Refining Developments

Consider improving refining and

petrochemical integration as
a revenue-generating option

F
FCC operations may be an increasing propylene source
due to new processing methods

O
P. K. Niccum, M. F. Gilbert, M. J. Tallman, and of the propylene consumed by the petrochemical industry,
C. R. Santner, Kellogg Brown & Root, Inc., as shown in Fig. 1. Petroleum refining, nearly all from fluid
Houston, Texas catalytic cracking (FCC), is the next largest propylene

O
source, supplying about 30% of the petrochemical require-

L
ight olefins—ethylene, propylene and butylenes—are ments.1 In the U.S., FCC supplies about one-half of the
the basic building blocks in the manufacture of vari- petrochemical propylene demand.

R
ous petrochemical products and fuels. Presently, light Propylene demand has been increasing at a faster rate
olefins are used to manufacture, polymers, antifreezes, pet- than that of ethylene. Steam crackers are limited in the
rochemicals, gasoline, explosives, solvents, medicinals, amount of propylene that they are able to produce. Thus,

P
fumigants, resins, synthetic rubber and many other products. alternate propylene sources are quickly becoming a great
Ethylene is the largest volume petrochemical feedstock, and interest, including increasing production from FCC units.
virtually all of the ethylene supply comes from thermal This paper discusses the evolution of the propylene market
(steam) cracking of hydrocarbon feedstocks such as ethane, and catalytic fluid-bed processes that are expected to meet
propane, naphthas and gasoils. the rising petrochemical industry demand.
Propylene is second in importance to ethylene as a raw
material. Petrochemical-grade propylene is a coproduct in Propylene market. The demand for propylene has increased
ethylene cracking. Liquid-feed ethylene plants typically rapidly during the last 20 years and is primarily driven by
produce about 15-wt% propylene and provide almost 70% polypropylene demand as shown in Fig. 2. Propylene

History of fluid catalytic light olefins production

During the late 1930s, propylene and butylene were largely supplied as a byproduct from the thermal cracking of petroleum.
The major use of these light olefins was for the manufacture of gasoline via catalytic polymerization.3 During WWII, FCC was
developed to produce high-octane aviation gasoline and C4 s (isobutylene and butadiene precursors for the rapidly expanding
U.S. synthetic-rubber industry).

The first commercial FCC unit was built for the Standard Oil of New Jersey’s Baton Rouge, Louisiana refinery and commissioned in
May 1942.* Between 1942 and 1944, 34 FCC units were constructed throughout the U.S. The refining process quickly became a
major contributor to worldwide propylene and butylene production.

Rare-Earth exchanged Y zeolite catalyst was first synthesized by Mobil in 1959. By the late 1960s, over 90% of the U.S. FCC units were
operating with this zeolite catalyst. The high activity of the zeolite catalysts, compared to the earlier amorphous catalysts, greatly
improved gasoline yield and reduced coke and dry gas production from the FCC units. However, the catalyst’s high-hydrogen transfer
characteristic greatly reduced light-olefin yield and gasoline octane.4 These changes in product selectivity are listed in Table 5.

In the 1970s, FCC unit design and operation evolved to regain some of the lost octane and light-olefin yield, primarily through
higher reactor operating temperature and riser cracking.5 Raising reactor temperatures increased light-olefin yield, but this came
at the expense of co-producing dry gas—a lower value FCC product. During the 1980s, two new technologies were developed
that increased octane and light-olefins yield while limiting incremental dry-gas production:

•  The ZSM-5 catalyst additive to crack low octane (linear) gasoline boiling range olefins and paraffins into light olefins
•  Closed cyclones that minimized product vapor residence time between the riser outlet and the main fractionator.
* The M.W. Kellogg Co.

HYDROCARBON PROCESSING / NOVEMBER 2001 47

 OF
Fig. 1. Worldwide propylene supply in 2000.

demand by the petrochemical industry has increased more

O
rapidly than the demand for ethylene. This trend is expected Fig. 2. Global propylene forecast by product.
to continue. In the next 20 years, the demand for propylene is
expected to more than double. During the next five years, the riser system, including a standpipe, catalyst control valve, feed
demand for ethylene, propylene and gasoline/distillates are injection system and riser. The mechanical layout of the FCC

R
expected to increase annually by 5.3%, 5.6% and 3%, respec- converter and its structure is studied to determine the optimum
tively.2 placement and configuration of the new riser system. Because of
Ethylene plants produce more ethylene than propylene, and the substantial increase in light-ends production, modifications to

P
the construction of ethylene capacity is directly tied to demand the FCC vapor-recovery unit (VRU) will be required (unless the
for ethylene not propylene. Accordingly, significant increases FCC feedrate is reduced while operating in the maximum propyl-
in FCC-produced propylene will be required to meet the higher ene mode.)
propylene demand. Equally important, new FCC capacity will As the FCC unit operates at higher reactor temperatures to
be driven by the demand for gasoline not propylene. Conse- raise propylene yield, the ethylene yield also increases. Previ-
quently, most of the new propylene supplies will be extracted ously, ethylene produced by the FCC unit was viewed almost
from investments in existing FCC installations. exclusively as a component for refinery-fuel gas. Ethylene
Chemical Market Associates, Inc., (CMAI), estimates that yield was minimized with such technologies as closed cyclones
during the next five years, 4.1 million tpy (4.1 MM tpy) of and proprietary feed-injection nozzles.
propylene must be “pulled” from existing refinery sources to However, a retrofitted FCC unit can produce an economic
meet the projected petrochemical-grade propylene demand. volume of ethylene for proximate petrochemical plant or eth-
This increased production from existing FCC units will be ylene-pipeline. For instance, while traditional FCC operations
obtained by: produced less than about 2-wt% ethylene, new FCC technolo-
• Increasing propylene yield from the FCC units gies can produce as much as 8-wt% ethylene, and other fluid-
• Raising the percentage of FCC-propylene recovered ized catalytic reactor systems can produce ethylene yields as
for petrochemical manufacture as opposed to other uses. high as 20-wt% from C4 to C8 olefin-containing feedstocks.7
Refiners can select investments to increase the propylene
FCC options to produce petrochemical feeds. Improved refin- purity relative to a traditional refinery-grade propylene product.
ing technologies can effectively maximize propylene yield from The higher-purity options include chemical-grade and polymer-
FCC feedstocks.6 While FCC operations typically produce less grade propylene, with typical specifications shown in Table 1.
than 6-wt% propylene, specialized fluidized catalytic processes Some options for increasing petrochemical feedstock produc-
can further raise propylene yields as high as 20-wt% or more tion from existing FCC units are:
from FCC feedstocks. • Recovery of refinery-grade propylene typically
To improve profitability, refiners can adjust their operations to requires investments in a depropanizer (C 3 /C4 splitter),
consciously process higher-demand petrochemical feedstocks metering systems, storage tanks, railcar/tank truck loading
such as propylene. Depending on their extent of participation in facilities, or pipeline connections.
this market, refiners may elect to increase light-olefins production • Recovery of chemical-grade propylene would typi-
through revamping and debottlenecking of the entire FCC unit cally involve installing a propylene splitter in addition to
(reactor, regenerator, main fractionator and vapor-recovery unit.) the other facilities listed previously.
In general, revamping an FCC unit may require adding a second- • Production of polymer-grade propylene is an expen-
sive option, but also offers potential for higher return.
Table 1. Typical propylene quality specifications In addition to the previously listed investments, this
option would typically include adding a deethanizer
Component Refinery Chemical Polymer
grade grade grade
Propylene, % min 65.0 92.0 99.5
H2, CO, CO2, N 2, ppm max 100 100 5 Table 2. New FCC technology circulating pilot-plant data for
C2 & lighter, ppm max 10,000 4,000 150 hydrocracked gasoil feedstock
Ethylene, ppm max 10,000 100 100
C4 & heavier, ppm max 10,000 800 150 Yields, wt% Run A Run B Run C Run D
Butadiene, ppm max 200 20 10 Ethylene 3.2 3.9 6.4 8.2
MAPD, ppm max 150 100 10 Propylene 16.0 18.7 19.1 21.5
Sulfur, ppm max 20 1 1 C5+ Gasoline 37.9 28.8 26.2 25.0
Water, ppm max 100 30 10 P/E (wt/wt) 5.0 4.8 3.0 2.6

48 HYDROCARBON PROCESSING / NOVEMBER 2001

 F
Fig. 3. Characteristic yields from cracking processes.

tower, as well as COS/arsine removal reactors and dry- Fig. 4. U.S. refinery propylene recovery economics—1994 to 2004.

O
ers. Hydrogenation reactors may also be required to
meet the tight specifications for dienes/acetylenes. anticipated peak in 2003 and 2004.
• Production of 99.9-vol% purity petrochemical-
grade ethylene requires installing more fractionation Technology options. Improved FCC technologies can maxi-

O
towers, as well as, equipment designed to remove mize propylene production from traditional FCC feedstocks
associated contaminants such as acetylene, water, and selected naphthas. Such developments increase propylene
oxygen, carbon monoxide, sulfur and nitrogen com- yield relative to that produced by conventional FCC units by

R
pounds. Design of a low-temperature FCC vapor combining enhanced catalyst technologies with proprietary
recovery system that meets all of these requirements hardware, including a second high-severity riser that is
will more than double the cost of a VRU installation designed to crack surplus naphtha into incremental light ole-
relative to that of a traditional absorption oil- based fins.6

P
VRU system. Alternatively, a dilute-ethylene stream In addition to processing recycled light naphtha, C4 and
may be sent directly to an adjacent petrochemical LPG, the riser also can accept naphtha from elsewhere in the
plant for feed to an ethylbenzene unit and/or steam- refinery complex, such as coker-naphtha streams. It can
cracker gas recovery system. upgrade these streams into additional light olefins. Olefinic
streams, such as coker naphtha, convert readily into light ole-
Balancing supply and demand. The relative production rates fins. Paraffinic naphthas, such as light straight-run naphtha,
of propylene and ethylene from units within the refinery and also can be upgraded, but to a lesser extent than olefinic feed-
petrochemical plant are key considerations in refinery/petro- stocks.
chemical plant integration. Within each process unit, selection Table 2 shows the result of pilot tests, which demonstrate
of feedstock and operating conditions, as well as catalyst sys- the flexibility of propylene-oriented FCC developments with
tems where applicable, can dramatically alter light-olefin respect to propylene to ethylene ratio. The tests were done on
production rates and the propylene to ethylene ratio (P/E). Fig. a single feedstock while operating conditions and catalyst
3 shows typical propylene and ethylene yield characteristics formulation differ between the runs. This processing flexibil-
of various processing options, including data on steam crack- ity allows unit performance to be tailored to the needs of the
ing.8 refinery/petrochemical complex. The number of product
Although propylene demand is high, refiners are still cau- streams, degree of product fractionation and several other
tious about committing large capital investments for intentional aspects of the vapor-recovery process will differ from unit to
propylene production due to the historically large swings in unit, depending on the market requirements of the applica-
propylene-fuels margins. A recent study by CMAI estimated tion.
the margin and Return on Investment (ROI) for facilities to Vapor recovery options. Design of a vapor-recovery unit to
recover polymer-grade propylene from existing FCC units. produce polymer-grade ethylene and propylene products
This analysis was based on estimates of the margin between includes several factors not typically addressed in FCC VRU
polymer-grade propylene value to a petrochemical plant and design:
the refinery propylene cost (assumed equal to the alternative • Cold fractionation train begins with a depropanizer
value of propylene in a fuels refinery). The ROI is estimated system, followed by a demethanizer, deethanizer and eth-
from the margin assuming a typical investment cost of $75 ylene-ethane splitter. Polymerization fouling is minimized
MM for a 250,000-tpy polymer-grade propylene recovery in both columns as a consequence of the low operating
facility.1 temperatures.
Fig. 4 summarizes the margins and resultant ROI estimated • Facilities are required to remove impurities from the
for U.S. refiners from 1994 to 2004. It shows the variability in process gas, and prevent freezing and hydrate formation at
the recovered propylene margin during this period. The data low temperature operations.
show a strong recovery from lower margin levels experienced • In the recycle tower, the heavy gasoline components
during 1998 and 1999. This figure and data also suggest that (200°F+) are removed from the overhead C4s and lighter
now is a very good time to invest in propylene-recovery gasoline components. The C4s and lighter gasoline com-
facilities that will come onstream to take advantage of the ponents are recycled to the FCC reactor.

HYDROCARBON PROCESSING / NOVEMBER 2001 49

 Table 3. Light hydrocarbon fluidized catalytic reactor system
ultimate yields
Yields, wt% C4 Raffinate Partially FCC Lt.
Hydrog. C5s Naphtha
Fuel gas 7.2 12.0 13.6
Ethylene 22.5 22.1 20.0
Propylene 48.2 43.8 40.1
Propane 5.3 6.5 6.6

F
Gasoline 16.8 15.6 19.7

Table 4. Steam-cracker integration with fluidized catalytic

reactor system

O
Steam-cracker complex Without With catalytic
material balance, Mtpy catalytic system system
Feed:
Naphtha 1,891 1,995

O
Products:
Ethylene 700 700
Fig. 5. Flow diagram of the light hydrocarbon fluidized catalytic reactor system. Propylene 419 535
Fuel gas 327 299

R
• The deethanizer separates the feed to the column into Gasoline 390 421
P/E ratio 0.60 0.76
an overhead C2 stream and a bottom C3 stream. The over-
head C2s are routed to the C2-splitter, where polymer-

P
grade ethylene product is produced. Table 5. Fixed-bed pilot plant data for waxy gasoil feedstock
• The feed to C3 splitter comes from the bottoms of the over commercial equilibrium FCC catalysts, 950°F cracking
deethanizer. This column produces a polymer-grade pro- temperature at constant conversion
pylene product from the mixed C3 feed.
Amorphous Zeolite
Yields catalyst catalyst
Light hydrocarbon fluidized catalytic reactor system. This Hydrogen, wt% 0.08 0.04
process provides an economic option for petrochemical produc- C1 + C2s, wt% 3.8 2.1
ers (or refiners) to increase propylene production and the overall Propylene, vol% 16.1 11.8
Propane, vol% 1.5 1.3
cracking complex propylene-to-ethylene ratio by using low- i-Butane, vol% 7.9 7.2
value, olefin-rich, light-hydrocarbon feedstocks generally in the n-Butane, vol% 0.7 0.4
carbon range of C4 to C8.7 This method coproduces ethylene at Butylenes, vol% 12.2 7.8
a typical ratio of 1 weight ethylene per 2 weights of propylene C5+ gasoline, vol% 55.5 62.0
LCO, vol% 4.2 6.1
(P/E ~ 2). The gasoline byproduct is highly aromatic and, thus, Bottoms, vol% 15.8 13.9
can contribute octane for gasoline blending or be valued as a Coke, wt% 5.6 4.1
chemical feedstock.
Feedstocks with the highest conversion and best selectivity to Gasoline octane
RON Clear 94.0 89.8
propylene are those rich in olefins. The ideal feedstocks are
pyrolysis generated C4 and C5 streams, which have been selec-
tively hydrogenated to convert acetylenes and diolefins into ole- light ends (C3–) are sent to the neighbor for purification, the
fins. If butadiene has a higher value relative to propylene and the middle cut is recycled to extinction and the gasoline cut is sent
producer has an extraction plant, then Raffinate-1 can be used. onto fuels or chemicals use.
Other possible feedstocks are MTBE Raffinate-2, aromatics-plant As shown in Table 3, a typical C4 raffinate, after butadiene
raffinate and refinery streams that are rich in olefins, such as extraction, yields approximately 65-wt% propylene plus ethylene.
naphthas from the FCCU, coker or visbreaker. Refinery steams do On a similar basis, almost 60% of a partially hydrogenated C5
not require pretreatment nor hydrogenation of dienes; there is no stream is converted to light olefins (P+E). Where refinery-cracked
limit on feed aromatic or diene content. Paraffins are partially naphthas may have low value to the blending pool, however, an
converted with each pass through the reactor and contribute to alternate use would be to upgrade the stream(s) to valuable pet-
ultimate light-olefins yield; thus allowing recycle to extinction rochemicals. Typical light FCC naphtha could ultimately yield
operation. more than 30-wt% propylene and 15-wt% ethylene.
The light hydrocarbon fluidized catalytic reaction system is The light hydrocarbon fluidized catalytic reactor system can
based on years of design and operating experience with FCCUs be used in concert with a steam cracker to increase the product
in the refinery. It is easily integrated into ethylene plants, sharing P/E of the petrochemical cracking complex. Ultimate olefin plant
a common product recovery section. Catalyst is continuously P/E ratios for a low-severity naphtha steam cracker are typically
regenerated, and it is quite robust in terms of feed impurities. No limited to 0.60 to 0.65. However, with newer processing methods,
feed pretreatment is required for typical trace components. The new plants can be designed for P/E ratios of about 0.80. Using
system is comprised of the riser reactor/regenerator vessel, air the fluidized catalytic reactor system within the olefins complex
compressor and catalyst handling, flue-gas system and feed/prod- provides a higher ultimate value as shown in the Table 4.
uct heat recovery equipment. Reactor effluent enters the ethylene Economic analyses show that integrating newer processing
plant recovery section at the main fractionator or process gas methods in the olefins complex has about a 11⁄2-year simple pay-
compressor suction. This effluent may also be processed in a out on gross margin and produces 700 Mtpy ethylene. Also, the
refinery VRU. A refiner in close proximity to a petrochemicals fluidizing catalytic reactor system can be used as a standalone
facility may consider a partial fractionation scheme, where the unit with its own separation section.

50 HYDROCARBON PROCESSING / NOVEMBER 2001

 Phillip K. Niccum is the chief technology engineer
Remarks. For more than 50 years, FCC has been a major of fluid catalytic cracking (FCC) at Kellogg Brown & Root.
contributor to worldwide production of propylene and butylene He has led the process engineering effort on more than
for the expanding petrochemical industry. Due largely to the 35 FCCU revamp projects and 2 grassroots resid FCC
units, and been involved with technology licensing,
rising demand for polypropylene, propylene demand continues proposal management, technical service, plant startup
to outpace the petrochemical industry demand for other light and technology development. Following graduation from
olefins. Much of this increased demand will have to be supplied California State Polytechnic University with a BS degree

F
from fluid catalytic processes that are particularly adept at in chemical engineering in 1980, Mr. Niccum began his
career with Texaco where he provided design and
propylene manufacture. In addition to traditional FCC and resid technical assistance to Texaco-owned and licensed FCC
FCC units, much of the increased demand will likely be satis- units.
fied with new generation fluid catalytic techniques. As these

O
processes proliferate, they are also expected to make significant Michael J. Tallman is a process manager in the
olefins process design group at Kellogg Brown & Root,
contributions to the supply of the most common, yet smallest, where he has worked for more than 21 years. He
petrochemical feedstock—ethylene.   graduated in 1980 with a BS degree in chemical engi-

O
neering from Rose-Hulman Institute of Technology in
Acknowledgment
Terre Haute, Indiana. His background includes experi-
Based on a paper originally presented at the NPRA Annual Meeting, New ence with various grassroots and revamp projects and
Orleans, March 18–20, 2001. studies for olefins plants and pyrolysis furnaces,

R
preparation of proposals and startup advisory services.
LITERATURE CITED
Recently, he has been involved with numerous process
1  inger, S. J., “The Critical Role of the Refinery in the Propylene Market,” 2000 World
Z studies for the Superflex process, including evaluation of both grassroots and
Petrochemical Conference, Houston, March 29–30, 2000.
2 Zinger, S. J., Chemical Market Associates, Inc., e-mail, September 2000. revamp applications.

P
3 Jahnig, C. E., D. L Campbell and H. Z. Martin, “History of Fluidized Solids Development

at Exxon,” Fluidization, Ed. J. R. Grace and J. M. Matsen, Perseus Books, January 1980. Chris Santner is Kellogg Brown & Root’s technology
4 Blazek, J. J, Oil & Gas Journal, 1971. director for FCC and refined products. His experience
5 Whittington, E. l., J. R. Murphy and I. H. Lutz, “Catalytic cracking—Modern designs,” includes both refinery and petrochemical work, but the
American Chemical Society, New York, Aug. 27–Sept. 1, 1972. majority of his career has been focused on refinery proj-
6 Niccum, P. K., R. B. Miller, A. M. Claude, M. A. Silverman, N. A. Bhore, K. Liu, G.
ects, specifically FCC technology. Mr. Santer has been
K. Chitnis and S. J. McCarthey, “MAXOFIN: A novel FCC process for maximizing involved in all phases of FCC including development, pilot
light olefins using a new generation of ZSM-5 additive,” NPRA annual meeting, San
plant, design, startup and troubleshooting. In his current
Francisco, March 15–17, 1998.
7 Gilbert, M. F., M. J. Tallman, W. C. Petterson and P. K. Niccum, “Light olefin produc- position, he is responsible for both the technical and
tion from SUPERFLEX and MAXOFIN FCC technologies,” ARTC Petrochemical business aspects of the technology. He received a BS
Conference, Kuala Lumpur, February 2001. degree in chemical engineering in 1972, served in the
8 Niccum, P. K., M. F. Gilbert, M. S. Tallman and C. R. Sanner, “Future refinery—FCC’s U.S. Navy, and then earned an MS degree in chemical engineering in 1977. Mr.
role in refinery, petrochemical integration, NPRA annual meeting, New Orleans, March Santer is a member of AIChE and is a registered profes-
18 –20, 2001. sional engineer.

Maureen F. Gilbert is the FCC team project leader

for Kellogg Brown & Root. She holds a BS degree in
chemistry from the Georgia Institute of Technology and
a MS degree in chemical engineering from the University
of Houston. Ms. Gilbert has worked in the engineering
and construction field for more than 23 years with
application in several process industries. Her process
and project experience has most recently been applied
to the commercialization of several petrochemical and refining technologies.

HYDROCARBON PROCESSING / NOVEMBER 2001 53