A new isoparaffin longer than most solid-acid catalysts.

alkylation process using Long catalyst life, in turn, permits ef-

a solid-acid catalyst has ficient and inexpensive process and
recently been scaled up reactor schemes.
and proven in a pilot Catalyst performance and stability
plant using a strategy were demonstrated on a pilot-scale re-
outlined in this article. actor that closely matched the commer-
This new process is predicated on cial reactor design. Tests showed that
two key components: a unique, engi- product quality and catalyst lifetime is
neered solid-acid catalyst' and a simple identical in the bench scale and pilot-
fixed-bed reactor scale reactors when operated at the
that is designed same temperature, feed composition,
S ale-up strateg applied to to enhance the
performance of
and olefin space velocity.
' Applying this scale-up strategy and

solid-acid alkylation process the solid-acid

catalyst. 2 The
engineered cata-
a fully integrated catalyst development
technique resulted in a high product
octane rating and a simple process
lytic system has operation.
a highly structured pore architecture, Microscale reactors were used for the
which enhances transport processes initial catalyst screening and to develop
Mitrajit Mukherjee within the catalytic system leading to mathematical models of the reaction
James Nehlsen better system stability. kinetics. Bench-scale reactors provided
Sankaran Sundaresan The larger pores provide efficient the data to optimize catalyst perfor-
G. Dan Suciu
access and quick diffusion of reagents mance and to design the reactor. These
John Dixon
Exelus Inc. to the microporous system and the tests also helped determine the effects
Livingston, NJ smaller pores provide the surface area of various process parameters.
required to contain the optimized acid Finally, a pilot unit was used to
sites. The integration of catalyst science verify all of the results with conditions
and reaction engineering allows catalyst identical to the commercial reactor. The
lifetimes that are an order of magnitude pilot unit was designed to match all of
the important transport
and reaction proper-
MICROREACTORRESULTS Fig. 1 ties of the commercial
100 ... ~

v
A A
v
A
v
A
v . ,. -v
reactor.
Olefin conversion
.
'it. 80
- -
- -
C8s
-
""'
Integrated
approach
s:i- Moving a new refin-
g ery process from incep-
CD
a; tion to commercializa-
(/)

0 60 tion requires many

.,,,
::J

ea. phases of development.

c0 Catalytic processes,
~ in particular, require
(]) 40
>
c: development at several
0
"c: scales in order to first
<;:::
CD - create and optimize the
6 20 catalyst, then to opti-
Lr l.J
-
~

--
Cs-C-z.
- ~ mize the reactor design
and process conditions,
0 -
~
'r - v
Cs+
.,. ~
- 0
and finallyto demon-
0
0 2 3 4 5 6 strate the performance
Catalyst age, kg olefin/kg catalyst on an appropriate pilot
scale.

48 Oil & Gas Journal I July 70, 2006

 To further complicate the develop-
PILOT TESTING Fig. 2
ment, these stages are not independent
100 A A A A A A A
100
l I
because the catalyst characteristics, reac- @lefin conversion
tor design, and process conditions all
contribute to the overall process design
and economics. 80 98
When developing the step-out
D
LI -
n
I
n
D
n
~
r\
LI
- --
RON
alkylation catalyst, we used a fully CJ)

integrated approach to process design '/!. r::

;::;
c0 60 96 ~
to achieve the best possible process per- ' Q)
r::
formance. The underlying principle of Q)
>
r:: tl0
this approach is that the catalyst, reactor,
and supporting unit operations are all
0
u
r::
:.=
Q)
40
- n
- n n n n
~ u
MON
- - "'
94 >.
~

-"'
parts of a single process that should be B <(

designed simultaneously rather than

linearly, and always considering the 20 92
practical application and cost of the
process.
This approach can lead to significant
improvements in process economics 0 90
0 20 40 60 80 100 120
and shorten the process development Time on stream, hr
time, with its own economic benefits.
Our goal was to develop a simple
swing fixed-bed reactor process using PRODUCT llUALITY Fig. 3
an environmentally safe solid-acid cata- 100

lyst. We wanted to use a fixed-bed reac-

<> <>
tor for two reasons: to simplify process <>
<>
<> <> , <> ,
scale up and to keep capital costs and
A
98 v
<>"" <> ~
vooe,;~<
"'"
v v
<> <>"
operating expenses as low as possible. <> RON

In addition to being a 1-3 mm diam-

"'iii 96 ---
eter particle to function in a fixed-bed
u0 a a
reactor, the solid-acid catalyst had to u a
a
- 'l: ~ ~ ... - -a a
be robust. Consequently, the ideas of "'e
-0 a a a
a u uu~ Bl! u a
~

"- 94 ' ~MON

replacing the conventional liquid acids
with a liquid acid supported on solid
particles or co-feeding toxic halogens 92
to a solid-acid catalyst to maintain its
activity were rejected right at the outset.
Only a chemically stable, solid acid 90
8 10 12 14 16 18 20
would meet the requirements.
Feed isoparaffin:olefin ratio, mole/mole
Our development approach was first
to understand the catalysis of the alkyla-
tion reaction by systematically studying tor design. product octane was more than 9 8 RON
the effects of physical and chemical We then studied the optimized solid- using a methyl tertiary butyl ether raf-
parameters of the solid-acid catalyst on acid catalyst in a bench-scale reactor finate feed in the pilot plant over mul-
product distribution and catalyst deacti- under commercially relevant conditions. tiple cycles of alkylation followed by
vation characteristics in a microreactor. Process conditions for operating the regeneration, confirming the viability
We developed a robust kinetic model alkylation unit along with the catalyst of the solid-acid alkylation technology.
from data obtained in the microreactor. design and reactor configuration were
The model was used to engineer the fine-tuned with the bench-scale reactor Microreactor studies
solid-acid catalyst to provide optimal data. We conducted initial catalyst screen-
performance in terms of catalyst life Finally, we used the optimized ing experiments in a once-through
and product octane. The model was also catalyst and chosen reactor configura- microreactor using 0.5-1.0 g of catalyst.
used to develop a viable fixed-bed reac- tion to design and run a pilot plant. The To achieve acceptable catalyst lifetimes,

Oil & Gas Journal I July 10, 2006 49

we conducted screening tests SOLID-ACID ALKYLATION PLANT fig. 4 alkylation process revealed
in the microreactor at feed that the solid-acid catalyst
isoparaffin-to-olefin (I:O) produces alkylate with a
ratios of 100-200 to suppress high octane rating over a
the olefin oligomerization wide range of operating
Olefin
reaction. We evaluated several feed temperatures (60-90 C.),
different catalyst formulations olefin space velocities (0.1-
encompassing a wide range 0.5 hr- 1), and feed composi-
of acid functions, number of tions (I:O ratios of 10-15).
acid sites, and acid strengths Its robustness minimizes
as part of this study. the requirements for feed
Through a systematic pretreatment.
study, we were able to iden- Table 1 summarizes
tify an optimal window of results from the bench-scale
design parameter values that testing of some recently
were then used to develop synthesized catalysts. These
TYPICAL ALKYLATESTESTED Table 1
the new solid-acid catalyst. catalysts are an improved
Parameter Untreated MTBE raffinate feed
Fig. 1 shows typical data version of the one used for
from the microreactor. Feed isoparaffin:olefin ratio, 15.78 11.85 10.11 pilot-plant testing. The octane
mole/mole
Product distribution, wt % values and Reid vapor pres-
Mathematical modeling c,-c, 8.3 9.2 9.5
sures were computed with
c, 87.5 85 .8 83.6
The microscale studies c 4.2 5.0 6.9 gas chromatography (GC)
TilAP:DMH ratio, mole/mole* 16.3 15.1 12.4
provided information for Product properties product analysis. 5 Typical
RON 100.5 99.6 99.0
initial correlations of catalyst MON 95.9 95.3 95.0 ASTM engine test octane
rvp, psia 2.23 2.36 2.34
performance with reaction values tend to be higher than
conditions. Analyzing the Molar ratio of the desired trimethylpentane isomers to the undesired dimethylhe><- those computed with the GC
ane isomers
rate data obtained from the method.
microreactor allowed us to the longest !ife with the highest -octane The trimethylpentane-
develop a robust model that can pre- product was chosen and catalyst synthe- to-dimethylhexane (TMP:DMH) molar
dict product octane as well as catalyst sized accordingly. ratio in the alkylate product indicates
lifetimes for a large range of reaction the relative degree to which the pri-
conditions. Bench-scale studies mary reaction is isobutane alkylation or
This model accounts for the effect Once the basis for the catalyst and olefin polymerization. Dimethylhexanes
of intraparticle diffusion, which is reactor design was established, larger arise from olefin dimerization followed
substantial, and can accurately pre- bench-scale studies were initiated. 4 by hydride transfer; trimethylpentane
dict the rate of catalyst deactivation To maintain a high I:O ratio (about formation largely depends upon paraf-
caused by the formation of heavy 400) in the reactor while using com- fin-olefin combination to form a carbo-
oligomers. The model was validated mercial grade feedstocks (I:O ratio of nium ion species followed by hydride
with experimental data. 10-15), we used a fixed-bed recycle abstraction from isobutane.
The analyses also revealed that the reactor to conduct most of the research A higher the TMP:DMH ratio in-
rate constant for the undesired second- work. Typically 5-10 g of catalyst were dicates a greater effect of alkylation
ary (oligomerization) reaction is two used for catalyst testing with a feed rate reactions, in contrast to olefin polym-
orders of magnitude higher than the of 50-100 g/hr. erization. With HF or with H 2SO 4 , the
rate constant for the primary alkylation The main purpose of the research TMP:DMH ratio is 3-6. These values are
reaction. 3 To reduce substantially the was to: also typical of most solid-acid catalysts
rate of oligomerization, therefore, the Fine-tune the solid-acid catalyst reported in literature.
olefin concentration near the catalyti- design. - The high TMP:DMH ratio for the
cally active sites must be two orders of Generate data to enable design of new catalyst is the reason for the long
magnitude lower than the isoparaffin viable fixed-bed reactors. catalyst lifetimes and high product oc-
concentration. Study the effect of temperature, tane; the olefin polymerization reaction
Using the reaction model, we simu- feed composition, and olefin space have been suppressed and the alkylation
lated several scenarios of pore size and velocity on catalyst stability and product reaction rates are increased. The long
acid site distribution within the catalyst quality. catalyst lifetimes enable the use of a
particle. The configuration that gave The bench-scale results for the simple fixed-bed reactor.

50 Oil & Gas Journal I July 70, 2006

 time of the alkyla-
PILOT-PLANT REACTOR DESIGN PARAMETERS Table 2
tion cycles only.
Scale-up Dimensionless Calculation* Bench-scale Pilot-scale Commercial Fig. 2 shows that
parameter group reactor reactor reactor
there is no activity
Chemical Damkohler T}k1l'llT 1,152 720 720 loss over repeated
reaction number
Momentum Reynolds d,vp/ 87 280 280 cycles of alkylation
transfer number
Mass Sherwood k,d/D 42 73 73 and regeneration as
transfer number
Heat Nusselt hd,lk 29 50 50 measured by olefin
transfer number conversion, product
Axial Peclet d,v/D, 0.21 0.69 0.69
dispersion number octane, or alkylate
Variables defined in nomenclature box. yield. The product
octane is signifi-
Designing the pilot unit Nomenclature
cantly higher than
After the catalyst and reactor design C ; Fluid specific heat. J/kg/K that obtained with liquid-acid processes
D = Fluid diffusivity, sq m/sec
were complete, we needed to test the d = Particle diameter, m under similar operating conditions
catalyst on a pilot scale. Pilot testing is rf = Axial dispersion, sq m/sec and feedstocks. The olefin feed, which
h' = Fluid heat-transfer coefficient,
essential to: W/sqm/K contained more than 3,500 ppm of
k = Fluid thermal conductivity, W/m/K
Generate enough alkylate to mea- k = Intrinsic kinetic rate constant, sec" diolefins, 700 ppm of oxygenates, and
k~ = lnterphase mass-transfer coefficient,
sure product octane with an engine test. m/sec 0.5 ppm of acetonitrile, was untreated
Demonstrate catalyst stability. v Superficial velocity, m/sec for these tests.
1] Catalyst effectiveness factor
Develop critical data to allow suc- Fluid viscosity, kg/m/sec Product quality was studied as a
p Fluid density, kg/cu m
cessful scale-up to a commercial-scale Residence time, sec function of process conditions, olefin
plant. feed concentration, and olefin type.
In anticipation of the commercial- Product octane is only moderately
reactor design, we chose a pilot-plant parameters (Table 3). In other words, sensitive to olefin space velocity but
configuration that included multi- scale-up of the new process from bench decreases 1 octane point for a 15 C.
staged olefin feed with product recycle. scale to a commercial plant is a com- rise in temperature.
We carefully designed the pilot unit by paratively simple exercise. Alkylate yields are typically 2. 15-2. 2
maintaining the same scale-up param- weight alkylate/ weight olefin con-
eters in the pilot unit as in the commer- Pilot-plant performance sumed for the range of process condi-
cial-scale reactor. Catalyst performance was validated tions studied. There is no difference in
Interphase heat and mass-transfer in the pilot unit for an extended time. catalyst performance between 1-butene
effects are important for liquid acid Catalyst stability, product octane, and and 2-butene feeds, which is in stark
alkylation reactors but do not play a alkylate yield were proven in the pilot- contrast to liquid-acid processes.
major role in the solid-acid alkylation scale reactor (Fig. 2) . Fig. 3 shows that product quality
process. We confirmed catalyst stability by is essentially unaffected by feed olefin
Table 2 compares the critical scale- measuring the catalyst performance concentration in the range of commer-
up parameters for the two reactors in multiple cycles of alkylation and cial interest (feed 1:0 ratios of 8-20) .
(see accompanying box: nomenclature regeneration using hydrogen. Alkylation Isobutylene has a slightly negative influ-
for dimensionless numbers). Table 2 cycles were 8-12 hr followed by a 2-hr ence on product quality. An FCC olefin
shows that the scale-up parameters are regeneration cycle. The time on stream mix containing 30% isobutylene pro-
matched, which should enable a seam- reported here refers to the cumulative duces a 3 octane-point loss compared
less scale-up from the pilot to an MTBE raffinate feed.
unit to a commercial-scale
plant. BENCH. PILOT-SCALE REACTOR PERFORMANCE Table 3
Commercial process
The critical parameters scheme
Parameter Bench-scale reactor Pilot-scale reactor
in the alkylation process are Scaling up the process
Temperature Base Base
reaction temperature, feed Feed Untreated MTBE raffinate Untreated MTBE
from a pilot-plant reactor to
composition, olefin space raffinate a commercial-scale reactor is
lsoparaffin:olef in rat io Base Base
velocity, and bed I:O ratio. Olefin space velocity Base Base a relatively straightforward
Olefin feed point Single Multiple
As long as these parameters Bed isoparaffin:olefin 400 400 process. Because the hy-
remain the same, catalyst ratio drodynamic, transport, and
Catalyst life, hr Base Base
performance is not a strong Product RON 97.5 97.3 chemical reaction parameters
Product MON 94 .3 94.5
function of the transport remain the same, there is

52 O il & Gas Journal I July 10, 2006

 - - - -------

little else that is different in these two tion section, which minimizes capital already in place. The primary changes
reactors except for the physical size of costs.6 The separation section resembles entail adding some optional feed-treat-
the vessel. those used in typical liquid-acid alkyla- ing equipment and installing a small
One added advantage of the isoparaf- tion plants and, as such, can be easily reactor recycle pump and a recycle
fin alkylation process is the low heat of retrofitted into existing liquid acid al- hydrogen compressor for the catalyst
reaction, which minimizes the need for kylation plants, catalytic polymerization reactivation section. Minimal modifica-
large heat-transfer surfaces to be inte- units, and idle MTBE units. tions to the existing distillation units
grated within the reactor volume. The revamp of a catalytic polymer- are required.
The new process scheme consists of ization unit is simple because two or Fig. 4 shows a revamp configuration
two swing-bed reactors and a separa- more multistage fixed-bed reactors are using multiple olefin feed stages with
product recycle. The reactor is operated
in a swing mode such that one reactor
is on alkylation while the other is on
catalyst regeneration.
CORRECTION Alkylation cycle lengths are designed
Figs. 2 and 6 in the article "Refiners to see strong returns near-term de-
to be 12-24 hr to simplify reactor op-
spite looming capacity buildup" by Aileen Jamieson (OGJ, Mar. 13, 2006,
eration. Every 12-24 hr, the alkylation
p. 52) contained errors. Fig. 2 mistakenly listed Dubai crude for the Singa-
pore margin calculation. Fig. 6 erroneously listed the blue bars as grassroots reactor is removed from service and the
and the green bars as expansions. The correct figures are shown here. solid-acid catalyst is regenerated with
excess hydrogen under mild conditions.
The regeneration step lasts 2 hr. Due to
REFINING MARGINS Fig. 2
the small amount of coke build up on
the catalyst surface, hydrogen consump-
-0- Northwest Europe (Brent) tion is minimized.
- - US Gulf Coast (West Texas Intermediate)
10
- - Singapore (Minas)
Acknowledgment
This work was partially funded
8 through grants from the US Department
of Energy and the National Science
Foundation. +

References
4
1. Mukherjee, M .. and Sundaresan,
S., "ExSact- A Step-out Isoparaffin
2 Alkylation Technology, Part 1. Catalyst
Development," World Refining, Janu-
ary /February 2005, pp. 28-31.
2000 2001 2002 2003 2004 2005 2. Mukherjee, M. and Sundaresan S.,
"ExSact-A Step-out Isoparaffin Alkyla-
tion Technology. Part 2. Reactor Devel-
opment," World Refining, March 2005,
REGIONAL REALISTIC ADDITIONS. 2005-15 fig. 6
pp. 22-26.
3.5
3. Simpson, M.F., Wei, J., and Sun-
'O
3.0 DExpansions daresan, S., "Kinetic Analysis oflsobu-
:0 2.5
c
0 Grassroots tane/ Butene Alkylation over Ultrastable
~E 2.0 H-Y Zeolite," Industrial Engineering
and Chemistry Research, 1996, No. 35,
i- 1.5 pp. 3861-73.
.~
'ij

1.0 4. Mukherjee, M., and Suciu, G.D.,

u
"ExSact-A Step-out Isoparaffin Al-
0.5
kylation Technology, Part 4. Solid-Acid
Asia Pacific Middle East North Africa Latin Europe Russia Catalyst Optimization," World Refining,
America America June 2005, pp. 28-31.
5. Hutson, T., and Logan, R., "Esti-

Oil & Gas Journal I July 70, 2006 53

 A World of Reasons
mate Alkylate Yield and Quality," Hydro-
to Read the 2006 IPE carbon Processing, September 197 5, pp.
107-10.
6. Mukherjee, M. and Porcelli,
The 2006 Jn_ternational Petroleum Encyclopedia is a time-honored and concise R., "ExSact-A Step-out Isoparaffin
source of information for the rapidly changing petroleum market. Inside these Alkylation Technology, Part 3. Process
Economic Evaluation," World Refining,
pages you'll find valuable information including:
April/May 2005, pp. 32-35 .
Atlas country reports offering 31 % more illustrations and four new
countries-Jamaica, Malta, Lithuania, and EastTimor
Key Stats providing 16% more illustrations
Statistics from the Oil & Gas Journal and BP Statistical Review of World Energy
2005 as well as the US Energy Information Administration in one section
The authors
Mitrajit Mukherjee ([email protected])
Chronology of Events is the founder and president of Exelus Inc., Livingston,
Guest Essay written by Steve Farris, President and COO of Apache Corp. NJ. He has spent most of his I 7-year career develop-
Directory of national oil companies and energy ministries ing viable solid-acid catalyst solutions for isoparaffin
alkylation. Before founding Exelus, he held positions at
Catalytica and ABB Lummus Global. Mukherjee holds
Available in a BS in chemical engineering from the Indian Institute
print or CD- of Technology and an MS from the Southern Illinois
University. He is a member of ACS and AIChE
ROM, the/Pf
contains more James Nehlsen [email protected])is a
than 350 pages
senior research engineer at Exelus Inc. special-
izing in the development of new catalytic processes.
of information He holds a PhD in chemical engineering from
on important Princeton University where he developed new desul-
activities and furization technologies. Nehlsen is a member of
ACS and AIChE.
developments
concerning Sankaran Sundaresan is director of research and
oilfields,
development at Exelus and also a professor of
pipelines, refineries, tanker . chemical engineering at Princeton University. He
has authored more than I 00 journal publications
terminals, and more. in many fields, including solid-acid alk{lation.
He holds a PhD from the University o Houston.
Sundaresan is an editor of the AIChE Journal. He
is a member of ACS and AIChE.
G. Dan Suciu is a senior research advisor at Exelus
Inc. He was previously vice-president of research
and development at ABB Lummus Global. He is
the inventor of many first-of-a-kind technologies
including fluidized bed maleic anhydride and liquid
ISSN 1-59370-083-0. . . . . . . . . . . . . . . . . . . . . . $195
phase ethylbenzene processes. Suciu holds numer-
ous patents and recently edited the book "Thermal
Hardcover/approx. 350 pages/June 2006 and Catalytic Processes in Petroleum Refining:'
/PECO 1-59370-084-9 . . . . . . . . . . . . . . . . . . . . $195 He holds a PhD from Polytechnica University.
Buy both & savell2006SET . . . . . . . . . . . . $312 Buucharest, Romania. He is a member of ACS

John Dixon is a senior research specialist at Exelus

Inc. He has more than 40 years experience in
There's no reason to travel any catalyst evaluation with companies including
Mobil Corp. and Exxon Corp., including 16 years
further for the information you need! working with alkylation technologies. Dixon also
Order your IPE 2006 book or CD today I holds a patent for synthesizing high viscosity
lubricants.

www.pennwellbooks.com

54 Oil & Gas Journal I July 10, 2006