CIP/CIL/CIC Adsorption Circuit Process Selection

Chris A. Fleming'

ABSTRACT
There are three basic carbon-adsorption processes used in the gold mining industry today, carbon-
in-pulp (CIP), carbon-in-Ieach (CJL), and carbon-in-columns (CIC). The first two processes
recover gold directly from pulps or slurries containing lip to 55% solids, while the Cle process is
used for treatment of solution, usually from a heap leaching operation. In choosing between CIP
and CIL, a number of factors need to be taken into consideration. The capital cost of a CIL plant
will generally be lower than CIP, but the process is not suited to all ore types, and operating costs
may be higher in CIL, particularly for slow leaching gold ores.

INTRODUCTION
The hydrornetallurgical process for the treatment of gold and silver ores remained unchanged for
the first 70 years of the twentieth century, and consisted essentially of leaching in cyanide solution
followed by solid-liquid separation, with the solid residue being washed as efficiently as possible,
and the leach liquor being treated by zinc cementation to recover the precious metals. This
became known as the Merrill Crowe process. While Merrill Crowe is generally extremely
efficient and fairly cheap, it does have limitations in the treatment of low-grade ores and certain
complex ore types. Fer example, ores with a high content of clay or other soft, fine minerals are
usually difficult to filter, and losses of soluble gold or silver in the residues can be unacceptably
high. In addition, ores with high concentrations of base metal oxides, particularly copper. are net
well suited to the zinc cementation process because of high consumptions of cyanide and zinc.
The ability (If activated carbon to adsorb gold from cyanide solution has been known for over
100 years (Davis, 1880). However, it wasn't until the 1950's that carbon-based processes began to
attract attention in the gold mining industry, when it was demonstrated by the US Bureau of
Mines that gold and silver could be recovered directly from pulp onto granules of activated
carbon, thereby bypassing the costly and inefficient solid-liquid separation steps, and when
methods of eluting the gold from the carbon were developed (Zadra, 1950, Zadra et aI., 1952).
The first full scale CIP plant was installed at the Homestake Mine in South Dakota, USA, in
the early 1970's [Hall, 1974], but the pace of change was fairly sluggish until the 1980's. By the
end of that decade, however, almost the entire gold mining industry worldwide had become
activated carbon disciples, to the extent that practically all new gold properties that were
developed during that decade incorporated carbon adsorption in the flowsheet. Total gold
production from the primary resource gold mining industry was about 1,300 tone in 1970, of
which90% (-1)00 tons), was recovered from cyanide leach solutions by cementation on zinc
dust, (the remaining - J 0% was recovered directly from products such as gravity concentrates and
copper concentratesj.vlsy 1990; total world gold' production, (from the primary sector), had
increased to 2,200t, but by this time only -30% (GOOt),was recovered by zinc cementation, while
-15% was smelted directly from primary gravity and copper concentrates, and a massive 55%
(1,200t), was recovered by the new processes of adsorption cnto activated carbon.

I Lakefield Research Limited. Lakefield. Ontario. Canada

1644
Two decades is a remarkably short period of time for a new process to assume a position of
such dominance, especially in an industry as historically conservative as the gold mining industry.
The reasons for this were, first, the fact that the early prediction of superior economics compared
to zinc cementation generally proved to be correct and. second, the carbon-based processes
proved to be extremely versatile and robust, both mechanically and chemically. Carbon-based
processes were applied to almost any feed, achieving higher recoveries of soluble gold, at lower
capital cost and generally lower operating cost than zinc cementation. The first half dozen CIP
plants, five of which were built in South Africa, treated everything from slirnes fractions to whole
ores, flotation concentrates, flotation tailings, roasted calcines, biologically' oxidized float
concentrates, re-pulped filter-plant residues and old tailings. ClP and ClL plants in North
America, Australia and the Pacific Rim treated feeds ranging from low-grade, open-pit ore bodies
to high grade flotation concentrates that had been' pre-oxidized by pressure leaching, while
adsorption on activated carbon became the process of choice for gold recovery from heap leach
.. '
solutions .
" Not only were carbon-based processes shown to accommodate a wide variety of feed
materials. the process was also shown to be robust through temporary plant upsets, such as
changes in head grade, surges in feed rate or losses in carbon activity, suffering only minor losses
in gold extraction efficiency. This was an important requirement for a new process being
introduced to an industry that had been lulled by 90 years' experience of the metallurgically
simple and chemically forgiving Merrill Crowe process. Most gold plant operators before CIP
were mechanically skilled but chemically unsophisticated, and this proved to be an adequate skiII-
set for the CIP process in the early days.
The new-technology pendulum probably swung too far, in that some projects that went for
carbon in the 1980's might have been better-suited to the zinc cementation process. There is now
a clearer understanding of the relative strengths and weaknesses of zinc and activated carbon, and
there is a more open-minded approach to flowsheet evaluation in the development of new gold
and silver ore bodies. For example, it is well known that the affinity of activated carbon for gold is
much greater than fur silver, and that ores that contain either high silver concentrations or a high
silver to gold ratio, are probably better suited to zinccementation than carbon adsorption. In
addition. because ef the relatively high cost of carbon elution and' reactivation, as well as
limitations in the extent to which gold can be loaded onto activated carbon (technical and
economical), it is also preferable to treat very rich gold cyanide solutions by zinc cementation,
rather than with activated carbon.

CARBON ADSORPTION PROCESSES

Carbon adsorption processes have become universally accepted as the standard method for gold
and silver recovery from solutions or pulps. The only exceptions, as pointed out above, are
solutions or pulps that, contain:

hign Ag tc Au ratios
very high Au concentrations
SPecies that interfere with carbon adsorption (typically organics),

There are threebaslC carbon-adsorption processes used in the gold mining industry today:

• carbon-in-pulp (CIP),
• carbon-in-leach (Cll.)
• carbon-in-columns (CIC).

The first two processes recover gold directly from pulps or slurries containing up to 55%
solids, while the elC process is used for treatment of solutions (typically from heap leach

1645
operations). The main differences between pulp and solution treatment are physical, and relate to
the separation and recovery of activated carbon granules from pulp or solution, after gold
adsorption has occurred. In CIP and CIL, this is done by screening in a wide variety of
mechanical or static (air-cleaned) screening devices, which have a screen mesh size that allows
the gold-depleted pulp to pass through, while retaining the carbon granules. In CIC adsorption, on
the other hand, the gold and silver-bearing solution generally flows upward through a series of
columns that are packed with activated carbon, at a flowrate that is sufficient to lightly fluidize the
bed of carbon, but not so fast that the carbon flows from the top of the column. Therefore,
physical separation of the carbon from the solution is not an issue. In all three processes (CIP,
CIL, CIC), solution or pulp flows by gravity from one contactor to the next, in a cascade of
between 4 and 8 adsorption tanks, while the carbon in the tanks is periodically pumped in a
countercurrent direction to the flow of solution or pulp.

CIP and CIL versus Merrill Crowe

The fundamental difference between CIP or CIL and the zinc-cementation process lies in the fact
that, in order to recover gold by reduction on zinc dust, it is first necessary to produce a clarified
filtrate, whereas in CIP and CIL it is possible to extract gold cyanide directly from the slurry. In
this way, the costly and inefficient solid-liquid separation processes of a Merrill Crowe gold plant
are replaced by the relatively simple and inexpensive screening procedures that are used in CIP
and CIL to recover the carbon granules from the leach slurry.
Early estimates of the capital and operating costs associated with the CIP process (Potter et
aI., 1974, Bhappu et aI., 1974, Bailey, 1987), indicated that savings of 20 to 50 percent could be
expected, compared to the conventional process, and these estimates have been vindicated, for the
most part, on the great many CIP and CIL plants operating around the world today. This has
meant that lower-grade gold ores can be economically treated than would previously have been
the case, and this was a major stimulus to new gold mine development around the world in the
latter part of the 20'h century, leading to an annual increase in gold production of about 3% from
1970 (-1,300 tons total), to 2000 (-'2,800 tons). Another important benefit or CIP and Cll.
processes is the generally improved efficiency of gold extraction that can be achieved. This stems
from two factors, Firstly, gold losses from a filtration or CCD plant usually amount to about 1
percent of the dissolved gold in the pregnant solution (i.e., 0.03 to 0.05 mg/L) , in the case of
easily filterable solids, while even higher soluble losses are suffered in the case of poorly filtering
or settling solids. By contrast, soluble losses of less than 0.01 ppm can be achieved on a well-
managed CIP or CIL plant. Secondly, the additional 5 or 6 hours leaching time in the CIP
adsorption contactors usually results in extra gold dissolution. This extra dissolution can be
considerable if there are preg-robbing constituents in the ore. Evidence for extra dissolution is
obtained by an analysis of the washed solids in the feed and the discharge from a CIP plant. In
most instances, gold concentrations in the residues are lower than in the feed solids, despite the
presence of fine abraided carbon from the CIP tanks in the residue.
The final advantage of the CIP and CIL processes is that they are far less vulnerable than zinc
cementation to impurities such as sulphide, arsenate and antimonate in the leach liquor, while
parameters such as cyanide concentration, pH and oxygen concentration do not have to be as
carefully controlled, because their influence on the loading of gold OD carbon is fairly minor.

CIC versus Merrill Crowe

While the major cost advantage of processes using activated carbon over zinc cementation lies in
the ability of granular carbon to extract gold and silver directly from pulps, there are also
compelling arguments for using carbon rather than zinc for the recovery of gold and silver from
clarified or unclarified solutions.

1646
The choice between carbon columns and zinc cementation is based on analyses of capital and
operating costs and consideration of the metallurgical efficiency. As a broad generalization,
treatment in carbon columns is more economical for large volumes of low-grade solutions
containing mainly gold, whereas zinc cementation is preferred for relatively small volumes of
high-grade solution, particularly those rich in silver.
Obvious examples of large-volume, low-grade solutions are tailings ponds of existing gold
plants where low concentrations of gold and silver are present in solution, either because of
inefficient solid/liquid separation and recovery in the main plant, or because of additional leaching
that has occurred in the tailings dam. There are a number of plants around the world recovering
gold from sources such as this, in packed carbon columns, and the payback period for this sort of
scavenging operation can be as short as a few weeks! Another example of a situation in which
carbon is usually preferred te zinc is in the treatment af scluiions from heap-leaching operations.
Here too, the main advantage of carbon over zinc cementation is that the pregnant solution need
not be clarified or deaerated prior to feeding to a carbon column. Another Important advantage of
carbon over zinc for heap leaching is the fact that the carbon process is "clean" and does not
introduce new ions into the recirculating leach liquor. In the cementation process zinc cyanide,
which is a by-product of the zinc cementation reaction, builds up in solution with continuous
recycling of barren solutions to the heap. In extreme cases, the build up of zinc must be controlled
by removing a bleed stream of leach liquor and treating it to destroy cyanide and precipitate zinc.
At best, the build up of zinc in heap leach liquors adds significantly to the time and cost of
washing and detoxifying a spent heap. In fact, even when Merrill Crowe has been selected as the
method of recovering gold and silver during a heap leach operation, carbon columns are often
installed at the end of the mine's life, to recover precious metals from the wash solutions. This is
because of the need to have free cyanide in solution for effective Merrill Crewe operation ( which
. would defeat the objective of wasting cyanide from the exhausted heap). Activated carbon does
not suffer from this problem, because the loading of gold and silver on carbon actually improves
as the cyanide in solution decreases ( Fleming, 1984).
For these and other reaSO!1S,elC has become the standard process for the recovery of gold
from solutions generated by heap leaching.

CIP versus CIL

The difference between CIP and Cll, lies in the extent to which the gold and silver are leached
from the ere prior to carbon adsorption. In eIP, most of the leachable gold and silver (typically
>98%) have been leached before the first carbon adsorption stage, and gold adsorption is carried
out in specially designed tanks that are situated after the leaching tanks in the overall process
flow. The adsorption tanks are typically a quarter to a tenth the size of the leaching tanks. In CIL,
carbon is added to the leaching tanks; and adsorption occurs simultaneously with leaching. In
some CIL plants, there are ne or two stages of leaching before the first CIL adsorption tank, and
the extent of gold leaching before carbon adsorption can be as high as 95% - in other words, not
much different from CIP. In other Cll, plants, carbon is present in all the leaching tanks, and both
the leaching and adsorption processes occur simultaneously throughout.
In choosing between CIP and CIL for a particular application,a number of factors must be
weighed up.

Advantages of CIL

The major advantage of ClL over CIF is the lower capital cost ef the plant, which results
from the removal of five or six CIP adsorption tanks and ancillary equipment from the
flowsheet. The additional operating costs associated with the CIP section would also be
eliminated. ,-

1647
Because CIL leaching tanks are much bigger than CIP adsorption tanks, the pulp surface
area available for interstage screening is much greater in CIL, and screening is less of a
bottleneck in the process.

Disadvantages of CIL

• On the basic assumption that the total inventory of carbon in CIL and CIP plants would
be the same (for the same overall metallurgical performance), the carbon concentration in
the pulp would be up to 5 times lower in CIL than in CIP. Hence, more pulp is
transferred upstream in ClL, and back-mixed during countercurrent transfer of the
carbon, which leads to metallurgical inefficiencies and places an increased load on the
carbon advance pumps and the interstage screens. In a Cll, plant, up to 25% cf the pulp
flow could be back-mixed with countercurrent carbon advancement, whereas in CIP
back-mixing will generally be less than 5% of the total pulp flow.
•. The kinetics of gold adsorption on carbon art slower in CIL than in CIP, because
leaching is incomplete when the pulp encounters carbon in the adsorption stages of CIL,
and the concentration of gold in solution is lower. To compensate for a decrease in the
kinetic driving force, more carbon is needed to match the metallurgical performance
achieved in CIP. Even for an ore from which the gold leaches at a relatively fast.rate, the
carbon inventory in CIL should be 10 to 15 per cent higher than in CIP, for the same
metallurgical performance.
• CIP has a subtle advantage over CIL in that extra carbon adsorption tanks can be
retrofitted into an existing eIP plant at relatively low cost, by simply converting the last
few leach tanks to adsorption tanks. Many of the early Clf' plants were designed on the
basis of data from laboratory tests and small pilot plants and, for a number of reasons, the
full scale plants tended to perform below design specification on startup. In such caSES, it
was pcssible to bring the plant performance back in line with design specification by
increasing the. inventory of carbon in the plant or increasing the number of adsorption
stages, A CIP plant obviously has far more flexibility than a CIL plant in this regard,

General
The mechanisms by which gold and silver cyanide load onto carbon, as well as their equilibrium
loading capacities, are believed to be the same for solutions and pulps, although the rate of
loading is significantly faster from solution than from pulp. This is because the average size of the
carbon granules is usually smaller in CIC - owing to the different screening requirements, which
leads to faster mass transfer kinetics. This typically allows cle plants to operate at 50 to 100
percent higher gold (and silver) loadings than CIF or CIL plants, for the same concentrations of
.gold and silver in the leach solution, and the same contact time between solution and carbon. The
processing methods for eluting gold and silver from carbon and regenerating the carbon for
recycling, are also the same for CIP, CIL and CIC.

OPTIMIZATION OF ADSORPTION
It is now well established that the gold cyanide ion (Au[CN];) loads on to' activated carbon by a
reversible.adsorption mechanism, without undergoing any chemical change, probably as an ion
pair [Adams, et al, 19~jQ,Fleming and Nicol, 1984]. Verification of the reversibility of adsorption
was achieved in an experiment in which loaded carbon from the first stage of a CIP plant was
isolated in a basket m the last stage of the same CIP plant. [Fleming and Nicol, 1984] Gold on the
carbon in the basket slowly desorbed (with tll2 -48 hours), in contact with the low gold tenor
solutions of the pulp in the last CIP contractors (Figure 1).

1648
It is expected that gold on very fine carbon granules would desorb far more rapidly than
depicted in Figure 1, and an important practical consequence of this is that fine carbon generated
by abrasion in a CIP circuit will fortuitously desorb most of its gold values before discharging
with the barren pulp from the circuit.
On the other hand, coarse carbon granules will des orb slower (III2 -48 hrs), than the average
residence time of pulp in a CIP or Cll, contacter (1 to 6 hours), so loss of coarse carbon (through
screen breakage or other mechanical failures), could result in significant gold loss.
The kinetic characteristics of aurocyanide adsorption and desorption on activated carbon are
quite unique in fact, and stem from the unusual physical properties of activated carbon. For
example, activated carbon is a highly porous substrate, with an enormous total surface area of
over 1000 rr/ per gram of carbon. Evidence suggests that as much as 90% of this surface area is in
very small micropores (i.e, with a crass sccuonal diameter of the pores of less than 100 A), and as
such are either inaccessible to the fairly large aurocyanide molecule, or becomes accessible only
after tortuously slow diffusion. This property helps to explain the observed kinetic phenomena in
gold cyanide processing.

4000 . ~ - Slope t% = 48 hr

Gold on
Carbon
(glt)
3500 i ~

3000

,
r,:'

;"'-
25i()0
0 10 20 30

Time (hr)

Figure 1: Themodynamic aspects of adsorption

Reversibility'of adsorption

PI2.nt experience and numerous laboratory investigations have shown that the initial rate of
aurocyanide adsorption onto carbon is fairly rapid, and that.it is controlled by the hydrodynamics
in the adsorption contactor. This initial, film-diffusion controlled reaction, which presumably
involves adsorption of gold cyanide molecules in the carbon macropores (which can be as large as

1649
00,000 A in diameter), continues for 24 per 48 hours.' Subsequently, gold cyanide continues to be
adsorbed slowly onto the carbon almost indefinitely and. in practice, the establishment of a true
equilibrium is difficult.
During the latter period of pore-diffusion-controlled adsorption, gold cyanide presumably
diffuses slowly into the carbon micropores, and, as the cross-sectional area of the micro-pores
approaches that of the aurocyanide ion, the resistance to mass transfer becomes infinite.
From a practical point of view, therefore, the aurocyanide-carbon interaction can be
considered to possess two thermodynamic regimes: the rnacropore-rnesopore pseudo equilibrium
and the total equilibrium, and the value of the former could be 10 times lower than the value of
the latter.
This aspect of the kinetics of gold cyanide loading or. activated carbon is illustrated in Figure
2. which depicts the result of an experiment conducted at the Western Areas CIP plant in the early
1980's, (Ni col et al., 1984).
In this experiment, fresh activated carbon was placed in an agitated basket inside the pulp of
the first CIP adsorption stage of the plant. The basket was made of 35 mesh stainless steel
screening material, which allowed slurry to flow through, but isolated the carbon in the basket
from that in the bulk of the adsorption tank. The experiment continued for 3 weeks, during which
time samples of carbon were frequently taken from the CIP tank and from inside the basket, and
analyzed for gold. The data in Figure '2 depicts the analytical data. .
The: loading of gold on the carbon in the CIP tank cycled upwards and downwards from
-2,000 to -4,000 g/t (with an average of -3,000 g/t), as carbon was periodically advanced in and
out of the tank during normal CIP operations. The gold on the carbon in the basket, however,
continued to increase over the 3 week period, reaching -1 i,000 g/t by the end of the test.
The gold grade in solution in the feed to the Western Areas CIP plant at that time was -2.5
mg/L, so the experiment showed that an upgrading ratio of 5,000: 1, can oe achieved.
In practice, the operating conditions in most CIP and ClL plants are set to achieve an upgrading
ratio in the range 750: 1 to 1,500: 1, which is similar to that shown for the plant operating
conditions in Figure 2. This range of up-grading ratio will likely be the correct operating strategy
in most cases. Not only is the rate of extraction relatively fast and responsive to good mixing
efficiency ill the adsorption contractors, but the rate of elution and ultimate elution efficiency
under a given set of conditions are also maximized. In addition, lock up of gold on carbon in the
adsorption plant is minimized, and the risk of gold losses through screen breakages or carbon theft
is also minimized.
However, the data in Figure 2 shows that these operating conditions are quite conservative,
and much higher gold loading could be achieved, which would benefit process economics by
reducing the amount of carbon reporting to the expensive carbon elution and regeneration
processes. If higher gold loadings on carbon were targetted for an operating CIP or Cll, plant, it
would be necessary to increase the inventory and concentration of carbon in the adsorption tanks
to compensate for the slower kinetics of loading at higher concentrations of gold on carbon, This
could become a limiting factor in CIP adsorption tanks, but not in CIL.
Finally, the unusual kinetic loading properties depicted in Figure 2 explain the rugged
efficiency of CIP, ,CIL and CIC processes. and their sluggish response to plant upsets. The fact
that the gold on the carbon is way below the equilibrium loading for each of the adsorption stages,
allows carbon adsorption plants to be operated very efficiently (most of the time), with low levels
of process control and phut surveillance.

2 This initial rate of loading responds to factors such as pulp density, pulp viscosity, mtxtng
efficiency, carbon particle size, and the presence, in the pulp, of species that adsorb onto and
poison the carbon.

1650
••
•• 11 •

Gold on 9000
Carbon
(g1t)
6000
Average

3000

100 200 300 400 500

Time (hr)

Figure 2 Kinetic aspects of adsorption

REFERENCES
Adams, M.D & Fleming, C.A. 1989. The mechanism of adsorption of aurocyanide onto activated
carbon, Metallurgical Transactions B, Vo!. 20, pp. 315-325.
Bailey, P.R. 1987. Application of activated carbon to gold recovery. In: 0.0. Stanley (Editor),
The Extractive Metallurgy of Gold in South Africa. S. Afr. Inst. Min. Metall., Monogr, Ser.,
M7,379-614.
Bhappu. R.H., Lewis, M.P. & McAl1ister, lA. 1474. Leaching of low-grade ores, economic
evaluation of available processes. AIME Annu. Meet. (Dallas, Texas).
Davis, W.N. 1880. Use of carbon for the precipitation of gold from solution and subsequent
burning. US Pat. 227,963.
Fleming, C.A. & Nicol, MJ. 1984. The adsorption of gold cyanide onto activated carbon. III.
Factors influencing the rate of loading and equilibrium capacity. IS Afr Inst Min Metall,
Vol. 84, No. 4, pp.85-93.
Hall, K.B~ 1974. Hornestake uses carbon-in-pulp to recover gold from slimes. World Mining,
27(12),44.
Nicol, M.L Fleming, C.A. & Crornberge, G. 1984. The adsorption of gold cyanide onto activated
carbon. I. The kinetics of adsorption
from pulps. J; South African Inst. Mining Metal!', 84(2),50-4.
Potter, O.M. & Salisbury, H.B. 1974. Innovations in gold metallurgy. Min. Congr. J., 54.
Zadra, J.B. 1950. A process for the recovery of gold from activated carbon by leaching and
electrolysis. USBM RI t!·672.
Zadra, J.B., Engel, A.L. & Heinen, H.J. 195?. Process for recovering gold and silver from
activated carbon by leaching ami electrolysis. USBM RI 4843.

1651