirbaş, 2001, ficiency standpoint the recovery boiler has several
Consonni et al., 2003). Higher yield of biofuel weaknesses; the main steam temperature and
based electricity fits with national and interna- pressure are low, the energy in smelt is not recov-
Table 1-2, Effectiveness of converting black liquor

HHV into fuel value in net product gas (Grace and
Timmer, 1995).
Original HHV, 14.7 13.3
MJ/kgbl
HHV in net gas, HHV in net gas,
% of original % of original
Commercial
Chemrec 55.9 50.7
StoneChem TCI) 49.0 42.6
Piloted processes
Figure 1 19, Estimated net power and heat outputs ABB 70.4 66.6
from recovery boiler and gasification of black liquor, Tampella 60.7 55.9
Pulp production 600 000 ADt/a (McKeough, 2004).
Conventional RB 61.1 58.6

ered and combustion temperature is only moder- medium, operation of gasifier at atmospheric
ate. Dealing with reductive/oxidative process, rare pressure, application of quench cooling and high
feature of smelt water explosion and corrosive steam content in gas from gasifier (McKeough,
process media cause high unit costs and require 2004).
extra safety features. The modern recovery boiler
has a very good record for environmental clean- The practical gasification process is still a long way
ness. In spite of this improvements are in need. of becoming a commercially attractive solution
(Hood and Henningsen, 2002). Industry visions of
The two main process alternatives for black liquor early nineties that black liquor gasification could
gasification are high temperature gasification and within next ten years become a viable alternative
low temperature gasification (Backman et al., have not materialized. Even though there are sig-
1993, Warnqvist et al., 2000, Whitty and Verrill, nificant gains to be made, there still remain a lot
2004). In low temperature gasification, the black of unresolved issues; finding materials that survive
liquor is kept below the melting point of ash. This in a gasifier, mitigating increased causticizing load
temperature usually corresponds to significantly and how to make startup and shutdown (Tucker,
less than 700 oC. Heat to low temperature gasifica- 2002).
tion can be supplied either directly or indirectly.
A typical process problem in low temperature
gasification pilot plants has been plugging due to 1.9 ALTERNATIVE RECOVERY
sintering. At low temperature the pyrolysis phase
is still relatively fast, but char gasification requires There are other processes developed to replace
considerable time. conventional evaporator, recovery boiler, causti-
cizing, lime kiln process. One driving force is the
In high temperature gasification the reaction relative inefficiency of the recovery boiler. Even
temperature is at about 1000 oC or higher. High though the recovery boiler compares well with
temperature gasification is fast but means of modern power production it is still the greatest
dealing with molten smelt and protection from entropy producing unit process in the kraft mill
corrosive athmosphere need to be developed. Both (Richards, 2001).
gasification processes require flue gas cleaning to
reduce sulfur and sodium emissions. There is less Na2CO3 + Fe2O3 → Na2Fe2O4 + CO2 11
sulfur release but more alkali release in the high Na2Fe2O4 + H2O → 2NaOH + Fe2O 12
temperature gasification process (Berglin and
Berntsson, 1998). Most of the proposed new processes involve a
number of new stages. This means that much
In gasification higher steam temperature and pres- experimenting and heavy investment is needed.
sure could be used if the product gas is cleaned. Maybe this is why there are only a few alternative
A part of the energy in smelt can be recovered processes that have evolved to the mill scale.
if dissolving is done under pressure. Still higher
combustion temperatures can be achieved with Direct alkali recovery
oxygen gasification.
Australian Paper has employed a process where
Gasification research has been done by all leading liquor from soda pulping is processed in fluidized
recovery boiler manufacturers. There have been bed with ferric oxide (Scott-Young and Cukier,
several key issues in the development. Energy 1995). The product, sodium ferrite is dissolved
conversion causes always extra losses. Processes in water. Sodium hydroxide and iron oxide are
have tended to go for multiple heat exchangers formed. The mill achieved several years of opera-
and recuperators to increase efficiency. It is very tion before it shut down due to uneconomical size
costly and technically difficult to clean gases to of the mill.
required cleanliness for gas turbine. Quench needs
to be applied. So far all commercial processes A mill in Denmark used straw to make pulp for
lack behind the energy efficiency of conventional various applications using direct alkali recovery
recovery boiler, Table 1-2 and Figure 1-19. with ferric oxide. In spite of the partially suc-
cessful recovery operation the mill closed down
Main technical obstacles in the low temperature because of economical reasons.
gasification (e.g. Thermochem) have been achiev-
ing high enough carbon conversion and keeping The development of direct alkali recovery system
the gasifier from plugging. The main factors which (DARS) process is slow because the use of fluid-
lower the overall efficiency of current processes ized bed and subsequent pelletizing and leaching
include; use of air in place of O2 as gasification operations are new for typical pulp mill personnel.
DARS has about 10 % lower thermal efficiency
than recovery boiler (Maddern, 1988). This is
due to extra heat in sodium iron leaving furnace
and higher amount of water in incoming liquor
per unit of virgin black liquor. This may be large
enough penalty to offset the savings when lime
kiln is not needed.
Other much studied component has been tita-

nium dioxide (Kiiskilä, 1979 and Richards, 2001).
Autocausticization
In autocausticization an additional component is
added to the liquor. This component needs to re-
act with sodium hydroxide at furnace and dissolve
back at green liquor. One such chemical is sodium
borate discovered by Janson (1978) and recently
studied by Tran et al., (2001).
Na2CO3 + NaBO2 → Na3BO3 + CO2 13

Na3BO3 + H2O → 2NaOH + NaBO2 14

2 Recovery boiler design
Figure 2-1, Typical recovery boiler in operation, Gruvön (Wallén et al., 2002).
In a pulp mill recovery boiler fulfills three main has increased because there is less water in black
functions. The first is to burn the organic mate- liquor, liquor spraying is now more uniform,
rial in the black liquor to generate high pressure new computer controls mean better stability and
steam. The second is to recycle and regenerate controllability and most importantly, new pulping
spent chemicals in black liquor. The third is to lines of corresponding capacity can be built.
minimize discharges from several waste streams
in an environmentally friendly way. In a recovery
boiler, concentrated black liquor is burned in the 2.1 KEY RECOVERY BOILER
furnace and at the same time reduced inorganic DESIGN ALTERNATIVES
chemicals emerge molten. A modern recovery
boiler, Figure 2-1, has evolved a long way from the There are alternative solutions to design of recov-
first recovery boilers, Figure 2-2.
One noticeable trend has emerged in recent years.

The average size of recovery boiler has grown
significantly in each year, Figure 2-3. The nominal
capacity of new recovery boilers at the beginning
of the 1980s was 1700 metric tons of dry solids per
day. This was regarded as the maximum at that
time (Pantsar, 1988).
By year 2000 more than ten recovery boilers,

capable of handling 2500 … 3500 metric tons of
dry solids per day were built. At 2004 recovery Figure 2-2, One of the first Scandinavian recovery
boilers with nominal capacity of 4450 and 5000+ boilers Korsnäs from 1943, 160 tds/d, 4 MPa, 400
tds/d were started. The maximum design capacity o
C (Sandquist, 1987).
Recovery boiler design 2-1

Figure 2-3, Size of recovery boiler versus startup year.
ery boilers. Major recovery boiler design options In addition to major design features the manu-
are; screen or screenless superheater area layout, facturers like to advertise their equipment with
single drum or two-drum, lower furnace tubing minor design features. In Figure 2-4, Babcock &
material, furnace bottom tubing material, vertical Wilcox, presents their design features.
or horizontal boiler bank, economizer arrange-
ment and number and type of air levels. Key design specifications
When sizing a recovery boiler some key design
specifications are usually given to the boiler ven-
dor to do the design. Typically given are dry solids
capacity, black liquor gross heat value, black liquor
elementary analysis, black liquor dry solids %
from evaporation, desired main steam conditions,
feed water inlet temperature and economizer flue
gas outlet temperature. Sometimes also desired
superheated steam temperature control point (%
of MCR) is given.
Black liquor dry solids flow is the key design cri-

teria. It establishes the required size of the boiler.
With black liquor heating value this defines the
recovery boiler capacity (Rickard, 1993). Using el-
ementary analysis and dry solids one can calculate
the heat released in the furnace. With water and
steam values the MCR (Maximum continuous rat-
ing) steam flow is established. It should be noted
that when black liquor is sprayed to the furnace it
contains ash collected from the electrostatic pre-
cipitator and ash hoppers. Because ash free black
liquor is the input flow to the recovery plant, it is
usually chosen as the design base.
Figure 2-4, Design features of Babcock & Wilcox

recovery boiler (Babcock & Wilcox, 2001).

Figure 2-5, Screen at left, screenless boiler at right.
Single drum covered with deposits.

- Screenless superheater section is higher
All modern recovery boilers are of single drum and so has higher building volume and cost.
Negative issues with the screen are
type. The single drum has replaced the two drum
- There has been number of cases where
(or bi-drum) construction in all but the smallest,
fallen deposits have caused the screen to
low pressure boilers. The same trend but 20 years rupture. This has caused boiler explosions
earlier happened with coal fired boilers. and long shutdown times for repairs.
- Superheater surfaces are more affected with
Screen or screenless boiler radiation behind the screen than behind the
nose
One of the key design issues is whether to have - Screen captures heat. This reduces
a screen in the recovery boiler. A screen is a low superheating.
temperature heat surface that is put in front of the
superheater area. Almost always the screen is an Fear of boiler accidents caused by fallen depos-
evaporative surface. There are a few screens with its caused the boiler purchasers in US to avoid
saturated steam entering them, but the experience buying new boilers with screen. In Scandinavia
has not been too favorable. boilers with screen have been bought all the time.
Even in US some new boilers with screen have
Benefits of the screen are been bought.
- Screen stops part of the carryover from
furnace
- Screen blocks radiation from the furnace 2.2 EVOLUTION OF RECOVERY
and reduces superheater surface BOILER DESIGN
temperatures. A screen protects superheater
from corrosion There have been significant changes in kraft pulp-
- Screen itself is cold surface with very minor ing in recent years (Lindberg and Ryham, 1994,
corrosion Vakkilainen, 1994b, Ryham, 1992). Increased use
- Screen captures unburnt liquor partcles. of new modified cooking methods and oxygen
Less unburnt reaches superheater surfaces, delignification have increased the degree of
especially lower bends. This decreases organic residue recovery. Black liquor properties
superheater corrosion rates. have reflected these changes, Table 2-1.
- Screen evens out the flow somewhat. This
blocking effect is small if the screen is not Changes in investment costs, increases in scale,

Table 2-1, Development of black liquor properties (Vakkilainen, 2000).
Property Two drum Modern Current

1982 1992 2002
Liquor dry solids,
kg dry solids/ton pulp 1700 1680 1780
Sulphidity, Na2S/(Na2S+NaOH) 42 45 41
Black liquor HHV, MJ/kg dry solids 15.0 13.9 13.0
Liquor dry solids, % 64 72 80
Elemental analysis, % weight
C 36.4 34 31.6
H 3.75 3.5 3.4
N 0.1 0.1 0.1
Na 18 18.4 19.8
S 5.4 5.9 6
Cl 0.2 0.4 0.8
K 0.75 1.0 1.8
Cl/(Na2+K2), mol-% 0.35 0.68 1.24
K/(Na+K), mol-% 2.39 3.10 5.07
Net heat to furnace, kW/kg dry solids 13600 12250 11200
Combustion air* required, m3n/kg dry solids 4.1 3.7 3.4
Flue gas* produced, m n/kg dry solids
3
4.9 4.3 3.9
* At air ratio 1.2
demands placed on energy efficiency and en- The most important factor in the recovery boiler
vironmental requirements are the main factors has been high availability. The electricity genera-
directing development of the recovery boiler tion in recovery boiler process and steam cycle
(Vakkilainen, 1994). Steam generation increases can be increased by elevated main steam pressure
with increasing black liquor dry solids content. and temperature or by higher black liquor dry
For a rise in dry solids content from 65% to 80% solids (Raukola et al., 2002).
the main steam flow increases by about 7%. The
increase is more than 2% per each 5% increase in Increasing main steam outlet temperature in-
dry solids. Steam generation efficiency improves creases the available enthalpy drop in the turbine.
slightly more than steam generation itself. This is The normal recovery boiler main steam tempera-
mainly because the drier black liquor requires less ture 480°C is lower than the typical main steam
preheating. temperature of 540°C for the coal and oil fired
utility boilers. The main reason for choosing a
There are recovery boilers that burn liquor with lower steam temperature is to control superheater
solids concentration higher than 80%. Unreliable corrosion. Requirement for high availability and
liquor handling, the need for pressurized storage use of less expensive materials are often cited as
and high pressure steam demand in the concen- other important reasons.
trator have frequently prevented sustained opera-
tion at very high solids. The main reason for the Two drum recovery boiler
handling problems is the high viscosity of black
liquor associated with high solids contents. Black Most of the recovery boilers operating today are
liquor heat treatment (LHT) can be used to reduce of two drum design. Their main steam pressure is
viscosity at high solids (Kiiskilä et al., 1993). typically about 8.5 MPa and temperature 480 °C.
The maximum design solids handling capacity of
For pulp mills the significance of electricity gener- the two drum recovery boiler is about 1700 tds/d.
ation from the recovery boiler has been secondary.
Three level air and stationary firing are employed.
Two drum boiler represents one successful stage
in a long evolutionary path and signified a design
with which the sulfur emissions could be success-
fully minimized. Main steam temperature was
increased to 480 °C using this design.
Two drum recovery boilers are constructed with

water screen to protect superheaters from direct
furnace radiation, lower flue gas temperatures and
to decrease combustible material carry-over to su-
perheaters. The two drum boiler was the first type
to use vertical flow economizers, which replaced
horizontal economizers because of their improved
resistance to fouling.
Currently the two drum boilers start to get modi-

fied with single drum vertical boiler bank design
(Lovo et al., 2004).
Modern recovery boiler

The modern recovery boiler is of a single drum
design, with vertical steam generating bank and
wide spaced superheaters. The most marked
change around 1985 was the adoption of single
drum construction. The construction of the verti-
cal steam generating bank is similar to the vertical
economizer. Vertical boiler bank is easy to keep
clean. The spacing between superheater panels in- Figure 2-6, Two drum recovery boiler.
creased and leveled off at over 300 but under 400
mm. Wide spacing in superheaters helps to mini-
mize fouling. This arrangement, in combination
with sweetwater attemperators, ensures maximum
protection against corrosion. There have been nu-
merous improvements in recovery boiler materials
to limit corrosion (Ahlers, 1983, Hänninen, 1994,
Klarin, 1992, Nikkanen et al., 1989).
The effect of increasing dry solids concentration

has had a significant effect on the main operating
variables. The steam flow increases with increasing
black liquor dry solids content. Increasing closure
of the pulp mill means that less heat per unit of
black liquor dry solids will be available in the
furnace. The flue gas heat loss will decrease as the
flue gas flow diminishes. Increasing black liquor
dry solids is especially helpful since the recovery
boiler capacity is often limited by the flue gas flow.
A modern recovery boiler, Figure 2-7, consists of

heat transfer surfaces made of steel tube; furnace-
1, superheaters-2, boiler generating bank-3 and
economizers-4. The steam drum-5 design is of
single-drum type. The air and black liquor are
introduced through primary and secondary air
ports-6, liquor guns-7 and tertiary air ports-8. Figure 2-7, Modern recovery boiler.
The combustion residue, smelt exits through smelt
spouts-9 to the dissolving tank-10.
Figure 2-8, One of the most modern boilers, Gruvön (Wallén et al., 2002).
The nominal furnace loading has increased during both the generating bank and the economizers
the last ten years and will continue to increase are arranged on generous side spacing. Two drum
(McCann, 1991). Changes in air design have boiler bank pluggage is often caused by the too
increased furnace temperatures (Adams, 1994, tight spacing between the tubes.
Lankinen et al., 1991, MacCallum, 1992, Mac-
Callum and Blackwell, 1985). This has enabled an The spacing between superheater panels has
significant increase in hearth solids loading (HSL) increased. All superheaters are now wide spaced
with only a modest design increase in hearth heat to minimize fouling. This arrangement, in
release rate (HHRR). The average flue gas flow combination with sweetwater attemperators,
decreases as less water vapor is present. So the ver- ensures maximum protection against corrosion.
tical flue gas velocities can be reduced even with With wide spacing plugging of the superheaters
increasing temperatures in lower furnace. becomes less likely, the deposit cleaning is easier
and the sootblowing steam consumption is lower.
The most marked change has been the adoption Increased number of superheaters facilitates the
of single drum construction. This change has been control of superheater outlet steam temperature
partly affected by the more reliable water quality especially during start ups.
control. The advantages of a single drum boiler
compared to a bi drum are the improved safety The lower loops of hottest superheaters can be
and availability. Single drum boilers can be built to made of austenitic material, with better corrosion
higher pressures and bigger capacities. Savings can resistance. The steam velocity in the hottest super-
be achieved with decreased erection time. There is heater tubes is high, decreasing the tube surface
less tube joints in the single drum construction so temperature. Low tube surface temperatures are
drums with improved startup curves can be built. essential to prevent superheater corrosion. A high
steam side pressure loss over the hot superheaters
The construction of the vertical steam generating ensures uniform steam flow in tube elements.
bank is similar to the vertical economizer, which
based on experience is very easy to keep clean Current recovery boiler
(Tran, 1988). Vertical flue gas flow path improves
the cleanability with high dust loading (Vakkilain- Recovery boiler evolution is continuing strongly.
en and Niemitalo, 1994). To minimize the risk for Maximizing electricity generation is driving in-
plugging and maximize the efficiency of cleaning creases in main steam pressures and temperatures.

Figure 2-9, Effect of black liquor dry solids and main steam parameters to electricity generation from recov-
ery boiler.
If the main steam pressure is increased to 10.4 Increasing recovery boiler main steam tempera-
MPa and temperature 520 oC, then the electricity ture affects the corrosion of the superheaters.
generation from recovery boiler plant increases Designing for higher recovery boiler main steam
about 7 %. For design dry solids load of 4000 tds/d pressure increases the design pressure for all
this means an additional 7 MW of electricity. boiler parts. The recovery boiler lower furnace
wall temperatures increase with higher operating
The current recovery boiler, Figure 2-8, can be pressure. New better but more expensive lower
much larger than the previous ones. Boilers with furnace materials are used. The air flow per unit of
over 200 square meter bottom area have been black liquor burned in the recovery boiler furnace
bought. Largest recovery boilers are challenging decreases. Therefore the number of air ports will
circulating fluidized boilers for the tittle of largest decrease.
bio-fuel fired boiler.
State of the art and current
The superheater arrangement is designed for
optimum heat transfer with extra protection to trends
furnace radiation. Mill closure and decreased
Recovery boiler design changes slowly. There are
emissions mean higher chloride and potassium
however some features that boilers bought today
contents in black liquor. Almost all superheaters
have in common. State of the art recovery boiler
are placed behind the bullnose to minimize the
has the following features;
direct radiative heat transfer from the furnace.
- One drum boiler with 3-part superheater
Increasing superheating demand with increasing
and water screen (optional)
pressure decreases the need for boiler bank and
- Steam design data 9.2 MPa / 490 oC
water screen arrangement.
- Design black liquor dry solids 80 % with
pressurized heavy liquor storage tank
The higher main steam outlet temperature
- Liquor temperature control with flash tank,
requires more heat to be added in the superheat-
indirect liquor heaters for backup
ing section. Therefore the furnace outlet gas
- DNCG burning in the boiler
temperature has increased. The alternative is to
- Low emissions of TRS, SO2 and particulates
significantly increase superheating surface and
- Flue gas cleaning with ESP (no scrubbers)
decrease boiler bank inlet flue gas. If boiler bank
inlet gas temperature is reduced the average
The design changes occurring can be listed. Cur-
temperature difference between flue gas and steam
rent trends for recovery boilers are
is also decreased. This reduces heat transfer and
- Higher design pressure and temperature
substantially more superheating surface is needed.
due to increasing demands of power
This approach has been abandoned because of
generation
increased cost. With increasing dry solids content
- Use of utility boiler methods to increase
the furnace exit temperature can safely increase
steam generation
without fear of corrosion caused by carryover.
- Superheater materials of high-grade alloys
Figure 2-10, Main steam temperature as a function of recovery boiler capacity.
Figure 2-11, Main steam temperature as a function of recovery boiler main

steam pressure.
Figure 2-12, Net heating values of typical kraft liquors at various concentra-
tions.

Figure 2-13,Virgin black liquor dry solids as a function of purchase year of the recovery boiler.
- Further increase in black liquor solids Main steam temperature

towards 90 % by concentrators using
elevated steam pressure Main steam temperature of recovery boilers is
- Burning of biological effluent treatment shown in Figure 2-10 as a function of MCR capac-
sludge and bark press filtrate effluent ity of that boiler.
- CNCG burner (LVHC gases)
Dissolving tank vent gases returned to the boiler The average steam temperature increases with
Advanced air systems for NOx control size. Small boilers tend to have low pressure to
reduce specific cost. There are many boilers with
main steam parameters higher than 500 oC. Most
2.3 CHOOSING RECOVERY of them are in Japan.
BOILER MAIN PARAMETERS
Main steam pressure
As stated the recovery boiler main parameters are
often given by the customer to the boiler vendor. Main steam temperature of recovery boilers is
So when the recovery boiler purchase is consid- shown in Figure 2-11 with corresponding main
ered these main parameters must be chosen after steam pressure. Increase in main steam tempera-
careful study. ture is usually accompanied with increase in main
steam pressure, to keep exhaust steam wetness in
Higher black liquor dry solids generates more control.
steam. This has been seen as a significant path to
increased steam generation (Ryham, 1992). The Main steam pressure of above 80 bar but below
main steam parameters (pressure and tempera- 90 has been the most typical chosen value in the
ture) can be increased from traditional values. In- recent years. Main steam pressure has been lim-
crease of main steam values results in significantly ited to about 60 … 65 bars in Sweden to control
more power generation, Figure 2-9. The change in lower furnace corrosion (Bruno, 1995). In Japan
steam data is as important as about 5 % change in several boilers have been recently built with more
black liquor dry solids. than 100 bar main steam pressure (Akiyama et al.,
1988).
The trend in recent years has been definitely in
favor of increased temperatures and pressures. Black liquor dry solids
Newest Scandinavian lines, have chosen main
steam values in excess of 8.0 MPa and 480 oC As fired black liquor is a mixture of organics in-
(Vakkilainen and Holm, 2000, Wallén et al., 2000). organics and water. Typically the amount of water
The overall mill heat balance should be used to is expressed as mass ratio of dried black liquor
optimize the feed water and flue gas temperatures to unit of black liquor before drying. This ratio is
(Suutela and Fogelholm, 2000). called the black liquor dry solids.

Figure 2-14, Specific steam generation kgsteam/kgBLdry solids as function of
black liquor dry solids.
Figure 2-15, Specific steam generation kgsteam/kgBLdry solids as function

of black liquor higher heating value.
If the black liquor dry solids is below 20 % or we note that on average dry solids has increased.
water content in black liquor is above 80 % the net This is especially true for latest very large recovery
heating value of black liquor is negative, Figure boilers. Design dry solids for green field mills have
2-12. This means that all heat from combustion of been either 80 or 85 % dry solids. 80 % (or before
organics in black liquor is spent evaporating the that 75 %) dry solids has been in use in Asia and
water it contains. The higher the dry solids, the South America. 85 % (or before that 80 %) has
less water the black liquor contains and the hotter been in use in Scandinavia and Europe.
the adiabatic combustion temperature.
Steam generation
Black liquor dry solids has always been limited
by the ability of available evaporation technology Steam generation will depend on recovery boiler
to handle highly viscous liquor (Holmlund and design parameters. A rough estimate can be seen
Parviainen, 2000). Virgin black liquor dry solids from Figure 2-14 and Figure 2-15. About 3.5
of recovery boilers is shown in Figure 2-13 as a kgsteam/kgBL dry solids is often used as a base
function of purchase year of that boiler. value. Specific steam production can be used to
presize other components in recovery boiler plant.
When looking at the virgin black liquor dry solids
Both black liquor dry solids and higher heat- must be scheduled, payments should be given
ing value affect the steam generation. Also black only against work done and purchases made. Cost
liquor sulfidity and main steam values affect the of project capital can be significant.
steam generation efficiency. For accurate steam
generation one should always calculate the mass In addition permitting process should be early
and energy balances. started and closely monitored. A boiler with
environmental performance acceptable in the
future should be purchased. Timing of permitting
2.4 PROJECTING A RECOVERY process is crucial.
BOILER
Equipment selection criteria
The designer/owner/operator needs to consider
several factors when projecting a new recovery Boiler equipment should be selected based on
boiler. The main tasks are satisfactory expected life span and acceptable
- Determine the steam/power requirement maintenance costs. This decision is often based on
- Determine the fuels available satisfactory previous references.
- Determine possible locations and place
ments All equipment should have adequate permanent
- Compare different types of equipment from strength. The equipment must be placed so that it
different vendors is accessible for inspection and repair. Especially
- Anticipate future needs easiness of future large maintenance tasks, such as
- Permitting changing pump and fan motors, seals etc. needs to
be considered.
Selection process is influenced by applicable
emission requirement, project time schedule and When making the purchasing decision availability
reliability. of spare parts and service should be looked at.
Often also several iterations need to be done. CE norms regulate the working environment in
Improving thermal efficiency and so electricity Europe and OSHA in North America. For other
generation requires extra investments. Invest- countries often similar equipment is proposed and
ments can be partially or totally offset by savings bought. In summary safe, reliable operation is the
in operating cost. Annual costs of owning and op- main target with pleasing working conditions.
erating a recovery boiler plant are a sum of annual
charges for capital, fuel, maintenance, manpower,
ash and waste disposal (Advances in …, 1986).
Permitting
Best configuration thus depends on actual site The operation of a recovery boiler requires per-
conditions. mits.
Boiler purchasing Zoning requirements place boiler to area that is

reserved for industry and especially energy gen-
The main purchase is the pressure vessel with the eration. Recovery boiler sites should be chosen so
steel structure. Boiler purchasing requires also as not to cause harm to zones where people live.
choosing and buying the boiler auxiliary equip-
ment such as choosing equipment for black liquor Boiler equipment needs to be authorized for use.
handling, pumps, fans and environmental equip- Boiler must be made according to the required
ment. Typically a building contractor does the pressure vessel code. The electrical system and
building and foundations. Similarly a company instrumentation need to fill applicable laws. Often
specializing in erection does the pressure part special requirements for chemical handling must
erection often as a subcontractor to recovery be met.
boiler vendor. A score of smaller subcontractors
works on the site installing e.g. electrification, Operating permits are needed. Environmental
automation and insulation. Depending on the permits are used to regulate loads of effluents to
project type chosen these subcontractors work for environment. These include gaseous effluents,
the boiler vendor or the purchaser. effluents to the water (also heat) and dumping of
solid waste (mostly ash). Boiler operators are often
Because of large sums involved there has to be a codified.
tight financial control for the whole project. Loans

Figure 2 -16, Delivery time of a complete recovery boiler (EPS).
Delivery time
One of the most critical factors in a successful
project is the delivery time of a new recovery
boiler. In some projects 18 months have been
achieved, Figure 2-16 (Vakkilainen and Holm,
2000). Delivery time is somewhat affected by the
amount of work the boiler manufacturer has. The
biggest factor is however whether one can start
construction work right after contract and how
easy the installation site is. From the pressure part
delivery point of view the steam drum is the most
critical part.

3 Material and energy balance
Figure 3-1, Black liquor air demand as function of the higher heating value.
Calculation of material and energy balances is Typical biofuels consist of carbon, hydrogen,
fundamental. Dimensioning of recovery boiler nitrogen, sulfur and oxygen. Black liquor contains
heat transfer surfaces can not be done if mass and also a high amount of ash; sodium, potassium and
energy flows are unknown. Proper calculation chloride compounds. It is often impossible to de-
is important for mill energy and mass balances. fine the individual chemical compounds that form
These are needed to evaluate economics and run- the fuel. The stoichiometric air demand, I, can be
ning costs. Recovery boiler mass and energy bal- calculated assuming that the fuel can be divided
ances have been presented by Gullichsen (1968b), into three fractions
Clement et al. (1963), Adams and Frederick 1. Organic portion, which combusts fully
(1988) and Vakkilainen (2000b). The only stand- 2. Reactive inorganic portion, which reacts to
ard that covers recovery boiler mass and energy predefined end products
balance calculation is the so called Tappi standard 3. Inactive portion, which passes through
(Performance, 1996). combustion system unchanged
In calculating recovery boiler inorganic reactions

3.1 MATERIAL BALANCE it can be assumed that e.g. ratio of sodium sulfide
to potassium sulfide equals the ratio of sodium to
Control of combustion requires air flow that potassium in smelt. Similarly all chloride reacts
matches fuel flow. The amount of air required to to form sodium and potassium chloride. To
burn a mass unit of black liquor depends mainly facilitate mass balances it can be assumed that all
on water content and the heating value of the fuel. sulfur that is not escaping with flue gas is reacted
Black liquor higher heating value increase, Figure either to sulfide or sulfate. It can also be assumed
3-1, means more air for combustion is needed. (mainly for the lack of thermodynamical data)
Figures like Figure 3-1 can be used to estimate that only sodiumborate compounds exists.
the required air flow for dimensioning the fans.
More importantly they should be used to estimate The basis for material calculations is typically one
the possible range of air flows. A recovery boiler mass unit of black liquor. If mass balance values
should be able to handle not only the design black for one mass unit are known it is easy to multiply
liquor, but a range of typical black liquors. Often these values for the desired load.
design and performance estimation is done using
only the actual black liquor elemental composi- The boiler house forms a convenient system
tion.
Material and energy balance 3-1
Figure 3-2, Modern recovery boiler with balance boundaries shown.
Table 3-1, Example black liquor analysis for flows

in Figure 3-2. boundary. It has been pointed out that losses and
heat input caused by fans, blowers and pumps
Element Value should not affect the boiler efficiency. On the
other hand, forced circulation pumps, flue gas re-
C 32.5
circulation fans and other internal process devices
H 3.3 should be taken into account, because they play
N 0.09 a role when boiler efficiency between different
types of boilers is compared. Therefore the system
S 6.1
boundary for boiler efficiency loosely includes
Na 20.0 some but not all equipment in the boiler house.
K 3.0 The balance boundaries for a recovery boiler are
shown in Figure 3-2. This system boundary for
Cl 0.25
boiler efficiency measurement is then determined
B 0.5 in equivalent way to definition in European stand-
Inert (NPE) 0.1 ard EN 12952-15. All flow values are recorded
when they cross the system boundary.
Oxygen (by difference) 34.16
Table 3-2, Example dust for flows in Figure 3-2. Example
Calculate required air flow for boiler shown in

Element Flow, g/kgds
Figure 3-2. As fired liquor dry solids analysis as
Na 0.1158 mass percent is shown in Table 3-1. Note that the
K 0.0260 corresponding virgin dry solids analysis would
not include the ash that is recycled back to the
Cl 0.0035
boiler.
-CO3 0.0662
-SO4 0.1682 In addition assume that the reduction degree is
96 % expressed as the mole ratio of sulfide to sum
-S 0.0004 of sulfate and sulfide in smelt. Black liquor dry
-B 0.0000 solids is 85 %. Air ratio is 1.1625 (~3% O2 in flue

Table 3-3, Example ash for flows in Figure 3-2. molS/kgds mass, end
gS/kgds product
Element Flow, g/kgds
1,903 61,0 available sulfur
Na 30.47
0,333 10,7 in NCG
K 6.84
-0,000 -0,0 SO2
Cl 0.93
-0,002 -0,1 dust as -SO4
-CO3 17.41
-0,000 -0,0 dust as -S
-SO4 44.25
-0,461 -14,8 ash as -SO4
-S 0.10
-0,003 -0,1 ash as -S
-B 0.00
Sum 1,770 56,7 S in smelt
0,04*1,770=0,071 2,3 -SO4 in smelt
gas), SO2 emissions are 0.052 mg/kgds and HCl 0,96*1,770=1,699 54,5 -S in smelt
emissions are 0.010 mg/kgds. Sootblowing steam

consumption is 150 g/kgds. The ESP ash to recycle Chloride balance can be calculated by subtracting
and the dust to stack have the same analysis. Dust from chloride in black liquor losses as dust, ash
loss is 380 mg/kgds and the ash recycle is 100 and HCl. This is chloride in smelt.
g/kgds.
molCl/kgds mass, end
Input flow per 1 kg of as fired black liquor dry gCl/kgds product
solids is from liquor analysis
0,071 2,5 available
mass, mol/kgds end chloride
g/kgds product 0,000 -0,0 NaCl and KCl
in dust
Carbon 325 325/12,011 CO2, Na2CO3
=27,059 -0,026 -0,9 NaCl and KCl
in ash
Hydro- 33 33/2,016 H2O, HCl
-0,000 -0,0 HCl in flue gases
gen =16,369
Nitrogen 0,9 0,9/28,0134 N2 Sum 0,044 1,6 Cl in smelt
=0,032
Borate balance can similarly be calculated with
Oxygen 341,6 341,6/31,999 CO2, SO2,
Na2CO3, Na2SO4, known flow in black liquor and losses to arrive
= 10,675 K2SO4, K2CO3,
H2O, NaBO2,
at borate in smelt. This can further be divided to
Na3BO3 sodiumtriborate and sodium metaborate.
Chloride 2,5 2,5/35,453 NaCl, KCl, HCl
=0,071 molB/kgds mass, end
Potas- 30 30/78,204 K2S, K2SO4, gB/kgds product
K2CO3, KCl,
sium =0,384 0,462 5,0 available boron
Sulfur 61 61/32,060 SO2, K2S, Na2S,

-0,000 -0,0 loss in dust
Na2SO4
=1,903
-0,000 -0,0 loss in ash
Sodium 200 200/45,980 NaCl, Na2S,
=4,350 Na2SO4, Na2CO3, Sum 0,462 5,0 boron in smelt
NaBO2, Na3BO3
0,80*0,462=0,370 4,0 Na3BO3 in smelt
Boron 5 5/10,811 NaBO2, Na3BO3
=0,462 0,20*0,462=0,092 1,0 NaBO2 smelt
Water 1000* 175,5/18,015 H2O

(1/0,85 - 1) =9,796
Sodium and potassium balances need to be
=175,5 calculated together. In addition it is assumed that
borate forms components only with sodium and
Sulfur balance can be calculated by first adding other smelt compounds are formed in proportion
up all sulfur inputs and losses to get sulfur flow of total sodium and potassium in smelt. The actual
with smelt. Then this sulfur flow can be divided to distribution between e.g. Na2S and K2S is not
sulfide and sulfate. known, but it does not affect the mass balance.

Sodium balance to smelt can be calculated by Potassium balance for smelt compounds can
subtracting from incoming sodium the sodium in similarly be calculated by first subtracting from
dust and ash. potassium in smelt the amounts of potassium-
sulfide and potassium sulfate. Then the amount
molNa2/kgds mass, end of potassiumchloride is subtracted to arrive at the
gNa/kgds product amount of potassiumcarbonate.
4,350 200,0 available sodium
molK2/kgds mass, end
-0,003 -0,1 Na2SO4, Na2CO3
and NaCl in gK2/kgds product
dust
Sum 0,296 23,1 potassium in
-0,663 -30,5 Na2SO4, Na2CO3 smelt
and NaCl in ash
0,074*0,071=0,005 0,4 K2SO4 in smelt
Sum 3,684 169,4 sodium in smelt
0,074*1,699=0,126 9,9 K2S in smelt
Potassium balance to smelt can be calculated by 0,074*0,022=0,002 0,1 KCl in smelt
subtracting from incoming potassium the potas- 0,296- Σ =0,162 12,7 K2CO3 in smelt
sium in dust and ash.
Other inorganic 1 g/kgds is assumed to pass
molK2/kgds mass, end through to smelt unreacted.
gK/kgds product
Carbon balance can be calculated after the smelt
0,384 30,0 available potas-
sium compounds are known by subtracting from
-0,000 -0,0 K2SO4, K2CO3 carbon in black liquor the amount of carbon in
and KCl in dust carbonates. This results in the amount of carbon
-0,087 -6,8 K2SO4, K2CO3 that is burned to carbon dioxide.
and KCl in ash
molC/kgds mass, end
Sum 0,296 23,1 potassium in
gC/kgds product
smelt
27,059 325,0 available carbon

Mole percent of potassium to sum of sodium and -0,001 -0,0 Na2CO3 and
potassium in smelt is K2CO3 in dust
-0,290 -3,5 Na2CO3 and

100*0,296/(3,684+0,296) = 7,4 % K2CO3 in ash
-1,425-0,162= -19,1 Na2CO3 and

K2CO3 in smelt
Mole percent of sodium to sum of sodium and -1,587
potassium in smelt is Sum 25,181 302,4 CO2
100*3,684/(3,684+0,296) = 92,6 % Oxygen balance requires subtracting from oxygen

in black liquor oxygen in all outgoing compounds.
Sodium balance for smelt compounds can be cal-
culated by first subtracting from sodium in smelt
molO2/kgds mass, end
the amounts of sodiumsulfide and sodium sulfate.
gO2/kgds product
Then amounts of sodiumborates are subtracted.
Finally the amount of sodiumchloride is subtract- 10,675 341,6 available oxygen
ed to arrive at the amount of sodiumcarbonate. -25,180 -805,7 CO2
-0,005 -0,2 CO3 and SO4

in dust
molNa2/kgds mass, end
gNa/kgds product -1,357 -43,4 CO3 and SO4
in ash
Sum 3,684 169,4 sodium in smelt
-0,000 -0,0 SO2
0,926*0,071=0,066 3,0 Na2SO4 in smelt
-0,142 -4,5 Na2SO4 and
K2SO4 in smelt
0,926*1,699=1,573 72,3 Na2S in smelt
-2,381 -76,2 Na2CO3 and K2
0,555 25,5 Na3BO3 in smelt
CO3 in smelt
0,046 2,1 NaBO2 in smelt

-0,555 -17,8 Na3BO3 in smelt
0,926*0,022=0,020 0,9 NaCl in smelt

-0,092 -3,0 NaBO2 in smelt
3,684-Σ=1,425 65,5 Na2CO3 in smelt

-8,185 -261,9 H2O
Sum -27,223 -871,1

The humid air demand is then Black liquor flue gas production depends on the
1.1625*0,8711/0.22925= 4.4173 kg/kgds black liquor heating value, Figure 3-3. For the
simple analysis estimation based on predeter-
If we apply the carbon content versus higher heat- mined mass ratio of flue gas to air can be done.
ing value equation (see appendix) we get for HHV An accurate flue gas flow can be calculated from
13.3 MJ/kgds. From Figure 3-1 the air demand is simple mass balance if air flow, black liquor flow
from 3.2 to 4 m3n/kgds or from 3.8 to 4.8 kg/kgds. and smelt flow are known.
The smelt flow can be calculated if we add up all Example

mass flows to smelt.
Calculate flue gas flow for the previous example.
Smelt balance
Mass balance
mol/ mass, g/kgds end
kgds product mass, g/kgds product
0.126 0.126*110.26=13.9 K2S in smelt 1000 dry black liquor
1.573 1.573*78.04=122.8 Na2S in smelt 175.5 water with black liquor
0.066 0.066*142.04=9.3 Na2SO4 in smelt 4417.3 air
0.005 0.005*174.25=0.9 K2SO4 in smelt 150 sootblowing steam
0.041 0.041*58.443=2.4 NaCl in smelt 0.4 loss in dust
0.003 0.003*74.55=0.2 KCl in smelt 100.0 ash recycle
1.424 1.424*105.99=150.9 Na2CO3 in smelt -377.3 smelt
0.163 0.163*138.2=22.5 K2CO3 in smelt Sum 5265 flue gas
0,370 0,370*127,76=47.3 Na3BO3 in smelt

The flue gas flow is then 5.265 kg/kgds. For
0,092 0,092*68,5=6,1 NaBO2 in smelt
environmental reporting it is often useful to know
- 1 Other inor-
the dry flue gas flow. The dry flue gas flow can be
ganics
calculated by subtracting water with black liquor,
Sum 377.3
sootblowing steam, water with air and water from
hydrogen in black liquor from flue gas flow.
The smelt flow is then 0.377 kg/kgds.
Figure 3-3, Black liquor flue gas production as function of the higher heating value.

Mass balance
As one can see from the Figure 3-4, most of
mass, g/kgds product the heat input comes from the heat that can be
released in the combustion of black liquor. Other
5.265 wet flue gas flow
large sources are sensible heat in black liquor and
-175.5 water with black liquor air preheating.
-150 sootblowing steam
Similarly looking at the Figure 3-5, most of the
-0.0135*4.417 water with air
released heat is converted to steam. Reduction
-16.369*18.02 water from hydrogen in black
reactions take a lot of heat as does outgoing
liquor
molten smelt. It should be noted that instead of
Sum 4.645 dry flue gas
traditional higher heating value, the lower heating
value of black liquor is used.
When calculating real recovery boiler air and flue
gas flows leakage should be taken into account.
Example
Air flow through air fans is less than calculated as
some of the air is infiltration air through various
Calculate main steam flow and feedwater flow for
openings. Flue gas flow at stack can be signifi-
the previous example. The main steam values are
cantly higher that through economizers because of
9.1 MPa(a) and 490 oC. Feedwater process values
the same reason.
are 11.0 MPa(a) and 115 oC. Flue gas flow exits
at 155 oC. Air enters at 30 oC and is preheated as
average value to 108.8 oC. The blowdown is 0.10
3.2 ENERGY BALANCE kg/kgds at drum pressure 103.6 MPa(a). Black
To calculate energy balance for a recovery boiler liquor enters at 140 oC. Sootblowing is from out-
one must make a boundary around the recovery side with enthalpy of 3054.8 kJ/kg. Radiation and
boiler and then calculate all energy flows in and convection losses are 0.283 % of total heat input.
out of the boiler. It is easy once all mass flows have Unaccounted losses and margin are 0.3 % and 0.5
been determined. % of total heat input respectively.
The basis for energy balance calculation is so Lower heating value of as fired black liquor is the
called heat loss method. First the sum of all input higher heating value corrected with the hydrogen
energy flows is calculated. Then from this the and water in fuel losses.
energy losses are subtracted. The result is the net
heat available for steam generation.
Figure 3 4, Example heat inputs to recovery boiler.

Figure 3-5, Example heat outputs form recovery boiler.
tables. The two main losses are the loss in heat of

Input mass flow Enthalpy
kg/kgds kJ/kg kJ/kgds smelt and the loss in wet flue gas.
Higher heating 1 13000 13000.0
value of BL Losses
Hydrogen in black 0.038 -21806.3 -719.6 Wet flue gas 5.265 155*1.107 903.8
liquor
Reduction to Na2S 0.123 13099 1607.3
Lower heating value 1 12280.4
of dry BL liquor Reduction to K2S 0.0139 9629 134.0
Water in black 0.1755 -2440 -430.3 Reduction to SO2 0.00052 5531 0.3
liquor
Lower heating value 1.1755 11849.8

Autocausticization 0.0473 1535 72.6
of wet BL of Na3BO3
Loss in smelt 0.377 0.377*1350 509.0

sensible
The total heat input to the boiler is the sum of
heat in fuel value, heat in fuel sensible, heat in air Radiation & con- - 0.283*12846.5 36.4
vection
preheating and heat in sootblowing.
Unaccounted losses - 0.3*12846.5 38.5
Margin 0.5*12846.5 64.2
Input mass flow Enthalpy
kg/kgds kJ/kg kJ/kgds Total losses 3366.1
Heating value of 1.1755 11849.8 11849.8
black liquor
Net heat available is then 12846.5-3366.1= 9480.4
Black liquor 1.1755 140*2.64 434.8
sensible
Enthalpy of steam (9.1 MPa,490 oC) is 3360.7 kJ/kg.
Air 4.196 (30-0)*1.0336 130.1 Enthalpy of water (11.0 MPa,115 oC) is 490.3 kJ/kg.
Air preheat 4.196 (140-30) 341.8 Enthalpy of sat. water (103.6 MPa) is 1423.3 kJ/kg.
*1.0336 Enthalpy of sootbl. steam (155 oC) is 2792.0 kJ/kg.
Infiltration 0.221 (30-0)*1.0336 6.8
Steam mass flow x can then be calculate from the
Sootblowing 0.150 3054.8-2500.9 83.1
simple balance
Total heat input 12846.5
x*3360.7+0.100*1423.3-(x+0.100)*490.3+0.000*2792.0
To find out the heat available to steam generation = 9480.4
one needs first to calculate the losses. Necessary
energy uses are the heats needed for reduction and Steam mass flow, x is then 3.270 kg/kgds.
autocausticization reactions. They can be found
for individual compounds from thermodynamical
Feedwater mass flow is
3-2
3.270+0.100+0.000 = 3.370 kg/kgds.
If we had based the efficiency to the higher heat-

3.3 RADIATION AND ing value in fuel heat input as is customary in
North America.
CONVECTION HEAT LOSSES
Radiation and convection heat losses from recov- 3-3
ery boiler are hard to measure because there are
many flows in and out to the space surrounding
boiler proper inside the recovery boiler build- A recovery boiler needs to spend part of the heat
ing. Radiation and convection heat loss can be in fuel to reduce chemicals. It is thus more cus-
estimated through equation 3-1. The coefficient tomary especially in Europe to look at boiler effi-
0.0257 is between coal (0.022) and lignite (0.0315) ciencies based on lower heating value and subtract
boiler values. This equation with Tappi loss func- the heat used to process purposes. This is called
tion (Performance, 1996) is presented graphically the net heating method. Net heating value or ef-
in Figure 3-6. fective heating value of a boiler can be calculated
by subtracting from total heat input, the reduction
ΦRC = 0.0257*Φs0.7 3-1 heats and the autocausticizing heat (Adams and
Frederick, 1988).
where
ΦRC is the rad. and convection heat loss, MW 3-4
Φs is the useful heat output, MW
Losses calculated from equation 3-1, losses from Efficiencies based on lower heating value are ones
Tappi (Performance, 1996) and measured losses that a well built boiler in practice can achieve.
are compared in Table 3-4. Tappi loss function is So actually recovery boiler is doing a very good
not valid for boilers above 3500 tds/d. job. Out of available heat it can transfer about 90
% to steam. A recovery boiler is well in range of
Table 3-4, Calculated heat losses (equation 3-1 efficiencies of the modern biofuel boilers.
and Tappi) compared to measured heat losses from
Finnish recovery boilers (Ahtiala, 1997).
3.5 HIGH DRY SOLIDS BLACK
Boiler Capac- Calc. Tappi Meas-
ity heat loss loss ured LIQUOR
tds/d kJ/kgds kW kJ/kgds kW
Sunila SK11 1000 63.3 734 72.9 870 One of the major trends of current years has
been the increase of dry solids from evaporators.
Kaukopää, 3000 45.6 1582 44.6 1590
SK6 The data presented is for study of a 3000 tds/d
Rauma, SK3 3200 44.7 1655 43.3 1600 recovery boiler. Even though this size of recovery
boiler is chosen as example, the results should be
Kaukas, SK3 3350 44.1 1709 42.6 2000
applicable to most of the current recovery boilers
(Lankinen et al., 1991, Vakkilainen and Niemitalo,
Most of the heat losses are from pipes (50 %).
Furnace accounts for roughly a part (22%). An-
other large portion of heat losses is from electrical
equipment (18%). Ducts and tanks account for
only a minor portion (10%) (Ahtiala, 1997).
3.4 STEAM GENERATION

EFFICIENCY
Steam generation efficiency depends on how it
is determined. Most general efficiencies are ones
based on total heat input and fuel heat input. Us-
ing the values in previous example we get
Figure 3-6, Radiation and convection losses equa-
tion 3-1, Tappi.

Table 3-5, 3000 tds/d recovery boiler main pa- 1994, Vakkilainen et al., 1998b).
rameters.
Max. continuous firing rate ton ds/day 3000 The design parameters for the studied recovery
boiler are shown in Table 3-5. Material balance
Steam pressure bar(g) 90.0
data from low dry solids value of 60 % to high
Steam temperature °C 490 value of 90 % is presented in Table 3-6. In these
Feedwater pressure bar(g) 109 calculations the black liquor elemental composi-
tion is assumed to remain constant. The typical
Feedwater temperature °C 115
change in existing recovery boilers has been
Primary air percentage % 35.0 increase of liquor solids content from 65 to about
Primary air temperature °C 120 80 % dry solids. For same cases the correspond-
Secondary air percentage % 50.0 ing energy balance values and steam flows are
presented in Table 3-7.
Secondary air temperature °C 120
Tertiary air percentage % 15.0 In the material balance the combustion air flow
Tertiary air temperature °C 50 and black liquor dry solids flow remain constant.
Flue gas temperature after eco °C 150
The flue gas flow decreases as less water enters the
furnace with increasing black liquor dry solids
Black liquor analysis content. The flow of smelt remains constant as
C % weight 34.6 reduction degree has been assumed to remain at
Na % weight 19.9 95 %.
S % weight 4.8
Even though the steam flow increases with
O2 % weight 35.9 increasing black liquor dry solids content, the
H2 % weight 3.5 sootblowing can remain constant. The blowdown
K % weight 1.0 has been increased with increasing steam flow as it
depends mainly on the quality of incoming feed-
Cl % weight 0,1 water. The total heat input falls slightly with in-
Others % weight 0.2 creasing dry solids, as the heat in the black liquor
Higher heating value MJ/kg ds 14.0 preheat will decrease. Because of this decrease, less
heat will be available in the furnace, the reduction
Dry solids content as fired % weight 65-90
and smelt losses being constant, Figure 3-7.
Liquor temperature as fired °C 140.0
Chemical loss to stack g/kgds 0.2 Flue gas heat loss will decrease as the flue gas mass
Reduction % 97 flow decreases. The heat loss due to water vapor
in flue gas will decrease as there is less water to be
Sootblowing steam flow kg/s 5.0
evaporated. The heat available to steam produc-
Balance reference temperature °C 0 tion will increase as the heat losses decrease more
Figure 3-7, Effect of black liquor dry solids at amount of flue gas, boiler efficiency, steam flow and heat input
to furnace.

Table 3-6, Material balance for the example boiler.
Liquor dry solids, % 65 70 75 80 85 90
Liquor flow, kg/s 53.4 49.6 46.3 43.4 40.8 38.5
Air flow, kg/s 161.9 161.9 161.9 161.9 161.9 161.9
Sootblowing, kg/s 5.0 5.0 5.0 5.0 5.0 5.0
Ash and dust, kg/s -3.5 -3.5 -3.5 -3.5 -3.5 -3.5
Smelt flow, kg/s -12.8 -12.8 -12.8 -12.8 -12.8 -12.8
Flue gas flow, kg/s -204.0 -200.2 -196.9 -194.0 -191.4 -189.2
Table 3-7, Energy balance for the example boiler.

Black liquor LHV, kJ/kgds 11923 12191 12423 12627 12806 12966
Sensible heat in BL, kJ/kgds 569 528 493 462 435 411
Air preheating, kJ/kgds 509 509 509 509 509 509
Sootblowing, kJ/kgds 80 80 80 80 80 80
Total heat available, kJ/kgds 13080 13307 13505 13677 13829 13965
Heat in smelt, kJ/kgds -545 -545 -545 -545 -545 -545
Reduction, kJ/kgds -1024 -1024 -1024 -1024 -1024 -1024
Heat in wet FG, kJ/kgds -976 -958 -942 -928 -916 -905
Unacc.etc. losses, kJ/kgds -142 -144 -146 -148 -150 -151
Total losses, kJ/kgds -2687 -2671 -2657 -2645 -2635 -2625
Heat for steam, kJ/kgds 10393 10636 10847 11032 11194 11339
Steam flow, kg/kgds 3.588 3.673 3.7465 3.811 3.868 3.918
Steam flow, kg/s 124.6 127.5 130.1 132.3 134.3 136.0
Efficiency, % 87.3 87.6 87.9 88.1 88.3 88.5
than the total heat input. ids the amount of flue gas generated falls by 7 %.
The flue gas passages can be made smaller as the
Steam generation increases with increasing dry flue gas flow decreases. At the same time, a smaller
solids. For a rise in dry solids content from 65 % flue gas fan will be required.
to 80 % the main steam flow increases by 7 %. The
increase is typically more than 2 % per each 5 % Liquor heat treatment and storage in high tem-
raise in dry solids. The superheater pressure loss perature cause some of the combustible material
also increases as main steam generation is raised. to be released as non-condensable gases. The main
To keep steam pressure losses in superheaters in effect is a decrease in black liquor sulphur content.
reasonable level larger tubes or more parallel tubes Release of NCG affects recovery boiler perform-
should be used. ance in the same way as the increase in black
liquor dry solids. The efficiency of steam genera-
Steam generation efficiency improves more than tion goes up and the amount of flue gas decreases.
steam generation itself. This is because preheat- The lower sulphur load to the furnace results in a
ing the higher dry solids black liquor requires less sharp decrease in the heat required for reduction
heat. For a rise in dry solids content from 65 % of sulfur compounds. If CNCGs are burned in the
to 80 %, the steam generating efficiency improves recovery boiler there is no net effect.
from about 65 % to close to 70 %. The decrease in
flue gas flow causes the greatest increase in steam
generating efficiency. For the same rise of dry sol-

4 Combustion of black liquor
Figure 4-1, Characteristic swelling behavior of black liquor during combustion.
The black liquor is composed of a large number of make combustion of black liquor difficult.
organic and inorganic compounds. The amount
and the composition of the black liquor depend Black liquor combustion occurs either as a droplet
on the wood species, the cooking method and the sprayed to the furnace from a liquor gun or in
pulping process. Properties and processes for e.g. the char bed at the bottom of the recovery boiler
sulfite liquor combustion can be different (Hupa furnace (Hupa and Solin, 1985). Black liquor is
et al., 1994). This text considers only kraft liquor sprayed into the furnace through a number of
combustion. The organic matter of black liquor is liquor guns. In many combustion applications the
combusted in the recovery boiler furnace, while aim is to produce very small droplets to maximize
part of the inorganic matter is recovered as smelt. combustion rates and temperature. Black liquor is
Black liquor has one of the lowest heating values not finely atomized as it enters the furnace. Rather
of industrial fuels. This is because of the large black liquor is sprayed as coarse droplets. The av-
inorganic portion of the black liquor. High water erage droplet diameter is about 2 ... 3 mm, so that
content, low heating value and huge ash content unburned material can reach char bed. The black
Table 4-1, Main stages in black liquor combustion in furnace for a 2 mm droplet
Stage Characterized by Time scale in furnace
Drying Water evaporation, constant diameter 0.1 … 0.2 s
after initial swelling
Devolatilization Appearance of flame, ignition, swelling 0.2 … 0.3 s
of the droplet, release of volatiles
Char burning Disappearance of flame, reduction reac- 0.5 … 1 s
tions, decreasing diameter
Smelt Constant or increasing diameter, Long
reoxidation
Combustion of black liquor 4-1

of moisture remains (Frederick and Hupa, 1993).
The drying rate for pine, birch and sodium sulfite
liquors is constant for black liquor droplets with
various dry solids contents at 700 oC and 800 oC.
Initially the black liquor is thought to consist of

water and pyrolysable material that remains after
drying. Black liquor mass can then be divided into
mo = mp + mw 4-1
where
mo is the initial droplet mass, kg
mw is the mass of water in droplet, kg
Figure 4-2, Example of drying black liquor drop- mp is the mass of pyrolysable material in
let, laboratory conditions (Åbo Akademi). droplet, kg
liquor droplet combustion is typically divided into The dry solids at start of combustion is
four stages, Table 4-1. This division is similar than
for other biofuels. As most of the combustible xd = mp/(mp+mw) 4-2
material in black liquor originates from wood
(Söderhjelm, 1994), this can be expected. where
xd is the droplet dry solids, -
During combustion, Figure 4-1, black liquor
swells (Hupa et al., 1994). No other industrial fuel The black liquor droplet temperature correspond
swells as much as black liquor during combustion. to boiling point rise ( BPRmax = 650 K)
The swelling behavior is caused by high volatiles
yields and suitable surface properties (Milanova, Td = 100+BPRmax*xd2.74 4-3
1988, Miller et al., 1986, Noopila et al., 1991).
Order of different combustion steps is shown as The heat to droplet is (cp = 2.5 kJ/kg, l = 2450
vertical lines in Figure 4-1. In reality the combus- kJ/kg)
tion stages overlap somewhat (Verrill and Wessel,
1995, Järvinen et al., 2000) at least for the larger 4-4
droplets. A reason for this is the different combus-
tion speeds at different parts of droplets. While
where
black liquor can already be dry and undergoing
volatiles release at surface, drying is not complete
at droplet centre. is the loss as water vapor
Heat to droplet is
4.1 DRYING
Qd = Qc + Qr 4-5
Drying is characterized by evaporation of water
from the black liquor droplet. It is often ex-
The liquor droplet diameter is constant Dd.
perimentally defined by absence of combustion
(visible flame), Figure 4-2. Evaporation of water
Dd = 1.54*Do 4-6
requires heat. Drying of black liquor droplet
proceeds as fast as the heat is transferred to the
The droplet area is
droplet. Even in the furnace temperatures drying
is limited by the heat flux to the droplet.
Ad = πDd2 4-7
The black liquor diameter increases to 1.3 - 1.6
The convective heat flux is (hc ~ 10 W/m2K)
times the original diameter during the first couple
of milliseconds after insertion into the furnace.
As water is evaporated, the density decreases, 4-8
but the diameter stays constant. Swelling restarts
For modelling of convective heat transfer a model
with the onset of the volatiles release. The black
of Ranz and Marshal (1986) can be used.
liquor droplet is not completely dry at the onset
of volatiles release. Typically about five percent

4-9
where
Dd particle diameter, m
h convective heat transfer coefficient, W/m2K
Red Reynolds number based on droplet
diameter and relative speed, -
Pr Prandtl number based on gas phase, -
The radiative heat flux (ε = 0.8)
4-10 Figure 4-3, Example of black liquor droplet at the

start of devolatilization, laboratory conditions (Åbo
The basic radiation heat transfer equation for Akademi).
droplet in gray gas is
Φr = εσAd(Tg4 - Td4) 4-11
where
Ad droplet surface area, m2
Tg gas temperature, K
Td droplet temperature, K
This type of equation is applicable to the real proc-

esses in the recovery boiler furnace.
In most of the experimental studies the main

radiative component is the hot furnace wall so
different type of radiative heat transfer equation Figure 4-4, Example of black liquor droplet at the
needs to be used. In the Åbo Akademi experi- end of devolatilization, laboratory conditions (Åbo
ments where the view port was small the resulting Akademi).
equation was (Frederick et al., 1989)
of volatile gases from the black liquor droplet and
Φr = εσ(Tw4 - Td4) 4-12 appearance of visible flame, Figure 4-3. The last
one is the most typical criteria for determining
In the IPC experiments where the view port was the length of devolatilization time in experimental
large the resulting equation was (Frederick et al., droplet combustion studies.
1989)
During devolatilization usually the gas release is
Φr = 1/2εσAd(Tw4 - Td4) + 1/3εσAd(Tg4 - Td4) 4-13 large enough so no oxygen can contact the droplet
surface. Therefore the conditions at droplet
resemble those of pyrolysis or heating in inert
4.2 DEVOLATILIZATION athmosphere. The term Pyrolysis is often used
incorrectly to mean devolatilization. So the release
As black liquor dries and temperature increases of volatile fraction occurs whether or not there is
reactions with lowest activation energies start oxygen present.
taking place. For organic fuels, release of low mo-
lecular weight component gases such as methane, Devolatilization of black liquor tends to be a fast
carbon dioxide, hydrogen and hydrogen sulfide proces and depends essentially on the heat trans-
starts. Hydrogen, carbon monoxide, carbon fer to the liquor. Devolatilization tends to occur
dioxide, light hydrocarbons, tar and light sulfur- in an outer shell of expanding thickness for the
containing gases have been reported to be main larger droplets, while a core of colder, unpyrolyzed
product gases during devolatilization (Jing and material remains within (Järvinen et al., 2000,
Iisa, 2001). Devolatilization is characterized by the Sricharoenchaikul et al., 2001).
increase of black liquor droplet volume, release

Figure 4-5, Relationships between the main organic constituents and the different combustion stages of a
kraft black liquor droplet (Alén, 2000).
Swelling the effect of wood species and cooking time. They

concluded that the swelling both for softwood and
During the devolatilization, the black liquor drop- hardwood is proportional to the ratio of lignin to
lets swell considerably. The swelling is continuous aliphatic acids. Typically a longer cooking time
from the onset of ignition until the devolatiliza- increases swelling (Alén, 2004). Liquor from soda
tion is complete, Figure 4-4. The swollen devolatil- cooks and NSSC cooks seem generally to swell
ized particles exhibit extensive macroporosity and less than kraft liquors. Adding of sodiumsulfate or
often a hollow central core is found. sodiumcarbonate to virgin black liquor decreases
swelling.
Two conditions have to be met for swelling; there
must be gas generation and the droplet must have The rate of combustion reactions is increased with
plastic surface properties. The maximum swelling higher swelling, Figure 4-6 (Vakkilainen et al.,
changes from one type of liquor to another. The 1998). This can be explained by the larger avail-
maximum swollen volume for kraft black liquors able surface area for highly swelling black liquors.
varies from less than 10 to 50 cm3/g. Backman et al. (1996) found that the swelling and
organic combustion time correlate for typical mill
During the devolatilization, the black liquor drop- liquors. Swelling is dependent on liquor specific
lets swell considerably. The swelling is continuous properties (Alén, 1997). Swelling decreases with
from the onset of ignition until the devolatiliza- increasing furnace temperature (Hupa et al.,
tion is complete. Noopila et al. (1991) have studied 1994).
Figure 4-6, Swelling versus organic combustion time for single droplet studies at 800 oC (=sum of devolatili-
zation and char burning times).

Figure 4-7, Sulfur species during pyrolysis of black liquor (McKeough et al., 1995).
Sulfur release Carbon conversion

During devolatilization a large number of reac- Devolatilization occurs very rapidly when black
tions take place. The main form of sulfur release is liquor solids are heated to temperatures sub-
dimethyl sulphide and methyl mercaptane, Figure stantially above 200 °C (Sricharoenchaikul et
4-7. Hydrogen sulfide, H2S is formed rapidly with al., 2001). During devolatilization a significant
decomposition reactions after gases are released portion of carbon in char is consumed, Figure 4-8.
from the droplet.Formation of dimethyl disulfide Increasing temperature results in higher pyrolysis
is small. yield. Volatiles release seems to decrease when
swelling increases (Whitty et al., 1997). This is
explained by increased heating rate with increased
Sodium release swelling.
There is some experimental evidence that during
devolatilization fragmentation though ejection As can be seen total of almost half of the char is
of small particles occurs (Frederick et al., 1995). reacted in normal recovery boiler lower furnace
This process is partly responsible for sodium loss temperatures. This includes the about 6 to 7 per-
from combusting black liquor droplets. In addi- cent that is involved in reduction reactions.
tion sodium associated with activated carboxylate
and phenolate sites can volatilize (Frederick et al., Kulas and Clay (1988) derived, using statistical
2004). methods, an equation for mass loss as a function
of initial droplet diameter and oxygen partial pres-
sure.
4-14
Frederick et al. (1989) models the black liquor

pyrolysis
Qp = mo[cp(Td-To)+l(1-xo/xd)]+Q∆p 4-15
Qp = Qc + Qr 4-16
Horton (1991) models the black liquor pyrolysis

Figure 4-8, Effect of temperature on char conver- as follows
sion (Järvinen et al., 2000).

inorganic ash. In laboratory tests char combus-
tion starts when the visible flame is extinguished,
Figure 4-9. In practice char combustion and
devolatilization overlap considerably. Often a
term organic combustion time, that is the sum
of devolatilization and char combustion times, is
used to characterize combustion.
Almost all the inorganic matter except sulfur

remains with majority of carbon in char. Black
liquor char contains carbon, sodium carbonate,
sodium sulfate and sodium sulfide. A representa-
tive composition of char is shown in Table 4-2.
Figure 4-9, Example of black liquor droplet dur- We can see that about two thirds of carbon and
ing char combustion, laboratory conditions (Åbo less than a fourth of hydrogen is present but
Akademi). no organic oxygen is left. The reduction rate
Table 4-2, Composition of kraft char (Grace, expressed as ratio of sodium sulfide to sodium
1990). sulfate is about 50 %. Char continues to burn with
the particle temperature increasing from outside
Component moles/ weight, % to inside. The inorganic residue eventually forms
mole Na2
molten smelt.
Sodium sulfide Na2S 1/6 9.0
Sodium sulfate Na2SO4 1/6 16.4 During the char combustion reduction reactions
Sodium carbon- Na2CO3 2/3 49.0 take place (Grace, 1985). The carbon has a major
ate role in the reduction reaction. Sodium sulfate
Na2SO4 reacts with carbon to form sodium sulfide
Carbon C 3 24.9
Na2S. So while the carbon in the char bed burns it
Hydrogen H 1 0.7 causes the reduction of sodium (Grace, 2004).
D(t) = Dd + (Dp-Dd)(Qp(t)/Qp)Nv 4-17
Na2S + 2O2 → Na2SO4 4-21
Qp = moxo/xp[cp(Tp-Td)+Hv(1-xd)] 4-18
Na2SO4 + 2C → Na2S + 2CO2 4-22
Qp = Qc + Qr 4-19
Na2SO4 + 4C → Na2S + 2CO 4-23
Frederick (1991) models the carbon remaining in
the droplet after the devolatilization as The rate of reduction reaction depends on the
char carbon content. This rate was found to be
xc = A-BT* 4-20
4-24
where
A = 0.513 for kraft liquors
B = 0.384 for kraft liquors The constants measured for kraft char were
and KRed = 1310+410, 1/s
T* = T[K]/1000 B = 0.022+0.008, kmol/m3
Ea = 122, kJ/kmol
Most of the char conversion cn be attributed to
formation of volatiles. Sricharoenchaikul et al. The reduction rate is dependent on temperature.
(2002) found in their LFR-experiments that tar In modern kraft recovery boilers high reduction
formation accounted to less than 5 % of cabon on efficiencies are typical. From thermodynamical
all conditions studied. equilibrium we note that there should be very
little of sodium oxides and thiosulphite. The rate
of reduction process is very slow when reduction
4.3 CHAR COMBUSTION degree is over 95%. Higher reduction degrees
require significant residence time for the smelt.
Char combustion of black liquor starts as the
volatiles release is finished. Often the combustible If there is enough oxygen to reach the char
material remaining after volatiles release is termed surface, the carbon in the char reacts with the oxy-
fixed carbon. Fixed carbon does not include
gen. If there is a deficiency of oxygen the char is
gasified with carbon dioxide CO2 and water vapor
H2O. Both carbon dioxide and water vapor react
with char to form carbon monoxide CO.
Cchar + CO2 → 2CO 4-25
Cchar + H2O → CO + H2 4-26
The CO is further oxidized to CO2 higher in the

furnace when it reacts with oxygen.
Figure 4-10, Example of black liquor droplet dur- 4.4 SMELT REACTIONS
ing smelt reactions, laboratory conditions (Åbo
Akademi). As the char combustion is finished the inorganic
residue remains. The black liquor droplet has
first enlarged and then shrunk to a liquid droplet,
Figure 4-10. If oxygen contacts smelt, the sulfide
in smelt is reoxidized to sodium sulfate Na2SO4.
In recovery boiler it is important to have enough
reacting material on top of smelt to avoid smelt
reoxidation.
4.5 EXPERIMENTAL
PROCEDURES TO LOOK AT
BLACK LIQUOR COMBUSTION
Laboratory-scale devices are used for studying the
combustion properties of the liquors: Single-drop-
let muffle furnace is maybe the most typical of
these, Figure 4-11.
The duration of the pyrolysis and char burning

times and swelling during combustion is recorded
Figure 4-11, Single droplet furnace for combustion with a video camera. Experiments are performed
experiments. in the single-droplet muffle furnace typically in
air at 700 and 800 oC. The sample size is usually
of order of 10 mg wet black liquor. The technique
has been used by Hupa et al. (1994) and Whitty et
al. (1997).
Another typical device is the single droplet tube

reactor, Figure 4-12. The main benefit of the tube
furnace to single droplet furnace is that gas analy-
sis of combustion products can be made. In the
Åbo Akademi one can follow the burning process
of individual droplets with on-line gas analysis
for CO2, CO, SO2 and NO. The experiments are
typically performed on 40 mg samples at 900°C in
different O2 compositions.
The single droplet tube reactors have been de-

scribed in more detail by Whitty (Whitty et al.,
Figure 4-12, Single droplet tube reactor. 1995) and Forssén (Forssén et al., 1997).

Figure 4-13, Recovery boiler lower furnace.
4.6 COMBUSTION OF BLACK Black liquor drying occurs close to the liquor gun
at the right wall. Droplet velocity is initially high,
LIQUOR DROPLET IN THE FUR- about 10 m/s. When volatiles release starts the
NACE black liquor droplets swell. Increased drag then
A typical lower part of recovery furnace is shown slows them down. At the same time droplets curve
in Figure 4-13. It consists of three air levels; upward because of the drag from flue gases. Dur-
primary, secondary and tertiary levels. The ing char burning the horizontal velocity is low.
furnace bottom is covered with a char bed. The As char burning is completed the droplet density
black liquor is sprayed from black liquor guns. increases. Then it starts falling down until it hits
Combustion of black liquor in the furnace can be the char bed.
simulated. Results of one such simulation can be
seen in Figure 4-14. In it black liquor droplets of Drops with higher initial dry solids start swelling
1.5 mm diameter, but of different dry solids con- faster and tend to burn higher in the furnace (Fre-
tents are sprayed into the furnace. Each droplet is derick and Hupa, 1992). All droplets in figure 4-14
drawn about 0.1 s time intervals. burn at about liquor gun level. In practice this is
too high and larger size droplets would be used.
4.7 COMBUSTION PROPERTIES

OF HIGH DRY SOLIDS BLACK
LIQUOR
The combustion properties of high dry solids
black liquors have been studied by Åbo Akademi
and VTT (Frederick and Hupa, 1993, McKeough
et al., 1995). The main interest has been the release
of sodium, sulphur and chloride and the combus-
tion properties of the drop.
Figure 4-14, Effect of increased dry solids to The black liquor droplet is a poor heat conductor.
combustion of 1.5 mm black liquor droplets fired to The burning process is controlled by the slow rise
furnace. of the temperature inside the droplet (Järvinen
Figure 4-15, Swelling as a function of black liquor Figure 4-16, LHT plant.
dry solids.
et al., 2000). The increase of the dry solids affects has been employed to fire and store black liquor
only slightly the black liquor combustion when up to 90 % dry solids in atmospheric pressure in
the dry solids of black liquor is raised. In Figure order to maximize the benefits of high dry solids
4-15, the small decrease of swelling as dry solids is in black liquor combustion (Ryham, 1990, Ryham
increased is shown. and Nikkanen, 1992).
The results of studies show that the increase in The main benefit of the liquor heat treatment
dry solids shortens the drying time, but it does (LHT) is viscosity reduction which allows the
not have an effect to the time for volatiles release evaporation up to 90 % dry solids and keeps the
nor to the char burning time. The increase of dry black liquor in a pumpable form even at atmos-
solids does not significantly affect the swelling of pheric pressure. This also allows storage of the
the black liquor, Table 4-3 strong liquor in atmospheric tanks. Viscosity
reduction is especially beneficial, if the raw mate-
Table 4-3, Effect of dry solids on combustion rial and cooking method at the mill is such that
properties the handling of strong liquor is a problem. One
Dry solids Pyrolysis Pyrolysis Swelling, Char, possible way to combine the LHT-process into an
% time, s m-loss, % cm3/g g/cm3 existing evaporation plant is shown in Figure 4-16.
70.1 15 36 18.1 0.0486
70.1 30 51 15.9 0.0418 The viscosity of the black liquor is determined by
its composition. The polysaccharides like xylan are
84.6 15 32 19.5 0.0502 dissolved during the cook as long chain molecules
84.6 30 48 14.1 0.0514 which increase the original viscosity level of black
93.2 15 42 12.5 0.0712 liquor. The relationship is clearly shown when we
compare the behavior of black liquors originat-
93.2 30 48 10.7 0.0689
ing from different mills, Figure 4-17. In practice
the effect of the lower viscosity and the reduced
The tests with a very high dry solids have been
polysaccharide content shows up as decreased
performed only with the Northern softwood and
energy consumption in pumping and less severe
hardwood liquors, so further research is needed
with black liquors generated from grass and tropi-
cal woods.
Effects of liquor heat treatment

on black liquor combustion
The heat treatment of black liquor can be defined
as a thermal treatment method where residual
alkali reacts with dissolved polysaccharides and
lignin. These reactions are very slow in typical
evaporator temperatures so the black liquor is
heated up to 180 ... 190 oC. The effect is destruc- Figure 4-17, Viscosity of black liquors from three
tion of high molecular weight compounds and mills using liquor heat treatment; Mills A and B
reduction of black liquor viscosity. The process before heat treatment, Mill C after heat treatment.

Figure 4-18, CO2 profiles for liquors 3 and 4 at about 80 % dry solids for combustion at 3% O2.
scaling in strong liquor pipes. increase the black liquor swelling during combus-
tion (Backman et al., 1996). The LHT can also
Simultaneously with the reaction of the polysac- affect beneficially to the NOx emissions (Forssén
charides and alkali there happens also another et al., 1997). The amount of released NO convert-
reaction. The sulfide in the liquor reacts with the ible nitrogen and nitrogen remaining in the liquor
lignin generating dimethylsulfide and methyl is reduced with LHT.
mercaptans. This gas is released from the liquor
when the pressure is decreased after the treatment. Burning rate
However, the amount of DMS/MM formed can be
controlled by adjusting the treatment conditions. For individual liquors an increased burning rate
The gas can be introduced into the malodorous with LHT can be observed. The increased rate is
gas treatment system and combusted in a lime more evident if we study the CO2 content profile
kiln or in an existing NCG combustor. in the flue gas from combustion of individual
droplets with on-line gas analysis. Figure 4-18
The most of the benefits of using LHT are related shows CO2 profiles for a mill liquor with and
to the boiler. Increasing the dry solids increases without heat treatment. In the figure the disap-
the boiler capacity and also efficiency. The benefits pearance of the CO2 curve shows the end of char
which are related to high dry solids combustion burning. The char combustion for the heat treated
can be summarised as follows; reduced sulfur liquor was approximately 30% faster than for the
dioxide emission and higher temperature at the untreated sample.
bottom part of the boiler. A well-known benefit
of raising the dry solid level of black liquor is the
reduced sulfur dioxide emission. The higher dry
Swelling
solids rises the temperature at the bottom part of Figure 4-19 shows the swelling during droplet
the boiler. This gives higher sodium release from combustion for all about 19 liquor samples (Vak-
black liquor droplets. The Na can form Na2SO4 kilainen et al., 1998). First line in each series is
with the SO2 present in the upper furnace. The swelling of a liquor sample without liquor heat
achievable level of SO2-emission is typically a treatment and subsequent 1 – 3 lines are for
function of two factors, the boiler load and the progressively more heat treatment. The swelling
black liquor sulfidity. increased significantly for liquor 1->2 and clearly
for liquors 3->6, 10->13, 14->16 and 17->19. Only
LHT treatment splits long lignin chains. It has for liquor 7->9 the swelling has decreased.
been observed (Backman et al., 1996, Alèn, 1997)
that increasing the amount of shorter (Mw 1500 Liquor heat treatment increases swelling and
… 3000) lignin chains increases swelling. Heat decreases pyrolysis time and mass loss during
treatment of black liquor has also been found to
Figure 4-19, Swelling of black liquor.
pyrolysis. The increased burning rate is probably Milanova, 1998 states that two conditions have to
mostly a result of the increased swelling: Increased be met for the char particle to swell. There must be
swelling increases the effective external surface gas generation inside the droplet and the droplet
area during the combustion process. This, again, must have plastic surface properties. The liquor
gives higher heat transfer from the hot environ- heat treatment affects the polysaccharide content
ment to the droplet and a higher heating rate and so also the surface properties of black liquor
during pyrolysis. Also, larger external surface droplet.
increases the amount of oxygen diffused for char
burning. The plot of polysaccharide versus swelling, Figure
4-20, shows decreased swelling with increased
During liquor heat treatment volatile organic polysaccharide content. It is suspected that the
gases are released. It would seem natural to change in liquor swelling can be attributed to
assume that the observed decrease in mass loss in- changes in polysaccharide and low molecular
dicates that same compounds are released during weight lignin content.
pyrolysis. The mass loss during heat treatment is
from 1 to 5 percent of incoming dry solids, Söder-
hjelm et al. (1998). The finding seems to suggest
Sulfur release
that in addition to mass loss, also other changes in Sulfur release during pyrolysis at 900 oC for a mill
the black liquor that affect combustion take place.
Figure 4-20, Swelling versus polysaccharides content.

Figure 4-21, Sulfur release during combustion at 3 % O2 and 900 oC for a mill liquors with and without LHT.
liquor is shown in Figure 4-21. The liquor heat This would seem to indicate that at least part of
treatment seems to decrease the sulfur release. the sulfur released during pyrolysis comes from
Sricharoenchaikul et al. (1995) measured that the the lignin demethylation reactions. Manninen and
main pyrolysis products of black liquor droplets Vakkilainen (1996) have modelled the black liquor
were dimethyl sulphide, methyl mercaptane and droplets and found that changes in swelling do not
hydrogensulfide. The same products are the main wholly explain the changes in sulfur release.
species released during heat treatment. McKeough
(1995) measured that of the main sulfur contain- Nitrogen release
ing pyrolysis products of black liquor droplets the
concentrations of dimethyl sulphide and methyl Figure 4-22 shows from the on-line measure-
mercaptane decreased for heat treated liquors. ments amount of NO formed during the pyrolysis
Figure 4-22, Nitrogen compound release as NO in flue gases during combustion at 3 % O2 and 900 oC.

Figure 4-23, Nitrogen release during combustion at 3 % O2 and 900 oC for a mill liquors with and without
LHT.
stage. For the treated liquors the decrease in the

NO formation is approximately 10 %. Decrease of
nitrogen release during pyrolysis of heat treated
liquors was also noted by Aho et al. (1993).
Base liquor nitrogen release (liquors 1, 3, 7 and

10) differs from each other. Heat treatment to
these liquors decreases nitrogen release (liquors 2,
4, 9 and 11).

5 Chemical processes in furnace
Figure 5-1, Some of the reactions in the lower furnace.
Recovery boiler processes efficiently inorganic are released. Finally char burning takes place
and organic chemicals in the black liquor. Efficient when carbon is burned off. In the lowest part
inorganic chemicals processing can be seen as of the furnace there are char bed reactions. In
high reduction rate. The furnace also disposes of addition to char burning, these consist mostly of
all organics in black liquor. This means stable and inorganic salt, especially melt reactions. In the up-
complete combustion. Reduction (removal of oxy- per furnace there is volatiles combustion. Sodium
gen) and combustion (reaction with oxygen) are sulfate and carbonate fume formation with other
opposite reactions. It is difficult to achieve both at aerosol reactions take place. The best way to study
same unit operation, furnace. the multitude of chemical reactions taking place
in the recovery boiler is to look at first what we
Other furnace requirements are even more know of the conditions in the furnace and then
complex. A recovery boiler should have a high individual reactions main component by main
thermal efficiency. It should produce low fouling component.
ash. Processes in the recovery boiler should be
environmentally friendly and produce a low level
of harmful emissions. In spite of successes, opti- 5.1 FURNACE GAS PHASE
mizing recovery boiler chemical processes is dif-
ficult. Processes are complex and there are several Above the char bed there is a mixture of air from
streams to and from the recovery boiler. primary and secondary air jets and combustion
gases. Relatively few measurements of gaseous
There are many simultaneous reactions going on components in the recovery boiler furnace have
in the lower furnace, Figure 5-1. First there are been done. Most notably Roos (1968), Borg, Teder
the black liquor burning processes described in and Warnquist (1974) measured low solids boiler
Chapter 4. Drying occurs when water is evapo- furnace gases. In addition equilibrium calcula-
rated, Devolatilization occurs when droplet size tions of furnace have been done (Perjyd and
increases and gases generated inside the droplet Hupa, 1984).
Chemical processes in furnace 5-1

Table 5-1, Concentration[ppm] in gas phase in Experimental results and equilibrium calculations
the lower furnace, (Roos, 1968, Borg et al., 1974, are compared in Table 5-1. It shows that there are
Perjyd and Hupa, 1984). large amounts of H2 in the lower furnace. Multiple
Roos Borg meas- Perjyd air levels can be seen as reducing atmosphere,
measured ured calculated which manifests itself with large concentrations of
unburned CO in the lower furnace. Presence of O2
H2 30 000 >10 000 as well as CH4 shows that kinetic and/or mixing
H2O 40 000 ~100 000 rates prevent reactions from reaching equilibrium.
CO 10..50 000 60 000 ~100 000 Use of equilibrium assumption in treating lower
recovery furnace should be avoided.
CO2 50..120 000 130 000 ~100 000
O2 1-40 000 50 000 ~0 Equilibrium calculations show that there are sig-
N2 680 000 >10 0000 nificant amounts of sodium compounds as fume
CH4 10 000 in the lower furnace. Because furnace tempera-
tures are rather low H2S is the predominant result
H2S 1..10 000 300 ~1 000
of free sulfur reactions. Formation of SO2 would
SO2 10..100 1 10..1 000 be significant with temperatures over 1100 oC.
Na <10 000
The temperature of the gas immediately above the
NaOH <10 000
char bed is lower than the temperature of the gas
NaCl <1 000 1 ... 2 m from bed surface. The IPC calculations
(Jones, 1989) predicted a temperature gradient
Table 5-2, Measurements at the ninth floor ( ~10 of 100 oC/m. Hellström (1977) and Jutila et al.
m above nose). (1978) measured similar gradients. When char
Component Emission bed temperatures are high, unprotected tempera-
ture probe will see almost constant temperature
SO2 ppm 1
(Tavares et al., 1998). In Figure 5-2 temperature of
TRS ppm 1 the char bed reaches close to 1100 oC. Distance 0
NOx ppm 100 equals solid char bed. Measurements higher than
distance D were made in lower furnace gas phase.
CO ppm 85
The measured temperature profile is similar to the
profile reported by Lundborg (1977) for an older
Table 5-3, Measurements in the stack. boiler. The temperature at distance 0 was 783 oC
Component Emission and the gas temperature away char bed 1013 oC.
Water content % 22.9
In all of the cases there was initially a temperature
Temperature °C 159.2
raise due to hot combustible gases from the char
O2 (wet) % 3.2 bed and the CO combustion. This was followed by
CO2 (wet) % 15.7 a temperature drop of 100 ... 200 oC. The prob-
CO ppm 7.1
able cause is the cooling effect of secondary air.
Then the temperature increased again to reach a
SO2 ppm 1.1 peak a few meters higher, where the main volatiles
TRS ppm 1.2 release/combustion takes place.
NOx ppm 105.6
The published measured temperature maximums
in the recovery boiler furnace are around 1100 ...
1200 oC (Jutila et al., 1978, Shiang and Edwards,
1985). It should be noted that the boiler in ques-
tion operated in extremely high load. It was also
operated at combustion air deficiency of 10 %.
Vakkilainen and Holm (2000) present in furnace

measurements in a modern high dry solids boiler.
Tests were conducted with continuous emis-
sion measurement equipment. O2-, CO-, SO2-,
TRS-, NO-, NO2-content and temperature were
recorded. Furnace gas measurements were done
Figure 5-2, Char bed temperature profile (Tavares with air-cooled probes about 1.5 m from the walls.
et al., 1998). Furnace measurements were conducted at five
Figure 5-3, SO2 (left) and TRS (right) from in furnace measurements.
Figure 5-4, CO (left) and NOx (right) from in furnace measurements.

locations along different heights. Distance inside 5.2 CHAR BEDS
the furnace was about 2 m.
The char bed is the heart of the recovery boiler. It
First emissions at approximately the nose level is a pile of material that includes carbon, partially
were measured. At the same time emissions were black liquor solids and smelt (Grace, 2001). The
measured in the stack, Tables 5-2 and 5-3. It can temperature in char bed changes from that of the
be seen that the emissions remain constant from water walls to that of flue gases. The materials in
the nose to the stack. This being a modern high char bed can be molten, soft or solid. Without the
solids boiler can be seen as very low SO2 and TRS char bed the main target of the recovery boiler,
contents. Earlier measurements (Boonsongsup high reduction would be hard to achieve.
et al., 1994) indicate higher emissions at the nose
level and decreasing emissions from the nose level Char bed serves as repository of material. How-
towards the boiler bank inlet. This might be ex- ever the amount of incoming material must equal
plained by the differences in boiler size and higher to the outflow of smelt and gases if the char bed
solids fired. This boiler is much higher than the is to remain stable. The bed grows locally if the
boilers in the earlier test. It has larger furnace and rate of material reaching the bed exceeds that of
longer residence time. material leaving the bed and decays if the rate of
material reaching the bed is less than the rate of
The formation of emissions was studied by meas- material leaving (Grace 2004).
uring emission profiles in the furnace. Similar but
more limited studies have been reported from an- The char bed receives new fresh include; liq-
other boiler by (Vakkilainen et al., 1998). By look- uor droplets from spraying, fallen residue from
ing at various measured emission profiles, Figure furnace walls and superheater area. In addition to
5-3 and Figure 5-4, conclusions can be drawn for falling char chunks, the material coming from the
the emission performance of large boilers. furnace walls includes molten smelt (Frederick,
Singbeil and Kish, 2003).
Sulfur emissions decrease to very low levels
quickly, Figure 5-3. The sulfur release from Char bed takes care of a significant part of black
sprayed black liquor droplets occurs mainly at liquor combustion. Dried and partially pyrolysed
or below the liquor gun level (Vakkilainen and liquor droplets fall on it. Some of the carbona-
Holm, 2000). Depending on the firing method it ceous material is gasified to be combusted in the
can extend down to the char bed level. The meas- upper furnace. The reduction of sodium sulfate in
ured values of both SO2 and TRS after the lower the char bed to sulfide is the main chemical proc-
tertiary level were insignificant. The air system ess step in the recovery cycle. Sodium, potassium
and consequently good mixing can explain the and chloride release takes place in the char bed.
disappearance of TRS as it is oxidized to SO2. It is
possible that higher levels of sulfur compounds Björkman and Warnquist (1985) state that only
could have been measured close to the centerline the region near the primary air ports is a site of
of the recovery boiler. These measurements, how- efficient gasification. The rest of the bed being
ever, clearly indicate that at or near the furnace only a storage of pyrolyzed liquor. If this were true
walls the SO2 will react with available fume. A full there would be large velocity and temperature
explanation of these results needs to wait until gradients at the char bed surface. Viewing the
the reaction rates of several known sulfurdioxide- char bed with the char bed imagining cameras
consuming reactions have been measured at high reveals that also the secondary jets reach the char
temperatures and a more comprehensive model of bed surface. Therefore the whole char bed surface
alkali release is available. serves as reaction zone.
Good mixing by the air system is further seen in Björkman & Warnquist further state that the
the decreasing trend for CO, figure 5-4. Most of gasification reaction creates inverted turbulent
the CO has reacted before the highest air level. diffusion flames. It is clear that there exists mixing
The curves for CO and TRS look alike, probably limitation above the char bed surface. Otherwise
because mixing heavily influences both. we could not measure high oxygen and high CO
concentrations at the same time.
Measured NOx profile is virtually identical to the
earlier measurements (Vakkilainen et al., 1998) in There have been different views of the impor-
another boiler, Figure 5-4. Fuel nitrogen species tance of the char bed reactions. Björkman and
are released very fast, virtually with the volatiles Warnquist (1985) consider the weight fraction of
release. It seems probable that much of the nitro- particles that spend their ‘life story’ in flight to
gen species react after the last air level. be insignificant. In the modern recovery boiler a
Figure 5-5, Char bed and boiler right and back wall through char bed camera (Saviharju and Pynnönen,
2003).
significant portion of the carbon release, maybe 60 Table 5-4, Thermal properties of char bed materi-
… 70 % is happening in flight. Recent measure- als (Adams and Frederick, 1988).
ments (Tamminen et al., 2002) have shown that Material Density cp λ a
maybe only 10 % of fume generation occurs in kg/m3 J/kgoC W/moC 109 m2/s
char bed.
Inactive 400- 1254 0.078 50-75
char 1330
There has been some char bed models produced
to study the operation of char bed, (Richardson Active char 290-460 1254 0.28- 500-
0.38 1000
and Merriam, 1978, Shiang and Edwards, 1986,
Sumnicht, 1989 and Sutinen et al., 2002). They Smelt, 923 1338 0.450 181
have provided insight on the basic events in the liquid
char bed. As there are only a few reported meas- Smelt, solid 2163 1421 0.882 284
urements of relevant char bed data those studies
have formed the basis of our char bed understand- Table 5-5, Measured char bed composition
ing. (Heinävaara, 1991).
Element S1 S2 S3
All these attempts have tried to model stationary
char beds. In practice we see the char bed height Carbon, C* % 0.5 2.4 0.3
and temperature changing. By creating a time Sulfur (tot), S % 10.5 11.1 11.6
dependent model of char bed it would be possible Sodium, Na % 40.8 40.1 42.2
to model the real behavior of char bed. To create
Potassium, K % 3.9 4.0 4.2
a time dependent char bed model, reaction rates
for main reactions have to be determined. There Chloride, Cl % 0.3 0.4 0.3
are some measurements of char bed heat transfer Magnesium, Mg % 2.7 2.6 0.7
behavior, from them time dependent behavior Calcium, Ca 10-6 45.3 47.3 68.5
can be extrapolated. The flows in the char bed
Iron, Fe 10-6 69.8 161 161
are a new challenge that has not been previously
modeled. Phosphor, P 10-6 18.7 9.6 32.5
Na2CO3 % 66.2 74.1 63.6
Char bed physical characteristics Na2SO4 % 6.8 3.7 8.3
Na2S % 21.6 24.9 23.3
Char bed contains mainly carbonaceous char,
sodium carbonate, sodium sulfate, sodium sulfide Na2S2CO3 % 0.9 1.0 0.8
and sodium chloride with a smaller percentage of * Measured as heating loss of the unsolvable portion of sample

potassium salts and unreacting material present. Table 5-6, Average composition of incoming
Typically the char bed shape is that of mostly material in char beds, weight percent. (Richardson
flat pile with a 1 ... 2 m height and sloping sides. and Merriam, 1977).
The top surface contains mounds and even smelt Substance Sample 1 Sample 2
pools. The shape isn’t constant, but changing
with time so there are several other types of beds C 10.7 7.1
that have also been reported, Figure 5-5. In high Na2S 8.6 3.7
beds there is a possibility of molten smelt pockets
Na2SO4 8.1 5.5
inside the bed. These pockets are not visible to
the outside. The question remains whether there Na2S2O3 6.2 2.6
is a situation in char bed operation, which can be Na2CO3 66.8 77.9
characterized as stationary.
NaCl - -
The actual black liquor char content of the char NaOH - -
bed is only 5 – 20 %. Most of it is inorganic salts K2 S - -
(Warnqvist, 1994). Heinävaara (1991) has made
K2SO4 - -
measurements of char beds, Table 5-5.
The chloride in the char bed is only half the chlo-

Table 5-7, Measured temperatures and densities
ride in the incoming black liquor. This is because
of char beds (Richardson and Merriam, 1977).
chloride volatilizes easily and thus participates in
the dust recycle. The magnesium seems to enrich. Mill Temperatures Densities
o
The magnesium content in the char bed is about C kg/m3
hundred times that of magnesium content of both Westvaco 1060-1070, surface 460, surface
the incoming liquor and the smelt. Reduction in CE, 1000 t/d 1120-1150, surface 2100, solid
the char bed is lower than at the outflowing smelt smelt
(Warnqvist, 1994). Brunswick 750-910, smelt BW
Georgia < 760, core, BW
During the ADL study (Merriam, 1978) a nitrogen 795-810, smelt CE
cooled probe was used to get samples of incom-
Ticonderoga 650-705, core 1300, char
ing material. As can be expected the results from
BW, 500 t/d 770, 0.150 m deep
B&W and CE furnaces differed. 870-1100, surface
Most of the material reaching the char bed is Androscoggin 960-1140, peak
CE, 600 t/d 770-830, core
either pyrolysed or gone through complete char
burning. Sometimes portions of liquor reach the
bed in a state of incomplete drying (Björkman & Table 5-8, Measured thermal conductivities and
Warnquist, 1985). diffusivities of char beds (Richardson and Merriam,
1977).
Mill Type, size Conductiv- Diffusivity
Char bed temperature and W/m3oC
ity, W/m2oC
density Ticonderoga B&W, 500 0.078 160
The structure of char bed is mostly a mixture of t/d
frozen and liquid sodium salts in a carbon matrix. Androscoggin CE, 600 t/d 0.28, char 290
The char bed contains sections ranging from very 0.45, smelt
porous to solid, glassy portions. Richardson and
Merriam (1977) measured properties of several
char beds during the ADL study. In their opinion
only the top layer of the bed is hot and reacting. in the bed was found to be up to 0.3 m deep in the
Measurements were made of temperatures, densi- char bed.
ties, thermal conductivities and thermal diffusivi-
ties. Some thermal properties of different char bed The surface temperatures measured in the ADL
materials are shown in the Table 5-7. Richardson study are somewhat higher than the ones reported
and Merriam (1977) observed that the bed surface by Borg et al. (1973) Figure 5-6. STFI values were
temperature wasn’t constant but there were cooler measured with a pyrometer.
and hotter spots. They also measured tempera-
tures and densities inside the bed, Table 5-7. In New devices such as char bed imagining cameras
some measurements the maximum temperature (Harrison and Ariessohn, 1985) have revealed the
cycling nature of char bed surface temperature. In
vated if molten smelt pockets remain in the bed.
Bed cooldown and heat up

When recovery boiler is shut down the smelt bed
remains. Typically oil firing is used to burn some
of the bed out. To save the recovery boiler furnace
bottom tubes all of it can’t be removed.
Char bed cooling has been studied by Richardson

and Merriam (1977, 1978). They conclude that,
conditions leading to rapid cooldown of the bed
are minimum height, high porosity in the top lay-
ers of the bed, and high content of solidified smelt
in the remainder of the bed.
Figure 5-6, Char bed surface temperatures (repro-
duced from Borg et al., 1973). Char bed cooling rate is very slow as porous bed is
a very good insulator. There can be molten smelt
pockets remaining after 1 … 5 days from stop
addition growing and falling of bed surface up to of firing. Bed inside temperature decay depends
1 m/min was observed. Cyclic surface temperature on the bed height, smelt content in the bed, and
variations of 80 oC/10 min were also observed. porosity and thickness of the top layers of the bed.
The surface temperature of the char bed has a Heat transfer in char beds isn’t solely conduction.
strong influence on emissions from recovery boil- Richardson and Merriam (1978) indicated that
ers. The higher the char bed surface temperature radiative heat transfer in the porous structure
the higher the dust emissions and the lower the plays a role. Heat transferred with the flowing
SO2 and H2S emissions (Chamberlain and Cairns, molten smelt is maybe the most significant heat
1972). Borg et al. (1973) have stated that the H2S transfer mode.
emissions from the char bed decrease from 750
ppm to 250 ppm when the bed surface tempera- Char bed is a large heat sink. The typical hearth
ture increases from 825 to 860 oC. This is consist- area loading of a modern recovery boiler is 0.2
ent with data from Lien et al, (2004) which state kgds/s /m2. With a HHV of 14.5 MJ/kgds this
that fume and particle production from char bed means heat input per char bed unit area of about
start at about 850 oC. 3 MW/m2. Kelly et al. (1981) measured the smelt
inventory of a CE boiler to be 250 kg/m2. With an
Furnace loading and black liquor dry solids average smelt enthalpy of 1350 kJ/kg the heat in
content have increased since 70’s. This means that smelt per char bed unit area is about 340 MJ/m2.
the surface temperatures are higher in the mod-
ern recovery boilers. The differences in density The average bed height is typically less than 2 m.
measurements were reportedly caused by uneven The average char bed temperature is less than 800
structures of char beds. The surface of the char o
C. With average heat capacity of 2000 J/kgoC and
bed is very porous. The deeper we go into the bed density of 1300 kg/m3 the heat in the char bed is
the denser it gets. less than 4200 MJ/m2.
Thermal conductivity and This means that the char bed contains about 1500
s worth of heat. The long cooling times observed
diffusivity are result of slow heat removal, not large heat
In some mills char bed thermal conductivity and content.
thermal diffusivity measurements were made,
Table 5-8. Safety
With high char beds cooldown times of 50 to 100 One of the main hazards in operation of recovery
hours have been reported. This situation is aggra- boilers is the smelt-water explosion. This can
vated if molten smelt pockets remain in the bed. happen if even a small amount of water is mixed
with the solids in high temperature. Smelt-water
With high char beds cooldown times of 50 to 100 explosion is purely a physical phenomenon.
hours have been reported. This situation is aggra-

Figure 5-7, Smelt flow from char bed (Saviharju and Pynnönen, 2003).
The smelt water explosion phenomena have been

studied by Grace (1988). The liquid - liquid type Smelt is the product of inorganic reactions in
explosion mechanism has been established as one recovery furnace. At the same time the carbon is
of the main causes of recovery boiler explosions. consumed the residual inorganic portion melts.
Inorganics flow out of the furnace through smelt
In the smelt water explosion even a few liters spouts, Figure 5-7. The amount of smelt inside
of water, when mixed with molten smelt can recovery boiler furnace has been measured by
violently turn to steam in few tenths of a second. Kelly et al. (1981). They found the smelt content
Char bed and water can coexist as steam blan- per furnace unit area to be about 250 kg/m2 for
keting reduces heat transfer. Some trigger event a decanting CE unit and about 140 kg/m2 for a
destroys the balance and water is evaporated B&W unit. The residence times found were 44 and
quickly through direct contact with smelt. This 25 minutes respectively.
sudden evaporation causes increase of volume and
a pressure wave of some 10 … 100 000 Pa. The Smelt temperature is about 100 oC higher than
force is usually sufficient to cause all furnace walls initial deformation temperature (Sandquist, 1987).
to bend out of shape. Safety of equipment and In older low solids boilers the smelt temperatures
personnel requires an immediate shutdown of the are 750 … 810 oC (Lundborg, 1977). In modern
recovery boiler if there is a possibility that water high solids boilers the typical smelt temperatures
has entered the furnace. All recovery boilers have are 800 … 850 oC.
to be equipped with special automatic shutdown
sequence. The smelt flow corresponds typically from 0.400
to 0.480 kg per kilogram of incoming black liquor
The other type of explosions is the combustible dry solids flow, Figure 5-8. The points highest
gases explosion. For this to happen the fuel and from average are caused by very high potassium
the air have to be mixed before the ignition. contents.
Typical conditions are either a blackout (loss of
flame) without purge of furnace or continuous Reduction and sulfidity
operation in a substoichiometric state. To detect
blackout flame monitoring devices are installed, The main process property of the smelt is the
with subsequent interlocked purge and startup. reduction. Reduction is usually expressed as the
Combustible gas explosions are connected with molar ratio of Na2S to Na2SO4,
oil/gas firing in the boiler. As also continuous O2
monitoring is practiced in virtually every boiler Reduction = Na2S/(Na2S+Na2SO4) 5-1
the noncombustible gas explosions have become
very rare. The higher the reduction the lower the amount of
sodium that reaches the cook unusable. Reduc-
5.3 SMELT
Figure 5-8, Smelt flow as function of sodium in black liquor.
tion rates of 95 ... 98 % are not uncommon in well

operated recovery boilers. Usually the reduction Often the typical mill analysis of reduction rate is
efficiency increases as the char bed temperature done for green liquor. Alkali in the green liquor
increases. From thermodynamical equilibrium will typically result in lower values that what is
calculations we note that there should be very lit- measured in smelt, Figure 5-9. Typically in mod-
tle of sodium oxides and thiosulfate. ern mills the reduction in green liquor is 2 … 3
percent points lower that in smelt.
Sulfidity is the molar ratio of sodium sulfide to the
total alkali content.
Smelt composition
Sulfidity = Stot/(Na2+K2) 5-2 Smelt compositions for modern recovery boilers
don’t vary much, Table 5-9. The amount of sodium
This equation is widely in use because of ease chloride and potassium compounds is a function
of measuring Sulfidity depends on the liquor of their concentration in the black liquor. When
circulation of the mill. Too high a sulfidity causes we think of smelt composition it must be borne in
operating problems for the recovery boiler. Espe- mind that smelt having greater than 2 % carbon
cially increased sulfidity increases SO2 and TRS content is not flowing but rigid (Grace, 1997).
emissions (Chamberlain and Cairns, 1972).
Figure 5-9, Effect of weak white liquor composition on reduction in green liquor, reduction is smelt 95 %,
sulfidity 35 %.

Table 5-9, Smelt composition. In poorly operating recovery boilers the dregs
Softwood Hardwood in green liquor reach 2 000 … 4 000 mg/l, while
improved operation results in 100 … 1 000 mg/l
Na2S % 25 – 28 19 - 21
(Clement et al., 1995). Saviharju and Pynnönen
Na2CO3 % 66 – 68 72 - 75 (2003) measured carbon contents in smelts from
Na2SO4 % 0.4- 1.0 0.6- 1.4 boilers equipped with vertical air. Carbon content
ranged from 0.07 to 5.2 gC/kgNa or 5 … 500 mg/l
Na2S2O3 % 0.3- 0.4 0.2- 0.4
while total dregs varied 400 … 4 000 mg/l. Carbon
Others % 5 -6 3 -5 content was in one sample more than 50 %. This
is significantly higher that the typical values
Table 5-10, Properties of smelt (Wozniak, 1985). reported by Lidén (1995). He found that carbon
Property Sub- Tempera- Measured content in dregs rarely exceeded 5 w-%.
stance ture, K value
Surface ten- Na2CO3 1218 - 1244 0.208 Some important smelt properties were measured
sion, N/m in the IPC by Wozniak (1985), Table 5-10. Typi-
cally the smelt solidifies totally at about 550 oC.
Na2SO4 1232 0.189
This temperature is however influenced by the mi-
Na2SO4 1261 0.188 nor constituents of the smelt. Ahlers has measured
Na2SO4 1285 0.187 a quick drop of melting point from 750 oC to 580
o
Density, Na2CO3 1218 1927 C, when the smelt NaCl composition increased
kg/m3 – 1244 – 1917 from 0 to 10 % (Ahlers, 1971).
Na2SO4 1232 2032
– 1261 – 2018
5.4 SODIUM
Table 5-11, Physical properties of selected compo- Sodium is released during the black liquor
nents (Himmelblau, 1989). burning and char bed reactions through vapori-
Component Melt. Boil. Density Formula zation and reduction of sodium carbonate (Li
K K kg/m3 wt
and van Heiningen, 1990). Sodium release from
Calcium carb., Ca2CO3 dec. 1098 2930 100.09 combusting black liquor increases as a function
Calcium oxide, CaO 2873 3123 2620 56.08 of temperature, Figure 5-10. At beginning of
Chromium, Cr 1888 2473 7100 52.01
combustion large portion of sodium is connected
to the organic portion of the black liquor. At the
Copper, Cu 1356 2855 8920 63.54 end of volatiles release almost all of it is inorgani-
Ferric oxide, Fe2O3 1883 dec. 5120 159.77 cally bound.
Magnesium, Mg 923 1393 1740 24.32
Magnesium oxide, MgO
Sodium release in kraft recovery boilers increases
3173 3873 3650 40.32
with increasing lower furnace temperature, Figure
Potassium, K 1032 dec. 860 39.10 5-11. It has been assumed that in industrial boilers
Potassium carb., K2CO3 1264 dec. 2290 138.27 all of the ESP dust is from reactions with vapor-
Potassium cyanide, KCN 907 - 1520 65.11 ized sodium. In addition the amount of sodium
released as a function of carbonate in ESP dust
Potassium chloride, KCl 1063 1773 1988 74.56
seems to increase. Increase in carbonate indicates
Potassium hydr., KOH 653 1593 2044 65.10 increase in lower furnace temperature (Vakki-
Potassium sulfate, K2SO4 860 - 2662 174.25 lainen, 2000). Sodium content in black liquors is
Potassium sulfide, K2S
around 20 w-%. This means that sodium release in
333 - - 200.33
recovery furnace is about 10 % of the sodium in
Silicon dioxide, Si2O 1883 2503 2250 60.09 black liquor.
Sodium, Na 371 1153 970 22.98
Sodium carb., Na2CO3 1126 dec. 2533 105.99 Much studied reactions involving sodium are
hydroxide formation, reduction reactions, and sul-
Sodium cyanide, NaCN 835 1770 - 49.01
fate formation with hydroxides, sulfate formation
Sodium chloride, NaCl 1081 1783 2163 58.45 with chlorides, sulfate formation with carbonate
Sodium hydroxide, NaOH 592 1662 2130 40.00 and carbonate formation.
Sodium sulfate, Na2SO4 1163 - 2698 142.05
Some of the sodium reactions occurring in recov-
Sodium sulfide, Na2S 1223 - 1856 78.05 ery furnace are shown in Figure 5-12 and listed
Sulfur, S 386 718 2070 32.07 below.
dec. = decomposes before this step

Figure 5-10, Sodium loss from combusting black Figure 5-11, Sodium release to ESP ash as function
liquor as a function of actual particle temperature of carbonate in ash in industrial boilers
in a LEFR for 100 μm particles (Wåg et al., 1997).
Na(c) → Na(g) 5-3 Carbonate reduction

Na2CO3 + 2C → 2Na(g) + 3CO 5-4 Carbonate reduction (reaction 5-4) seems to be
the one of the main carbon consuming reactions
Na2S + CO2 + H2O → Na2CO3 + H2S 5-5 after volatiles release above 900 oC (Järvinen,
2002). If carbonate reduction is the main sodium
Na(g) + H2O → NaOH + H 5-6 release reaction, then sodium and sulfur release
are independent.
Na2SO4 + 4C → Na2S + 4CO 5-7
Reactions with oxygen
NaOH + SO3 → Na2SO4 + H2O 5-8
In the char bed, high concentrations of oxygen
NaOH + NaHSO4 → Na2SO4 + H2O 5-9 immediately above the bed has been measured.
Borg et al. (1974) measured oxygen concentra-
Na2CO3 + 2SO2 + 0.5O2 → Na2SO4 + CO2 5-10 tions of 3 ... 5 %. The two implications of this are
that either there was not enough mixing or the
reactions are kinetically limited.
Figure 5-12, Reactions of sodium and potassium in the lower furnace (Reactions of potassium are similar to
reactions of sodium).

Inside char particles there are gas producing The rate is fairly independent of sulfate content up
reactions. Oxygen is consumed by reactions to about 94 % reduction, but then reduces sharply.
with carbon dioxide and hydrogen. Oxygen stays This in addition of smelt surface oxidation ex-
mainly outside the char (Wåg et al., 1995, 2002). It plains why reduction rates of more than 95 % are
seems that direct reactions of char with oxygen are hard to achieve. The air requirement to complete
rather limited. this cycle is about 20 ... 40 % of total air.
Reactions with carbon dioxide Sjöberg and Cameron (1984) studied the sodium
sulfate reduction with CO.
and monoxide
Na2SO4 + 4CO → Na2S + 4CO2 5-16
Na2S + CO2 + H2O → Na2CO3 + H2S 5-11
They found the rate equation to be
Sodium sulfide can react with carbon dioxide and
produce hydrogen sulfide. 5-17
Reduction The constants measured were

The carbon has a dual role in the char bed (Grace, KCO = 31.7 kmol/m3atms
Cameron and Clay, 1985). During the sulfate- ECO = 115 kJ/kmol
sulfide cycle the carbon reacts with Na2SO4 to
form Na2S. If smelt contacts air it can oxidize. The Birk et al. (1971) have measured the sodium sul-
reduction process is slow requiring significant fate reduction by hydrogen.
residence time to be completed to a high degree.
Na2SO4 + 4H2 → Na2S + H2O 5-18
The sodium sulfate - sulfide cycle was discovered
by Grace, Cameron and Clay. In it the three com- They found the hydrogen reduction rate to be
peting reactions take place; Oxidation of sulfide about six times faster than with CO at 840 oC.
Because of water shift reaction there is hydrogen
Na2S + 2O2 → Na2SO4 5-12 inside char bed.
and reduction of sulfate by char In addition to carbon monoxide and hydrogen,

water vapor can reduce sodium sulfide. Rates
Na2SO4 + 2C → Na2S + 2CO2 5-13 of these reactions are more than two orders of
magnitude slower than with carbon at recovery
Na2SO4 + 4C → Na2S + 4CO 5-14 boiler conditions (Grace, 1997). Their role in
kraft recovery boilers is insignificant (Sjöber and
It can be observed that to achieve high reduction Cameron, 1984).
efficiencies the rate of sodium sulfide oxidation
has to be much lower than the rate of sodium Reactions with water vapor
sulfate reduction. It is generally accepted that in
kraft recovery boilers the sulfide oxidation is mass Water vapor can react with sulfide and create
transfer controlled (Wåg et al., 1997). High reduc- sodium bisulfide. This reaction is responsible for
tion can be achieved by having reactive, fresh char decrease in causticization rate. So it is very prob-
on top of char bed. able that until the black liquor droplet has dried
there is some sodium hydroxide present.
Cameron and Grace (1982) measured the rate of
reduction reaction with the presence of carbon. Na2S + H2O → Na2S + 4H2O 5-19
The rate equation was found to be
Sodium carbonate can react with water vapor and
produce sodium hydroxide (Blackwell and King,
5-15 1985), but probably the rate controlling step is
sodium carbonate decomposition.
The constants measured for kraft char were
Na2CO3 + H2O → 2NaOH + CO2 5-20
KRed = 1.31+0.41*103, 1/s
B = 0.022+0.008, kmol/m3
Ea = 122, kJ/kmol

Figure 5-13, Equilibrium composition curves (Eskola et al., 1998).
Formation of sodium vapor Na2SO4 → 2Na + SO2 + O2 5-22
Based on Figure 5-13 the most probable gaseous This reaction would produce sulfur dioxide and
components are sodium chloride and sodium oxygen in addition to volatile sodium. The reac-
hydroxide. Existence of metallic sodium is not tion is more favorable at high temperatures and
thermodynamically favorable. There have been low oxygen concentrations.
discussions as what is the form of sodium release.
Gaseous sodium, sodium atoms and sodium oxide Sodium carbonate can decompose
have been suggested as products. There is a large
amount of water vapor in recovery boiler furnace. Na2CO3 → Na2O + CO2 5-23
After release most of the sodium reacts fast to
sodium hydroxide (NaOH) (Jokiniemi et al., 1993, There is an abundance of water vapor present so
Jokiniemi et al., 1995). Furnace tests have detected probably sodium hydroxide is formed
as much as 26 w-% alkali hydroxides in lower
furnace fume deposits (Tavares et al., 1995). Na2O + H2O → 2NaOH 5-24
Iisa and Jing (2000) studied sodium vaporization

in a laminar entrained flow reactor. In their ex- 5.5 POTASSIUM
periments with nineteen black liquors the sodium
Potassium in the black liquor originates from the
release seemed to be fairly independent of the
wood supply. Typically its content is 3 ... 5 % of the
liquor Na content. They also added sodium salts
sodium content (Tran 1989).
to one black liquor. They found that the sodium
release could best be correlated to amount of fixed
Potassium is chemically related to sodium. Practi-
C at the resulting liquor. This seems to indicate
cally all sodium compounds will also exist as
that none of the additions substantially changed
potassium compounds. They behave similarly in
sodium release reactions and that sodium release
kraft recovery furnace. Potassium release seems
seems to be dependent on organic portion of the
to correlate well with sodium release in industrial
liquor.
boilers, Figure 5-14.
Cameron (1988) showed that sodium can be
Potassium release is on average about 1,5 times
released by oxidation of sodium sulfide.
higher than sodium release and a bit more than
15 w-% of potassium in black liquor. It has been
Na2S + O2 → 2Na + SO2 5-21
suggested that significant amounts of potassium
can be released through potassium chloride
Sodiumsulfate and -carbonate vaporization (Verrill and Wessel, 1998). Data from
decomposition industrial boilers indicates that potassium enrich-
ment and chlorine do not correlate, Figure 5-15.
Sodium sulfate can decompose The potassium in the black liquor will enrich close

Figure 5-14, Sodium release versus potassium release in industrial boilers.
to the tube surface, Table 5-12 (Klarin 1992). Table 5-12, Elementary analysis of char bed close
to the tube surface (Klarin, 1992).
Distance from 2-5 5-10 10-30 >30
5.6 SULFUR tube surface, mm
Sulfur emissions play a crucial role in the char bed Color Black Grey Yellow Red
reactions. The one main concern in any recovery Solubility to Very Good Fair Slow
boiler is; will we get low enough sulfur emissions. water good
The sulfur in black liquor is for the most part in Aluminum, Al 0 0.27 0.33 0.08
inorganic sulfur compounds. The main inorganic
Chloride, Cl 1.6 1.85 0.35 0.32
compounds are sulfide and sulfate. In addition mi-
nor amounts of thiosulfate, sulfite and polysulfates Iron, Fe 0 0 0.4 0.14
are present (Grace, 2001). 30 to 40 % of sulfur is Potassium, K 12.0 5.77 5.09 7.15
bound to organosulfur compounds. Thiosulfate Silica, Si 0.30 0.22 0.33 0.77
and organic sulfur are rapidly released from gase-
ous sulfur compounds (Jing and Iisa, 2001). Sodium, Na 38.1 67.6 67.8 64.7
Sulfur, S 48.0 24.3 25.8 26.8
The sulfur as such is not stable compound at S/Na2 2.52 0.72 0.76 0.83
recovery boiler furnace conditions. Björkman
K/Na2 0.32 0.09 0.08 0.12
Figure 5-15, Potassium release versus chloride in black liquor in industrial boilers.

Figure 5-16, Sum of sulfur in SO2 and sulfur in ESP dust in industrial boilers, upper figure % of sulfur re-
leased, lower figure total release as function of black liquor sulfidity. rofile, dry gases.
Figure 5-17, Sulfur release as function of temperature (Wessel and Kaufman, 2000).

and Warnquist (1985) claim that about 30 % of Figure 5-18 shows theoretical predictions
incoming sulfur leaves the recovery boiler furnace compared to experimental values in one furnace
in the flue gas and fume. Ångpanneföreningen’s (McKeough and Janka, 2001). Actual parameter
measurements from ten boilers in Sweden indi- that varied was boiler load, but measured data was
cated about 15 gS/kgds sulfur release when sulfidi- plotted versus lower furnace temperature. Increase
ties were between 30 and 45 mol-% and about 20 in temperature of 100 oC would decrease sulfur
gS/kgds sulfur release when sulfidity was around release by 25 %. Magnusson (1977) states that
50 mol-% (Moberg, 1973). Recent industrial data recovery boiler load affects the sulfur release. Data
from modern boilers, Figure 5 – 16, show that from industrial boilers suggests that load does not
sulfur release is between 10 and 20 gS/kgds. Dry affect the sulfur release above 2 MW/m2 load.
solids does not have a strong correlation with
sulfur release. It seems that sulfur release as sum Contrary to Figure 5-18 general trend from boil-
of S in ESP dust and SO2 is very insensitive to dif- ers seems to be that increase in dry solids does
ferences in virgin sulfidities of black liquor. not affect the sulfur release, Figure 5-19. Similarly
the boiler specific load does not seem to affect the
In upper part of Figure 5-16 released fraction of sulfur release. Manninen and Vakkilainen (1996)
sulfur decreases with increased sulfidity. In lower found through their theoretical sulfur release
part of Figure 5–16 total release is hardly affected model that droplet size affected somewhat the
by black liquor sulfidity. Sulfur release is about 15 sulfur release. It is probable that differences on the
g/kgds or 20 … 40 w-% of sulfur. results from various sources reflect also the way
the boilers are fired.
It should be pointed out that the form by which
sulfur leaves the recovery boiler is very sensitive to Reactions leading to formation
dry solids. At high dry solids hardly any gaseous
SO2 is leaving the boiler. of H2S
Sulfur release from the black liquor has been stud-
In laboratory conditions the measured sulfur
ied extensively. It is released as organic volatiles,
release as function of temperature resembles a
through reduction of thiosulfate and from reac-
bell shape (Forssén et al., 1992). Single droplet
tion of sulfide with gaseous species. The formation
data, Figure 5-17, tends to overpredict the sulfur
of hydrogen sulfide H2S is considered the main
release. Therefore somewhat lower values have
pathway to the formation of sulfur emissions.
been used to arrive to realistic levels (Wessel and
Kaufman, 2000). Generally it is believed that sul-
Li and Heiningen (1991) found that during
fur release is inversely proportional to the furnace
pyrolysis about 22 % of all sulfur in the incoming
temperature and that it is a function of black liq-
black liquor was released. In addition after the
uor droplet heating rate. Faster heating produces
sample reached about 470 oC the sulfur emissions
lower sulfur release.
in the nitrogen atmosphere stopped. The sample
Figure 5-18, Sulfur release as function of furnace temperature (McKeough and Janka, 2001).

Figure 5-19, Sulfur release as function of black liquor dry solids.
must have also been dried by then so neither H2O Na2S2O3 + 3CO → Na2S + S + CO2 5-27
nor CO2 gasification could take place. Typically
sulfur release after volatiles release stops. In pyrolysis tests about 30 % of released sulfur was
hydrogen sulfide. Most of the sulfur release came
McKeough et al. (1995) attribute inorganic sulfur when methoxyl groups of dissolved lignin in black
reactions to reaction of sulfide and reduction of liquor react with hydrogen sulfide ions. Primary
thiosulfate. Thiosulfate in black liquor seems to reaction is formation of methyl mercaptan (MM)
decompose very fast < 0.3 seconds (Sricharoen- and secondary reaction is formation of dimethyl
chaikul et al., 1995). In laboratory experiment sulfide (DMS) (McKeough et al., 1995).
about 40 – 50 % of sulfur in thiosulfate and 20
– 70 % of sulfur in sulfide was released (Kulas et Liginin-OCH3+NaHS→Lignin-ONa+CH3SH 5-28
al., 1989).
CH3SH+NaOH → CH3SNa + H2O 5-29
Na2S + CO2 + H2O → Na2CO3 + H2S 5-25
Liginin-OCH3+CH3SH→Lignin-ONa+(CH3)2S
Na2S2O3 + CO2 → Na2CO3 + SO2 + S 5-26 5-30
Figure 5-20, Reactions of sulfur in the lower furnace.

Figure 5-21, Chloride reactions in the lower furnace.
In high temperature conditions at lower furnace Sulfation of sodium carbonate

methyl mercaptan and dimethyl sulfide will
Sulfation of sodium carbonate has been suggested
quickly decompose to carbohydrates and hydro-
as the capture mechanism for sulfur dioxide
gen sulfide.
(Blackwell and King, 1985). Reaction rate of gase-
ous sulfur dioxide or trioxide with solid sodium
carbonate is too small to affect significantly to
Gaseous reactions of sulfur sulfur capture (Frederick et al., 1995).
Oxidation of hydrogen sulfide is in furnace condi- Na2CO3 + 2SO2 + 0.5O2 → Na2SO4 + CO2 5-38
tions fast and requires mostly proper mixing.
2H2S + 3O2 → 2SO2 + 2H2O 5-31

5.7 CHLORIDE
Other gas stream reactions involve fume reactions
with sodium. Sulfate formation with hydroxides, Chloride can be assumed to be entirely of NaCl.
sulfate formation with chlorides and sulfate for- Chloride compounds found in the flue gas are
mation with carbonate. mostly released due to direct vaporization.
Some of the sulfur reactions occurring in recovery The chloride is released from black liquor char by
furnace are shown in Figure 5-20 and listed below. vapor pressure. It is unclear whether it is released
as pure Cl(g), chloride gas Cl2(g) or sodium chlo-
Na2S + -OCH3 → CH3SH,(CH3)2S 5-32 ride NaCl(g). It however reacts very fast to form
hydrogen chloride HCl(g).
Na2S2O3 → Na2S + SO2 5-33
Some of the sodium chloride reactions occurring
Na2S + CO2 + H2O → Na2CO3 + H2S 5-34 in recovery furnace are shown in Figure 5-21 and
listed below.
2H2S + 3O2 → 2SO2 + 2H2O 5-35
NaCl(s,l) → NaCl(g) 5-39
NaOH + SO2 → NaHSO3 5-36
NaCl + NaHSO4 → HCl + Na2SO4 5-40
Na2CO3 + 2SO2 + 0.5O2 → Na2SO4 + CO2 5-37
NaCl + H2O → HCl + NaOH 5-41

Figure 5-22, Chloride release from operating recovery boilers.
Chloride release to completely convert gaseous sodium chloride to

sulfate before the superheater section (Iisa, 2004).
Iisa and Jing (2000) studied chloride vaporiza-
tion in a laminar entrained flow reactor. In their 2NaCl + SO2 + 0.5O2 + H2O → Na2SO4 + 2HCl
experiments with nineteen black liquors the 5-42
chloride release seemed to be linearly dependent
of the liquor Cl content. They also added sodium Hydrochloric acid found in recovery boiler flue
chloride to two black liquors. The chloride release gas corresponds roughly to 1/3 of sulfurdioxide
increased almost linearly to chloride content in present (Someshwar and Jain, 1995). This strongly
the resulting liquor. Same linear regression in supports the existence of this type of mechanism
laminar entrained furnaces was observed for black at recovery boiler.
liquor and black liquor with KCl addition (Khalaj
et al., 2004). Linear behavior is observed also for Sodium chloride – hydrochloric
industrial boilers, Figure 5-22.
acid
Increasing temperature increases chloride release
Sodium chloride can react with water vapor to hy-
in laminar entrained furnaces (Khalaj et al., 2004).
drochloric acid. Equilibrium calculations suggest
However the recovery boiler furnace temperature
that equilibrium amount would be very small.
has weaker effect on chloride release than on
sodium release (Janka et al., 2004).
NaCl + H2O → NaOH + HCl 5-43
In earlier laminar entrained flow reactor experi-
ments more than 60 % of chloride in black liquor
was released (Wåg et al., 1997). Because of large 5.8 REACTIONS INVOLVING
flue gas flow and small liquor flow the chloride
release in LEFR experiments is not limited by CARBON
sodium chloride vapor pressure, so results can be
The main reactions involving carbon in the lower
misleading.
furnace belong to two classes. Carbon is released
from combusting black liquor as organic volatiles,
Sulfation of sodium chloride through char combustion and then with reduc-
tion of sulfate. Possibly also reactions of carbonate
Sulfation of sodium chloride has been suggested and char gasification are significant. Gas stream
as a capture mechanism for sulfur dioxide (Black- reactions occurring mainly above the char bed
well and King, 1985). Reaction rate of gaseous include oxidation of carbon monoxide to dioxide,
sulfur dioxide with solid sodium chloride at 400 formation of carbonate.
… 600 oC is too small to affect significantly to
sulfur capture to deposits (Frederick et al., 1995). Some of the carbon reactions occurring in recov-
Equilibrium studies suggest that it is not possible ery furnace are shown in Figure 5-23 and listed

Figure 5-23, Reactions involving carbon in the lower furnace.
below. Char combustion

In recovery boiler furnace the carbon in the char
Na2S + -OCH3 → CH3SH,(CH3)2S 5-44 is combusted to CO2. The required oxygen for
char combustion can come from oxygen, O2, el-
-C(s) + H2O → CO + H2 5-45 emental oxygen, O, sodium sulfate, Na2SO4, water
vapor, H2O or carbon dioxide, CO2. For coal com-
-C(s) + CO2 → 2CO 5-46 bustion the main reactions are the surface C - O2
and C - CO2, which produce CO. The CO is used
2CO + O2 → 2CO2 5-47 by the gas phase CO - O2 reaction (Makino, 1990).
The reaction rate of C - CO2 is several orders of
CO + H2O → CO2 + H2 5-48 magnitude slower than the C - O2 reaction. In the
charbed the oxygen concentration is very small,
Na2SO4 + 4C → Na2S + 4CO 5-49 because CO will consume the oxygen.
Na2S + CO2 + H2O → Na2CO3 + H2S 5-50 -C(s) + 0.5O2 → CO 5-53
2NaOH + CO2 → Na2CO3 + H2O 5-51 -C(s) + O2 → CO2 5-54
Na2CO3 + 2SO2 + 0.5O2 → Na2SO4 + CO2 5-52 Järvinen (2002) has modeled black liquor com-
bustion. During char combustion water vapor
and carbon dioxide gasification seem to be the
Main carbon containing products from black dominant reactions, Figure 5-24. Direct oxidation
liquor pyrolysis in laminar entrained flow reactor of char was negligible.
tests are tars, CO, CO2 and various hydrocarbons
(Sricharoenchaikul et al., 1995). Main hydrocar- The water vapor can react according to the Lang-
bon product was methane which had yield 3 … 6 muir-Hinshelwood kinetics with C to form CO
w-% of carbon. Other smaller hydrocarbon spe- and H2.
cies were methanol, ethylene, acetylene, acetalde-
hyde, acetone and formaldehyde. McKeough et al., -C(s) + H2O → CO + H2 5-55
1995 noted that increasing temperature increased
carbon monoxide production. Adomeit et al. (1976) noted first the big effect of
water vapor on char combustion rate of coal char.

Figure 5-24, Contributions of different mechanisms, 4 % O2, 12 % CO2, 3 mm particle, 1200 °C (Järvinen et
al., 2001).
In recovery boilers the char combustion by water athmosphere with carbon dioxide present will
vapor gasification is one of the largest reactions increase the char combustion rate from the char
(Järvinen, 2002). From one third to one half of all bed surface. Both carbon dioxide and water vapor
char conversion occurs with water vapor. Most of can react with carbon. Conversely the addition
this vapor comes from inside the particle. of water vapor seemed to decrease somewhat the
char consumption rate. They concluded that pres-
The char combustion rate, in flight, for coal parti- ence water vapor increased CO oxidation to CO2
cles is independent of surrounding CO2 concen- and thus decreased overall reaction rate.
tration (Mitchell and Madsen, 1986).
Reaction rates of carbon at char beds are much
-C(s) + O2 → CO2 5-56 lower than the overall combustion rate (Sutinen et
al., 2002, Hupa, 2004). Probably more than 60 %
In recovery boilers char gasification by carbon of all carbon reactions occur above the char bed.
dioxide is one of the main reactions (Järvinen et This means that suspension firing is the dominant
al., 2001). firing mode.
Sutinen et al (2002) modeled experiments by

Brown et al (1998). They concluded that an

6 Dimensioning of heat transfer
surfaces
Figure 6-1, Main parts of a recovery boiler.
Recovery boiler dimensioning reflects on the design also improves and changes by time. This
properties of the design fuel, black liquor. The text aims to present some general guidelines that
evolution of recovery boiler design follows the can be followed when purchasing and projecting a
evolution of industrial boilers in general. Black recovery boiler. They can not give for all cases an
liquor is a low calorific fuel, therefore the furnace optimum answer, because of differences in fuels,
is large. Fume deposits on surfaces have total construction details and materials. So caution
emissivity about 0.5 (Wessel et al., 1998). As should be used when using the presented set of
normal boiler tubes have emissivity of 0.8 this guidelines.
makes radiative heat transfer low. Black liquor
has extremely high ash content; therefore the heat A typical example of modern recovery boiler is
transfer surfaces have been designed fouling in shown in Figure 6-1. Pressure parts are made from
mind. Black liquor is corrosive, therefore moder- steel tubing. Lower furnace part (1) is made from
ate flue gas temperatures, steam temperatures and compound tube and has start-up burner openings,
steam pressures are used. primary and secondary air ports. Above these in
the furnace (2) are liquor guns and tertiary air
As with many industrial devices exact design rules ports. Example boiler is of single drum (7) design
and especially operating margins implemented by with vertical flow boiler bank (5). Feedwater from
different manufactures vary. The recovery boiler feedwater tank (14) is fed to economizers (6) and
Dimensioning of heat transfer surfaces 6-1

Figure 6-2, Evaporation. Figure 6-3, Superheating.
further to drum. Steam from drum is superheated to steam is called evaporation or boiling. The
in panel superheaters (4). Dolezahls (13) condense whole process of evaporation from saturated water
steam, which is sprayed to control main steam to saturated steam takes place at almost constant
temperature. Superheaters are protected by screen temperature. At low heat flux conditions this
tubes (3). temperature is called the saturation temperature
and can be found from e.g. steam tables (Schmidt,
Electrostatic precipitator and hopper ash is mixed 1989, Wagner and Kruse, 1998). Most of evapora-
with black liquor in a mix tank (11). Black liquor tion takes place in the furnace below the nose.
temperature is controlled by indirect black liquor Less evaporation is done on the upper walls of the
heater (12). Air for combustion is supplied by fans furnace. Often about 10 … 15 % of evaporation is
(8) and heated by steam air heaters (9). Heated air done in a dedicated evaporator. This surface is also
is inserted to the furnace through air ports. Smelt called boiler generating bank or boiler bank.
from furnace flows to the dissolving tank (10).
Inside heat transfer coefficient for boiling in
natural circulation boilers is high. It is in the rage
6.1 MAIN HEAT TRANSFER of 30 000 … 120 000 W/m2K. This means that the
SURFACES tube average temperature is close to the saturation
temperature.
The main heat transfer surfaces in recovery boiler
are evaporative surfaces (2, 3 & 5), superheating Because of natural circulation there is no pressure
surfaces (4) and water preheating or economizing loss across an evaporating surface.
surfaces (6). Each type of surface has its own func-
tion in boiler design. The main characteristic of all
these surfaces is that they are preliminary made of
Superheating
seamless tube. Increase of steam temperature from saturated
temperature is called superheating. Superheating
There are other heat transfer surfaces that belong was first used to dry steam. Soon it was real-
to the auxiliary equipment. Dolezahl steam attem- ized that superheating allowed for higher power
perator has typically tube surface. Air preheaters production (Hills, 1989). Typically the total su-
are finned tube surfaces. Smelt spout cooling wa- perheating surface is divided into stages. Large re-
ter and dissolving tank vent circulation exchang- covery boilers have four to six superheaters. These
ers are often plate type surfaces. superheaters are arranged from two to four stages.
Water is sprayed to steam between superheating
Evaporating stages to control the final steam temperature.
In the first steam generators, phase change con- Steam density is rather low. So the inside heat
sumed most of the heat. Phase change from water transfer coefficient in superheaters tends to be low.

Therefore we need high velocities on the order of extremely low. Inside heat transfer coefficient is in
10 … 25 m/s to improve heat transfer. Typically the rage of 500 … 2 000 W/m2K.
superheater inside heat transfer is in the rage of
1 000 … 5 000 W/m2K. Pressure loss in economizers is very low because
of low water velocities. It is often so low that to
High velocities mean high pressure loss. Typically ensure even water distribution orifices need to be
the pressure loss over the superheating section is used.
more than 10 % of the main steam pressure.
Economizers 6.2 HEAT LOAD CALCULATION

The steam boiler design starts with heat load
If flue gases would leave boiler after the evapora-
calculation. In heat load calculation each type of
tive surfaces they would have temperature of
heat transfer surface is assigned a load or heat flow
more than 350 oC. Heat in these flue gases can be
based on desired total steam generation.
used to preheat water. Preheating decreases final
flue gas temperature and improves boiler heat
efficiency. Discovery of this principle in 1800s h-p diagram for water/steam
lead to greater fuel economy. Therefore the water
preheating surfaces are called economizers. h-p diagram for water, Figure 6-5 can be used to
determine operating conditions and the required
Ideally preheating is done up to saturation tem- heat loads for heat transfer surfaces. Starting
perature. In practice some margin for evapora- from chosen steam outlet pressure and tem-
tive temperature is needed to prevent boiling. If perature a line is drawn down and right towards
boiling occurs, unstable water surges can seriously desired feedwater inlet pressure and temperature.
harm boiler operation and even damage it. In Enthalpies where saturated steam and saturated
practice temperature margin of about 30 oC at water lines are crossed are marked.
MCR is often used.
Evaporation (boiling)
Water is dense and generally economizers have
high inside heat transfer coefficients. Inside heat The first estimation of heat required for trans-
transfer coefficient in recovery boiler economizers forming water to steam is the difference between
is very low, because the used water velocities are the points where the line crosses saturated steam
and saturated water lines.
As can be seen there is no phase change above

critical pressure (22.12 MPa). This means no
evaporative surfaces are needed. Water behaves
continuously in this region. In general, lower
operating pressure means larger latent heat re-
quirement for the phase change. Higher operating
pressure means that more superheat and preheat
is required.
Superheat
The first estimation for the heat required to super-
heat steam is the enthalpy difference between the
main steam outlet point and the point where the
line crosses saturated steam line, Figure 6-6.
With higher pressures the portion of heat required

for superheating increases, up to critical pressure,
Figure 6-6. Because of some 1 MPa pressure loss
in the superheaters the line angles to the right.
This pressure loss slightly increases the superheat-
ing heat requirement.
Figure 6-4, Economizing

Figure 6-5, h-p diagram for water/steam.
With supercritical cycles the temperature increase feedwater to boiling is the enthalpy difference be-
from enthalpy corresponding to critical tempera- tween the points where the line crosses saturated
ture and pressure decreases with increasing pres- water and the feedwater inlet point.
sure. As can be seen the enthalpy needed around
critical pressure forms an unstable saddle point. When pressure increases the portion of heat
Controlling final temperature is a demanding task required for economizers becomes larger.
there. This is one of the reasons why few boilers
operate at close to critical pressure. Temperature - heat input profile
Water preheat (economizer) Temperature - heat input profile can then be
formed based on heats to each part of the steam
The first estimation for heat required to bring water cycle. Figure 6-7 shows a real tempera-
Figure 6-6, Enthalpy for ideal superheat as a function of main steam values.

Figure 6-7, Temperature – heat input profile for a
natural circulation boiler.
Figure 6-8, Principle of natural circulation.
ture - heat input profile for a natural circulation Driving force

boiler. In the diagram the individual heat transfer
surfaces are arranged in the order of water/steam ∆Plosses = (ρwater-ρmixture)gh 6-1
flow.
where
The heat required for evaporation is smaller than ∆Plosses is the flow losses in the circulation
the initial estimate would give us. Economizing is ρwater is the density of the water in the
not usually completed. Part of the water preheat- downcomers
ing is then replaced by evaporative heat. ρmixture is the density of the water steam
mixture in the heated section
There are sharp edges in the superheating part. h is the height of the circulation
Between superheating sections water is sprayed to
control steam temperature. This is called desuper- Driving force is static head difference between
heating. In practice this means that 30 to 90 oC water in downcomers and steam/water mixture
superheating must be redone. This increases the in furnace tubes. Saturated water in downcom-
heat required for superheating and decreases the ers weighs more than the steam water mixture in
heat required for economizing and evaporating. raiser tubes. This creates a force which starts water
circulation. Equilibrium is established when pres-
sure losses from friction and pipe fittings equal
6.3 NATURAL CIRCULATION the driving force. For safety reasons no valves are
used in natural circulation arrangement.
Natural circulation is based on the density dif-
ferences. The same principle can be observed in Pressure increase decreases driving force because
e.g. room in a cold winter day. Heated air rises on the density difference becomes smaller. When
the wall where room is heated and subsequently driving force becomes smaller the effect of flow
cooled more dense air falls downwards on the instabilities becomes larger. Maximum pressure
opposite wall. where natural circulation boilers have operated is
17,5 MPa (Effenberger, 2000).
Natural circulation is caused by the density differ-
ence between saturated water and heated water
partially filled with steam bubbles. In a natural
circulation unit water tubes are connected to a Optimization of natural
loop, Figure 6-8. Heat is applied to one leg called circulation design
raiser tubes, where water steam mixture flows
upward. Denser saturated water flows downward Natural circulation design is affected by several
in unheated leg called downcomer. design factors. If we increase furnace height the
driving head increases. Natural circulation is also

Figure 6-9, Natural circulation arrangement of modern recov-
ery boiler, 1 - water flows down, 2 - steam water flows up water
walls, 3 - steam water mixture flows back with risers, 4 - boiler
bank has separate downcomers and risers.
Figure 6-10, Optimization of natural circulation.

improved by higher heat flux in lower part of the temperature, decreases superheater fouling but
tubes. This is one of the reasons that in some older costs more (MacCallum, 1982). In practice fur-
designs boiler bank had upwards flow. Key design nace sizing is based on parameters like Effective
success factor is trying to balance flows in all parts projected surface area (EPRS), Hearth heat release
of the boiler. This is done by careful arrangement rate (HHRR) and Hearth solids loading (HSL)
of distribution piping and high enough flow veloc-
ity at MCR. Furnace floor area = width*depth
Cross sectional area = width*depth
We must try to avoid steam in downcomers. Furnace volume = width*depth*height
HSL = ds firing rate / floor area
Efficient steam separators reduce the fraction of
Heat input = ds firing rate * HHV
steam inside the drum. Inserting feedwater to HHRR = heat input / floor area
drum at subsaturated state cools down and col- Volumetric input = heat input / volume
lapses remaining bubbles. Minimized axial flow EPRS = projected wall area
inside drum helps creating equal flow in all parts excluding furnace floor
of boilers. Heat flux = heat transferred / EPRS
In addition furnace performance can be looked at

6.4 KEY FURNACE SIZING by furnace residence time and furnace gas average
CHARACTERISTICS velocity. Typical design parameters for modern
recovery boiler furnaces are shown in Table 6-1.
Recovery boiler furnace needs to be sized so that In boilers with older air systems and low dry
combustion has time to be completed for low solids somewhat more conservative values need to
emissions. Large furnace lowers furnace outlet be used.
Table 6-1, Typical values for recovery boiler fur- We know that at least three things can limit the
nace design parameters, MCR load. furnace capacity (Adams, 1994). If the flue gas
velocity is excessive, then the carryover that accu-
Parameter Value mulates on heating surfaces can plug gas passages.
Heat release rate kW/m 2
2700 .... 3300 If the char bed combustion rate limit is reached
Furnace outlet temperature C
o
950 ... 1000 then the char bed growth can be uncontrollable.
If there is poor air jet mixing then the emissions
Generating bank inlet C
o
550 ... 650
from the boiler can be unacceptable.
Table 6-2, Effect of dry solids on maximum
Major changes have been made to the recovery
furnace capacity, if furnace upwards velocity limits
boiler. Current recovery boilers have improved
capacity.
air systems and so more combustion capacity.
Dry solids HHRR Maximum Velocity Minimum The as fired black liquor is in new boilers close to
% kW/m2 capacity m/s capacity
80 % dry solids concentration. This increases the
65 3 150 100 % 4.5 60 % of max furnace temperatures and the rate of dry solids
75 3 750 120 % 4.5 50 % of max burning. New liquor firing practices have in-
90 4 100 130 % 4.5 40 % of max creased the fraction of black liquor that is burned
in suspension. Finally, a higher char bed tempera-
ture is seen. All this has significantly increased the
combustion rate. In the following sections we limit
the discussion to the modern recovery boilers.
Carryover
The most dramatic effect on furnace capacity has
been from the increase in dry solids concentra-
tion. Furnace capacity can be limited by carryover.
In such a case the average velocity in the lower
furnace gives an indication of maximum firing
rate. An example of the effect of dry solids can
be seen in Table 6-2. The increase in dry solids
from 65 to 90 % would increase the dry solids
firing capacity from 3150 to 4100 kW/m2 or 30 %.
Figure 6-11, Two different size boilers with same
So even with somewhat increased temperatures
residence time.
the required furnace cross section reduces with
Figure 6-12, Effect of boiler MCR capacity on furnace relative heights.
increased dry solids if we design for a constant If we assume that the design heat flux to the
lower furnace velocity. This example illustrates furnace walls is constant, then for same hearth
why carryover because of an excessive lower fur- loading the relative height should be constant. On
nace velocity should not limit the performance of the other hand, the larger the recovery boiler, the
high dry solids recovery boilers. lower is the relative height needed for complete
combustion. This is because doubling the width
Furnace height quadruples the capacity. If we also double the
height then the furnace volume increases eight
Furnace height must provide for the time needed times. Then there would be two times more time
for complete combustion. Furnace height depends to complete the combustion. Larger boilers have
on the required residence time. From industry ex- consequently more time to finish combustion.
perience it can be determined that approximately
4.5 seconds is required. If we apply this rule to Mixing of air jets
some typical recovery boiler hearth release rates
we find that the minimum height corresponding The penetration depth of a single air jet into the
to a 4.5 s residence time increases with increased furnace is given by Adams and Frederick (1988)
dry solids, even if the applied hearth heat release
rate is higher, Table 6-3.
6-2
In practice this would mean that two boilers of
same specific loading, but widely different capac-
where
ity would have an equal height, Figure 6-11.
Lp is the jet penetration length, m
k is an experimental constant, ~1.5
Does the 4.5-second rule limit recovery boiler ca-
Dn is the hydraulic diameter of air port
pacity? We can plot actual recovery boiler capacity
opening, m
versus their height and the height required for
Vn is the jet velocity, m/s
complete combustion at MCR load, Figure 6-12.
Vf is the furnace gas upward velocity, m/s
For nearly all recovery boilers the furnace height
Tn is the jet gas temperature, K
seems to be about three times the furnace width.
Tf is the furnace gas temperature, K
Table 6-3, Furnace height required to ensure a
4.5 s residence time. Evaluating the equation we find that most of
Dry solids HHRR Minimum height the parameters are independent of boiler size.
% kW/m2 m The ratio of jet velocity to furnace velocity has
65 2800 17
increased since modern air systems use higher
velocities (pressures) than before. But the design
75 3300 19 velocities are approximately the same independ-
85 3800 22 ent of furnace size. Similarly the ratio of furnace
90 4000 25 temperature to jet temperature is approximately
constant.
Figure 6-13, Effect of dry solids capacity on HHRR (redrawn after McCann, 1991).
One parameter that changes with boiler size is the A major dimensioning criterion for recovery
hydraulic diameter of air port opening. If we de- boilers is the furnace floor area. Typical hearth
sign large and small furnaces with the same hearth solids loading is close to 0.2 kgds/sm2 (McCann,
loading, then we need the same amount of air per 1991), Figure 6-13. Hearth solids loading may
unit of furnace bottom area. The airflow require- be misleading as it does not take into account
ment per unit of wall increases linearly with wall the variation in liquor properties, especially the
width. If we keep the number of airports constant heating value. A better indicator is the hearth heat
then doubling the wall width would double each release rate.
air port area. The hydraulic diameter of the air
port opening increases as the square root of the Processes in recovery boiler are not limited by the
wall width increase. furnace volume. After furnace bottom area has
been chosen the furnace height is determined to
In summary, based on the above equation the correspond to desired furnace exit temperature.
single jet penetration depth would increase slower
than the furnace size, so at some capacity the mix-
ing would not be enough. Analyses of modern air
Furnace outlet temperature
systems indicate that currently the jet penetration The adiabatic combustion temperature, Figure 6-
is high enough even for the largest boilers (Adams 14, represents the maximum temperature achiev-
and Frederick, 1992, Vakkilainen, 2000). able by black liquor combustion. It is easy to see
that a minimum practical operating dry solid is 55
Furnace loading %, where adiabatic combustion temperature is still
above 1000 oC.
There is very little data on the effect of recovery
boiler size on its performance. McCann (1991) In recovery furnace the combustion black liquor
states that the recovery boiler hearth heat release radiates heat to furnace walls, so maximum
rate, HHRR increases with boiler size. Typical temperatures are about a third lower than adi-
MCR furnace design parameters for new boilers abatic temperatures. Char bed operation requires
are 2700 … 3300 kW/m2 (La Fond et al., 1993). temperatures in excess of 900 oC. Typical design
If a boiler is well tuned, operates with high dry furnace outlet temperatures are 950 ... 1050 oC,
solids black liquor and is equipped with mod- but in highly loaded furnaces they can exceed this
ern air system, the maximum operating rate is by several hundred degrees.
between 3100 and 3700 kW/m2 (Burton, 2000).
The recommended values have increased as design Recovery boiler furnace exit temperature depends
has improved. Barynin and Dickinson (1985) strongly on black liquor dry solids and heating
recommended 2680 … 2840 kW/m2 and super- value, Figure 6-15. So low MCR loading means
heater inlet temperature below 925 oC. Walsh and lower furnace and high MCR loading requires
Strandell (1992) noted 2500 … 3200 kW/m2 and higher furnace for the same furnace outlet tem-
superheater inlet temperature 925 … 950 oC. perature. Furnace outlet temperature affects the

Figure 6-14, Adiabatic combustion temperatures of typical kraft liquors at vari-
ous concentrations.
Figure 6-15, Effect of furnace loading on furnace exit temperature.
Figure 6-16, Effect of dry solids on furnace outlet temperature (Vak-

kilainen, 1994).

superheater size. Low furnace outlet temperatures The net heat input per unit area of furnace floor
mean large and expensive superheating surface. increases. The average flue gas flow decreases with
On the other hand high furnace outlet tempera- less water. This lowers flue gas velocities in furnace
tures mean high heat fluxes to first superheaters by less than 0.5 m/s. The main impact with evolv-
and consequently higher corrosion rates and more ing design is the increase in heat transfer surface
maintenance. required after the furnace. Changing liquor
properties require more heat to be transferred
The choice of furnace outlet temperature affects especially in the superheating section.
the ash characteristics. Control of ash behavior
and corrosion at superheaters is a key design The processes and reactions taking place on the
parameter. Key sizing parameters are first ash surface and just above the char bed have the great-
melting temperature, ash softening temperature est impact on boiler operation and performance.
and ash flowing temperature (Hupa et al., 1990). About 90% of the total combustion air (equal to
all theoretical air required) is introduced at or just
Furnace outlet temperature affects the dimension- above the char bed to maintain a high temperature
ing of the rest of the heat transfer surfaces and has in the lower furnace. This high temperature leads
a major effect on emissions. It is often practical to low emissions and high reduction efficiency.
to find the proper value through several itera- With increased dry solids concentration the main
tive rounds. Calculating furnace height requires problem is keeping enough char bed for proper
determination of where heat release takes place. operation, not keeping it hot to achieve low sulfur
This depends on the air splits and liquor firing emissions and high reduction.
strategy. Furnace heat flux profiles differ from one
boiler manufacturer to another. Typically furnace Example
heights correspond to average heat flux of 60 ...
100 kW/m2. Estimate the required furnace dimensions for the
3000 tds/d recovery boiler. The black liquor dry
One way to look at changes in furnace design is solids is 70 %.
to examine the different design basis for the three
recovery boilers discussed earlier, Table 6-4. The Dry solids firing rate is 3000 / 86.4 = 34.72 kgds/s.
nominal furnace loading has increased and will If we choose the HSL of 0.2 kgds/(sm2), then the
continue to increase. Improved air designs to furnace floor area is 34.72 kgds/s / 0.2 kgds/(sm2)
increase firing rates have been introduced by all = 173.6 m2. This corresponds to 13.2m*13.2m
major recovery boiler manufacturers. A major fac- square.
tor in helping these innovations has been the use
of new three dimensional computational methods The black liquor higher heating value is 15 MJ/
for furnace calculation. Raise of black liquor dry kgds. Heat input is then 15*34.72 = 520.8 MW.
solids from 65% to 80% has increased the furnace (Note! In addition to HHV heat is added with air
temperatures. Increase in HSL has been about 30 preheat etc.) The HHRR is then 520800/173.6 =
%, but increase in HHRR has been some 10 %. 3000 kW/m2.
Table 6-4, Design basis for the three recovery boilers (For design liquors see Table 2-1).
Two drum Modern Current
Design dry solids capacity, ton dry solids/day 1700 2400 4000
Design black liquor HHV, MJ/kg dry solid 15.0 13.9 13.0
Design dry solids, % 64 74 80
Design steam temperature, C o
480 480 520
Design steam pressure, bar(absolute) 85 85 104
Steam production, kg steam/kg dry solids 3.584 3.357 3.004
Specific steam production*, kg steam/MJ 0.238 0.242 0.231
Furnace design HSL, ton dry solids/m2 day 15.6 17.1 20.1
Furnace design HHRR, kW/m 2
2715 2760 3031
Superheater design area, m2 day/ton dry solids 6.0 5.2 6.8
Boiler bank design area, m day/ton dry solids
2
3.5 3.0 3.5
Economizers design area, m2 day/ton dry solids 7.1 10.1 9.1
* steam production per unit of black liquor divided by the HHV of unit of black liquor

Figure 6-17, Effect of steam velocity on difference between recovery boiler superheater steam and outer tube
temperatures.
The area chosen assures that the boiler is of aver- are small. Temperature differences between
age HSL load. However the HHRR is high. It is in adjacent tubes can be somewhat controlled by
fact higher than what is typical for new boilers. changing tube lengths between passes. Outermost
Heat available in the furnace is (13986-1861-290) tube which receives the most of radiative heat
kJ/kgds*34.72 kgds/s = 410.9 MW flux should be shorter than the rest of the tubes.
Proper superheater arrangement also eliminates
If we choose the furnace exit temperature on the much of the problems with uneven or biased flue
low side, 936 oC, then with flue gas flow of 209.6 gas flow.
kg/s and a cp value of 1.355, the heat transferred
in the furnace is 410900 - 209.6*1.355*936 = Superheater tube does not transfer heat equally
145.1 MW. If the furnace height corresponds to around its circumference. Much of the heat
heat flux 80 kW/m2, then the effective projected transfer is blocked by adjacent tubes and deposits
area is 1800 m2. The furnace height is 1800/ between tubes. Heat transfer is decreased to about
(13.2+13.2+13.2+13.2) = 34.3 m. the half of the value for clean tubes (Roos, 1968).
It is best to use projected heta transfer surface
when comparing superheater area. Projected
6.5 SUPERHEATER SECTION heat transfer surface is calculated assuming the su-
DIMENSIONING perheater side surfaces are flat. For a superheater
element it then equals to element height multi-
The superheater surfaces are dimensioned based plied by the sum of two times element width and
on radiative heat transfer and ash properties. Su- one tube diameter. Use of projected heat transfer
perheater dimensioning plays major role in boiler area instead of true tube area is further based on
cost. Superheater is one of the most expensive heat the fact that most of the heat transfer is radiative,
transfer surfaces. It is usually made from alloyed which is proportional to projected area.
steels. Heat transfer in superheaters is mainly
radiative, but in the primary superheaters the One of the most difficult tasks is to account for
convection often plays major role. uneven temperature and flow fields in the super-
heater area. An example of upper furnace CFD
To achieve good steam temperature control, the calculation can be seen as Figure 6-18. Size and
last superheater can not have too large tempera- placement of bullnose affects flow (Vakkilainen et
ture rise. Often last superheater section transfers al., 1992). Typically the superheater needs to be
15 … 33 % of total superheating heat requirement. reduced to a series of heat transfer surfaces, which
Placement of superheaters is important. Super- are then solved separately.
heater tubes that receive furnace radiation run
much hotter than protected tubes, Figure 6-17. Superheater size
Superheater design should also be such that Typically the total superheating area is divided
temperature differences between adjacent tubes into sections for manufacturing reasons, for

Figure 6-18, Left - temperature profile in a superheater, right – flow profile in the same superheater (Vakki-
lainen, 2000)
effective sootblowing and to enable desuperheat-

ing. Lifting and handling a very large number of
tubes during erection is costly. In most cases two
desuperheating points are needed. It is custom-
ary to call superheaters with roman numerals. All
superheating sections have roman numeral I from
drum up to first desuperheating, roman numeral
II from first desuperheating to second desuper-
heating etc. The superheating sections in between
Figure 6-19, Effect of furnace size and superheater
are numbered alphabetically at direction of steam
outlet temperature on specific superheater projected
flow, IA, IB etc.
area.
The amount of superheater surface required de-
pends on the furnace size, steam outlet tempera-
ture and desired range of superheating, Figure
6-19. Typically recovery boiler superheaters have
specific projected heat transfer surface about 3
m2/(tds/d).
Effect of load to superheating

Decreasing load (as fired black liquor dry solids
rate) decreases flue gas flow and temperatures to
Figure 6-20, Effect of load to temperatures (Adams
and from the superheater, 6-20. Thus the super-
and Frederick, 1988, Haynes et al., 1988)
heating reduces because both the heat transfer co-
efficient and the available temperature difference
decrease. Usually this means that at some 70 % of
MCR load the recovery boiler does not superheat
enough to reach MCR main steam temperature.
Superheater spacing
Typical superheater spacing can be seen in Figure
6-21. For low temperature, nonfouling zone tight
spacing can be used. For high temperature, where
fouling is probable, wide spacing is used.
Figure 6-21, Superheater spacing (Singer, 1981)

Recovery boiler screens and superheater area have
typically spacing from 300 to 400 mm. Wider
spacing is less resistant to fouling, but leads to
more expensive boiler. Recovery boiler vertical
flow boiler banks have spacing from 150 to 250
mm. Economizers have spacing from 100 to 150
mm.
Panel superheater
When tubes in an inline superheater are placed
close to each other in gas flow direction a panel or
platen superheater is created, Figure 6-22. Panel
superheaters were introduced to use in 1952 by
Combustion Engineering.
In panel superheaters the tube longitudinal spac-

ing is smaller than 1.25 *Ød. Panel superheater is
resistant to fouling and can withstand high heat
flux. It is used in most demanding applications,
typically as the first gas side superheater.
Arrangement of parallel tubes

Figure 6-22, Panel superheater Often to limit the steam side pressure loss in a
superheater one must use several parallel tubes.
Each of these tubes has for construction reasons
different length, different total heat absorbed and
necessarily somewhat different outlet temperature.
For corrosion reasons one often uses arrangement

that results in shorter length of the outermost
tube, Figure 6-23. First, outermost or edge tube re-
ceives more radiation heat especially from furnace
than the other tubes. Also often the corrosive ash
has the most difficult properties in the first tube.
Shorter flow path helps keep first tube cooler than
the others and decreases the first tube corrosion,
but increases corrosion of the rest of the tubes.
Superheater hanging
To limit the amount of ash that can escape
Figure 6-23, Cooler edge tubes. through superheater openings in the boiler roof,
protective boxes are used, Figure 6-24. They seal
the opening to the roof to form gas tight construc-
tion.
Superheaters need to be supported, so they hang

from these boxes. Hanger rods connect boxes
to secondary beams. This arrangement helps to
control the stresses caused by tube heating and
cooling.
Figure 6-24, Superheater hanging.

6.6 VERTICAL BOILER BANK
DESIGN
Most boiler banks are of vertical design, Figure
6-25. Earlier generating banks had common up-
per and lower headers. Modern boiler banks are
formed from finned tube panels. Panels are then
connected to common headers. This stiffens the
construction and limits tube movement, which
was problem in the earlier non finned designs.
Ample space is reserved for sootblowing. All

major recovery boiler manufacturers use straight
sootblowing space. Some place flow distribution
plates to this space to improve heat transfer. Flow
in the space is hotter and faster that flow in the
rest of the surface, which decreases heat transfer
(Vakkilainen and Anttonen, 1994).
B&W reports that they have increased the cross

flow spacing from original 152.4 mm first to 203.2
mm and then to 228.6 mm to improve fouling be-
havior (Dickinson et al., 2000). Minimum spacing
has varied a lot but new boilers are bought with
spacing of 164 ... 230 mm.
Figure 6-25, Vertical boiler bank.
Typically boiler generating bank inlet tempera-
ture is 540 ... 600 oC. Maximum flue gas velocity
should be under 20 m/s (Recovery, 1993). Typical
design flue gas mass flow densities at MCR are 6
… 8 kg/sm2. Special emphasis should be placed
on achieving low inlet velocities. Normally in
modern boilers the free flue gas flow area is 0.17
… 0.23 m2/m2bottom and the finned heat transfer
surface area is 45 … 70 m2/m2bottom.
Boiler bank inlet temperature is one of the key

design factors. High inlet temperature means that
dust entering boiler bank is still considerably mol-
ten. Low inlet temperature mens heavy investment
to superheating surface. Barynin and Dickinson
(1985) recommend maximum entering gas tem-
perature 625 … 700 oC. McCann (1991) recom-
mends 565 … 625 oC if Cl and K are present.
6.7 VERTICAL ECONOMIZERS

Most economizers are of vertical design, Figure
6-26. They are formed from finned tube panels
connected to common headers. Special openings
are reserved for sootblowing. In modern large,
high dry solids boilers the sootblowing require-
ment for economizers is very minimal.
Typically economizer inlet temperature is 350

... 450 oC. Maximum flue gas velocity should
Figure 6-26, Vertical flow economizer. be under 30 m/s. The design flue gas mass flow
density at MCR are 8 … 10 kg/sm2. The free flue
gas flow area is typically 0.13 … 0.17 m2/m2bottom arrangement, etc.), -
for economizer 2 and 0.11 … 0.16 m2/m2bottom fo is the overall correction (fouling correction,
for economizer 1. The total finned heat transfer etc.), -
surface area in both economizers is 150 … 230 ki is the inside heat transfer coefficient
m2/m2bottom. Minimum spacing has varied a lot but referred to the outside surface, W/m2K
new boilers are bought with economizer spacing ko is the outside heat transfer coefficient,
of 110 ... 180 mm. Inlet velocity is important. High W/m2K
inlet velocity leads to plugging. Sandquist (1987) kr is the radiative heat transfer coefficient,
reports that reduction of inlet velocity by 30 % W/m2K
solved economizer plugging at Leaf river mill. kex is the external heat transfer arranged to re-
Barynin and Dickinson (1985) recommend using present a heat transfer coefficient, W/m2K
MCR design velocities lower than 11 m/s and kc is the convective heat transfer
entering gas temperature below 425 oC. coefficient, W/m2K
λ is the heat conductivity of tube material
Economizers with low design flue gas mass flow do is the outside tube diameter, m
densities and wide spacings are less prone to ds is the inside tube diameter, m
fouling. Low flue gas velocities lead to low heat s is the tube wall thickness, m
transfer rates and high heat transfer areas.
The correction for number of rows, fn can be
expressed as
6.8 HEAT TRANSFER IN
BOILERS
6-5
The purpose of this section is to give the reader an
overview of some possible methods to calculate
heat transfer in boilers. The treatment is short and
the reader is requested to study other material The correction for overall fouling, temperature
if he is interested in heat transfer phenomena in deviations, flow deviations etc. fo, is usually
detail (Blokh, 1988, Brandt, 1985, Rahtu, 1990, dependent on fuel type and properties, steam gen-
Brandt, 1999), erator configuration, type of surface and expected
calculation error. The exact numerical value must
so be determined empirically from previous
Overall heat transfer experimental data.
Overall heat transfer in a boiler heat transfer
surface can be expressed with the general heat The heat transfer resistance through a tube of
transfer equation as uniform material for a tubular construction where
tubes are separated from each other is
Ø = kA∆T 6-3
6-6
where
Ø is the heat transferred, W
k is the overall heat transfer
coefficient, W/m2K Radiation heat transfer
A is the heat transfer surface, m2
∆T is the temperature difference, K Radiation heat transfer can be calculated accord-
ing to the formulas specified in VDI – heat atlas,
The overall heat transfer coefficient k is a function Kc1-Kc12 (1993). Even though the concept of
of convective and radiative heat transfer. radiation heat transfer coefficient has only a very
weak connection to physical reality, it is used here
so that the radiative heat transfer can be expressed
with the convective heat transfer.
6-4
Radiation heat transfer coefficient can be deter-
mined through radiation heat flow.
where
6-7
fn is the form correction (correction for
number of rows, correction for

Radiation heat flow Φr can be expressed if tem- εCO2 = αo + α1µ + α2µ2 + α3µ3 6-18
peratures and emissivities are known as
µ = µ(Tg) 6-19
6-8
6-20
αi = αi(pCO2,s) 6-21
where
εw is the emissivity of the wall, for fully
oxidized boiler tubes 0.8 is typical 6-22
εdg is the emissivity of the dusty gas
αdg is the absorptivity of the dusty gas where
εa is the background emissivity, -
For dusty gas we can express the emissivity and
absorptivity assuming no band overlapping, For dust emissivity εd a fully dispersed approach
through pure gas and pure dust emissivities is used
εdg = εg + εd - εdεg 6-9
6-23
αdg = αw + εd - αwεd 6-10
where
The emissivities and absorptivities of pure gas αd is the absorption coefficient of dust
mixtures can be calculated with pure gas proper- Bd is the dust loading, kg/m3
ties s is the radiation beam length, m
εg = εH2O(Tg,spH2O)+ εCO2(Tg,spCO2)+ The exact expression for radiation beam length s is

εSO2 (Tg,spSO2)- ∆εg(εH2O,εCO2,εSO2) 6-11 very complicated and involves integrating in space
of surface element. Usually graphical approxima-
αdg = αH2O + αCO2+ αSO2 - ∆αg 6-12 tions can be found to all relevant problems.
where Kraft recovery boiler deposits have low emissivity,

∆εg is the overlapping correction for emissivity, . Figure 6-27 (Wessel et al., 1998, Sinquefield et al.,
∆αg is the overlapping correction for 1998, Baxter et al., 2004). Care should be taken to
absorptivity, - obtain true surface properties.
px is the partial pressure of substance x, MPa
s is the radiation beam length, m
Outside convection heat transfer
As no better approximations have been developed Convection heat transfer is calculated by e.g. Gn-
we must express absorptivities through emissivity ielinski equation VDI – heat atlas, Gf1-Gf3 (1993)
functions, applicable when total gas side pressure
is about 0.1 MPa.
6-24
6-13
The correction factor fA for outside heat transfer
is an empirical coefficient. It is dependent of fuel
ptot = pH2O + PCO2 + pSO2 6-14 type and properties, steam generator configura-
tion, type of surface and expected calculation
6-15 error. The exact numerical value must so be
determined empirically from previous experimen-
tal data.
6-16
Gas side pressure drop, inline
6-17 Gas side pressure drop can be calculated accord-
ing to the Gaddis equation in VDI – heat atlas,
Ld1-Ld7 (1993).
The individual emissivity functions can be math-
ematically approximated as
∆pgs = ∆pgf 6-25

Figure 6-27, Effect of deposit thickness on emissivity, fume deposited on top of tube (Sinquefield et al., 1998).
gas side pressure losses = friction losses Inside heat transfer, fluid inside
For in line arrangement the pressure drop coef-
tube
ficient for heat transfer surface with horizontal Turbulent convection heat transfer for steam and
tubes is water flowing inside an circular tube can be cal-
culated according to the Hausen equation; 2300 <
∆p = nrζr∆pd 6-26 Re < 107, 0.6 < Pr < 500, d << L, Tw ~ Tf (Hausen,
1974)
where
nr is the number of rows in heat transfer
surface 6-30
∆pd is the dynamic pressure, calculated at the
gas side mean temperature and smallest Nui = 0.0235(Re0.8 - 230)(1.8Pr0.3 - 0.8) 6-31
area
For laminar flow case, Re < 2300 a value of
The single row pressure drop ζr for in line tubes is
sum of laminar and turbulent coefficients Nui = 3.64 6-32
can be used, Hausen (1974)

6-27
Usually no correction for inside fouling is used,
6-28 as tube inside is kept relatively clean by normal
boiler operation.
6.9 EXAMPLE CALCULATION

6-29 OF HEAT TRANSFER SURFACE
We have a single heat transfer surface arrange-
where ment, Figure 6-28, consisting of main heat transfer
ζl is the laminar part of the pressure drop surface and side walls situated transverse to the
coefficient, - gas flow. We can make the following simplifica-
ζt is the turbulent part of the pressure drop tions applicable to most of the cases in practice
coefficient, - - dust is dispersed uniformly to the gas flow
sl is the dimensionless transverse pitch, - temperatures of the dust Tdi and Tdo are
st is the dimensionless longitudinal pitch, equal to the respective temperatures of the
Re is the Reynolds number, calculated at the gas flow Tgi and Tgo
gas side mean temperature and smallest - temperature of the side walls Tsw can be
area, - treated as constant i.e. Tgi - Tgo >>Tswo - Tswi

Figure 6-28, Recovery boiler superheater arrangement
θln = θln(Tgi,Tgo,Tsi,Tso) 6-35

- no combustion inside the heat transfer To aid us in heat transfer calculations, when solv-
surface volume ing unknown temperatures for known area and
- heat losses Φl and external heat flow Φex can overall heat transfer coefficient we use nondimen-
be determined separately from the sional heat transfer parameters ε, R and z (Ryti,
superheater heat transfer calculations 1969).
We want to determine the heat transfer from the εln = εln(Tgi,Tgo,Tsi,Tso) 6-36
flue gas flow to the steam flow. The basic equation
involved is Rln = Rln(Tgi,Tgo,Tsi,Tso) 6-37
6-33 zln = zln(Tgi,Tgo,Tsi,Tso) 6-38
For the case of superheaters we usually determine We do this by starting from the definition of the
an overall heat transfer coefficient kt which is heat flow to the steam.
treated as an independent value. The temperature
distribution as reference to area can be solved for Φs = qmscps(Tso - Tsi) 6-39
simple parallel and counterflow cases. Usually we
express the solved equation in basic form. For the simple no external heat flow case the heat
flow to the steam equals the heat flow from the
flue gas
6-34
Φg = qmgcpg(Tgi - Tgo) = Φs 6-40
heat to steam flow = overall heat transfer coef-
ficient * logarithmic temperature difference * Then the parameters describing the heat transfer
effective heat transfer surface can be defined
where G = ktAeff 6-41

Cmax = Max(qmgcpg,qmscps) 6-42 6-58
Cmin = Min(qmgcpg,qmscps) 6-43

z = G/cpgqmg 6-59
∆Tmax = Max(Tgi - Tgo,Tso - Tsi) 6-44
It must be noted that above is valid only for the
∆Tmin = Min(Tgi - Tgo,Tso - Tsi) 6-45 case when flue gas heats the steam.
θ = ∆Tmax - ∆Tmin 6-46 Heat balance

ε = ∆Tmax/θ 6-47 For the superheater the first principle must hold
true
R = ∆Tmin/∆Tmax 6-48
Φg - Φs - Φsw + Φex - Φl = 0 6-60
z = G/Cmin 6-49
heat from gas - heat to steam - heat to sidewalls +
Then we can express the heat flow to steam also external heat - heat losses = 0
with the dimensionless parameters where
Φs = Φs(ε,R,z) = qmscps(Tso - Tsi) 6-50 Φs = Gθln = qmscps(Tso - Tsi) 6-61
For the complex heat transfer situation occurring Φg = (qmgcpg + qmdcpd)(Tgi - Tgo) 6-62
in the recovery boiler superheaters, where there is
also several additional heat flows, we can redefine Φsw = kswθswAsw 6-63
the nondimensional heat transfer parameters with
aid of correction factor b to be able to treat this External heat Φex and heat losses Φl are usually
multi heat flow case as single heat flow case in the constant, fixed values.
heat transfer calculations (Vakkilainen, 1994)
G = ktAeff 6-51
Heat capacities
To make use of the dimensionless parameters ε,
θ = Tgi - Tsi 6-52 R and z possible we must refer the heat transfer
situation to the simple two massflows exchange
6-53 heat case with the aid of correction factor b. We
have chosen the steam side heat flow to be the
determining heat flow. Then for recovery boiler
parallel flow
superheaters
6-54 Cmin = qmgcpg 6-64
C*min = qmscps 6-65
6-55 C*min = bqmgcpg 6-66
R = C*min/C*min 6-67
countercurrent flow
6-56
6-68
6.10 EFFECT OF HIGH DRY

6-57 SOLIDS TO RECOVERY BOILER
DIMENSIONING
The lower furnace plays a key role in terms of
lowering emissions and raising capacity (Borg
Table 6-5, Furnace operating conditions for example boiler, constant HHRR.
Sum of heat inputs, MW 559.4 557.9 556.7 555.6 554.7 553.8
Heat in furnace, MW 393.0 402.6 410.9 418.1 424.4 430.0
cp gas, kJ/kgoC 1.391 1.372 1.355 1.340 1.324 1.314
Liquor flow, kg/s 57.9 53.4 49.6 46.3 43.4 40.9
Flue gas flow, kg/s 221.0 216.9 212.8 209.5 206.7 204.2
Furnace outlet temp., oC 913 926 936 943 949 953
Firing capacity, kg dry sol- 0.2 0.2 0.2 0.2 0.2 0.2
ids/m2s
HHRR, MW/m2 3.0 3.0 3.0 3.0 3.0 3.0
Heat release rate, MW/m2 2.26 2.32 2.37 2.41 2.44 2.48
et al., 1974). The key is to control the reactions furnace loading would be higher with the 65 %
taking place in the lowest part of the furnace, dry solids content black liquor than with the 80
on the bed surface and just above it. Increasing % dry solids black liquor. Higher furnace loading
black liquor dry solids will increase lower furnace indicators include hearth heat release rate, solids
temperatures (Backman et al., 1995, Ryham and loading, total heat input and net heat input. This
Nikkanen, 1992, Haynes et al., 1988). Temperature is because the flue gas losses are high for the 65 %
will have a significant impact on emissions, reduc- liquor.
tion efficiency and on the fouling and corrosion of
the convective heat surfaces. Liquor heat treatment reduces the furnace load-
ing. The net heat input remains almost constant.
The air system introduces most of the combustion The liquor dry solids flow decreases, as does the
air into the lower furnace, just above the furnace black liquor volumetric flow. The average flue gas
floor. As the dry solids increases, the penetra- flow falls more than 10 %. The hearth heat release
tion of air jets increases. With increasing dry rate falls by 11 %, and the nominal firing capacity
solids and improved design, emissions have been by 8 %. Because the higher black liquor heating
reduced significantly. The nominal furnace load- value and the dry solids firing rate decrease, the
ing changes only slightly with an increase in black net heat release rate remains constant. Even by a
liquor dry solids content. The furnace loading is conservative estimate, the furnace bottom area
usually determined by either the hearth heat re- could be reduced by 4-8 %.
lease rate (HHRR) or the firing capacity. Because
the higher black liquor heating value and the dry Furnace temperature profiles have been pub-
solids rate remain constant, these indicators show lished by Jones (1989) and Jutila et al (1978). A
constant loading. The hearth heat release rate in heat release profile can be found through fitting
the furnace remains constant. to achieve a known temperature profile. Making
correct furnace outlet temperature predictions
As we raise dry solids from 65 % to 80 %, the total require fine tuning to available measurements.
heat input per plan area is reduced. This is caused
by a reduction in the black liquor sensible heat In typical furnace temperature profiles the tem-
input. The net heat input to the furnace increases perature reaches a maximum in the lower furnace
noticeably, as less water needs to be evaporated. ant then starts dropping toward the nose. The
The average flue gas flow decreases as less water lower furnace temperature increases about 100 oC
vapor is present. If the furnace temperature with an increase in dry solids from 65 % to 80 %.
remained constant, this would reduce the flue gas This is significantly less than the 200 oC change in
flow rates, but by less than 0.5 m/s. adiabatic flame temperature. On the other hand,
the radiative heat flux to the walls increases by
Loading values for the example 3000 tds/d recov- about 20 %. Increased lower furnace temperatures
ery boiler for a furnace of similar size are listed in can speed up local corrosion in recovery boilers.
Table 6-5. Steam generation with higher dry solids
increases. The steam generation with 80 % dry The furnace outlet temperatures increase as the
solids equals the steam generation at 106.9 % flow dry solids content rises from 65 % to 80 %. The
of the 65 % dry solids content black liquor. measured increase is 20-35 oC. It must be borne
in mind that the measurement error of furnace
If we were to fire at the same steam flow, the outlet temperatures is at least +20 oC. The increase
Table 6-6, Heat available for superheating.
Flue gas flow, kg/s 221.0 216.9 212.8 209.5 206.7 204.2
Furnace exit
temperature, oC 913 926 936 943 949 953
Cp gas, kJ/kg Co
1.391 1.372 1.355 1.340 1.324 1.314
Steam flow, kg/s 122.8 126.6 129.9 132.8 135.3 137.5
Heat for superheat, MW 109.9 110.4 109.7 108.8 107.5 106.4
Heat for superheat, kW/kg 894.6 871.3 844.0 819.2 794.3 773.6
If furnace outlet temperature
constant = 926 oC
Heat for superheat, MW 113.9 110.4 106.8 103.9 101.2 99.1
Heat for superheat, kW/kg 927.5 871.3 821.7 782.5 748.1 721.2
in furnace outlet temperatures is caused by If the boiler bank inlet temperature is reduced, the
changes in the mass flow and the net heat input to average temperature difference between the flue
the furnace. gas and the steam is also reduced. This reduces
heat transfer and we need substantially more
The calculated furnace outlet temperature varia- superheating surface. Heat treatment changes the
tions for a fixed-size furnace are shown in Table heat available for superheating, Table 6-6. This
6-5. Understandably, liquor heat treatment lowers behavior is similar to the decrease in superheating
the furnace exit temperature. The change in with falling boiler load.
furnace exit temperature will show up, albeit with
reduced magnitude, in the flue gas temperature Mixed measurements on the effect of high dry
entering the boiler bank. solids on superheating have been presented. Jones
has recorded a significant superheating tempera-
The basic dimensioning of recovery boiler super- ture drop with the Arkansas Kraft boiler (Jones
heaters will not change much in response to the and Anderson, 1992). The Metsä-Sellu Äänekoski
change in black liquor dry solids content. As these boiler does not show such behavior (Ryham and
surfaces are radiative surfaces, their heat transfer Nikkanen, 1992). One possible reason for this
does not depend on flue gas velocity. The radia- is the smaller amount of fouling of heat transfer
tive heat transfer at the superheaters increases by surfaces.
10-20% along with the increase in black liquor
dry solids. The heat flow increase will not require Example
changes in the design of the superheater. More im-
portant than the change in radiative heat transfer During operation with 65 % dry solids and at
is the change in fouling of superheater surfaces. 90 bar steam drum pressure, total superheat-
ing equals to 230 oC of which 50 oC is used for
With increasing dry solids content there should be attemperating. If we assume that furnace exit
no obstacle to increasing the furnace exit tempera- temperature remains constant at 950 oC and boiler
tures. For new recovery boilers, less than 1000 oC bank inlet temperature is now 580 oC, how would
is typically specified even though nose tempera- superheating change when black liquor dry solids
tures of up to 1100 oC have been measured. With is 80 %.
higher dry solids the observed superheater corro-
sion rates are much reduced. A possible cause is We can use non dimensional calculation and as-
the lack of reducing conditions and the increase in sume heat transfer coefficient does not change.
carbonate content in deposits.
Total temperature difference θ is about
The flue gas flow is lower with higher dry solids
liquor and the heat flux required to superheat 950 - 310 = 640 oC.
steam increases. This means that the temperature
drop of flue gas across superheater will increase. Temperature difference of flue gas is
To maintain constant superheating either the
furnace outlet temperature must be increased or 950 -580 = 370 oC.
the boiler bank inlet temperature can be lowered.

Temperature difference of steam is 1-1.05*0.231/*0.266 = 9 % or 20 oC.
480 - 310 = 170 oC. Example
Non dimensional parameters for the current case How would then increase in furnace exit tempera-
are ture by 23 oC affect the boiler bank inlet tempera-
ture and heat to superheat.
R = 170/370 = 0.459
We can again use non dimensional calculation
ε = 370/640 = 0.578 and assume that heat transfer coefficient does not
change.
then
Total temperature difference θ is
Rε = 0.459*0.578 = 0.266
973 - 310 = 663 oC.
and we can solve z
All dimensionless parameters, R, ε and z remain
z = ln((1-Rε)/(1-ε))/(1-R) = 1.025 constant so from previous example R = 0.406, ε =
0.569 and z = 1.025
Now the flue gas flow will decrease 5 % (Figure 6-
3) , so z will decrease 5 % to 0.974. The steam flow Temperature drop of flue gas from nose to boiler
will increase 7%, so R will decrease by 7% and by bank inlet is εθ = 0.569*663 = 377 oC. The boiler
5 % to 0.406. bank inlet temperature is then 973 - 377 = 596 oC.
The heat capacity ratio R is the same so heat to
We can then solve the new ε by assuming that z steam increases by 377/370-1 = 2 %.
remains constant
From table 6-6 we see that for example boiler the
ε =1-(1-R)/ (e -z(1-R)-R) = 0.569 available heat to superheat decreased 107.5/110.4
= 2.6 %, steam flow increased 135.3/126.6 = 6.9 %
then and superheating decreased 20 oC. Our assump-
tion of constant heat transfer coefficient caused
Rε = 0.406*0.569 = 0.231 underprediction of change.
so heat to superheat decreases

7 Recovery boiler processes
Figure 7-1, Recovery boiler overall process (Ahlstrom Machinery).
The processes occurring in boilers are divided system handles smelt and vent gas system handles
according to the main media to several subproc- wet exhaust from green liquor system. Starting
esses. The main categories are air system, flue gas and stopping black liquor recovery boiler requires
system, black liquor system and water and steam auxiliary fuel firing, which is usually oil or gas.
system. The ash system takes care of the ESP and
convective surfaces ashes. Kraft process produces smelly odorous gases.
These gases classified as concentrated NCG and
Recovery boiler has processes that are typically diluted NCG are often burned at recovery boiler.
missing from other types of boilers. Green liquor
Figure 7-2, Specific air for black liquor combustion as function of black liquor higher heating value.
Recovery boiler processes 7-1

Air is typically delivered at several horizontal
elevations to ensure complete combustion and
minimize emissions. A typical boiler air system,
Figure 7-3, consists of 1 - inlet duct with silencer,
2 - venturi for air flow measurement, 3 - air
blower, 4 - air heater, 5 - distribution ducts, 6&7
- air ducts. In addition pressure and temperature
measurement devices are installed to proper loca-
tions.
Air flow is controlled with dampers in furnace

openings and by control of air blowers. The air
flow trough blowers can be controlled by damper
in the duct, changing of rotative speed of the
blower or by inlet vane control dampers. The air
flow control with dampers in furnace openings
produces high duct air pressures. This increases
blower power consumption but ensures uniform
flow though each opening.
Air intake is typically located high inside the

boiler house. This ensures even temperature
Figure 7-3, Air system for a small boiler. profile in the boiler house and utilization of heat
losses. Upstream location of air blower from air
7.1 AIR SYSTEM preheaters decreases air blower power consump-
tion. After air heater the air is split into separate
Black liquor consists of combustible organic ducts for even distribution to different sides of the
material and inorganic material that react in the boiler. It is important to use low design velocities
furnace. To be able to sustain combustion air is as even small variations in parallel air duct lengths
needed. As the organic content in the black liquor otherwise cause large differences.
increases more air is needed per unit of black liq-
uor dry solids. The main requirements for boiler Air temperature is controlled by air heaters. There
air system are maximum mixing and proper air are several air heater types. When flue gas is rela-
distribution. tively clean it can be used to preheat air. In bark
Figure 7-4, Air and flue gas process diagram.

clean it can be used to preheat air. In bark boilers
several pass steel tube cross flow air heaters are
used. In dirty environments, recovery boilers,
steam and feedwater air heaters are used. The role
of air heater is to increase the furnace temperature
which decreases required heat transfer surfaces
and increases heat transfer to steam.
Air system - requirements

The newest air systems introduce most of the
combustion air into the lower furnace, just above
the furnace floor. The main target is to maintain a
high temperature and uniform gas velocity distri-
bution in the lower furnace. Most of the innova-
tive new air systems concentrate on increasing the
mixing and jet penetration in the lower furnace
(Wallen et al., 2002).
Higher black liquor dry solids content increase the

furnace temperature. With a higher temperature
Figure 7-5, Furnace of a large recovery boiler (Män- difference the penetration of air jets increases. Im-
tyniemi and Haaga, 2001). proved understanding of mixing requirements in
the lower furnace led to redesign of the airports.
boilers several pass steel tube cross flow air heaters All secondary airports are designed for both pres-
are used. In dirty environments, recovery boilers, sure and flow area control. This allows the opera-
steam and feedwater air heaters are used. The role tor to change airflow pattern depending on the
of air heater is to increase the furnace temperature changes in black liquor properties for improved
which decreases required heat transfer surfaces lower furnace control. Airport shapes are chosen
and increases heat transfer to steam. to maximize operating life and minimize fouling.
Air is typically delivered at several horizontal The main target of the design of the air system
elevations to ensure complete combustion and is to maintain a high and uniform temperature
minimize emissions, Figure 7-5. Forced draft and gas distribution in the lower furnace. With
fans are used to supply air to furnace. Airflow is optimized air introduction the emissions are low
controlled with dampers in furnace openings and and the reduction efficiency is constantly high.
ducts and by control of air fans. Damper in the With a proper design the combustion takes place
duct can control the airflow to furnace. The same in the lower part of furnace. The whole height is
can be done by changing of rotative speed of the effectively utilized for cooling the flue gases before
blower or by inlet vane control dampers. The air they enter the superheaters.
flow control with dampers in furnace openings
produces high duct air pressures. This increases The combustion air is introduced into furnace
blower power consumption but ensures uniform from at least three air levels; primary, second-
flow though each opening. ary and tertiary. The primary and secondary are
located close to the furnace floor beneath black
Air intake is typically located high inside the liquor guns. The tertiary air is introduced above
boiler house. This ensures even temperature liquor guns. With this system reducing conditions
profile in the boiler house and utilization of heat are maintained close to the char bed in the lower
losses. Upstream location of air blower from air part of the furnace. Effective oxidizing environ-
preheaters decreases air blower power consump- ment in the upper part is generated. This means
tion. After air heater the air is split into separate that 80 to 90% of total air is introduced through
ducts for even distribution to different sides of the primary and secondary air nozzles and the rest
boiler. It is important to use low design velocities and excess air via tertiary air nozzles.
as even small variations in parallel air duct lengths
otherwise cause large differences. Each air level has a separate air fan (forced-draft
fans). In smaller boilers only two air fans are used;
Air heaters control air temperature. There are one for primary and one for both secondary and
several air heater types. When flue gas is relatively tertiary. Having separate fans for each level, air
pressure in windbox and airflow can be controlled
Depending on the black liquor analyses a typical
value of specific airflow is 3.6-4.0 m3n/kgds. While
calculating the actual volumetric air flow the tem-
perature, pressure and altitude corrections have to
be made. In order to increase the controllability 20
to 30% design margin is used at the primary and
secondary air levels. When operating at design
load the airports are about 80 percent open. At
the tertiary air level design margin is up to 100%.
The big margin is due to willingness to control the
share of tertiary air in wide scale. Design margin
is also needed to take into account possible chang-
es in black liquor composition. If the calculated air
flow from given black liquor analysis differs a lot
from typical specific value wider design margins
have to be reconsidered.
Windbox pressure is a function of boiler capac-

ity. Different design pressures at different size
furnaces are used so that secondary and tertiary
air jets reach the center of furnace. Design static
pressure in a primary air windbox varies between
Figure 7-6, Area of penetration for primary air.
800 …1500 Pa. The lower value is the minimum
necessary to avoid char bed and smelt intrusion.
independently. Variable inlet vanes or speed of the Secondary and tertiary air windbox pressures
fans controls desired windbox pressure. vary between 3000 … 5000 Pa. Due to furnace
draft the pressure drop over the nozzle is a little
Each air level has usually own air ducts. These are higher (+200 … 400 Pa). For larger boilers higher
located symmetrically around the centerline of windbox pressures are used. High windbox pres-
the furnace. Much attention is paid to design of sure means high jet velocity, high mixing, high
air ducts. Several inlet locations and low design efficiency combustion and higher electricity usage.
velocity are targeted. This arrangement ensures a
uniform air distribution to all walls eliminating
the risk for uneven loading of the upper furnace. Primary air system - design
In older systems the airflow can vary up to 30%
The primary air should be delivered at constant
between two adjacent nozzles at the same wall.
pressure and velocity to maintain combustion
stability. It should be delivered uniformly to the
Air flow used in airport calculations is a theoreti-
boiler to create symmetry (Adams, 1988). Primary
cal air plus excess air (10 … 20 %), from which
airflow has lesser effect to furnace combustion
sum the infiltration air (4 … 8 %) is subtracted.
than other air levels. With primary air charbed is
kept away from the walls by increasing burning
at the edges of the char bed. The effective area of
primary air reaches only to about 1 m from the
wall, Figure 7-6.
Primary air openings are equally spaced on each

wall. Higher spacing requires higher primary air
pressure. Side spacing of up to about 650 mm
has been successfully used. Side spacing in the
corners of the furnace is 300 … 700 mm. In that
way accumulation of black liquor in the corners
is prevented. This charred heap can collapse and
flow through primary air openings to the air duct
and even into the boiler house. The airports are
designed so that height (straight portion) to width
ratio is 1.5 … 2. Plugging tendency of modern
wide openings is lower than the narrower ones.
Figure 7-7, Cast primary airport opening (Shen-
Use of high pressure and individual dampers de-
assa et al., 2002).
creases the possibility of primary air plugging. and Hastings, 1992). After few hole diameters the
jet has coalesced to round one.
The primary air openings are directed from 10 to
45° downwards. The purpose of this is to prevent The first openings are placed at 0,5 … 2 m
smelt flowing to air openings. Higher opening distance from the corner. The spacing between
angled leads to ‘hole digging’ in front of the ports. secondary openings can vary up to 2.5 m. The
Molten areas can be seen, which lead to low re- larger the opening the larger is the optimum spac-
duction and in extreme cases possibly to primary ing (Tse et al., 1994). When planning secondary
port damage. airport opening arrangement the number and
placement of start-up burners has to be taken into
Secondary air system - design consideration. The number of secondary airports
is often equal in the opposite walls. Unequal
Secondary air is introduced just above the char number causes an uneven gas flow at the furnace
bed to maintain a specified temperature in the outlet. It is also recommended that the number of
lower furnace. By controlling this temperature low the secondary airports at each wall is even.
emissions and high reduction efficiency can be
maintained. If the boiler is using high dry solids Tertiary air system - design
the main challenge is often how to keep a high
enough char bed, not how to keep it hot. With The tertiary air burns the remaining combustibles
a proper design the combustion takes place on rising from the lower section. Its effective area
the lower part of the furnace. The whole furnace should extend to center of furnace. In practice this
height is thus effectively utilized for cooling the is quite easy because the tertiary air is unheated
flue gases before they enter the superheaters. This and provided at high velocity.
improved temperature profile (reduced furnace
outlet temperature variation) reduces fouling Tertiary openings are located typically on the
of heating surfaces and the risk for superheater front and rear wall. Maximum spacing between
corrosion. Stable combustion in the lower furnace nozzles can be up to 3.5 m. The number of the ter-
requires efficient mixing by high velocity jets. Ad- tiary airports can be uneven. Often front wall has
justable dampers and precise flow control enable one port more than rear wall. The tertiary airports
operation at low load levels. are then interlaced. Interlacing increases mix-
ing (Jones in Adams et al., 1997). By introducing
Secondary air level has the biggest effect on recov- more air through front wall the upflow is forced
ery boiler furnace operation. Its main purpose is to rear wall side. Properly designed nose can then
to burn up flowing gases rising from the char bed. function as effective flow turner.
This burning radiates heat back to the bed, which
helps to maintain high char bed temperature. In Opening shape does not play a major role. Jet
addition the secondary air prevents the char bed penetration is function of port hydraulic diameter
from growing too high. Complete combustion (Collin and Vaclavinec, 1989).
means that the mixing of the combustibles and the
air should be as effective as soon as possible above Often the number of tertiary airports tends to
the whole char bed. Upper section heat surface increase with boiler size, Figure 7-8. It should be
plugging can be decreased by proper introduc- remembered that 3/2 and 4/3 arrangements are
tion of secondary air. Using improved secondary the best for mixing (Adams, 1994).
air arrangement can decrease carryover, lower
furnace exit temperature, lower gaseous emis-
sions and increase hearth temperatures (Jones and
Chapman, 1992).
The secondary airport arrangement should be as

flexible as possible. Then different air distribution
models can be used during boiler operation. The
secondary airports are usually placed along the
whole length of the wall. Most modern air systems
use two wall secondaries. The idea of secondary
air port arrangement is to get as high coverage
of the horizontal cross sectional area as possible.
Another item is to decrease the collisions of the
air jets in the corners of the furnace. Air port Figure 7-8, Different tertiary airport placements.
shape has little effect on jet penetration (Blackwell
The airflow through primary airports is adjusted
either individually or in groups. Secondary and
tertiary airports are adjusted independently.
Airports at the secondary and tertiary air levels

should be as open as possible. If a number of
openings needs to be throttled it is more prefer-
able to close a few airports in order to get others
more open. Dampers are needed at the primary
air level. With them a lower windbox design pres-
sure can be utilized. Leaving them out leads in
problems with low load operation.
During boiler operation pressures fluctuate at the

primary air level +-200 Pa, at the secondary air
level +-500 Pa and at the tertiary air level +-700
Pa. Reason for the fluctuation is that the airports
are getting dirty. When automatic airport rodders
are installed this deviation can be kept below 100
Pa.
Figure 7-9, Different primary air airport damper The individual dampers of secondary and tertiary
arrangements (Ahlstrom Machinery). airports are connected to a governor shaft that
runs along each wall. The system is provided with
Airport flow adjustment an actuator and the dampers can be operated
remotely wall by wall. Since the connection to the
Airflow to each level is accurately measured with a shaft is adjustable it is still possible to individually
venturi either on fan inlet or on fan outlet. Airflow adjust a damper even if it is connected to the main
is controlled with dampers in furnace openings shaft. Governor shaft can be used in air control-
and by control of air blowers. ling if requested.
In all the airports the adjusting takes place

smoothly just in front of the furnace wall. Adjust- Multilevel air systems
ing is performed by dampers. In addition, those
Theoretically multilevel air looks nice but in
are protected at the primary air level ring duct
practice it is difficult to operate, because the flow
in the case of a falling char bed. Guillotine type
conditions in the furnace are not stable. Especially
dampers can also be used. In Figure 7-9 several
at the low load the penetration of the air jets de-
types of Ahlstrom Machinery primary air airports
creases considerably. In literature and in practice
are shown.
four level air can be introduced by doubling the
Figure 7-10, Furnace temperature with three and four level airs.

secondary air level or by doubling the number of
tertiary air level.
The most typical application of multilevel air is

NOx control. By increasing the number of air lev-
els we can shift the combustion down. The lower
furnace temperatures decrease increasing with
lower stoichiometry the sulfur species concen-
tration. This arrangement is beneficial for NOx
control, but detrimental for SOx and reduction
control.
Low concentration NCG systems

Typically dedicated nozzles are used to burn
HVLC NCG (or DNGC). These nozzles are of spe-
cial design. To allow rodding of each airport dur-
ing operation ball valves and tightly closeable disk
valves are used to separate each nozzle from NCG
stream. Figure 7-11 shows and example of DNCG
nozzle design, courtesy of Andritz. Another man-
datory requirement is condensate removal lines.
DNCG’s contain often 20 – 30 % moisture.
High concentration NCG

systems
Dedicated burners are used for LVHC NCG (or
Figure 7-11, Example of HVLC NCG system before
CNCG), turpentine and methanol. Typically all
insulation and installation of separation valves.
three are burned using same burner opening.
Turpentine and methanol are fired at liquid form
using lances.
High concentration odorous gases increase sulfur

load in the furnace. This increase can be up to
30 % of all sulfur released in the furnace. If high
concentration odorous gases are fired, then high
dry solids (>75%) is preferred. The main criterion
is that LVHC NCG can be fired without plugging
of economizer section.
Burning dissolving tank vent

gases
There are several recovery boilers in operation Figure 7-12, Example of dissolving tank vent gas
that introduce dissolving tank vent gases to the system.
recovery boiler furnace. In a typical system a fan use vent gases are routed trough scrubber to roof
draws vent gas from dissolving tank. Condenser or stack. Figure 7-12 shows and example dissolv-
is used to reduce moisture. Controlling moisture ing tank vent gas system courtesy of Kvaerner.
levels is essential. Adding water vapor to furnace
unnecessarily cools down flue gases and ads to
flue gas volume. 7.2 FLUE GAS SYSTEM
A scrubber is sometimes used to reduce particu- The flue gas system transports combusted material
late loading and to keep ducts from plugging. Be- from the furnace safely to the atmosphere. Flue
fore introduction to furnace these gases are heated gases from the furnace pass through heat transfer
to prevent accidental introduction of water to surfaces. The flue gas then flows through flue gas
recovery boiler furnace. When the system is not in ducts, dampers, flue gas fan, electrostatic precipi-
Figure 7-13, Flue gas system of a recovery boiler (Ahlstrom Machinery, 1998).
tator, possible scrubber and stack. Flue gas system of an example boiler is shown in
figure 7-13. Similarly to air system flue gas fans
Natural draft caused by density differences in draw air from the furnace to the stack. In old days
stack is not adequate to draw flue gases out from draft in the stack was enough. Flue gas ducts con-
recovery boiler furnace. The flue gas flow is nect pieces of equipment. Electrostatic precipita-
controlled by induced draft of ID fans. Flue gas tors and other emission reduction equipment are
fan controls also the furnace draft. The furnace employed to decrease pollution.
pressure at liquor gun level must be kept below
boiler house pressure.
7.3 WATER AND STEAM
Boiler water and steam circulation starts with low
pressure feedwater and ends with high pressure
and temperature steam. The water and steam
circulation components transport, pressurize,
preheat, vaporize and superheat the water into
steam, Figure 7-14.
The starting point of boiler water and steam

circulation is the feedwater system. For reliable
operation of boiler the feedwater must be low
in oxygen and minerals. The feedwater system
consists of feedwater tank-1, deaerator, boiler
feedwater pump-2, control valve and the feedwa-
ter piping-3. The feedwater enters then first heat
transfer surface, economizer.
The economizer recovers heat available in the flue

gases. This heat is used to preheat the water close
to the boiling point. Recovery boilers surfaces foul
easily so typically vertical gas flow is used. Coun-
terflow arrangement where hot flue gases and cold
feed water enter at opposite ends, is preferentially
Figure 7-14, Water circulation system. used.

Figure 7-15, Steam and water process diagram.
The water from the economizer is used to generate sion due to the presence of molten smelt in a
clean attemperating or desuperheating water-5 porous char bed. To minimize possible damage an
from steam in a sweet water condenser-4. The emergency system, rapid drain system is installed.
feedwater then flows to the steam drum-6. In The rapid drain system consists of fast valves and
steam drum steam is separated from water. pipes to major pressure parts. If the rapid drain is
Separation is achieved by gravitation, screens and opened pressure inside boiler drives water from
cyclone separators. the pressure part. In rapid drain system there are
separate lines from furnace walls, from the boiler
Downcomers-7 feed saturated water to evapo- generating bank and from economizers. Water
rative surfaces. Partially evaporated water is is left at the lowest part of the furnace to provide
collected with risers to the steam drum. Furnace cooling for the floor tubes. If all water from floor
walls do most of evaporation. Depending of boiler tubes is drained, the heat from char bed damages
operating pressure between 10 to 25 % evapora- the tubes.
tion is done in boiler bank. The two predomi- Water circulation is caused by density differ-
nant boiler bank arrangements are vertical and ence between steam filled heating surfaces and
horizontal flow. Both can be used for dirty and downcomers. The circulation is limited by friction
clean gases. losses in surfaces, headers, risers and downcom-
ers. Increasing boiler operating pressure decreases
Saturated steam exits from drum and flows the driving force, density difference. High pressure
through superheaters. The role of superheaters operation requires either pump assisted circula-
is to heat the steam well above the saturation-8. tion or once through operation.
Because turbines can only operate down to some
95 % steam content, proper choice of main steam Downcomers and raisers
pressure and temperature is required. Steam flow
is controlled by main steam valves-9. The steam Downcomers are tubes that start from the drum
from superheaters flows trough main steam line- and lead water to lowest points of furnace walls
10 to turbine. and boiler banks, Figure 7-16.
The steam temperature control is done by at- Downcomers can be divided to main downcomers
temperating. Either condensed steam or feedwater and connecting tubes. It is customary that from
is sprayed into steam line between superheaters. two to five large tubes go straight down from the
Typically only a two stage desuperheating is used. drum. From these tubes smaller tubes take water
The source of the spray water must be determined to wall headers.
by the available feedwater purity.
Separate downcomers are often constructed to
Recovery boiler has a risk of smelt/water explo- the boiler bank and sometimes also to the side
wall extension between furnace tubes proper and
boiler bank.
Wall tubes
Wall tubes are the membrane wall tubes that ex-
tend from the lower headers to the upper headers.
Usually front and back wall are bent together to
form the furnace bottom, Figure 7-16. In addi-
tion front wall is often bent towards the boiler
bank to form the furnace roof. In lower furnace
compound membrane construction is used. In
middle and upper furnace carbon steel membrane
construction is almost always used.
For airports, manholes, liquor gun openings,

sootblower openings, smelt spout openings and
instrument openings tubes must often be bent.
Headers
Typically each heat transfer surface starts and
ends with header, Figure 7-16. Headers are larger
tubes that connect all parallel tubes of a heat
transfer surface together. Their role is to divide the
incoming liquid evenly to all tubes. They also try
to collect the fluid without causing uneven flow.
Headers were earlier of various shapes. Even

square cross section was used. Nowadays only
tubular headers are used.
External pipes
Figure 7-16 Main tubes in a recovery boiler (Kvaer- Most of the parts of the pressure vessel belong to
ner). the heat exchanger surfaces and circulation sys-
tem. There are is a system of pipes, equipment and
vessels that handles the feedwater. There is also a
system of pipes that handles main steam. In ad-
dition there are pipes that belong to the desuper-
heating system, blowdown system, air preheating
system, draining system, air venting system and
piping assisting the operation of the safety valves.
All these tubes and pipes are called external pipes.
Feedwater system
Feedwater system consists of piping tanks and
equipment that pump water to the inlet of the
economizer, Figure 7-15. Demineralized water is
pumped to the feedwater tank. Feedwater tank
has a storage capacity corresponding to 15 … 45
minutes of feedwater usage at MCR. In addition to
storage, deaeration is carried out to the incoming
water. The purpose is to remove all traces of gases
especially oxygen and carbon dioxide from the
Figure 7-17, Superheater connection tubes water

Figure 7-18, Feedwater tank, 1- deaerator, 2 - steam in, 3 – trays, 4 - water in, 5 - water out.
Feedwater is pumped from the feedwater tank to spond to some 20 minutes to 1 hour reserve water
the economizers with feedwater pumps. Typically calculated from MCR feedwater.
the feedwater pumps are driven by electric mo-
tors. To ensure safety during pump failure there Feedwater tank is also used to separate gases from
are usually more than one pump. Feedwater flow the feedwater. This is done by steam stripping
is controlled by feedwater valves. incoming water in the deaerator.
Superheater connecting tubes Feedwater pump

Each superheater must be connected to the fol- Feedwater pumps are commonly centrifugal
lowing superheater. To even out left right flow and pumps, Figure 7-19. To ensure high enough head,
temperature imbalances connecting tubes cross several even dozens of individual pumps have
sides. been arranged in series to same axle. A feedwater
pump is most often driven by electric motor for
energy efficiency. Often also a spare steam turbine
Feedwater tank driven feedwater pumps included for safety
reasons.
Feedwater tank sits usually in the middle of the
boiler room. Feedwater tank is situated at height To ensure adequate feedwater flow a boiler must
of 10 – 30 meters so that the feedwater pumps can have at all times a little bit of higher capacity for
have the necessary head to operate. Feedwater is feedwater than for steam generation. Feedwater
stored in the feedwater tank, Figure 7-18. Return-
ing condensate is mixed with make-up water. The
feedwater flow is not even and the boiler can have
disturbances. The tank capacity needs to corre-
Figure 7-19, Feedwater pump Figure 7-20, Dolezahl – attemperator, 1 - Steam

flows from drum, 2 - condensation in heat ex-
changer, 3 - condensate sprayed to desuperheat, 4

Figure 7-21, Spray water group
pumps can be common for several boilers. For Spray water group
safety normally one needs full feedwater flow even
if one feedwater pump fails. In a spray water group feed water is injected into
steam. This increases the steam massflow and
Dolezahl – attemperator decreases the steam temperature. Typically spray-
ing to different tubes is arranged close to another
Dolezahl attemperator is used especially in the for maintenance purposes. Also automatic spray
industrial boilers where feed water quality is lower water valve is usually doubled or tripled with
than in the utility boilers. In a Dolezahl feedwater hand valves. Often, because of location airing and
condenses steam. This condensate is sprayed using emptying valves need to be installed.
the static head formed to attemperate the pro-
duced steam. Normally saturated steam is used. Steam temperature increases in each superheating
High quality of spray water is ensured as saturated surface. Between surfaces water is sprayed to con-
steam has much higher purity than corresponding trol final steam temperature. Forcing steam flow to
feedwater. change sides helps balance uneven temperatures.
Drainage and air removal

When boiler is brought to operation it must be
Figure 7-22, Operation of attemperating group Figure 7-23, Steam drum.

Figure 7-24, Erection of steam drum under wintery conditions.
filled with water and air removed. All top points of

each heat exchanger surface need to be equipped
with small lines with a hand valve to be able to
remove air from the pressure vessel.
Steam drum
Steam drum consists of a circular section welded
from bent steel plate with two forged ends. Cir-
cular section has a number of inlets and outlets.
A drum can be made of a single thickness sheet if
the pressure is low or with plates of multiple thick-
nesses if pressure is high. One reason for different
thicknesses is that sections where are openings
need to be thicker than sections without openings.
Steam water mixture enters from raisers. Cyclone
Figure 7-25, Operation of drum. separators force water down. Steam exits through
Figure 7-26, Placement of water separators in a drum, 1- feed water in, 2 – downcomers, 3 – raisers, 4 - steam
out, 5 - droplet separator.

Figure 7-27, Black liquor and ash process diagram
baffle separators from top. Water exits through 7.4 BLACK LIQUOR AND ASH
downcomers. Additional separation is done by a
wire mesh or baffles at the top of the steam drum. In modern recovery boiler the black liquor from
heavy black liquor storage is pumped through
Only <<0.01 % water in steam is allowed. Water heavy liquor pipe to ring header. Ring header is a
contains impurities. Impurities deposit on tubes large diameter pipe that circles the furnace. The
and cause superheater overheat. Especially purpose of the ring header is to ensure equal flow
carbonate (CO3) and sulfate (SO4) form hard to to all liquor guns. Collected fly ash form precipita-
remove deposits with low thermal conductivity. tor is mixed with about 50 % dry solids liquor in a
Impurities will also deposit on turbine blades. Na mix tank and pumped back to the evaporator. Ash
+ K are the most harmful impurities for turbine. recycle helps evaporator operation in high solids.
As the steam drum is the heaviest piece of equip- Black liquor system of an example boiler is shown
ment it is often lifted as the first piece of pressure in figure 7-28. It needs to have flow, temperature
part equipment. and pressure measurement. Typically there are
more than one insertion points of the black liquor
into the furnace, with black liquor guns.
Figure 7-28, Black liquor system of a recovery boiler (Ahlstrom Machinery, 1998).

Figure 7-29, Ash system of a recovery boiler (Ahlstrom Machinery, 1998).
To control the liquor temperature the liquor are used to control ash flow. Ash can is handled
can be heated in a liquor heater. Heating is done dry or sluiced and pumped away.
either directly with steam or indirectly with a heat
exchanger. In older boilers make up chemicals were added to
mixing tank. Now as there often is excess sulfur in
Black liquor system design is thoroughly covered the mill the make up is added as sodium hydrox-
by standards and safety recommendations. This ide.
is because if it malfunctions, fire or explosion
outside or inside the furnace could happen with
loss of life and property. 7.5 OIL/GAS SYSTEM
When we burn black liquor chemical ash is Oil/gas fired fuel system of an example boiler is
formed. Ash system of an example boiler is shown shown in Figure 7-30. If several fuels are used
in Figure 7-29. Ash system consists of ash trans- each fuel has its own fuel system.
port equipment like screws and conveyors. There
are hoppers where ash is collected. Rotary feeders Oil firing system has built in safety systems. Each
Figure 7-30, Oil/gas fired fuel system of a recovery boiler (Ahlstrom Machinery, 1998)

Figure 7-31, Green liquor process diagram.
burner needs it own flame monitoring. Instead Mixing molten inorganics to water creates heat.
of main flame monitoring an auxiliary or support In a 3000 tds/d boiler the heat input is more than
flame can be monitored. Normally double valve 10 MW. Heat usually escapes dissolving tank as
arrangement is required to ensure closing of oil steam. With steam green liquor droplets and H2S
line. gases formed escape. To reduce gas volume, dust
and TRS content the vent gases are scrubbed with
Flame is lighted with an igniter. Often gas igniters alkaline liquid.
are used. Especially heavy fuel oil is heated to
lower its viscosity. To ensure sufficient atomiza- Smelt spouts run from furnace to dissolving tank
tion of oil pressurized air or steam is used to assist transporting molten smelt. They are cooled by ~60
atomization. o
C water that circulates inside them.
If natural gas is used as fuel, venting to air is often Because green liquor forms pirssonite that will
used to ensure that gas leak to furnace is impos- quickly block transport pipes, weak wash and
sible. green liquor pipes are exchanged to one another
once a shift.
7.6 GREEN LIQUOR

7.7 AUXILIARY EQUIPMENT
Green liquor is formed when smelt and weak
white liquor are mixed at dissolving tank. Green Recovery boiler needs for operation a number of
liquor gets its name from iron oxide impurities equipment in addition to main equipment. This
which color green liquor. equipment is called auxiliary equipment. There are
number of recent changes and additions to recov-
ery boiler auxiliary equipment that are improving
the process.
Automatic cleaning system

The target of airport cleaning system is to keep all
the airports as open as possible most of the time.
The port rodders are typically provided to airports
below liquor guns. Sometimes it has also been
provided to secondary airports. They are not used
at all levels primarily because of expensive price.
Plugging tendency at tertiary and quartenary air
levels is low because of bigger airports and higher
Figure 7-32, Roddingmaster automatic primary pressure.
port air rodders (Sheyneassa et al., 2002).

Figure 7-33, Radial air fan (Singer, 1981). Figure 7-34, Steam air preheater.
The primary airports are cleaned in groups of two In operation, they normally cause a small under
or three openings. At the secondary air level they pressure into the furnace.
are cleaned separately. All this is performed ac-
cording to pre-programmed intervals. The selection of fan is made with the performance
curves provided by the fan manufacturer. The
curves are based on experiments that the manu-
Pyrometers and char bed facturers have made in laboratories for different
cameras types of fans. The curve illustrates the change in
the total pressure created by the fan as a func-
Separate pyrometers are not needed if bed cam-
tion of volume flow and speed of rotation. When
eras are used. Instead cameras with integrated py-
choosing a fan, the required volume flow and
rometer should be used. Cameras and pyrometers
pressure difference must be known. Other factors
are located in secondary air openings. One camera
influencing the choice are
on a side wall close to rear wall and the other one
- Efficiency
in the opposite corner in the front wall.
- Required space
- Shape of the characteristic curve for the fan.
Air fans
Most recovery boiler fans are radial type. Some-
In boiler plants, fans supply primary and second- times two sided air inlet is preferred. Axial fans
ary air to the furnace. The air is primarily used are avoided as more expensive.
for combustion of fuels, but can also be utilized
in pneumatic transport of fuels and other solid
materials to the furnace. The induced draft (ID) Steam air preheater
fans exhaust combustion gases from the boiler.
Steam air preheaters are used to preheat steam.
Figure 7-35, Retractable sootblower.

Figure 7-36, Electrostatic precipitator being erected.
Steam flows inside tubes. Air flows outside finned

or bare tubes and is heated by the condensing
steam. Condensate flows out from the bottom
headers.
Sootblowing
Most typical sootblower in the recovery boiler has
a rotating lance which is inserted from wall to the
heat exchanger surface, Figure 7-35. Tip of the
lance has a hole. Steam at sonic speed is injected
from tip. Sootblowers are pulled outside the boiler,
when not in use.
This type of sootblower gradually rotates around
while it is inserted and retracted. The sootblower
is then able to clean a large area of a heat ex-
changer surface.
The steam mass flow through the nozzle is directly

linear to the steam pressure (Jameel et al., 1994).
Figure 7 37, Schematic arrangement of collecting Nozzle design and flow area at nozzle throat affect
surfaces with electrodes (F.L.S. Miljø, 2002). the mass flow.
Electrostatic precipitators
Electrostatic precipitator efficiency depends on
7-1
Where
e is the collection efficiency, -
A is the collector surface, m2
w is the collection migration velocity, m/s
V is the actual gas flow, m3/s
k is an empirical constant (0.4 .. 0.5 ..0.6), -
Figure 7-38, Section of an ash conveyor. The principle of electrostatic precipitator is simple.
Figure 7-37. Dust laden flue gas flows along col-
Figure 7-39, Typical ash conveyor before insulation.
lecting surfaces formed of plates. Strings of wire

are suspended between the platens. When high
voltage electric current is applied between the
electrodes (-) and the collector plates (+), the dust
particles become charged. The charge draws them
towards the collecting plates where they deposit.
Ash conveyors
Ash is collected to hoppers. Motors turn ash con-
veyors. Collected ash is dropped to ash chutes.
Figure 7-39 shows a section of an ash conveyor.

Ash is dropped to the conveyor from the top. It is
dragged along the bottom to the right with mov-
ing scraper blades. These are attached to a belt.
The belt returns on the top of the conveyor.
The return leg can be also below the conveyor

Figure 7-40, Ash mixing tank.
bottom.
causes problems in the evaporation.

Mixing tank
Ash from electrostatic precipitator, economizers Boiler bank, economizer and electrostatic pre-
and boiler banks is mixed with 50 … 65 % black cipitator salt ash is dropped from the top through
liquor. Recovery boiler ash is reusable chemicals, one or several openings to the mix tank. Incoming
mostly sodium sulfate. The amount of recovery black liquor is sprayed from special nozzle so that
boiler ash is 5 … 12 % of black liquor dry solids it wets the ash at the liquid surface, Figure 7-40.
flow. Higher ash flow corresponds to higher black Rotating rotor, often with blades mixes ash with
liquor dry solids. In the older designs the black black liquor. Ash containing black liquor is drawn
liquor from the mix tank was sprayed to the boiler through a screen plate (not shown in the picture)
furnace. In more modern, high dry solid designs with a pump and pumped back to the evaporator.
the black liquor is returned to the evaporation
plant for final concentration. If black liquor dry Mixing hot ash with black liquor causes evapora-
solids inside the mix tank gets too high the black tion of water. At the same time sulfur containing
liquor stops flowing out from the mix tank. If gases in the black liquor are released. Formed
black liquor dry solids is too low the ash is dis- NCG’s must be removed from the mix tank and
solved into the black liquor. Increasing dissolved treated appropriately. Part of the ash is coming
sulfate in black liquor makes it more fouling and from economizer hoppers which operate under

Figure 7-41, Dissolving tank.
like popcorn. This popping action is responsible

for the noise. To decrease the noise the dissolv-
ing tank is often lined with concrete. In addition
steam jets or green liquor jets are used in trying
to disperse smelt from smelt spouts to make very
small smelt droplets.
Dissolving tanks often operate at high enough

alkali content so pirssonite is formed at all
surfaces (Frederick et al., 1990). To limit scaling
incoming weak white liquor and outgoing green
liquor pipes are exchanged once a shift. Heat in
Figure 7-42, Startup burners. smelt forms vapor in dissolving tank. This vapor
contains small droplets of green liquor and TRS.
This stream must be treated before it is burned of
atmospheric pressure. To prevent backflow all released to atmosphere.
ash chutes are equipped with rotary feeders and
air purge. Mix tank operates at about 115 oC. To Sometimes an operating error leads to a large
prevent condensation and plugging all surfaces smelt flow. This flow can cause a dissolving tank
must be insulated. explosion. A dissolving tank stack is constructed
so that it will lead pressure wave away from the
boiler room. To prevent personnel accidents a
Dissolving tank concrete platform is built above the dissolving
The molten smelt flows from the recovery boiler tank.
furnace through smelt spouts to the dissolving
tank, Figure 7-41. When dissolved to water or Startup burners
weak wash the smelt forms green liquor. Constant
level and constant alkali content are maintained Startup burners are placed usually at all four
by pumping in weak white liquor to replace out walls, close to the recovery boiler bottom. The role
pumped green liquor. Control of alkali content is of the recovery boiler startup burners is to heat up
done by measuring the density of the out flowing the char bed during startup and low loads. In ad-
green liquor. dition the startup burners are used to burn away
excess charbed before the boiler shutdown.
Dissolving is a violent action with noise, heat and
vibration. This is because a smelt droplet is a poor Startup burners use light fuel oil or natural gas.
heat conductor. After it drops to the tank the liq-
uid cools the surface of the droplet. Cool surface Direct contact evaporator
contracts, but hot molten inside doesn’t. When
tension becomes high the smelt droplet pops open There are still many boilers that use direct contact
Figure 7-43, Cascade evaporator (Tampella).
(DCE) or cascade evaporators, Figure 7-43. Cas- Other problems that plagued the cascade evapora-
cade evaporator looks like old mill’s waterwheel, tor were fire (black liquor is burnable after all) and
where flat surfaces are replaced with rows of tubes. crusting of liquor to tubes because of too high flue
Incoming black liquor sticks to the tubes when gas temperatures.
they are submerged in a black liquor pool. When
these tubes emerge from the black liquor, flue There was another type of direct contact evapora-
gases evaporate water from the liquor that sticks tor promoted by Babcock & Wilcox. In it the black
to the tube surface. Black liquor dry solids 70 … liquor was sprayed to flue gases in a cyclone. Be-
74 % can be reached. Cascade evaporator was a cause this type is based on equilibrium and the CE
rather cheap way of increasing black liquor con- type is based on countercurrent drying the B&W
centration at a time when multiple effect evapora- type achieved lower solids 60 … 65 %.
tors could produce only 45 to 50 % dry solids.
Last new boiler to be equipped with cascade
A big problem with cascade evaporators is that evaporator started operating 1991 in US. At that
TRS in the black liquor is released during the time cascade evaporator was almost phased out
process. Even if black liquor oxidation, which de- in Scandinavia. Boilers with cascade evaporators
creases the TRS levels to close to the 10 ppm level, have been supplied recently to India and China.
is used the resulting sulfur emission is considered
unacceptably high for a modern mill.

8 Fouling
Black liquors contain inorganic chemicals. During Often a cause for fouling can be found by look-
combustion and reduction reactions significant ing at black liquor properties and changes in the
amounts of alkali compounds vaporize. Small char pulping process. Especially changes in sulfidity,
fragments and liquor particles may entrain to flue chloride and potassium levels can cause increased
gas flow. These phenomena cause deposits on heat fouling.
transfer surfaces.
To combat fouling black liquor properties can
The heat flow in a heat transfer surface can be be changed. Decreasing flue gas temperatures or
expressed as function of heat transfer area and carryover rate helps. Improving surface clean-
temperature difference between flue gas and ability or more frequent sootblowing can be used
steam/water. As the heat transfer surface gets to reduce fouling. Their ability to solve the current
fouled this heat flow decreases. Reduction in per- fouling problem should be weighed carefully.
formance is caused by the increased heat transfer
resistance of deposits.
8.1 ASH DEPOSITS ON HEAT
Fouling and fouling related phenomena have long TRANSFER SURFACES
been of concern in recovery boiler design and op-
eration. Several studies of recovery boiler plugging To understand fouling the first step is to identify
and fouling mechanisms have been reported (Tran the location and growth rate of the problematic
et al., 1983, Tran, 1988). Even with improved air deposit. Then one needs to find the cause for the
systems and advances in recovery boiler design, problem. It is often difficult to clearly identify
fouling remains one of the big operating problems. the degree of fouling without shutting down the
boiler, but instrumentation can help in pinpoint-
It is typical to talk about recovery boiler fouling ing the progress of fouling (Vakkilainen, 2000).
as it would be a consistent type of process with a
single cause. Like many industrial problems there The two main methods to observe the fouling of
are number of different causes that lead to same the heat transfer surfaces are visual observation of
end condition. To determine the cause of fouling the heat transfer surfaces by experienced operators
the exact location and the fouling rate should be and prediction of fouling from operating param-
determined. It is also recommended to look at eters; i.e. temperature and pressure measurements
black liquor properties and changes in the pulping (Vakkilainen and Vihavainen, 1992). Prediction
process. Changes in sulfidity, chloride and potas- of fouling through estimation from operating
sium levels of black liquor will cause changes in parameters is the most frequently used method in
recovery boiler fouling. modern sootblowing control systems (Bunton and
Moskal, 1995, Jameel et al., 1995). The main foul-
For a modern recovery boiler the crucial piece of ing indicators are pressure loss measurements, flue
equipment is often the boiler bank. When black gas temperature measurements and steam / water
liquor dry solids is over 70 %, the economiz- temperature measurements.
ers plug far less often. Superheater fouling has
dramatically reduced. Modern recovery boilers Modern CFD-programs offer some help in
have less superheater deposits because the new estimating the amount of carryover that ends up
improved air systems have halved the carryover. to superheater area (Fakhrai, 1999, Kaila and Savi-
harju, 2003). The amount of carryover is subject to
Finding the cause for recovery boiler foul- changes in air distribution, black liquor properties
ing can be started by examining fouling at that and liquor gun operating practices. Because of
boiler. There are several questions that need to be limitations in input CFD can only be used to high-
answered; How fast is the recovery boiler foul- light the direction of changes to carryover and not
ing. What measurements indicate the progress to predict actual operating levels.
of fouling. Is the fouling continuous or are there
equilibrium stages. Are there any upset or abnor- Several groups have attempted to formulate equa-
mal operation conditions that correlate with times tions describing deposit properties in recovery
of fast fouling. Does the furnace operate smoothly boiler heat transfer surfaces (Backman et al., 1996,
and well. McKeough and Vakkilainen, 1998, Frederick et
Fouling 8-1
Figure 8-1, Deposits on superheaters seen along cavity between SH3 and SH2.
al., 2001, Janka et al., 2004). With the aid of these Niemitalo, 1994, Janka et al., 1998).
programs and measured black liquor properties
one is able to predict from given operating condi- The flue gas velocities at superheater area are
tions which surfaces are prone to fouling (Forssén rather low 3 – 5 m/s. This creates good conditions
et al., 2000). for separation of large particles. For particles over
10 µm the collection efficiency at superheater
As the heat transfer surface gets fouled the build- surfaces is high (Jokiniemi et al., 1996).
up of deposits decreases available gas flow area.
This causes an increase in flue gas flow velocity Superheater deposits decrease the heat transfer
and pressure loss across a heat transfer surface. and lower the steam temperature. Low super-
The pressure loss is often more important for heater main steam temperature requires boiler to
maintaining operation than the decreasing heat be shut down for water was. Superheater deposits
transfer. Usually the recovery boiler maximum can significantly reduce the recovery boiler avail-
capacity is limited by the flue gas side pressure ability (Clement et al., 1995). As superheater heat
loss. Fouling increases until the flue gas fan can transfer decreases the temperature entering boiler
not supply enough pressure and flow. Then the re- bank increases. Superheater deposition can thus
covery boiler load must be decreased or the boiler negatively influence the fouling in downstream
shut for water washing. heat transfer surfaces.
Superheater fouling Boiler bank fouling

Superheater deposits are formed mainly by two Boiler bank is the crucial piece of equipment for a
mechanisms, inertial impaction and fume deposi- modern recovery boiler. Especially after increased
tion (Backman et al., 1987). Deposit proper- mill closure the boiler bank plugging has been
ties vary widely depending on the location of the most problematic fouling in modern recovery
sampling. The high carbonate and sulfide contents boilers.
as well as low chloride and potassium enrichment
can be explained by assuming different mixing There are several indicators for boiler bank plug-
ratio of carryover and fume (Vakkilainen and ging. Some of the typical ones are low melting

Figure 8-2, Deposition in the middle of the boiler bank.
temperature of ash deposited, increased black erick et al., 1998). As sulfur release stays rather
liquor chloride content, too high flue gas tem- flat (Perjyd and Hupa, 1984) the excess shows as
perature into boiler bank and fast sintering rate carbonate in ash, Figure 8-3.
of deposits (Frederick et al., 2001). All of these
indicators point to problems in controlling the With low dry solids firing there is no carbonate.
fouling rate. As less and less sodium is released there is excess
of sulfur. Part of the dust remains sodium bicar-
Economizer fouling bonate (Backman et al., 1984). Increased sodium
bicarbonate decreases pH of dust. Figure 8-4
Economizer fouling is typically caused by low ash shows actual simultaneous mill measurements of
pH. Higher dry solids increases lower furnace dust carbonate and pH. To keep dust pH above 10
temperature. Increased lower furnace temperature one needs at least 2 to 3 % of carbonate in ash.
increases sodium and potassium release (Fred-
Figure 8-3, Carbonate in ESP ash as a function of dry solids.
Fouling 8-3
Figure 8-4, Effect of dust carbonate to pH of ESP ash.
Figure 8-5, Pressure losses in heat transfer surfaces during 14 days after startup, 60 equals situa-
tion after 60 days (Vakkilainen and Niemitalo, 1994).
Figure 8-6, Temperature raises in superheaters during 14 days after startup, 60 equals situation
after 60 days (Vakkilainen and Niemitalo, 1994).

It has been found that ash with low pH is sticky drops of superheater section are always lower than
and hard to remove form surfaces (Tran, 1988, 10 Pa. The pressure losses in economizers and
Uloth et al., 1996). In practice it has been found boiler bank increase during the first few days after
that economizer ash is not sticky and is easy to start-up because boiler load is increasing. Then a
remove if the pH of 10 % of dust in water solution level of stable pressure difference is achieved. The
is at least 10. pressure difference in the superheater area and
economizers changes only a little after that. Vak-
With higher dry solids the economizers plug far kilainen and Niemitalo (1994) observed that after
less often and are seldom the cause of serious two months of operation no significant change
problems. in superheater, economizer and boiler banks had
occurred, Figure 8-5. The boiler bank fouling
Use of temperature seems to be slower and a more continuing process.
This increase is often very slow. With modern in-
measurements to predict fouling strumentation the overall fouling of boiler banks
and economizers can be predicted by pressure loss
Flue gas temperatures are measured in several lo-
measurements. The pressure loss measurement
cations. These typically include char bed tempera-
does not tell where exactly in the heat transfer
tures, furnace temperatures from different heights,
surface is fouling.
boiler bank inlet and flue gas exit temperatures.
The furnace temperature measurements are done
mainly for operating purposes and are not nor- Predicting superheater fouling
mally used to predict fouling. Temperature meas-
urements at superheaters, boiler bank and econo- Superheater fouling can be predicted by using
mizers can be used to predict fouling. Operating steam side temperature raise. These temperatures
temperatures rise when heat transfer surfaces get are reliably and easily measured. At a constant
fouled. The main difficulty in predicting fouling steam flow rate, the steam temperature raise is di-
with flue gas side temperature measurements is rectly proportional to the heat flow to that surface.
the limited number of measuring points. Accurate When the superheating decreases, the fouling in
prediction of fouling at individual superheaters that heating surface has increased.
needs far more gas side temperature measure-
ments than are usually available. The daily averages of superheating right after
the start-up of a recovery boiler are presented in
Figure 8-6, (Vakkilainen and Niemitalo, 1994). It
Use of pressure measurements to shows that each superheater achieves its equilib-
predict fouling rium fouling degree in two weeks. After that the
superheating in all superheaters is fairly constant.
If we position two gas side pressure measurement
elements one before the heat transfer surface and The attemperating flow is another good index of
one after the heat transfer surface we can use the overall superheater fouling. As the heat flow to su-
pressure difference information to indicate foul- perheaters decreases less attemperating is needed.
ing. Measured pressure loss is affected also by the The attemperating water flow describes how much
flue gas flow. The flue gas flow can be assumed to the superheated steam has to be cooled to keep
be directly proportional to the boiler load. The the main steam flow temperature constant. Clean
most reliable boiler load measurement is main heat transfer surface conducts heat better and
steam flow. To compensate for the changes in the superheating is more effective. Attemperating
boiler load, the measured pressure differences can also be measured as the temperature decrease
can be scaled with the square of the steam flow as between one superheater outlet to next inlet.
follows:
Figure 8-7 (Vakkilainen and Niemitalo, 1994),
dp* = dp (qnom /qsteam)2 8-1 presents the attemperating flows between the
superheaters as daily averages after start-up. The
where lowest curve, attemperating flow I, is between
dp* is the scaled pressure difference, Pa primary and secondary superheaters. The middle,
dp is the measured pressure difference, Pa attemperating flow II, is between secondary and
qsteam is the measured steam flow, kg/s tertiary superheaters. The highest curve is the
qnom is the nominal steam capacity, kg/s sum of these two. Directly after start-up, when
the superheaters are clean, the attemperating flow
In practice only the pressure drops of economiz- is at its maximum. Then the attemperating starts
ers and boiler bank are significant, Figure 8-5 decreasing as superheaters are fouled. The changes
(Vakkilainen and Niemitalo, 1994). The pressure in attemperating flow are small after two weeks of
Fouling 8-5
operating. The boiler operating mode affects the furnace there are small metal oxides and other
superheating and attemperating flow. impurities present, which will serve as start-
ing nucleus for the vapors. This process is called
heterogeneous condensation. After formation of
8.2 FORMATION OF ASH nucleus the vapor condensation to these particles
PARTICLES IN RECOVERY keeps supersaturation low.
BOILER Formed very small particles grow to a uniform
In kraft recovery boilers the ash is formed by size. Much of the particle size growth is result of
several mechanisms. It is known that black liquor agglomeration. Small liquid particles collide with
properties, operation practices and furnace condi- each other and form new spherical particles. Re-
tions affect the ash content and ash properties in sulting particle size is less than 10 µm. Measured
the flue gases (Frederick and Vakkilainen, 2003, mass averaged fume particle average sizes are 0.5
Hupa, 1993, Janka et al., 2004, Tran et al., 2000, … 1.3 µm, Figure 8-8 (Bosch 1971). Furnace con-
Vakkilainen, 2002). ditions especially the black liquor dry solids with
load and liquor species affect the average particle
Alkali metals in burning black liquor droplets va- size (Mikkanen, 2000).
porize. They subsequenly react with furnace gases
and condense to small particles. This process is Measurements of fume particle mass distribu-
called fume formation. Most of the ash in recovery tion indicate that the mean particle size does not
boiler flue gases is fume. Almost all ash collected change from the size seen at the furnace outlet
in the electrostatic precipitators is fume. level (Mikkanen et al., 1999, Baxter et al., 2001).
After flue gas temperature decreases below about
Some of the black liquor fired into recovery boiler 550 oC fume particles that hit each other do not
furnace gets caught with flue gas. Furnace gases stick. But particles that touch each other can
drag these burning particles with them out of the grow together through a process called sintering
furnace. This process is called carryover particle (Duhamel et al., 2004). Sintering takes place at
formation . Rate of carryover depends on firing particles deposited on heat transfer surfaces. There
practices especially droplet size. In addition fume can form large agglomerates. These agglom-
material that can form carryover is physically erates can re-entrain to flue gas when sootblowed.
ejected from burning black liquor droplets (Verrill They show as larger 20 … 30 µm particles at elec-
and Wessel, 1998). Most of the ash depositing to trostatic precipitator inlet, Figure 8-9 (Mikkanen,
superheater surfaces is carryover. 2000, Janka et al., 2000). Amount of re-entrained
dust is 10 … 30 % of total dust flow (Tamminen et
It has also been observed that during char com- al., 2000).
bustion intermediate size particles. are formed
in the char bed (Kochesfahani et al., 1998). They Fume deposits at tube surfaces consist initially of
probably contribute both to fume and carryover fine filaments with large aspect ratios and high
formation and deposit on superheater and boiler porosity (Frederick and Vakkilainen, 2003, Baxter
bank(Mikkanen, 2000). et al., 2004).
Most of the fume tends to end up in electrostatic

Fume precipitator. Figure 8-10 shows ESP dust as func-
tion of black liquor dry solids from about twenty
In the kraft recovery furnace the alkali elements
recovery boilers. Furnace temperature increases
are vaporized. When flue gases are cooled the va-
with increased black liquor dry solids. This in-
pors become supersaturated. Another mechanism
creases sodium and potassium release (Frederick
of supersaturation is formation of species by reac-
et al., 1995). ESP ash flow forms an estimate of
tions. E. g. sodium hydroxide and sulfur dioxide
fume formation in recovery boiler. Black liquor
may react to form sodium sulfate, which has lower
sodium and potassium release depends also on the
vapor pressure than the reactants (Jokiniemi et al.,
liquor i.e. wood species and the pulping condi-
1996).
tions (Backman et al., 1999). In addition the firing
conditions affect the lower furnace temperature.
Supersaturated vapors may form particles via
As can be seen the actual fume formation rate as a
two mechanisms. In homogeneous nucleation
function of black liquor dry solids can vary a lot.
two vapor molecules stick together to form a
small particle. For this a vapor pressure much
Dust (or fume) formation is almost directly pro-
higher than unity is needed. Vapors then start to
potional to the black liquor firing rate, Figure 8-11
condense on these particles. In recovery boiler
(Tamminen et al., 2002). Chemical equilibrium
Figure 8-7, Attemperating flow as the sum of temperature drops
between the superheaters II - III and IB – I (Vakkilainen and
Niemitalo, 1994).
Figure 8-8, Measured fume size distribution from the furnace,

sampled at nose before screen tubes (Mikkanen et al., 1999).
Figure 8 9, Size distribution of dust samples, collected before electrostatic precipitator,

at different operating conditions (Janka et al., 2000).
Fouling 8-7
Figure 8-10, Measured ESP dust concentration from recovery boilers.
calculations (Perjyd and Hupa, 1984) suggest that light streaks against darker background. Detec-
fume formation increases as the lower furnace gas tion devices to monitor combusting carryover are
temperature increases. This is shown as slightly available and are widely used. Flue gas tempera-
upward trend in the figure 8-10. tures drop fast after the boiler nose. This means
that the carbon reactions practically stop when the
Briscoe et al. (1991) reported that in their meas- carryover particle reaches superheater. Carryover
urements chemical fume production remained that is combusting at nose shows as unburned car-
about constant even though lower furnace gas bon residue in the deposits. It can be seen as black
temperature varied. Increased temperatures in color on boiler bank ash coming out of hoppers.
lower furnace didn’t result in decreased TRS emis- Carryover that has combusted totally has high
sions. reduction and contains sodium sulfide. Even if
combustion was over early, sulfation of sulfide
is slow for carryover size droplets (DeMartini,
Carryover 2004). High flow of fully combusted carryover can
Carryover is burned and unburned particles that be seen as pink or red color at superheater soot-
get caught with the flue gases. Very small particles blower lances. In modern high dry solids boilers
burn fast and form liquid droplets which don’t get carryover carbon content is very low. Carryover
entrained with the flue gas. Large particles do not particles are depleted of potassium and chlorine
have time to swell and hit furnace walls or char (Khalaj et al., 2004).
bed before they entrain. It is the fraction around
initial 1 mm size that is responsible for entrainmet Intermediate size particles
(Horton and Vakkilainen, 1993).
The actual coarse material that is seen on super-
For kraft recovery boilers running close to MCR heater section is very diverse. There are particles
equipped with modern air systems the typical that are larger than fume but smaller than the
carryover rate is 2 – 4 g/m3n (Kaila and Saviharju, residue of medium sized black liquor droplet.
2003). The amount of carryover depends on the These particles are called the intermediate size
air flow settings and the chosen air model (Costa particles. In industrial measurements (Mikkanen
et al., 2004). Typically carryover decreases for et al., 2001) at least five different types of interme-
lower loads. At about 85 % load and 70 % dry sol- diate particles have been identified based on their
ids liquor carryover rates in three boilers were 0,5 assumed formation path.
– 1,5 g/m3n (Backman et al., 1995). For older boil- 1. Large agglomerates which consist of 0.3±0.7
ers with conventional air systems carryover rates µm diameter primary particles. These
can be as high as 5 – 8 g/m3n (Metiäinen, 1991). particles form when the fine fume particles
deposit onto the heat exchangers, sinter and
Carryover can be broadly classified into two dis- re-entrain to the flue gas.
tinct groups. Carryover that is still burning when 2. Extensively sintered irregular particles
it reaches the boiler nose elevation can be seen as 30±250 µm in diameter. The appearance

Figure 8-11, Dust generation as a function of boiler load (Tamminen et al., 2002).
and the composition of these particles uted most of this loss to physical ejection during
suggest that they were formed from drying and volatilization. The release of ISP from
extensively sintered fine fume particles and char bed seems to increase with increasing char
subsequently removed from the heat bed temperature (Verrill and Lien, 2003). Total
exchangers by soot blowing. mass release at 1000 oC seems to be about 8 % of
3. Spherical particles 5±100 µm in diameter, initial char mass. About 75 % of the released mass
which appeared highly porous inside. These seemed to be ISP (Kochesfahani et al., 1998). This
particles are enriched in potassium. would mean that 10 … 30 % of Na loss or 2 … 4
Formation of these particles is most likely g/m3n is ISP from char bed.
through residue of burning black liquor.
4. Dense spherical particles 5±250 µm in The amount of intermediate size particles in typi-
diameter. These particles are enriched in cal recovery boilers is still somewhat open. Wessel
Na and K and depleted in Cl and C. and Kaufman (2000) indicated in their recovery
Furthermore, these particles deposit boiler model that about half of all particles could
efficiently in the superheater area. be ISP. In field studies at two US boilers between
5. Irregular, rough particles 3±40 µm in 1 and 2 g/m3n have been found (Schaddix et al.,
diameter. These particles have Si, Ca and 2003, Baxter et al., 2001, Lind et al., 2000). In an-
sometimes traces of Mg in their spectra other set of field studies at three North American
suggesting that they originate from the boilers 20 to 40 % of large material at superheater
mineral impurities in the black liquor. inlet was ISP (Kochesfahani and Tran, 2000). The
Some of these particles might have formed measured ISP amount equaled 1 … 3 g/sm2. Mass
on the evaporator. averaged size of ISP was about 50 µm.
Type four intermediate particles can form two

ways. Small particles can be physically ejected 8.3 DEPOSITION OF PARTICLES
during black liquor droplet combustion (Ver- AND VAPORS
rill and Wessel, 1998). During char combustion
release of 10 to 100 µm char fragments happens Deposition of particles and vapors on heat trans-
in laboratory experiments and pilot plant tests fer surfaces can occur through multiple pathways
(Kochesfahani et al., 1998). In mill trials this size (Goerg-Wood and Cameron, 1998). Thermo-
particles have been found, Figure 8-12 (Kochesfa- phoresis or temperature gradient between flue gas
hani and Tran, 2000, Mikkanen et al., 2001). and deposit surface transfers small particles to
surface. Larger particles are deposited by inertial
The release of ISP from combusting black liquor or turbulent impaction.
has not yet been measured. Verrill et al. (1994)
measured sodium loss from combusting black Vapor or gaseous species can condense on the
liquor droplets and found 5 … 20 % of sodium surface and react with deposits on the surface.
was released during the first seconds. They attrib- Concentration differences play role near the heat
Fouling 8-9
the ratio of the particle stopping distance, lp, to
the characteristic cylinder dimension, that is to
the cylinder radius Rc. Impaction of particle on
the cylinder surface will occur only, when Stk >
0.125 (Friedlander, 1977). The stopping distance
is a finite distance, which a particle will travel
before coming to rest as t → ∞, when the particle
is projected into a stationary fluid with a velocity
Up,0. The stopping distance is usually calculated
by applying Stokes linear drag law (Friedlander,
1977)
lp = Up,0τ 8-4
where τ is the particle relaxation time, defined as

Figure 8-12, Intermediate size particles (Kochesfa-
hani and Tran, 2000). τ = mp/3πdpη = ρpbdp2/18η 8-5
transfer surface. For very small particles (<0.1 and where

µm) molecular diffusion and Brownian motion mp is particle mass,
can play a role. Eskola et al. (1998) have presented dp is particle diameter,
formulas of deposition rates. η is the dynamic viscosity of host gas and
ρpb is the particle bulk density.
Thermophoresis However, when particles do not obey Stokes linear
drag law, that is when Rep >1, stopping distance
Thermophoresis plays significant role in capturing
also depends on the particle Reynolds number,
the small 0.2 … 1 µm fume particles suspended
defined as
in the flue gas. When hot flue gas flow past colder
heat transfer surfaces temperature gradient forms.
Rep = Udp/ν 8-6
This temperature gradient forces small particles
towards cold surface. Thermophoretic deposition
where U is the free stream velocity. Fortunately,
rate is proportional to the temperature gradi-
this dependency can also be taken into account,
ent, particle diameter and particle concentration
when stopping distance is defined as (Israel and
(Goerg-Wood and Cameron, 1998).
Rosner, 1983)
DT = KBddp(Tg-Ts)/Tl 8-2
8-7
and
K is constant, -
Bd is dust concentration, kg/m3
dp is particle diameter, m where r is the gas density. Using this form of the
Tg is gas temperature, K stopping distance, we are able to generalize the
Ts is surface temperature, K Stokes number for particles, which do not obey
Tl is gas temperature on the outer edge of the Stokes linear drag law. This new effective Stokes
laminar layer, K number is defined as
Inertial impaction
8-8
For particle sizes larger than about 20 µm the
main deposition mechanism is the inertial impac- where Ψ(Rep) is the non-Stokesian drag correction
tion. The capture efficiency of a collector(cylinder) factor (Israel and Rosner, 1983, Wessel and Righi,
in a cross flow of combustion gases is mainly 1998), which is monotonously decreasing function
determined by inertia, which can be characterized of Rep and Stk is the same Stokes number, which
with the Stokes number was already defined in equation (8-4).
Stk = lp/Rc 8-3 Collection efficiency, which is our interest, is

defined as a ratio of the actual particle mass col-
As can be seen, the Stokes number is defined as lection rate of the collector to the rate at which
mainstream particles flows through the projected for a particle and particles initial velocity in the
area of the target, that is model is the same as mean fluctuation velocity in
the fluid mainstream, . Particles fluctuation-
imparted stopping distance, calculated at the outer
8-9
edge of laminar sub layer, is
where ρp(i),∞ is the mass content of particles in a lpf = t. 8-13

flue gas. Next we need to be able to form the fol-
lowing relationship between collection efficiency Because gas turbulent velocity is negligible in the
and effective Stokes number: laminar sub-layer, particle can deposit at the wall
only if
ηcoll = f(Stkeff(Rep)) 8-10
lpf > l 8-14
Fortunately, using a potential flow field approxi-
mation to the fluid motion about the collector, the where is the thickness of the laminar sub-layer.
collection efficiency of the cylindrical target has Substituting equation (8-14) into equation (8-15)
been calculated numerically (Israel and Rosner, we get a following form
1983, Wessel and Righi, 1998) as a function of
effective Stokes number. These numerical results
have also been correlated as a function of Stkeff. 8-15
One recommended collection efficiency curve-fit
equation for Stkeff > 0.14 is (Israel and Rosner, This inequality clearly shows that the smaller the
1983): particles are, the greater is the mean fluctuation
velocity in the fluid mainstream , required for
ηcoll(Stkeff ) = {1+1.25(Stkeff-0.125)-1- their deposition at the wall. Theoretical models for
0.014(Stkeff-0.125)-2+ particles turbulent impaction, which are shown
0.508*10-4(Stkeff-0.125)-3}-1 8-11 in the literature, are modelled assuming fully
developed turbulent flow in pipes. All of these
However, more accurate two-equation correlation models include parameters, which are based on
for the collection efficiency of the cylinder is given experiments. Most of these models fall close to
by (Wessel and Righi, 1998). They give a cor- the experimentally determined dimensionless
relation for the local collection efficiency, which deposition velocity. Comprehensive treatment of
enables to calculate local deposition rates on the particles turbulent impaction used in our calcula-
cylinder surface as a function of cylinder angle. tions can be found from (Im and Chung, 1993).
Turbulent eddy impaction Vapor deposition

When host gas and particles flow over the plane Vapors may condense by direct condensation on
parallel with in the isothermal system, one can heat transfer surfaces or on the particles inside
observe notable deposition of large particles the boundary layer. Condensing vapor may also
although Brownian diffusion coefficient of these react with surfaces. For direct vapor condensation
large particles is very small. This is possible, not flux to the surface we get a following expression
by way of Brownian or turbulent diffusion, but by (Eskola et al., 1998)
way of turbulent impaction, which is important
deposition mechanism for large particles, when
boundary layer between surface and host flow is 8-16
turbulent. Inside the turbulent boundary layer
turbulent eddies have velocity components, which where
are normal to the main flow and are capable to Sh is Sherwood number,
give enough momentum for particles to cross the Dv is vapor diffusivity,
laminar sub-layer and finally deposit on the wall. dc is cylinders characteristic dimension,
This phenomenon can be understood using previ- Rg is the gas constant,
ously defined stopping distance model. In this pv is actual vapor pressure,
case eddies lose momentum pv,s is the saturation vapor pressure
Tg is gas temperature, K
mp 8-12 Tw is wall temperature, K
Fouling 8-11
Deposition to surfaces boiler heat transfer surfaces determines many of
the deposit properties. The first melting tempera-
Deposition rate is controlled by particle size ture (FMT) is the temperature at which the melt
distribution, temperature difference, heat transfer appears in a recovery boiler deposit. The tempera-
surface construction and ash content in the flue ture where deposit melting appears marks also the
gases. It is therefore obvious that deposition rate region where fouling problems occur. Under-
to different recovery boiler surfaces is not constant standing and predicting first melting temperature
but varies. is particularly important in terms of superheater
corrosion (Tran et al., 1999).
Fume tends to deposit on economizers and boiler
banks, Figure 8-13 (Jokiniemi et al., 1996). Fume Liquid content
deposition occurs mainly through thermophore-
sis. Carryover tends to deposit on superheater When recovery boiler dust is heated it at first
surfaces. If we look at deposits on these surfaces behaves like solid. At first melting temperature liq-
we see a clear difference on particle size that forms uid appears. Above that temperature there exists
these deposits. Boiler bank dust contains larger, a state where the deposit is partially molten and
carryover particles mixed with smaller fume, partially solid. Depending on the type of system
Figure 8-14. Economizer deposits tend to contain after small or larger temperature difference the
mostly fume, small submicron size spheres. whole system is molten.
Figure 8-16 presents two typical binary systems.

8.4 PROPERTIES OF RECOVERY Left is eutetic mixture. Typical eutetic mixtures are
BOILER ASH for example systems Na2SO4-Na2S and Na2SO4-
NaCl. At right there is a minimum melting system.
Deposits in recovery boiler heat transfer surfaces Typical minimum melting systems are for example
are not static but depend on process conditions. systems Na2SO4-K2SO4 and Na2SO4-Na2CO3. In
The ash deposition is often time dependet i.e. the eutetic mixtures the components do not mix
ash properties change in a deposit cross section. together but remain separate through heating and
The most important ash property is its strength. cooling. In minimum melting mixtures the com-
It depends on ash melting behavior and sintering pounds are reasonably similar so they can form a
rate. complete solid structure with each other.
First melting temperature It is typical that in salt systems melting tem-

peratures of mixtures are lower than individual
Melting behavior of deposits in kraft recovery melting temperatures. In Figure 8-16 the melting
Figure 8-13, Fume deposition on recovery boiler heat transfer surfaces (Jokiniemi et al., 1996).

Figure 8-14, At the left precipitator dust (small spheres), at the right boiler bank dust (mixture of different
sizes).
Figure 8-15, Effect of carbonate content to first melting temperature of synthetic

dust containing chloride at various potassium contents (Tran et al., 1999).
Figure 8-16, Basic phase diagrams for binary systems: left eutectic and right minimum melting (Tran et al.,
1999).
Fouling 8-13
Figure 8-17, Predicted liquid contents as a function Figure 8-18, Effect of potassium and chloride on
of temperature for recovery boiler dusts (Backman recovery boiler dust sticky temperature (Backman et
et al., 1996). al., 1995)
temperatures of component B is higher than com- the wall. If there is too much water the plaster will
ponent A and the mixture melting temperatures again will fall off the wall But between these states
are well below the melting temperature of pure the plaster will stick to the wall.
compound B.
Liquid content can be predicted with computer
The temperatures where 15 w-% of the deposit programs (Backman et al., 1996).
is at molten state and 85 w-% is at solid state is
called T15 or sticky temperature. The temperatures Sintering
where 70 w-% of the deposit is at molten state
and 30 w-% is at solid state is called T70 or flow Sintering is a process where dust particles that
temperature. Recovery boiler dusts tend to stick to have stuck to a heat transfer surface gradually
the heat transfer surfaces and form hard deposits densify and harden. Originally loose and porous
when they are between T15 and T70. An example material becomes solid almost crystal like. As
showing the effect of water content on material deposits densify, their strength increases exponen-
behavior is plaster. If it is dry it will fall off from tially, and they become more difficult to remove
Figure 8-19, Progression of neck development and grain growth during sintering of an ash sample from the
electrostatic precipitator of a kraft recovery boiler. Conditions were 450 oC for three minutes upper right fifteen
minutes lower left, and sixty minutes lower right (Frederick et al., 2001).

Figure 8-20, SEM photomicrograph of a cross-section of a deposit formed on a deposition probe inserted in
the mid-boiler bank region of an operating kraft recovery boiler (Frederick et al., 2001).
by sootblowing. Because of the end product the In real recovery boiler deposits the sintering
process is often called deposit hardening. behavior is much more complex. These deposits,
Figure 8-20, contain fine fume, large carryover
Fume particles sinter at temperatures above particles and intermediate particles. Each particle
300 oC (Tran et al., 1988). The rate of sintering type has its own composition and so its own
increases rapidly with temperature (Skrifars et sintering behavior. The actual deposit sintering
al., 1991). The recovery boiler deposits sintering behavior is then sum of sintering of all of these
and hardening is affected also by deposit packing different materials.
density, dust and flue gas composition.
These fine particles apparently sinter and harden 8-17

into difficult-to-remove boiler bank deposits
by two mechanisms: (a) evaporation of sodium
chloride from the particles and recondensation, Lien et al. (1999) measured the rate of linear
forming a neck at the point of contact between shrinkage (∆L/L) of pellets pressed from particles
two particles, and (b) transport of mass from collected in the electrostatic precipitators of eight
the particles to the neck by solid-state diffusion different recovery boilers. They correlated the
(Frederick et al., 2001). Figure 8-19 illustrates the linear shrinkage data with the chloride content of
changes that occurred during sintering of fine the dusts, the temperature at which the particles
particles. As sintering proceeds, the neck between were sintered, and time. Equation 8-17, the cor-
the fine particles grows until ultimately individual relating equation, was adapted from Kingery and
particle shapes can not be seen. Beyond that point Berg (1955), who used it to correlate sintering
the grain size of the structure grows as mass from rates for particles that sintered by the evaporation-
smaller, individual particles is transported to the condensation mechanism. Raoult’s law was used
larger ones. These phenomena have been observed to express po, the partial pressure of NaCl in this
both in laboratory studies (Skrifars et al., 1991, case, in terms of the pure component vapor pres-
Tran et al., 1988) and in deposits formed in oper- sure of NaCl and the mole fraction NaCl (yNaCl) in
ating recovery boilers (Frederick and Vakkilainen, the dusts.
2003).
Fouling 8-15
Figure 8-21, Examples of ash adhesive forces as function of temperature (Fujisaki et al., 2003).
Strength of deposits capacity is limited, it is important to find out how

different operating parameters affect the fouling.
Deposit strength depends only on its porosity
(Kaliazine et al., 1997). The following equation
has been derived by fitting the bending strength
Enrichment of potassium
versus porosity data of Piroozmand et al. (1998) Potassium compounds are more volatile than
for sintered pellets made from recovery boiler ESP sodium compounds. They are typically enriched
catch (Frederick et al., 2003) in the precipitator dust (Vakkilainen et al., 1995).
The amounts of sodium and potassium depend on
σ = 42800 e6.64 P 8-18 the amounts of other compounds in the sample.
Therefore potassium content is typically expressed
where as the molar ratio of potassium to the sum of
P is the deposit porosity, - sodium and potassium. The enrichment or enrich-
σ is the deposit strength, Pa ment factor is usually defined as the ratio the of
potassium content of the dust sample to the potas-
At some temperature a part of the deposit starts sium content in virgin black liquor, i.e.
to melt and forms liquid. This point is called the
initial melting point or the first melting tempera- EK = Ksample/(Nasample+Ksample) / KBL/(NaBL+KBL) 8-19
ture (FMT). Compression strength reaches its
maximum around the first melting temperature, where
Figure 8-21. Kxx is the potassium content at xx as mol-%
Naxx is the sodium content at xx as mol-%
8.5 DEPOSIT BEHAVIOR Vakkilainen et al. (1995) indicated that potas-
sium enrichment factors range from 1 to 2.5. This
In recovery boiler furnaces alkali metals are va-
means that the weight fraction of potassium in
porized. They react with sulfur, chloride and car-
ESP dust is about 2.5±0.5 times the potassium in
bon to form fume. In combination with combus-
virgin black liquor. Another way to express this is
tion residue from lower furnace they combine to
the mole ratio of potassium to the sum of sodium
form deposits on boiler surfaces. Several studies of
and potassium in ESP dust is about 4.6±0.9 times
recovery boiler plugging and fouling mechanisms
the mass weight percent of potassium in virgin
have been reported. The effect of potassium and
black liquor.
chloride on recovery boiler fouling mechanisms
has been studied by (Backman et al., 1987, Tran,
The data from twenty-one operating boilers,
1988, Skrifvars, 1989). A preliminary experimen-
Figure 8-22, show that the potassium enrichment
tal investigation of potassium and chloride enrich-
decreases with increasing carbonate content.
ment was reported by (Reis et al. 1994).
This data implies that the potassium enrichment
decreases with increasing furnace temperature.
Significant progress has been made in understand-
They found no correlation between potassium
ing recovery boiler fouling. When recovery boiler
enrichment and other operating variables.
Figure 8-22, Potassium enrichment factors in ESP dust versus carbonate.
Enrichment of chloride of less than 1.5. Another way to express this is the
mole ratio of chloride to the sum of sodium and
Chloride compounds in the lower furnace have potassium in ESP dust is about 8±1 times the mass
higher vapor pressures than other sodium and weight percent of cloride in virgin black liquor.
potassium compounds (Hupa, 1993). They are This means that the weight fraction of chloride in
typically enriched in the precipitator dust. Chlo- ESP dust is about 4±2 times the chloride in virgin
ride content is typically expressed as molar ratio black liquor. As with the potassium enrichment
of chloride to sum of sodium and potassium. The factors, this is less than data reported in other
enrichment or enrichment facor is usually defined sources (Tran and Reeve, 1991, Adams and Fred-
as the ratio of chloride content in dust sample to erick, 1988).
chloride content in virgin black liquor, i.e.
Chloride enrichment versus furnace temperatures
ECl = Clsample/(Na2,sample+K2,sample) / ClBL/(Na2,BL+K2,BL) indicated by the CO3 content of the dust gives two
separate enrichment regions. When the ESP dust
where CO3 content was below 4 weight-% the data was
Clxx is the chloride content at xx as mol-% scattered. Above an ESP dust CO3 content of 4
weight-%, the data were much less scattered and a
The data from twenty-one operating boilers, Fig- decrease in enrichment with increasing carbonate
ure 8-23 shows that in these boilers the chloride content was noted. This implies that chloride en-
enrichment ranges from 0.3 to 6. The average richment decreases with increasing temperature.
chloride enrichment is 2.9 with standard deviation Smaller fume particles are enriched in chloride
Figure 8-23, Chloride enrichment factors in ESP dust versus carbonate
Fouling 8-17
Figure 8-24, SO2 content of gases concentrations at the boiler flue gas exit level versus temperature at the
secondary air level for four different kraft recovery boilers.
(Mikkanen, 2000, Välimäki and Salmenoja, 2004). SO2 and furnace temperature affected the chloride
enrichment.
Effect of SO2 on chloride
Replotted Figure 8-25, shows enrichment values
enrichment that first decrease to less than unity and then
increase sharply when SO2 decreases to 0 to start
(Frederick et al., 1998) reported data on SO2
another decline with further temperature increase.
concentrations at the boiler flue gas exit level and
This is consistent to data presented in Figure 8-23.
temperature at the secondary air level (~2.5 m) in
In real boilers, because of mixing limitations and
the lower recovery furnace for four kraft recovery
non uniformity of temperature fields the sharp
boilers. The SO2 concentration decreased with
increase should be much more gradual.
increasing temperature in the lower furnace for
each boiler, Figure 8-24. These characteristic SO2
versus temperature curves were shifted to different ESP dust enrichment
locations along the temperature axis for the differ-
ent boilers. The probable reason for the shift is the Based on more than 200 tests with analysis of ESP
different sulfur release rate for each liquor. dust and black liquor, it is possible to predict ESP
dust composition as a function of black liquor
Using data from Figure 8-24 and measurements composition and dry solids;
of chloride enrichment factors in a laboratory - Prediction of SO2 is based on (Maso, 1988)
combustor Frederick et al. (1998) studied how - Depending on lower furnace temperature
Figure 8-25, Prediction of enrichment of chloride to dust as function of lower furnace temperature.

Figure 8-26, Prediction of potassium and chloride enrichment in a recovery boiler as a function of black
liquor dry solids.
there are two operating regions; with and data prevents from verifying this trend.
without SO2
- When boiler operates without SO2 then
ESP dust analysis is based on enrichment
BB dust enrichment
versus carbonate content data (Vakkilainen Boiler bank dust has a composition which is
et al., 1995) a mixture of fume and carry-over. Carry-over
- When boiler operates with SO2 then particles deplete the potassium and chloride levels
enrichment data is based on (Frederick et in dust (Vakkilainen and Niemitalo, 1994). The
al., 1998) mixture assumption seems to fit a wide range of
liquors and boilers (Janka et al. 1998, Mikkanen et
Predicted enrichment, Figure 8-26, shows gradual al., 1999). By setting carry-over to a known level,
decrease of potassium enrichment as a function of we can calculate the properties for this mixture in
black liquor dry solids for a recovery boiler with a typical recovery boiler (Eskola et al. 1998).
S/(Na2+K2) = 27 mol-%, Cl/(Na+K) = 0.90 mol-%,
K/(Na+K) = 12 mol-% and black liquor dry solids The predicted enrichment behavior, Figure 8-27,
75 %. This is consistent with data from (Reis et al., seems to fit the data presented by (Janka et al.
1994). The chloride enrichment shows a mini- 1998). The notable exception is the 60 % dry solids
mum enrichment at 62 % dry solids, followed black liquor. Because of chloride losses as HCl the
by an increase and then decline. The enrichment chloride enrichment should drop down as shown.
behavior of chloride at very high dry solids seems Trends like this need more industrial verification.
to reach a plateau. Lack of industrial operating
Figure 8-27, Prediction of potassium and chloride enrichment for carryover and fume for several black liquor
dry solids levels. Lower left carryover, middle boiler bank, upper right ESP ash.
Fouling 8-19
Figure 8-28, Prediction of effect of dry solids to boiler bank dust properties in a recovery
boiler, sulfidity = 27 %.
Figure 8-29, Prediction of effect of sulfidity to boiler bank dust properties in a recovery
boiler, dry solids =75 %.
Figure 8-30, Prediction of sintering index as a function of dry solids.

Based on the above described assumptions the In high solids boilers there seems to be low sulfide
model has been used to predict the effect of content in deposits (Raukola and Haaga, 2004).
changes in black liquor properties on different
fouling indicators. The predictions should not be Predicting deposit behavior
assumed to show absolute values, but rather show
trends which are applicable to interpreting operat- In addition to traditional indicators, ash sintering
ing data on recovery boilers. index can be calculated (Frederick et al., 2004).
Ash sintering index is based on the model by
Dust properties, such as sticky temperature can (Lien et al., 1999). In their model the sintering
be predicted based on data from (Backman et al., index was mainly a function of the first melting
1995). In the model flow temperature T70, sticky temperature. Lowest sintering rate was set as unity
temperature T15 and first melting temperature T0 and higher sintering rates are shown as ratios to
are based on ash chloride, potassium and carbon- the lowest rate. Figure 8-30 shows the effect of the
ate contents. Changes in these temperatures have dry solids on this sintering index for a recovery
been show to correlate with changes in fouling boiler.
behavior (Tran, 1988).
Dust sintering index seems to increase for ESP
Effect of dry solids to boiler bank dust. This is because of higher chloride and potas-
sium enrichment factors. Industrial practice seems
dust properties to indicate that in spite of this economizers do not
Figure 8-28 shows boiler bank ash behavior as a show increased fouling. It might be that because
function of dry solids for S/(Na2+K2) = 27 mol- of low temperature the economizer fouling is not
%, Cl/(Na+K) = 0.90 mol-% and K/(Na+K) = controlled by sintering rate. Figure 8-30 shows
12 mol-%. Increasing dry solids eliminates SO2 decreasing sintering rates with increasing dry
and HCl emissions after about 65 % dry solids. solids for boiler bank and superheaters. Industrial
One sees the sticky and flow temperatures in this practice indicates that boiler bank fouling decreas-
recovery boiler to be well above industry practice es with increasing dry solids.
for designing boiler bank inlet temperatures (Mc-
Cann, 1991). If the aerosol fume would strongly
affect the boiler bank fouling then this kind of
8.6 DEPOSIT REMOVAL
limit would make sense. Sootblowing steam exits the sootblower at or
above the sonic speed of steam. The velocity
Effect of sulfidity to boiler bank degrades fast, Figure 8-31. Decay is caused by the
dust properties interaction of turbulent jet with surrounding gas
(Tandra et al., 2001). As can be seen the effective
Sulfidity affects the recovery boiler fouling cleaning radius of a sootblower jet is 1 … 1,5 m.
through the same mechanism as dry solids. In-
crease in sulfidity shows the same way as decrease One of the most problematic cases are the older
in dry solids, Figure 8-29. The figure seems to types of heat transfer surfaces. Wrongly designed
suggest that for boiler that operate at low SO2 tube anti vibration ties and too large distance
emissions, lowering sulfidity might mean lower between sootblowers can cause problems. The
fouling. There are no industry cases where this remaining solution is to add more sootblowers to
has been tried. Higher solids firing should at first the offending heat transfer surface.
increase chloride enrichment as gaseous HCl is
converted to NaCl as sodium becomes available. Superheater deposit removal can be improved by
After a zero SO2 level is reached, enrichment chill-and-blow. Stopping black liquor firing but
should decrease because higher furnace tempera- using auxiliary burners decreases superheater
tures cause higher sodium release (Vakkilainen, temperatures. Because of differences in thermal
2000b). Potassium enrichment should decrease expansion stresses are created with superheater
with increasing dry solids because of the higher deposits and superheater tubes. Deposit removal
release. can then be done by mechanical means or by us-
ing sootblowers extensively at superheater area.
Higher solids firing should affect recovery boiler
the same way than decreasing sulfidity. Higher dry After recovery boiler has been on line for some
solids should worsen the economizer fouling, con- time the recovery boiler economizer area fouling
trary to industry observation, unless the fouling is appears to level out. Sootblowing does not seem
mainly controlled by the ash pH (Skrifars, 1989). to affect the heat transfer but lack of sootblowing
Fouling 8-21
Figure 8-31, Decline of exit velocity of supersonic turbulent sootblower jet (Kaliazine et al., 1998).
leads to fast increase of draft loss (Uloth et al., Change black liquor properties
1996b).
Black liquor properties affect the fouling rate
of a recovery boiler. The three main options are
8.7 HOW TO DECREASE increasing dry solids, decreasing chloride content
FOULING RATE in black liquor or lowering sulfidity. In addition to
these, changes in pulping can affect the black liq-
Fouling rate is a combination of three parameters; uor combustion properties, which, in turn, affect
amount of incoming particles, their ability to stick the furnace outlet temperature and so recovery
to the surface and the removal rate by sootblow- boiler fouling. An example of development of
ing (Tran, 2004). Many of these parameters relate potassium and chloride levels in a Scandinavian
to the patrticle composition especially potassium mill is shown in Figure 8-32
and chlorine content. There are several measures
that can be used to decrease fouling in a recovery Increasing potassium and chloride content will
boiler. Their effectivity to solve the current fouling increase recovery boiler dust sintering rate at tem-
problem should be weighed carefully. The cat- peratures above 450 oC (Skrifars, 1989, Hupa et al.,
egories are; affect the liquor properties, decrease 1990). Chloride concentration in white liquor can
flue gas temperatures, decrease carryover rate and be calculated from the approximate mill balance;
improve surface cleanability.
Clwl = 0.25*Clin*(1-ηS)/[1-(1-ηS)(1-ηW)] 8-21
Figure 8-32, Potassium and chloride levels in a Scandinavian mill.

Figure 8-33, Effect of sulfidity to boiler bank ash properties.
where Most effective means of reducing temperatures is

Clwl is the white liquor chloride content, g/l to reduce boiler load. If furnace performance can
Clin is the chloride input, kg Cl/tpulp, (1 … 6) be improved by a new air system of better liquor
ηS is the fraction of chloride lost through firing, then furnace nose temperatures can be
recovery boiler stack loss, (0.00 … 0.02) decrease. Addition of heat transfer surface helps
ηW is the fraction of chloride flow lost through situation downstream.
washing loss, (0.005 … 0.02)
The fact is that many older recovery boilers are
The best way to affect the chloride level is to try operated more than +30% over design capac-
to reduce the chloride input to the mill. Typically ity. Often the upper furnace heat surface area is
the main source of chlorides is the chips (Ferreira, too small for this capacity. The boiler bank inlet
2003). Surprisingly in many problematic cases temperature will be too high no matter how well
the main input has been as impurity in purchased the furnace is operating.
NaOH. Chloride level can also be reduced by
increasing losses. The most used method is to Decrease carryover
purge ESP dust (Nunes et al., 1995). Chloride can
be purged as HCl if SO2 emissions to stack are Improving furnace behavior by decreasing car-
allowed (Ibah, 1995). If chloride input to mill is ryover rates is often very beneficial. Carryover
high, then the chloride removal process should affects the fouling through two phenomena. Pos-
be considered. If the chloride is less than 0.5 w-% sible sulfide decreases dramatically dust sintering
in black liquor dry solids, no extra measures are temperature. Increased flow increases superheater
usually required. deposit rate, which can lead to increased fouling.
Sulfidity affects the recovery boiler fouling

through the same mechanism as dry solids. In-
Improve sootblowing
crease sulfidity shows the same way as decrease in If fouling is local the first priority is to improve
dry solids, Figure 8-33 (Vakkilainen, 2000b). cleanability at that site. This can be done by
improving sootblowing. Improvement can be
Dry solids increase decreases fouling, (Niemitalo, achieved with additional sootblowers, high
1993). The main improvement comes as the SO2 efficiency nozzles and a change in the sootblow-
emission is eliminated. Typically high sulfur emis- ing program. In practice the location and timing
sions indicate low ash pH. appear to be the most important factors when
determining the sootblowing efficiency (Uloth et
Decrease temperatures al., 1996).
Decreasing temperatures in the flue gas side of the

recovery boiler decreases the deposit strengths.
Fouling 8-23
9 Firing black liquor
Figure 9-1, Black liquor gun nozzle.
Black liquor can be fired in a recovery boiler in a

number of ways. The aim of firing strategy is to
obtain a reasonable fouling rate, operate within
the permitted emission limits and maintain stable
and secure operation (Poon, 2001). The opera-
tor can choose an air model that suits him. With
air model it is meant the division of air into the
different air levels and also how the ports in each
level are configured. The operator can also choose
how he wants to fire the black liquor guns.
To illustrate the firing of black liquor a number of

simulations has been made using a computer pro-
gram developed for that purpose (Järvinen, 2002,
Järvinen et al., 1997, Manninen and Vakkilainen,
1996, Horton and Vakkilainen, 1993).
9.1 LIQUOR GUN OPERATION

Black liquor is delivered to a recovery boiler
furnace by using liquor guns, Figure 9-1. The end
part of the liquor gun is called the nozzle. The
main types of nozzles are the splashplate nozzle,
U- and V-type nozzles, fan nozzle, willow flute Figure 9-2, Flow from a splash plate liquor gun,
and the swirl cone nozzle. There are many other spray from middle of the picture upwards (TKK)
Firing black liquor 9-1

perforation breakup the liquid sheet forms holes
randomly which rapidly enlarge (Spielbauer and
Aidun, 1992). Resulting black liquor strands break
to droplets. No matter how the droplets form, it
seems that the shape of droplet size distribution is
the same (Kankkunen and Miikkulainen, 2003).
When hot steam containing liquid is sprayed to at-

mospheric pressure the steam bubbles play a dual
role. They can form an instability that will break a
hole into the sheet causing perforation breakup. If
the steam amount is sufficient the sheet formation
Figure 9-3, Pressure and flow in a liquor gun fir-
does not occur and liquid droplets are formed
ing liquor above the BPR (Kankkunen et al., 1998).
directly (Helpiö and Kankkunen, 1995).
When the pressure of black liquor falls in the
types that are used and have been used, but mill liquor gun handle under the pressure that cor-
experience seems to lean towards these types. responds to the boiling temperature, bubbles
Usually, the nozzle size and type varies from one will begin to form, Figure 9-3. The boiling is not
recovery boiler to another. The spray properties usually sudden, but an increasing part of the black
from these nozzles differ to a great extent, but liquor flow is steam bubbles. The increase in the
every nozzle has an optimum operating point. firing temperature will be seen as an increase in
Nozzles are selected, based on previous experi- the spray velocity and as a decrease in drop size
ence, to keep the boiler operation effective and the and mass flow, Figure 9-4. Typically this means
droplet size optimal. that larger nozzles are required to maintain load.
The purpose of the black liquor gun is to from Boiling temperature of black liquor is defined as
small black liquor droplets to facilitate burning. sum of boiling temperature of water at same con-
Liquor flow inside a tube is either constricted or ditions and black liquor boiling point rise. Boiling
bounced of a plate to form a sheet of liquid, Figure point rise is liquor dependent but is mostly a func-
9-2. This sheet will break up and create droplets. tion of black liquor dry solids. Typically constant
With liquid sheet the droplet formation occurs difference to boiling point temperature produces
with two mechanisms. In wavy formation the ‘optimum’ firing conditions for a particular liquor
surface tension and drag help form waves from gun type and black liquor. At very high dry solids
local instabilities (Dombrowski and Johns, 1963). this temperature must be increased to compen-
These waves will break into ribbons, which will sate for the decrease in water volume in the black
collapse into cylindrical strands and ultimately liquor, Figure 9-5 and to keep the liquor viscosity
form droplets (Spielbauer and Aidun, 1992). In low enough for pumping.
Figure 9-4, Liquor droplet velocity at liquor gun exit as a function of the firing temperature (Vakkilainen
and Holm, 2001).

Figure 9-5, Black liquor as fired temperatures as function of black liquor dry solids.
After selecting the nozzle size and type, the liquor correlated with a simple two term equation.
flow is controlled by the liquor pressure and tem-
perature. A higher firing temperature has another Cf = a + bRenc 9-2
effect. The liquor flashes in the nozzle tube and
causes a bigger pressure loss due to the acceler- where
ated velocity, Figure 9-5. Thus, we have to raise Ren is Reynolds number at nozzle outlet, -
the liquor firing pressure to keep the boiler load a, b, c are constants, -
unchanged.
Reynolds number is calculated on the basis of the
Flow from black liquor gun smallest flow area of the nozzle. The minimum
flow area is a function of plate installation angle.
The nozzle pressure loss can be expressed through The values for the constants a, b and c are pre-
flow coefficient and flow dynamic pressure at sented in table 9-1 (Spielbauer and Adams, 1990).
nozzle. The flow coefficient takes into account the
friction and acceleration losses (Spielbauer and Table 9-1, Nozzle flow coefficient constants (Em-
Adams, 1990). pie et al., 1992).
Nozzle type a b c
Δpn = Cf ρu2/2 9-1
splashplate nozzle 1.17 373 -0.92
where fan spray nozzle 1.16 2780 -1.65

Cf is nozzle flow coefficient, -
Usually, the liquor firing pressure, temperature
The flow coefficient for industrial nozzles has been and volume flow rate are measured in the liquor
ring. For some boilers, only the total flow rate is
measured. A better knowledge and controllabil-
ity of liquor firing is achieved by measuring and
controlling the volume flow rate on each wall or
even through each of the liquor guns.
Effect of black liquor type

Black liquor has different properties in different
mills; even the same mill can have quite variable
liquor properties. Chemical cooking and wood
Figure 9-6, Black liquor as fired pressures as func- source affect the liquor properties. To illustrate the
tion of black liquor dry solids. effect of firing parameters we can make simula-

Figure 9-7, Black liquor droplet size for both liquor types and different firing pressures as a function of firing
temperature at 70 % dry solids.
tions on gun behavior. Softwood and hardwood cantly. This is because of a larger mass flow rate
liquor were studied (Järvinen et al., 1997). A range and higher outlet velocities.
of typical values is firing temperature of 110, 120
and 130 oC and pressures of 120, 140 and 160 kPa. In order to enlarge the softwood droplet size equal
to hardwood values, the firing temperature must
Table 9-2, Black liquor properties at 70 % dry to be lowered. Approximately a temperature de-
solids and 110 oC. crease of 5…7 oC is needed at the firing tempera-
Liquor type Hardwood Softwood ture range of 110…120 oC. A decrease of 2…3 oC
with a firing temperature range of 120…130 oC is
Viscosity, mPas 344 206
needed. Another way of enlarging the droplet size
Diameter ratio during 2.9 3.2 is to lower the firing pressure. Thus, the gun load
devolatilization, - would be smaller.
Boiling point rise, oC 13.9 13.9
When flashing, an increasing pressure decreases
A single throttled splashplate nozzle of 27 mm, the nozzle outlet quality meaning that the change
situated at the level of 6.2 m was used. The dry of the outlet velocity is less significant. The calcu-
solids content was 70 % all the time. The refer- lated initial velocities of the black liquor droplets
ence values of the properties of black liquor are are presented in Figure 9-8.
presented in Table 9-2. The diameter ratio during
drying was 1.5. An increasing firing pressure and temperature
Firing temperature and boiling point rise

determine the velocity at gun throat (Helpiö
and Kankkunen, 1995). Higher velocity leads to
smaller diameter droplets. The mass median drop-
let diameter is presented as a function of firing
temperature, Figure 9-7.
The black liquor droplet size of hardwood liquor is

larger than the droplet size of softwood liquor due
to its larger viscosity. The droplet size decreases
with increasing temperature due to the accelerated Figure 9-8, Black liquor droplet initial velocities
flow and the smaller black liquor viscosity. When for both liquor types as a function of firing tempera-
firing below the liquor boiling point, an increasing ture at 70 % dry solids for different firing pressures.
firing pressure decreases the droplet size signifi-
decreases the fraction of the liquor sprayed to the
char bed and increases suspension burning and
the fraction of the liquor reaching the furnace
walls. When firing liquor below the boiling point,
hardwood liquor increases the liquor amount
reaching the char bed. When firing under flashing
condition at 130 oC, fraction of liquor reaching
the char bed is smaller with hardwood liquor
compared with softwood liquor. This is due to the
increased liquor fraction ending up on the furnace
walls.
Figure 9-9, Flight paths of hardwood and softwood
It is known that hardwood liquor swells less
liquor droplets as a function of droplet size, t = 120
that softwood liquor. In order to compare liquor o
C, u = 8 m/s, horizontal spraying.
droplet combustion and flight path differences,
was combustion of two different droplets with a
diameter of 2 and 3 mm simulated. The flue gas The larger the nozzle, the larger the droplet size
upward velocity used was 5 m/s and the droplets and the smaller the nozzle outlet velocity. The
were fired horizontally with a velocity of 8 m/s. lower values of the quality with the smaller noz-
The flight paths of the droplets are presented in zles results from a higher pressure in the nozzle
Figure 9-9. Swelling of the black liquor affects inlet. On the basis of these results, combustion
the flight path of the droplet. Black liquor swells simulations for each case were made. The amount
normally by a factor of 2 to 3 in diameter during of the liquor sprayed to the walls, to the char bed
devolatilization. Liquors which swell much, loose and the carryover amount were calculated.
their momentum faster than non swelling liquors
(Manninen and Vakkilainen, 1996). There are no significant differences in drying. The
droplet has not swollen yet so the droplets tend to
Due to the larger swelling and the smaller droplet stay in the same vertical plane. Water is released
size of softwood liquor, softwood liquor droplets further away from the liquor gun as the droplet
have bigger friction forces with the result that size increases.
the penetration of the spray into the furnace is
smaller. This is considered the reason for the The center of gravity of carbon release is pre-
larger hardwood liquor amount ending up on the sented in Figure 9-10. This puts the whole carbon
furnace walls at higher firing temperatures. releases to a single point so half of release occurs
below and half above. Similarly half of the release
Table 9-3, Spray properties calculated as a func- occurs behind and half in front of the point. A
tion of liquor gun type, qm = 7.4 kg/s, tf = 136 oC, larger nozzle size moves combustion towards the
dry solids content 83.5 % opposite wall of the liquor gun due the longer
Nozzle Firing Outlet Droplet Outlet combustion times. A smaller nozzle size moves
size, press., velocity, size, quality, the combustion up higher. Smaller droplets tend
mm kPa m/s mm % to burn faster, break faster and are more carried
38 100 11 3.07 0.055 with upflowing flue gas.
33 128 12 2.66 0.039
The amount of liquor reaching furnace walls and
28 184 14 2.22 0.024 char bed is presented in Figure 9-11. The amount
of the liquor hitting the furnace walls decreases as
The effect of the liquor gun type on black liquor
combustion can be modeled. Järvinen et al. (1997)
studied firing with three different size liquor guns.
The boiler process values and dimensions were
based on typical Scandinavian practice. The objec-
tive was to find out how gun size affects firing.
The firing pressure was adjusted to keep the gun
load unchanged, i.e. the smaller the nozzle, the
higher the firing pressure. The firing temperature
was kept constant. A single throttled splashplate
nozzle, situated at the level of 8.6 m from furnace
bottom, was used. The liquor firing parameters are Figure 9-10, Center of gravity of carbon release as a
presented in Table 9-3. function of gun size, qm = 7.4 kg/s, t = 136 oC.

Figure 9-11, Fraction of the black liquor sprayed to Figure 9-12, Fraction the sulfur released of the total
the char bed and on furnace walls as a function of sulfur during devolatilization.
liquor gun size, qm = 7.4 kg/s, t = 136 oC.
a function of decreasing droplet size. This is be- at 73 % dry solids. This temperature is above the
cause of the shorter combustion times due to the atmospheric boiling point of the liquor.
smaller droplet size. The amount of liquor hitting
walls could also be controlled by gun type and The liquor was sprayed from guns without vertical
spraying angle, but here these were kept constant. velocity distribution. The sheet velocities, droplet
sizes and liquor mass flow distributions for all
The release of sulfur during the devolatilization nozzles were modeled as continuous functions of
was studied by calculating the total sulfur amount spray angle. These models are based on interpola-
released for each case. The result is presented in tion and extrapolation of experimental measure-
Figure 9-12. The larger the nozzle, i.e. the larger ments. Use of data from small scale model work
the droplet size, the larger the sulfur release. was needed for some parameters.
Five different nozzle types were studied (Helpiö

9.2 LIQUOR GUN TYPE and Kankkunen, 1995b, Adams, 1994, Helpiö and
Kankkunen, 1994, Empie et al., 1995). Nozzles
Järvinen et al. (1997) have studied the effect of one, two, three and five were of the splashplate
liquor gun type to the firing of typical softwood type. The orifice diameter and the position of
black liquor. Boiling point rise is a function of the splashplate varied depending on the manu-
black liquor dry solids content. Black liquor sheet facturer. Nozzle four was a fan spray nozzle. The
velocity increases due to liquor flashing in the dimensions of the nozzles are presented in Table
nozzle opening. Typical droplet burning param- 9-5. Operating parameters of the nozzles were
eters for softwood are listed in Table 9-4. selected by Reynolds number criteria i.e. the same
value for Re was used for every nozzle. In all cases,
Table 9-4, Softwood liquor burning properties. only one nozzle was used which was placed at the
Swelling factor during drying 1.54*Do center of the front wall and at height of 6 m. The
Devolatilization swelling factor 2.3*Do vertical spraying angle used was constant for every
case and it had a value of -10o.
Boiling point rise 16 oC
The horizontal direction of liquor droplets is

The firing temperature used was 123 oC, which is
determined by the liquor mass flow distribution.
typical for recovery boilers firing softwood liquor
If the nozzle opening angle is wide, liquor is usu-
C27 D28
Figure 9-13, Two typical splash plate liquor guns, left partially blocked, right totally open (TKK).

Figure 9-14, Mass flow distribution of the nozzles at 123 oC, Re = 3400, sector width 5 o.
Table 9-5, Nozzle properties. Droplet size

Type Description Open- Plate Diam-
ing angle, eter, One of the most important parameters in control-
angle, o o
mm ling recovery boiler furnace operation is the drop-
let size. Plenty of research work has been done in
1 Splashplate 85 25 21.6
throttled
order to study black liquor droplet formation and
liquor sheet disintegration. Full scale measure-
2 Splashplate 210 35 15 ment data was available for the droplet sizes of
nonthrottled
nozzles 1 and 2. The droplet size of nozzle 5 was
3 Splashplate 120 25 15 calculated based on nozzle 1. The droplet sizes of
at a distance nozzles 3 and 4 were based on measurements with
4 Fan spray 60 - 15.5 small scale models. Droplet sizes were then scaled
nozzle to full size values based on scaling criteria from
5 Splashplate 120 25 42 small and full scale measurement results using
at a distance dimensional analysis, Table 9-6. The nozzle 5 has a
droplet size much bigger than the others.
Table 9-6, Mass median diameter, MMD of the
nozzles. Release of water
Nozzle type MMD, mm
The evaporation of water consumes energy. If the
1 Ahlstrom 1.89
droplets are not dry when reaching the char bed,
2 Kvaerner 1.79 the temperature of the bed will decrease causing
3 Pantsar 1.79 unstable furnace operation. In the worst case, the
4 CE 1.97
water in the bed may lead to bed extinction.
5 B&W 3.78 It was observed that the water reaching the char
bed was insignificant with the nozzles spraying
ally sprayed to adjacent furnace walls. Mass flow smaller size droplets. Only the largest nozzle 5
distributions of the nozzles are presented in Figure delivered some water to the char bed, 5 % of total
9-12. The mass flow distribution of the nozzle 5 water amount. The maximum spraying of black
was similar to nozzle 3. The mass flow distribution liquor to adjacent walls occurred with nozzle 2
varies between geometrically different nozzles. due to a wide opening angle. The amount of water
Especially, the fan spray nozzle differs from the landing the walls with this nozzle was 13.5 % of
others because of the small opening angle used. total water.
Another exception is the nozzle type 2. Because
of wide angle of spray a great deal of liquor is The drying takes place in the vicinity of the noz-
sprayed close to walls. The mass flow rate distribu- zle. With all nozzles, most of the water is released
tion of a splashplate nozzle is shaped as a triangle within the first second. With increasing droplet
while the fan nozzle has a flat distribution. size, the drying moves farther from the nozzle in

Figure 9-15, Release of water in vertical direction. Figure 9-16, Release of mass during devolatilization
in vertical direction.
horizontal direction.
Devolatilization
During devolatilization gases are released from the
black liquor droplet and the diameter of the black
liquor droplets increases by a factor of 3 (Horton
et al., 1992). This means larger drag forces due to
larger cross-sectional area. If the friction forces are
high enough, entertainment takes place.
A clear difference between the nozzles can be seen

in the mass release during devolatilization, Figure Figure 9-17, Release of sulfur in vertical direction.
9-14. With nozzles 1, 2, 3 and 4, devolatilization is
completed in flight. With nozzle 5, many droplets
are still going through pyrolysis when they reach
the char bed.
With the small nozzles devolatilization occurs at

the vicinity of the liquor gun. The release distribu-
tion is flatter with bigger droplet size.
Sulfur release
The release model from (Manninen and Vak-
kilainen, 1995) was used for calculating sulfur
release during pyrolysis. It can be seen that the
larger the droplet size, the lower in the furnace the
sulfur release occurs, Figure 9-17. Figure 9-18, Release of sulfur in vertical plane, noz-
zle 2.
When comparing the total release of sulfur in the
furnace nozzle 2 released sulfur closest to the wall.
At the same time, the release distribution is flatter.
With wider opening, the maximum release point
moves towards the liquor gun wall.
The minimum sulfur release occurred with nozzle

5. The reason for this is the larger droplet size.
Calculated sulfur release was cut off when droplets
landed on the char bed. When the droplet size is
bigger, the release point of sulfur moves towards
the opposite wall from the liquor gun.
Sulfur release seems to occur along the flight path,

Figure 9-19. Sulfur release starts when the droplet
surface is dry. With larger MMD of droplets the Figure 9-19, Release of sulfur in vertical plane, noz-
sulfur release occurs farther away from the liquor zle 5.

temperatures and uncontrollable char bed growth
would result.
9.3 AIR DISTRIBUTION

Air is added to lower furnace for the combustion
and gasification reactions to proceed. There is
typically less than stoichiometric air added to keep
up reducing atmosphere. Because of the limited
reaction speeds there is however some 3 ... 12 % of
Figure 9-20, Release of carbon in the direction of O2 in the lower furnace gases.
nozzle center line.
Air addition happens typically in three stages.
- Primary air, which is added at the root of
gun.
the char bed.
- Secondary air, which is added at height of
Char combustion 1 ... 3 m above primary air.
- Tertiary air, which is added some 3 ... 6 m
Char combustion is important for proper furnace above the liquor gun level.
operation. If the droplet size is too small the char
combustion is completed before the droplets reach The two level air staging is in use only in older
the char bed. When liquor is fired with small noz- boilers. It typically produces higher emissions and
zles, the maximum release point is at the level of results in lower operating furnace loads than the
liquor gun. When droplet size is bigger, the release three level air.
point moves lower and more carbon is delivered
into the char bed. There have been some reported schemes to
increase the number of air levels from three to five
Horizontal mass flow pattern affects how effi- by splitting the primary and secondary air into
ciently the furnace cross sectional area is utilized. two levels. No benefits of these tries have been
The wider the opening angle and the bigger the reported.
droplet size, the better one liquor gun covers the
whole furnace. This is why gun 5 can be large. There are many conflicting results on how air
distribution affects sulfurous gas and particulate
Figure 9-20 shows the carbon release as a function emissions. If we assume that combustion is mix-
of distance from the liquor gun. With the small ing limited in the recovery boiler furnace, then
nozzles the release is rather flat. With bigger guns changing air flow rates and patterns affects the
much of the release occurs at the vicinity of the location of the main combustion zone. This affects
opposite wall and char bed. the furnace temperature pattern and so the char
bed temperature distribution. It is the temperature
The carbon release patterns for nozzle types 1 and profile that then controls the emissions.
3 are fairly similar. There are two main release
locations. For small and medium size droplets the
release occurs at the vicinity of the liquor gun. Primary air
Larger droplets have just completed evapora-
The role of the primary air is to shape the bed
tion and started devolatilization as they hit the
so that the sides angle towards the juncture of
opposite wall. The fan spray nozzle 4 behaves
the furnace wall and the bottom and provide air
like splashplate nozzles 1 and 3. Most of the
for char burning. There have been other tasks
carbon release is at the liquor gun level. Nozzle 2
attributed to primary air such as keeping the TSR
releases carbon fairly close to the gun. The liquor
emissions low (Jutila et al., 1978) and controlling
is sprayed in wide angle so carbon release occurs
the loading level.
along the liquor gun wall. This type of spraying
decreases the penetration of the liquor spray to
The amount of primary air flow depends on the
the center of the furnace.
boiler manufacturer and the way the boiler is
operated. Range of reported operating values is as
The biggest nozzle 5 seems to have a different type
high as 25 ... 60 % primary air of total air (Kippo,
of operation. A lot of the droplets land on char
1979). In modern boilers with high dry solids the
bed without significant carbon release during
primary air is usually well below 30 % of all air.
flight in the furnace. In practice, low char bed

Figure 9-21, Effect on secondary air model on furnace temperature.
Table 9-7, Effect of air model on deposition rate at nose level

Air model Softwood, Hardwood,
mm/h mm/h
Straight 7-9 5–7
Swirl 3-5 4-5
Interlaced 3-5 3-4
Figure 9-22, Effect on secondary air model on carryover.

thus increasing corrosion.
Depending on boiler operation both increasing
and decreasing primary air temperatures have Interlaced models seem to give much of the swirl
been reported in the literature. Harrison and benefits without its adverse effects. Left and right
Ariessohn (1985) report that char bed surface has superheater section temperatures can be kept
a maximum temperature corresponding a primary fairly even. The carryover measurements show
air value. Both higher and lower primary air that high velocity core exists with conventional
flows result in lower char bed temperatures. This firing, Figure 9-22. This increases carryover and
phenomenon has been attributed by (Adams and decreases furnace effective cross section. Swirl
Frederick, 1988) to the limited reaction rate. If air firing seems to decrease carryover because of
more than what is needed to produce only CO2 is disappearance of the high velocity core. According
provided, then the excess will cool the bed. It has to the measurements, Table 9-7, interlaced pattern
been observed that high air flow will blow off and can minimize carryover.
prevent light char particles from landing.
Tertiary air
Secondary air
The role of the tertiary air is to provide mixing
The main tasks of secondary air are to keep the of unburned flue gas with air to complete the
size of the char bed desired and provide air for combustion and even temperature distribution at
volatiles burning and for the combustion reactions the bullnose level. Tertiary air has little effect on
on the top section of the char bed. Secondary air is char bed operation.
the main air responsible for mixing in the furnace
(Adams, 1994a). Secondary air is the main control Optimum operating temperature
air level.
Many of the lower furnace operating parameters
Furnace behavior can be affected by changing depend on the lower furnace temperature. Low
secondary air model (Metiäinen, 1991). The furnace temperatures cause high TRS emissions.
main effects are seen at the furnace temperatures Increasing lower furnace temperature decreases
and the amount of carryover. The swirl seems to the sulfur dioxide and TRS emissions. But when
increase lower furnace temperatures significantly, the sulfurous emissions are lowered the nitrogen
Figure 9-21 (Vakkilainen, 1996). Even modest oxides increase and the hydrochloric acid emis-
adjustments in air port damper settings change sions decrease.
the temperature profile. It is possible to decrease
the flue gas temperature at nose level by some 20 High sulfur emissions cause increased lower
- 30 oC. In the beginning of 1990 some boilers furnace corrosion. High furnace temperatures
used very much swirl (Bergman and Hjalmarson, increase heat fluxes and corrosion caused by high
1992). Increasing swirl too much results in very metal temperatures. High heat fluxes decrease or
high temperature imbalance between left and even remove the protective frozen char bed and
right superheater sections. High swirl deposits smelt layers covering the tubes. Composite tube
unburned material on air and liquor openings floor cracking can be partially caused by low char
Figure 9-23, Optimum temperature window.

Figure 9-24, Reduction efficiency after new boiler startup (Vakkilainen and Holm, 2000).
bed height. 9.5 NOX CONTROL IN

High sulfur dioxide emissions mean sticky, low
RECOVERY BOILERS
pH, sodiumbisulfite containing dust that is foul- Wood contains some 0.05 … 0.15 w-% of nitrogen
ing. Increased lower furnace temperature and (Verveka et al., 1990). During kraft cooking most
sodium release binds the SO2 and creates excess of it is released and transferred to black liquor.
sodiumcarbonate. Resulting dust is non sticky and It seems that dissolution of nitrogen containing
easy to remove. A further increase in the lower compounds occurs easily during the first stages of
furnace temperature increases chloride enrich- cooking (Niemelä and Ulmgren, 2002). Very little
ment in dust causing boiler bank pluggage and nitrogen exits with pulp (Kymäläinen et al., 1999).
superheater corrosion. Most of the ammonia and other volatile conden-
sates in black liquor is released during evaporation
and ends up in methanol and NCG. Heat treat-
9.4 REDUCTION CONTROL ment can increase this release of nitrogen (Aho et
Good reduction efficiency requires high tempera- al., 1994).
ture, lack of oxygen and presence of carbon. For
good reduction proper char bed that covers smelt Some of the main forms of nitrogen in black
is essential. After new boiler startup the reduction liquor are pyrrole, pyridine and amino acids
rate will initially vary a lot, Figure 9-24. Modern (Niemelä, 2001), Figure 9-25. Different wood
boilers can achieve reduction efficiencies of 95 … species have nitrogen in different forms. In pine
97 mol-% measured in smelt. (pinus sylvestris) about fifteen different amino
acids can be found.
There are several reasons for poor reduction.
Reduced smelt can reoxidize. If char bed does Nitrogen release from black
not cover the smelt spout openings, air will react liquor during combustion
with molten smelt. Low char bed temperature will
result in low reduction and high sulfur emissions. About two thirds of nitrogen in black liquor is
Suspension firing with too small droplets will released as ammonia during volatiles release (Aho
affect reduction. Droplets will burn fast and fur- et al., 1993). The rest remains in char and exits the
nace can be hot, but lack of carbon will slow down recovery furnace with smelt probably as sodium
reduction reactions. cyanate NaOCN (Kymäläinen et al., 2002). In
recovery boiler furnace nitrogen in black liquor
Lower load will reduce furnace temperatures and has three pathways. It ends up as NO, elemental
decrease reduction (Verloop and Jansen, 1995). nitrogen or nitrogen in smelt, Figure 9-26.
Often low loads are run with high excess air levels.
This further decreases reduction. The formation of NOx in coal, gas and oil flames
has been extensively studied. The main chemical
mechanisms that can be used to control formation

Figure 9-25, Some of the nitrogen compounds in black liquor.
and reduce formed NO have been established for

some time now. These same principles apply to Flue gas volume
kraft recovery boilers. The effect of the presence
of volatile alkali metals such as sodium and potas- Vfg = 663.5*22.4m3/kmol/27.4kg/kmol = 542 m3
sium seem to have minor effect. The main sources
of NOx in boilers are fuel and thermal NOx. The Nitrogen oxides
fuel NOx is formed during pyrolysis and volatiles
evolution. The source is the nitrogen in the fuel. VNOx/Vfg = 0.16 m3 / 542 m3 = 300 ppm
Example The typical NOx emissions from a recovery boiler

are about 100 ppm. We can conclude that there is
Calculate NOx from nitrogen in black liquor us- enough fuel nitrogen in the black liquor to pro-
ing typical values for recovery boiler material and duce all of the NOx emissions if the conversion is
energy balance and a typical nitrogen content of 25 ... 30 %. Similar conversion fractions have been
0.1 w-% in the dry solids. We can get for possible measured in laboratory conditions. This fraction
NO generation per 100 kg of dry solids. seems to be about constant for black liquor of
various wood species, type of cook and dry solids
Nitrogen oxides moles concentration (Forssén et al., 1997).
MNOx = 0.001kg/kgds*100kgds/14kg/kmol = 7.1mol Thermal NOx is formed during combustion.

The source is the elemental nitrogen coming in
Nitrogen oxides volume with the air. Typically about up to 60 % of NOx
emissions in PFC boilers are from thermal NOx.
VNOx = 7.1mol*22.4m3/kmol = 0.16 m3 The temperatures in recovery boiler furnace are
Figure 9-26, Nitrogen reaction paths from black liquor.

Figure 9-27, Formation of thermal NO in recovery boilers (after Nichols et al., 1993).
too low for significant thermal NOx production Ernola et al. (1989) have presented an extensive
(Nichols et al., 1993). Formation of thermal NOx study on NOx reduction using CHEMKIN com-
is very low until the furnace temperatures reach puter code in coal combustion. They found H2 and
1400 oC, Figure 9-27. Even increase of black liquor CO to be effective reducing agents. They conclud-
dry solids from 67 to 80 % does not in typical ed that the main destructive reactions were
boiler produce significant amounts of thermal
NOx (Adams et al., 1993). NO + CO → N + CO2 11-3
Other smaller sources of NOx, such as prompt NO + H2 → N + H2O 11-4

or Fennimore NOx can be ignored for recovery
boilers. Lyon et al. (1989) have studied the effect of H2
as NOx destruction agent in a pilot scale tower
NOx destruction in recovery furnace. They found hydrogen to be less effective
than propane. In the presence of CO and when
boilers temperature is over 1400 oC, the NO formed will
The main effort on NOx control has concentrated react to N fairly rapidly (less than 100 ms) (Aho et
on firing and air system modifications as these al., 1993).
offer almost the only viable option (Anderson
and Jackson, 1991). NOx formation is affected by Thompson and Empie (1993) have proposed that
the amount of air and combustible products. The some reaction of NO by sodium species to NaNO3
lower portion of recovery boiler furnace is at or could take place. Analyses of mill ESP samples
little below stoichiometric level. The fluctuating have shown this to be of minor importance.
concentrations, temperatures and velocities affect
strongly to the formation of NOx. Effect of fuel nitrogen content to
NOx emissions
The recovery boiler differs from typical coal and
oil fired boilers in the respect that the inert frac- When boilers are run with liquors from pulping of
tion in the fuel is very large. In addition the fuel different wood species, high nitrogen containing
heating value is low. So the temperatures encoun- woods (hardwood) produce more NOx than soft-
tered in recovery boilers are low (1000 - 1400 oC ). wood liquors. Field tests support the conclusion
that recovery boiler NOx emissions correlate with
Some of the possible main pathways for NOx nitrogen in the black liquor (Clement and Barna,
reduction are 1993). The highest reported nitrogen content has
- reduction by hydrocarbons been 0.5 w-% of bagasse black liquor dry solids
- reduction by CO (Aho, 1994). The detection accuracy of nitrogen
- reduction by H2 in the fuel from commercial laboratories is very
- reduction by H2S poor so direct correlations are often misleading.
- reduction by residual carbon in char For example two black liquors of nitrogen content
- reverse Zeldovich mechanism 0.072 and 0.114 mass % get both typically a value
- reduction by sodium species of 0.1 mass-%.

Figure 9-28, NO during trials at a North American mill.
Aho et al., (1994) studied the reduction of General observations of mill trials seem to indi-
nitrogen in the black liquor. One can somewhat cate that
change the nitrogen content by treating chips, - flue gas dust content and NOx emissions
making cooking changes and treating black liquor. have no significant correlation
The results however have so far showed this to be - black liquor HHV and NOx emissions have
uneconomical. no significant correlation
- air split between primary and secondary
Part of the nitrogen in black liquor is in the form level has no significant effect to NOx
of NH3 and so easily vaporized. Parts of nitrogen - air pressure has no significant correlation
can then be extracted from black liquor in the with NOx
stripper area. This product may enrich to some - CO emissions affect NOx
parts of the recovery cycle so great care should be - staging air higher (to tertiary) can reduce
taken when this waste stream is destroyed. NOx a little
- adding a upper tertiary or quaternary air
Effect of operating parameter level can reduce NOx
changes to NOx emissions Decreasing furnace temperature seems to decrease
Changing air and liquor distribution affects NOx NO. This change can be attributed to higher SO2
levels. In trials done in a North American boiler formation. In trials with actual operating recovery
the NOx level could be varied from 75 to 55 ppm. boilers typically after SO2 emissions go to zero an
The nitrogen content in the black liquor was low. increase in NOx emissions is seen, Figure 9-29.
20 to 30 % reduction from highest values can be
done by proper air staging and firing. In the NOx tests performed by manufacturers the
Figure 9-29, Effect of furnace temperature to NO during trials at a North American mill.

Figure 9-30, Effect of quaternary air flow to NO during trials at a Scandinavian mill (Wallén et al., 2002,).
number of liquor guns hasn’t affected the NOx

emission level. This means that for typical kraft re-
covery boilers the uneven fuel and oxidant mixing
isn’t a significant source of NOx emissions. How-
ever in some tests it has been seen that uneven fir-
ing can cause increase in other nitrogen emissions
than NO. High gun pressure and resulting small
black liquor droplet size has also been suggested
as source of some reports of high NO excursions
(Björklund and Warnqvist, 1989). Recovery boiler
load increase has been reported to increase NOx
emission (Salmenoja 1998).
Trials in Sweden with oxygen enriched air to Figure 9-31, Effect of residence time to quaternary
recovery boiler have resulted in unchanged con- air to NO reduction during trials at a Japanese mill
centration of NOx in flue gas. As flue gas flow is (Wallén et al., 2002).
reduced, the total emission of NOx per ton of pulp
did reduce (Verloop et al., 2001).
level and the first tertiary air level improve NOx
removal.
Reduction of NOx in the upper
furnace NOx behavior in a large Scandinavian boiler
has been studied by Vakkilainen et al. (1998a).
Recovery boiler NOx emission can be decreased The recovery boiler employs partially interlaced
by applying air in stages (Forssen et al., 2000a). secondary air injection model. This type of air
Typically this has been done by adding fourth model is capable of achieving high mixing rates
level of air. The effect of amount of quaternary in the lower furnace resulting at a fairly even
air can be seen in Figure 9-30. More than 10 % of temperature profile. The average temperature at
total air is needed to achieve significant reduction. the furnace exit is about 950 oC.
Also residence time in furnace before the final air The modelling results showed fairly typical flow,
injection seems to play a role, Figure 9-31. More temperature and gaseous species concentration
than 10 s residence time was needed for NOx re- patterns. Nitrogen is released primarily during the
moval. This seems to indicate that sufficiently low pyrolysis phase. Half of the fuel nitrogen was as-
temperature is reached before the final combus- sumed to be released during pyrolysis as ammonia
tion takes place. Brink et al. (2004) predict that and half as molecular nitrogen, Figure 9-32. The
long residence time between the black liquor gun nitrogen release occurs between the secondary
Figure 9-32, Furnace vertical NH3 concentration Figure 9-34, Furnace NO- measurement profile
profile, dry gases. converted to dry gases and 3 % O2.
This boiler was fitted with upper tertiary air to

reduce NOx emissions. The retrofit was done in a
pre planned three day shut down. Results from the
calculations were compared to measurements per-
formed on two days A and B. For practical reasons
the experimental cases differ to some extent from
simulated case. The furnace measurements were
conducted by inserting an air cooled sampling
tube through air ports, 2,5 m into the furnace
at the primary, secondary and tertiary air levels.
The temperatures were measured by a shielded
thermocouple and the NO profile by an on-line
analyzer. The measured temperatures for two days
A and B are shown as Figure 9-33 and the furnace
NO profile for the same cases as Figure 9-34.
The experimental temperatures show a fairly con-

stant lower furnace temperature. The computed
temperature field is consistent with the experi-
Figure 9-33, Furnace temperature measurements. mental data. The nitrogen species measurements
in the furnace reached similarly only the furnace
perimeter. Main nitrogen species measured by
and tertiary air levels at a rather large furnace the on-line analyzer was NO. Significant NO2
cross section. This would indicate that lower fur- concentrations could not be found. It was not
nace conditions are the most relevant to nitrogen possible to measure NH3 nor HCN concentrations
oxides formation and possible reduction reactions. at the same time so the values reported at Figure
There is a small peak in the middle of the char 9-34 show only NO. The NO concentration is very
bed surface showing a local release of nitrogen small at the primary air level. This indicates that
containing species. This is not indicative of char the portion of NO released from the char bed is
nitrogen, which was not modelled, but of liquor fairly low. The overall exit NO concentration (92
droplets landing onto the char bed at this site.
- 112 ppm) represents typical values when operat- Table 9-8, Effect of solids increase to recovery
ing with mixed wood liquors. boiler operation
Dry solids % 65 77
The measurements showed that NOx reduction
Capacity tds/d 2130 2130
is practical with splitting the tertiary air into two
or more separate levels (in vertical direction). In Steam temp. o
C 480 480
practice the NOx can be lowered by 10 … 15 % Steam pressure bar 86 86
compared to traditional system.
Economizer exit o
C 168 157
The study shows that this kind of modeling can Reduction eff. % 93–96 96-97
be applied to a 3150 tds/d recovery boiler. The Sootblowing % 2.5 – 3 <2
modeling results increased the understanding of steam
the in-furnace processes and helped in interpreta- SO2 emission ppm 200 – 300 0
tion of the measurements and of the observations
CO emission ppm ~100 30 - 60
in the operation of the boiler. The results also
support the role of fuel nitrogen as a major source NOx emission ppm 60 – 80 ~70
of the recovery boiler NO emissions. HHRR MW/m2 3.1 3.1
Gross steam % 100 105.5
Other means of controlling NOx
emissions
droplet size might have a beneficial effect to NO
We can control the formation of NOx by. emissions. In modelling studies fine black liquor
- Reducing the free oxygen concentration droplets produces more NOx than coarse black
(Boström, 1990). This decreases both liquor droplets (Forssén et al., 2000b).
the thermal and fuel NOx generation. All
new improved air systems use this as part
of their NOx reduction strategy. Reduction 9.6 SOX AND TRS CONTROL
of oxygen level should be continued until
unacceptable CO emissions result. In the early 1980’s it was discovered that increas-
- Reducing the mean and the peak ing black liquor dry solids decreased the sulfur
temperature levels in the boiler. This emissions (Maso, 1988). It was quickly understood
reduces the thermal NOx. that this happened because increased dry solids
- Improving mixing, which affects increased furnace temperatures. This increased
temperature and concentration profiles sodium release, which then could bind sulfur di-
decreases usually NOx emissions. oxide to sodium sulfate (Perjyd and Hupa, 1984).
- Staging the air to more levels than
conventionally. Increased furnace temperatures also increase reac-
tion rates decreasing the reduced sulfur emissions.
Some studies indicate that increasing black liquor Lankinen et al. (1991) have outlined how the third
Figure 9-35, Effect of black liquor dry solids content to sulfur emissions (A. Ahlstrom Oy).

Figure 9-36, Effect of black liquor dry solids on ash behavior and emissions of SO2 and HCl (Vakkilainen,
2000).
generation recovery island will affect the boiler dried liquor droplets will have only a marginal
operation, Table 9-8. How the char bed processes influence on reaction and temperature conditions.
will be affected is not as obvious. They based this assumption on the large mass and
high temperature of the char bed. The modern use
Earlier it was common practice to oxidize black of char bed imagining cameras has concluded that
liquor. Oxidation decreased TRS emissions from even small low temperature areas will significantly
black liquor. With modern air systems and higher increase sulfur emissions. Practical experience
dry solids, black liquor oxidation is becoming with char bed imagining cameras indicates that
obsolete. creation of ‘cool spots’ in the bed correlates with
increased SO2 emissions. If poor mixing or low
In recent years the dry solids concentration from temperatures prevail in furnace also higher TRS
evaporators has increased from 65 to about 80 % emissions result.
dry solids. Increasing the black liquor dry solids
concentration does not change the combustion Firing high dry solids black liquor has decreased
air flow or the black liquor dry solids flow (Vak- recovery boiler fouling (Vakkilainen, 2000, Vak-
kilainen, 2000). The flue gas flow decreases as less kilainen et al., 1995). This effect can be attributed
water enters the furnace. The same amount of heat to decreased sulfur emissions and the diluting
from burning organics can increase the flue gas effect of sodium carbonate, Figure 9-36. It should
temperature. Even a moderate increase in black be noted that there are two regimes of fouling; low
liquor dry solids increases the lower furnace tem- solids fouling is controlled by SO2, and high solids
perature and decreases the SO2 emission, Figure fouling is controlled by the ash sticky temperature.
9-35 (Ryham and Nikkanen, 1992, Vakkilainen et
al., 1998).
9.7 FIRING BLACK LIQUOR
It is also obvious from Figure 9-35 that increasing – MILL EXPERIENCE
sulfidity increases sulfordioxide emissions. Even
boilers that fire 80 % dry solids will have sulfur Actual black liquor firing practices differ between
emissions if sulfidity is high enough (Ibach, 1995). different boilers. This is illustrated by studying
firing practices in five modern boilers (Järvinen et
Sulfur dioxide and NOx have cross correlation. al., 1997). These boilers are large and from both
This cross correlation exists also with high dry main Scandinavian vendors. The furnace dimen-
solids boilers (Jones and Stewart, 1993). Increas- sions and liquor gun locations for each boiler are
ing furnace temperatures decreases sulfur dioxide presented in Table 9-9. Dimensionless height is
emission. Increase of NOx emissions is not be- the actual height divided by the square root of
cause of thermal NO formation. Recovery boiler bottom area.
furnace temperatures hardly reach required 1400
o
C (Nichols et al., 1993). NOx emissions increase
because H2S is much better scavenger of oxygen.
Firing parameters
To record operating practices inquiries were sent
Björkman and Warnquist (1985) claimed that non to operators and then the answers were reviewed

Figure 9-37, Air distribution models of the boilers (Järvinen et al., 1997).
Table 9-9, Boiler geometries, gun locations and

loads. Black liquor dry solids in use at these mills reflect
Boiler Furnace Furnace Furnace Liquor Load, the Scandinavian practice. Older mills have dry
width depth height gun tds/d solids close to 70 %. The newest mills have dry
height
solids in excess of 80 %. It is also typical that
A 10.1 10.4 27.8 6.0 2550 the target black liquor dry solids is the same for
B 12.6 11.8 34.5 9.0 3300 all wood species. In addition the difference of
C 9.9 9.9 23.0 4.8 2200 firing temperature to boiling point temperature
at atmospheric pressure is shown. Boiling point
D 12.6 11.8 34.5 8.6 3500
rises were based on actual liquor analysis at mill
E 10.9 10.8 31.6 6.2 2750 and represent averages of analysis for the mill in
question.
Table 9-10, Liquor parameters.
There are four different types of liquor guns in use
Boiler Firing Firing Liquor Dry Diff. to at these five mills. It is usual that liquor guns differ
press. temp. - solids BP from one mill to another. Liquor gun selection at a
kPa o
C % o
C
mill is typically based on long time learning curve.
A1 150 124 softw. 68.6 +10.7 Mills A and E use splashplate with a narrow liquor
A2 150 124 hardw. 68.6 +10.7 firing angle. Mills B and D use splashplate with a
A3 150 124 mixed 68.6 +10.7 large diameter and wide liquor firing angle. Mill C
uses small size liquor gun
B1 95 118 mixed 74.4 +2.0
B2 125 125 mixed 74.4 +9.0 Difference between liquor firing temperature
C1 220 133 softw. 75.0 +16.7 and the atmospheric boiling point temperature
C2 220 133 hardw. 75.0 +16.7
describes the flashing tendency. As can be seen
in Figure 9-3, the black liquor partly forms steam
C3 202 133 mixed 75.0 +16.7 inside the liquor gun if the operating temperature
D1 170 136 softw. 83.5 +15.2 is above the atmospheric boiling point. The larger
D2 100 134 mixed 83.5 +13.2 the value of temperature difference, the larger the
driving force of evaporation. If firing temperature
E1 140 125 softw. 72.6 +9.8
exceeds the liquor atmospheric boiling tempera-
E2 120 115 hardw. 72.6 -0.2 ture, flashing occurs. Practically all boilers operate
E3 130 120 mixed 72.6 +4.8 above the atmospheric boiling temperature.
and discussed with copies of parameters from In order to compare the liquor droplet combus-
automation systems. Typical liquor firing param- tion and flight path between different boilers, an
eters at full load are shown in Table 9-10. All of average flue gas velocity was calculated. Flue gas
the mills pulp several wood species and operators density was scaled with respect to the adiabatic
use different parameters to fire black liquor from flame temperature. The furnace temperature when
pulping of different wood species. These values firing high solids black liquor is higher due to a
reflect the actual average values in use at the mill, smaller flue gas amount. Furnace temperature has
and are not necessarily manufacturer’s recom- a great effect on char bed processes and on droplet
mendations. combustion. The liquor mass flow rates and flue

gas flow parameters are presented in the Table observed if we use CFD calculations to look at
9-11. where black liquor burns. For the presented five
boilers this was done (Järvinen et al., 1997). Black
It can be seen that there are no significant differ- liquor combustion was calculated for one nozzle,
ences between the mean flue gas velocities. The situated at the center of the front wall, for every
calculated adiabatic flame temperatures differ boiler and each wood species. Actual liquor gun
mainly because of different dry solids. This means operating height position and vertical spraying
that some furnaces will naturally operate some- angle were used. The sheet velocity was assumed
what hotter that the others. to be constant as a function of sheet angle. The
mass flows varied as a function of sheet angle.
The air distribution models differ from boiler
to boiler. The model is typically changed when Carbon combustion is important for a proper
firing different types of liquors. The air distribu- furnace operation. For example, the reduction
tion models for the boilers are presented in Figure requires carbon to proceed. If carbon combustion
9-37. As can be seen there is a marked difference is completed before the droplets reach the char
between the different primary air flows and terti- bed, there is no reducing carbon for the reactions
ary air flows. in the char bed.
Liquor spray properties The black liquor combustion level was calculated
on the basis of the vertical carbon release distribu-
As it was previously observed, the boilers differ tion in the furnace. The results are presented as
from each other as far as liquor firing is con- a function of the liquor droplet size, Figure 9-38.
cerned. Various types of nozzles are used. There A clear correlation between the combustion level
are differences in the boiler geometry and the and the droplet size can be seen. The larger the
liquor dry solids content. The liquor spray proper- droplet size, the lower the combustion level.
ties; nozzle outlet velocity and mean droplet size
were calculated with respect to the real operating If combustion of different wood species is com-
parameters. The results are presented in Table pared, the effect of the liquor type is obvious.
9-12. Softwood liquors burn at a higher level compared
to hardwood liquors. This is because of softwood
liquor swells more.
Black liquor combustion height
Differences in black liquor firing practices can be The horizontal carbon release level is presented
Table 9-11, Black liquor and flue gas flows Table 9-12, Liquor spray properties (Järvinen et
(Järvinen et al., 1997). al., 1997).
Boiler Gun / do / ho mass No of Flue gas Flue gas Firing Firing Nozzle Droplet Furnace
flow guns flow velocity temp. press. velocity size temp.
- /mm/mm kg/s - m3n/s m/s o
C kPa m/s mm K
A1 sA/27 /22 3.98 8 101.0 4.5 A1 124 150 11 2.05 1400

A2 sA/27 /22 3.98 8 101.0 4.5 A2 124 150 11 2.09 1400
A3 sA/27/22 3.98 8 101.0 4.5 A3 124 150 11 2.08 1400
B1 sB/30/30 5.25 8 132.0 4.3 B1 118 95 6 3.91 1455
B2 sB/30/30 5.25 8 132.0 4.3 B2 125 125 7 3.32 1455
C1 pp/22.5/19 3.96 7 91.0 4.6 C1 133 220 17 1.39 1468
C2 pp/22.5/19 3.96 7 91.0 4.6 C2 133 220 17 1.49 1468
C3 pp/22.5/19 3.96 7 91.0 4.6 C3 133 220 17 1.44 1468
D1 sB/38/38 5.72 8 139.0 4.9 D1 136 150 10 3.34 1545
D2 sB/38/38 5.72 8 139.0 4.9 D2 134 100 8 3.96 1545
E1 sA/27/22 4.33 8 114.0 4.7 E1 125 140 10 2.20 1441
E2 sA/27/22 4.33 8 114.0 4.7 E2 115 120 6 3.31 1441
E3 sA/27/22 4.33 8 114.0 4.7 E3 120 130 8 2.72 1441
sA = splashplate type A, sB = splashplate type B, pp
= ”willow flute” type, do = nozzle tube diameter, ho =
height of the nozzle opening

Figure 9-39, Horizontal distance of carbon release from the level of the liquor gun as a
function of droplet size for the different boilers.
Figure 9-38, Vertical carbon release level as a function of droplet size.
Figure 9-40, Fraction of the liquor delivered to the char bed and the fraction of the
liquor burned in suspension.

Figure 9-41, Total amount of sulfur released during devolatilization as a function of droplet size.
in Figure 9-39. The general trend in the horizon- Sulfur release

tal carbon release location seems to be that the
smaller the droplet size, the closer the location of The release of sulfur in different boilers was
the carbon release from the liquor gun. In boilers studied. The total amount the sulfur released was
B and D, with wide angle guns and large droplet calculated. The result is presented in Figure 9-41.
size, the liquor is burned closer to the nozzles. Generally, the amount of sulfur released seems
This is due to the larger opening angle of the noz- to increase with increasing droplet size. Boilers B
zles used in these boilers. In them a greater liquor and D show an opposite trend, as before, during
amount is sprayed near the adjacent furnace walls char combustion. The reason for a smaller amount
of the liquor gun, which reduces the penetration of sulfur released is wall spraying.
of the spray into the furnace. When liquor types
are compared, the softwood liquor, that swells In these calculations sulfur release is cut off when
more, seems to burn closer the liquor gun wall the liquor hits the wall. In reality, the release
due to larger drag forces and the smaller droplet continues on the furnace wall and as the liquor
size. droplet falls from the furnace wall to the char bed.
This means a large part of the sulfur release is
Combustion in suspension, in probably unaccounted for.
char bed and on furnace walls
Firing 90 % liquor in a recovery
The amount of liquor reaching the char bed, the boiler
furnace walls and the fraction of liquor form-
ing carryover were calculated on the basis of the Tests at the mill level to spray very high dry solids
droplet flight path data. The results are presented black liquor were performed in the Stora-Enso
in Figure 9-40. mill at Varkaus (Vakkilainen and Holm, 2000). In
these tests different dry solids black liquors were
There are major differences between the boilers. sprayed in to the furnace. Mass flow and liquor
The boilers B and D seem to spray liquor to the gun types were varied. Typical commercial nozzles
furnace walls to a great extent. The boilers A and were used. The evaporative potential of the black
E show the opposite trend. liquor at the very high dry solids is small. There-
fore it was noticed, perhaps unexpectedly, that the
There is a variation between suspension burning liquor sheet formed from firing 90% black liquor
and bed burning. In the boilers A, C and E ap- looked quite similar to the corresponding sheet at
proximately 60 % of the liquor combustion takes firing 75% black liquor.
place during flying and the rest of the combustion
occurs in the char bed. With the boilers B and D Burning of 80 to 85% dry solids black liquor is in
the fraction of the inflight combustion is approxi- commercial operation in several kraft recovery
mately 40 % of the total liquor and the bed and boilers. Spraying of the very high dry solids black
wall burning consumes nearly 60 % of the liquor liquor has not led to any unsolvable problems.
sprayed.

Figure 9-42, Picture of 90 % liquor sprayed at recovery boiler furnace with liquor gun type B22 (Vakkilainen
and Holm, 2000).
The main advantage of the very high dry solids is pentine is more typically fired in the lime kiln.
that black liquor can be fired at substantially lower Main use of methanol has lately been as a support
partial loads (Mäntyniemi and Hartley, 1999). The fuel for CNCG burner. There are mainly positive
char bed is more stable and easily controlled and experiences from this type of operation. However
the boiler is able to handle process disturbances from the mill energy balance point of view also
more smoothly (Vakkilainen and Holm, 2001). In methanol could be better used at the lime kiln to
addition emissions can be controlled also when reduce fossil fuel consumption.
burning odorous gases (Mäntyniemi and Haaga,
2001). Removal of turpentine from NCG is important
because turpentine is very explosive. Collec-
Some problems in burning of very high dry solids tion lines that contain turpentine are prone to
black liquor have been that the membrane mate- explosion hazard. Several incidents with NCG in
rial of manometers does not last long, valves tend Finland have probably been caused by high tur-
to stick and they have other mechanical problems pentine concentrations. If NCG contains gaseous
and there are difficulties to get big enough droplet turpentine, the line should have droplet separation
sizes. and flame arrester for safety reasons.
9.8 BURNING WASTE STREAMS Firing soap

IN A RECOVERY BOILER Soap has been separated from black liquor as it
naturally ascends to the surface of a black liquor
Modern boilers that burn waste streams have tank since 1910’s (Niemelä, 2004). Modern mills
typically liquor dry solids at least in the 70’s and a have often so low sulfur loss that it is uneconomi-
modern or modernized air system that comprises cal to manufacture tall oil from soap by acidifica-
of primary, secondary and often two-level tertiary tion with sulfuric acid. In several mills firing soap
air. One noticeable feature in these air systems mixed with black liquor has been practiced with
is that they have been improved by using larger good success.
and wider spaced airports. Modern air system is
a necessity. Excess streams require high furnace Black liquor that contains soap burns slower.
temperature and rather stable combustion. Soapy liquor also swells less. Soapy liquor needs
to be fired with smaller droplet. It is harder to
Firing methanol and turpentine achieve zero SO2.
Turpentine and methanol can be fired in recov- It is important to note that soap affects the refrac-
ery boiler furnace. Most often turpentine and tometer reading, Figure 9-43.
methanol are fired at liquid form using separate
lances using the CNCG burner opening. Tur-
Figure 9-43, Effect of addition of soap to black liquor dry solids reading by refractometers.
Burning biosludge Effect of waste stream burning to

Biosludge has been burned successfully in several sulfur emissions
recovery boilers by mixing it with black liquor Whether dry solids is enough is easily checked.
(Dahlbom, 2003). Biosludge contains organic dead Assume that sulfur input with NCG forms SO2.
bacteria. This makes water removal very difficult. Further assume that SO2 is reacted to sulfate
Biosludge dry solids is typically only 5-15 g/l from replacing carbonate in ESP ash. Check that there
waste water plant. It needs to be concentrated with is enough carbonate remaining in the ash to keep
e.g. centrifuges to > 10 % dry solids. pH high, Figure 9-44, even after high concentra-
tion odorous gases are being fired.
Biosludge contains inorganic residue from waste
water and chloride from bleaching. It has a nega- Boiler SO2-emission and so carbonate in ESP ash
tive effect to fouling, plugging and corrosion. can be additionally controlled by air model, liquor
gun type and air ratios. Mill sulfur balance will
Combustion of bleach plant affect the sulfur release and the recovery boiler
effluent operation. High sulfidity means LVHC NCG can
not be fired. It has been found that even with +80
There have been mill trials of combustion of % dry solids in as fired black liquor, LVHC NCG
bleach plant effluents in US and Sweden. The aim burning was impossible when sulfidity was over
is to decrease bleach plant effluent loading by 45 %.
recycling part of black plant effluent to recovery.
Mill water usage can decrease significantly. Burning HVLC NCG or DNCG in a recovery
boiler has only a minor but negative effect on
The main concern has been the increased NPE recovery boiler sulfurous gases emissions, if
load to recovery cycle. Especially the high chloride black liquor dry solids is over 68 %. It has been
content of bleach plant effluents has been men- found that in most cases when sulfidity is below
tioned. 40 % and the black liquor dry solids is over 72 %,
DNCG can be burned with minimum sulfur emis-
Adding bleach plant effluents to black liquor sions. Finnish recovery boiler committee does not
seems to affect the liquor viscosity and combus- recommend burning NCG in a recovery boiler if
tion properties only slightly (Ledung and Ulm- sulfur emissions increase significantly.
gren, 1998, Ledung and Ulmgren, 1997, Nichols,
1992,). The results were similar to both acidic and Effect of waste stream burning
alkali stage additions. This is because the com-
position of bleach plant filtrate is fundamentally to NOx
similar to black liquor. CNCG, biosludge and methanol can contain sig-

Figure 9-44, Effect of carbonate on ESP ash pH.
nificant amounts of nitrogen compounds.
LCHV NCG has not been reported to affect

the NOx levels in recovery boilers. HCLV
NCG contains sometimes up to 2000 ppm of
nitrogen compounds, mostly ammonia. The
flow if high concentration odorous gases is
between 0,2 and 3 % of total flue gas flow. If
one assumes all nitrogen compounds to react
to NOx then the NOx level would increase less
than 5 ppm.In practice the measured increase
has been 2 ppm or less.
Crawford and Jain (2002) found that in ther-

mal oxidizers at temperatures above 850 oC
about 25 to 30 % of ammonia in CNCG and
stripper off gas was converted to NOx.
Recommendations for
combustion of waste streams
Burning waste streams in a recovery boiler
causes risks. Analysis of accidents with NCG
burning helped to quantify these risks. The
following risks were identified; leaking of toxic
streams to working environment, disruptions
in condensate removal, explosion in line or
duct containing gases, gaseous explosion in
recovery boiler furnace, smelt-water explosion
in recovery boiler furnace and corrosion in Figure 9-45, Example of how recovery boiler looks
ducts, lines or pressure parts. after NCG handling accident.
Several recovery boiler users associations

including BLRBAC, Swedish and Finnish
associations have approved recommendations
concerning combustion of waste streams in
recovery boilers. The target has been to avoid
accidents that can happen when NCG are
burned in a recovery boiler, Figure 9-45. Startup burners are located near furnace floor.
They help to heat recovery boiler char bed during
The recommendations deal exclusively with waste startup. Autonomous black liquor firing needs
stream combustion in a recovery boiler. Collec- lower furnace temperatures in excess of 800 oC.
tion of waste streams and handling them in other
systems (e.g. lime kiln) are not covered by the Startup burners need to be used when char bed
recommendations. is out of control or blackout is eminent. Char bed
can if the black liquor droplet size landing on the
bed is too large to grow uncontrollably. The rem-
9.9 AUXILIARY FUEL FIRING edy is to decrease black liquor firing rate and start
using startup burners until the situation is under
Auxiliary fuel can be fired in a recovery boiler control again. If the bed growth proceeds too long,
from startup burners and load burners. Typical a blackout result. This means liquor firing must be
auxiliary fuels are natural gas, light and heavy fuel stopped and bed burned away.
oil.

10 Material selection and corrosion
Figure 10-1, Corrosion rate of a row of superheater lower bend tubes.
When recovery boilers are designed one of the If corrosive environment is known proper mate-
most difficult questions that arise is; what kind rial can be selected. Often only experience can
of materials should one use for different parts give the answer to this. The recovery boiler is
of the boiler. Corrosion is typically divided into physically large. Corrosion rate in one side of the
areas based on location of corrosion; water side boiler can significantly differ from corrosion on
corrosion, high temperature corrosion and low another side, Figure 10-1.
temperature corrosion
Water side corrosion occurs in the steam/water

Normal tube environment
side of the boiler tubes. Most often the cause is Before we start the discussion about corrosion,
impurities in the feedwater. High temperature cor- we should look at how a normal tube surfaces in a
rosion occurs typically in the superheaters. Low
temperature corrosion occurs in the economizers
and air heaters. Low temperature corrosion is
often associated with formation of acidic deposits.
10.1 GAS SIDE CORROSION OF

HEAT TRANSFER SURFACES
Gas side corrosion in recovery boilers is caused
by formation of deposits, which have corrosive
properties. A typical cause is formation of molten
alkali phase. Gas side corrosion can also be caused
by gaseous components and is then usually associ- Figure 10-2, Typical tube surface towards furnace
ated with reducing atmosphere. (MPSP, 1991).
Material selection and corrosion 10-1

known corrosive environment.
Often the outermost tube is corroded iron oxide

(Fe2O3). This is porous and flaky, so gases can
easily pass through. Corroded iron oxide has
very low strength. It then flakes and breaks away
at fast rate. Often there forms an ash layer on
the tubes, especially superheater tubes. Ash is
formed through direct deposition of ash particles,
condensation of vapours and reaction of gaseous
species with deposited material. Depending on
conditions this layer is thin and reactive or thick
and rather static. For corrosion the layer closest to
the tube is the most important.
Sulfidation
Sulfidation is the most common form of corrosion
in the lower furnace, in openings and in super-
heater tubes. In particular hydrogen sulfide (H2S)
contents close to the wall in the lower furnace
Figure 10-3, Effect of temperature and H2S concen- can be around twenty mole percent (Singh et al.,
tration on corrosion rate of carbon steel and alloy 1999). In modern, high solids boiler the concen-
AISI 304 (Salmenoja and Tuiremo, 2001). tration close to the wall above the tertiary air is
only some tens of ppm (Vakkilainen and Holm,
typical boiler look like. 2001). In sulfidation the iron reacts with sulfur in
the gas phase forming iron sulfide FeS. This means
In Figure 10-2, we see that the boiler tube oxidizes that the protective oxidized surface of metal tube
fast and forms protective iron oxide (Fe3O4) layer. reduces in presence of sulfur containing species.
Protection is typically higher if tube Cr, Ni and
Mn content is increased. Increase in corrosion Corrosion due to gaseous H2S is low up to 250 oC,
prevention is often associated with tighter depos- but at carbon steel it increases rapidly above 310
its, which prevent passage of corrosive agents. o
C, Figure 10-3. Metal temperatures of the furnace
Specifying a proper alloy and thus a proper pro- tubes are typically 15-30 oC hotter than the satura-
tective layer composition helps to protect against tion temperature of the boiler water. Sulfidation
Figure 10-4, Corroded superheater tube (Suik, 2001).

corrosion decreases when chrome containing met-
als are used. Sulfidation corrosion does not occur
at significant rates with stainless steel until ~480
o
C is reached. Stainless steel 304L is the most typi-
cal metal used to control sulfidation corrosion.
Sulfidation attack can be seen in a tube as rough

uneven wear on tube surface, Figure 10-4. Corro-
sion pattern is often uneven, Tube surface is rough
and larger, deeper pockets can be found (Eilersson
et al., 1995). Without adequate protection sulfida-
tion corrosion occurs fast under partially dried
black liquor at furnace walls (Bruno, 2003).
High temperature corrosion

Figure 10-5 shows thinning of superheater tube
side surfaces. Ash is sticking to the windward side
of the tube. Molten layer exists at tube sides.
Fuels contain minor amounts of impurities. These

impurities can cause high temperature corro- Figure 10-5, Superheater corrosion (MPSP, 1991).
sion in the superheaters. Especially vanadium,
sulfur and alkali metals cause high temperature
corrosion. Figure 10-6 shows actual deposits on a
superheater. As can be seen many different types
of particles exist.
One of the most typical means of fighting high

temperature corrosion is to use alloyed materials.
Figures 10-7 and 10-8 show the effect of chro-
mium to a corrosion rate. Other popular metals to
add are nickel and molybdenum. Choice of proper
material is in practice often a matter of trial and
error.
Alkali corrosion
Recovery boiler ash in contact with the tubes is
Figure 10-6, Forest residue superheater ash FB
normally solid. Sodium and potassium may form
(Kurkela et al., 1998).
ash with low melting points 500 … 600 oC. Con-
tact of tube surface with molten alkali salts causes
fast corrosion. This is because chemical reactions
Figure 10-7, Effect of chromium on corrosion. Figure 10-8, Effect of material temperature on cor-
rosion rate, A – D different materials (Salmenoja
and Tuiremo, 2001).

Figure 10-9, First melting temperature of recovery boiler superheater deposit containing some chloride.
occur faster with molten than solid phase, a liquid Chloride corrosion
phase provides an electrolyte for electrochemical
reactions, a liquid phase can readily dissolve the Chloride in fuel forms NaCl, Cl2 and HCl in gase-
corrosion products and liquid phase can fluidize ous from. In active oxidation these gases react
the ash. Carbon steel corrodes very fast in molten with iron to FeCl2 (Grabke et al., 1995). This when
smelt, 1 … 3 mm/h. So maintaining protective, in contact with oxygen, reacts back to Fe2O3 re-
frozen layer is essential for safe operation. leasing chloride in gas form. Reformed iron oxide
layer is porous and does not offer protection from
Typical first melting temperatures of recovery corrosion. Released chloride can react with fresh
boiler deposits are shown in Figure 10-9. With iron oxide and the cycle repeats. Active oxidation
high enough liquid content ash flows down the leads to rapid waste of superheater tubes, Figure
superheater surface. First melting point is a func- 10-10.
tion of boiler operation. Increasing lower furnace
temperature increases fuming and carbonate in Chloride corrosion needs continuous gaseous
deposit (Stead et al., 1995). Alkali corrosion usu- chloride species present to proceed. Salmenoja
ally leaves a smooth, even surface. Uniform thin- et al. (1999) found in laboratory tests that it took
ning can be seen in e.g. superheater tube bends. A some 20 min to start the corrosion and some 15
special case for alkali corrosion prevention is the min to stop when HCl in oxidizing conditions was
smelt spouts. By keeping tube temperature close used.
to 100 oC carbon steel can survive the molten, free
flowing smelt. Chloride corrosion has been studied a lot in con-
nection with waste burning and recovery boilers.
It can be identified from sharp chloride layer
between deposit and the uncorroded tube (Nishio
et al., 2001), Figure 10–11.
Chloride corrosion is affected by gaseous chloride

content near superheaters, Figure 10-12. Even
with higher chloride content at deposits the high
carbonate (i.e. high dry solids) boiler shows lower
chloride corrosion rate (Koivisto, 2000). Laborato-
ry studies show that chloride corrosion rate below
Figure 10-10, Chloride corrosion in superheater
bank (Koivisto, 2000).

Molten hydroxide corrosion
Typically in the furnace conditions sodium
hydroxide vapor reacts with available sulfur and
chloride species. It can also react with carbon
dioxide
2NaOH(l) + CO2(g) → Na2CO3(s) + H2O(g) 10-1
Close to air and liquor ports the carbon dioxide

partial pressure can be low. There it is possible that
sodium hydroxide condenses to the tube surface.
Lately it has been suggested that hot fluxing of
metal into the salt (Holcomb, 2001) can cause this
attack.
Fe2O3 + Na2O → 2NaFeO2 10-2
This kind of corrosion has been found especially

Figure 10-11, Chloride corrosion in superheater behind the furnace tubes close to air ports where
bank (Koivisto, 2000). gases can leak. In addition it has been found at liq-
uor gun openings. Stainless steel 304L has about
10 times the corrosion rate in molten hydroxide
than normal carbon steel. In high pressure boilers
450 oC is very low (Salmenoja, 2000). Salmenoja
the wastage of composite layer can occur in less
et al. (1999) report that chloride corrosion rate is
than two years (McGurn, 1993). Alloys with high
low if metal chrome content is above 12 w-%.
nickel to chrome ratio seem to be able to resist this
corrosion (Paul et al., 1993).
Laboratory studies done by Japanese boiler maker
Mitsubishi show that deposit properties affect
the chloride corrosion rate. Decreasing deposit Acidic sulfate corrosion
melting temperature from 580 to 520 oC by adding
potassium to carbonate containing deposit, with Boilers running at high sulfur dioxide emission
SO2 present in flowing gases doubled corrosion can experience acidic sulfate corrosion. Formation
rate. Increasing chloride content in the deposit of acidic sulfates, sodium bisulfate or pyrosulfates
from 1 to 5 w-%, which decreases FMT by 4 … occurs at flue gas temperatures below 400 … 450
o
6 oC increased corrosion rate only marginally C and causes problems especially in boiler banks
(Nishio et al., 2001). (Backman et al., 1984). Formation of acidic sul-
fates is reduced by running boiler with low level of
Figure 10-12, Effect of operating conditions on corrosion rates of superheater tubes from two boilers.

Figure 10-13, Cracked 304L floor.
excess air and low sulfur dioxide emissions. the main cause. Several means of control were at-
tempted to reduce corrosion in carbon steel walls.
Sulfur dew point Studding the wall or using refractory helped, but
required extensive maintenance. It wasn’t until
When flue gases cool their capability of hold- the composite tubing appeared after 1972 that this
ing species in gaseous form weakens and such problem decreased. In late 1980’s it was thought
species start condensing. The temperature where that corrosion with composite tubing was non
condensed species appear is called dew point existent (Sandquist, 1987).
temperature. Acid dew point means conditions
where this condensate is acidic. Typically sulfur The second wave of large problems hit when
dew point refers to conditions where sulfur triox- composite floor tubes started cracking, Figure
ide starts condensing. Also HCl condensation is 10-13, in the late 80’s. In the beginning it was
possible. thought to be only a specific problem occurring
in some individual boilers, but around 1990 it
was recognized that it was a wide spread industry
10.2 FURNACE CORROSION problem. Much progress has been made. Several
possible reasons for cracking of stainless steel 304
Recovery boiler furnace walls needs to be water- have been discovered. Alloy 825 and modified
cooled, gas tight, easy to maintain, very safe, low alloy 835 (Sanicro 38) tubes seem much less prone
cost and easy to manufacture (Björklund 1963). to cracking in floor conditions than 304L.
Even today the furnace walls lack in safety, cost
and ease of manufacture. Main reason for this is The third wave of furnace problems hit in the
the difficult process conditions. 1990’s, when primary airport cracking started
causing large problems (Falat, 1996). Currently
One of the most researched corrosion phenomena this is still a question awaiting answer. Use of
in recovery boilers has been furnace wall and floor highly alloyed steels (e.g. alloy 825 or alloy 625)
tube corrosion. The first major corrosion prob- has not solved this problem.
lems emerged in early 1960s. Heavy wall corro-
sion in two recovery boilers at Kaukopää mill was
discovered in 1961 and later in others (Soodakat- Cracking of floors
tiloiden, 1968). Recovery boiler furnace walls were
After 70s almost all recovery boiler floors were
corroding rapidly and corrosion rates of up to 1
made of 304L. Cracking of 304L was first identi-
mm/year were recorded. At that time there were
fied in Scandinavia and then in early 90s in North
number of accidents and wall tubes of new higher
America (Keiser et al., 2004). Klarin (1993) at-
pressure boilers were experiencing short lifetime.
tributed the cause of these attacks to higher black
Sulfidation of iron oxide layer was found to be
liquor dry solids and mill closure. These changes
that form can form a molten phase in floor tube
operating temperature.
Poor flow at some floor tubes has been suspected

to cause high temperatures and floor cracking.
Simultaneous measurements of tube tempera-
ture excursions and flow at the same tubes have
confirmed that in measured boilers there is flow at
all times in the tubes exhibiting temperature peaks
(Koivisto and Holm, 1998).
Figure 10-14, Floor tube temperature excursion.

General corrosion
work to create hotter shallower char bed which Kraft recovery boiler furnace has very reducing
is richer in potassium and chloride. All this will environment. Water shift reaction causes high
increase the probability of smelt tube contact. levels of carbon dioxide and hydrogen. In addition
the burning black liquor droplets can cause very
304L and carbon steel have different thermal high levels of reductive gases. But general corro-
expansion behavior. Therefore stresses occur when sion seems not to be the limiting factor of current
composite tubes made of 304 are subjected to recovery boiler furnace life.
thermal cycles, Figure 10-14. Thermal spikes dur-
ing operation were found (Abdullah et al., 1999).
This caused early suspicion that thermal stresses Smelt corrosion in floors
were the main cause for floor tube cracking. Ther-
In kraft recovery boiler floors there appear large
mal stress may assist, but is no longer considered
patches of metal loss. These spots are so big that
the main culprit (Keiser et al., 2001).
it cannot be called pitting. One probable cause is
molten smelt corrosion. If steel tubes are subjected
The cause of thermal spikes in the recovery boiler
to molten sodium compounds, very rapid corro-
floor remain unknown. At first it was assumed
sion occurs. Normally contact is prevented by a
that problems in water circulation were causing
frozen layer of smelt on the tubes.
thermal spikes. Extensive bottom monitoring has
found no correlation between abnormal flow and
Layer on the tubes can melt if subjected to high
thermal spikes (Hogan, 1999). Current theory as-
heat flux e.g. from startup burners. There is smelt
sumes that cracks in the frozen layer cause vertical
flowing along the tubes and out from smelt spouts.
smelt flow in the bed, which is seen as thermal
Large flow could eat away protective smelt layer.
spikes in bed.
The layer close to the tubes can contain low melt-
Most of the analyzed cracks seem to exhibit stress
ing compounds and so corrodes the tubes. There
corrosion cracking (SCC). In SCC strain in the
is evidence that high sulfidities are found at the
tubes aids progression of chemical corrosion. For
surface of floor tubes. This indicates in turn that
the time being nobody has proved which chemical
polysulfides exist. Smelts with polysulfide in them
is causing it and how this corrosion can proceed
have very low melting temperature (Backman et
under the frozen smelt layer.
al., 2002).
Some researchers attribute floor tube cracking to
phenomena during water wash. During a water
wash there is a liquid electrolyte (water + bed Galvanic effects
material especially sulfide). This solution has been
show to cause SCC in laboratory conditions (Tran It’s well known that carbon steel floor tubes cor-
et al., 1999). rode fastest close to the weld to composite tubing,
Figure 10-15. This is one disadvantage of carbon
There is evidence that sulfur, chlorides and steel floors. Koivisto and Holm (2000) made stud-
potassium enrich to recovery boiler tube surface ies with different material combinations trying to
(Backman et al., 2002). All these can lower the find out how strong this so called galvanic effect is
melting temperature of the smelt. They can make in each case. Results show that compared to 304L,
the frozen layer thinner and easier for the smelt to Sanicro 38 is slightly better but Sanicro 65 in-
penetrate. They also decrease the viscosity of the creases the corrosion of carbon steel significantly.
molten smelt, which then is able to flow more eas- One coating ( 45CT ) was also tested and its’ effect
ily towards the floor tubes. Especially polysulfides on corrosion was similar to that of Sanicro 65 (i.e.

Figure 10-15, Galvanic corrosion of carbon steel.
strong galvanic effect). of thermal cycles as compared to 304L type com-

The experience is that with an adequately pro- posite before cracks are initiated (Karhula, 1992).
tected and properly cooled floor the corrosion rate This is in accordance with the thermal expansion
of the carbon steel floor is low. Highest rates that properties of the cladding materials i.e. the bigger
have been measured are of the order of 0.3 mm/a the difference in the thermal expansion between
near the weld to the composite (304L). Thinning the substrate and the cladding, the lower the fa-
can be monitored with regular thickness measure- tigue limit is. If carbon steel is used, then thermal
ments and necessary precautions can be taken stresses are of much less concern.
when they are needed.
When the lower walls are of Sanicro 38 type

Furnace environment
composite then the corrosion of carbon steel due Floor tube cracking is in most cases found to be
to galvanic effect will be even smaller according to environmentally assisted. This means that there
laboratory tests.
Thermal fatigue
Thermal cycles take place in recovery boiler
floors. A cycle is initiated each time boiler is shut
down and started up. Such spikes have also been
observed during normal operation. Cracks can
be created when a certain amount of cycles has
occurred. The higher the stresses or strains as-
sociated with a thermal cycle the fewer cycles are
needed for a crack to initiate.
Typical thermal fatigue cracks, Figure 10-16, are

found in recovery boiler walls. Normally these
failures can be traced to occurrence of very large
temperature peaks. The main risk with thermal
fatigue is the fracture doesn’t stop at the interface
between carbon steel and cladding but continues
to grow into the carbon steel.
Test results show that Sanicro 38 type composite

Figure 10-16, Thermal fatigue crack.
material can withstand three times the number

must be some chemicals, which at certain tem- Recovery boiler airports suffer corrosion. Some of
perature assists crack initiation and propagation. the recorded corrosion phenomena include cold
In addition to this chemical and temperature an side corrosion, membrane cracking and recently
electrically conductive electrolyte is needed. tube cracking (Keiser et al., 2004).
The Savcor company in Finland has measured the Cold side corrosion is caused by molten hydrox-
bed electrical resistance with a probe they have ide corrosion. There vaporized sodium hydrox-
developed and they have seen that bed resistance ide deposits on cold back side of air port tubes.
varies a lot from “normal” between 2 and 5 k cm Especially boilers with cast air tubes have suffered
to as low as 0.1 k cm and sometimes even lower from this corrosion mechanism. As rate with com-
(Laitinen et al., 1998). This indicates the exist- pound is much higher than with carbon steel, this
ence of molten phases in the bed material during doesn’t normally present great concern.
normal steady state operation.
At upper and lower end of air ports there is larger
Doug Singbeil and his group in Paprican have width of membrane, because of tube bends. This
made SCC tests in simulated water wash condi- membrane is then at higher temperature and
tions both with artificial salt mixtures and with exhibits higher corrosion rates. Corrosion can be
real smelt taken from a boiler. Tests show that augmented by periodic temperature excursions
these environments (salt + water) can cause crack- (Tay et al., 2004).
ing at temperatures ~170 oC. They have found a
threshold stress about 170 MPa and a threshold There are differences with external conditions
temperature about 160 oC for 304L (Keiser et al., between air ports even close to each other. The
2001). It seems that the presence of sodium sulfide following affect the corrosive environment: pres-
will prevent complete dryout until about 180 oC ence of molten smelt on the char bed surface, flow
(Tran et al., 1999). of molten smelt towards the air port tubes and
presence of large pieces of char near primary air
Preet Singh from IPST has made numerous tests ports. It has been suggested that a combination
in different artificial water wash solutions using of several events may lead to tube temperature
the slow strain rate test method. These experi- excursions at primary air ports (Vafa, 2003).
ments show that 304L steel may suffer from SCC
already at very low temperatures like 50 oC. These
tests, however, show only if the alloy is prone to 10.3 EROSION
SCC in test environment but don’t tell if SCC re-
ally takes place in real recovery boiler water wash Gas side heat transfer surfaces in boilers are
environment. subjected to passage of ash particles. Erosion
is caused as these particles hit the heat transfer
surface. Erosion rate increases with kinetic energy
Airport corrosion of the particles. Erosion depends strongly on ash
Figure 10-17, Corroded load burner sleeve.

Table 10-1, Requirements for boiler feed water according to DENÅ requirements.
Drum maximum pressure MPa(a) 2.5 3.6 6.8 9.1 12.6 16.1
Superheated steam pressure MPa(a) 2.3 3.3 6.3 8.5 11.1 14.1
pH25 upper limit 15.1 15.0 13.5 12.8 11.8 10.5
pH25 lower limit 9.5 9.5 9.0 9.0 8.5 8.5
p-value mval/kg 8.0 6.0 2.0 0.75 0.20 0.05
p-value if phosphates mval/kg 1…8 1…6 1…2
Conductivity25 mS/m 400 350 80 40 15 4
Na+K mg/kg 800 650 150 80 30 8
PO4 mg/kg 10…20 <15 <15 2…6 2…6 2…6
Silicates mg/kg 108 56 7.0 3.0 1.0 0.35
characteristics and local velocity field. high pH. This corrosion is characterized by loss
of magnetite film, irregular patterns and is often
The erosion problems caused by ash in recovery referred to as “caustic gouging”.
boilers are small. Firstly the flue gas velocities are
moderate (4 … 12 m/s). Secondly the ash in kraft Hydrogen attack results in loss of metal strength.
recovery boilers is typically rather soft. There are no visible signs to indicate wether this is
happening.
A significant source of erosion in recovery boilers
is erosion caused by sootblowers. Sootblowing Main means of steam/water side corrosion protec-
steam contains condensate droplets. When blown tion is creation of a protective oxide film on the
these droplets hit the tubes. If the distance to the metal surface. Steam/water circuits in power plant
tube is short (velocity is high) tube erosion will depend on such films for their integrity. The prime
occur. Erosion can be partly controlled by adding aim of water treatment is to maintain conditions
high resistance erosion control plate on top of the where these protective oxide films are formed and
tubes. are stable (MPSP, Vol. E, 1992).
In addition to heat transfer surface erosion the Oxygen corrosion

fuel transport and handling equipment suffer
from erosion. There are number of cases related to Oxygen corrosion causes pitting of tube inside
erosion corrosion in black liquor pipes especially surface. Rapid failure of a tube can occur with
in elbows and contractions. relatively small overall metal loss. Even small
amounts of dissolved oxygen can cause severe
damage. Oxygen is extremely corrosive when
10.4 WATER SIDE CORROSION present in hot water. Components that raise water
temperature rapidly, such as feedwater heaters
The water side of boiler system is subject to several
and economizers are particularly susceptible to an
types of corrosion. These types of corrosion can
oxygen corrosion attack. The temperature increase
result in severe damage that will affect the safety
provides the driving force that accelerates the cor-
and operational reliability of the boiler. To prevent
rosive reaction.
corrosion feedwater purity has strict limits, Table
10-1.
Out of service boilers are also susceptible to an
oxygen corrosion attack. Such an attack may typi-
Oxygen corrosion causes pitting of tube inside
cally be found at the water/air interface. If boiler is
surface. Rapid failure of a tube can occur with
shut down for longer periods, special procedures
relatively small overall metal loss. Even small
such as using nitrogen are necessary to avoid cor-
amounts of dissolved oxygen can cause severe
rosion.
damage.
Almost complete removal of dissolved oxygen is
Acid corrosion occurs when feedwater has too low
necessary. Oxygen is removed mechanically in
pH. This corrosion is characterized by a general
deaerator and chemically by a suitable chemical
wastage of the metal.
dosage. Small residual oxygen content is needed to
prevent flow assisted corrosion.
Caustic corrosion occurs when feedwater has too
Figure 10-18, Rate of corrosion of carbon steel in boiler water as a function of pH, (Navitsky and Gabrieli,
1980).
Acid corrosion boilers, the pH will be maintained above 8.5. One

typical method of control is to feed caustic solu-
When low pH water attacks steel, the attack is tion into the feedwater.
characterized by a general wastage of the metal,
as opposed to the localized pitting nature of an Caustic corrosion
oxygen attack. Since acidic conditions during the
operation will produce rapid and severe corro- High pH boiler water can be equally as corrosive
sion, pH control of the boiler water is critically as acid. High-pressure boilers are more susceptible
important. to caustic attack than the acidic attacks. Deionized
feedwater provides no buffering capacity to react
In normal operating pH range of boiler water very with the OH-ion and inhibit caustic concentra-
little corrosion occurs. Under these conditions, tion. Predominate contaminate in demineralized
steel readily reacts with water to form magnetite: makeup is sodium hydroxide (caustic).
3Fe + 4 H2O → Fe3O4 + 4H2 10-3 The caustic will concentrate under scale where
evaporation of water occurs. As steam contains
Magnetite forms as a thin, tenacious film on boiler almost no caustic the caustic concentration is
tube. This film protects boiler steel against attacks. increased. The concentrated caustic readily “dis-
Highly acidic or caustic conditions can dissolve solves” the protective magnetite film, forming
the magnetite film. complex caustic-ferritic compounds. The exposed
If pH in boiler water decreases below 4.5 magnet-

ite film is loosened from the tubes. ESP is required
as separated magnetite layer will block tubes.
pH control in high purity feedwater systems

is problematic. Feedwater as such is subject to
extreme pH swings from minor changes in acid or
in alkali concentrations. Usually some buffer-
ingchemical is used. A typical primary constitu-
ent of internal chemical treatment is phosphate.
Phosphate will serve to buffer the deionized water
to dampen pH swings.
Figure 10-19, Caustic corrosion of carbon steel tubes
To control acid corrosion in the high-pressure
(Kearns et al., 2001).
steel then reacts with water to reform the magnet- To prevent feedwater related problems dem-
ite film, which is again dissolved by the caustic. ineralization equipment must be operated with
The attack on the boiler steel continues for as long care. Good control of demineralizer operation,
as the concentrated caustic exists. regeneration and maintenance is needed to ensure
minimum sodium and silica leakage. There needs
The resultant metal loss assumes irregular pat- to be a pH buffer in the boiler water. This is done
terns and is often referred to as “caustic gouging”. with dosing continuously with suitable chemical.
When deposits are removed from the tube surface Some internal water treatment chemicals provide
during examination, the characteristic gouges an operating range.
are evidence of caustic corrosion. The gouges are
accompanied by a typical white salt deposit which High pressure boilers with high makeup are prone
is a combination of sodium carbonate (which is to deposits from the precipitation of chemical
the residue of the caustic after contact with the compounds. Substances such as calcium bicar-
carbon dioxide in the air) and crystallized sodium bonate if not properly removed from the makeup
hydroxide. water will decompose in the boiler water to form
calcium carbonate as follows:
Prevention of caustic corrosion is obtained by
maintaining both proper pH and a control of Ca(HCO3)2 → CaCO3 + CO2 + H2O 10-4
molar ratio of sodium in the boiler water.
Calcium carbonate has very limited solubility
Flow assisted corrosion and thus, the precipitated particles will agglomer-
ate at the heated surfaces to form a scale. Other
Flow assisted corrosion (FAC) causes wall thin- substances such as calcium sulfate are also scale
ning in carbon steel. Wall thinning occurs as producing. In this case, the scaling mechanism is,
protective oxide layer is dissolved into the water. however, a function of a retrograde solubility or a
This process is continuous and can be seen if decrease in solubility in water with an increase in
tube inside is examined. FAC has caused a lot of temperature. These substances have a low thermal
condensate line problems, economizer leaks and conductivity, and if left untreated, even a thin
boiler shutdowns for repairs. Flow assisted corro- scale will overheat the boiler tube.
sion does not affect superheated steam lines.
Table 10-2, Summary of waterside corrosion
Flow assisted corrosion rate increases with in- (adapted from Makkonen, 1999).
creasing water velocity. Especially bends, orifices Cause Corrosion mechanism
and headers, where flow is turbulent are prone to
Salts General corrosion
FAC. Corrosion rate is affected by temperature.
The highest rates have been recorded between 100 Oxygen General corrosion,
and 250 oC. If steels with more than 0.5 w-% are pitting
used the rate is much lower. Lower pH promotes Too low pH Acid corrosion
FAC. Presence of 10 to 20 ppb oxygen will help
maintaining protective layer and reduce corrosion Too high pH Base corrosion
rate. Non heterogeneous Pitting
oxide layer
Corrosion control Insufficient pre treat- General corrosion,
ment pitting
A major factor in preventing water/steam side
Chlorides General corrosion,
corrosion is keeping internal surfaces clean.
pitting
Major contributors to the formation of a heat
deterrent scale or deposit are contaminating Stress Stress corrosion crack-
elements present in the makeup water, metal ing
oxides transported to the boiler with feedwater, Vibrations Corrosion fatigue
contaminants from rest of the plant introduced
Undesired turbu- Erosion corrosion
into the condensate returned to the boiler and
lence / high velocity
solids present in condenser leakage. Having clean
surfaces means operating boiler with feedwater
and boiler water within the applicable guidelines.
The blowdown flow must be sufficient to prevent Summary of waterside corrosion
furnace side deposits.
Summary of waterside corrosion is shown in Table
10-2. Waterside corrosion is generally understood
Figure 10-20, Critical exposures as function of
elevation (Bauer and Sharp, 1993).
well. In spite of this waterside corrosion causes
problems frequently because the operators tend to
ignore all waterside issues. Keeping the feedwater
purity high is essential.
10.5 FURNACE DESIGN AND

MATERIALS
Recovery boiler furnace walls and floors have long
been under investigation for better materials. Es-
pecially the floor construction and materials affect
the recovery boiler safety Figure 10-20 (Bauer
and Sharp, 1991). Most of the critical leaks in the
furnace occur in the lowest 3 m of furnace walls.
Figure 10-20 shows some of the used possibilities Figure 10-21, Different recovery boiler walls, lowest
for lower furnace construction. refractory with studs, middle protective welded
cladding, highest finned membrane wall made from
The lowest is studding and refractory. Corrosion composite tubes (Tampella).
protection with studs is excellent, but this solution compound tubing, top, which is the most used re-
requires large amount of maintenance and repair covery boiler wall construction. All new recovery
work. The middle picture shows membrane wall boilers are of membrane design. Tangent tubing
with welded corrosion protection of alloyed mate- was phased out late 1980’s (Sandquist, 1987).
rial. Welded furnace wall is of comparable price to
It is important to protect the floor tubes from high
temperatures. Proper design of water circulation
lowers maximum temperatures. Sufficient water
flow needs to be maintained in the tubes to cool
them and to remove created steam bubbles. Usu-
ally the requirement is flow velocities > 0.5 m/s in
all tubes.
The floor angle in modern boilers needs to be

upwards with the flow. As bottom tubes are sup-
ported by steel beams they hang a little. Floor an-
gle helps to avoid parts where steam bubbles could
get stuck. Depending on the distance between the
Figure 10-22, Different recovery boiler corner support tubes, the angle needs to be from 2.5 to 4
designs; upper left corner tube completely blocked degrees. Smelt spouts need to be high enough so
from heat ransfer, upper right no corner tube, lower that all floor is covered with frozen smelt layer. Es-
left rounded corner, lower right another rounded pecially critical is the area farthest from the smelt
corner. spouts and the area right in front of the smelt
spouts. In practice it seems that 200 ... 300 mm
Table 10-3, Properties of typical floor tube materials.
Carbon steel 304L Sanicro 38 Sanicro 65

(Alloy 825) (Alloy 625)
Main elements Fe 20Cr–10Ni 20Cr–40Ni 20Cr–60Ni
Thermal expansion, 10–6 /oC 13.5 17.5 14.9 13.9
o 41 19 16 14
Thermal cond., W/m C
SCC resistance Excellent Low High Excellent
Corrosion resistance Low Moderate Excellent High
is enough. Too much height will cause problems struction. Carbon steel has also been lately used
when we try to empty the bed for shutdown. as floor material, Figure 10-23. Floors with carbon
tube are not susceptible to SCC corrosion.
Four different corner designs are shown in Figure
10-22. The amount of heat transferred to corner Extensive research in Finland has been carried out
tubes differs from one design to another. Of the where the corrosion of different materials in mol-
different recovery boiler corner designs only the ten polysulfides was studied. This research showed
design at the upper left is not recommended. The that Sanicro 38-type composite material had the
lower designs are safer than the upper designs. best corrosion resistance among the steels studied
(Mäkipää and Backman, 1998). Test panels made
Furnace tube materials of Sanicro 38 installed in 1991 and 1994 have not
shown any alarming corrosion. Nor have there
Some of the most typical furnace tube materials been any reported cracking found in recovery
are listed in Table 10-3. Many more have been boiler bottoms made from Sanicro 38 since 1995.
tried and for one reason or another abandoned. This highly alloyed material seems to have good
Carbon steel was the material of choice before the corrosion resistance, but it is fairly expensive.
compound tubing. Upper furnace from above the
highest air level is always made from carbon steel. Stainless steel 304L seems to last well in the fur-
Carbon steel seems to resist most corrosive condi- nace walls above the char bed. It is very resistant
tions at oxygen rich conditions. But bare carbon to sulfidation. SCC in the tubes at the furnace bot-
tubes can not resist black liquor burning on them tom tubes has made manufactures and recovery
nor smelt contact. Most modern recovery boilers boiler owners search for replacement materials in
use highly alloyed compound tube as floor con- that area (Keiser et al., 2000).
Figure 10-23, Modern carbon steel furnace bottom (Andritz).

Suppliers’ current recommendations are to use
modified alloys in the front and rear bends and
close to the side walls. To facilitate weld inspection
the whole lower furnace is often made of modified
alloys up to and over the primary air ports. The
present favorite is Sanicro 38 composite tube. In
addition of high content of chrome and nickel the
tube has about the same thermal expansion coef-
ficient that the carbon steel.
Sanicro 65 (Alloy 625) composite tubing is an-

other possibility. It has very favorable properties
considering thermal fatigue and stress corrosion
cracking. There are some reports of failure. So use
of 625 needs more study at the moment. Another Figure 10-24, Burning of membrane material (Riv-
area under study is the air port cracking (Keiser ers et al., 2002)
et al., 2002, Keiser, et al., 2004). Primary air ports
and smelt openings seem to exhibit cracking.
Thermal cycling and smelt contact are suspected
causes.
Use of compound tubing has about 30 year his-

tory in recovery boilers. Compound tubing is
expensive and the selection of materials is limited.
Some competing alternatives are chromizing of
tubes (Plumley et al., 1989, Labossiere and Henry,
1999). Another popular method is spray of plasma
coating. Compound surface can also be replaced
by welded surface. Of all above methods quality
control is easiest with compound tube.
Membrane materials
Membrane materials should be similar to the tube
material used. Carbon steel fin is used in case of
carbon steel tubes. Either composite membrane or
totally stainless steel membrane is used in case of Figure 10-25, Refractory with studs to protect the
composite floor tubing. furnace wall (Quest Integrated Inc)
anymore widely used in recovery boilers.
Fins receive thermal radiation and need to con-
duct heat to the tube proper. Fin surface is thus
at higher temperature than the tube surface. In 10.6 SUPERHEATER DESIGN AND
high heat flux areas and with wide fins this can MATERIALS
lead to tube cracking, Figure 10-24. A composite
membrane has better thermal conductivity as Recovery boiler superheaters suffer from corro-
compared to solid material, which is important sion, design issues and operation issues. Opera-
especially in case of wide tube spacing. tion problems that cause superheater failure can
be (McMillan, 2004)
1) Condensate blocked tubes during start up
Refractory and studs 2) Water from hydrotest not being evacuated
Small studs can be welded to tube and then during start up
covered with refractory, Figure 10-25. Refrac- 3) Water carryover from the drum
tory is also a fair corrosion protection. It should 4) Desuperheater spray water quality control
be remembered that both refractory and studs issues
need regular replacement. It is also impossible to 5) Sootblower action
inspect a floor for faults after it has been studded. 6) High temperature cycling caused by poor
Because of this neither refractory nor studs is superheating control
Even the most modern recovery boilers suffer
from superheater corrosion (Salmenoja, 2004).
Corrosion is the main problem that limits the
ability of kraft recovery boiler to produce electric-
ity (Bruno, 1997). In e.g. coal fired boilers much
higher superheater temperatures are typically
used. In comparison to coal fired boilers kraft
recovery boilers have higher rates of alkali metals,
chloride in gaseous form and often highly reduc-
ing conditions caused by carryover particles On
the other hand levels of some high temperature
corrosion causing substances like antimony, vana-
dium and zinc are typically low.
Loss of tube thickness can be caused by sulfida-

tion, alkali or chloride corrosion. Typically super-
heaters exhibit higher corrosion resistance if their
tube materials have higher contents of chromium
(Blough et al., 1998).
Figure 10-26, Effect of material and temperature
on recovery boiler superheater corrosion, effect of
tube temperature on corrosion rate in field tests Effect of steam outlet
(Fujisaki et al., 1994).
temperature
Main steam temperature is the main parameter
that affects the choice of superheater materi-
als. The rule of thumb is to keep the superheater
surface temperature below the first melting
temperature of deposits (Salmenoja and Tuiremo,
2001). Corrosion rates in final superheaters are
increased because superheater material tempera-
tures are highest, Figure 10–26. As can be seen
there typically is some temperature range where
the corrosion rate is acceptable. Increasing tube
temperature by some tens of degrees can signifi-
cantly increase corrosion rate.
Figure 10-27, Effect of steam velocity and tube
position on superheater tube surface temperature, Steam side heat transfer coefficients in typical
51 mm tube, behind nose calculated with average recovery boiler superheaters are low. Superheater
heat flux, furnace radiation calculated with heat surface temperature can be tens of degrees higher
flux from furnace. than the bulk steam temperature, Figure 10-27.
It can easily be seen that surface temperatures
and thus corrosion rates are greatly affected by
superheater positioning. Furnace radiation can
effectively be reduced by placing a screen to block
radiation heat flux. Therefore placing the hottest
superheaters behind the nose or screen will sig-
nificantly decrease corrosion.
Typical materials
Typical primary superheater materials, when they
are protected from direct furnace radiation are
carbon steel. Secondary and tertiary superheater
materials contain often 1 to 3 % Cr. These kinds
of materials are easy to weld and have good cor-
rosion protection. T22/10CrMo910 material can
Figure 10-28, Effect of chromium content on corro- usually be used up to 495 oC steam outlet
sion rate in laboratory tests, (Fujisaki et al., 1994).

Table 10-4, Example compositions of various superheater materials (Note that actual composition will vary
depending on the tube manufacturer and the production batch).
Cr Ni Mo Si Cu Al Mn C Fe Other
15Mo3 0.3 0.3 0.3 0.26 0.3 0.65 0.16 97.6
SA213 - T12 1.0 0.5 0.45 0.45 0.1 97.5 P,S
13CrMo44 1.0 0.3 0.3 0.26 0.3 0.55 0.14 97.1
SA213 - T11 1.25 0.50 0.75 0.45 0.15 96.9 P
10CrMo910 2.3 1.00 0.50 96.1
SA213 - T22 2.25 1.00 0.75 0.45 0.15 95.4 P
X10CrMoVNb91 9,0 0,3 1,00 0,40 0,50 88,7 V, Nb
SA213 - T91 9,0 0,33 0,99 0,24 0,03 0,45 0,11 88,7 Cb, N
HCM 12 12.0 1.00 86.9 V, Nb
X20CrMoV121 12.0 0.5 1.00 0.50 0.50 0.23 85.2 V
AISI 304 19.0 9.5 71.4
SS3338 18.0 12.0 0.72 0.15 69.2 Nb, Ta
AISI 316 17.0 11.0 2.70 69.2
AISI 321 18.5 10.2 0.75 2.00 0.06 68.5
AISI 347 17.5 10.3 1.30 0.70 1.60 0.08 68.4 Nb, Ta
253 MA 21.0 11.0 1.70 66.2 N, Ce
2205 22,0 5,5 3,20 1,00 2,0 66.2
Esshete 1250 16.0 11.0 1.25 1.00 7.0 62.5 V, Nb
AISI 309 23.0 14.5 0.70 1.50 0.10 60.2 P
HR2M 22.2 14.4 1.49 0.55 3.14 0.03 58.1 N
YUS170 24,4 13,2 1,50 0,78 0,58 0.02 59.5 N, P
AISI 310 25.0 20.5 0.50 0.08 53.8
Alloy 800 HT 21.0 32.5 0.50 0.40 0.40 0.75 0.05 44.3 Ti
HR3C 25.0 29.5 0.42 1.28 1.10 0.06 42.6 N
AC66 27.0 32.0 0.25 0.02 0.8 0.06 39.2 Ce, Nb
Sanicro 28 27.0 31.0 3.50 1.00 0.05 0.01 37.3
Sanicro 38 21.0 38.0 2.50 0.30 1.70 0.80 35.6 Ti
Alloy 825 22.0 39.0 3.50 0.36 0.50 34.5 Ti
HR11N 28.5 41.2 1.06 0.12 0.50 0.01 28.6 N
Nicrofer 45 TN 27,0 46,9 2,70 0,20 0,08 23,0 Re
Super 625 21.0 52.7 8.97 0.21 0.28 0.20 0.01 15.8 Nb
Alloy 600 15.5 74 0.25 0.26 0.50 0.07 9.3 V, Nb
Sanicro 65 21.0 61.0 8.40 0.35 0.38 8.8
Alloy 625 21.0 58.0 9.0 0.50 0.4 0.50 0.10 5.0 Nb,Ti
temperatures (Clement, 1990). With higher chrome content of the superheater tube is in-
temperatures and higher chloride and potassium creased, Figure 10-28. Similar trend was found
contents in the black liquor it is advisable to use from Swedish studies in Norrsundet recovery
higher chromium containing tubes. boiler (Eriksson and Falk, 1999). They found that
alloyed austenitic materials 304L and Sanicro
Fujisaki et al. (1994) found that recovery boiler 28 had much better corrosion resistance than
superheater corrosion is much reduced when high alloyed ferritic materials SS2216 and X20.

pressure loss, lower mass flow and higher end
temperature. The shorter routes will have lower
pressure loss, higher mass flow and lower end
temperature. We can balance corrosion by using
shorter routes for outside bends. These outside
tubes experience high heat flux and have thus
higher temperature difference between tube and
steam. But reduced path can lower steam tempera-
ture and balance corrosion rate. This procedure is
in more detail described by LaFond et al (1992).
Sealing the roof

Superheater tubes and the furnace roof need to for
a gas tight construction. This is usually done by
box made of steel plate, Figure 10-30. If the seal-
ing is not tight or leaks corrosive salt builds on top
of the roof. Superheaters need to be supported.
Hanger rods are tied to seal box. Individual tubes
Figure 10-29, Outermost tube temperature can be hang from horizontal supports inside the box.
decreased by arranging shorter flow path. Thermal movement needs to be accounted for.
This means that superheater tubes can not hang
from headers.
10.7 SCREEN DESIGN AND

MATERIALS
Screens are built as first heat transfer surface after
the furnace to protect other heat transfer surfaces
from radiation and carryover. Screens are typi-
cally just tube panels, where tube spacing is very
narrow. Most of the screens are made of carbon
steel as their temperature is close to the tube wall
temperature. Corrosion of screen tubes is low
because of the material temperature.
Figure 10-30, Superheater roof seal box arrange-
ment. The most problematic screen design element is
the damage that can be caused by falling deposits
Stainless steel lower bends in hottest superheaters (Villarroel et al., 2004). Especially at the corner
have been used for tens of years. From field tests it of front wall and roof large deposits can build up.
seems that no appreciable corrosion occurs up to If a deposit drops from the roof down to a screen
the metal temperature of 540 oC (Arakawa et al., it can break the tubes. Deposit damage can be
2004). minimized but not eliminated by building fins or
extra thickness to screen top surface.
Effect of tube arrangement

10.8 BOILER BANK DESIGN AND
Recovery boiler superheaters typically consists
of some 10 … 40 panels that have some 20 ... 40 MATERIALS
tubes each. To control steam velocity in the tubes
Two drum boiler banks have very good reliability
and so the pressure loss, parallel tube passes are
record. They, however, suffer form mud drum
used. Most typically from three to five tubes are
corrosion (Labossiere and Henry, 1999). This type
used. These superheater tube parallel passes are
of corrosion is caused by steam from sootblow-
often all of same or close to same length.
ing wetting the salt at tube joints in lower drum.
The progress of the near drum corrosion can be
Some control of tube surface temperatures can be
monitored with ultrasonic equipment (Soar et
made by choosing shorter and longer tube routes,
al., 1994). Two drum boiler banks are made with
Figure 10-29. The longer routes will have higher

plain tubes tapered at both ends. The joints can
be made without welds up to the pressure of over
8.0 MPa.
Another problematic failure type is caused by vi-

brations from gas flow or sootblowing. The longer
the free tube length the higher the resulting stress
at joins. Industry practice states that maximum
length of free tubes is about 8 meters. Typically
longer tubes are too flexible and will vibrate too
much. This will create cracks and faults in few
years. All larger recovery boilers have vertical
boiler banks.
Vertical boiler banks consist of hundreds of tubes.

Often you find from 60 to 100 tube rows between
Figure 10-31, Upper end of vertical flow boiler
sidewalls. Each tube row consists of tubes welded
generating bank showing left straight fin and right
together with wide fins to improve heat transfer
cut fin, which minimizes thermal stresses around
and create rigidity. In modern boilers the vertical
weld (Bruno and Hägerström, 1998).
boiler bank has one or two sootblowing cavities
open from top to bottom.
Finned design causes temperature differences In economizers the loss of tube thickness can be
between fin and tube. This will create high stresses caused by gas side corrosion; sulfidation and acid
at fin ends. To prevent these stresses cut fins are dew point corrosion or water side erosion corro-
preferred, Figure 10-31. sion.
Some pluggage problems have been reported on Lower ends of economizers in recovery boilers
boiler bank lower end (Sandquist, 1987). If lower suffer from water side erosion corrosion. Typically
headers are located too close to each other they the symptoms are worst in the first few meters of
trap falling material. Placing a sootblowers close economizer tube.
to the lower end is also critical to keep it free from
fouling. Recovery boiler economizers have hundreds of
weld joints. Each weld even after inspection is
potentially problematic. Therefore the preference
10.9 ECONOMIZER DESIGN AND was to avoid unnecessary welds and use only
MATERIALS continuous tubes without butt welds. Largest boil-
ers have economizer lengths of 27 meters. Carbon
Modern economizers are of vertical design. Earli- steel tubes maximum length is some 23 meters.
est horizontal economizers had severe plugging So in the newest boilers this preference can not be
problems and were replaced by cross flow design. adhered to. Attention should be paid to qualifica-
Cross flow economizer had lower heat transfer tion of welds in economizer tube joints.
coefficients and was more prone to plugging than
the modern vertical economizer.

11 Emissions
Environmental issues matter to the pulp and Table 11-1, Typical emissions (dry, 3 % O2) to air
paper industry (Vasara, 2001). In reducing emis- from recovery boilers (IPPC, 2001).
sions to air the performance of the recovery boiler
ppm mg/m3n mg/MJ kg/ADt
is the key factor. The pulp and paper industry has
in the recent past significantly reduced emissions Sulphur 35 - 275 100 - 800 60 - 250 1-4
from recovery boilers (Bruce and Van der Voored, dioxide
2003). Total elimination of all emissions to air is TRS <7 < 10 <5 < 0,05
impossible, but proper recovery boiler design and NOx 50– 125 100 - 260 50 - 80 0,8 – 1,8
operation can minimize the impact of harmful
Dust 10 - 200 0,1 – 1,8
emissions.
Table 11-2, BAT emission levels from kraft pulping
Emissions to air from recovery boilers can be
process and recovery boilers (IPPC, 2001).
broadly categorized with the age of the boiler,
the black liquor dry solids and the boiler load. Emission Kraft pulping Recovery boiler
But emissions do not only depend on the process Particles kg/ADt 0.2 … 0.5 0.2 … 0.5
conditions and the type of equipment. Significant SO2 kg(S)/ADt 0.2 … 0.4 < 0.1
emission decreases can be achieved when trained,
NOx kg(NO2)/ADt 1.0 … 1.5 0.7 … 1.1
educated and motivated personnel run a properly
maintained unit with high level process control TRS kg(S)/ADt 0.1 … 0.2 < 0.1
optimization (IPPC, 2001).
It is possible to reduce the sulfur dioxide emis-
sions from recovery boilers to near zero levels.
11.1 TYPICAL EMISSIONS Additional sulfur releases occur from lime kiln,
NCG-boiler and as uncontrolled gaseous releases.
Kraft recovery boiler is the largest producer of The role of uncontrolled gaseous releases in TRS
gaseous effluents from the pulp mill. Typical emissions can clearly be seen. Lime kiln contrib-
emission levels of main controlled substances are utes some 25 … 35 % of the mill NOx release. A
shown in Table 11-1 (IPPC, 2001). It is evident significant source of mill ammonia is the green
from that table that at present stage some kraft and white liquor system (Kymäläinen et al., 1999).
pulp mills still produce significant sulfurdioxide
and dust releases. The NOx levels from kraft pulp-
ing are well below the average NOx releases from 11.2 REDUCED SULFUR SPECIES
energy production. Kraft pulping gets accused
of producing bad smell. The smelly component Reduced sulfur species are smelly gases. They
release (TRS) from recovery boilers is however of include hydrogen sulfide, methyl mercaptan,
minor magnitude. dimethyl sulphide and dimethyl disulphide. The
main reason for limiting their emission is to
When determining emission levels it is often improve local environment around the mill. The
advantageous to look at total emissions from pulp main source of TRS in modern kraft pulp mills
mills. BAT emission levels from kraft pulping are uncollected vent gases and other points where
process and recovery boilers are shown in table mill liquors are handled in contact with the air.
11-2 (IPPC, 2001). These figures should include Therefore in addition to NCG, the mixing tank
lime kiln, NCG-combustion and uncontrolled vent gases and dissolving tank vent gases need to
gaseous releases. In integrated mills the bark be burned in the furnace.
and auxiliary boiler emissions are not typically
included to pulp mill releases, because modern The TRS emission is often caused by cold char
pulp mills can produce all their heat from just bed and poor mixing in the furnace. When black
black liquor. liquor dry solids is increased furnace temperature
Emissions 11-1
Figure 11-1, TRS emissions from modern recovery boilers.
gets higher. Typical emission levels for modern 11.3 CARBON MONOXIDE
recovery boiler main stack at over 65 % liquor are
1 … 5 mg/m3n or 5 ppm at 3 % O2 and dry gas, Incomplete combustion produces carbon mon-
Figure 11-1. Reduced sulfur species (TRS) will oxide (CO). It is a colorless, odorless, poisonous
oxidize in a modern furnace to sulphur dioxide. gas. CO formation increases dramatically when
This requires sufficient temperature and mixing. black liquor is burnt at very low excess oxygen
Improving recovery boiler air system below the conditions. Carbon monoxide is emitted from
liquor gun level can be used to lower TRS emis- all combustion sources including motor vehicles,
sions (La Fond et al., 1994). power stations, waste incinerators, domestic gas
boilers and cookers.
Figure 11-2, Typical carbon monoxide levels from modern boilers.

Figure 11-3, Renewable carbon dioxide from black liquor combustion.
Black liquor is a biofuel. Carbon dioxide produced

Carbon monoxide has limited effect on human when burning organics from wood on black liquor
health. It attaches itself to red blood cells prevent- is assumed to be reabsorbed back to organic
ing the up-take of oxygen. Carbon monoxide matter in forests. Therefore the only fossil carbon
contributes indirectly to an enhancement of global dioxide production from kraft recovery boilers is
warming through reactions with other gases in the from use of auxiliary fuel during startup, upsets or
lower atmosphere. for extra steam production.
CO correlates well with O2 content in the flue

gases. The higher the air ratio and the better the
mixing the lower is the CO emission. Higher 11.5 NOX
furnace temperature and longer residence time
decrease CO emission. The NOx and TRS emis- Nitrogen oxides (NOx) emissions from recovery
sion increase with decreasing CO (Vakkilainen boilers are mainly nitric oxide (NO). In the atmos-
and Holm, 2000). It is typically advantageous to phere it reacts to nitrogen dioxide (NO2). Nitric
run at some hundreds of ppm of CO to lower oxide is a colorless, odorless gas. Nitrogen dioxide
other emissions and keep reasonable flue gas heat is a reddish-brown gas with a pungent odor.
losses. Nitrogen oxides contribute to acid rain formation.
Nitrogen dioxide affects human respiratory system
With modern large boilers the CO level is not aggravating asthma and other lung diseases.
greatly affected by as fired black liquor dry solids
or boiler load. In small retrofitted boilers the fur- Major man-made releases of nitrogen oxides are
nace residence time might be too low to complete from automobiles and fuel combustion. Forest
combustion. fires and biological processes in soils and waters
produce also NOx.
11.4 CARBON DIOXIDE NOx levels in recovery boilers are typically 70

… 120 ppm at 3% O2.(Aho 1994). The emissions
Carbon dioxide (CO2) is odourles gas that is not from Swedish recovery boilers in 1998 were be-
harmful to humans. Carbon dioxide absorbs tween 60 and 140 ppm with average 92 ppm at 3%
radiation. It has been postulated that increasing O2 (Kjörk and Herstad, 2000). Typical emission
carbon dioxide levels in atmosphere cause global from modern recovery boiler is 60 ... 80 mg/MJ.
warming. Therefore reduction of carbon dioxide This level is significantly lower than most of the
production is of importance. coal and oil fired boilers can achieve.
Emissions 11-3
Figure 11-4, Typical recovery boiler NOx emission values.
(DeMartini et al., 2004). The highest reported

nitrogen content has been 0.5 w-% of bagasse
black liquor dry solids (Aho, 1994). The detection
accuracy of nitrogen in the fuel from commercial
laboratories is very poor so direct correlations are
often misleading. For example two black liquors of
nitrogen content 0.072 and 0.114 mass % get both
typically a value of 0.1 mass-%.
Aho et al. (1994) studied the reduction of nitrogen

in the black liquor. One can somewhat change
the nitrogen content by treating chips, making
Figure 11-5, Correlation of NO emissions in recov- cooking changes and treating black liquor. The
ery boilers to nitrogen in fuel (Plotted from data in results however have so far showed this to be
Forssén et al, 1997). uneconomical.
Recent changes in recovery boiler operation have Part of the nitrogen in black liquor is in the form
increased the NOx level. Examples are; increasing of NH3 and so easily vaporized. Parts of nitrogen
black liquor dry solids, combustion of NCG in can then be extracted from black liquor in the
recovery boiler furnace and combustion of dis- stripper area. This product may enrich to some
solving vent gases in the recovery boiler. parts of the recovery cycle so great care should be
taken when this waste stream is destroyed.
Effect of fuel nitrogen content to
NOx emissions NOx - reduction
When boilers are run with liquors from pulping The following reduction means have been tried in
of different wood species, high nitrogen contain- recovery boilers in an industrial scale
ing woods (hardwood) produce more NOx than - Optimisation of the feeding of air and fuel
softwood liquors. Field tests support the conclu- - Increasing number of air levels
sion that recovery boiler NOx emissions correlate - Feeding of the ammonia to the top of the
with nitrogen in the black liquor (Clement and furnace
Barna, 1993). Laboratory tests show that the - Vertical air
nitrogen release from black liquor seems to be - Oxidation /reduction processes
independent of wood species and cooking method
Several other methods can be used to reduce the Ammonia consumption is at 3000 tds/d
NOx of recovery boilers. Among the methods
known from the general combustion processes are NH3 kg/24h = 3000*3,86*0,771*100/1000 = 893
the following kg/24 h
- electron beam method
- catalytic conversion (SCR) NOx removal is
SCR can not be used in recovery boilers as the NO2 kg/24h = 3000*3,86*2,053*25/1000 = 594
fine aerosol particles and the sodium containing kg/24 h
gases pose challenges to the lifetime of the catalyst
(Anderson and Jackson, 1991). This comes as about 1,5 tons of NH3 per one ton
of NO2 removed.
Selective noncatalytic reduction
Approximate operating cost is then 2.000 – 2.500
Selective noncatalytic reduction is the next pos- €/t NO2. Ammonia in water solution costs 2 – 3
sibility to reduce NOx from recovery boilers. The € per kg of ammonia. For a 3000 tds/d recovery
use of noncatalytic reduction is difficult in recov- boiler the operating cost of ammonia would be
ery boilers as it requires a rather narrow operating about 3 million Euro per year.
temperature range around 950 oC to function.
NOx is removed by injecting some reducing sub- Oxidation-/reduction processes
stance into flue gas. Typical chemical components
used are ammonia, urea and their derivatives. The From chemical industry there are known several
main reaction path is. methods to oxidize NO to NO2 followed by cap-
ture of NO2 and reduction to N2. Known oxidizers
2NH3 + 2NO + 1/2O2 → N2 + 3H2O 11-1 are peroxide, ozone and chlorinedioxide. This
process can occur in a recovery boiler (Janka et
Ahlstrom Machinery and Gotaverken have tried al., 1998). The scrubber material needs to be able
the feeding of the ammonia to the top of the to withstand the oxidizer. Published operating
furnace in a recovery boiler (Lövblad et al., 1991). costs are from 2.300 to 3.100 €/t NO2. Trials of this
During the trials the gas temperature at injection process have been conducted in several mills.
point was about 1000 oC. With this technique a 15
… 35 % or 22 … 60 ppm reduction was reached. Use of chlorine dioxide is about 4 ppm for 1 ppm
The NOx level in the flue gases lowered from de- of NOx.
pending on how much reducing agent is used and
how big ammonia slip is allowed. To restrict the ClO2 kg/24h = 3000*3,86*3,149*400/1000 = 14586
ammonia, discharge below 5 ppm level only about kg/24 h
25% reduction was reached.
Removed NOx is
Table 11-3, Achievable NOx reduction levels
Nitro-
NO2 kg/24h = 3000*3,86*2,053*75/1000 = 1786
Normal Reduced
gen kg/24 h
w-% ppm mg/ mg/MJ ppm mg/ mg/MJ
m3n m3n required chlorine dioxide is about 8 tons per
0,08 83 170 55 62 128 41 removed ton of NO2.
0,10 97 200 65 73 150 49
Chlorine addition to recovery cycle would be
0,15 130 266 87 97 200 65
Cl kg/t pulp = 14586 kg/24 h / 1670 t / 24h = 9
With the spraying of the ammonia there exists kg/ t pulp
always some ammonia slip caused by the uneven-
ness of the spraying and NOx. Ammonia is known Normal chlorine input is 1 - 2 kg/ t pulp, so the
to cause low temperature corrosion and is in itself addition might increase the chloride level in the
a corrosive agent. mill significantly. In one Finnish mill the chlorine
level in black liquor would have increased from
A recovery boiler has 100 ppm of NOx at dry flue 0.3 w-% to 1.8 w-%. So the additional capacity
gases and dry flue gases 3,86 m3n/kgds. Level of reduction in recovery boiler must be considered if
ammonia in flue gases is 100 ppm/m3n. The am- chlorine based oxidizer is used.
monia content is then 1 ppm of NH3 to 1 ppm of
NO. 1 ppm NH3 = 0.771 mg(asNH3)/m3n.
Emissions 11-5
Typical mill emission levels
Table 11-5, NOx and SO2 emissions from Swedish
Typical emission levels of NOx (and SO2)in Scan- pulp mills in 2001 (Collected from company web
dinavia are listed in Tables 11-4 and 11-5. pages in 2002).
Produc- NOx SO2 NOx SO2
Table 11-4, NOx and SO2 emissions from Finnish tion t/a t/a kg/ADt kg/ADt
ADt/a
pulp mills in 2001 (Environmental report, 2002).
Husum 591000 1199 768 2.03 1.30
Produc- NOx SO2 NOx SO2
tion Väröbacka 327000 660 400 2.02 1.22
t/a t/a kg/ADt kg/ADt
ADt/a
Piteå 395000 743 126 1.88 0.32
Äänekoski 397000 963 465 2.43 1.17 Billingsfors 50621 93 49 1.84 0.97
Sunila 300000 594 106 1.98 0.35 Iggesund 308000 558 508 1.81 1.65
Kaskinen 368000 705 587 1.92 1.60 Munksund 215000 360 76 1.67 0.35
Joutseno 334000 618 377 1.85 1.13 Skärblacka 395000 654 356 1.66 0.90
Kemi 443000 801 22 1.81 0.05 Karlsborg 275000 449 377 1.63 1.37
Varkaus 178000 323 322 1.81 1.81 Aspabruk 159555 243 87 1.52 0.55
Valkeakoski 142000 243 107 1.71 0.75 Valvik 169400 256 142 1.51 0.84
Pietarsaari 527000 852 155 1.62 0.29 Norrsundet 245000 351 384 1.43 1.57
Kuusankoski 420000 666 69 1.59 0.16 Obbola 241000 336 271 1.39 1.12
Uimaharju 560000 844 155 1.51 0.28 Gruvö 630000 871 109 1.38 0.17
Lappeenranta 678000 1011 392 1.49 0.58 Östrand 395000 541 537 1.37 1.36
Rauma 470000 681 433 1.45 0.92 Mörrum 390000 520 360 1.33 0.92
Kemijärvi 178000 248 69 1.39 0.39 Skutskär 467000 613 602 1.31 1.29
Veitsiluoto, 312000 421 10 1.35 0.03 Mönsterås 658000 770 350 1.17 0.53
Kemi
Frövi 222400 213 61 0.96 0.27
Imatra 727000 981 410 1.35 0.56
Gävle 654500 622 196 0.95 0.30
Oulu 299800 353 24 1.18 0.08
Skoghall 528000 489 321 0.93 0.61
Kotka 121000 83 280 0.69 2.31
Average 340776 524 309 1.54 0.91
Average 379694 611 234 1.61 0.62
Table 11-6, Nitrogen oxide emissions from pulp
The emissions per ton of pulp produced vary a
mills as kg/ADt.
lot from mill to mill. The national average mill
size and the average emissions per ton of pulp Range Average
produced are very close to each other. If we count Finland 2001 1,2 – 2,0 1,61
only those mills that achieve sulfur emission be- Sweden 2001 1,0 – 2,0 1,54
low 0,5 kg/ADt, the corresponding NOx emission
USA (Pinkerton, 1,38
is 1,2 … 1,8 kg/ADt.
1998)
The lowest emission of NOx per ton of pulp IPPC BAT 1,0 – 1,5
produced has been achieved at a mill where the New pulp mill, hard- 1,4
corresponding sulfur emissions are the highest. wood
The nitrogen oxide emissions can be plotted. Ta- New pulp mill, 1,2
bles 11-4 and 11-5 have been plotted as function softwood
of total pulp production and scaled to 100 %.
Nitrogen oxide emissions from pulp mills Tables

11-4 and 11-5 are plotted as function of total pulp
Addition of nitrogen containing
production and scaled to 100 % in Figure 11-6. flows
One can see from that typical nitrogen oxide
emissions are around 1.5 kg / ton of pulp. Normal Burning of NCGs and dissolving tank vent
emission limits from pulp mills can be seen in gases will increase the nitrogen flow to furnace.
Table 11-6. Changes in N to furnace are directly related to
NOx emissions.

Figure 11-6, Nitrogen oxide emissions from pulp mills Tables 11-4 and 11-5 plotted as function of total pulp
production and scaled to 100 %.
Dissolving vent gases Table 11-7, BAT NOx emissions from recovery
boilers (IPPC report, 2001).
Dissolving vent gas flow is about 0.25 m3n(dry)/ Mill Country kg/ADt
kgds. Dissolving vent gases containing nitrogen
as ammonia at about 100 ppm/m3n(dry). One can Äänekoski Finland 1,748
assume that 30 % ammonia is converted to NOx Pöls Austria 1,600
(Crawford and Jain, 2002). The increase in NOx Värö Sweden 1,510
emission from recovery boiler is in the order of
Iggesund Sweden 1,500
NOx+ % = (3,67*100 + 0,30*0,25*100)/(3,67*100) Vallvik Sweden 1,440
* 100 -100 = ~2 % Kaskinen Finland 1,366
Örstrand Sweden 1,360
DNCG
Aspa Sweden 1,290
The flow of high volume low concentration NCG Skutskär Sweden 1,260
is typically 5 … 10 % of total air flow. The nitrogen Dynäs Sweden 1,240
in DNCG gases is mainly ammonia. Level of
Bäckhammar Sweden 1,220
nitrogen compounds in DNCG can be assumed
to be 10 ppm/m3n(dry) (Janka, and Tamminen, Husum Sweden 1,210
2003). One can assume that 30 % ammonia is Enocell Finland 1,186
converted to NOx even if the CNCG burner is in- Skoghall Sweden 1,140
stalled above the liquor guns. The increase in NOx
Sunila Finland 1,029
emission from recovery boiler is in the order of
Obbola Sweden 1,020
NOx+ % = (3,67*100 + 0,30*0,1*3,67*10)/ Stora Celbi Portugal 1,020
(3,67*100) * 100 -100 = <1 %
Joutseno Finland 1,013
CNCG Frövi Sweden 1,010

Mönsteras Sweden 0,950
The flow of low volume high concentration NCG Huelva Spain 0,880
is 0,010 … 0,025m3n/kgds. The nitrogen in CNCG
Wisaforest Finland 0,864
gases is mainly ammonia. Level can be assumed
to be 2000 ppm/m3n. One can assume that 30 % Oulu Finland 0,810
ammonia is converted to NOx. The increase in N Skärblacka Sweden 0,660
to boiler is
NOx+ % = (3,67*100 + 0,30*0,025*2000)/

(3,67*100) * 100 -100 = ~4 %
Emissions 11-7
Table 11-8, Volatile organic emissions from US kraft recovery boilers with dry bottom ESP and without
direct contact evaporator (NCASI, 2001).
MILL # C1 C2 J K M Avg.
Acetaldehyde ppm 1.8-2.3 0.5-0.9 0.2-0.4 0.5
Methanol ppm 7-22 4.70 2.4-2.8 4.3-9.7 1.7-2.0 2
Acetone ppm 0.14-0.21 0.33 0.17-0.20 0.10-0.29 0.1-0.2 0.1
Methyl ethyl ketone ppm 0.35 0.29 0.09-0.30 0.07-0.1 0.1
Benzene ppm 0.26-0.27 0.72 1.79 0.3-2.7 0.5
Methyl isobutyl ppm 0.81 0.5
ketone
Toluene ppm 0.12 0.02-0.03 0.5
m,p-xylene ppm 0.04 0.03 0.05
o-xylene ppm 0.04 0.02 0.05
Xylenes ppm 0.09 0.18 0.1
Styrene ppm 0.06 0.12 0.08 0.1
Alpha-pinene ppm 0.10 0.1
Beta-pinene ppm 0.06-0.21 0.1
Terpenes ppm 0.11-0.38 0.13-2.3 0.14-0.17 0.3
Formaldehyde ppm 0.98 1
VOC ppm 152 119 3-193 100
BAT recovery boiler NOx suspended in the air. They are one of the most
noteworthy air quality problems in urban areas
emission (Ohlström et al., 2000). Typically particulate
BAT recovery boiler emission levels in Europe matter comes from a variety of different sources.
are listed in Table 11-7. The best recovery boilers Especially smaller particle size can directly affect
achieve emissions little below 1.0 kg/ADt. NO human blood through respiratory system. Par-
emission strongly correlates with N-content in the ticulate matter comes from combustion sources,
fuel. road traffic, dusts, sea salt and biological particles.
Particulate matter is usually reported as total solid
mass flow. Additionally small size particular mat-
11.6 VOC ter is reported as PM10 or PM2.5 (particles with
diameter less than or equal to 10 or respectively
Volatile organic compounds (VOC) contributes 2.5 micrometers). 80 … 99 % of recovery boiler
to the formation of ground-level ozone in the particulate matter emission belongs to PM2.5
atmosphere. Some high molecular mass VOC are category (Mikkanen et al., 1994, Mikkanen et al.,
carcinogens (Froste, 1996). Common sources of 2001).
VOC emissions include vehicle exhausts and the
chemicals industry, with other minor releases also
occurring naturally from forest fires. The majority 11-1
of industrial emissions to air are from leaks during
production, use, storage, transport and disposal of where
carbohydrates. Volatile organic compound emis- e is the collection efficiency, -
sions from recovery boilers are low and mainly of A is the collector surface
low molecular weight. w is the collection migration velocity, m/s
V is the actual gas flow, m3/s
Most of the measured VOC (CxHy, CxHzOH, …) k is an empirical constant (0.4 .. 0.5 ..0.6), -
emission seems to come from the mixing tank
vent gases and dissolving tank vent gases. Even at high dry solids the emission after the
recovery boiler economizer is typically below 20
g/m3n. The dust emission is strongly dependent on
11.7 DUST EMISSIONS boiler load. It has been noted that firing with high
Dust or particulate matter is small particles black liquor droplet velocities and small droplets

Figure 11-7, Total uncontrolled dust emission from recovery boiler furnace.
causes excess carryover and high dust loading. oC fume particles that hit do not stick to each
Dust emission can be reduced by using electro- other. Sintering can occur at heat transfer surfaces.
static precipitators. There fume can form large agglomerates. These
agglomerates can re-entrain to flue gas when soot-
Effect of operating conditions to blowed. They show as larger 20 … 30 μm particles
at electrostatic precipitator inlet (Mikkanen, 2000,
dust emissions Janka et al., 2000). Amount of re-entrained dust
is 10 … 30 % of total dust flow (Tamminen et al.,
In kraft recovery boilers the ash is formed by
2000).
several mechanisms. Most of the ash is formed
through fume formation. Almost all ESP ash is
Chemical equilibrium calculations (Perjyd and
typically fume. Some of the black liquor fired into
Hupa, 1984) suggest that fume formation increas-
recovery boiler furnace gets caught with flue gas.
es as the lower furnace gas temperature increases.
Furnace gases then drag these particles with them
When boilers are run with higher dry solids the
out of the furnace. It has also been postulated that
furnace temperature increases (Vakkilainen and
during char combustion some intermediate size
Holm, 2000). Higher dust emission is then a trade
particles are formed.
off to lower sulfur emissions.
In the kraft recovery furnace the alkali elements
Most of the fume tends to end up in electro-
are vaporized. When flue gases are cooled the
static precipitator. Figure 11-7 shows ESP dust
vapors become supersaturated. Another mecha-
as function of black liquor dry solids. Furnace
nism of supersaturation is formation of species
temperature increases with increased black liquor
by reactions. E. g. sodium hydroxide and sulfur
dry solids, which increases sodium and potas-
dioxide react to sodium sulfate, which has lower
sium release (Frederick et al., 1995). ESP ash flow
vapor pressure than the reactants (Jokiniemi et al.,
forms an estimate of fume formation in recovery
1996). Resulting particle size is 0.5 … 1.3 μm.
boiler. Black liquor sodium and potassium release
depends also on the liquor i.e. wood species and
In recovery boiler furnace there are small metal
the pulping conditions (Backman et al., 1999). In
oxides and other impurities present, which will
addition the firing conditions affect the lower fur-
serve as starting nucleus for the vapors. This proc-
nace temperature. As can be seen then the actual
ess is called heterogeneous condensation. After
fume formation rate can vary a lot.
formation of nucleus the vapor condensation
keeps then supersaturation low.
Typical mill emission levels
Much of the particle size growth is result of ag-
glomeration. Small liquid particles collide with Typical emission levels in Scandinavia are listed in
each other and form new spherical particles. After Tables 11-9 and 11-10.
flue gas temperature decreases below about 550
Emissions 11-9
Table 11-9, Dust emissions from Finnish pulp mills BAT recovery boiler emission
in 2001 (Environmental report, 2002).
levels
Production Dust Dust
ADt/a t/a kg/ADt
The typical emissions, according to the IPPC
Äänekoski 397000 304 0,77 document, from recovery boiler are shown in
Sunila 300000 133 0,44 Table 11-11.
Kaskinen 368000 227 0,62
BAT recovery boiler emission levels in Europe
Joutseno 334000 93 0,28 are below 50 mg/m3n. The best recovery boilers
Kemi 443000 29 0,07 achieve emissions little below 0.2 kg/ADt.
Varkaus 178000 158 0,89
Valkeakoski 142000 547 3,85 11.8 SULPHUR DIOXIDE
Pietarsaari 527000 647 1,23
Sulphur dioxide (SO2) is a colorless gas. About
Kuusankoski 420000 125 0,30
half of world sulphur dioxide emissions are from
Uimaharju 560000 161 0,29 burning of fossil fuels (Lefohn et al., 1996). Large
Lappeenranta 678000 87 0,13 part of the rest comes from emissions from metal
Rauma 470000 191 0,41
production and large industrial plants. Pulp mills
used to be a major source of sulfur emissions, with
Kemijärvi 178000 127 0,71 most of it coming from recovery boiler.
Veitsiluoto, Kemi 312000 87 0,28
Imatra 727000 37 0,05 Sulphur dioxide is a major contributor to acid
rain. Sulphur dioxide irritates the eyes and air
Oulu 299800 98 0,33
passages. It may increase asthma or lung disease
Kotka 121000 87 0,72 symptoms. Sulphur dioxide is transported by air
Average 379694 185 0,67 for long distances.
The sulphur dioxide emissions are dependent

Table 11-10, Dust emissions from some Swedish on the dry solids and the black liquor sulphur to
pulp mills in 2001 (Collected from company web sodium and potassium molar ratio (Wallén et al.,
pages, 2002). 2004). The higher the dry solids the higher sulfur
Production Dust Dust
emission one can produce. Similarly the higher
ADt/a t/a kg/ADt the sulfidity the higher is the possible emission.
Husum 591000 428 0,72 Sulfur emissions from about 30 different boilers
running different loads at different dry solids and
Väröbacka 327000 840 2,57 black liquor sulfidities are shown in Figure 11-9.
Billingsfors 50621 36,9 0,73 Even at high sulfidities the sulfur dioxide emis-
Iggesund 308000 141 0,46 sions tend to decrease to levels below 10 ppm.
Munksund 215000 531 2,47
Correlating black liquor sulfidity and dry solids to
Obbola 241000 157 0,65 sulfur dioxide emission results in a graph like 11-
Östrand 395000 163 0,41 10. In it we can see three trends. Sulfur emission
Mörrum 390000 220 0,56 is steadily decreasing until at some point it rapidly
falls to a very low figure. For increased sulfidity
Mönsterås 658000 140 0,21
this fall point is at higher solids.
Gävle 654500 215 0,33
Average 340776 295 0,98 If the boiler load decreases the furnace tempera-
ture decreases. The lower the furnace temperature
the lower is the alkali release. It seems that below
2500 kW/m2bottom the furnace starts to cool and
Table 11-11, Typical dust emissions from recovery SO2 emission increases, Figure 11-10.
boilers (IPPC report, 2001).
Location Emission 11.9 HCL
after ESP mg/m3n 10 … 200
Chlorine can be found in ESP dust as NaCl and
kg/ADt 0.1 … 1.8
KCl and in recovery boiler flue gases as HCl. The
enrichment factor of chlorine is 1-4.5 from BL
Figure 11-8, Typical sulfur dioxide emissions at different dry solids levels.
Figure 11-9, Correlation of sulfur dioxide emissions with dry solids.
Emissions 11-11
Figure 11-10, Correlation of sulfur dioxide emissions with Heart Heat Release Rate at sulfidity between
40 to 45 % (Wallén et al., 2004).
Figure 11-11, Examples of measured HCl concentrations in recovery boiler flue gas.

to ESP dust. Released chlorine in the recovery Dioxins
furnace reacts to form HCl. If there is residual
free sodium after sulfur species have reacted then Organic compounds can combine with available
HCl will react to NaCl. NCASI has found HCl to chlorine to form dioxins. No single component
be about one third of SO2 emission. In modern can be found. There are hundreds of such ele-
Scandinavian boilers the level of HCl was lower ments but only a handful have been analyzed as
for the same sulfur dioxide emission. The probable their content in typical recovery boiler furnace
cause is better mixing. stack is typically very low. The various dioxins can
be converted to TCDD equivalent based on their
concentrations in recovery boiler flue gas. toxicity. The measured dioxin emissions from
Sonnenberd and Nichols (1995) studied the effect recovery boilers are about two magnitudes lower
of adding bleach plant effluents to the black liquor. than measured emissions from the same mills
They found that only about 5 % of chlorine in power boilers (Luthe et al., 1997).
bleach plant effluent was released as HCl. Most
organic chlorine is trapped as NaCl rather than NACASI measured dioxin emissions from six
emitted as HCl during pyrolysis of bleach plant recovery boilers. The levels varied from 1.2–2.5
solids that contain a molar excess of sodium over pg TEQ/m3 (NACASI, 1996). This corresponds to
chlorine. 0.38 to 5.2 pg TEQ /m3, levels measured by Luthe
et al (1997). The typical emission level for dioxin
Rules of thumb emissions is in the order of <100 pg/m3n at 3 %
1 wt-% of Cl in black liquor is 3.0 mole-% O2 or <10 pg(as TCDDequivalent)/m3n at 3 % O2.
1 wt-% of Cl in ESP dust is 2.0 mole-% Note! 1 000 000 000 pg = 1 mg
HCl emission is less or about 1/3 of SO2 emission.
Main components in one measurement were
Calculation of chlorine emission per unit mass of 2,3,7,8-TCDD, 1,2,3,7,8,9-HxCDD and 1,2,3,7,8-
black liquor fired PeCDD.
10 ppm HCl => 4.5 m3n*10*10-6*1.63 kg/ Sonnenberd and Nichols (1995) studied the effect
m3n*35.5/36.5 = 71 mg of adding bleach plant effluents to the black liquor.
250 mgdust/ m3n and 0.5 m-% Cl in BL => 4.5 They found that adding 1 mass-% bleach plant
m3n*250*0.005*3*4.5/2 = 38 mg effluent to black liquor increased dioxin levels
Total emission = (71+38)= 110 mg/kgblds from black liquor droplet combustion by about a
decade.
11.10 MISCELLANEOUS Furans

MINOR EMISSIONS There are very little reliable measurements on
furan contents of recovery boiler emissions. A
One can in addition to major emissions measure typical value for furan emissions is in the order
some very little streams that are of environmental of 1000 pg/m3n at 3 % O2 or 100 pg(as TCD-
interest. Dequivalent)/m3n at 3 % O2. Note! 1 000 000 000
pg = 1 mg.
Sulfuric acid
Main components in one measurement were
Sulfuric acid, H2SO4 emissions correlate with SO2 2,3,7,8-TCDF, 1,2,3,6,7,8-HxCDD and 1,2,3,7,8-
emissions. Some of the sulfur dioxide in flue gases PeCDF.
reacts with molecular oxygen in the recovery
furnace and forms SO3. Typically the amount of Fluorine
SO3 is much less than 5 %. During sampling or at
surface the SO3 reacts with water vapor to form Fluorine in black liquor can be found as trace
H2SO4. metal. The amounts are typically much smaller
than 5 ppm. Most of the fluorine in black liquor
In some measurements the sulfuric acid content is expected to be released in the recovery furnace.
has been found to be tens of ppms if the SO2 emis- Because fluorine levels in black liquor are very
sion is very high. In modern boilers the H2SO4 low the fluorine emission from recovery boiler is
emission is expected to be practically zero. expected to be negligible.
Emissions 11-13
11.11 HEAVY METALS Protection Agency has listed eleven metals as
hazardous air pollutants. These are antimony,
Heavy metal emissions are harmful as they may arsenic, beryllium, cadmium, chromium, cobolt,
collect to human body. Especially inhaling small lead, manganese, mercury, nickel and selenium.
particles can increase heavy metal levels in blood. The European Community has signed a protocol
Heavy metal levels in recovery boiler dusts are to limit the total annual emissions into the atmos-
very low. Most of the matter in dust is harmless phere of cadmium, lead and mercury. Directive
sodium sulfide, sodium carbonate and sodium 2000/76/EC list twelve heavy metals that require
chloride. limits; antimony, arsenic, cadmium, chromium,
cobalt, copper, lead, manganese, mercury, nickel,
Heavy metals include antimony (Sb), arsenic thallium and vanadium. The most problematic
(As), barium (Ba), beryllium (Be), cadmium ones listed are cadmium, thallium and mercury.
(Cd), chromium (Cr), cobalt (Co), copper (Cu), Expected levels from recovery boilers are listed in
gold (Au), iron (Fe), lead (Pb), manganese (Mn), Table 11-2.
mercury (Hg), niobium (Nb), nickel (Ni), sele-
nium (Se), silver (Ag), tellurium (Te), thallium The actual level in black liquor varies significantly
(Th), zink (Zn), vanadium (V) and other even from mill to mill (Control, 1999). The main source
more minor elements. They can be found as trace of these trace elements is wood chips. So the in-
elements in black liquor. The US Environmental flow is mainly controlled by the soil conditions in
Table 11-12, Expected heavy metal emissions.

Metal Concentration in BL Expected tot.emission Expected in gas Expected in dust
mg/kgds mg/Adtp μg/ m3n μg/ m3n
Al 10 - 350 158 3-8 1 - 20

Antimony 1 – 20
As 1 - 10 <21 <3 <0,4 - 2
0,1 – 200
Au 0,2 – 15 <1
Ba 5 - 30 2–4 8 - 12
Be 0 - 10 0,2 – 3 <1
Cb <15
Cd 0 - 10 0,92 0,04 - 0,13 0,1
0,4 – 30
Cr 0.1 - 20 8,7 0,13 - 0,7 1,5
0,5 – 25
Cu 1 - 10 3 – 100 <1
F 1 - 10 <100
Fe 10 - 100 145 3 - 11 2 - 30
Hg 3 0,4 - 0,5 <0,02
1 – 30
Mg 50 - 1000 66 2,4 - 9,7 3 - 40
Mn 20 - 150 22 0,4 - 1,6 1 - 20
3 – 40
Ni 1 - 10 <21 <1,3 <2
1 – 35
Pb 1 - 10 8,8 1 - 1,3 0,1 - 0,3
5 – 70
Se 0 - 10 <6,3 <0,7 <0,3
1 – 200
Tl <21 <1,7 <1,5
<15
Zn 10 - 40 15 – 40 4-8
360

the area where the wood is brought in. Significant flow is about 0,3 - 0,2 m3n(dry)/kgds or 0,5 - 0,3
increase of levels can be expected if bleach plant m3n(wet)/kgds (lower values for higher capaci-
effluent is recycled to brown stock washing. main ties). Most of the dissolving tank flow is thus water
exits are with pulp and dregs. Highest concen- vapor. Increasing air infiltration decreases specific
trations are with Mg 50 ... 1000 mg/kgblds, Mn emissions (mg/m3n) but does not affect the emis-
20 ... 150 mg/kgblds and Ba 5 ... 30 mg/kgblds. sions as mass per fired solids (kg/tds).
Lower concentrations are Zn 10 ... 40 mg/kgblds,
Cu 0,1 ... 10 mg/kgblds, Cr 0 ... 20 mg/kgblds, Ti The typical output level for sulfur dioxide is 0.0050
0 ... 4 mg/kgblds, V, 0 ... 2 mg/kgblds, Be 0 ... 8 … 0.0075 kg/tds or 8 … 12 mg/m3n, dry. The
mg/kgblds. Total concentration is 1000 ... 2000 typical output level for TRS is 0.0050 … 0.0075
mg/kgblds. kg/tds or 6 … 10 mg/m3n, dry. TRS emission can
increase if mix tank vent gases are introduced to
The emission is mostly as dust in fluegases after dissolving tank vent gases, the scrubbing media
ESP. The heavy metals can be divided into three contains volatile sulphur compounds (e.g. is not
groups based on their volatility (Iisa, 2004). As, condensate) or dissolving tank receives conden-
Cd and Pb are highly volatile and they vaporize sates from NCG handling or evaporator.
completely or nearly completely (Backman et al.,
2004). Se, Sb and Be are of medium volatility. Mn, The typical output level for particulate with scrub-
Ni, Cr and Co are of low volatility. bing as only control is 0.035 … 0.050 kg/tds or 100
… 150 mg/m3n, dry.
11.12 DISSOLVING TANK

EMISSIONS
Dissolving tank emissions are caused by non
reacted material in the smelt and reaction of water
with sulfur species. Dissolving tank dust emissions
are caused by droplets of green liquor which when
dried contain inorganic salts. Dissolving tank
Emissions 11-15
12 References
A study of kraft recovery furnace total gaseous pp. 643 – 647.
non-methane organic emissions, 1996, NCASI, Aho, Kaisa, 1994, Nitrogen oxides formation
Atmospheric Quality Improvement Technical in recovery boilers. Lic. Tech. thesis, Report 94-7,
Bulletin, No. 105, National Council of the Paper Combustion Chemistry Group, Åbo Akademi,
Industry for Air and Stream Improvement, New Åbo. 118 p.
York, NY.
Aho, Kaisa, Hupa, Mikko and Nikkanen, Sam-
Abdullah, Z., Gorog, J. P., Keiser, J. R., Meyers, uli, 1993, Release of nitrogen compounds during
L. E. and Swindeman, R.W., 1999, Pattern of ther- black liquor pyrolysis. Proceedings of 1993 TAPPI
mal fluctuations in a recovery boiler floor. NACE Engineering Conference, Orlando, September, pp.
International CORROSION ‘99, San Antonio, TX, 377 - 384.
April 25-30, 1999
Aho, Kaisa, Vakkilainen, Esa and Hupa,
Water tube boilers and auxiliary installation Mikko, 1994, Fuel nitrogen release during black
– Part 15: Acceptance Test on Steam Generators, liquor pyrolysis, Part 1: Laboratory measurements
2003, 12952-15:2003 E, CEN European Commit- at different conditions. Tappi Journal, Vol. 77, No.
tee for Standardization, 89 p. 5, May 1994, pp. 121 – 127.
Adams, Terry, 2001, Sodium salt scaling con- Ahtila, Pekka, 1997, Optimization of recovery
trol in black liquor evaporators and concentrators. boiler plant air systems by new design method
Tappi Journal, Vol. 84, No. 6, June 2001, 18 p. and process integration. Acta Polytechnica Scan-
Adams, Terry N., 1994a, Air flow, mixing and dinavica, Me, Mechanical engineering series, 127,
modelling for recovery boilers. 30 Years Recovery Finnish Academy of Technology, 137 p. ISBN
Boiler Co-operation in Finland, International 952-5148-34-3.
conference, Baltic sea, 24 - 26 May, pp. 61 - 78. Akiyama, H., Uchimura, M., Osawa, S., Shino-
Adams, Terry N., 1994b, Black liquor spray hara, M., Hasegawa, T., Iwanaga, T. and Iwahashi,
nozzles for kraft recovery boilers. TAPPI Kraft K., 1988, Recovery boiler with higher pressure,
Recovery Short Course, Orlando, January 9 - 13, temperature and thermal efficiency. MHI technical
pp. 287 - 302. Review, Vol. 25, No. 2, pp. 84 - 90.
Adams, Terry N., 1988, Air jets and mixing in Alava, Paavo, 1955, Kokemuksia JMW-
kraft recovery boilers. Tappi Journal, Vol. 71, No. soodakatilan käytöstä. (Operation experiences of a
1, January 1988, pp. 97 – 100. JMW recovery furnace) in Finnish, Paperi ja Puu,
Adams, Terry N. and Frederick, William J., Vol. 37, No. 5, pp. 209 – 215.
1988, Kraft recovery boiler physical and chemical Albrecht, Melvin, 2002, Enhancing the
processes. American Paper Institute, Inc., New circulation analysis of a recovery boiler through
York. 256 p. the incorporation of 3-D furnace heat transfer
Adams, Terry N., Frederick, Wm. James, results from COMO. Proceedings 2002 TAPPI
Grace, Thomas M., Hupa, Mikko, Iisa, Kristiina, Fall Technical Conference, September 8 - 11, San
Jones, Andrew K. and Tran, Honghi, 1997, Kraft Diego, Ca. 7 p.
recovery boilers. AF&PA, Tappi Press, Atlanta, Alén, Raimo, 2004, Combustion behavior of
381 p. ISBN 0-9625985-9-3. black liquors from different delignification condi-
Adams, Terry N., Stewart, Robert I. and tions. 40th Anniversary International Recovery
Jones, Andrew K., 1993, Using CFD calculations Boiler Conference, Finnish Recovery Boiler Com-
to estimate thermal-NOx from recovery boilers mittee, Haikko Manor, Porvoo, May 12-14, 2004,
at 67% and 80% dry solids. Proceedings of 1993 pp. 31 – 42.
Tappi Engineering Conference, Orlando, Florida, Alén, Raimo, 2000, Basic chemistry of wood
September 20 – 23, Tappi Press, pp. 625 – 634. delignification. in Forest Products Chemistry,
Adomeit, G., Mohiuddin, G. and Peters N., editor Stenius, P., Fapet, Helsinki, Finland, pp.
1976, Boundary layer combustion of carbon. Six- 58-102.
teenth Symposium on Combustion. The Combus- Alén, Raimo, 1997, Mustalipeän koost-
tion Institute, Pittsburgh, pp. 731 - 743. umuksen vaikutus lipeän poltto-ominaisuuksiin.
Ahlers, Per E., 1983, Investigation of alloyed (Effect of black liquor composition to combustion
steels for use in black liquor recovery boilers. properties) in Finnish, LIEKKI2- yearly report,
Swedish Corrosion Institute, Stockholm. Proceed- part II, Åbo Akademi, Department of Chemical
ings 1993 Environmental Conference, Tappi press, Engineering, pp. 751 – 759.
References 12-1
Anderson, Peter H. and Jackson, James C., Juha, 1995, Flue gas chemistry in recovery boilers
1991, An analysis of best available control technol- with high levels of chlorine and potassium. Pro-
ogy options for kraft recovery furnace NOx emis- ceedings 1995 International Chemical Recovery
sions. Tappi Journal, Vol. 74, No. 1, January 1991, Conference, Toronto, CPPA, pp. A95 – A103.
pp. 115 - 118. Backman, R., Frederick, W. J. and Hupa, M.,
Arakawa, Yoshihisa, Taguchi, Yuzo, Maeda, 1993, Basic studies on black-liquor pyrolysis and
Takayuki and Baba, Yoshitaka, 2004, Experience char gasification. Bioresource Technology, Vol. 46,
of high pressure and high temperature recovery No. 1-2, pp. 153 - 158.
boilers for two decades. Proceedings of 2004 Backman, Rainer, Hupa, Mikko and Uppstu,
International Chemical Recovery Conference, Erik, 1987, Fouling and corrosion mechanisms in
June 6-10 2004, Charleston, South Carolina, USA, the recovery boiler superheater area. Tappi Jour-
pp. 35 - 43. nal, Vol. 70, No. 6, June 1987, pp. 85 - 89.
Arakawa, Yoshisa, Ichinose, Toshimitsu, Backman, Rainer, Hupa, Mikko and Uusikar-
Okamoto, Akiyasu, Baba, Yoshitaka and Sakai, tano Timo, 1985, Kinetics of sulphation of sodium
Toshiyuki, 2003, Application for in-furnace NOx carbonate in flue gases. Proceedings of 1985 In-
removal system for recovery boilers. Pulp & Paper ternational Chemical Recovery Conference, New
Canada, Vol. 104, No. 2, pp. T42 – T46. Orleans, Louisiana, Book 3, pp. 445 - 450.
Backman, Rainer V., Wikstedt, Henrik, Skrif- Backman, Rainer, Hupa, Mikko and Hyöty,
vars, Bengt-Johan, Hupa, Mikko M., Ruohola, Paavo, 1984, Corrosion relating to acidic sulfates
Tuomo and Haaga, Kari, 2004, Trace element in kraft and sodium sulfite recovery boilers. Tappi
distribution in and around the recovery boiler. Journal, Vol. 67, No. 12, December 1984, pp. 60
Proceedings of 2004 International Chemical - 64.
Recovery Conference, June 6-10 2004, Charleston,
Barynin, J. A. and Dickinson, J. A., 1985,
South Carolina, USA, 503-517 p.
Considerations for the uprating of recovery boil-
Backman, Rainer, Hupa, Mikko, Kymäläinen, ers. Proceedings of 1985 International Chemical
Marita and Pulliainen, Martti, 2002, Formation Recovery Conference, New Orleans, Louisiana,
of sulfur-rich melt near the floor tubing of kraft Book 1, pp. 49 – 53.
recovery boilers. Tappi Journal, Vol. 85, No. 4,
Bauer, Donald G. and Sharp, W. B. A., 1991,
April 2002, 7 p.
The inspection of recovery boilers to detect factors
Backman R., Skrifvars B.-J., Hupa M., Siisko- that cause critical leaks. Tappi Journal, Vol. 74, No.
nen P. and Mäntyniemi, J. 1996, Fluegas and dust 9, September 1991, pp. 92 - 100.
chemistry in recovery boilers with high levels of
Baxter, Larry, Hatch, Greg, Sinquefield, Scott
chlorine and potassium. Journal of Pulp and Paper
A. and Frederick, Wm. James, 2004, An experi-
Science, Vol. 22, No. 4, pp. J119-J126.
mental study of the mechanisms of fine particle
Backman. Rainer, Hupa, Mikko, Söderhjelm, deposition in kraft recovery boilers. Proceedings
Liva, 1996, Mustalipeän polttotekniset ominaisu- of 2004 International Chemical Recovery Confer-
udet. (Combustion properties of black liquors) in ence, June 6-10 2004, Charleston, South Carolina,
Finnish, Åbo Akademi University, Department of USA, pp. 393 - 412.
Chemical Engineering, Report 96-9.
Baxter, Larry, Lind, Terttaliisa, Rumminger,
Backman, Rainer, Forssén, Mikael, Hupa, Marc and Kauppinen, Esko, 2001, Particle size
Mikko and Uusikartano, Timo, 1996, New black distributions in recovery boilers. Proceedings of
liquor combustion characteristics II. Liekki II, 2001 International Chemical Recovery Confer-
project 511, final report. Finnish Recovery Boiler ence, June 11-14, 2001, Whistler (BC), Canada,
Committee, SPL-20803-007. pp. 65 - 69.
Backman, R., Eriksson, G. and Sundström, K., Berglin, Niklas and Berntsson, Tore, 1998,
1996, The recovery boiler advisor, Combination of CHP in the pulp industry using black liquor
practical experience and advanced thermodynam- gasification: thermodynamic analysis. Applied
ic modeling. 3. colloquium on process simulation, Thermal Engineering, Vol. 18, No. 11, November
Espoo, Finland, 11-14 Jun 1996, pp. 357. 1998, pp. 947 - 961.
Backman, Rainer, Hupa, Mikko, Söderhjelm, Bergman, Jan and Hjalmarsson, Lennart,
Liva, 1995, Liquor-to-liquor differences in com- 1992, Rotafire – a new recovery boiler opera-
bustion and gasification processes: alkali release tion concept. Proceedings of 1992 International
and fume formation. Journal of Aerosol Science, Chemical Recovery Conference, Seattle, Washing-
Vol. 26, September 1995, pp. S681 - S682. ton, June 7-11, pp. 35 - 43.
Backman, Rainer, Skrifars, Bengt-Johan, Bergroth, Nici, 2004, Char bed processes in a
Hupa, Mikko, Siiskonen, Pentti and Mäntyniemi, kraft recovery boiler – A CFD based study. M. Sc.
thesis, Åbo Akademi, Kemisk-Tekniska Fakulteten Hiner, L., 1994, SO2 capture and HCl release at
Processkemiska Centret, Report 04-2, 149 p. ISBN kraft recovery boiler conditions. AIChE Symp. Ser.,
952-12-1331-0. Vol. 90, No. 302, pp. 39 – 45.
Bernath, P., Sinquefield, S. A., Baxter, L. L., Borg, A., Teder, A. and Warnquist, Björn,
Sclippa, G., Rohlfing, C. and Barfield, M., 1996, 1973, Inside a kraft recovery furnace - studies on
In situ analysis of ash deposits from black liquor the origins of sulfur and sodium emission. Tappi
combustion. 26. International symposium on Journal, Vol. 57, No. 1, pp. 126 - 129.
combustion, Naples (Italy), 28 Jul - 2 Aug, 1996, Borg, A., Nilsson, C., Teder, A. and Warnquist,
24 p. B., 1974, Process betingelsernas inverkan på emis-
Bhattacharya, P. K., Parthiban, V. and Kunzru, sionen från sodahusaggregatet. (Effect of process
D., 1986, Indian Engineering Chemical Process conditions to recovery boiler sulfur emissions)
Design and Development, Vol. 25, p. 420. STFI Meddelande, serie B nr. 203 (MA B:40).
Björklund, Henry, Warnquist, Björn and Pet- Bosch, J. C., 1971, Tappi Journal, Vol. 54, p.
tersson, Bertil, 1991, Inside a kraft recovery boiler 1871.
- combustion of high-sulphidity black liquor at Boström, Curt-Åke, 1990. Aktuellt om mas-
high dry solids. Pulp & Paper Canada, Vol. 92, No. saindustrins NOx utsläpp. (News of pulp industry
8. pp. T206 - 208. NOx emissions) in Swedish, Sodahuskonferenssen
Björklund, Henry, 1963, Synpunkter på olika 1990, ÅF-IPK, Stockholm, pp. 103 -108.
väggkonstruktioner för sodahusaggregat. (View- Brandt, Fritz, 1999, Dampferzeurger, kes-
points of different wall constructions for recovery selsysteme energiebilanz stömungstechnik. (Steam
plants) in Swedish, Sodahuskonferens 1963, ÅF- generators, boiler systems, energy balances, fluid
IPK, Stockholm, pp. A1 – A9. mechanics) in German, Vulkan-Verlag, Essen, 283
Björkman, Anders and Warnquist, Björn, p. ISBN 3-8027-3504-8.
1985, Basic processes in kraft liquor reductive Brandt, Fritz, 1985, Wärmeübertragung in
gasification/burning. Proceedings of 1985 Inter- dampferzeugern und wärme-austauschern, (Heat
national Chemical Recovery Conference, New transfer in steam boilers and heat exchangers)
Orleans, Louisiana, Book 1, pp. 13 - 23. in German, Vulkan-Verlag, Essen, 281 p. ISBN
Blackwell, Brian and Hastings, Calvin, 1992, 3-8027-2274-4.
Simple mathematical analysis of air jets in recov- Brown, Craig A., Grace, Thomas P., Lien, Steve
ery and power boilers. Paperi ja Puu (Paper and J. and Clay, David T., 1989, Char bed burning rates
Timber), Vol. 74, No. 3, pp. 216 – 222. - experimental results. Tappi Journal, Vol. 72, No.
Blackwell, Brian and King, Tracy, 1985, Chem- 12, December 1989, pp. 175 - 181.
ical reactions in kraft recovery boilers. Sandwell Brink, Anders, Coda Zabetta, Edgardo, Hupa,
and Company Ltd. 206 p. Mikko and Saviharju, Kari, 2004, NOx chemistry
Blokh, A. G., 1988, Heat transfer in steam in recovery boilers under staging conditions. Pro-
boiler furnaces, Hemisphere Publishing Corpora- ceedings of 2004 International Chemical Recovery
tion, 283 p. ISBN 0-89116-626-2. Conference, June 6-10, 2004, Charleston, South
Blough, J. L., Seitz, W. W. and Girshik, A., Carolina, USA, 491-501 p.
1998, Fireside corrosion testing of candidate Briscoe, B., Nichol, M., Blackwell, Brian,
superheater tube alloys, coatings, and claddings Brewis, K. and MacCallum, C., 1991, Application
- Phase 2 field testing. Technical report, ORNL/ of modern recovery boiler instrumentation.
Sub-93-SM401/02, Oak Ridge National Lab., TN, Bruce, D. and Van der Vooren, T., 2003,
United States, 75 p. Trends in air emission limits for world class mill.
Blue, Jerry D., Heiner, Larry A. and Wessel, Pulp & Paper Canada, Vol. 104, No. 7, July 2003,
Richard A., 2000, Advanced combustion and flow pp. 51 - 54.
modeling: comprehensive model study improves Bruno, Fredrik, 2003, Corrosion as a cause for
recovery boiler operation. Tappi Journal, Vol. 83, recovery boiler damages. Pulp & Paper Canada,
No. 7, July 2000, 13 p. Vol. 104, No. 6, June 2003, pp. T 143 - 151.
Boiler plant handbook, 1968, Sulzer Brothers Bruno, Fredrik, 1999, The significance of
Limited, Switzerland, 163 p. superheater tube corrosion on steam temperature
Boniface, Arthur, 1985, Introduction. in on kraft recovery boilers. Proceedings 8th Inter-
Chemical recovery in alkaline pulping processes, national Symposium on Corrosion in the Pulp &
editor Hough, Gerald, Tappi Press, Atlanta, pp. 1 Paper Industry, Stockholm, Sweden, May 16-19,
– 7. ISBN 0-89852-046-0. Swedish Corrosion Institute, pp. 138 – 149.
Boonsongsup, L., Iisa, K., Frederick, W. J. and Bruno, Fredrik, 1999, On the influence of
References 12-3
chlorides and sulphurous compounds on the cor- 1986, Contract No. AC02-83CE40637, DOE. 227
rosion of superheater tubes in boilers with special p.
consideration on kraft recovery boilers. Technical Clement, Jack L., 1996, Black liquor firing
Report, SVF-664, Värmeforsk Service AB. 36 p. presentation. Presented to: Black Liquor Recovery
Bruno, Frederick and Hägerström, Joachim, Boiler Advisory Committee, Spring 1996 Techni-
1998, Skador på sodapannor i Sverige 97-98. (Inci- cal Session, April 3, 1996, Atlanta, Georgia, U.S.A.
dents at Swedish recovery boilers 97-98) in Swed- Clement, Jack L., 1990, High pressure and
ish, Sodahuskonferens 1998, ÅF-IPK, Stockholm, temperature recovery boilers. Proceedings of Ba-
pp. 33 – 55. bcock & Wilcox, Pulp & Paper Seminar, February
Bunton, Mark A. and Moskal, Thomas E., 12 – 14, 1990, Portland, Oregon, 7 p.
1995, Increasing boiler efficiency through soot- Clement, J. L., Hiner, L. A., Moyer, S. C., Cox,
blower optimization. Proceedings of 1995 TAPPI C. and Harris, R., 1995, Fundamental approach to
Engineering Conference, Dallas, Texas, Book 2, black liquor combustion improves boiler opera-
pp. 706 - 705. tion. Proceedings of 1995 International Chemical
Burton, Dave, 2000, Practical experiences with Recovery Conference, Toronto Canada, CPPA, pp.
ABB recovery boilers. UEF Conference on Behav- A47 – A57.
ior of Inorganic Materials in Recovery Boilers, Bar Clement, Jack L. and Barna, Joan L., 1993, The
Harbor, Maine, 4 - 9 June 2000. United Engineer- effect of black liquor fuel-bound nitrogen on NOx
ing Foundation, 28 p. emissions. Proceedings of 1993 Tappi Environ-
Bäckman, Kaj, Lindberg, Hans and Sjöberg, mental Conference, Tappi Press, pp. 653 – 660.
Hans, 1999, Recovery modernization at Stora Clement, J. L., Coulter, J. H. and Duda, S.,
Enso’s Skutskär kraft mill. Tappi Journal, Vol. 82, 1963, B&W kraft recovery unit performance cal-
No. 12, December 1999, pp. 90 - 98. culations. Tappi Journal, Vol. 46, No. 2, p. 153.
Cameron, John, 1998, Chem. Eng. Comm. Vol. Coelho, P. J., 1999, An engineering model for
59, pp. 243 – 257. the calculation of radiative heat transfer in the
Casale, F. S. and Friz, P. A., 1990, Start up and convection chamber of a utility boiler, Journal of
operation of a high-solids recovery boiler at S.D. the Institute of Energy, Vol. 72, pp. 117 – 126.
Warren Company Westbrook, Maine. Proceedings Colebrook, C. F., 1939, Turbulent flow in pipes
of 1990 Engineering Conference, Tappi Press. pp. with particular reference to the transition region
687 - 688. between the smooth and rough pipe laws. Journal
Chamberlain, R. E. and Cairns, C. E., 1972, of Institution Civil Engineers, Vol. 11, 1938/39, pp.
Analysis of recovery unit operation and control. 133 – 156.
Pulp and Paper Magazine of Canada, Vol. 73, No. Collier, J. G., 1983, Boiling and evaporation, in
1, pp. 97 - 104. Heat exchanger design handbook, HEDH, Part 2,
Chemical recovery in the alkaline pulping Fluid mechanics and heat transfer, VDI, Düssel-
processes, 1985, a project of the Alkaline Pulping dorf, pp. 2.7.1 – 2.7.4-12. ISBN 3-18-419082-X.
Committee of the Pulp Manufacture Division, edi- Collin, Rolf and Vaclavinec, Jiri, 1989, Injec-
tor Gerald Hough, Tappi Press. tion of secondary and tertiary air in the spent
Chemical recovery in the alkaline pulping kraft liquor recovery boiler – model study. The
processes, 1992, Prepared by the Alkaline Pulping midnight sun colloquium on recovery research,
Committee of the Pulp Manufacture Division, edi- June 14 - 16 1989, STFi, pp. 86 - 96.
tors Green, Robert P. and Hough, Gerald, TAPPI Combustion Engineering A reference book on
PRESS. 196 p. fuel burning and steam generation, 1949, editor de
Churchill, S. W., 1972, Friction factor equation Lorenzi, Otto, Combustion Engineering – Super-
spans all fluid regimes. Chemical Engineering, heater, Inc, New York.
November 1972, pp. 91 – 92. Combustion fossil power, 1991, 4th edition,
Clay, David T., 1987, Fundamental studies in editor Singer, Joseph G., Asea Brown Boveri, 977
black liquor combustion. in Proceedings of the p. ISBN 0-9605974-0-9.
Black Liquor Research Program Review Fourth Combustion fossil power, 1981, 3rd edition,
Meeting, editor Sobczynski, Stanley F., July 28- 30, editor Singer, Joseph G., Combustion Engineering
Contract No. DE-AC01-87CE40762, pp. 177 - 225. Inc., 866 p. ISBN 0-9605974.
Clay, David T., Lien, Steve J., Grace, Thomas Compilation of “air toxics” emission data for
M., Macek, Andrej, Semerjian, Hratch G., Amin, boilers, pulp mills, and bleach plants, 1993, NCASI
N. and Charagundla, S. Rao, 1987, Fundamental Technical Bulletin No. 650, National Council of the
studies in black liquor combustion. Report No. 2 Paper Industry for Air and Stream Improvement,
- Phase I for the period October 1984 - November
New York, NY, June 1990, 128 p. generation, combustion, firing, steam boilers)
Consonni, Stefano and Larson, Eric and in German, Springer Verlag, Berlin, 362 p. ISBN
Katofsky, Ryan E., 2003, Integration issues, per- 3-540-13771-8.
formances estimates and environmental benefits Dombrowski, N. and Johns, W. R., 1963,
of black liquor gasification combined cycles for Chem. Eng. Sci., Vol. 18, pp. 203 - 214.
the pulp & paper industry. Colloquium of black Donovan, Joe and Brown, Robert, 2003, BLG:
liquor combustion and gasification, Park City, Implications for the industrial gases industry. Col-
Utah, 13-16 May 2003, 14 p. loquium of black liquor combustion and gasifica-
Costa, A. O. S., Biscaia, Jr. E. C. and Lima, E. tion, Park City, Utah, 13-16 May 2003, 14 p.
L., 2004, Mathematical description of the kraft Duhamel, Melanie, Tran, Honghi and Freder-
recovery boiler furnace. Computers & Chemical ick, W. James Jr., 2004, The sintering tendency of
Engineering, Vol. 28, No. 5, May 2004, pp. 633 recovery boiler precipitator dust. Tappi Journal,
– 641. Vol. 3, No. 10, October 2004, pp. 25 - 29.
Crawford, Robert J. and Jain, Ashok K., 2002, Edling, Gustaf, 1937, Några synpunkter beträf-
Effect of stripper off-gas burning on NOx emis- fande sodahus av olika system. (Some observa-
sions. Tappi Journal, Vol. 85, No. 1, January 2002, tions concerning different types of recovery) in
6 p. Swedish, presented in the Swedish Paper and
Dahlbom, Johan, 2003, Effects of non proc- Pulpengineer’s winter meeting 6th march 1937,
ess elements in the chemical recovery system in Sodahuset 1930 – 1945, Ångpannaföreningen,
of a kraft pulp mill from the incineration in the Stockholm, pp. 27 - 36 p.
recovery boiler of biological sludge. (Effekter av Edling, Gustaf, 1981, Tekniken i Svenska so-
PFG vid indunstning och förbränning av bioslam dahus 1870 – 1935. Sodahuskonferenssen, Svensk
i ett massabruks sodapanna) in Swedish, Report Papperstidning, No. 6, pp. 19 – 26.
Number SVF-798, Project Värmeforsk-S2-226,
Edling, Gustaf, 1983, Sodahuset 1930 – 1945.
42 p.
Ångpannaföreningen, Stockholm, 77 p.
Darmstad, W. J., Wangerin, D. D. and West, P.
Edwards, Louis L., Shiang, Niann T., Damon,
H., 1968, Combustion of black liquor. Chapter 3
Robert and Strothmann, Brian, 1986, Kraft recov-
of Chemical recovery in alkaline pulping process,
ery furnace capacity and efficiency improvement.
editor Whitney, Roy P., TAPPI Monograph series,
Nordic Pulp and Paper Research Journal, No. 2, pp.
No. 32, Mack Printing Company, Easton, Pa., pp.
18 - 23.
59 – 99.
Effenberger, Helmut, 2000, Dampferzeugung.
Deeley, E. and Kirkby, A. H., 1967, The devel-
(Steam boilers) in German, Springer Verlag, Ber-
opment of chemical recovery boiler. Journal of the
lin, 852 p. ISBN 3-540-64175-0.
Institute of Fuel, September 1967, pp. 417 – 424.
Eilersson, Thomas, Leijonberg, Anders and
DeMartini, Nikolai, Forssén, Mikael, Niemelä,
Praszkier, Paul, 1995, Black liquor recovery boiler
Klaus, Samuelsson, Åsa and Hupa, Mikko, 2004,
superheater gas side corrosion. Proceedings of
Release of nitrogen species from the recovery
8th International Symposium on Corrosion in the
processes of three kraft pulp mills. Tappi Journal,
Pulp & Paper Industry, Stockholm, Sweden, May
Vol. 3, No. 10, October 2004, pp. 3 - 8.
16-19, Swedish Corrosion Institute, pp. 158 – 164.
Demirbaş, Ayhan, 2001, Pyrolysis and steam
Empie, Jeff H., Lien, Steven J., Yang, Wen Rui
gasification processes of black liquor. Energy
and Adams, Terry N., 1992, Spraying characteris-
Conversion and Management, Vol. 43, No. 7, pp.
tics of commercial black liquor nozzles. Proceed-
877 - 884.
ings of 1992 International Chemical Recovery
Dickinson, James A., Hiner, Larry A. and Conference, Seattle, Washington, June 7-11, pp.
Wessel, Richard A., 2000, practical solutions 429 - 440.
to improving recovery boiler operation. UEF
Empie, H.J., Lien, S. J., Yang, W. and Samuels,
Conference on Behavior of Inorganic Materials in
D. B., 1995, Effect of black liquor type on droplet
Recovery Boilers, Bar Harbor, Maine, 4 - 9 June
formation from commercial spray nozzles. Journal
2000. United Engineering Foundation, 15 p.
of Pulp and Paper Science, Vol. 21, No. 2, pp. J63
Directive 2000/76/EC of the European Parlia- – J67.
ment and of the Council of 4 December 2000 on
Environmental pollution control pulp and pa-
the incineration of waste. 21 p.
per industry part 1 Air, 1976, U.S. Environmental
Doležal, Richard, 1967, Large boiler furnaces, protection agency, Technology Transfer, Prepared
Elsevier Publishing Company, 394 p. October 1976, by EKONO Inc. and EKONO Oy,
Doležal, Richard, 1985, Dampferzeugung: 260 p.
verbrennung, feuerung, dampferzeuger, (Steam
References 12-5
Environmental report statistics for 2001, 2002, Alan R., 1992, Engineering analysis of recovery
editors Sieppi, Susanna and Heino, Petri, Finnish boiler superheater corrosion. Tappi Journal, Vol.
Forest Industries Federation, Kirjapaino Oy West 75, No. 6, June 1992, pp. 101 - 106.
Point, Rauma, 6.2002, ISSN 1236–1097. Forssén, Mikael, Backman, Rainer, Wal-
Eriksson, Thomas and Falk, Ivan, 1999, Över- lén, Jonas and Hupa, Mikko, 2000a, Flygaskans
hettarmaterial för energi-effectivare, miljövänlig- sammansättning och nedsmutsande tendens i
are och bränsleflexiblare sodahus och barkpannor. sodapannan. (Flue gas dust composition and foul-
(Superheater materials for better energy efficiency, ing tendency in recovery boilers) in Swedish, SVF-
lower emissions and better fuel flexibility in 680, Technical Report, Värmeforsk, Stockholm,
recovery and bark boilers) in Swedish, Studsvik Sweden, Project Värmeforsk-S6-612, S9-804, ISSN
Material AB, S6-615, Värmeforsk Service AB. 47 p. 0282-3772, 78 p.
Ernola, P., Hupa, Mikko, Kjäldman, Lars and Forssén, Mikael, Kilpinen, Pia and Hupa,
Oksanen, P., 1989, Detailed modelling of NOx Mikko, 2000b, NOx reduction in black liquor
emissions in fuel staging. Åbo Akademi, Depart- combustion - reaction mechanisms reveal novel
ment of Chemical Engineering, Combustion operational strategy options. Tappi Journal, Vol.
Chemistry Research Group, Report 89 - 9. 16 p. 83, No. 6, June 2000, 13 p.
Eskola, Arkke, Jokiniemi, Jorma, Vakkilainen, Forssén, Mikael, Hupa, Mikko and Hellström,
Esa and Lehtinen, Kari, 1998, Modelling alkali salt Peter, 1997, Liquor-to-liquor differences in com-
deposition on kraft recovery boiler heat exchang- bustion and gasification processes: Nitrogen oxide
ers in the superheater section. Proceedings of formation tendency. Tappi Journal, Vol. 82, No. 3,
1998 International Chemical Recovery Confer- March 1999, pp. 221 - 227.
ence, Tampa, USA, 1 - 4 June 1998, Vol. 3, TAPPI, Forssén, Mikael, Frederick, W. J., Hupa, Mikko
pp. 469 – 486. and Hyöty, Paavo, 1992, Sulfur release during
Eskola, Arkke, Jokiniemi, Jorma, Pyykönen, pyrolysis from single black liquor droplets. AIChE
Jouni and Lehtinen, Kari, 1997, Modelling alkali- 1991 Forest Products Symposium Proceedings,
vapour/particle deposition on cooled surfaces. Tappi Press, Atlanta, pp. 11 – 22.
Proceedings of 1997 European Aerosol Confer- Frederick, Laurie, Singbeil, Doug and Kish,
ence. Hamburg 15 - 19 Sept. 1997, Journal of Joey, 2003, Red hot video of combusting liquor
Aerosol Science, Vol. 28, No. S1, pp. 453 - 454 & smelt at primary airports. Proceedings of Col-
Fakhrai, Reza, 1999, Modelling of carry-over loquium of black liquor combustion and gasifica-
in recovery furnaces. Lic. Tech. thesis, Dept. tion, Park City, Utah, 13-16 May 2003, 9 p.
of Metallurgy, Royal Inst. of Tech., Stockholm, Frederick, Wm James Jr., Iisa, Kristiina, Sin-
Sweden, 68p. quefield, Scott A. and Sricharoenchaikul, Viboon,
Falat, Lad, 1996, Mechanism and prevention 2004, The role and fate of alkali metals in black
of localized corrosion of recovery boiler tubes at liquor combustion. 40th Anniversary Internation-
air ports. Tappi Journal, Vol. 79, No. 2, February al Recovery Boiler Conference, Finnish Recovery
1996, pp. 175 -185. Boiler Committee, Haikko Manor, Porvoo, May
Ferguson, Kelly H., 1991, L-P Samoa mill im- 12-14, 2004, pp. 43 – 49.
proves air quality with recovery island moderniza- Frederick, Wm. James, Ling, Alisa, Tran,
tion. Pulp & Paper, September, pp. 180 - 185. Honghi N. and Lien, Steven J., 2004, Mechanisms
Ferreira, Licínio Manuel G. A., Soares, Mi- of sintering of alkali metal salt aerosol deposits
caela A. R., Egas, Ana Paula V. and Castro, José in recovery boilers. Fuel, Vol. 18, No. 83, pp. 1659
Almiro A. M., 2003, Remoção seletiva de cloreto – 1664.
e potássio em fábricas de cellulose kraft. (Selective Frederick, Wm. J. Jr. and Vakkilainen, E. K.,
removal of chloride and potassium in kraft pulp 2003, Sintering and structure development in
mills) in Portuguese, O Papel / Tappi Journal, Vol. alkali metal salt deposits formed in kraft recovery
3, No. 1, July 2003, pp. 13 – 20. boilers. Energy & Fuels, Vol. 17, No. 6, November
La Fond, John, F., Jansen, Johan H. and Eide, 2003, pp. 1501-1509.
Perry, 1994, Upgraded recovery boiler meets low Frederick, W. J., Lien, S., Vakkilainen, E. K.
air emissions standards. Tappi Journal, Vol. 77, and Tran, H., 2001, A method to predict the con-
No. 12, December 1994, pp. 75 - 80. ditions for boiler bank plugging by sub-micron
La Fond, John F., Verloop, Arie and Jansen, sodium salt (fume) particles. Proceedings of 2001
Johan H., 1993, Recovery boiler fireside capacity: International Chemical Recovery Conference,
an update of theory and practice. Tappi Journal, June 11-14, 2001, Whistler (BC), Canada, pp. 311
Vol. 76, No. 9, September 1993, pp. 107 – 113. - 322.
La Fond, John F., Verloop, Arie and Walsh, Frederick, W. J. Jr., Dent, Gary and Vakki-

lainen, Esa, 1998, Effects of temperature and SO2 Yoshitaka and Arakawa, Yoshisa, 2003, Plug-
on chloride and potassium enrichment factors ging prevention of recovery boiler by character
in kraft recovery boilers. Proceedings of 1998 improvement of the ash which used potassium re-
International Chemical Recovery Conference June moval equipment (Part II). Pulp & Paper Canada,
1 - 4, Hyatt Westshore, Tampa, Fl, pp. 945 - 954. Vol. 104, No. 1, pp. T1 – T3.
Frederick, W. J., Reis, V. V., Wag, K. J. and Iisa, Fujisaki, Akihiro, Takatsuka, Hiromu and
K., 1995, A laboratory study of alkali metal salt Yamamura, Misao, 1992, World’s largest high pres-
aerosol formation in black liquor combustion. sure and temperature recovery boiler. Proceedings
Journal of Aerosol Science, Vol. 26, September of 1992 International Chemical Recovery Confer-
1995, pp. S693 - S694. ence, Seattle, Washington, June 7-11, pp. 1 – 20.
Frederick, W. J., Iisa, K., Wag, K., Reis, V. V., Given, P. H. et al., 1986, Calculation of
Boonsongsup, L., Forssen, M. and Hupa, M., 1995, calorific values of coals from ultimate analysis:
Sodium and sulfur release and recapture during Theoretical basis and geographical implications.
black liquor burning. DOE/CE/40936-T2, Oregon Fuel, June 1986, Vol. 65, No. 6, pp. 849 - 854.
State University, Corvallis, OR, United States, 225 Goerg-Wood, Kristin and Cameron, John
p. H., 1998, Fume deposition with kraft recovery
Frederick, William J. and Hupa, Mikko, 1993, boiler implications. Innovative advances in the
Combustion properties of kraft black liquors. Åbo forest products industry, the 1997 Forest Products
Akademi, Department of Chemical Engineering, Symposium, editor Brogdon, Brian N., New York,
Report 93-3. 112 p. American Institute of Chemical Engineers, AIChE
Frederick, William J. and Hupa, Mikko, 1992, symposium series 319, Vol. 94, pp.138 – 145.
The effect of swelling on droplet trajectories, Grabke, H. J., Reese, E. and Spigel, M., 1995,
carbon burn out, and entrainment in black liquor The effects of chlorides, hydrogen chloride, and
combustion. Proceedings of 1991 Forest Products sulfur dioxide in the oxidation of steels below
Symposium, November 17-22, 1991, Los Angeles, deposits. Corrosion Science, Vol. 37, No. 7, pp.
California, Tappi Press, pp. 79 – 89. 1023 – 1043.
Frederick, W. J., Krishnan, R. and Ayers, R. J., Grace, Thomas M., 2004, A review of char
1990, Pirssonite deposits in green liquor process- bed processes (What goes on in the char bed).
ing. Tappi Journal, Vol. 73, No. 2, pp. 135-140. 40th Anniversary International Recovery Boiler
Frederick, William J., 1990, Combustion Conference, Finnish Recovery Boiler Committee,
processes in black liquor recovery: Analysis and Haikko Manor, Porvoo, May 12-14, 2004, pp. 21
interpretation of combustion rate data and an en- – 29.
gineering design model, Report no. One. Contract Grace, Thomas M., 2001, A review of char bed
no. AC02-83CE40637, U.S. Department of Energy, combustion. Proceedings of 2001 International
DOE/CE/40637-T8. 130 p. Chemical Recovery Conference, June 11-14, 2001,
Frederick, William J, Noopila, Taina and Whistler (BC), Canada, pp. 213 - 220.
Hupa, Mikko, 1989, Modelling of black liquor Grace, Thomas M., 1997, Char bed process.
droplet combustion. Soodakattilapäivä, Helsinki, Chapter 6 in Kraft recovery boilers, AF&PA, TAP-
19 p. PI PRESS, Atlanta, 381 p. ISBN 0-9625985-9-3.
Fricke, A. L., 1987, Physical properties of kraft Grace, Thomas M. and Timmer, William M.,
black liquors: Interim report Phase II, report No. 1995, A comparison of alternative black liquor
DOE/CE 40606-T5 (DE88002991), U.S. Dept. of recovery technologies. Proceedings of 1995 Inter-
Energy, Washington, 1987, 66 p. national Chemical Recovery Conference, Toronto
Friedlander, S. K., 1977, Smoke Dust and Canada, CPPA, pp. B269 - B275.
Haze, John Wiley & Sons, New York. 317 p. ISBN Grace, Thomas M., 1990, Char bed burning.
0-471-01468-0. Kraft Chemical Recovery, University Center, May
Froste, H., 1996, Volatile organic compounds 14, 1990.
and oxides of nitrogen - Further emission reduc- Grace, Thomas M., Cameron, J. H. and Clay,
tions. Technical Report, SNV-4689, Swedish Envi- David T., 1989, Role of the sulfate/sulfide cycle in
ronmental Protection Agency, Stockholm Sweden, char burning - experimental results and implica-
ISSN 0282-7298. 109 p. tions. TAPPI Kraft Recovery Operations Seminar,
Fryling, Glenn R., 1967, Combustion engi- pp. 159 - 167.
neering, A reference book on fuel burning and Grace, Thomas M., Cameron, J. H. and Clay,
steam generation. 2nd edition, The Riverside Press, David T., 1988, Role of the sulfate/sulfide cycle in
Cambridge, Massachusetts, 831 p. char burning - experimental results and implica-
Fujisaki, Akihiro, Tateishi, Masakazu, Baba, tions. TAPPI Kraft Recovery Operations Seminar,
References 12-7
pp. 157 - 165. Heinävaara, Antero, 1991, Recovery boiler
Grace, Thomas M., Cameron, J. H. and Clay, analysis. Ahlstrom Machinery, Internal report.
David T., 1985, Role of the sulfate/sulfide cycle in Hellström, Bengt, 1982, En-doms sodapanna.
char burning - experimental results and implica- (Single-drum recovery boiler) in Swedish, Soda-
tions. Proceedings of 1985 International Chemical huskonferensen ‘82, ÅF-IPK, Stockholm, pp. 133
Recovery Conference, New Orleans, Louisiana, – 135.
Book 3, pp. 371 - 379. Hellström, Bengt, 1977, Styr sodapannans
Green, Robert P. and Grace, Thomas M., lufttillförsel. Kemisk Tidskrift, No. 9, pp. 112 - 114.
1984, A method for calculating the composition Hellström, O., 1970, Rapport från verksam-
and heating value of black liquors from kraft and heten inom den amerikanska kommittén för
polysulfide pulping. Tappi Journal, June 1984, pp. sodahusfrågor (BLRBAC) under 1970. (Report
94 - 98. of action in the american committee for recovery
Greenwood, Larry D., Ip, Trevor, Dhak, Janice, boiler questions (BLRBAC) in 1970) in Swedish,
Kennard, Gary E., Towers, Michael T., Karidio, Sodahuskonferens 1970, ÅF-IPK, Stockholm, pp.
Ibrahim, Uloth, Vic C. and Weiss, Klaus, 2000, Re- A1 – A15.
verse superheater fouling in the Al-Pac recovery Helpiö, Tomi and Kankkunen, Ari, 1995, The
boiler. Proceedings of 2000 Tappi Engineering effect of black liquor firing temperature on atomi-
Conference, TAPPI Press, Atlanta, 19. zation performance. Proceedings of 1995 TAPPI
Grönblad, Kjell and Sandegård, Ragnar, Engineering Conference, Dallas, 11 - 15 Septem-
2000, Ombyggnad av sodapanna SP-7 i Skutskär. ber, pp. 717 - 722.
(Rebuild of recovery boiler SP-7 in Skutskär) in Helpiö, T. and Kankkunen, A., 1995b, Meas-
Swedish, Sodahuskonferens, ÅF-IPK, Stockholm, urements of the Effect of Black Liquor Firing
pp. 53 – 62. Temperature on Atomization Performance, Part
Gullichsen, Johan, 1968, Heat values of pulp- 2: Drop Size Distributions, Helsinki University of
ing spent liquors. Proceedings of the Symposium Technology, Laboratory of Energy Engineering
on Recovery of Pulping Chemicals, Helsinki, The and Environmental Protection. 34 p.
Finnish Pulp and Paper Research Institute and Helpiö, T. and Kankkunen, A., 1994, Measure-
Ekono, pp. 211 - 234. ments of the Effect of Black Liquor Firing Temper-
Görner, Von K., 1986, Strömungsvorgänge in ature on Atomization Performance, Part 1: Nozzle
feuerräumen von dampferzeugern. (Flow predic- Flow Properties and Sheet Breakup. Helsinki
tion in steam boiler furnaces) in German, VGB University of Technology, Laboratory of Energy
Krafrtwerkstechnik, Vol. 66, No. 3, March 1986, Engineering and Environmental Protection.
pp. 224 – 233. Henderson, Pamela, Kjörk, Anders, Ljung, Per,
Harila, Pauli, 1993, Biolietteen käsittely, mus- Nyström, Olle and Skog, Erik, 2000, Överhet-
talipeän superväkevöinti ja hajukaasujen poltto tarkorrosion i bioeldade anläggningar – status.
soodakattilassa. (Biosludge treatment, super-con- (Superheater corrosion in combustion of biofuels
centration of black liquor and NCG burning in - a status) in Swedish, Technical report, SVF-700,
recovery boiler) in Finnish, Report 19.10.1993, Oy Värmeforsk Service AB. 36 p.
Metsä-Botnia Ab, Kemi mill, 10 p. Himmelblau, David M., 1989, Basic princi-
Harms, T. E., Haynes, Jim B. and Edwards, ples and calculations in chemical engineering. 5th
Louis L., 1990, Evaluation of process alternatives edition, Prentice-Hall International Series in the
to improve recovery boiler performance. Pulp & Physical and Chemical Engineering Sciences.
Paper Canada, Vol. 91, No. 7, pp. T267 - T270 Hills, Richard L., 1989, Power from steam.
Harrison, Ray E. and Ariessohn, Peter C., Cambridge University press, 338 p. ISBN 0 521
1985, Application of a smelt bed imaging system. 34356 9.
Proceedings of 1985 TAPPI International Chemi- Hiner, L.A., 1994, Kraft Recovery Boiler Op-
cal Recovery Conference, pp. 153 - 158. eration for Control of Ash Chemistry. Proceedings
Hausen, H., 1974, Erweiterte gleichung für of 1994 TAPPI Engineering Conference, Press,
den wärmeübergang bei turbulenter strömung. Atlanta, GA, p.179 - 186.
(Expanded equation for heat transfer in turbulent Hochmuth, Frank W., 1953, New develop-
flow) in German, Wärme- und Stoffübertragung, ments in recovery unit design. Tappi Journal, Vol.
Vol. 7, pp. 222 – 225. 36, No. 8, p. 359.
Haynes, Jim B., Adams, Terry N. and Edwards, Hoddenbagh, J. M. A., Wilfing, K., Miller, K.,
Lou L., 1988, Recovery boiler thermal perform- Hardman, D., Tran, H. and Bair, C., 2002, Borate
ance. Proceedings of 1988 Tappi Engineering autocausticizing: a cost effective technology. Pulp
Conference, Book 2, pp. 355 - 367. & Paper Canada, Vol. 103, No. 11, November
2002, T283 – T288. Porvoo, May 12-14, 2004, pp. 125 – 132.
Hogan, Edward F., 1999, Investigation of Hupa, Mikko, Backman, Rainer and Fred-
chemical recovery unit floor tube overheat fail- erick, William J., 1994, Black liquor combustion
ures. Tappi Journal, Vol. 82, No. 2, February 1999, properties. 30 Years Recovery Boiler Co-operation
pp. 130 - 137. in Finland, International conference, Baltic sea, 24
Holcomb, Gordon R., 2001, Synergistic air - 26 May, pp. 37 - 60.
port corrosion in kraft recovery boilers. Tappi Hupa, Mikko, 1993, Recovery boiler chemis-
Journal, Vol. 84, No. 8, August 2001, pp. 35 - 37. try – the picture becomes sharper. Paperi ja Puu,
Holmlund, Karl and Parviainen, Kari, 2000, Vol. 75, No. 5, pp. 310 – 319.
Evaporation of black liquor. Chapter 12 in Chemi- Hupa, Mikko, Backman, Rainer, Skrifvars,
cal Pulping, Book 6, series editors Gullichsen, Bengt-Johan and Hyöty, Paavo, 1990, The influ-
Johan and Fogelholm, Carl-Johan, Finnish Paper ence of chlorides on the fireside behavior in the
Engineers’ Association and TAPPI. ISBN 952- recovery boiler. Tappi Journal, Vol. 73, No. 6, June
5216-06-3. 1990, pp. 153 - 158.
Hood, Kenneth T. and Henningsen, Gunnar Hupa, Mikko, 1989, Recovery boiler chem-
B., 2002, Black liquor gasification: evaluation of istry. Kraft Recovery Operations Seminar, Tappi
past experience in order to define the roadmap of Press, pp. 153 - 158.
the future pathway. Proceedings of 2002 TAPPI Hupa, Mikko, Skrifvars, Bengt-Johan and
Fall Technical Conference, September 8 - 11, San Moilanen, A., 1989, Measuring the ash sintering
Diego, Ca, 17 p. tendency by a laboratory method. Journal of the
Hooper, William B., 1984, The two-K method Institute of Energy, September, pp. 131 – 137.
predicts head losses in pipes and fittings. Chemical Hupa, Mikko and Solin, P., 1985, Combustion
Engineering, August 24, pp. 96 – 100. behavior of black liquor droplets. Proceedings of
Horton, Robert R. and Vakkilainen, Esa K., 1995 International Chemical Recovery Confer-
1993, Comparison of simulation results and field ence, New Orleans, Louisiana. Book 3, pp. 335
measurements of an operating recovery boiler. - 344.
Proceedings of 1993 TAPPI Engineering Confer- Hyöty, Paavo, 1994, The history of recovery
ence, Orlando, September, pp. 20 - 23. boiler. 30 Years Recovery Boiler Co-operation in
Horton, Robert R., Grace, Thomas M. and Finland, International conference, Baltic sea, 24
Adams, Terry N., 1992, The effect of black liquor - 26 May, pp. 9 - 22.
spray parameters on combustion behaviour in Hyöty, Paavo A. and Ojala, Sakari T., 1987,
recovery furnace simulations. Proceedings of 1992 Super combustion of black liquor. Proceedings of
International Chemical Recovery Conference, 1987 Tappi Engineering Conference, pp. 49 – 53.
Seattle, Washington, June 7-11, pp. 85 - 99.
Hänninen, Hannu, Tiitinen, Eero, Tanaka,
Horton, Robert R., 1991, In-flight black liquor Toshi and Kiesi, Timo, 2003, Soodakattila tul-
combustion modelling. in Black liquor chemical evaisuudessa – Japanin matkakertomus. (Recovery
recovery research, Program review meeting, Wey- boilers of future – Trip to Japan) in Finnish, Finn-
erhaeuser Technology Center, Tacoma, Washing- ish recovery boiler committee, Lipeätyöryhmä,
ton, April 8-9, 15 p. 84 p.
Horton, Robert R., 1991b, Computer simula- Hänninen, Hannu, 1994, Cracking and corro-
tion of black liquor combustion. DOE Annual Re- sion problems in black liquor recovery boilers. 30
port on Recovery Furnace Modeling, September. Years Recovery Boiler Co-operation in Finland.
Huhtinen, Markku, Kettunen, Arto, Nurmin- International conference, Baltic sea, 24 - 26 May,
en, Pasi and Pakkanen, Heikki, 1999, höyrykatti- pp. 121 - 132.
latekniikka. (Steam boiler technology) in Finnish, Ibah, Sheldon, 1995, Conversion to high solids
EDITA, Helsinki, 316 p. ISBN 951-37-1327-X. 1. firing. Proceedings of 1995 International Chemi-
Hultin, Sven O., 1968, Physical properties cal Recovery Conference, Toronto Canada, CPPA,
of Finnish sulphite liquors and black liquors. pp. A173 – A180.
Proceedings of the Symposium on Recovery of Iisa, Kristiina, 2004, A comparison of the
Pulping Chemicals, Helsinki, The Finnish Pulp sulfation rates of different fume species in recov-
and Paper Research Institute and Ekono, pp. 167 ery boilers. Proceedings of 2004 International
- 182. Chemical Recovery Conference, June 6-10 2004,
Hupa, Mikko, 2004, Research highlights in Charleston, South Carolina, USA, 1043-1057 p.
recovery boiler chemistry. 40th Anniversary Iisa, Kristiina, 2004, Emissions of trace ele-
International Recovery Boiler Conference, Finn- ments from recovery boilers. 40th Anniversary
ish Recovery Boiler Committee, Haikko Manor,
References 12-9
International Recovery Boiler Conference, Finnish Skrifars, Bengt-Johan and Hupa, Mikko, 1998,
Recovery Boiler Committee, Haikko Manor, Por- The roles of black liquor composition and boiler
voo, May 12-14, 2004, pp. 115 – 124. combustion parameters on the fouling tendency of
Iisa, Kristiina, Horenziak, Steven A. and Jing, recovery boiler. Proceedings of 1998 International
Qun, 2000, Release of sulfur during gasification of Chemical Recovery Conference, Tampa, USA, 1
black liquor in CO2 atmospheres. UEF Conference - 4 June 1998, TAPPI Press, Atlanta, pp. 629 – 640.
on Behavior of Inorganic Materials in Recov- Janka, Kauko, Ruohola, Tuomo, Siiskonen,
ery Boilers, Bar Harbor, Maine, 4 - 9 June 2000, Pekka and Tamminen, Ari, 1998, A comparison
United Engineering Foundation, 10 p. of recovery boiler field experiments using various
Iisa, Kristiina and Jing, Qun, 2000, Effect of NOx reduction methods. Tappi Journal, Vol. 81,
liquor composition on potassium and chloride No. 12, December 1998, pp. 137 - 141.
enrichment. UEF Conference on Behavior of Jing, Qun and Iisa, Kristiina, 2001, Black
Inorganic Materials in Recovery Boilers, Bar Har- liquor devolatilization kinetics. Proceedings of
bor, Maine, 4 - 9 June 2000, United Engineering 2001 International Chemical Recovery Confer-
Foundation, 10 p. ence, June 11-14, 2001, Whistler (BC), Canada,
Im, K. H. and Chung, P. M., 1993, Particu- pp. 221 - 229.
late Deposition from Turbulent Parallel Streams. Jokiniemi, Jorma K., Pyykönen, Jouni,
AIChE J., Vol. 29, No. 3, p. 498. Mikkanen, Pirita and Kauppinen, Esko I., 1996,
Incropera, Frank P. and DeWitt, David P., Modeling fume formation and deposition in kraft
1996, Fundamentals of Heat and Mass Transfer. 4th recovery boilers. Tappi Journal, Vol. 79, No. 7, July
edition. John Wiley & Sons. ISBN 0-471-30460-3. 1996, pp. 171 - 181.
Integrated pollution prevention and control Jokiniemi, Jorma, Kauppinen, Esko and
(IPPC), reference document on best available Mikkanen, Pirita, 1993, Rikki ja natrium sitoutu-
techniques in the pulp and paper industry, 2001, vat soodakattilassa aerosoliin. (Sulfur and sodium
European Commission, European Integrated Pol- are caught to aereosols in a recovery boiler) in
lution Prevention and Control Bureau, December Finnish, Paperi ja Puu (Paper and Timber), Vol.
2001, 475 p. http://eippcb.jrc.es/pages/FActivities. 75, No. 3, pp. 122 – 125.
htm. Jones, Andrew K., 2004, 70 years of advances
Ishigai, Seikan, 1999, Historical develop- in recovery boiler design. 40th Anniversary
ment of strategy for steam power. in Steam Power International Recovery Boiler Conference, Finn-
Engineering, editor Ishigai, Seikan, Cambridge ish Recovery Boiler Committee, Haikko Manor,
University Press, 394 p. ISBN 0 521 62635 8. Porvoo, May 12-14, 2004, pp. 15 – 20.
Israel, R. and Rosner, D.E., 1983, Use of Jones, Andrew K. and Chapman, Paul J., 1992,
generalized Stokes number to determine the CFD combustion modeling – a comparison of
aerodynamic capture efficiency of non-Stokesian secondary air system designs. Proceedings of 1992
particles from a compressible gas flow. Aerosol Sci. Tappi Engineering Conference, Boston, MA, Sep-
Technol. (AAAR), Vol. 2, pp. 45 – 51. tember 14 – 17, Tappi Press, pp. 559 – 579.
Jameel, M. I., Cormack, Donald E., Tran, Jones, Andrew K., 1989, A model of kraft
Honghi and Moskal, Thomas E., 1994, Sootblower recovery furnace. Doctoral Dissertation, Apple-
optimization. Tappi Journal, Vol. 77, No. 5, pp. 135 ton, Wisconsin, The Institute of Paper Chemistry.
– 142. 157 p.
Jameel, M. I., Schwade, H. and Easterwood, Jones, Andrew K. and Stewart, R. I., 1993, The
M. W., 1995, A field study on the operational im- high solids breakpoint: a trade off between SO2
pact of improved sootblower nozzles on recovery and NOx. Pulp & Paper Canada, Vol. 94, No. 12,
boilers. Proceedings of 1995 TAPPI Engineering pp. T479 – T482.
Conference, Dallas Texas, Book 2, pp. 695 - 705. Jones, Andrew K. and Anderson, Michael J.,
Janka, Kauko, Wallén, Jonas and Backman, 1992, High solids firing at Arkansas Kraft. 78th
Rainer, 2004, Prediction of dust content and Annual Meeting, Technical section, CPPA, pp.
properties in kraft recovery boilers: comparison A39 - A49.
of theory and experimental results. Pulp & Paper Jutila, E., Pantsar, Ossi and Uronen, Paavo,
Canada, Vol. 105, No. 1, January 2004, pp. 46 – 50. 1978, Computer control of a recovery boiler. Pulp
Janka, Kauko and Tamminen, Ari, 2003, & Paper Canada, Vol. 79, No. 4, pp. 61 - 65.
Recovery boiler furnace as concentrated NCG Järvinen, Mika, 2002, Numerical modeling of
incinerator. Tappi Journal, Vol. 2, No. 2, February the drying, devolatilization and char conversion
2003, 9 p. processes of black liquor droplets. Ph. D. thesis,
Janka, Kauko, Heinola, Ari, Heinola, Marja, Acta Polytechnica Scandinavia, Mechanical Engi-
neering Series, 94 p. Kays, W. M., 1950, Transactions of ASME, Vol.
Järvinen, Mika, Zevenhoven, Ron and Vak- 72, pp. 1067 – 1074.
kilainen, Esa, 2002, Implementation of a detailed Kearns, Jim, Levesque, Denise and Navitsky,
black liquor combustion model for furnace Gary, 2001, Reducing CRU downtime – a circula-
calculations, IFRF Electronic Combustion Journal, tion and water chemistry perspective. Proceedings
Article Number 200206, June 2002, 34 p of 2001 TAPPI Engineering/Finishing & Convert-
Järvinen, M., Forssén, M., Vakkilainen, E., ing Conference, San Antonio, TX. 21 p.
Zevenhoven, R. and Hupa, M., 2000, Black liquor Keiser, James R., Singbeil, Douglas L., Sarma,
devolatilization and swelling detailed droplet Gorti B., Kish, Joseph R., Choudhury, Kimberly
model. UEF Conference on Behavior of Inorganic A., Hubbard, Camden R., Frederick, Laurie A.,
Materials in Recovery Boilers, Bar Harbor, Maine, Yuan, Jerry W. and Singh, Preet M., 2004, Causes
4 - 9 June 2000. United Engineering Foundation, and solutions for recovery boiler primary air port
14 p. composite tube cracking. Proceedings of 2004
Järvinen, Mika and Kankkunen Ari, 2000, International Chemical Recovery Conference,
Ideal solution approximation for vapor pressure June 6-10 2004, Charleston, South Carolina, USA,
of water in black liquor using Raoult’s law. To be 527-542 p.
published. Keiser, James R., Singbeil, Douglas L., Sarma,
Järvinen, Mika, Vakkilainen, Esa and Gorti B., Kish, Joseph R., Choudhury, Kimberly
Kankkunen Ari, 1997, Effects of liquor gun type A., Frederick, Laurie A., Gorog, Peter, Jetté,
on black liquor combustion. Proceedings of 1997 François, R., Hubbard, Camden R., Swindeman,
Tappi Engineering Conference, Nashville, Tennes- Robert W., Singh, Preet M., Yuan, Jerry and
see, pp. 1349 - 1356. Maziasz, Philip J, 2004, Cracking and corrosion of
composite tubes in black liquor recovery boilers.
Jönsson, S-E, 1961, Synpunkter på indun-
40th Anniversary International Recovery Boiler
stningsanläggningen och dess ekonomi vid
Conference, Finnish Recovery Boiler Committee,
indunstning till högre torrhalt. (Viewpoints of
Haikko Manor, Porvoo, May 12-14, 2004, pp. 59
evaporators and their economy with high dry
– 89.
solids evaporation) in Swedish, Sodahuskonferens,
ÅF-IPK, Stockholm, pp. 54 – 59. Keiser, James, Singbeil, Douglas L., Sarma,
Gorti B., Choudhury, Kimberly, Singh, Preet M.,
Kaila, Jarmo and Saviharju, Kari, 2003, Com-
Hubbard, Camden R., Swindeman, Robert W., Ely,
parison of recovery boiler CFD modeling to actual
Thomas, Kish, Joseph, Kenik, Edward, Maziasz,
operations. PAPTAC Annual meeting, January 28,
Philip J. and Bailey, Sarah, 2002, Comparison of
Montreal, Canada, 13 p.
cracking in recovery boiler composite floor and
Kaliazine, Andrei, Cormack, Donald E., primary air port tubes. Tappi Journal, Vol. 1, No.
Eibrahimi-Sabet, Abdolreza and Tran, Honghi, 2, February 2002, pp. 1 - 7.
1998, The mechanics of deposit removal in kraft
Keiser, J. R., Sarma, Gorti B., Wang, Xun-Li,
recovery boilers. Proceedings of International
Hubbard, Camden R., Swindeman, Robert W.,
Chemical Recovery Conference. Tampa, USA, 1
Maziasz, Philip J., Singbeil, Douglas L. and Singh,
- 4 June 1998, Vol. 3, TAPPI, pp. 641 – 654.
Preet M., 2001, Why do kraft recovery boiler
Kaliazine, Andrei L., Piroozmand, Farshad, composite floor tubes crack? Tappi Journal, Vol.
Cormack, Donald E. and Tran, Honghi, 1998, 84, No. 8, August 2001, 12 p.
Sootblower optimization II. Tappi Journal, Vol. 80,
Keiser, J. R., Aramayo, G. A., Goodwin, G. M.,
No. 11, pp. 201 – 207.
Maziasz, P. J. and Gorog, J. P., 2000, Improved ma-
Kankkunen, Ari, Helpiö, Tomi and Rantanen, terials for use as components in kraft black liquor
Pekka, 1994, Small scale measurement of black recovery boilers. Oak Ridge National Lab., Report
liquor spraying with splashplate nozzles. Proceed- C/ORNL95-0375, 34 p.
ings of 1994 TAPPI Engineering Conference, San
Kelly, P. A., Frederick, William. J. and Grace,
Francisco, 19 - 22 September, pp. 207 - 214.
Thomas M., 1981, The residence time distribution
Karhula, M., 1992, Soodakattilan compound- of inorganic salts in kraft recovery boilers. Tappi
putkien säröilytutkimus. (Study of recovery boiler Journal, Vol. 64, No. 10, October, pp. 85 - 87.
compound tube cracking) in Finnish, M. Sc.
Khalaj, Asghar, Kuhn, David and Tran,
thesis, Lappeenranta University of Technology,
Honghi, 2004, Composition of carryover particles
Finland, 79 p.
in recovery boilers. Proceedings of 2004 Interna-
Kawada, Shin and Ibuka, Hideo, 1991, The tional Chemical Recovery Conference, June 6-10
newest Babcock-Hitachi recovery boiler. in Japa- 2004, Charleston, South Carolina, USA, 1009-
nese, Japan Tappi Journal, Vol. 45, No. 6, pp. 644 1020 p.
– 655.
References 12-11
Kiiskilä, Erkki, Lääveri, Anu, Nikkanen, Sam- bed burning. Proceedings of 1998 International
uli and Vakkilainen, Esa, 1993, Possibilities for Chemical Recovery Conference, Tampa, USA, 1
new black liquor processes in the pulping industry - 4 June 1998, TAPPI Press, Atlanta, pp. 599 - 614.
energy and emissions. Bioresource Technology, Vol. Koivisto, Lasse, 2002, Recovery boiler corro-
46, pp. 129 - 134. sion and materials issues. Presentation, Andritz
Kilpinen, P., Hupa, M. and Aho, M., 1997, Corporation.
Selective non-catalytic NOx reduction at elevated Koivisto, Lasse and Holm, Ralf, 1998, Furnace
pressures: studies on the risks for increased N2O floor design and materials for recovery boilers.
emissions. Proc. 7th Int. Workshop on Nitrous Proceedings of 1998 TAPPI Engineering Confer-
Oxide Emissions, Cologne, April 21-23, 1997, Ber- ence, September 14-18, 1998, Atlanta, Georgia, pp.
gische Universität Gesamthochschule Wuppertal, 1091 - 1098.
Physikalische Chemie, Bericht Nr. 41, September,
Kraft pulping, 1989, A compilation of notes,
1997.
editor Agneta Mimms, Tappi Press.
Kilpinen, Pia, 1992, Kinetic modeling of gas
Kulas, Katherine A., 1990, An overall model
phase nitrogen reactions in advanced combustion
of the combustion of a single droplet of kraft black
processes. Ph. D. thesis, Åbo Akademi University,
liquor. Ph. D. thesis, The Institute of Paper Sci-
350 p, ISBN 951-650-132-X.
ence and Technology. 159 p.
Kingery, W. D. and Berg, M., 1955, Study of
Kulas, Katherine A., Harper, Frank and Clay,
the initial stages of sintering solids by viscous flow,
David, 1989, Black liquor burning advances in
evaporation-condensation, and self-diffusion.
understanding. The midnight sun colloquium on
Journal of Applied Physics, Vol. 26, No. 10, pp.
recovery research, June 14 – 16, 1989, STFi, pp.
1205 - 1212.
11 - 22.
Kittredge, C. P. and Rowley, D. S., 1974,
Kulas, Katherine A. and Clay, David T., 1988,
Resistance coefficients for laminar and turbulent
An empirical rate equation describing the volatiles
flow through ½ inch valves and fittings. Transac-
burning stage of kraft black liquor. AIChE Forest
tions of ASME, Journal of Engineering for Power,
Products Division, pp. 53 - 57.
pp. 1579 - 1586.
Kurkela, J., Latva-Somppi, J., Tapper, U., Kaup-
Kippo, Asko, 1979, Simulation studies of
pinen, E. I. and Jokiniemi, J., 1998, Ash forma-
dynamics and multivariable controllers of the re-
tion and deposition onto heat exchanger tubes
covery boiler. Lic. Tech. thesis, University of Oulu.
during fluidized bed combustion of wood-based
Kjörk, Anders and Herstad Swärd, Solvie, fuels. Proceedings of the ABC’98 International
2000, Atmospheric emission of nitrogen oxide Conference on Ash Behaviour Control in Energy
from kraft recovery boilers in Sweden. (Kartläg- Conversion Systems, Yokohama, Japan, March 18-
ging av NOx-utsläpp från sodapannor i Sverige) in 19, 1998, pp. 110 - 118.
Swedish, Technical Report, SVF-679, Värmeforsk,
Kymäläinen Marita, Forssén Mikael, Jansson,
Stockholm Sweden, Project Värmeforsk S9-807,
Mette and Hupa Mikko, 2002, The fate of nitrogen
ISSN 0282-3772, 58 p.
in the chemical recovery process in a kraft pulp
Klarin, Anja, 1993, Floor tube corrosion in mill, Part IV: Smelt nitrogen and its formation
recovery boilers. Tappi Journal, Vol. 76, No. 12, in black liquor combustion. Journal of Pulp and
December 1993, pp. 183 - 188. Paper Science, Vol. 28, No. 5, pp. 151 - 158.
Klarin, Anja, 1992, Analysis of char bed mate- Kymäläinen, Marita, Forssén, Mikael and
rial. in Finnish, Ahlstrom Machinery, Internal Hupa, Mikko, 1999, The fate of nitrogen in the
report. 3 p. chemical recovery process in a kraft pulp mill.
Knacke, O. Kubaschewski, O. and Hessel- Part I: A general view. Journal of Pulp and Paper
mann, K., 1991, Thermochemical properties of Science, Vol. 25, No. 1, pp. 410 - 417.
inorganic substances. II, Springer-Verlag, Berlin. Källberg, Y., 1982, Simulation of formation
2412 p. and depletion of NOx in high temperatures. in
Kochesfahani, Saied H. and Tran, Honghi, Swedish, M. Sc. thesis, Åbo Akademi.
2000, Field studies on intermediate sized particles Labossiere, James and Henry, Jeff, 1999,
(ISP) in recovery boilers. UEF Conference on Chromizing for near-drum corrosion protection.
Behavior of Inorganic Materials in Recovery Boil- Tappi Journal, Vol. 82, No. 9, September 1999, pp.
ers, Bar Harbor, Maine, 4 - 9 June 2000, United 150 - 157.
Engineering Foundation, 22 p.
Laitinen, A., Pulliainen, M. and Heinävaara,
Kochesfahani, S. Tran, H. N., Jones, A. K., A., 1998, On-line monitoring of recovery boilers.
Grace, T. M., Lien, S. J. and Schmidl, W., 1998, 9th International Symposium on Corrosion in the
Particulate formation during black liquor char Pulp and Paper Industry, May 26 – 29, 1998, Ot-
tawa, Ontario, Canada, 199 – 202. Lien, Steven J., Ling, Alisa, Frederick, Wm.
Lankinen, Matti, Paldy, Ivan V., Ryham, Rolf James Jr. and Tran, Honghi, 1999, Sintering
and Simonen, Liisa, 1991, Optimal solids recovery. and densification rates for recovery boiler fume
Proceedings of CPPA 77th Annual meeting, pp. deposits. Paper 5-5, Proceedings of 1999 Tappi
A373 - A378. Engineering Conference, September 12 – 16,
Anaheim, California.
Ledung, Lars and Ulmgren, Per, 1998, Viscos-
ity changes in black liquor when bleach plant Lind, Terttaliisa, Rumminger, Marc, Baxter,
filtrates are added. Proceedings of 1998 Interna- Larry, Kauppinen, Esko, Wessel, Rick, Roscoe, Bob
tional Chemical Recovery Conference, Tampa, and Ariessohn, Peter, 2000, A field investigation
USA, 1 - 4 June 1998, TAPPI Press, Atlanta, pp. of recovery boiler salt particle characteristics. UEF
157 - 168. Conference on Behavior of Inorganic Materials in
Recovery Boilers, Bar Harbor, Maine, 4 - 9 June
Ledung, Lars and Ulmgren, Per, 1997,
2000. United Engineering Foundation, 17 p.
Combustion properties of black liquors contain-
ing additions of totally chlorine free bleach plant Lindberg, Hans and Ryham, Rolf, 1994,
filtrates. Nordic Pulp and Paper Research Journal, Chemical recovery technology for future fiber
Vol. 12, No. 3, pp. 145 - 149. production. CPPA 80th Annual meeting, pp. A73
- A78.
Ledung, Lars and Kaul, Vikaram, 1996,
Sodapanneprocessen: en sammanställning av Lindblom, Josefina, 1999, Sintering in the
forskningsresultat från litteraturen. (Recovery kraft pulp mill – studies in the rotary kiln and re-
boiler process: compendum of research results covery boiler. Ph. D. thesis, Doktorsavhandlingar
from litterature) in Swedish, STFI-rapport BF 4, vid Chalmers tekniska högskola, Ny serie 1554,
Juni 1996, 67 p. Chalmers University of Technology, Göteborg,
Sweden, 88 p., ISBN 91-7197-864-X.
Lefohn, A. S., Husar, J. D., Husar, R. B. and
Brimblecombe, P., 1996, Assessing historical glo- Lindsley, David, 2000, Power-plant control
bal sulfur emission patterns for the period 1850- and instrumentation: the control of boilers and
1990. Technical report, DOE/ER/30234—1, A.S.L. HRSG systems, Institution of Electrical Engineers,
and Associates, Helena, MT United States. 190 p. London, 222 p., ISBN 0-85296-765-9.
Li, Jian, 1989, Rate processes during gasifica- Llinares, Jr. V. and Chapman, P. J., 1989,
tion and reduction of black liquor char. Ph. D. Stationary firing, three level air system retrofit
thesis, McGill University. experience. Proceedings of 1989 Tappi Engineer-
ing Conference, Atlanta, Georgia, September
Li, Jian and van Heiningen, A. R. P., 1991,
10-13, 10 p.
Sulfur emissions during slow pyrolysis of kraft
black liquor. Tappi Journal, Vol. 74, No. 3, March Lobben, Peder, 1930, Handboken för mekan-
1991, pp. 237 - 239. iker. (Mechanic’s handbook) in Swedish, 4th edi-
tion, Albert Bonniers Förlag, Stockholm, 748.
Sodium emission during pyrolysis and gasification Lobscheid, H., 1965, Dampf Babcok – Hand-
of black liquor char. Tappi Journal, Vol. 73, No. 1, buch. (Steam Babcock – handbook) in German,
December 1990, pp. 213 - 239. 4th edition, Vulkan – Verlag, Essen, 322 p.
Li, Jian and van Heiningen, A. R. P., 1989, Lovo, Augusto Cesar, Puig, Floreal and Santos,
Sulfur emission during gasification of black liquor Saulo, 2004, Presentation of recovery boiler’s
char. Proceedings International Chemical Recov- retrofits – Aracruz Celulose S/A. Proceedings of
ery Conference, CPPA, pp. 209 - 216. 2004 International Chemical Recovery Confer-
ence, June 6-10 2004, Charleston, South Carolina,
USA, pp. 35 - 43.
Mass transfer limitations in the gasification of
black liquor char by CO2. Journal of Pulp and Lundborg, Sten, 1977, Undersökning av
Paper Science, Vol. 12, No. 5, September, pp. J146 smältatemperaturer. (Study of smelt temperatures)
- 151. in Swedish, Sodahuskonferensen ´77, ÅF-IPK,
Stockholm, pp. 43 – 50.
Lidén, Jan, 1995, Green liquor dregs, its origin
and effects in the lime cycle. Proceedings of 1995 Luthe, Corinne E., Uloth, Vic, Karidio,
International Chemical Recovery Conference, Ibrahim and Wearing Jim, 1997, Are salt-laden
Toronto Canada, CPPA, pp. A291 – A298. recovery boilers a significant source of dioxins?
Tappi Journal, Vol. 80, No. 2, February 1997, pp.
Lien, Steven J. and Verrill, Christopher L.,
165 - 169.
2004, Particulate emissions from char bed burn-
ing. Proceedings of 2004 International Chemical Lyon, Rickhard K., Cole, Jerald A., Kramlich,
Recovery Conference, June 6-10 2004, Charleston, John C. and Chen Shi L., 1989, The selective
South Carolina, USA, 1023-1041 p. reduction of SO3 to SO2 and the oxidation of NO
References 12-13
to NO2 by methanol. Combustion and Flame, Vol. & PCE&I, Chicago, Il, USA, 8 p.
81, No. 1, pp. 30 - 39. Mannola, Lasse and Burelle, Richard, 1995,
Lövblad, Roland, Moberg, Göran, Olausson, Operating experience of a 7,270,000 lb d.s./day
Lars and Boström, Curt-Åke, 1991, NOx-reduc- recovery boiler. Pulp & Paper Canada, Vol. 96, No.
tion from a recovery boiler by injection of an 3, March 1995, pp. 70 – 72.
enhanced urea solution. Proceedings of 1991 Mao, Xiaosong, Tran, Honghi and Cormack,
Tappi Environmental Conference, Tappi Press, pp. Donald E., 2000, Effects of chemical composition
1071 – 1075. on the removability of recovery boiler fireside
MacCallum, Colin, 1992, Towards a superior deposits. Proceedings of 2000 Tappi Engineering
recovery boiler air system. Proceedings of 1992 Conference, TAPPI Press, Atlanta, 11 p.
International Chemical Recovery Conference, Maso, Vesa, 1988, Formation and reduction of
Seattle, Washington, June 7-11, pp. 45 - 56. emissions. Seminar on Black Liquor Combustion,
MacCallum, Colin and Blackwell, Brian, R., February 24 - 25, Lappeenranta, Lappeenranta
1985, Modern kraft recovery boiler liquor-spray University of Technology. 12 p.
and air systems. Proceedings of 1985 International Masse, M. A., Kiran, E. and Fricke, A. L., 1986,
Chemical Recovery Conference, New Orleans, LA, Polymer, Vol. 27, p. 619.
pp. 33 - 47.
Matsuda, Takao, Masuda, Takeharu and Sue-
MacCallum, Colin, 1982, Tailor-made recov- mitsu, Nobuo, 199x. Outline of the world’s largest
ery boilers. Proceedings of Black liquor recovery soda recovery boiler with high pressure and high
boiler symposium 1982, Helsinki, August 31 temperature steam. pp. 234 – 248.
– September 1, The Finnish Recovery Boiler Com-
McCann, Colin, 1991, A review of recovery
mittee – EKONO Oy, 12 p.
boilers process design. CPPA 77th Annual meet-
Maddern, K. N., 1988, Mill-scale development ing, pp. A49 - A58.
of the DARS direct causticization process. Pulp &
McCarthy, J. H., 1968, Recovery plant design
Paper Canada, Vol. 87, No. 10, pp. T395 – T399.
and maintenance. Chapter 5 of Chemical recovery
Maddox, R. N., 1983a. Properties of mixtures in alkaline pulping process, editor Whitney, Roy
of fluids. in Heat Exchanger Design Handbook, P., TAPPI Monograph series, No. 32, Mack Printing
Part 5. Physical properties. Hemisphere Publish- Company, Easton, Pa., pp. 159 – 199.
ing Corporation, pp. 5.2.1-1 - 5.2.5-4.
McDonald, Kevin L., 1977, Calorific analysis
Maddox, R. N., 1983b. Properties of super- of spent liquors by carbon analysis. Tappi Journal,
heated gases. in Heat Exchanger Design Hand- Vol. 60, No. 12, December 1977, pp. 107 – 109.
book, Part 5. Physical properties. Hemisphere
McGurn, J. F., 1993, Composite tubes in black
Publishing Corporation, pp. 5.5.2-1 - 5.5.2-11.
liquor recovery boilers: an update. Pulp & Paper
Magnusson, Hans, 1977, Sodapannan som Canada, Vol. 94, No. 12, December 1993, pp. T428
kemisk reaktor. (Recovery boiler as a chemical – T432.
reactor) in Swedish, Sodahuskonferensen ´77, ÅF-
McKeough, Paterson, 2004, Evaluation of
IPK, Stockholm, pp. 26 – 29.
potential improvements to BLG technology.
Makino, A., 1981, A theoretical and experi- Proceedings of PulPaper 2004 Energy and carbon
mental study of carbon combustion in stagnation management, 1-3 June 2004, Helsinki Finland, pp.
flow. Combustion and Flame, Vol. 73, pp. 166 69 - 76.
- 187.
McKeough, Paterson, 2003, Evaluation of
Makkonen, Pasi, 1999, Artificially intelligent potential improvements to BLG technology. Pro-
and adaptive methods for prediction and analysis ceedings of Colloquium of black liquor combus-
of superheater fireside corrosion in fluidized bed tion and gasification, Park City, Utah, 13-16 May
boilers. Ph. D. thesis, Lappeenranta University of 2003, 12 p.
Technology, Acta Universialis Lappeenrataensis,
McKeough, Paterson and Janka, Kauko, 2001,
89, 187 p. ISBN 951-764-375-6.
Sulphur behaviour in the recovery boiler furnace:
Manninen, Jussi and Vakkilainen, Esa, 1996, theory and measurements. Proceedings of 2001
Reduction of sulfur emissions from kraft recovery International Chemical Recovery Conference,
furnace with black liquor heat treatment. Finnish- Whistler, PAPTAC Publications, pp. 231 – 237.
Swedish Flame Days, Naantali.
McKeough, P. J. and Vakkilainen, E. K., 1998,
Malmberg, Barry, Edwards, Lou, Lundborg, Effects of black liquor composition and furnace
Sten, Ahlroth, Mikael and Warnqvist, Björn, 2003, conditions on recovery boiler fume chemistry.
Prediction of dust composition and amount in Proceedings of 1998 International Chemical Re-
kraft recovery boilers. Proceedings of 2003 TAPPI covery Conference, Tampa, USA, 1 - 4 June 1998,
Fall Technical Conference: Engineering, Pulping
TAPPI Press, Atlanta, pp. 487 - 504. 13, No. 4, pp. 778 - 795.
McKeough, P. J., Kurkela, M., Arpiainen, V., Mikkanen, Pirita, Kauppinen, Esko J., Joki-
Mikkanen, P., Kauppinen, E. and Jokiniemi, J., niemi, Jorma K., Sinquefield, Scott and Frederick,
1995, The release of carbon, sodium and sulphur William, J., 1994, Bimodal fume particle size
during rapid pyrolysis of black liquor. Proceedings distributions from recovery boiler and laboratory
of 1995 TAPPI International Recovery Confer- scale black liquor combustion. Tappi Journal, Vol.
ence, Toronto Canada, CPPA, pp. A217 - A226. 77, No. 12, December 1994, pp. 81 - 84.
McMillan, John, 2004, Superheater problems, Mikkanen, P., Kauppinen, E. I., Jokiniemi, J.
their causes and solutions. Proceedings of 2004 K., Sinquefield, S. A., Frederick, W. J. and Mäk-
International Chemical Recovery Conference, inen, M., 1994b, The particle size and chemical
June 6-10 2004, Charleston, South Carolina, USA, species distributions of aerosols generated in kraft
pp. 953 – 964. black liquor pyrolysis and combustion. AIChE
Merriam, Richard L., 1980, KRAFT, version Symp. Ser., Vol. 90, No. 302, pp. 46 - 54.
2.0 computer model of a kraft recovery furnace, Milanova, E. and Kubes, G. J., 1985, The com-
Vol. III: users manual. prepared for The American bustion of kraft liquor chars. Proceedings of 1985
Paper Institute, New York, Arthur D. Little, Inc., International Chemical Recovery Conference,
216 p. New Orleans, Louisiana, Book 3, pp. 363 - 370.
Merriam, Richard L., 1977, KRAFT, version Milanova, E., 1988, Variables affecting the
2.0 computer model of a kraft recovery furnace, swelling of kraft black liquor solids. Journal of
Vol. II: engineering manual. prepared for The Pulp and Paper Science, Vol. 14, No. 4, July, pp. J95
American Paper Institute, New York, Arthur D. - J102.
Little, Inc. 136 p. Miller, P. T., Clay, David T. and Lonsky, W.
Metiäinen, Sami, 1991, Soodakattilan ilman- F. W., 1986, The influence of composition on the
syöttömallien kokeellinen tutkimus. (Experimen- swelling of kraft black liquor during pyrolysis.
tal investigation of recovery boiler air models) in Proceedings of Tappi Engineering Conference, pp.
Finnish, M. Sc. thesis, Lappeenranta University of 225 - 234.
Technology. 84 p. Mitsubishi chemical recovery boiler, 1997,
Miikkulainen, Pasi, Kankkunen, Ari and Sales brochure, Mitsubishi Heavy Industries, Ltd.
Järvinen, Mika, 2002, Mustalipeän ruiskutusko- Tokyo, Japan, 18 p.
keet soodakattilassa ja ruiskutuskammiossa. Moberg, Orvar, 1974, Recovery boiler cor-
(Black liquor firing tests in recovery boiler and rosion. in Pulp and paper industry corrosion
firing test rig) in Finnish, Energy Engineering problems, National Association of Corrosion
and Environmental Protection publications, Engineers, Houston, Texas, pp. 125 – 136.
Helsinki University of Technology, Department of
Moberg, Orvar, 1973, Sodahusets roll i svavel-
Mechanical Engineering, No. 8, 34 p. ISBN 951-
balansen. (Recovery boilers role in sulfur balance)
22-6026-3.
in Swedish, Sodahuskonferens, 29 November
Mikkanen, P., Jokiniemi, J. K., Kauppinen, E. I. 1973, Ångpannaföreningen, Stockholm, pp. 8
and Vakkilainen, E. K. 2001, Coarse ash parti- – 14.
cle characteristics in a pulp and paper industry
Moberg, Orvar, 1967, Eftereldytornas
chemical recovery boiler. Fuel, Vol. 80, No. 7, pp.
utformning i sodahusaggregat. (Form of after
987 – 999.
furnace surfaces in a recovery plant) in Swedish,
Mikkanen, Pirita, 2000, Fly ash formation in Sodahuskonferens, 27 November 1967, ÅF-IPK,
kraft recovery boilers. Ph. D. thesis, VTT Publica- Stockholm, pp. M1 – M18.
tions 421, Technical research centre, VTT, Espoo,
Modern power station practice, Vol. B, Boil-
116 p. ISBN 951-38-5583-X.
ers and ancillary plant, 1991, 3rd edition, Perga-
Mikkanen, Pirita. Kauppinen, Esko, Jokiniemi, mon press, Singapore, 184 p. ISBN 0-08-040512-6.
Jorma, Moisala, Antero, Vakkilainen, Esa K. and
Modern power station practice, Vol. E,
Hämäläinen, Jouni, 1999, Ash particle deposition
Chemistry and metallurgy, 1992, 3rd edition,
study in an operating recovery boiler. American
Pergamon press, Singapore, 576 p. ISBN 0-08-
Aerosol Association Research Conference, Octo-
040515-0.
ber 11-15, 1999, Tacoma, Washington.
Modern power station practice, Vol. 2,
Mikkanen, Pirita. Kauppinen, Esko,
Mechanical boilers, fuel and ash-handling plant,
Pyykönen, Jouni, Jokiniemi, Jorma, Aurela,
1971, 2nd edition, Pergamon press, Hungary, 432 p.
Minna, Vakkilainen, Esa K. and Janka, Kauko,
ISBN 08-016060-3.
1999, Alkali salt ash formation in four Finnish
industrial recovery boilers. Energy and Fuels, Vol. Moody, L. F., 1944, Friction factors for pipe
References 12-15
flow. Journal of Heat Transfer, Vol. 66, pp. 671 Niemelä, Klaus, 2004, Early history of black
– 684. liquor research: recovery of by-products and
Myslicki, T. P. and Orr, Alex, 1992, New 3000 cooking chemicals. 40th Anniversary Internation-
tons high pressure / high temperature recovery al Recovery Boiler Conference, Finnish Recovery
boiler. Proceedings of 1992 International Chemi- Boiler Committee, Haikko Manor, Porvoo, May
cal Recovery Conference, Seattle, Washington, 12-14, 2004, pp. 7 – 14.
June 7-11, pp. 413 – 420. Niemelä, Klaus, Järvinen, Risto, Ulmgren, Per,
Mäkipää, M., Kauppinen, E., Lind, T., Samuelsson, Åsa, Forssén, Mikael, DeMartini,
Pyykönen, J., Jokiniemi, J., McKeough, P., Oksa, Nicolai, Kilpinen, Pia, Hupa, Mikko, Saviharju,
M., Malkow, Th., Fordham, R. J. Baxter, D., Kari, Ferreira Pinho and Rebola, João, 2002,
Koivisto, L., Saviharju, K. and Vakkilainen, E., Reduction of air emissions at kraft pulp mills
2001, Superheater tube corrosion in recovery - overview of a multidisciplinary EU project. Pro-
boilers. VTT Symposium 214, 10th International ceedings of the 7th New Available Technologies
Symposium on Corrosion in the Pulp and Paper Conference, Stockholm, Sweden, June 4-6, 2002,
Industry (10th ISCPPI), Helsinki, 21 - 24 August pp. 78 - 81.
2001, Vol. 1, editor Hakkarainen, Tero, VTT Niemelä, Klaus and Ulmgren, Per, 2002, Be-
Manufacturing Technology, Espoo, 2001, pp. 157 haviour of nitrogen during kraft pulping of wood.
– 180. ISBN 951-38-5720-4. Proceedings of the 7th European Workshop on
Mäkipää, Martti and Backman, Rainer, 1998, Lignocellulosics and Pulp, Turku, Finland, August
Corrosion of floor tubes in reduced kraft smelts: 26–29, 2002, pp. 71 - 74.
studies on effects of chlorine and potassium. Niemelä, Klaus, 2001, Identification of novel
Proceedings of 9th International Symposium on volatile sulfur and nitrogen compounds in kraft
Corrosion in the Pulp and Paper Industry, May 26 pulping process streams. Proceedings of the 11th
– 29, 1998, Ottawa, Ontario, Canada. International Symposium on Wood and Pulping
Mäntyniemi, J. and Hartley, C. E., 1999, Wid- Chemistry, Nice, France, June 11-14, 2001, Vol. I,
ening the operating envelope of recovery boilers pp. 125 - 128.
by using high dry solids firing, Proceedings of Niemitalo, Hanna, 1993, Soodakattilan likaan-
1999 TAPPI Engineering/Process and Product tuminen ja nuohouksen ohjaus. (Fouling and
Quality Conference, Vol. 1, p. 479. sootblowing control of recovery boiler) in Finnish,
Mäntyniemi, Jussi and Haaga, Kari, 2001, Op- M. Sc. thesis, University of Oulu.
erating experience of XL-sized recovery boilers. Nikkanen, Samuli, Tervo, Olavi, Lounasvuori,
Proceedings of 2001 Tappi Engineering, Finishing Risto and Paldy, Ivan V., 1989, Experience of
& Converting Conference, Tappi Press, Atlanta, recovery boiler modernizations. Proceedings of
GA, 7 p. 1989 International Chemical Recovery Confer-
NCASI summary of PCDD/F emissions from ence, Ottawa, Ontario, pp. 39 - 44.
wood residue and black liquor combustion - Lo- Nikuradse, J., 1933, Strömungsgesetz in
cating and estimating air emissions from sources rauchen rohren. (Flow loss in rough tubes) in
of dioxins and furans, 1996, U.S. EPA Draft Final German, Forshungsheft 361, Ausgabe B, band 4,
Report, January 1996. VDI Verlag Gmbh, Berlin NW7.
Navitsky, G. J. and Gabrieli, F., 1980, Boiler Nishikawa, Eiichi, 1999, General planning
water treatment, feedwater treatment, and chemi- of boiler gas side heat transfer surfaces. in Steam
cal cleaning of drum-type utility steam generators. power Engineering, editor Ishigai, Seikan, Cam-
Combustion, August, p. 19. bridge University Press, 394 p. ISBN 0 521 62635
Nichols, Kenneth M. Thompson, Laura M. 8.
and Empie, H. Jeff, 1993, A review of NOx forma- Nishio, Toshiaki, Matsumoto, Hirotoshi, Shi-
tion mechanisms in recovery furnaces. Tappi nohara, Masatomo and Arakawa, Yoshihisa, 2001,
Journal, Vol. 76, No. 1, January 1993, pp. 119 Influence of ash composition on corrosion rate
- 124. of superheater tubes. Proceedings of 2001 Tappi
Nichols, Kenneth M., 1992, Combustion Engineering, Finishing & Converting Conference,
of concentrates resulting from ultrafiltration of Tappi Press, Atlanta, GA, 6 p.
bleached-kraft effluents. Tappi Journal, Vol. 75, Noopila, Taina, Alén, Raimo and Hupa,
No. 4, April 1992, 153 - 158. Mikko, 1991, Combustion properties of labora-
Nichols, Kenneth, M., 1987, Nitrogen pollut- tory-made black liquors. Journal of Pulp and Paper
ant formation in a high pressure entrained- coal Science, Vol. 17, No. 4, July, pp. J105 - J109.
gasifier. Ph. D. thesis, Department of Chemical Nunes, Walter R., Puig, Floreal and Lindman,
Engineering, Brigham Young University. 242 p. Nils, 1995, Purging of ESP ash to control steam

temperature. Proceedings of 1995 International and paper mill SO2 and NOx emissions, 1980–
Chemical Recovery Conference, Toronto Canada, 1995. Tappi Journal, Vol. 81, No. 4, April 1998, pp.
CPPA, pp. A59 – A61. 114 - 122.
Näretie, V. and Arpalahti, Esko, 1972, Höyry- Piroozmand, F., Tran, H. N., Kaliazine, A. and
tekniikka. (Steam technology) in Finnish, 277 p. Cormack, D. E., 1998, Strength of Recovery Boiler
ISBN 951-1-00009-8. Fireside Deposits at High Temperatures. Pro-
Ohlström, Mikael O., Lehtinen, Kari E. J., ceedings of 1998 Tappi Engineering Conference,
Moisio, Mikko and Jokiniemi, Jorma K., 2000, TAPPI Press, Atlanta, pp. 169-179.
Fine-particle emissions of energy production in Plumley, A. L., Lewis, E. C., Barker, T. J. and
Finland. Atmospheric Environment, Vol. 34, No. Esser, F. A., 1989, Chromizing for recovery boiler
22, pp. 3701 - 3711. corrosion protection. Proceedings of 1989 Tappi
Ohtomo, Kenji, 2000, The experience of high Engineering Conference, Atlanta, Georgia, Sep-
temperature and pressure recovery boiler opera- tember 10-13, 8 p.
tion. Japan Tappi Journal, Vol. 54, No. 1, January Pohjanne, Pekka, Hänninen, Hannu, Mäkipää,
2000, pp. 109-114. Martti and Ehrnsten, Ulla, 1992, Cracking of
Oulu Osakeyhtiö’s new sulphate mill, 1937, compound tubes in black liquor recovery boilers.
The Finnish Paper and Timber Journal, Vol. 19, No. Proceedings of the 17th International Sympo-
7A, pp. 276 – 278. sium on Corrosion in the Pulp & Paper Industry,
Orlando, FL, USA, 16 - 20 November 1992, TAPPI
Paju, Raimo, 1992, Features of safe and reliable
Atlanta, pp. 319 – 324.
recovery boiler. Premier symposium international,
Prevention sur les chaudieres de recuperation a Poon, William, Barham, David and Tran,
liqueur noire, Bordeaux, France, 10 p. Honghi, 1993, Formation of acidic sulfates in kraft
recovery boilers. Tappi Journal, Vol. 76, No. 7, July
Pantsar, Ossi, 1988, Comparison of firing
1993, pp. 187 - 193.
methods, Effect of furnace size on the process.
Seminar on black liquor combustion, 24 – 25 Puumassan valmistus, 1983, (Manufacture
February 1988, Lappeenranta. of pulp) in Finnish, editor Virkola, Nils-Erik,
Suomen Paperi-insinoorien Yhdistyksen oppi- ja
Park, Song Won, 2001, Análise de modelos de
käsikirja II, osa 1, 2nd edition, 1106 p. ISBN 951-
calderira de recuperação. (Analysis of recovery
99117-3-1.
boiler models) in Portuguese, 14th Latin-Ameri-
can Recovery Boilers meeting, August 7 – 10, Sao Pyykönen, Jouni, 2002, Computational simu-
Paulo, Associação Brasileira Technica de Celulose lation of aerosol behaviour. Ph. D. thesis, VTT
e Papel. 8 p. Publications 461, Espoo, 154 p. ISBN 951-38-
5977-0.
Paul, Larry D., Barna, Joan L., Danielson,
Michael J. and Harper, Sharon L., 1993, Corro- Pyykönen, Jouni and Jokiniemi, Jorma, 2002,
sion-resistant tube materials for extended life of Modelling alkali chloride superheater deposition
openings in recovery boilers. Tappi Journal, Vol. and its implications. Fuel Processing Technology,
76, No. 8, August 1993, pp. 73 - 77. Vol. 80, No. 3, 15 March 2003, pp. 225 - 262.
Passinen, K, 1968, Chemical composition of Rahtu, Jari, 1990, Lämmönsiirto soodakattilan
pulping spent liquors. Proceedings of the Sympo- tulistinalueella. (Heat transfer in recovery boiler
sium on Recovery of Pulping Chemicals, Helsinki, superheaters) in Finnish, M. Sc. thesis, Lappeen-
The Finnish Pulp and Paper Research Institute ranta University of Technology, Department of
and Ekono, pp. 184 - 210. Energy Technology, 92 p.
Performance test procedure: sodium base Raukola, Antti and Haaga, Kari, 2004,
recovery units, 1996, CA Report No. 84041601, Expandable recovery boiler, a novel method for
March 1996, Tappi Press, 54 p. ISBN 0-89852338. future capacity increase. Proceedings of 2004
International Chemical Recovery Conference,
Perjyd, L. and Hupa, Mikko, 1984, Bed and
June 6-10 2004, Charleston, South Carolina, USA,
furnace gas composition in recovery boilers - ad-
pp. 241 - 246.
vanced equilibrium calculations. Proceedings of
1984 TAPPI Pulping Conference, San Francisco, Raukola, Antti T., Ruohola, Tuomo and
November. 11 p. Hakulinen, Aki, 2002, Increasing power genera-
tion with black liquor recovery boiler. Proceed-
Pettersson, Bertil, 1983, Korsnäs sodapan-
ings of 2002 TAPPI Fall Technical Conference,
nor under 40 år. (40 years of recovery boilers at
September 8 - 11, San Diego, Ca, 11 p.
Korsnäs) in Swedish, Sodahuskonferensen ‘83,
ÅF-IPK, Stockholm, pp. 57 – 68. Raymond, Del, 2003, The compelling case
for gasification technology in the forest products
Pinkerton, John E., 1998, Trends in U.S. pulp
industry. Proceedings of Colloquium of black liq-
References 12-17
uor combustion and gasification, Park City, Utah, rosion i sodahuseldstäder i Finland. (Recent cor-
13-16 May 2003, 14 p. rosion experiences of recovery boiler furnaces in
Recovery boiler design review – Functional Finland) in Swedish, Sodahuskonferens, ÅF-IPK,
design considerations. 1993, Babcock and Wilcox Stockholm, pp. C1 – C18.
Recovery Boiler Users’ Association meeting, Sep- Roos, Tom, 1968, Design and operation
tember 14 – 15, 1993, Atlanta Georgia, 15 p. experience of reductive recovery boilers. Proceed-
Reis, V. V., Frederick, W. J., Wåg, K. J., Iisa, K. ings of the Symposium on Recovery of Pulping
and Sinquefield, S. S., 1994, Potassium and chlo- Chemicals, Helsinki, The Finnish Pulp and Paper
ride enrichment during black liquor combustion. Research Institute and Ekono, pp. 437 - 489.
Paper 104c, 1994 AIChE Spring National Meeting, Rosner, D. E., 1986, Transport processes in
Atlanta, April 17-21. chemically reacting flow systems. Butterworth-
Reis, Victor V., Frederick, William J., Wåg, Kaj Heinemann, Stoneham, MA.
J., Iisa, Kristiina and Sinquefield, Scott A., 1995, Rönnquist, Eva-Marie, 2000, Överhetta-
Effects of temperature and oxygen concentration rkorrosion i bioeldad panna - teorier och prov i
on potassium and chloride enrichment during Västermalmsverket, Falun. (Superheater corrosion
black-liquor combustion. Tappi Journal, Vol. 78, in biomass boiler - theories and tests in Väster-
No. 12, December 1995, pp. 67 - 76. malmsverket, Falun) in Swedish, Technical report,
Richards, Tobias, 2001, Recovery of kraft black SVF-708, Värmeforsk Service AB. 94 p.
liquor - alternative processes and system analysis. Rydholm, Sven A., 1965, Pulping processes.
Ph. D. thesis, Doktorsavhandlingar vid Chalm- Interscience Publishers, a division of John Wiley
ers tekniska högskola, Ny serie 1732, Chalmers & Sons, Inc., London. 1269 p. ISBN 0 471 74793 9.
University of Technology, Göteborg, 88 p. ISBN Ryham, Rolf, 1992, A new solution to third
91-7291-048-8. generation chemical recovery. Proceedings of 1992
Richardson, David L. and Merriam, Richard International Chemical Recovery Conference,
L., 1977, Study of cooling and smelt solidification Seattle, Washington, June 7-11, pp. 581 - 588.
in black liquor recovery boilers. Phase I report, Ryham, Rolf and Nikkanen, Samuli, 1992,
prepared for The American Paper Institute, New High solids firing in recovery boilers. Tappi Kraft
York, Arthur D. Little, Inc., 133 p. Recovery Operations Short Course, pp. 233 - 238.
Richardson, David L. and Merriam, Richard Ryham, Rolf, 1990, High solids evaporation of
L., 1978, A study of black liquor recovery furnace kraft black liquor using heat treatment. Proceed-
firing conditions, char bed characteristics and ings of 1990 TAPPI Engineering Conference,
performance. Phase II report, prepared for The Westin Seattle, Seattle, Washington, September 24
American Paper Institute, New York, Arthur D. - 27, Book 2, TAPPI Press, pp. 677 - 681.
Little, Inc., 165 p.
Ryti, Henrik, 1969, Lämmönsiirto. (Heat
Richter, W., 1985, Scale-up and advanced per- transfer) in Finnish, Tekniikan käsikirja, Osa 4, K.
formance analysis of boiler combustion chambers. J. Gummerus, Jyväskylä, pp. 595 – 600.
American Society of Mechanical Engineers winter
Saarinen, P. and Hänninen, H., 1997, Termin-
annual meeting, Miami, FL, United States, 17-21
en väsyminen soodakattilan kompound-putkissa.
Nov 1985, 21 p.
Finnish recovery Boiler Committee’s Recovery
Rissanen, Erkki, 1965, Economic design Boiler Day, October 22, 1997, Helsinki, Finland.
aspects of recovery boilers. Papper och Trä, No. 10,
Salmenoja, Keijo, 2004, Superheater corrosion
pp. 553 – 566.
in modern recovery boilers. 40th Anniversary
Rivers, Keith B., Roherty, Arnie, Anderson, International Recovery Boiler Conference, Finn-
Blaine A., Milbury, Cindy A., Mott, Dan C., ish Recovery Boiler Committee, Haikko Manor,
Piroozmand, Farshad, Tran, Honghi and Vafa, Porvoo, May 12-14, 2004, pp. 103 – 113.
Shermineh, 2002, Modifying boiler operation to
Salmenoja, Keijo and Tuiremo, Johanna, 2001,
reduce primary air port cracking in a recovery
Achievements in the control of superheater corro-
boiler. Proceedings of 2002 TAPPI Fall Technical
sion in black liquor recovery boilers. Proceedings
Conference, September 8 - 11, San Diego, Ca, 14
of 2001 Tappi Engineering, Finishing & Convert-
p.
ing Conference, Tappi Press, Atlanta, GA, 9 p.
Roschier, R. H., 1952, Sooda osasto. (Soda de-
Salmenoja, Keijo, 2000, Field and laboratory
partment) in Finnish, in Hiokkeen ja Selluloosan
studies on chlorine-induced superheater corrosion
valmistus, editors Pellinen H. and Roshier, R. H.,
in boilers fired with biofuels. Ph. D. thesis, Report
Suomen Paperi-Insinöörien Yhdistys, Helsinki,
00-1, Process Chemistry Group, Åbo Akademi,
pp. 359 – 456.
Turku, Finland, 102 p. ISBN 952-12-0619-5.
Roos, Tom, 1963, Senaste erfarenheter av cor-
Salmenoja, Keijo, Hupa, Mikko and Backman, furnaces and comparison with measurements
Rainer, 1999, Laboratory studies on the influence made in an experimental combustion chamber,
of gaseous HCl on the fireside corrosion of super- Gas Wärme International, Vol. 24, No. 2, pp. 64
heaters. Journal of the Institute of Energy, Vol. 72, – 75.
December 1999, pp. 127 – 133. Scott-Young, R. E. and Cukier, M., 1995,
Salmenoja, Keijo, 1998, The role of the black Commercial development of the DARS proc-
liquor recovery boiler in the closure of chemical ess. Proceedings of 1995 International Chemical
cycles in the pulp mill. Appita Journal, Vol. 52, No. Recovery Conference, Toronto Canada, CPPA, pp.
2, March 1998, pp. 88 - 92. B263 - B267.
Sandquist, Kent, 1987, Operational experi- Schunk, M., 1983. Properties of pure fluids. in
ence with single drum recovery boilers in North Heat Exchanger Design Handbook, Part 5 Physi-
America. Tappi 1987 Kraft Recovery Operations cal properties, Hemisphere Publishing Corpora-
Seminar, Orlando, Fl, January 11 – 16, pp. 221 tion, pp. 5.5.2-1 - 5.5.2-11
– 231. Sebbas, Eva, Ahonen, Aki and Haasiosalo,
Sandquist, Kent, 1987, Operational experience Taisto, 1983, Jäteliemien poltto, kemikaalien
of composite tubing in recovery boiler furnaces. talteenotto ja keittoliuosten valmistus. (Burn-
Götaverken Energy Systems, technical paper TP- ing of waste liquors, chemicals recovery and
2-87, 15 p. manufacture of cooking liquor) in Finnish, in
Saroha, Anil K., Wadhwa, Rajesh K., Chandra, Puumassan Valmistus (Manufacture of pulp),
Sandip, Roy, Maheshwar and Bajpai, Pramod, K., editor Vikola, Nils-Erik, 2nd edition, Teknillisten
2003, Soda recovery from non-wood spent pulp- tieteiden akatemia, Turku, pp. 1189 – 1300. ISBN
ing liquor. Appita Journal, Vol. 56, No. 2, pp. 107 951-99452-0-2.
– 110. Shenassa, Reyhaneh, Tuiremo, Johanna and
Saviharju, Kari and Pynnönen, Petri, 2003, Haaga, Kari, 2002, Primary air port tube integrity
Soodakattilan keon hallinta. (Control of kraft re- - a critical review of primary air port design and
covery boiler char bed) in Finnish, Finnish Recov- the effect of boiler design parameters. Proceed-
ery Boiler Users Association, Konemestaripäivä, ings of 2002 TAPPI Fall Technical Conference,
23.1.2003, Oulu, Finnish Recovery Boiler Users September 8 - 11, San Diego, Ca, 11 p.
Association, 16 p. Shiang, Niann T., 1986, Mathematical mod-
Schaddix, Chris, Holve, Don and Wessel, Rick, eling and simulation of recovery furnace. Ph. D.
2003, Recent measurements of ISP concentration, thesis, University of Idaho, Moscow.
size, and chemistry in a recovery boiler. Collo- Shiang, Niann T. and Edwards, Louis L., 1988,
quium of black liquor combustion and gasifica- Understanding and controlling fireside deposits
tion, Park City, Utah, 13-16 May 2003, 10 p. in kraft recovery furnaces. Chemical Engineering
Schmidt, Ernst, 1989, Properties of water and Technology in Forest Products Processing, AlChE
steam in SI-units: 0-800 oC, 0-1000 bar. 4th, enl. Forest Products Division, Vol. 2, pp. 105 - 114.
pr., Berlin, Springer, 206 p. ISBN: 0-387-09601-9. Shiang, Niann T. and Edwards, Louis L., 1986,
Sricharoenchaikul, Viboon, Frederick, Wm. Kraft recovery furnace modeling and simulation:
James, Jr. and Agrawal, Pradeep, 2002, Black heat transfer and gas flow. Forest Products, AIChE
liquor gasification characteristics, 1. Formation Symposium Series, pp. 85 - 94.
and conversion of carbon-containing product Singh, Preet M., Al-Hassan, Safaa J., Stalder,
gases. Ind. Eng. Chem. Res., Vol. 41, No. 23, pp. Sloane and Fonder, Greg, 1999, Corrosion in kraft
5640 - 5649. recovery boilers – in-situ characterization of cor-
Sricharoenchaikul, Viboon, Frederick, Wm. rosive environments. Proceedings of 1999 Tappi
James and Agrawal, Pradeep, 2002, Black liquor Engineering Conference, September 12 – 16,
gasification characteristics, 2. Measurement of Anaheim, California, 14 p.
condensable organic matter (tar) at rapid heating Sinquefield, S. A., Baxter, L. L. and Frederick,
conditions. Ind. Eng. Chem. Res., Vol. 41, No. 23, W., J., 1998, An experimental study on the mecha-
pp. 5650 - 5658. nisms of fine particle deposition in kraft recovery
Sricharoenchaikul, Viboon, Hicks, A. L. and boilers. Proceedings of 1998 International Chemi-
Frederick, Wm. James, Jr., 2001, Carbon and char cal Recovery Conference, Tampa, USA, 1 - 4 June
residue yields from rapid pyrolysis of kraft black 1998, Vol. 3, TAPPI, pp. 443 – 467.
liquor. Bioresource Technology, Vol. 77, pp. 131 Sjöber, M. and Cameron, J. H., 1984, AIChE
- 138. Symposium Series, Vol. 80, No. 239, p. 35.
Schupe, W. and Jeschar, R., 1975, Simplified Skrifvars, Bengt-Johan, Hupa, Mikko and
calculation of radiant heat transfer in industrial Hyöty, Paavo, 1991, Composition of recovery
References 12-19
boiler dust and its effect on sintering. Tappi Jour- 1995 International Chemical Recovery Confer-
nal, Vol. 74, No. 6, pp. 185-189. ence, Toronto Canada, CPPA, pp. A227 - A236.
Skrifvars, Bengt-Johan, 1989, Recovery boiler Stead, N. J., Singbeil, D. L. and Forget, C.,
dust sintering and dust composition. The mid- 1995, Formation of low melting deposits in a
night sun colloquium on recovery research, June modern kraft recovery boiler. Proceedings of 1995
14 - 16 1989, STFi, pp. 8 - 10. International Chemical Recovery Conference,
Smith, E. L., 1964, Kraft mill recovery units Toronto Canada, CPPA, pp. A105 – A109.
– the third generation. Paper Trade Journal, Vol. Steam its generation and use, 1992, 40th edi-
140, No. 44, p. 30. tion, editors Stultz, Steven C. and Kitto, John B.,
Soar, Roger J., Bardutz, Ronald W. and Guzi, 929 p. ISBN 0-9634570-0-4.
Charlie E., 1994, Characterizing the wastage on Steam power engineering, 1999, editor Seikan
near-drum generator tubes using full-coverage Ishigai, Cambridge University Press, New York,
ultrasonic scans. Tappi Journal, Vol. 77, No. 8, 394 p. ISBN 0-521-62635-8.
August 1994, pp. 201 - 209. Streeter, Victor L. and Wylie. E. Benjamin,
Someshwar, Arun V. and Jain, Ashok K., 1995, 1983, Fluid mechanics. McGraw-Hill Book Com-
Hydrochloric acid emissions from kraft recovery pany. ISBN 0-07-066578-8.
furnaces. Tappi Journal, Vol. 78, No. 12, December Suik, Heinrich, 2000, Catastrophic corrosion
1995, pp. 77 - 83. of superheater tubes. UEF Conference on Behav-
Sonnenberg, Gerhard Siegfried, 1968, Hun- ior of Inorganic Materials in Recovery Boilers, Bar
dert jahre sicherheit. (Hundred years of safety) in Harbor, Maine, 4 - 9 June 2000, United Engineer-
German, Technikgeschichte in Einzeldarstellun- ing Foundation, 13 p.
gen, 6, VDI-Verlag, Düsseldorf, 338 p. Sumnicht, Daniel W., 1989, A computer model
Sonnenberg, Lucinda B. and Nichols, Kenneth of a kraft char bed. Doctoral Dissertation, Atlanta,
M., 1995, Emissions of hydrochloric acid, PCDD Georgia, The Institute of paper Chemistry, 193 p.
and PCDF from the combustion of chlorine-con- Sunil, Anita, 2003, Fluidity of molten smelt in
taining kraft pulp mill bleach plant waste. Chem- recovery boilers. Web page at www.pulpandpaper.
osphere, Vol. 31, No. 10, pp. 4207 – 4223. utoronto.ca/Research/Kraft/Summaries/sunil.
Soodakattiloiden syöpymistutkimus 1965 html.
– 1968, 1968, (Corrosion study of recovery boilers Sutinen, J., Karvinen, R. and Frederick, W.
1965 – 1968) in Finnish, Final report 23.10.1968, J. Jr., 2002, A chemical reaction engineering and
Ekono, 151 p. transport model of kraft char bed burning. Indus-
Spielbauer, Thomas M. and Aidun, Cyrus, K., trial & Engineering Chemistry Research, Vol. 41,
1992, Mechanisms of liquid sheet breakup and the No. 6, pp.1477 - 1483.
resulting drop size distributions, Part 1: types of Suutela, Jukka and Fogelholm, Carl-Johan,
spray nozzles and mechanism of sheet disintegra- 2000, Heat and power co-generation. Chapter
tion. Tappi Journal, Vol. 75, No. 2, February 1992, 16 in Chemical Pulping, Book 6, series editors
pp. 136 - 142. Gullichsen, Johan and Fogelholm, Carl-Johan,
Spielbauer, Thomas M. and Aidun, Cyrus, K., Finnish Paper Engineers’ Association and TAPPI.
1992, Mechanisms of liquid sheet breakup and the ISBN 952-5216-06-3.
resulting drop size distributions - Part 2: strand Swartz, J. N. and MacDonald, R. C., 1962,
breakup and experimental observations. Tappi Alkaline pulping. Chapter 9 in Pulp and Paper
Journal, Vol. 75, No. 3, March 1992, pp. 195 - 200. Science and Technology, Vol. I, editor Libby, C.
Spielbauer, T. M. and Adams, T. N., 1990, Flow Earl, McGraw-Hill, New York, pp. 160 – 239.
and pressure drop characteristics of black liquor Sychev, V. V. et al., 1987a, Thermodynamic
nozzles. Tappi Journal, Vol. 73, No. 9, September properties of nitrogen. National Standard Refer-
1990, pp. 169 -174. ence Data Service of the USSR, A Series of Prop-
Sricharoenchaikul, V., Frederick, W. J. and erty Tables. Hemisphere Publishing Corporation.
Grace, T. M., 1995, Thermal conversion of tar to 341 p.
light gases during black liquor pyrolysis. Proceed- Sychev, V. V. et al., 1987b, Thermodynamic
ings of 1995 International Chemical Recovery properties of methane. National Standard Refer-
Conference, Toronto Canada, CPPA, pp. A209 ence Data Service of the USSR, A Series of Prop-
- A216. erty Tables. Hemisphere Publishing Corporation.
Sricharoenchaikul, V., Frederick, W. J., 341 p.
Kymäläinen, M. and Grace, T. M., 1995, Sulfur Sychev, V. V. et al., 1987c, Thermodynamic
species transformation and sulfate reduction dur- properties of oxygen. National Standard Reference
ing pyrolysis of kraft black liquor. Proceedings of
Data Service of the USSR, A Series of Property Ta- Teir, Sebastian, 2003, Steam boiler technology.
bles. Hemisphere Publishing Corporation. 307 p. 2nd ed, Energiatekniikan ja ympäristönsuojelun
Söderhjelm, Liva, Sågfors, Pär-Erik and laboratorio, Konetekniikan osasto, TKK, Espoo.
Kiiskilä, Erkki, 1998, Factors influencing heat 216 p. Electronic publication http://eny.hut.
treatment of black liquor. Proceedings of 1998 fi/library/e-books.htm
International Chemical Recovery Conference, Thomas, H.-J., 1975, Thermische kraftanlagen.
Tampa, USA, 1 - 4 June 1998, TAPPI Press, At- (Thermal powerplants) in German, 386 p. ISBN
lanta, pp. 169 – 184. 3-540-06779-5.
Söderhjelm, Liva, 1994, Black liquor proper- Thompson, Laura M. and Empie, H. Jeff, 1993,
ties. 30 Years Recovery Boiler Co-operation in A proposed mechanism for the depletion of NOx
Finland, International conference, Baltic sea, 24 in a kraft recovery furnace. Proceedings of 1993
- 26 May, pp. 23 - 36. Environmental Conference, Tappi press, pp. 643
Talbot, L., Cheng, R. K., Schefer, R. W. and – 647.
Willis, D. R., 1980, Thermophoresis of particles in Tomlinson, C. L. and Richter, F. H., 1969, The
a heated boundary layer. ASME Journal of Fluid alkali recovery system. Chapter 10 in Pulp and
Mechanics, Vol. 101, No. 4, p. 737. paper manufacture, Vol. I, The Pulping of Wood,
Taljat, B., Zacharia T., Wang X.-L., Keiser J. editors MacDonald, Ronald G. and Franklin,
R., Swindeman R. W. and Hubbard C. R., 1998, John, H., McGraw-Hill, New York, pp. 576 – 627.
Mechanical design of steel tubing for use in black Tran, Honghi, 2004, Fouling of tube surfaces
liquor recovery boilers. 9th International Sympo- in kraft recovery boilers. 40th Anniversary Inter-
sium on Corrosion in the Pulp and Paper Indus- national Recovery Boiler Conference, Finnish Re-
try, May 26 – 29, 1998, Ottawa, Ontario, Canada. covery Boiler Committee, Haikko Manor, Porvoo,
Tamminen, Tarja, Kiuru, Jukka, Kiuru, Reijo, May 12-14, 2004, pp. 91 – 101.
Janka, Kauko and Hupa, Mikko, 2002, Dust and Tran, Honghi and Arakawa, Yoshisa, 2001,
flue gas chemistry during rapid changes in the Recovery boiler technology in Japan. Proceedings
operation of black liquor recovery boilers: Part of 2001 Tappi Engineering, Finishing & Convert-
1—Dust composition. Tappi Journal, Vol. 1, No. 5, ing Conference, Tappi Press, Atlanta, GA, 13 p.
July 2002, pp. 27 - 32. Tran, Honghi, Frederick, William J. Jr., Iisa,
Tamminen, Tarja, Hupa, Mikko, Kujanpää, Kristiina, Tavares, Alarick J. and Villarroel, Rob-
Jukka and Tikkanen, Juha, 2000, An on-line meth- erto, 2000, Relationship between SO2 emissions
od for measuring transient changes in the amount and precipitator dust composition in recovery
and composition of dust in recovery boilers. UEF boilers. Proceedings of 2000 TAPPI Engineering
Conference on Behavior of Inorganic Materials in Conference, September 17-21, 2000, Atlanta.
Recovery Boilers, Bar Harbor, Maine, 4 - 9 June Tran, Honghi and Villarroel, Roberto, 1999,
2000, United Engineering Foundation, 13 p. Effect of CNCG combustion on recovery boiler
Tandra, Danny S., Kaliazine, Andrei, Corma- operation and dust composition. Proceedings
ck, Donald E. and Tran, Honghi, 2001, Numerical of 1999 Tappi Engineering/Process and product
simulation of sootblower jet and deposit interac- Quality Conference, September 12 – 16, Anaheim,
tion. 14th Annual Chemical Recovery Research California, pp.179 - 186.
Review Meeting, University of Toronto, November Tran, Honghi, Habibi, Bahman and Jia,
14-15, 2001, 3 p. Charles, 1999, Drying behaviour of waterwash
Tavares, Alarick J., Tran, Honghi and Reid, solution and the effect on composite floor tube
Timothy P., 1998, Effect of char bed temperature cracking in recovery boilers. Proceedings of 1999
and temperature distribution on fume generation Tappi Engineering/Process and product Quality
in a kraft recovery boiler. Tappi Journal, Vol. 81, Conference, September 12 – 16, Anaheim, Cali-
No. 9, September 1998, pp. 134 - 138. fornia, pp. 1061 – 1070.
Tay, Chee Vui, Kawaji, Masahiro, Tran, Tran, Honghi, Gonsko, Meredith and Mao,
Honghi and Grace, Thomas M., 2004, Recovery Xiaosong, 1999, Effect of composition on the first
boiler tube temperature excursions caused by melting temperature of fireside deposits in recov-
smelt contact. Proceedings of 2004 International ery boilers. Tappi Journal, Vol. 82, No. 9, Septem-
Chemical Recovery Conference, June 6-10 2004, ber 1999, pp. 93 - 100.
Charleston, South Carolina, USA, 543-553 p. Tran, Honghi, 1988, How does a kraft recov-
Techakijkajorn, U., Frederick, W. J. and Tran, ery boiler become plugged. TAPPI 1988 Kraft
H. N., 1999, Sintering of fume deposits in kraft re- Recovery Operations Seminar, pp. 175 - 183.
covery boilers. Journal of Pulp and Paper Science, Tran, H. N., Barham, D. and Reeve, D. W.,
Vol. 25, No. 3, pp. 73 - 83. 1988, Sintering of fireside deposits and its impact
References 12-21
on plugging in kraft recovery boiler. Tappi Journal, Proceedings 2002 TAPPI Fall Technical Confer-
Vol. 70, No. 4, pp. 109 - 113. ence, September 8-11, San Diego, Ca, 9 p.
Tran, H., Reeve, D. and Barham, D., 1983, Vakkilainen, Esa, Holm, Ralf and Simo-
Formation of kraft recovery boiler superheater nen, Liisa, 2001, Emission performance of large
fireside deposits. Pulp & Paper Canada, Vol. 84, recovery boilers. Tappi Engineering / Finishing &
No. 1, January 1983, pp. 36 - 41. Converting Conference, September 16-20, 2001,
Truelove, J. S., 1983, Furnaces and combustion San Antonio, TX, Tappi Press, pp. 118 – 127.
chambers. in Heat exchanger design handbook, Vakkilainen, Esa K., 2000, Predicting ash
HEDH, Part 3, Thermal and hydraulic design of properties in recovery boilers. UEF Conference on
heat exchangers, VDI, Düsseldorf, pp. 3.11.4-1 Behavior of Inorganic Materials in Recovery Boil-
– 9. ISBN 3-18-419082-8. ers, Bar Harbor, Maine, 4 - 9 June 2000. United
Tse, D., Salcudean, M. and Gartshore, I., 1994, Engineering Foundation, 11 p.
Multiple jet interaction with relevance to recovery Vakkilainen, Esa, 2000b, Recovery boiler.
boiler flow fields. Pulp & Paper Canada, Vol. 95, Chapter 13 in Chemical Pulping, Book 6, series
No. 2, February 1994, pp. T79 – T82. editors Gullichsen, Johan and Fogelholm, Carl-
Tsuchiya, Keiichi, Yamada, Hidetsugu and Johan, Finnish Paper Engineers’ Association and
Hosoya, Minoru, 2002, Investigation of KP opera- TAPPI. ISBN 952-5216-06-3.
tional conditions in Japanese KP mills (phase 3): Vakkilainen, Esa and Holm, Ralf, 2000, Firing
chemical recovery process (black liquor evapora- very high dry solids in recovery boilers. Proceed-
tor, recovery boiler, recausticizing, lime calcining). ings of 2000 Tappi Engineering Conference,
Japan Tappi Journal, Vol. 56, No. 2, February 2002, September 17-21, 2000, Atlanta
pp. 59-74. Vakkilainen, Esa K., Taivassalo, V., Kjäldman,
Tucker, Paul, 2002, Changing the balance of L., Kilpinen, P. and Norström, T., 1998, High
power. Solutions! February 2002, pp. 34 – 38. solids firing in an operating recovery boiler - com-
Uloth, V. C., Marcovic, C. M., Wearing, T. J. parison of CFD predictions to practical observa-
and Walsh, A., 1996, Observations on the dynam- tions in the furnace. Proceedings 1998 Interna-
ics and efficiency of sootblowing in kraft recov- tional Chemical Recovery Conference, Tampa,
ery furnaces, part II - Dynamics. Pulp & Paper USA, 1 - 4 June 1998, TAPPI Press, Atlanta, pp.
Canada, Vol. 97, No. 7, July 1996, pp. 35 – 38. 245 – 256.
Uloth, V. C., Marcovic, C. M., Wearing, T. Vakkilainen, Esa, Backman, Rainer, Forssén,
J. and Walsh, A., 1996b, Observations on the Mikael and Hupa, Mikko, 1998b, Effects of liquor
dynamics and efficiency of sootblowing in kraft heat treatment on black liquor combustion prop-
recovery furnaces, part I: Dynamics. Pulp & Paper erties. Proceedings 1998 International Chemical
Canada, Vol. 97, No. 6, June 1996, pp. 59 – 65. Recovery Conference, Tampa, USA, 1 - 4 June
1998, TAPPI Press, Atlanta, pp. 229 – 244.
Uppstu, Erik, 1995, Soodakattilan ilmanjaon
hallinta. (Control of recovery boiler air distribu- Vakkilainen, Esa, 1996, Recovery boiler
tion) in Finnish, Soodakattilapäivä 1995, Finnish adjustable air. Presentation at Spring BLRBAC,
Recovery Boiler Committee, 6 p. Atlanta, Georgia. 7 p.
Vafa, Shermineh, 2003, Effect of char bed Vakkilainen, Esa, Frederick, William J., Reis,
dynamics on recovery boiler primary air port V. V. and Wåg, Kaj. J., 1995, Potassium and chlo-
tube temperature excursions. Web page at www. rine enrichment during black liquor combustion.
pulpandpaper.utoronto.ca/Research/Kraft/Sum- Laboratory and mill measurements and a mecha-
maries/vafa.html. nistic model. Proceedings 1995 International
Chemical Recovery Conference, Toronto, CPPA,
Vakkilainen, Esa K., Kankkonen, Sebastian
pp. B63 - B76.
and Suutela, Jukka, 2004, Advanced efficiency op-
tions - increasing electricity generating potential Vakkilainen, Esa, 1994a, Limits of recovery
from pulp mills. Proceedings of 2004 International boiler technology. 30 Years Recovery Boiler Co-
Chemical Recovery Conference, June 6-10 2004, operation in Finland, International conference,
Charleston, South Carolina, USA, pp. 1 - 12. Baltic sea, 24 - 26 May.
Vakkilainen, Esa K., 2004, Future of recovery Vakkilainen, Esa, 1994b, High black liquor
boiler technology. 40th Anniversary International solids changes operation of recovery boiler. Paperi
Recovery Boiler Conference, Finnish Recovery ja Puu (Paper and Timber), Vol. 76, No. 8.
Boiler Committee, Haikko Manor, Porvoo, May Vakkilainen, Esa and Niemitalo, Hanna,
12-14, 2004, pp. 51 – 58. 1994, Measurement of high dry solids fouling and
Vakkilainen, Esa K. and Pekkanen, Markku J., improvement of sootblowing control. Proceed-
2002, Burning excess streams in recovery boilers. ings of 1994 Tappi Engineering Conference, San
Francisco, California. devolatilization - application of modeling results
Vakkilainen, Esa and Anttonen, Timo, 1993, to understanding laboratory data. Proceedings of
Effect of economizer central pass design on 1995 International Chemical Recovery Confer-
thermal performance. Proceedings of 1993 Tappi ence, Toronto, Canada, CPPA, pp. B89 - B103.
Engineering Conference, Orlando, Florida, Tappi Verrill, Christopher L. and Wessel, Richard A.,
Press. 14 p. 1995, Model of alkali release during black liquor
Vakkilainen, Esa and Vihavainen Esa, 1992, combustion. Journal of Aerosol Science, Vol. 26,
Long term fouling of recovery boiler surfaces. September 1995, pp. S171 - S172.
Proceedings of 1992 International Chemical Verveka, Peter, Nichols, Kenneth M., Horton,
Recovery Conference, Seattle, Washington, June Robert R. and Adams, Terry N., 1993, The form of
7 - 11. nitrogen in wood and its fate during kraft pulping.
Vakkilainen, Esa, Adams, Terry N. and Hor- Proceedings of 1993 Environmental Conference,
ton, Robert R., 1992, The effect of bullnose designs Tappi press, pp. 777 – 780.
on upper furnace flow and temperature profiles. Villarroel, Roberto, Gonçalves, Cláudia and
Proceedings of 1992 International Chemical Tran, Honghi, 2004, Experience of screen tube
Recovery Conference, Seattle, Washington, June damage caused by falling deposits in kraft recov-
7-11, pp. 101 - 112. ery boilers. Proceedings of 2004 International
Vannérus, Torbjörn, Svensson, Gustav and Chemical Recovery Conference, June 6-10 2004,
Forslund, Folke, 1948, Ångpannor. (Steam genera- Charleston, South Carolina, USA, pp. 341 - 349.
tors) in Swedish, Chapter 19, in Maskinteknik, Viteri, João Arlindo, 2001, Recentes rejetos de
Ingenjörshadboken 2, Nordisk Rotogravyr, Stock- novas caldeiras de recuperação no Brasil. (Recent
holm, Sweden, pp. 937 – 1072. projects of new chemical recovery boilers in Bra-
Vasara, Petri, 2001, Through different eyes: zil) in Portuguese, 14th Latin-American Recovery
Environmental issues in Scandinavia and North Boilers meeting, August 7 – 10, Sao Paulo, Associ-
America. Tappi Journal, Vol. 84, No. 6, June 2001, ação Brasileira Technica de Celulose e Papel. 10 p.
pp. 46 - 49. Välimäki, Erkki and Salmenoja, Keijo, 2004,
VDI heat atlas, 1993, Verein Deutscher Ing- Measured data for sootblowing optimization.
enieure, VDI-Gesellschaft Verfahrenstechnik und Proceedings of 2004 TAPPI Fall Technical Confer-
Chemieingenieurwesen, Düsseldorf, VDI-Verlag, ence, October 31 - November 3, Atlanta, Georgia,
ISBN 3-18-400915-7. 13 p.
Vegeby, Anders, 1966, Scandinavian practices Wagner, Wolfgang and Kruse, Alfred, 1998,
in the design and operation of recovery boilers. Properties of water and steam the industrial
Tappi Journal, Vol. 49, No. 7, p. 103. standard IAPWS-IF97 for the thermodynamic
properties and supplementary equations for other
Vegeby, Anders, 1961, Synpunkter på indun-
properties tables based on these equations. Berlin,
stning till högre torrhalt – inverkan i sodahuset.
Springer, 354 p., ISBN 3-540-64339-7.
(Viewpoints of high dry solids evaporation – effect
to recovery operation) in Swedish, Sodahuskonfe- Wallén, Jonas, Ruohola, Tuomo and Aikio,
rens, ÅF-IPK, Stockholm, pp. 60 – 69. Anne, 2004, Sulfur dioxide emission dependency
on kraft recovery boiler operation parameters.
Verloop, Arie and Jansen, Johan H., 1995,
Proceedings of 2004 International Chemical
Operating a recovery boiler at reduced solids
Recovery Conference, June 6-10 2004, Charleston,
throughput. Proceedings 1995 International
South Carolina, USA, pp. 483-490.
Chemical Recovery Conference, Toronto, CPPA,
pp. B1 – B4. Wallén, Jonas, Viiala, Juhani and Björkström,
Ingvar, 2002, Operational performance of Billerud
Verrill, Chris and Lien, Steven, 2003, The
Gruvon’s new recovery boiler with environmental
significance of char bed-generated particulate on
friendly solutions. Proceedings of 2002 TAPPI
recovery boiler fouling relative to other sources of
Fall Technical Conference, September 8 - 11, San
combustion areosols. Proceedings of Colloquium
Diego, Ca, 10 p.
of black liquor combustion and gasification, Park
City, Utah, 13-16 May 2003, 9 p. Walsh, Alan R. and Srandell, Olof, 1992, Use
of design and operating data to determine recov-
Verrill, Christopher L. and Wessel, Richard
ery boiler upgrade potential. Proceedings of 1992
A., 1998, Detailed black liquor drop combustion
International Chemical Recovery Conference,
model for predicting fume in kraft recovery boil-
Seattle, Washington, June 7-11, pp. 421 – 428.
ers. Tappi Journal, Vol. 81, No. 9, September 1998,
pp. 139 - 148. Walsh, Alan R. and Grace, Thomas M., 1989,
TRAC: A computer model to analyze the trajec-
Verrill, Christopher L. and Wessel, Richard A.,
tory and combustion behavior of black liquor
1995, Sodium loss during black liquor drying and
References 12-23
droplets. Journal of Pulp and Paper Science, Vol. vari, Vesa, 1995, Liquor to liquor differences in
15, No. 3, May 1989, pp. J84 - J89. combustion and gasification processes pyrolysis
Warnqvist, Björn, Delin, Lennart, Theliander, behavior and char reactivity. Proceedings of 1995
Hans and Nohlgren, Ingrid, 2000, Technoeco- International Chemical Recovery Conference,
nomical evaluation of black liquor gasification Toronto, CPPA, pp. A245 - A253.
processes. Technical report, S6-614, Värmeforsk Wozniak, John C., 1985, Measurements of the
Service AB, June 2000, ISSN 0282-3772. 37 p. physical properties of kraft smelt. The Institute of
Warnqvist, Björn, Ihrén, Niclas and Berglin, Paper Chemistry, Appleton Wisconsin. 69 p.
Niklas, 1994, Combined power plant technol- Wåg, Kaj J., Reis, Victor V., Frederick, William
ogy in the recovery cycle of kraft pulp mills. 30 J. and Grace, Thomas M., 1997, Mathematical
Years Recovery Boiler Co-operation in Finland, model for the release of inorganic emissions dur-
International conference, Baltic sea, 24 - 26 May, ing black liquor char combustion. Tappi Journal,
pp. 95 – 106. Vol. 80, No. 5, May 1997, pp. 135 – 145.
Warnqvist, Björn, 1994, Char bed and smelt Wåg, Kaj J., Grace, Thomas M. and Kymäläin-
properties mill studies. Pulp and Paper Canada, en, Marita, 1995, Sulfate reduction and carbon
Vol. 95, No. 8, pp. 313 – 315. removal during kraft char burning. Proceedings of
Wessel, Richard. A. and Kaufman, Keith C., 1995 International Chemical Recovery Confer-
2000, Computer modeling of kraft recovery boil- ence, Toronto, Canada, CPPA, pp. B35 - B50.
ers. UEF Conference on Behavior of Inorganic Zaman, A. A. and Fricke, A. L., 1995, Effects
Materials in Recovery Boilers, Bar Harbor, Maine, of pulping conditions and black liquor composi-
4 - 9 June 2000. United Engineering Foundation, tion on the heat of combustion of slash pine black
16 p. liquor. in Advances in pulping and papermaking,
Wessel, Richard A., Denison, Martin K. and AIChE Symposium Series, Vol. 91, No. 307, pp. 154
Samretvanich, Artit, 1998, The effect of fume on – 171. ISBN 0-8169-0687-4.
radiative heat transfer in kraft recovery boilers. Zeldovich, J., 1946, The oxidation of nitrogen
Proceedings 1998 International Chemical Recov- in combustions and explosions. Acta physiochimi-
ery Conference, Tampa, USA, 1 - 4 June 1998, ca U.R.S.S., Vol. 21, No. 3, pp. 577 – 628.
TAPPI Press, Atlanta, pp. 425 – 441. Zevenhoven, Ron and Kilpinen, Pia, 2001,
Wessel, R. A. and Righi, J., 1998, Generalized Control of pollutants in flue gases and fuel gases.
correlations for inertial impaction of particles on Picaset Oy, Espoo, 298 p. Electronic publication
a circular cylinder. Aerosol Sci. Technol. (AAAR), http://www.hut.fi/~rzevenho/gasbook.
No. 9, pp. 29-60.
Westerberg, E. Norman, 1983, Kraft mill
recovery units examined. Pulp and Paper Interna-
tional, March 1983, pp. 53 – 54.
Whitney, Roy P., 1968, Introduction. Chapter
1 of Chemical recovery in alkaline pulping proc-
ess, editor Whitney, Roy P., TAPPI Monograph
series, No. 32, Mack Printing Company, Easton,
Pa., pp. 1 – 14.
Whitty, Kevin and Verrill, Christopher L.,
2004, A historical look at the development of
alternative black liquor recovery technologies and
the evolution of black liquor gasifier designs. Pro-
ceedings of 2004 International Chemical Recovery
Conference, June 6-10 2004, Charleston, South
Carolina, USA, pp. 13 - 34.
Whitty, Kevin, Backman, Rainer, Forssén,
Mikael, Hupa, Mikko, Rainio, Janne and Sor-
vari, Vesa, 1997, Liquor-to-liquor differences in
combustion and gasification processes: pyrolysis
behaviour and char reactivity. Journal of Pulp and
Paper Science, Vol. 23, No. 3, March 1997, pp. J119
- J128.
Whitty, Kevin, Backman, Rainer, Forssén,
Mikael, Hupa, Mikko, Raino, Janne and Sor-
Index
acid dew point 10-6 oxidation 1-8, 5-12, 7-21, 9-19
acidic sulfate 10-5 piping 7-14, 9-20, 9-23
adiabatic combustion temperature 2-10, 9-20 spraying 4-1, 4-8, 9-1, 9-2, 9-6
air swelling 4-1, 4-4, 4-8, 4-10
distribution 7-2, 7-4, 9-9 virgin dry solids 2-9, 3-2, 4-4, 8-16
emission 11-1 viscosity 4-9, 9-4
fan 7-3 blackout 5-8, 9-27
flow control 7-2 bleach plant effluent 2-9, 9-25
heater 7-2, 7-3 boiler bank
intake 7-2 design 6-15, 10-18
level 1-4, 4-8, 7-3, 7-6, 11-4 spacing 6-15
multilevel 7-6 boiling heat transfer coefficient 6-2
opening 7-5 boiling point 9-20
primary 7-4, 9-10 boiling point raise 9-2
secondary 7-5, 9-11 Brownian diffusion 8-10, 8-11
sticky 8-4 bullnose 2-7, 6-12, 6-13, 9-11
tertiary 7-5, 9-12 burner 7-15
air system 1-4, 1-7, 7-2, 7-3 NCG 9-13
four level 1-9 oil 7-20
multilevel 1-9 startup 7-20
multiple level 1-4 cadmium 11-14
tangential 1-8 capacity 6-7, 6-7, 6-8, 6-9, 9-20
three level 1-8, 2-5 carbon release 5-18
vertical 1-9 carbonate
airport 1-8, 7-3, 7-5, 10-9 decomposition 5-12
cleaning system 7-16 reduction 5-11
corrosion 10-9 carryover 6-7, 8-7, 9-11
placement 7-6 cascade evaporator 1-6, 1-7, 7-21
spacing 7-4 causticizing 1-1
antimony 11-14 char 4-4, 4-6
arsenic 11-14 combustion 4-6, 4-10, 5-19, 8-6, 9-9
ash composition 4-6
conveyor 7-19 char bed 1-4, 5-2, 5-3, 5-4, 9-10
mixing 1-7 cooling 5-7
purging 1-1 heat content 5-7
sticky 8-5, 8-14, 8-21, 9-12, 9-19 oxygen concentration 5-11
system 7-15 chemical reaction 5-2
attemperating 7-9 chill-and blow 8-21
auxiliary fuel 9-13, 9-26, 11-2 chloride 8-11, 8-15
balance concentration 8-18
energy 3-6 release 5-17
mass 3-2 chloride removal 8-23
barium 11-14 chromium 5-10, 10-3, 10-16, 10-7, 11-14
beryllium 11-14 chromizing 10-14
biosludge 9-25 circulation 1-6, 7-8, 10-13
black liquor forced circulation 1-3
air for combustion 3-1 CNCG 2-9, 7-7, 9-24, 9-25
burning 5-1 CO 4-6, 5-2, 9-11, 9-14, 11-2
char burning 5-2 CO2 1-10, 4-6, 5-12, 9-13, 11-3
devolatilization 4-3, 5-1, 9-8 cobalt 11-14
drying 4-1, 4-2, 4-8, 5-1 condensate 7-7, 7-11, 9-26, 10-10
high dry solids 4-8, 9-22 combustion 1-2, 2-10, 5-1
inorganics 2-9, 3-5, 5-8, 7-16 devolatilization 4-3
net heating value 1-1 drying 4-1, 4-2, 4-8
organics 1-1, 2-9, 5-1, 9-19 high dry solids 4-9, 4-10
Index I-1
laboratory scale 4-8 dry solids
pyrolysis time 4-8 effect on dimensioning 6-21
stages 4-1 effect on steam generation 3-8
compound tubing 1-6, 7-10, 10-13 duct 7-4
copper 10-10, 11-14 dust 9-14
corrosion early recovery technology 1-2
acid 10-11 economizer 1-7, 7-9, 10-16
airport 10-8 cross flow 10-16
alkali 10-4 design 10-16
caustic 10-10 fouling 8-3
chloride 10-5 heat transfer coefficient 6-6
control 10-11 horizontal 1-7
flow assisted 10-11 vertical 1-7
gas side 10-2 efficiency
general 10-7 steam generation HHV 3-8
high temperature 10-3 steam generation LHV 3-8
molten hydroxide 10-5 electrostatic precipitator 1-5, 7-18
oxygen 10-10 elemental composition 3-1, 3-9
smelt 10-7 emission 9-13, 11-1
sulfidation 10-2 carbon dioxide 11-3
thermal fatigue 10-8 carbon monoxide 11-2
crack 9-12, 10-6, 10-8, 10-15 dioxins 11-12
cross sectional area 6-7 dissolving tank 11-14
current recovery boiler 2-6 dust 11-8
cyclone evaporator 1-6, 1-7, 7-21 fluorine 11-14
cyclone separator 7-9, 7-13 furan 11-14
damper 7-3 HCl 11-11
guillotine 7-6 heavy metal 11-14
deaerator 7-8, 7-11, 10-10 nitrogen oxides 11-3
decanting furnace 1-6 reduced sulfur species 11-2
delivery time 2-12 sulfuric acid 11-12
deposit sulphur dioxide 11-10
hardening 8-13 volatile organic compounds 11-7
melting 8-11 energy balance 3-6
strength 8-14 enrichment 8-14
desuperheating 6-5 BB dust 8-17
dew point 10-6, 10-19 chloride 8-15, 10-7
devolatilization 4-3, 5-11, 9-8 ESP dust 8-16
dimethyl disulfide 4-5 factor 8-14, 8-15
dimethyl sulfide 4-5, 4-10, 4-12 potassium 8-14, 10-7
direct contact evaporator 1-4, 1-6 sulfur 10-7
direct alkali recovery system 1-11 entrainment 8-7
dissolving tank 1-3, 2-5, 7-7, 7-16, 7-20 EPRS 6-7
explosion 7-20 erosion 10-9
stack 7-20 evaporation 1-1, 6-2
vent gas 7-7 evaporator
DNCG 2-7, 7-7, 9-25 direct contact 1-4
Dolezahl 7-11, 7-12 horizontal 1-5
downcomer 1-6, 6-5, 7-9, 7-13 excess air 3-2, 7-4, 9-12
draft 7-3, 7-8 explosion 9-22, 9-25
forced 7-3 fan
induced 7-8, 7-17 air 7-17
droplet 1-4, 4-1, 4-2, 4-7, 5-4, 9-2, 9-4, 9-7 flue gas 7-8
drying 4-2 induced draft 7-17
flight path 9-5 feedwater 10-11
size 9-8 pump 7-11
drum 1-6, 2-4, 7-9, 7-13 system 7-10
single 1-6 tank 7-10, 7-11
two 1-8, 2-4 fin 10-14, 10-18
I-2 Kraft recovery boilers
firing effect on fouling 8-16
pressure 9-3 hoppers 7-15
stationary 1-8 HSL 2-6, 6-7, 6-11
temperature 9-3 hydrochloric acid 5-18
first melting temperature 8-17 hydrogen sulfide 4-5, 4-12, 5-15, 10-2
fixed carbon 4-6, 5-13 oxidation 5-16
flame oven 1-2 h-p diagram 6-4
flashing 9-5, 9-7, 9-20 ID fan 7-8
floor 10-12 igniter 7-15
construction 10-13 infiltration 3-5, 7-4
decanting 1-6 inertial impaction 8-2, 8-9, 8-10
sloping 1-6 inorganic matter 4-1
tube 1-4, 1-6, 10-6 intermediate size particles 8-6, 8-8, 8-10, 11-9
flow temperature 8-12, 8-16 interlacing 1-9, 7-5
flue gas iron 11-14
system 7-8 jet penetration depth 6-8
fly ash 7-14 key design specifications 2-2
forced draft fan 7-2, 7-7 lead 11-14
fouling 1-5, 8-1 lignin 5-16
predicting 8-2 lime kiln 1-1, 1-11, 4-10, 9-24
fume formation 5-3, 5-5, 8-5, 8-6, 8-7, 11-8, 11-9 liquor gun 9-2
furnace 1-3, 1-5 liquor heat treatment 3-9, 4-9
brick lined 1-6, 10-13 liquor heater 2-7, 6-2, 7-15
compound tube 1-6, 7-0, 10-13 lower furnace 6-2, 10-12
decanting 1-6 main steam
design 10-12 pressure 2-1, 2-8
height 6-7 temperature 2-8, 10-14
materials 10-12, 10-13 major design features 2-2
membrane wall 1-6, 7-10, 10-9, 10-13 make up chemicals 7-15
outlet temperature 6-10 manganese 11-14
requirements 5-1 magnesium 10-10, 11-14
roof 10-17 material balance 3-1
Sanicro 1-7 membrane 1-6, 7-10, 10-9, 10-13, 10-15
sizing 6-6 mercury 11-14
sloping botttom 1-6 methanol 5-20, 7-7, 9-22
studded 1-6, 10-13 methyl mercaptane 4-5, 4-12
tangent 1-6, 10-13 mixing 1-9, 5-4, 5-18, 6-8, 7-5, 9-11
wall 1-6, 10-12 modern recovery boiler 2-4
gasification 1-10 natural circulation 6-2, 6-5
generating bank 6-15 NCG
gold 11-14 CNCG 2-9, 7-7, 9-24, 9-25
governor shaft 7-6 DNCG 2-7, 7-7, 9-25
green liquor 7-16 HVLC 7-7
hanger rods 6-14, 10-18 LVHC 7-7
HCl 5-18, 5-19, 8-21, 9-19 nickel 11-14
header 6-15, 7-9, 7-10 niobium 11-14
square 1-6 nitrogen
hearth heat release rate 6-7, 6-9 in black liquor 9-13
hearth solids loading 6-7, 6-9 release 4-13, 9-13
heat loss non process element 1-1
radiation and convection 3-8 NOx
heat transfer emission 9-13, 11-3
convection 6-19 fuel 9-13
radiation 6-16 prompt 9-14
steam side 10-15 reduction 11-4
heating value 3-1, 3-5 thermal 9-14
heavy metals 9-14 nozzle
HHRR 2-6, 6-7, 6-8, 9-17 fan 9-2
high dry solids 1-7 splash plate 9-2
Index I-3
swirl cone 9-2 secondary air 7-5, 9-11
U-type 9-2 model 9-11
V-type 9-2 selenium 11-14
willow flute 9-2 silver 11-14
oil sintering 1-11, 8-6, 8-14, 8-15, 8-21, 8-22
burner 7-15 sintering index 8-15, 8-21
system 7-15 single drum 1-6, 2-2, 2-3, 2-5, 6-1
organic matter 4-1 smelt
oxidation 9-37 composition 5-8
particulate matter 11-6 flow 5-8
permit heat 7-16
environmental permit 2-11 inventory 5-7
zoning requirements 2-11 properties 5-9
phosphor 5-5 smelt spout 1-4, 2-5, 5-8, 7-10, 7-16, 7-20
pitting 10-7, 10-9, 10-10 smelt-water explosion 5-7, 7-9, 9-25
platen 6-14 SO2 11-1, 11-6, 11-10
plugging 8-2, 8-3 soap 1-1, 9-24, 9-25
polysaccharides 4-9, 4-11 sodium
polysulfate 5-14 bicarbonate 8-3
port rodder 7-16 bisulfate 10-5
potassium 5-13, 8-24 metallic 5-12
release 5-23 release 4-5, 4-10, 5-10
precipitator 1-4, 7-14, 7-18, 8-6 vaporization 5-12
pressure drop sodium carbonate 4-6, 5-2, 5-10, 5-11, -18, 7-14,
gas side 6-18 8-3, 8-13, 8-17, 9-12
pressure sulfation 5-18
main steam 1-10, 2-1, 2-8 sodium chloride 5-8, 5-18
pressure vessel code 2-11 sodium sulfate 5-2
primary air 1-6, 1-8, 2-5, 5-4, 7-3, 9-10 decomposition 5-13
projecting recovery boiler 2-11 sootblower 1-5, 7-17, 7-18
pump sootblowing 8-21, 8-22
feedwater 7-11 specific
steam turbine driven 7-11 air requirement 7-2
purchasing 2-11 flue gas production 3-5
pyrolysis 4-3 steam production 2-10
pyrosulfate 10-5 spray water 7-12
quatenary air 9-15 group 7-12
radiation 2-3, 4-3, 6-12, 6-16, 10-16 spraying 1-4, 4-1, 4-8, 9-1, 9-2, 9-6
raiser tubes 6-5 state of the art boiler 2-7
recovery boiler steam
evolution of design 2-3 generation 1-1, 1-7, 2-4, 2-9, 3-6, 3-8, 6-3, 6-21
process 1-2 steam drum 1-5, 1-6, 2-3, 2-4, 7-12, 7-13, 10-18
purposes 2-1 erection 7-14
recycle 7-14 sticky temperature 8-12, 8-17
reduction 5-1, 5-8, 5-11 stoichiometric 3-2
efficiency 5-8, 9-12 studding 10-12
in green liquor 5-8 studs 10-14
reduction rate 4-6, 5-1 sulfate-sulfide cycle 5-11
refractory 1-6, 10-12, 10-14 sulfidity 5-8
reoxidation 4-7 sulfur
residence time 6-7 dew point 10-6
ring header 7-14 emission 9-18
rotary oven 1-3 release 4-5, 4-12, 5-14, 9-6, 9-22
safety 1-9, 2-6, 5-7, 5-8, 6-5, 7-11, 7-15, 9-24 sulfurdioxide 5-2, 5-15, 11-1, 11-6
saturated superheater 7-9, 10-4
steam 6-2 connection tube 7-11
water 6-2 corrosion 9-12, 10-15
screen 2-2, 2-3, 6-2, 6-14 design 10-15
screen plate 7-20 fouling 8-1, 8-2, 8-5, 8-21
materials 10-16
panel 6-14
spacing 6-13
tube arrangement 10-17
superheating 6-3
area 6-12
heat required 6-4
heat transfer coefficient 6-3
sweet water condenser 7-9
swelling 4-1, 4-4, 4-8, 4-10
tall oil 9-24
tangent tube 1-6, 10-13
tank
dissolving 7-16, 7-20
feedwater 7-11
tellurium 11-14
temperature
boiling 9-2
gradient 8-8
liquor firing 9-7, 9-20
main steam 2-8, 10-15
spikes 10-7
tertiary air 7-5, 9-12
thallium 11-14
thermophoresis 8-8
thiosulfate 5-14, 5-15
TRS 7-20, 9-12, 9-18
turpentine 7-7, 9-24
two drum recovery boiler 2-4
valve 7-18
double 7-15
safety 7-16
spray water 7-12
vanadinium 11-14
vapour condensation 8-10
venting 7-10, 7-16
vertical flow economizers 1-7, 2-5
virgin black liquor dry solids 2-9, 3-2, 4-4, 8-16
viscosity 4-9, 9-4
VOC 11-7
wall drying 1-6
water
circulation 7-8
preheating 6-3
wash 1-5, 10-7
windbox 7-4, 7-6
zink 11-14
Index I-5
A Emission conversions
Figure A-1, Typical emissions from recovery boiler.
Figure A-2, Same emissions ordered from large to small and scaled so average equals 100.
A.1 CONVERSION OF AVERAGE

EMISSIONS TO TIME NOT TO
EXCEED When converting average data to peak data at least
the corrections in the Table A-1 should be applied.
Recovery boiler emissions values are normally E.g. Average SO2 emission is 150 mg/MJ. Value for
based on continuous operation. Short-term values not to exceed for more than 15 min/24h for SO2 is
exceed these average values. then 2,00*150 mg/MJ = 300 mg/MJ.
Emission conversions A-1

Table A-1, Multiplicator for conversion from average to time not to exceed.
Time CO CxHy TRS SO2 NOx Dust
24 h 100% 1,00 1,00 1,00 1,00 1,00 1,000
8 h/24 h 67 % 1,05 1,03 1,02 1,02 1,01 1,000
1 h/24 h 95 % 2,00 1,70 1,30 1,40 1,05 1,300
30 min/24 h 98 % 2,50 2,20 1,40 1,60 1,07 1,500
15 min/24h 99 % 3,00 2,50 1,60 2,00 1,10 3,000
A.2 CONVERSIONS where

EY is the emission at dry flue gases and
The following equations can be used to convert oxygen content Y, mg/m3n
emissions from one value base to another. EZ is the emission at dry flue gases and oxygen
content Z, mg/m3n
Emission from mg/m3n to mg/MJ O2Y is the corresponding oxygen content Y at
dry flue gases, vol-%
E[mg/MJ] = 0.24 n k E[mg/m3n] A-1 O2Z is the corresponding oxygen content Z at
dry flue gases, vol-%
where
E[mg/m3n] is the emission at dry flue gases and The same equation can be used to change ppm at
air ratio n, mg/m3n one O2 level to another O2 level at dry flue gases.
E[mg/MJ] is the emission per heat input, mg/MJ The emission as mg/MJ at one O2 level is the same
n is the air ratio at which emission as at another O2 level (no correction needed).
[mg/m3n] is given
k is a correction factor to correct
emission to lower heating value, -
Emission form ppm to mg/m3n
and Emission at dry or wet flue gases can be converted
H (100-x)
to mg/m3n at the same O2 level and at dry or wet
k = qHHV/(qHHV-21.987 /100-2.443 /x) A-2 conditions respectively by multiplying with a coef-
ficient. It is customary that many emissions can
where be expressed also as equivalent species (e.g. sulfur
H is the black liquor hydrogen content emissions are often expressed as S or as SO2).
as fired, w-%
qHHV is the black liquor higher heating 1 ppm H2S = 1.535 mg(as H2S)/m3n
value as fired, MJ/kgds 1 ppm H2S = 1.441 mg(as S)/m3n
x is the black liquor dry solids as fired, 1 ppm CO = 1.253 mg(as CO)/m3n
w-% 1 ppm NO2 = 2.053 mg(as NO2)/m3n
1 ppm NO = 2.053 mg(as NO2)/m3n
Example 1 ppm NO = 1.229 mg(as NO )/m3n
1 ppm NOx = 2.053 mg(as NO2)/m3n
NO 260 mg/m3n at 1.2, HHV = 14.0, ds = 75%, H 1 ppm NOx = 1.375 mg(as 95% NO
= 3.5%. and 5 % NO2)/m3n
k = 14.0/(14.0-21.987*3.5/100-2.443*((100-75)/75) 1 ppm SO2 = 2.926 mg(as SO2)/m3n
= 14.0/(14.0-0.77-0.81) = 1.13 1 ppm SO2 = 1.461 mg(as S)/m3n
NO[mg/m3n] = 260*1.2*1.13*0.24 = 85 mg/MJ 1 ppm HCl = 1.628 mg(as HCl)/m3n
1 ppm HCl = 1.583 mg(as Cl)/m3n
Emission form one O2 level to 1 ppm NH3 = 0.771 mg(as NH3)/m3n
another O2 level 1 ppm NH3 = 2.053 mg(as NO2)/m3n
1 ppm CH4 = 0.719 mg(as CH4)/m3n
A good approximation of air ratio n is 1 ppm H2SO4 = 4.375 mg(as H2SO4)/m3n
1 ppm CH4S = 2.148 mg(as CH4S)/m3n
n = 20.9/(20.9-O2) A-3 1 ppm CH4S = 2.860 mg(as SO2)/m3n
1 ppm CH4S = 1.428 mg(as S)/m3n
O2 is the oxygen content at dry flue gases, vol-% 1 ppm C2H6S = 2.774 mg(as C2H6S)/m3n
1 ppm C2H6S = 2.860 mg(as SO2)/m3n
EY = EZ (20.9-O2Y)/(20.9-O2Z) A-4 1 ppm C2H6S = 1.428 mg(as S)/m3n
A-2 Kraft recovery boilers

KRBFull

Uploaded by

Copyright:

Available Formats

KRBFull

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

KRBFull

Uploaded by

Copyright:

Available Formats

Kraft recovery boilers -

Principles and practice

Principles and practice

This book was created as postgraduate lecture

The purpose of this text is to give the reader an

To understand the processes happening in the

Many thanks to Marja and Ari Heinola, who not

In 1989 Esa Vakkilainen found himself involved

ii Kraft recovery boilers

iv Kraft recovery boilers

Figure 1-1, Kraft mill unit operations.

Recovery of black liquor has other advantages. 1.1 FUNCTION OF RECOVERY

There are other minor operations to ensure con-

Principles of kraft recovery 1-1

which is used to generate electricity in a turbine. green liquor

Combustion in the recovery boiler furnace needs

1-2 Kraft recovery boilers

Tampella was among the first manufacturers to

The main recovery equipment itself remained

The boilers parts were improved on. In 1930s even

Hand operated recovery grew more complicated

Use of rotary oven improved the heat economy.

Principles of kraft recovery 1-3

Recovery boiler were soon licensed and produced

The construction utilized was very successful.

The first recovery boilers had severe problems

Principles of kraft recovery 1-5

Sanicro 38 is a widespread material that offers

1.5 HIGH DRY SOLIDS 1.6 IMPROVING AIR SYSTEMS

Air system Main target But also should

1-8 Kraft recovery boilers

ground. Use of about 50 % secondary seemed to

Fourth generation air systems are the multilevel

Principles of kraft recovery 1-9

recovery boiler at Iwakuni mill (Ohtomo, 2000).

Table 1-2, Effectiveness of converting black liquor

1-10 Kraft recovery boilers

Other much studied component has been tita-

Na2CO3 + NaBO2 → Na3BO3 + CO2 13

1-12 Kraft recovery boilers

One noticeable trend has emerged in recent years.

By year 2000 more than ten recovery boilers,

Recovery boiler design 2-1

Black liquor dry solids ﬂow is the key design cri-

Figure 2-4, Design features of Babcock & Wilcox

2-2 Kraft recovery boilers

Single drum covered with deposits.

Recovery boiler design 2-3

Property Two drum Modern Current

Two drum recovery boilers are constructed with

Currently the two drum boilers start to get modi-

Modern recovery boiler

The eﬀect of increasing dry solids concentration

A modern recovery boiler, Figure 2-7, consists of

2-6 Kraft recovery boilers

Figure 2-11, Main steam temperature as a function of recovery boiler main

2-8 Kraft recovery boilers

- Further increase in black liquor solids Main steam temperature

Recovery boiler design 2-9

Figure 2-15, Speciﬁc steam generation kgsteam/kgBLdry solids as function