1

ConocoPhillips S Zorb Gasoline Sulfur Removal Technology:

Unique Chemistry, Proven Performance, and Optimized Design
By
Jeff Vander Laan
Technology Product Director
ConocoPhillips
Introduction
ConocoPhillips Company breakthrough technology, S Zorb Sulfur Removal
Technology (SRT), substantially lowers the sulfur concentration in FCC gasoline
while still protecting octane-rich olefins and aromatics. In the S Zorb process
sulfur-containing molecules react with the novel S Zorb sorbent that retains the
sulfur atom from the molecule while the hydrocarbon portion of the molecule is
released back into the process stream. This unique reaction pathway does not
generate free H
2
S; therefore, preventing recombination of hydrogen sulfide and
olefins to form mercaptans. This technology was first demonstrated at the
ConocoPhillips refinery in Borger, Texas, USA in April 2001 and a second
commercial unit began operation in November 2003 and ConocoPhillips refinery
in Ferndale, Washington, USA.
Recent work conducted as part of a process optimization study led
ConocoPhillips to further optimize the S Zorb process resulting in a reduction of
capital and operational costs. In addition, this work led to the discovery of an
interesting S Zorb phenomenon. Pilot plant studies showed that olefin
conversion, hydrogen consumption and octane loss all decrease with increasing
reactor temperature while selectivity for sulfur removal increases. Subsequent
testing in the Borger unit confirmed these results on a commercial scale. The
resulting process enhancements are incorporated into an optimized design for
the unit that is being built at the ConocoPhillips Lake Charles Refinery, located in
Westlake, Louisiana, USA as well as a 4
th
ConocoPhillips unit that is in the
advanced stages of project approval. This paper will focus on the optimized
design, improved sulfur selectivity, additional enhancements to the technology,
and operational details from the two commercial S Zorb units.
Overview of S Zorb Sulfur Removal Chemistry
The ConocoPhillips S Zorb SRT is an advanced processing technology for
extracting sulfur atoms from gasoline blend stocks. It is not a
hydrodesulfurization technology. Rather, it uses a novel sorbent that specifically
absorbs sulfur-containing molecules and removes the sulfur atom from the
molecule. The sulfur atom is retained on the sorbent while the hydrocarbon
portion of the molecule is released back into the process stream. This unique
reaction pathway does not generate free H
2
S, and, therefore avoids problems
associated with H
2
S inhibition and mercaptan recombination. Figure 1 compares
S Zorb SRT chemistry with hydrotreating.
2
Figure 1: S Zorb vs. Hydrotreating
The S Zorb chemistry leaves the hydrocarbon stream virtually unchanged except
for the removal of sulfur. Hence the ConocoPhillips S Zorb SRT process can
economically reduce the sulfur content of gasoline to less than 10 ppm with
minimal octane loss, minimal hydrogen consumption and near zero volume loss.
Hydrogen consumption is kept low by minimizing the extent of olefin saturation.
Another interesting aspect of the S Zorb chemistry is the order of sulfur reactivity.
Figure 2 shows a comparison of the reaction rates for different species of sulfur
compounds that are typically found in cracked naphthas. This figure does not
compare the rates on an absolute basis, but rather on a relative basis where the
first-order rate of sulfur removal has been normalized relative to thiophene. On
an absolute basis, the sulfur removal rates observed with S Zorb SRT are
several times faster than those reported for HDS. As the figure illustrates,
thiophene is the hardest molecule to remove for either process; however, there is
an enhancement in activity for alkylated thiophenes, alkylated benzothiophene,
and benzothiophene for the S Zorb sorbent. As a result, S Zorb SRT offers much
more flexibility and forgiveness with regard to gasoline endpoint changes and
other operational upsets which may lead to significant variations in the feed sulfur
speciation.
3
Figure 2. Comparison of relative reaction rates for different sulfur
compounds. The HDS data were taken from Hatanaka et al. (1).
The ability to retain high sulfur conversion while minimizing olefin hydrogenation
and the subsequent octane loss presents a significant advantage of S Zorb SRT
over hydrotreating technologies, as shown in Figure 3. (HDS data are from
McGuff (2) and S Zorb SRT data are from the Borger, Texas commercial unit.)
Relative
Reactivity for S Zorb SRT
1.0
2.1
3.7
4.0
7.0
10.0
1
10
Thiophene C1-C2
Thiophenes
C3+ Thiophenes Alkyl-
Benzothiophenes
Benzothiophene Mercaptans
& Sulfides
Relative
Reactivity for HDS
1.0 1.0
0.3
2.0
5.5 5.5
1
10
Thiophene C1-C2
Thiophenes
C3+ Thiophenes Alkyl-
Benzothiophenes
Benzothiophene Mercaptans
& Sulfides
Relative
Reactivity for S Zorb SRT
1.0
2.1
3.7
4.0
7.0
10.0
1
10
Thiophene C1-C2
Thiophenes
C3+ Thiophenes Alkyl-
Benzothiophenes
Benzothiophene Mercaptans
& Sulfides
Relative
Reactivity for HDS
1.0 1.0
0.3
2.0
5.5 5.5
1
10
Thiophene C1-C2
Thiophenes
C3+ Thiophenes Alkyl-
Benzothiophenes
Benzothiophene Mercaptans
& Sulfides
4
Figure 3. Olefin Retention Comparisons
0
20
40
60
80
100
80 85 90 95 100
Sulfur Conversion, %
O
l
e
f
i
n

C
o
n
v
e
r
s
i
o
n
,

%
Conventional fixed bed Selective HDS
Selective HDS S Zorb SRT
Octane loss at a specified sulfur conversion is determined by the type of olefins
saturated, not just by the total olefin conversion. As sulfur conversion
requirements increase, the more severe operating conditions required will cause
octane loss to increase. This is true of all technologies, but less so for the S Zorb
process. Data compiled throughout the development of S Zorb SRT and from
commercial experience demonstrate the high selectivity of the S Zorb
technology, which results in superior octane retention capabilities.
A recently completed optimization study resulted in the discovery of an
interesting S Zorb phenomenon, which does not occur in hydrotreating. S Zorb
pilot plant work showed that olefin conversion, hydrogen consumption, and
octane loss all decrease with increasing reactor temperature. Figure 4 shows the
decrease in olefins conversion with temperature observed in pilot plant
operations and confirmed through commercial operations at the first S Zorb unit
located in Borger, TX. Figure 5 shows the temperature effect on RON and MON
observed in commercial trials.
5
Figure 4: Temperature Effect - Olefin Conversion
Figure 5: Temperature Effect - RON & MON
Temperature Temperature
O
c
t
a
n
e

L
o
s
s

RON Loss
MON Loss
RON Loss
MON Loss
RON Loss
MON Loss
Commercial Unit
Figure 6 summarizes the effect of temperature on octane loss for various types of
feeds with varying levels of feed sulfur. In all cases there was a significant
reduction in octane loss with the higher reactor temperature and a very slight
increase in light ends formation (less than 0.2 wt%).
Olefin Conversion with Temperature
O
l
e
f
i
n

C
o
n
v
e
r
s
i
o
n
Temperature Temperature
Pilot Plant
Commercial Unit
6
Figure 6: Temperature Effect Feed Variation
Process Description, Revamp Options, and Flexibility
Process Description
To take advantage of the unique selectivity for sulfur removal at low levels of
olefin hydrogenation, the S Zorb process was designed using a fluidized bed
reactor with continuous regeneration to control the steady-state level of sulfur on
the sorbent in the reactor. The use of a continuously regenerated, fluid bed
reactor system means that an S Zorb unit achieves true steady state operation,
with no degradation in performance throughout the run.
Gasoline feed is combined with a hydrogen stream that is vaporized in the
charge heater and injected into the bottom of the fluidized bed reactor. As the
vaporized feed stream moves through the bed, the sorbent removes sulfur
compounds from the hydrocarbon vapor. The sulfur-free hydrocarbon portion of
the molecule stays in the process stream. The sulfur atom remains with the
sorbent, which is transported to the regenerator. In the regenerator, sorbent is
oxidized to produce sulfur dioxide (SO
2
). The regenerator effluent is treated and
then recycled back to the regenerator. The regenerated sorbent is transferred to
a reducing vessel before being returned to the reactor. A general overview of
the S Zorb process is illustrated in Figure 7.
Feed 2
915 ppm S
Feed 3
1300 ppm S
Feed 4
1990 ppm S
Feed 1
735 ppm S
HCN of
Feed 1
1000 ppm S
O
c
t
a
n
e

L
o
s
s

Base Temperature
Higher Temperature
Feed 2
915 ppm S
Feed 3
1300 ppm S
Feed 4
1990 ppm S
Feed 1
735 ppm S
HCN of
Feed 1
1000 ppm S
O
c
t
a
n
e

L
o
s
s

Base Temperature
Higher Temperature
Base Temperature
Higher Temperature
!Pilot Plant Runs at Similar Sulfur Conversion Levels
7
Figure 7: S Zorb SRT Process Overview
Revamp Options
Refiners evaluating revamp options should give serious considerations to S Zorb
SRT. A review of the S Zorb flow sheet Figure 7 shows that, outside of the
reactor-regenerator section shown within the dashed lines, the process looks
similar to a conventional hydrotreater. Existing hydrotreaters can be converted to
use S Zorb technology, usually with higher throughput. The recycle hydrogen gas
rates used for the S Zorb process are much lower than those typically used for
hydrodesulfurization, which results in additional hydraulic capacity that can be
used for incremental liquid feed. Existing hydrotreater equipment can be easily
revamped for use with S Zorb SRT. The S Zorb process also has the unique
ability to optimize octane retention capabilities across a fairly wide range of
reactor pressures making it a good fit for many revamp scenarios where the
operating pressure is set by existing equipment. As a result of this unique
flexibility, there are many cases where a revamp using S Zorb SRT will be able to
use existing compression, heat exchange, and separation equipment without the
extensive modifications that would be required for a comparable HDS revamp. Of
course, the refiner will also benefit from all the operating advantages and
superior operating performance inherent in S Zorb SRT. If a refiner has already
installed a gasoline desulfurization technology that cannot economically meet the
existing and future limits, replacing the reactor with an S Zorb reactor system is a
viable option.
8
Flexibility
The S Zorb fluidized bed reactor holds an even temperature distribution and
maintains start of run activity throughout the run cycle. Because sorbent is
continually regenerated, there is no measurable coke build up through the run
cycle. Reactor fouling due to diolefin or other contaminants is not a problem for S
Zorb SRT. The S Zorb sorbent has proven to be remarkably tolerant to
contaminants and poisons found in fluidized catalytic cracker (FCC) and coker
naphthas. For example, S Zorb SRT has demonstrated the ability to handle
typical silicon levels found in most coker naphthas with no loss in performance.
During silicon upsets, full activity can be restored either via the natural
replacement of sorbent due to slow attrition or by employing a controlled addition
and withdrawal strategy. In this manner, the S Zorb process design gives refiners
a number of options to insure that optimum sulfur removal activity is maintained,
offering an unprecedented level of flexibility.
It is a well-known fact that nitrogen compounds are poisonous to hydrotreating
catalysts. Refiners who routinely process crude oils that are high in nitrogen
content are well acquainted with the challenges associated with desulfurization of
distillate, naphtha, and other feedstocks using current hydrotreating technologies.
In contrast to hydrotreating, S Zorb pilot plant experiments conducted with FCC
naphtha feeds high in total and basic nitrogen have shown no negative effect on
sulfur removal activity, octane retention, run length, or sorbent requirements. This
is just one more example of how the unique S Zorb chemistry produces a very
robust and flexible desulfurization process.
In addition, the unit is easily adjusted for feed changes and hydrogen purity has
little effect on unit performance. In S Zorb SRT units these features combine to
give stable yet flexible operations and run lengths for gasoline units that can
easily match an FCC unit. Matching the run lengths of the S Zorb unit and the
FCC unit provides refiners with an economic advantage, eliminating the need for
additional tankage to store untreated FCC gasoline during sulfur treatment unit
downtime.
Commercial Experience
Borger Refinery S Zorb Unit
The first commercial installation of S Zorb Sulfur Removal Technology has been
in service at ConocoPhillips refinery in Borger, Texas, USA since April 2001.
The unit was designed to process full range debutanized FCC naphtha with feed
sulfur content up to 1000 ppmw. The unit has very successfully demonstrated
the impressive capabilities of this technology and confirmed our ability to scale
up pilot plant results to a commercial scale. Table 1 shows a side by side
comparison of typical results from the Borger S Zorb unit vs the pilot plant. This
is just one of many examples of how the Borger unit exceeded ConocoPhillips
performance expectations. During the last 3 _ years of operation, this unit has
shown the ability to effectively remove sulfur from feeds ranging from 200 to
9
1400 ppmw and produce product with less than 10 ppmw sulfur. This unit has
demonstrated that S Zorb SRT is a proven, reliable, and flexible process for FCC
Naphtha desulfurization.
Table 1 Comparison of typical performance of the Borger S Zorb unit vs
pilot plant results with Borger feed
Borger S Zorb Unit Pilot Plant
Feed sulfur, ppmw 248 218
Product sulfur, ppmw 14 14
Sulfur Conversion 94.4% 93.6%
Octane Loss, !(R+M)/2 +0.1 -0.1
Charge Rate, mtpd 750 0.14
Reactor Temperature, C 400 413
Reactor Pressure, kb/cm
2
g 9.3 10.6
WHSV, hr
-1
8.0 6.0
Ferndale Refinery S Zorb unit
The largest commercial S Zorb SRT unit was brought on stream in the fourth
quarter of 2003 at ConocoPhillips refinery in Ferndale, Washington, USA. The S
Zorb unit was part of a refinery upgrade project that included the start up of a
new fluidized catalytic cracking unit in the first quarter of 2003. The Ferndale
refinery has a crude oil processing capacity of 95,000 bpd. The S Zorb unit at
Ferndale has a design capacity of 20,000 bpd. It is designed to reduce the sulfur
in a debutanized full range FCC gasoline stream from about 1,500 ppmw down to
10 ppmw.
The Ferndale S Zorb SRT unit was designed and constructed in 17 months and
has proven capable of running at its design capacity and meeting product
specifications. The startup of the S Zorb unit at Ferndale took place in November
of 2003 and was completed successfully and safely. The unit was commissioned
ahead of schedule and came in 10% under budget. Gasoline from the FCC
debutanizer bottoms was introduced to the S Zorb unit on November 15, 2003.
Within 24 hours, the unit was running at design capacity. Sorbent regeneration
began about 48 hours after gasoline was introduced into the unit. In less than 72
hours, the S Zorb unit was stable at steady state conditions.
10
Soon after the initial startup, the Ferndale S Zorb unit began to achieve a high
degree of sulfur removal. Sulfur levels of less than 10 ppmw were observed with
minimal octane loss. Figure 8 is a plot of octane loss versus sulfur conversion
obtained during the first several months of operation. The data show the
excellent capabilities of the S Zorb unit at Ferndale and confirm the ability of S
Zorb SRT to achieve low product sulfur levels and minimum octane loss even at
high sulfur conversions.
Figure 8. Gasoline sulfur conversion and corresponding octane loss
observed at the S Zorb SRT unit at Ferndale refinery
-3.0
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
94% 95% 96% 97% 98% 99% 100%
% Desulfurization
O
c
t
a
n
e

L
o
s
s

(
P
r
o
d
u
c
t

-

F
e
e
d
)
Delta RON
Delta MON
Delta Road
S Zorb Process Optimization
Recognizing the importance of continuous improvement, ConocoPhillips
conducted a Process Optimization Study for the S Zorb process. The study was
based on the lessons learned from more than two years of the Borger S Zorb unit
operating experience and the improved process conditions demonstrated in our
pilot plants. Over 150 value-engineering ideas were systematically evaluated,
validated, and, if approved, incorporated into the optimized design.
The objectives for the process optimization work included fine-tuning the basic
technology, reducing the ISBL cost by at least 20%, and matching or exceeding
the reliability, measured as onstream availability, of a fixed-bed
hydrodesulfurization process. Figure 9 depicts the flow scheme of reactor,
regeneration, and sorbent transfer sections of the optimized design.
11
Figure 9: S Zorb Optimized Flow Scheme
Reactor Section
The most significant change in the reactor section was an increase in reactor
pressure. The increase in pressure increases gas residence time in the reactor,
allowing a decrease in reactor sorbent inventory and a decrease in the molar
ratio of hydrogen to hydrocarbon. Higher reactor pressure also increases sorbent
efficiency, which allows an increase in sulfur on sorbent load. The lower gas
volume allows the use of a reciprocating compressor or the possibility for the
refiner to select a once-through hydrogen option. Also, lower gas volumes result
in a reduction of the reactor filter cost. The lower sorbent circulation rate required
as a result of the higher level of sulfur retained on sorbent allows a reduction in
the size of equipment and piping.
Regeneration Section
Better understanding of sorbent regeneration chemistry combined with over two
years of commercial operating experience has enabled optimization of
regenerator operating conditions. The major change is the use of once-through
air regeneration, as opposed to a recycle nitrogen system with makeup air. This
change eliminated the regeneration recycle gas compressor, and the lower gas
volume allows use of a smaller air compressor. The lower gas volume also
reduces the size of both the SO
2
treatment package and the regenerator system
equipment sizes including vessels, piping, and valves. In addition, the
regenerator temperature has been lowered to allow the use of preferred vessel
metallurgy. The lower regenerator temperature shows no apparent effect on
either sulfur removal activity or octane loss.
12
Solids Transfer Section
The solids transfer system, which moves sorbent to and from the reactor and
regenerator systems, has been simplified. The use of one lockhopper system, as
opposed to two-lockhopper operation, has resulted in lower capital requirements
and operating costs (lower nitrogen and hydrogen requirements) without
compromising operability, reliability, or safety. HAZOP reviews have been
conducted to assure the safe operation of the One-Lockhopper System. The
results of the HAZOPs did not show any safety issues associated with the
design.
Reliability
The S Zorb process has the ability to shutdown and isolate the sorbent
regeneration system long enough to perform typical maintenance tasks. This
allows the reactor system to stay online at full charge making low sulfur product.
One objective of the Process Optimization Study was to ensure that the
optimized design matched or exceeded current reliabilities. The Process
Optimization team examined each section of the S Zorb process and made
recommendations to further improve unit reliability. In parallel with the Process
Optimization Study, KBC Advanced Technologies Inc. was requested to perform
an independent reliability assessment for the optimized design. In particular,
KBC was asked to compare S Zorb SRT reliability with hydrodesulfurization
reliability. KBC used their own methodology, equipment reliability database, and
set all the reliability assumptions. They concluded that there was no significant
difference in reliability between S Zorb SRT and hydrodesulfurization.
Conclusions and Future Outlook
The Ferndale units safe and successful start-up and reliable performance
demonstrate the advantages of S Zorb SRT for desulfurization of gasoline
streams. The demonstrated commercial success at Ferndale is complemented
by further reductions in capital and operating costs and improvements in the
technologys performance and reliability. A 38,500 bpd S Zorb unit featuring
these technical improvements is in the advanced stages of design/construction
for the ConocoPhillips refinery at Lake Charles, Louisiana. The combined feed
sulfur level is approximately 1,000 ppm and the unit is being designed to produce
a 10 ppm product. Startup of the S Zorb unit at Lake Charles is scheduled for late
2005. In addition, a 4
th
ConocoPhillips S Zorb unit is in the advanced stages of
project definition and approval. The anticipated design throughput is around
30,000 BPD and is targeted to start-up in late 2006.
The future promises to bring continued downward pressure on the sulfur content
of gasoline even beyond the near term targets that have been established in
many countries around the world. S Zorb technology is strongly positioned to
produce fuels that comply with much more stringent sulfur regulations with little,
13
or potentially no, octane loss, minimal hydrogen consumption, strong operating
reliability, and flexibility.
ConocoPhillips is committed to an active technology development program
focused on additional enhancements to S Zorb SRT. These enhancements
promise to deliver additional improvement to the technology, including further
improvements to the impressive octane retention capability. These
enhancements are expected to be available near term and are considered drop-
in improvements. So, the benefits are immediate requiring little or no capital to
implement.
References
1) S. Hatanaka, M. Yamada, and O. Sadakane, Hydrodesulfurization of
Catalytic Cracked Gasoline. 1. Inhibiting Effects of Olefins on HDS of
Alkyl(benzo)thiophenes Contained in Catalytic Cracked Gasoline, Ind.
Eng. Chem. Res., 1997, 36, 1519-1523.
2) Tim McGuff, Ultra Low Sulfur Gasoline via Catalytic Distillation, CD
Tech, GDS Seminar, July 18, 2001, Houston.