Propylene Production

FCC PROPYLENE PRODUCTION AND TECHNOLOGY ALTERNATIVES TO OPTIMIZE PROPYLENE YIELDS Keith A. Couch Senior Manager UOP LLC  Jim P. Glavin  Product Line Manager UOP LLC Dave Wegerer  Senior Associate UOP LLC Jibreel Qafisheh  Process Specialist UOP LLC

 Abstract: According to Purvin Purvin & Gertz(4), Gertz(4), global FCC capacity is projected to grow by approximately 19% between 2006 and 2015. Across that same time, polymer grade propylene production from FCC units is projected to grow by nearly 32%(3). Although the current average propylene yield from the installed FCC base is 4 to 5 wt%, (weight percent of fresh feed to the FCC unit) many of the new FCC units that come on line over the next ten years will produce even higher propylene yields, some projected to be as high as ~20 wt%.

Pushing higher propylene production from the FCC unit requires the refiner to make some significant economic decisions. The optimum cash cost of production for propylene is an intricate balance between CAPEX, OPEX, and throughput, versus operating severity to produce the highest value product slate. To effectively optimize overall economics, it is important to understand the dynamic impact of varying feed pretreatment severity on FCC product yields. In this paper UOP will discuss the expected world demand for propylene, present a study on the synergies associated with proper integration of Hydroprocessing and FCC technologies on overall product yields, and evaluate the factors that influence the selection of a propylene propylene yield design point for for the FCC unit. UOP will also present other other routes to on purpose propylene, such as our MTO, Olefin Cracking and Oleflex™ technologies.

609 2

Propylene Market and the FCC Unit

The demand for petrochemical feedstocks, particularly the light olefins for the production of polypropylene is expected to increase in the future. According to Purvin & Gertz(4), the global FCC capacity will grow by approximately 19% between 2006 and 2015. Over this same time frame, according to CMAI(3)  the polymer grade propylene demand filled by FCC units will grow by nearly 32%. The FCC process has proven to be a very flexible process and is in a unique position to help satisfy the expected increase in propylene demand. Although its principal application has been to produce gasoline, the FCC unit is frequently operated to maximize other products, such as distillates or LPG. The LPG mode can be considered a step towards a petrochemical mode of operation since it provides enhanced yields of petrochemical feedstocks. With the strong market demand for propylene and the capability of a FCC unit to achieve elevated propylene yields, there is a natural desire to maximize the propylene yield from new FCC units. However, there are competing economic forces that suggest that the optimal yield of propylene from a FCC unit is on the order of 10 – 11 wt%, which is substantially lower than what current technology can produce. One of the primary questions being asked of UOP today is how to properly balance design and operation of the FCC unit between Maximum Gasoline and Maximum Propylene production. The optimum is typically somewhere between these two extremes.  At the same time as the crude diet is changing, the demand for light, clean, highquality fuels has risen to an all-time high due to increasingly stringent fuel specifications. Shifts toward gasoline and diesel fuels combined with regulations requiring ppm levels of sulfur and demand for improved combustion properties to help reduce emissions of polyaromatics and NOx NOx are driving this increase. increase. The challenge for today’s refiner is to identify and implement technology which produces lighter, higher-quality products containing virtually no sulfur from crudes containing large concentrations of sulfur, low native yields of naphtha and distillate or both.  As is the case today, propylene production p roduction is projected to come from a number of sources, both refinery and petrochemical-complex petrochemical-complex based. based. On the refining side, increased propylene production from FCC units is expected to be a major contributor to the onpurpose requirement. requirement. On the petrochemical petrochemical side, there there are more more alternative routes to propylene available than than ever before. These alternatives include include propane dehydrogenation, dehydrogenation, methanol-to-olefins, and olefin conversion including metathesis and olefin cracking processes. Each of these alternatives alternatives can offer competitive competitive economics in certain certain situations.

610 3

Refinery Configuration Analysis

The optimum configuration for a refinery is very specific to geography, local markets and individual core business objectives. While the optimum complex configuration will vary from client to client, understanding some of the integration synergies will help establish the fundamental heart of the complex. Refiners that operate with an objective of “maximum propylene” are most often integrated with a petrochemicals complex, both for the production of polypropylene and the production of benzene, toluene, and xylene (BTX). With this objective, the core of the complex becomes a Coker, hydroprocessing unit, a FCC unit, and a Platforming™ unit; see Figure 1.

Coker

Dry Gas Propylene LPG

Hydroprocessing Complex VGO Hdt (Option 1) Raw VGO

C5 & C6

FCC

C7 – C9 Platforming BTX Unit C10 – C11

MHC (Option 2)

LCO

HCK (Option 3)

MCB

Figure 1: Refinery Conversion Configuration

The optimum design and operation of this conversion complex is an economic balance between the severity of hydroprocessing applied to the VGO feeding the FCC unit, the severity of FCC operation to make petrochemical feedstocks, and the severity of the Platforming unit to optimize BTX production. The proper design envelope for each of these units is heavily impacted by answering two questions; 1) how much hydrogen should be put in the FCC feedstock, and 2) how severely should the FCC Unit be operated”.

611 4

Impact of VGO Hydroprocessing on FCC Performance

To address the question of how much hydrogen should be put in the FCC feedstock, we considered three different hydroprocessing options: (1) hydrotreating, (2) mild hydrocracking, and (3) hydrocracking. The higher the hydrogen content of the FCC feedstock, the higher the propylene selectivity on feed; however, the higher conversion operation of the hydroprocessing unit reduces the remaining feedstock to the FCC unit. To demonstrate this relationship, UOP conducted a set of sequential pilot plant studies that quantified the results very effectively. In this pilot plant study, a single source raw Light Arabian VGO was used for all of the testing. The raw VGO was processed at different levels of severity: 15 lv-% conversion to emulate hydrotreating, 45 and 55 lv-% conversion to emulate mild hydrocracking, and 70 lv-% conversion to emulate hydrocracking. All of these runs were completed using commercially available hydrotreating and hydrocracking catalyst systems available through UOP. The full-range of products from each of the hydroprocessing runs were collected, then fractionated in an Oldershaw lab to collect the 343 C+ material as feedstock to an FCC pilot plant. The product properties of this material are shown in Table 1. 

Table 1 –  Hydroprocessing Conversion and 343 C+ Properties

HP Conversion, lv%  API Gravity UOP K 

15

45

55

70

30.25 12.29

37.88 12.92

39.06 12.91

40.37 13.01

The FCC pilot plant was inventoried with a commercially produced “maximum LPG” equilibrium catalyst that was high in ZSM-5 content. Additional lab deactivated ZSM-5 additive (25% ZSM-5 content), was used to blend the total level of ZSM-5 in the equilibrium catalyst system to 20%. The FCC pilot plant was run in RxCat mode of operation, with a riser outlet temperature of 566 C (1050 F) utilizing a regenerated catalyst-to-oil ratio of 10, and a carbonized catalyst-to-oil ratio of 10. The results are shown in Table 2. 



Table 2 –  FCC Product Selectivities with Hydroprocessing Severity HP Conversion, lv-%

15

45

55

70

Methane, wt%

2.8

2.8

3.1

2.9

Ethylene, wt%

4.7

5.9

6.2

6.3

Ethane, wt%

2.1

2.4

2.7

2.6

Propylene, wt%

16.6

22.4

23.1

23.7

Propane, wt%

2.0

2.3

2.4

2.4

Total C4’s, wt%

19.1

25.3

25.8

27.0

612 5

 Although progressively higher severity hydroprocessing resulted in greater propylene selectivity, when considered on a constant crude basis, the total tonnes of propylene produced were dramatically reduced; see Table 3 and Figure 2.

Table 3 –  FCC Product Selectivities with Hydroprocessing Severity HP Conversion, lv-%

0

15

45

55

70

Low Severity

High Severity

Hdt VGO Feed to FCC, BPSD

130,000

110,500

71,500

58,500

39,000

Propylene Produced, k-MTA

N/A

895

743

623

423

(PetroFCC Propylene Yields)

   %    t   w    d    l   e    i    Y   e   n   e    l   y   p   o   r    P

30.0

1000

28.0

900

26.0

800

24.0

700

22.0

600

20.0

500

18.0

400

16.0

300

14.0

200 Propylene

12.0 10.0 0

10

20

30

40

50

60

K-MTA 70

   A    T    M      K

100 80

0

Volume Percent Hydrocracking Severity

Figure 2: FCC Propylene Yields with Hydroprocessing Severity 

This data was subsequently used to support a commercial project, with 130,000 BPSD of raw vacuum gas oil feeding a hydroprocessing complex. In this case the refiner determined that increased hydrogen content of the VGO and contaminant removal improved the conversion capability in the FCC unit and decreased downstream product treating requirements. Further optimization of the complex identified that optimized product

613 6

values were obtained by loading the VGO hydrotreater with a catalyst system that maximized hydrogen uptake for hydrodenitrification and improved aromatics saturation with minimal barrels converted. To further validate this analysis, the data was used to help populate the overall complex LP model. This LP model included every process unit in the overall refinery and the petrochemicals complex. Each of the configurations was ranked on a basis of NPV divided by CAPEX, and the clear winner was the VGO Hydrotreater; see Table 4. Table 4 –  FCC Propylene Yields with Hydroprocessing Severity Primary Configuration

Coking + VGO Hydrotreating

NPV/ CAPEX 0.70

+ FCC Coking + Mild Hydrocracking

0.60

Coking + Hydrocracking

0.40

+ FCC

It is clear from the data herein that for refiners targeting high propylene production, the VGO should only be hydrotreated to the extent required to achieve contaminant removal targets. Over-conversion of the VGO results in a barrel loss to the FCC unit that dramatically outpaces the increase in propylene selectivity. Catalyst systems that maximize hydrogen uptake with minimal converted barrels should also be considered. FCC Design Conditions and Yields as a Function of Operating Severity

The second portion of this analysis focused on optimizing the design and operating severity of the FCC unit. Increased propylene production from FCC units has been a widely discussed topic over the past three to five years, not only with new unit construction, but also with existing units. One of the primary questions being asked of UOP today is how to properly balance design and operation of the FCC unit between Maximum Gasoline and Maximum Propylene production. The optimum is typically somewhere between these two extremes. The average propylene yield from the installed FCC base is around 5 wt% (on fresh FCC unit feed, exclusive of recycle) on a global basis. Many of the new FCC units that come on line over the next ten years will produce even higher propylene yields, some with design points as high as ~20 wt%. With the strong market demand for propylene and the capability of a FCC unit to achieve elevated propylene yields, there is a natural desire to maximize the propylene yield from new FCC units. However, there are competing economic forces that suggest that the

614 7

optimal yield of propylene from a FCC unit is on the order of 10 – 11 wt%, which is substantially lower than what current technology can produce. The yield pattern for the FCC unit is a continuum of operating severity and process design that can be optimized for refinery specific economics and objectives. The optimum process design provides the refiner with the flexibility to move up or down the most economic range of the propylene yield curve; see Figure 3.

60

   F    F

   %      t   w20  ,    d    l   e    i    Y   e   n 10   e    l   y   p   o   r 5    P

   F    F

   %      t   w  ,    d    l   e    i    Y 30   e   n    i    l   o   s   a    G

 f (Rx T, P  P  , C/O, P T  )

Maximum Gasoline

Gasoline + LPG

Maximum Propylene

Figure 3: FCC Unit Design and Operating Modes 

While propylene generation from a FCC unit certainly varies with feedstock quality, catalyst formulation and the use of LPG additives such as ZSM-5, from a mechanical design perspective it is primarily a function of reactor temperature, hydrocarbon partial pressure, catalyst-to-oil ratio and total pressure. With a full range hydrotreated VGO the technology exists to operate over a range from about 5 wt% to around 20 wt% propylene on feed. It is important to note that higher propylene production comes at the expense of gasoline. As we work with refiners to meet their processing objectives, we see that three design modes emerge: 1) Maximum Gasoline  which is traditional of most U.S. refiners, 2) Gasoline + LPG   for refiners that want the most market flexibility, and 3) Maximum Propylene   for true petrochemical applications. The inflection point typically defines the optimum, because any increase above this point requires a greater change in the operating severity for the same change in propylene yield. It is important to note that the inflection point occurs in the Gasoline + LPG mode of operation, not maximum Propylene.

615 8

Table 5 below provides the yields for 3 different modes of operation for a FCC unit processing a typical 24 °API full range hydrotreated Vacuum Gas Oil (VGO) feedstock.

Table 5 –  FCC Product Yield Comparisons across Operating Modes

Ethylene Ethane Propylene Propane Butylenes i -Butane n-Butane Debutanized Gasoline LCO Clarified Oil Naphtha Composition  Aromatics Benzene in Gasoline

Gasoline 0.83 0.90 4.76 1.84 6.62 3.92 1.21

wt% Yield on Fresh Feed Gasoline + LPG 1.42 0.94 10.50 3.52 9.62 4.87 1.51

Propylene 7.10 1.21 18.10 2.18 9.83 2.98 0.82

54.36 11.57 7.93

43.94 10.10 6.89

35.21 8.32 5.59

34.70

44.00

54.70

0.46

0.59

1.29

 Yield Comparison

 As the operation of the FCC unit is shifted towards the higher propylene production there is a coincident increase in ethylene and butylene. Along with this shift towards light products, there is also a decrease in gasoline yield and a change in the gasoline composition. While most refiners expect that higher propylene generation comes at the expense of gasoline yield, what is often not understood is that the quality of the gasoline is progressively reduced the more severe the operation of the FCC unit. This is due to both existing aromatics being concentrated in less gasoline, as well as the production of additional aromatics. In this case, total aromatics increased by 58%, and benzene increased by 280%. With gasoline benzene limits already in force, the high benzene content of the propylene mode is often not suitable for gasoline blending without either extraction or saturation. For most refiners, maximum propylene operation reduces gasoline quality and devalues the product. Refiners that practice propylene mode operation typically process the FCC naphtha through a naphtha hydrotreater and a Platforming unit as feed preparation upstream of a Petrochemical Complex for the production of benzene, toluene, and xylene (BTX). However, high severity FCC operation actually reduces the overall aromatics production by reducing

616 9

precursors that would be more selectively converted to BTX in the Platforming unit; see Table 6 and Figure 4. Table 6 –  FCC + Platforming Unit Aromatics Production across Operating Modes Modes of Operation Gasoline + LPG 12,840

Gasoline 18,230

Feed to Platforming Unit, BPSD  Aromatics Production C8 Aromatics, t/hr C9 Aromatics, t/hr C10+ Aromatics, t/hr Total Aromatics, t/hr FCC Aromatics, % of Production FCC Propylene, t/hr FCC Dry Gas, t/hr Total Aromatics + C3= - Dry Gas, t/hr

Propylene 10,015

21.77 17.78 18.83 79.01 46.15 14.31 8.43

16.18 14.32 15.70 61.81 62.71 31.62 11.11

13.31 12.10 12.07 51.06 77.63 54.49 31.44

84.90

82.32

74.06

*Basis: 50,000 BPSD feedstock to the FCC Unit

90   r    h    /   s   e   n   n   o    t  ,    t   c   u    d   o   r    P    l   a    t   o    T

90

80

80

70

70

60

60

50

50

40

Total Aromatics

40

30

Propylene

20

Aromatics + C3= - Dry Gas

10

30

FCC Aromatics, % of Total

4

6

Maximum Gasoline

8 10 12 14 Wt% Propylene Yield from FCC

Gasoline + LPG

16

18

20 20

  s   c    i    t   a   m   o   r    A    l   a    t   o    T    f   o    %  ,   s   c    i    t   a   m   o   r    A    C    C    F

Maximum Propylene

Figure 4: FCC + Platforming Unit Aromatics Production across Operating Modes 

 As the operating severity of the FCC unit is increased through the Gasoline + LPG mode of operation, the percentage of aromatics produced in the FCC with respect to the total production (FCC + Platforming unit) steadily increases. However, the higher operating

617 10

severity of the FCC unit reduces the naphtha feedstock to the Platforming unit faster than the increased rate of aromatics from the FCC unit, resulting in a net lower aromatics production from the complex. This relationship is shown by the pink and blue lines in Figure 4. Polymer-grade propylene has higher product value than that of mixed BTX. Although the aromatics production does go down with higher FCC severity the propylene production goes up, thereby offsetting the reduced value from the production loss on aromatics. This remains true as the FCC severity is increased through the gasoline + LPG production mode. However, as severity is increased towards maximum propylene generation, the dry gas (C2-) production increases by a factor of 3.7 times base. Although a large percentage of the dry gas production is ethylene, recovery of the ethylene is expensive and most refiners consider the ethylene market to be outside of their core business objectives, and as such, devalue ethylene to fuel gas. Worse yet, refiners can push themselves into situations where the higher severity operation of the FCC unit results in a fuel gas long situation requiring them to either cut back on severity or capacity. For most petrochemicals producers, the gasoline + LPG mode appears to be a reasonable balance between propylene and BTX production with negligible product devaluation to fuel gas. Similarly, for the fuels focused refiner, the gasoline + LPG mode appears to be a reasonable balance between the world demand for higher propylene production and the need to maintain acceptable gasoline blend stock quality. In the Gasoline + LPG mode we are able to obtain a 180% increase in propylene production with only a 28% increase in the benzene content of the gasoline. Although the benzene content increases, more than 80% of this increase is due to the concentration of existing benzene production in the gasoline as a result of selectively cracking olefinic naphtha to LPG with the use of ZSM-5 additives. It is also important to note in Table 5 that butylene production hits a plateau around medium severity; so if the FCC unit is being operated to produce alkylation feedstock, medium severity operation is sufficient. Whether operating for fuels for petrochemicals, with the intricate balance of hydrotreating, FCC, and Platforming unit technology integrations, optimizing the overall complex LP is critical to defining the proper design and operating envelope for the refinery. In the analyses completed by UOP, although the technology currently exists to operate the FCC for maximum propylene production, this does not appear to be an economic optimum for most. To cross-check this result, the next section of this paper will discuss what UOP has observed in the market place with respect to several recent commercial FCC projects.

618 11

FCC Commercial Production Targets

With the strong market demand for propylene and the capability of a FCC unit to achieve elevated propylene yields, there is a natural desire to maximize the propylene yield in terms of wt% from the design of new FCC units. There has been a significant increase in FCC capacity licensed over the past few years, and a clear trend in refiners’ requests for greater propylene production from those units. With these projects, many refiners have started with the objective to push propylene yields towards the upper limits of what the equipment, catalyst, and feedstock can produce. However, there are competing economic forces at hand and real market data that suggests an optimal yield of propylene from a FCC unit is on the order of 10 – 12 wt%, which is substantially lower than the theoretical limits associated with most feeds and operating systems. Table 7 shows the initial propylene targets as originally cited in the request for quotation (RFQ) and the final design basis for seven new FCC projects over the past two years. While propylene production from a FCC unit is desired, the information in Table 7 strongly suggests that there are economic forces that push the refiner from maximum propylene back towards the Gasoline + LPG mode.

Table 7 –  Initial RFQ Propylene Requests versus Final Design Point  wt% Propylene Yield on Fresh Feed Initial RFQ Final Design Unit “A” --10.5 Unit “E” 20.0 8.0 Unit “J” 20.0 10.5 Unit “P” 18.0 15.1 Unit “R” 21.0 11.3 Unit “S” 15.0 11.1 Unit “T” 16.0 10.0

The optimum cash cost of production for propylene from a FCC unit is an intricate balance of capital, throughput, operating severity and overall product values. Refiners often optimize their FCC unit by maximizing converted barrels (throughput), minimizing their operating and capital costs and producing a flexible product slate. The main problem with pushing the limits of propylene production from a FCC unit is that in doing so, all of the above optimization factors are negatively impacted. The operating conditions needed to maximize propylene yield as a weight percent of fresh feed require significantly larger equipment per barrel processed, resulting in a higher capital cost. To move the operation from Gasoline to Propylene mode derates the operating capacity by approximately 50%; i.e. if a unit is designed for a Gasoline mode throughput of 50,000 BPSD, to maximize severity to Propylene mode operation in the same equipment, the feed rate would need to be reduced to approximately 25,000 BPSD. The operating costs associated with maximum propylene production are higher than for Gasoline mode operation. This further penalizes the economics of maximum propylene production. Lower hydrocarbon partial pressure to maximize propylene selectivity requires additional steam use, and maximizing C3= over C4= requires lower absolute operating pressure, both contributing to larger vessel requirements per barrel throughput. The

619 12

catalyst systems for maximum propylene production command a premium; and although all patents and royalty requirements for ZSM-5 expired globally as of January 01, 2007, the costs for ZSM-5 additives remain high. Lastly, the increased production of LPG and net lower molecular weight of the reaction products increases the overall compression costs for product recovery.  As a net result of detailed CAPEX, OPEX, and product value evaluations, we are not surprised to see the market selecting an optimal propylene production that is far less than theoretical limits. While there will always be exceptions, the ideal balance point across a variety of feedstocks and market regions appears to be around 11 wt% propylene yield. When determining the optimum design and operating point for a unit, it is important to understand the relationship between propylene selectivity as a function of percentage of feed processed, and yield as a function of tons produced. Within a fixed unit size, the economic influence of feed rate is much greater than that of operating severity. In almost all cases, it is better to maximize throughput over severity. To help demonstrate this relationship, consider a FCC unit that is constrained based on maximum cyclone velocity. At this point, there is a trade-off to be realized between operating severity and throughput. If throughput is increased, the operating severity must be reduced; likewise, if severity is increased, throughput has to be reduced. By carefully evaluating this relationship, UOP can work with the refiner to determine the most economic unit design. This relationship is demonstrated in Figure 5.

Existing Reactor Constrained at 65 ft/sec Cyclone Inlet     )    F   o   r   0   o   0  ,   0   r   )    1   g   o   i     s  ,    (    t   p   T   n   (   e   e  e  r   c   u   r   r    t   e  u   a   s    P  s  r   e   e   r   p    P  m   e    T

100 90 80 70 60 50 40 30 20 10 0

30

40

50

60

70

80

1300 1250 1200 1150 1100 1050 1000 950 900 850 800

  y   a    d    /   s   e   n   n   o    T  ,   n   o    i    t   c   u    d   o   r    P   =    3    C

Hydrotreated Feed, kBSD C3= wt-% FF Reactor Temp (T–1000°F) C3= Production Tonnes/day

Reactor Pressure Conversion vol-%

Figure 5: Unit Size versus Throughput and Severity

Figure 5 shows a set of operating variables for a unit that is constrained on cyclone velocity. In this example we start with a unit designed to operate at 35,000 barrels per day, at a maximum severity to achieve 20 wt% propylene on feed. The combination of operating variables results in a maximum recommended cyclone inlet velocity. As the feed rate is

620 13

increased, one of the other operating variables (reactor temperature, steam rate, or total pressure) has to be changed, to maintain the same flowing volume of products to the cyclones. Increased reactor pressure has the least detrimental effect on propylene selectivity; as such, it is the first variable to be moved. The propylene yield from feed rate outpaces the selectivity loss with increased reactor pressure. Although the weight percent propylene on feed goes down, the actual tonnes per day of production go up. This relationship can be maintained until the reactor nears a design pressure limit. At this point, any additional throughput requires the reactor temperature to be reduced and propylene production starts falling rapidly. In this example, we have been able to double the feed rate to the unit and end up with nearly the same tonnes per day of propylene production. The peak propylene production (tonnes/day) occurs at a propylene yield of approximately 15 wt%, which is well under the maximum theoretical propylene yield of the feedstock. Since the other key products such as gasoline have good market value, refiners have the incentive to push capacity higher and accept lower propylene yields.

FCC Propylene Production

There have been many published market projections over the past few years indicating a shortage of propylene supply to the market. Purvin and Gertz project the world-wide FCC capacity to grow at an average annual rate of 2.7%, while the 2005 CMAI data indicates that propylene production from FCC units will increase at an average annual rate of 4.3 % from 2005 to 2015. This is consistent with the data presented in Table 7 showing that the weight percent of propylene produced from the average new FCC unit is expected to increase over historical norms. However, it is also important to note that not all of the future growth has to come from “new” FCC units. There is definitely a place for leveraging existing assets to help close the market gap on propylene. The average propylene production from FCC units in Europe is about 4.7 wt%. This is significantly lower than the 10 – 11 wt% optimum discussed herein. There are two ways to increase propylene production from existing FCC units: 1) improve the recovery capability of the existing gas concentration unit, and 2) increase the quantity of propylene produced in the FCC reactor. CMAI estimates that on average the recovery of polymer grade propylene from FCC units across the world is approximately 67%. Irrespective of where the propylene product is sent, the economic incentives to improve recovery have increased dramatically over the past few years. The incentive for the refiner to invest in gas con recovery projects is the value differential between propylene and natural gas. Prior to 2003, the value gap was only about $215 per tonne, which made it difficult to justify recovery projects. However, since that time, the value gap has dramatically increased to around $900 per tonne, making C3 slippage to fuel gas much more costly to the refiner; see Figure 6.

621 14

1400 1200   e   n   n   o    T    /    $

1000

Polymer Grade Contract Value Refinery Grade Contract Value Natural Gas

800 600

$900

400

$215

200 0 Dec ‘98

Dec ‘99

Dec ‘00

Dec ‘01

Dec ‘02

Dec ‘03

Dec ‘04

Dec ‘05

Dec ‘06

Figure 6: Propylene and Natural Gas Values

 An optimized FCC unit should operate between 3 to 5 mol-% C3+ in the fuel gas. When operating below 3 mol-% C3+ there is a risk of over-absorption of H 2S in the gas concentration unit which can result in a pH imbalance that can lead to elemental sulfur precipitation in the gasoline and/or accelerated corrosion in the system. Operating the Gas Con above 5 mol-% C3+ in the fuel gas results in a downgrade of valuable product. If we consider the operation of a 50,000 BPSD FCC unit, at a value gap of $900 per tonne, slipping from 3 to 7 mol% C3+ in the fuel gas results in a product value loss of about $4.6 MM per year; see Table 8. Table 8 –  Propylene Product Down-Grade(1) C3+ mol-% in Dry Gas(2)

3 mol-% 5 mol-% 7 mol-%

C3= Slippage to Dry Gas (tonne/day) 10.2 17.4 24.8

C3= Downgrade to Natural Gas(4) ($/day) $0 $6,480 $13,460

=$4.6 MM/year

Notes:

(1) Economics based on a 50,000 BPSD FCC unit. (2) Additional slippage of C 3+ was considered at 61.2% C 3= and 38.8% C3. (3) C3= was valued at “Refinery Grade” as a feedstock to an alkylation unit valued at $680 per metric tonne per CMAI for October 2006. (4) C3= was valued as natural gas at $195 per metric tonne per CMAI for October 2006.

With respect to increased operating severity, the incentive to move higher on the propylene production curve is the value gap between propylene to alkylate feed and regular gasoline; see Figure 7. There have been substantial periods of time over the past 3 years when this differential has been quite significant, extending to around $450 per metric ton. The market does appear to be taking advantage of this opportunity. UOP has recently

622 15

completed over 500 k-BPSD of revamp designs for existing FCC capacity with the objective of increasing propylene yields in the range of 7 to 10 wt%.

1400 1200

C3= to Alkylate Regular Gasoline

$450

1000

  e   n   n 800   o    T    / 600    $

400 200 0 Dec ‘98

Dec ‘99

Dec ‘00

Dec ‘01

Dec ‘02

Dec ‘03

Dec ‘04

Dec ‘05

Dec ‘06

Figure 7: CMAI U.S. Short-Term Propylene Prices 

Revamp Unit Designs

The feasibility to employ the absorption based gas con with high LPG and low naphtha yields has occasionally been challenged, based on a perception that there simply is not enough naphtha absorbent available to maintain system efficiency. While in fact, there is more than enough. The primary absorber is designed to circulate both stabilized and wild naphtha absorbent; see Figure 8. In gasoline mode operation, 80% of the absorbent is wil d naphtha from the main column overhead receiver and 20% is stabilized naphtha from the debutanizer. As the unit is moved to maximum propylene generation, the absorbent ratio changes towards 80% stabilized naphtha and 20% wild naphtha, but the net liquid leaving the bottom of the absorber stays essentially the same. This is a very important point, as the same equipment capacity is applicable over a large range of operating severity.

623 16

Operating  Mode

Gasoline: C3= + Aromatics:

Wild  Naphtha

Stabilized  Naphtha

0.8 X

0.2 X

0.2 X

0.8 X

 Absorber/  Stripper 

C 1 / C 2  Main Column

X

C 3 / C 4 / C 5 

Figure 8: Absorption Based Gas Con Flexibility

This flexibility in the absorption based FCC gas con design has enabled many refiners in recent years to shift their operations toward higher propylene yields in the range of 7 to 12 wt% within existing main equipment constraints while maintaining high levels of light olefin recovery. Balancing the target propylene generation, percent recovery and the cost of modifications to achieve the refiner’s objectives are an important part of every revamp study. When trying to maximize the value of existing assets, the additional capital required to achieve a propylene recovery of 97% may not be economic when compared to that of a 96% recovery case. A single percentage change in design recovery can require substantial equipment modifications or replacement. A process study normally reveals unit limitations and provides the refiner with answers that lead to the most economic case for revamp implementation. UOP has been involved in many unit revamps with high propylene recovery targets. In one of the more aggressive cases, UOP worked with the client to achieve a 58% increase in throughput while simultaneously achieving a 30% increase in propylene yield over the original nameplate design. To achieve this goal, the unit was progressively debottlenecked in stages so as to identify the true limits of equipment components enabling the best use of capital investment to achieve the refiner’s objectives. While no two units are ever designed or operated exactly the same, there are some typical capacity constraints that require revamp to enable higher severity or higher throughput operation of the gas concentration unit. These include:



Wet Gas Compressor Capacity:  The increase in light ends and LPG make will increase the load in the wet gas compressor. However, for most revamps the reactor pressure is normally increased to accommodate the reactor cyclone design within

624 17

existing reactor shell constraints. Increasing the reactor pressure results in an increase of the compressor suction pressure, which often offsets the decrease in molecular weight to the compressor. Due to the high cost of wet gas compressor replacement, in many cases the operating conditions for the revamp are set within the maximum capacity of the existing compressor casing. It is important to note that rotating equipment vendors have also improved their technical offerings over time, and can often re-rate existing equipment beyond previously considered limitations.









Fractionators/Absorbers Capacity: In most revamps, the capacity constraints of trayed columns can be overcome with the use of high capacity trays, such as MD™ trays, or packing. The typical limiting areas are the HCO section of the main fractionator, top section of the debutanizer, the stripper, and the absorber. Cooling Temperature and Absorber Lean Oil Circulation: For propylene yields less than 12 wt%, a traditional cooling water heat removal system is generally sufficent to meet the desired propylene recovery. A chilled water system can be applied if the unit has other constraints that can not be overcome; however, they are not typically required. Absorber lean oil circulation is normally the primary variable that can be adjusted for a revamped unit to achieve a higher propylene recovery. Heat Integration with the FCC Main Fractionator: Proper design of the main fractionator heat integration with the gas concentration unit is typically the most complex part of most revamps. Different scenarios of heat exchange with the main fractionator are normally identified to stay within the main fractionator limitations and reduce equipment modifications.  Alternative Absorber Configuration:  The conventional UOP scheme can be modified to increase propylene recovery to the stripper bottoms and C2  minus rejection to the primary absorber lean gas. The traditional recycle of stripper overhead vapor back to the inlet of the high pressure condenser is eliminated. This off-loads the condenser and the high pressure receiver, allowing for more economic new unit design, and greater potential to retain existing equipment in revamp situations. This also provides a means to revamp units for higher throughput or severity of operation that are constrained by plot space to increase high pressure condenser duty. The stripper feed preheater can also be eliminated to improve feed conditioning to the stripper to improve propylene recovery. These changes can result in greater than 99% propylene recovery and C2 minus rejection sufficient to make polymer grade propylene specification while eliminating the need for a downstream deethanizer column.

625 18

 Alternative Routes to Propylene

The many alternatives available for propylene production offer a wide variety of opportunities but determining the right route to propylene can be confusing. There are certain factors that apply to each alternative that can narrow the choices and help determine which one is best for a particular project.

Propane Dehydrogenation (PDH)

Since 1990, propane dehydrogenation has been providing a growing source of propylene for petrochemical applications. There are currently eight plants in operation producing approximately 2.5% of the worldwide propylene supplied for petrochemicals. Six of these plants use the Oleflex™ process licensed by UOP. Five additional Oleflex process units are currently in design or construction phases which will add another 2 million metric tons of propylene capacity when these units come on-stream over the next few years. The Oleflex process uses a proprietary platinum on alumina catalyst. Four adiabatic reactors are operated in series as shown in Figure 9. The dehydrogenation reaction is endothermic so interheaters are included between each reactor to maintain the desired reactor temperatures. The Oleflex process uses a CCR™ regenerator to continuously regenerate the catalyst and maintain high conversion and selectivity.

C  C   R

Rx

Rx

Rx

Rx

Heaters

Propane

Propylene Recovery H2

Figure 9: C3 Oleflex Process

626 19

The yield of propylene from propane feedstock is over 85 wt% with the Oleflex process. The amount of ethylene produced is very small, such that it is usually not recovered. The ethylene together with the other reactor byproducts is typically used to supplement the fuel consumption for a propane dehydrogenation unit. This means that propylene is the only product produced from a propane dehydrogenation unit unless there is a local demand for the hydrogen produced by the dehydrogenation reaction. PDH offers opportunities for simple back-integration for propylene derivative producers looking for a secure, economical source of propylene. PDH produces a single product (propylene) so there is no need to market any co-products unless there is a local need for hydrogen, which can easily be recovered from an Oleflex unit. The key for any PDH project is the propylene-propane price differential. About $200/MT is the minimum long-term average price differential required between propylene and propane in order to achieve good economics with an Oleflex complex. The price differential has been ranging between $300/MT to more than $500/MT over the past few years, which has driven the growth in PDH capacity. It is not true that PDH is only feasible when tied to discounted propane. In fact, most of the PDH plants that have been installed are located where propylene is needed as opposed to where “cheap” propane is located. PDH offers three specific advantages to propylene derivative producers. First, a PDH plant makes a single product, propylene. A company specifically interested in producing propylene derivatives may not want to produce ethylene or C4+ co-products that are made from naphtha crackers, or gasoline and fuels from refineries. PDH focuses an investment specifically on propylene capacity. Second, production costs for a PDH plant are tied to the cost of propane. Propane prices are not tied directly to naphtha prices or the propylene market; therefore, PDH allows large propylene derivative producers to diversify the overall cost structure of their feedstock. Finally, some of the best locations for propylene derivative plants do not have good access to byproduct propylene. Given the high cost of shipping and storing propylene, PDH is generally more cost-effective than buying propylene for these locations, if propane is available. PDH requires a relatively low capital investment compared to other grass-roots alternatives for producing similar amounts of propylene. Good economies of scale are achieved with unit capacities of 250,000 MTA or larger. UOP’s Oleflex process offers experience and reliability with six commercial units in operation. Over 1.25 million MTA of propylene capacity has come on-stream using the Oleflex process since 1990 and the onstream Oleflex process capacity will exceed 3 million MTA of propylene within the next few years.

627 20

MTO

The UOP/HYDRO MTO process (Figure 10) was jointly developed by UOP and Norsk Hydro for the selective production of ethylene and propylene from methanol derived from natural gas or coal. MTO combines well proven FCC and naphtha cracker technologies with a proprietary new catalyst from UOP. The catalyst used in the process is based on a silicoaluminophosphate, SAPO-34. The technology has been extensively demonstrated in a demo plant by Norsk Hydro and more than fifteen years of development have been completed. The MTO process converts methanol to ethylene and propylene at close to 80% carbon selectivity in a fluidized bed reactor. The carbon selectivity approaches 90% if MTO is combined with olefin cracking technology as discussed later in this paper.

Quench  Reactor Regenerator  Tower 

Caustic Wash

 De-C 2

 De-C 1

C 2 C 3  Splitter   De-C 3  Splitter   De-C 4  Tail Gas Ethylene

Regen Gas

Propylene  Dryer 

Mixed C4

DME Recovery C 2 H 2  Reactor  C5+

Air

Propane

Water

Ethane Methanol Water

Figure 10: UOP/HYDRO MTO Proces s

The MTO reaction is exothermic. Carbon or coke accumulates on the catalyst and must be removed to maintain catalyst activity. The coke is removed by combustion with air in a catalyst regenerator system. Other co-products include very small amounts of C1-C4 paraffins, hydrogen, CO and CO2, as well as ppm levels of heavier oxygenates that are removed to ensure that the product olefins meet polymer-grade specifications. The UOP/HYDRO MTO process offers the greatest flexibility of any propylene producing technology. The ratio of propylene/ethylene product can range from less than 0.8 to more than 1.3. When combined with the TOTAL PETROCHEMICALS/UOP Olefin Cracking process (to be discussed later) to convert the heavier olefins, the propylene/ethylene product ratios can range from less than 1.0 to more than 2.0.

628 21

MTO is part of a two-step process, which converts natural gas to methanol followed by the conversion of methanol to light olefins. MTO projects are driven by the desire to minimize costs of production by utilizing cost-advantaged natural gas or coal and the market demands for ethylene and propylene and their derivatives. In many locations, natural gas prices are not directly tied to crude oil and naphtha market prices so MTO provides another means for olefin derivative producers to diversify the cost structure for their feedstocks. MTO can provide much lower costs of production and higher returns on investment than naphtha crackers when crude oil market prices are above $20 per barrel. MTO projects are tied to mega-scale methanol projects for two reasons. First, the amount of methanol required for an MTO project is consistent with the methanol production from today’s latest world-scale methanol projects (5,000 to 10,000 t/d methanol). Second, the costs of methanol production from a mega-scale methanol unit are low enough to support alternative uses for methanol such as fuel applications or MTO. MTO plants can be located near or integrated with the methanol plant or they can be located separately with the methanol plant located near the gas source or coal fields and the MTO plant located near the olefin markets or olefin derivative plants.

Olefin Conversion

Olefin cracking technology offers an alternative to upgrade C 4 to C8  olefins to propylene and ethylene at high propylene to ethylene ratios. In mid-2003, the TOTAL PETROCHEMICALS/UOP Olefin Cracking process was introduced after TOTAL PETROCHEMICALS and UOP completed extensive development and demonstration activities. Following initial work by TOTAL PETROCHEMICALS in the mid 1990s, UOP and TOTAL PETROCHEMICALS formed a joint-development alliance in late 2000. The development activities included successful operation of a demonstration unit, catalyst performance testing in pilot plants, feed-yield determination, catalyst manufacturing scaleup and process design development. The demo unit, started in 1998 at an industrial facility associated with TOTAL PETROCHEMICALS in Antwerp, Belgium, processes commercial feedstocks from operating plants. The demo unit includes feed pretreatment, a reactor section, catalyst regeneration facilities, and internal recycle capabilities. The TOTAL PETROCHEMICALS/UOP Olefin Cracking process (Figure 11) features fixed-bed reactors operating at temperatures between 500 and 600°C and pressures between 1 to 5 bars gauge. The process utilizes a proprietary zeolitic catalyst from UOP and provides high yields of propylene at propylene/ethylene product ratios of about 4.0. The catalyst minimizes the reactor size and operating costs by operating at high space velocities and high conversions and selectivities without requiring an inert diluent stream.

629 22

 A swing reactor system is used for catalyst regeneration. The layout of the product separation facilities is dependent on how the olefin cracking unit is integrated with other processing plants such as a naphtha cracker, an FCC or an MTO unit (Figure 12).

Light Olefin Product

Olefinic C4 -- C C8 Feed

Depropanizer  Column

SHP OC  OC Reactor 

C4 & & C5 Purge

Recycle

Rerun Column

C6+ Purge

Figure 11: TOTAL PETROCHEMICALS/UOP Olefin Cracking Process

Naphtha

Product Recovery & Product Purification Recovery &

Furnace Furnace

C4/C5 C5/C6 Paraffin-Rich

Light Olefins

 Refineries

Gas Oils

FCC

 

FCC

 Naphtha Crackers

C4-C5 Olefins

C2= C3=

Olefin Cracking

C3=

Olefin Cracking

 MTO Plants

Olefin Olefin Recovery LPG Light Olefins

Figure 12: Olefin Cracking Applications

630 23

Light Olefins

Gasoline Cycle Oils

Olefin Cracking

C4-C8

C2=

MTO

MeOH

C2= C3= Fuel Gas Pygas

C4+

Olefin conversion technologies offer opportunities for upgrading low-value heavier olefins to light olefins via either metathesis or olefin cracking. These technologies favor high yields of propylene so they are especially well suited for projects looking to maximize propylene production. Olefin conversion units integrate with other olefin production and recovery facilities such as ethylene plants (steam crackers), refinery FCC units, or MTO units. The TOTAL PETROCHEMICALS/UOP Olefin Cracking process was developed to utilize low value byproduct streams containing C4 to C8 olefins from steam crackers, fluid catalytic cracking (FCC) refineries, and/or methanol-to-olefins (MTO) plants.5 When combined with steam crackers, the process allows the steam cracker to expand capacity and achieve higher propylene/ethylene production ratios. For FCC refineries, the technology utilizes FCC byproduct streams to increase propylene and ethylene production while reducing the olefin content of gasoline blending streams with little or no loss of octane. For MTO complexes, the technology increases the yield of light olefins for a given amount of methanol feedstock and reduces or eliminates the C4+ byproduct streams. (Refer to Figure 12 for an overview how Olefin Cracking can integrate with these other processes) The TOTAL PETROCHEMICALS/UOP Olefin Cracking process is generally economical when propylene to ethylene price ratio is between 0.8 and 1.2. Approximately 0.25 tons of ethylene are produced per ton of propylene produced. Olefin cracking feedstocks can include Raffinate-1, Raffinate-2 and/or C5, C6, C7, or C8 olefins from naphtha crackers, FCC, cokers, or MTO. The feedstocks can be pooled together from various sources to achieve good economies of scale while reducing the amounts of C5 olefins that are blended into gasoline. The C4-C8 purge streams from olefin cracking can be recycled to cracker furnaces to further supplement olefin production. Summary

Global propylene demand trends remain strong, and with the change towards lighter feedstocks in new steam crackers there will be a growing reliance on FCC units to balance the supply side of the propylene equation. The technology exists today to help make this happen. UOP believes that the FCC process is flexible enough to meet the challenge associated with closing the global market gap for propylene. We believe that this will happen through a combination of new units designed for elevated propylene yields and revamps of existing facilities to increase propylene yield and recovery within logical equipment constraints. Refiners that operate with the objective of “maximum propylene” are most often integrated with a petrochemicals complex for the production of polypropylene, benzene, toluene, and xylene (BTX). With this objective, the core of the complex becomes a Coker, hydroprocessing unit, a FCC unit, and a Platforming unit. It is clear from the data herein that the VGO should only be hydrotreated to the extent required to achieve contaminant removal targets. Over-conversion of the VGO results in a barrel loss to the FCC unit that

631 24

dramatically outpaces the increase in propylene selectivity. Catalyst systems that maximize hydrogen uptake, hydrodenitrification, and aromatics saturation with minimal converted barrels should be used. There are competing economic forces and real market data that suggests that the optimal yield of propylene from a FCC unit is on the order of 10 – 12 wt%, which is substantially lower than the theoretical limits associated with most feeds and operating systems. The premium that propylene and butylene command over other products, such as gasoline, is sufficient to provide many refiners the incentive to invest in their existing assets. UOP’s expertise in overall refinery complex design enables us to work with the refiner to optimize new units and revamp designs for the best overall economic return on investment.

References

1. Thakkar, Ackelson, Rossi, Dziabala, and McGehee, Innovative Hydrocracking  Applications for Conversion of Heavy Feedstocks”, 2007 NPRA Annual Meeting, Publication AM-07-47 2. Chemical Market Associates Inc. (CMAI), “2004 World Light Olefins Analysis”, www.cmaiglobal.com 3. Chemical Market Associates Inc. (CMAI), “2006 World Light Olefins Analysis”, www.cmaiglobal.com 4. Purvin & Gertz, “2006 Global Petroleum Market Outlook”, www.purvingertz.com 5. U.S. Energy Information Administration (EIA), “Annual Energy Outlook”, www.eia.doe.gov 6. Couch, Glavin, Wegerer and Qafisheh, “FCC Propylene Production – Closing the Market Gap by Leveraging Existing Assets”, 2007 NPRA Annual Meeting, Publication AM-07-63 7. Houde, Mark, “Rebalance Gasoline Surplus by Maximizing FCC Propylene”, ERTC 2005, www.uop.com/objects/RebalanceGasolineSurplus.pdf  8. Leliveld, RG, “STARS Ketjenfine 860 – Exceptional New Catalyst for Hydrocracking Pretreat”, Albemarle Catalyst Courier, Spring 2007, Issue 67 9. Houdek, and Anderson, “On-Purpose” Propylene – Technology Developments, ARTC 8th  Annual Meeting

632 25