Co-processing renewables in a

hydrocracker
Understanding and controlling the impact on catalyst performance, yields, and product
properties is key to renewables co-processing in hydrocracking units

Peter Andreas Nymann and Pronit Lahiri

Topsoe

T
he demand for having part of the product slate renewable feedstock should cause severe performance
from refineries consisting of renewable materials is deficits in the hydrocracker, co-processing in the hydro-
increasing. Hydrotreating units have been the pri- cracker might not be feasible.
mary choice so far, but hydrocracking units are also quite Understanding the impact on the performance of the
suitable and will be unavoidable in the future, especially hydrocracking unit when introducing new feedstock
when processing more demanding second- and third-gen- types, such as first-, second- and third-generation renew-
eration feedstocks. able feedstocks, is therefore of utmost importance for
Extensive pilot plant testing and a fundamental under- this evaluation.
standing of chemistry supported by industrial operation
from several units worldwide have put Topsoe in a position Feedstocks and chemistry
to support refineries in this challenging effort. Topsoe has In recent years, the most applied renewable feedstocks
been collaborating with refineries navigating through the in refinery units have been triglycerides. These typically
energy transition to reduce CO2 emissions originating from originate from what is referred to as virgin oils, more com-
the use of fossil-based feedstocks. monly known as first-generation feedstocks. More recently,
second-generation feedstocks like used cooking oil (UCO),
Renewable feedstock challenges palm oil mill effluent (POME), and others have been con-
The world energy sector is going through an energy transi- sidered. Second-generation renewable feedstocks are typi-
tion and looking for solutions to decarbonise energy produc- cally byproducts from the production of virgin oils. Going
tion. Part of this is facilitated by replacing fossil feedstocks forward, third-generation feedstocks like pyrolysis oils
with renewable resources and treating these in the existing derived from municipal sewage waste, byproducts from the
refinery framework to produce fuels that are fully compat- paper and pulp industry, wood, plastic, or tyres are being
ible with the hydrocarbon fuels in the market today. investigated.
Processing of renewable feedstocks, like first-generation The triglycerides are characterised by consisting of three
feedstocks such as palm oil and other triglyceride-based glycerol chains with a typically even carbon number and
hydrocarbons, has therefore been widely applied in recent a backbone of propane connected by oxygen atoms (see
years. Several projects constructing new standalone units Figure 1). Hydrotreating reactions include saturation of
or revamping existing hydroprocessing units to either double bonds inside chains, breaking of triglyceride into pro-
co-process or solely process renewable feedstocks have pane from the backbone and paraffins from the side chains,
materialised. and the reaction of oxygen atoms with hydrogen, forming
Co-processing renewable feedstocks in hydrotreating water, CO or CO2. They occur readily over hydrotreating
units is now well established, and refineries are looking for catalysts and are well known in several hydroprocessing
more ways to include renewable feedstocks in their exist- units worldwide.
ing units. Furthermore, the increased demand for includ- Many second- and third-generation feeds are more com-
ing renewable feedstocks in the energy system and the plex than the first-generation feeds and hold more complex
availability of renewable feedstocks directs the industry molecules. Consequently, they result in a different boiling
to incorporate second- and third-generation renewable range of the products as well as different properties. Unlike
feedstocks. This introduces new and more demanding the triglycerides that generate straight chained molecules
requirements to convert these challenging feedstocks into boiling in the distillate boiling range, the second- and third-
on-spec fuels. generation feedstocks contain ring structures with different
One way is to co-process renewables in hydrocracking heteroatoms that need more hydrogenation and eventually
units. The hydrocracker has advantages and disadvan- cracking to bring them into the transportation fuel boiling
tages when it comes to co-processing renewables. If the range and to ensure that these products fulfil other product
economic uplift from processing renewable feedstocks is specifications, such as cold flow properties and densities.
lower than that of normal hydrocracker feedstock, or the Hydrotreating of triglycerides requires significant

www.digitalrefining.com Catalysis 2023 25

Triglyceride molecule
Better cetane index
O Paraffins Higher cloud point

H2C O
Backbone – O
Formation of CO, CO2 & H2O
propane Oxygen
HC O Higher H2 consumption
formation O
9 12

H2C O 3
Ole-sat gives higher H2 consumption
Olefins
Higher heat release

Figure 1 First-generation renewable feedstocks mainly consist of triglycerides that consist of a propane backbone linking
the paraffin chains with oxygen bonds. Breaking of the oxygen bonds will result in yields of paraffins, propane, and
oxygen-containing molecules like water, CO and CO2

amounts of hydrogen, and the reactions are highly exo- therefore be less. Likewise, the associated heat release
thermic. The product will be boiling in the middle distillate originating from these reactions is also more in the same
range and suitable for diesel. If jet fuel is needed, the lon- order of magnitude as in hydrocrackers designed to handle
ger normal paraffin chains need to be converted to lower high temperature rises by having multiple beds.
boiling material by either cracking or isomerisation. Normal Contaminants present in the renewable feeds pose a
paraffins have low density and high cetane index, but also challenge to the hydrotreaters, and significant amounts
high cloud point. Isomerisation provided by a hydrocrack- of bulk catalysts need to be replaced to manage the con-
ing catalyst is therefore sometimes needed to achieve cloud taminants. In hydrocrackers, the VGOs and HCGOs often
point specifications. processed also have considerable amounts of contami-
Hydrotreating of aromatics and oxygen-containing ring nants, and hydrocrackers are therefore normally designed
structures produce naphthenes with higher density and to include volume for grading catalysts. They may therefore
boiling range than what is suitable for diesel. Selective allow greater quantities of renewables to be processed
hydrocracking reactions involving ring opening will there- without compromising catalyst cycle length.
fore be needed for processing these more demanding Different types of renewable feeds may produce differ-
second- and third-generation feedstocks to produce high ent renewable products, and the fractionation section in
yields of high-quality jet and diesel products. hydrocrackers makes it possible to separate these differ-
ent products. The products from renewables often require
Overcoming operational challenges more than hydrotreating, either to achieve better cold flow
Operational challenges are better overcome in hydrocrack- properties by isomerisation or, in the case of second- and
ers. Considering the high content of oxygen in renewable third-generation feedstocks, hydrogenation of rings and
feeds, the H2 consumption associated with these reactions ring opening. Hydrocracking catalysts are multifunctional
is in the range of 300-500 Nm3/m3, which is orders of mag- and may be selected to include both isomerisation and
nitude greater than what is normally seen in hydrotreaters. selective ring opening activity to ensure the fulfilment of all
However, compared to hydrocrackers, the difference is less, required end-product properties.
and the impact from processing in a hydrocracker would
Real-world examples
Several hydrocrackers worldwide are processing renewable
Vegetable oil, 0% Vegetable oil, 5%
Feed C, Feed C, feedstocks. Some hydrocrackers are part of standalone units
Feed A,
3% Feed A,
36%
3%
36% comprising hydrotreating, isomerisation, and hydrocracking
sections, and others are conventional fossil fuel hydrocrack-
ers co-processing renewable feed batches. Topsoe has
Planned throughput supported refineries in preparing for and conducting test
Planned throughput
during fossil run,
co-processing run, runs, and the following discussion gives examples of the
vol%
vol% observations from one of these test runs showing industrial
data supporting the theoretical basis previously discussed,
confirming learnings from pilot plant studies.
Feed B, Feed B,
61% 56% Test run co-processing vegetable oil in a European refinery
The refinery operates a single-stage once-through hydro-
Figure 2 Unit throughput was kept stable during the test cracking unit, designed for 80 wt% gross conversion of
run processing typical feedstock for baseline comparison. various low-value gas oil streams to high-quality products.
5 vol% of the feedstock was slowly replaced with vegetable The primary objective of the unit is to maximise middle dis-
oil and then all unit operational parameters were kept stable tillate yield. In 2021, the refinery considered co-processing
to evaluate performance during co-processing bio-based vegetable oil feedstocks in the hydrocracking unit.

26 Catalysis 2023 www.digitalrefining.com

As a first preparation step, pilot plant tests were conducted
in the refinery’s research facility using Topsoe catalysts HC Bed 1 HC Bed 2
installed in the industrial unit. Pilot plant results were anal-
PTR Bed 2
ysed and interpreted in comparison with estimates made by PTR Bed 3
PTR Bed 1
Topsoe. The refinery then decided to conduct a test run in HC Bed 1
HC Bed 2

the hydrocracking unit using 5 vol% refined palm oil feed- PTR Bed 3
PTR Bed 2
stock co-processed during normal operation (see Figure 2).
At operating conditions of the hydrocracking unit, the Fossil run Co-processing run

hydrotreating of triglycerides in vegetable oil are relatively PTR Bed 1 5˚C

easy reactions and are expected to occur over the first NiMo
pretreat catalyst bed (see Figure 3).
The weighted average bed temperatures (WABT) for Figure 3 The recorded temperatures show that almost all
pretreat and cracking beds show that catalytic severity reactions related to triglycerides in the vegetable oil occur
was lower during the co-processing run. This is in line with in the first pretreat bed during the co-processing run.
expectations due to the following factors: Thereafter, the temperature profile in the catalyst beds is
 Effect over pretreat catalysts: Vegetable oil replaces almost identical for fossil and co-processing runs
VGO feed. Pretreat reactions on triglyceride are easier com-
pared to HDS, HDN, and HDA reactions on VGO molecules composition of diesel stream in the following product prop-
that are more complex. erties discussion.
 Effect over hydrocracking catalysts: Pretreat effluent  Unconverted oil (UCO) yield/gross conversion:
during co-processing has a higher number of C16-C18 mol- Vegetable oil replaces VGO in feed during the co-pro-
ecules from triglyceride side chains. These boil in the distil- cessing run. Molecules in vegetable oil are relatively easy
late fraction. Hence, the true conversion over hydrocracking to treat at hydrocracking unit conditions. Hence, the feed
catalysts is lower during co-processing compared to the during the co-processing run is relatively easy to process
fossil run to achieve a similar UCO yield/gross conversion compared to 100% VGO feed. Even at directionally lower
(see Figure 4). catalytic severity, UCO yield is lower, i.e., gross conversion
is higher during the co-processing run
PTR WABT HC WABT  Off-gas make: Higher off-gas flows are recorded dur-
Fossil run Base Base ing the co-processing run. As explained in the feedstocks
Co-processing run Base – 1°C ↓ Base – 1°C
and chemistry section, C3 is expected from the triglycer-
ide backbone, and C1 is expected from the methanation
Yield structure reaction. These are also consistent with higher C1 and C3
Mass balance during both fossil and co-processing runs recorded in the treat gas composition during the co-pro-
closed well within the accepted tolerance for commercial cessing run.
hydrocracking units (see Figure 5). Hence, the yield struc-
ture could be evaluated in more detail with a good degree Product properties
of confidence. Vegetable oil feed has predominantly C16 and C18 chain
The following are salient points from the fossil and co- length side chains. These are expected to fully saturate to
processing runs: n-paraffins and boil in the middle distillate product fraction.
 Hydrogen consumption: Co-processing vegetable oil The following plots compare chromatograms of distillate
consumes additional hydrogen as expected and explained in samples recorded during the test run. The peaks in the
the previously discussed feedstocks and chemistry section. chromatograms are representative of the concentration of
 Diesel yield: Diesel yield is higher during the co-pro- molecules boiling at the specified temperature.
cessing run. This is an indication that C16 and C18 molecules Following are the theoretical boiling points of isomers of
expected from the triglyceride side chain in the feed are varying chain lengths:
retained in the diesel fraction. See further discussion on the

2 wt% Fossil run Co-processing run

Co-processing run

Fossil run 5 wt%FF

True conversion Gross conversion

Figure 4 Even at lower catalytic severity during Naphtha Kero Diesel UCO
co-processing test run, higher gross conversion was
recorded, further validating the two factors mentioned in Figure 5 Mass flows recorded during the test run as a
WABT discussion above percentage of fresh feed processed

www.digitalrefining.com Catalysis 2023 27

1.4 Biogenic carbon Fossil carbon
Biogenic carbon, 2%

1.2

1
Fossil run, Co-processing
kero product run, kero product
0.8
wt%

0.6

Fossil carbon,
0.4 98%

0.2 Figure 7 Result from ASTM D6866 test for distinguishing

carbon isotopes in kerosene sample drawn during the test
0 run
100 150 200 250 300
Temp, ˚C
no carbon-14 isotope). In this application, the amount of
Kero product - 100% fossil run Kero product - 5% co-processing run
biogenic carbon in the product fractions can be directly
correlated to the relative amount of product resulting from
Figure 6 Chromatogram of kerosene product sample vegetable oil during co-processing.
drawn during the test run Results from the biogenic carbon analyses in the prod-
uct fractions are discussed in conjunction with the product
Chain length Boiling point, °C chromatograms in Figures 6, 8 and 10.
nC14 254
nC15 271 Kerosene product
nC16 287 Interpreting a simulated distillation curve as a probability plot
nC17 303
iC18 303 is shown in Figure 6, where the temperature on the x-axis
nC18 317 refers to the average boiling point of lumps of material boil-
nC19 330 ing in 1°C intervals, making identification of different quanti-
ties of specific components possible. We can, for example,
Product samples were also analysed for biogenic carbon infer from this chromatogram that there was a higher con-
content by ASTM D6866. This is done by distinguishing centration of nC15 (BP 271°C) in the kerosene product dur-
carbon isotopes from biogenic sources (containing well- ing the co-processing test run compared to the fossil run.
characterised amounts of carbon-14 isotope) against car- The presence of carbon sourced from vegetable oil during
bon from fossil sources (containing carbon-12 isotope and co-processing is further validated by biogenic carbon analy-
ses showing a 2% carbon-14 isotope (see Figure 7).

1.8 Diesel product

Two of the most well-defined peaks can be seen in the
1.6
diesel chromatogram in Figure 8, corresponding to nC16
1.4 (BP 287°C) and nC18 (BP 317°C). The other two smaller
peaks correspond to nC15 (BP 271°C) and nC17 / iC18
1.2 (BP 303°C).
1
Biogenic carbon, 9%
wt%

Biogenic carbon Fossil carbon

0.8

0.6

0.4
Fossil run, Co-processing
diesel product run, diesel product
0.2

0
100 200 300 400
Temp, ˚C
Diesel product - 100% fossil run Diesel product - 5% co-processing run Fossil carbon,
91%

Figure 8 Chromatogram of diesel product sample drawn Figure 9 Result from ASTM D6866 for distinguishing
during the test run carbon isotopes in diesel sample drawn during the test run

28 Catalysis 2023 www.digitalrefining.com

1.2
Biogenic carbon Fossil carbon

0.8 Co-processing
Fossil run,
UCO product run, UCO product
wt%

0.6

0.4

Figure 11 Result from ASTM D6866 for distinguishing

0.2 carbon isotopes in unconverted oil sample drawn during
the test run
0
250 300 350 400 450 500 550 600
consumption, temperature rises, and contaminants, as well
Temp, ˚C
as separate the renewable products into different product
UCO product - 100% fossil run UCO product - 5% co-processing run
streams.
Hydrocracking catalyst options also provide the func-
Figure 10 Chromatogram of unconverted oil sample drawn tionality needed to make products with desired properties,
during the test run such as good cold flow properties or high degree of satu-
ration, as demonstrated with the recent industrial test run
Biogenic carbon analyses of diesel from fossil and co- co-processing 5 vol% vegetable oil in a hydrocracking unit
processing runs validated significant retention of vegetable using Topsoe catalysts. The bulk catalysts already loaded in
oil feed in diesel product fraction, as shown in Figure 9. the reactors were found adequate for co-processing mod-
erate quantities of renewable feedstock.
Unconverted oil Most exothermic hydrotreating and hydrodeoxygen-
Unconverted oil product during both fossil and co-process- ation reactions occur in the first/grading bed. There was a
ing runs appears to have similar boiling range distribution. slight increase in exotherm in this bed, but this could eas-
This is also validated by biogenic carbon analyses that show ily be handled by reducing the reactor inlet temperature.
no presence of carbon-14 isotope during co-processing Thorough product analyses showed that the kero/jet prod-
(see Figure 10). uct stream had 2% biogenic carbon, and the diesel stream
The most significant UCO properties recorded during the had 9% biogenic carbon during the co-processing test run.
test run bear testimony to the lower true conversion over Most of these carbon molecules are n-paraffins, as noted
the hydrocracking catalyst during co-processing. Typically, by chromatogram peaks.
it is seen that the UCO density and viscosity reduce with Activity requirement during co-processing was lower
increasing conversion severity (see Figure 11). The higher compared to normal operation, as noted by lower pretreat
density and higher viscosity of UCO during the co-process- WABT, as well as higher gross conversion during co-pro-
ing run are an indicator of lower conversion severity on the cessing, indicating feed during co-processing is easier to
VGO part during this operation. treat compared to typical fossil-based VGO feed in this unit.
This test run has helped establish base-line performance
Test Method Fossil run Co-processing run for processing biogenic feedstocks in commercial applica-
SG 60/60°F D4052 0.870 0.872 tions and validated estimates made by Topsoe based on
Viscosity at fundamental research. The hydrocracking unit will remain a
100°C, cSt D7042 5.903 6.247 critical component of the renewable transition.
Peter Nymann is Senior Solution Specialist at Topsoe, involved in
Conclusions technical service for hydrocrackers, hydrotreaters, and renewable units
Processing renewable feedstocks in existing refinery assets worldwide. He graduated as a chemical engineer from Technical Uni-
is a strategy being used by many refiners globally to reduce versity of Denmark. Email: [email protected]
CO2 emissions originating from fossil-based feedstocks. In Pronit Lahiri is Principal Technical Service Engineer, Clean Fuels, at
addition to the local availability of renewable feedstocks, Topsoe in Denmark, involved in technical sales and service, as well as
another critical consideration has been to evaluate which R&D for making estimates of catalyst performance and planning pilot
unit in an existing refinery is best suited for co-processing plant tests. He holds a BTech in chemical engineering from National
biogenic feeds together with fossil-based feedstock. Institute of Technology in Jaipur, India. Email: [email protected]
Hydrotreaters are widely used, but hydrocracking units,
being one of the most versatile units in the refinery, are ide- LINKS
ally suited to co-process renewable feedstocks. They can
better overcome the operational challenges related to H2 More articles from: Topsoe
More articles from the following categories:
Catalysts and Additives Decarbonisation
Emissions Reduction
www.digitalrefining.com