Grace - High-Performance Specialty Chemicals and …...The GENESIS TM Catalyst System www. grace.com 7500 GraceDrive¥Columb ia,MD21044USA Mos trefiners need flexiblecatalystsystems

NUMBER 105 Spr ing 2009

Upgrading Bottoms

The GENESISTM Catalyst System

www.grace.com 7500 Grace Drive • Columbia, MD 21044 USA

Most refiners need flexible catalyst systems that allow them to

take advantage of changing economic scenarios. Grace Davison

delivers this flexibility with the GENESISTM catalyst system. GENESISTM

catalysts provide a means to maximize yield potential through optimization of

discrete cracking catalyst functionality. GENESISTM catalysts are a blend of

two catalyst types. The key component is MIDAS®, which maximizes conver-

sion of bottoms and improves coke selectivity by eliminating coke precursors.

The other GENESISTM component is most often an IMPACT® catalyst. The

inclusion of IMPACT® provides critical zeolite surface area and activity as well

as superior coke and gas selectivity in a broad range of applications, from

severely hydrotreated gasoils to heavy resid feeds.

GENESISTM catalysts demonstrate a true yield synergy with a superior

coke to bottoms relationship than either component alone. The synergy

exists because each component cracks specific feed species.

Maximize yield withGrace Davison’s

GENESISTM FCC catalyst.

GENESISTM catalyst providesthe ultimate in formulation

and FCC operational flexibility.

For more information, contact your Grace Davison Technical Sales Manager

A message from the editor...

Dear Refiners:

During this current global economic crisis, petroleum refiners face different,

more difficult challenges than they have in past recessions. Gasoline demand is

falling in many of the developed countries, leading to decreased refinery utiliza-

tion. At the same time, more capacity will come online in the developing regions,

further straining refinery margins.

One constant has been the refiner’s need to crack bottoms. Once again,

Grace Davison’s innovative research and development team has delivered a state-

of-the art bottoms cracking catalyst in MIDAS®-300 technology. MIDAS®-300 is an

extension of our successful MIDAS® platform, which has been commercially

proven to upgrade the bottom of the barrel in over 100 applications. While our

lead article features commercial applications desiring max LCO make, MIDAS®-

300 can also be formulated with higher zeolite to matrix ratios to maximize conver-

sion to gasoline while maintaining excellent bottoms selectivity.

In this issue, we also discuss processing of opportunity feedstocks, espe-

cially synthetic crude, in hydrotreating and fluid catalytic cracking. Future refiners

will be called on to deliver optimum yields from these feeds on short notice. Both

ART and Grace Davison have developed catalyst and processing strategies to

meet these demands.

We hope you find these articles offer solutions to your challenging refining

environment. For more information, please contact your Grace Davison or

Advanced Refining Technologies representative.

Joanne DeadyVice PresidentMarketing/R&DGrace Davison Refining Technologies

IN THIS ISSUE

Recycle Strategies and MIDAS®-300 for Maximizing FCC Light Cycle OilBy Ruizhong Hu, Hongbo Ma, Larry Langan, Wu-Cheng Cheng, David HuntGrace Davison Refining TechnologiesTo take full advantage of the increased value of FCC light cycle oil relative togasoline, the refiner should optimize product cutpoints, operating conditionsand FCC catalyst technology. Strategies to maximize LCO include proper FCCcatalyst selection and defining optimum conditions for recycle of heavy cycleoils. Davison Circulating Riser Pilot Plant Studies reveal the optimum recyclestream compositions and rates taking into consideration the complete FCCproduct slate and the unit operating constraints.

Distillate Pool Maximization by Exploiting the use of OpportunityFeedstocks Such as LCO and SyncrudeBy Brian Watkins and Charles OlsenAdvanced Refining TechnologiesThe use of opportunity feedstocks in the FCC feed such as LCO, dieselstreams from Hydroprocessing Units and feeds from Synthetic Crudes hashelped refiners to maximize their diesel pool in light of increasing ULSDdemands. We highlight differences in feed reactivity for various feed compo-nents considering the discrete chemical composition. We also explore theimpact of various Hydrotreating Catalysts and operating conditions on dieselproduction.

Characterization and Catalytic Cracking of Synthetic Crude FeedstocksBy Michael Ziebarth and Rosann SchillerGrace Davison Refining TechnologiesThere will be a dramatic increase in the production of synthetic crude from oilsands in the coming years. This paper explores the effect of these feedstockson the FCC operation in terms of catalyst selection and FCC product slate.We present laboratory development work as well as commercial experience.

Factors Influencing ULSD Product ColorBy Greg Rosinski, Charles Olsen and Brian WatkinsAdvanced Refining TechnologiesProduct Color of petroleum products such as kerosene, jet fuel, diesel fuel andlube oils is a concern. Unit cycle length can be shortened due to productcolor degradation. In this paper we identify components that contribute tocolor degradation and report on the effects of feedstocks, chemical composi-tion, operating conditions (EOR temperatures), product cut points and catalystselection on ULSD color.

1

CATALAGRAM 105Spring 2009

Managing Editor:Joanne Deady

Contributors:Wu- Cheng Cheng

Ruizhong HuDavid HuntLarry LanganHongbo MaCharles OlsenGreg RosinskiRosann SchillerPhyl StrawleyKristen WagnerBrian Watkins

Michael Ziebarth

Please addressyour comments to

[email protected]

©2009W. R. Grace & Co.-Conn.

The information presented herein is derived from our testing and experience. It is offered, free of charge, for your consid-eration, investigation and verification. Since operating conditions vary significantly, and since they are not under our con-trol, we disclaim any and all warranties on the results which might be obtained from the use of our products. You shouldmake no assumption that all safety or environmental protection measures are indicated or that other measures may not berequired.

NUMBER 105 Spr ing 2009

Upgrading Bottoms

Grace Davison Refining Technologies7500 Grace Drive • Columbia, MD 21044 • 410.531.4000

www.e-catalysts.com

15

23

34

Catalagram 105 Spring 2009 1

David HuntFCC Technical Service Manager

Ruizhong HuManager of Research & TechnicalSupport

Hongbo MaResearch Engineer

Larry LanganResearch Engineer

Wu-Cheng ChengDirector of R&D

Grace DavisonRefining Technologies

aximizing FCC light cycle oil(LCO) yield to take advantageof high diesel prices relative

to gasoline requires re-optimization ofproduct cut point, operating condi-tions and catalyst technology. It is wellknown that the LCO-to-gasoline ratiocan be increased through loweringconversion by adjusting FCCU operat-ing conditions and decreasing cata-lyst activity.[1,11] The drawback of thisapproach is the increase in bottomsyield. Recycle is often required to fullymaximize LCO while maintaining bot-toms yield consistent with a traditionalmaximum gasoline operation.

This article is a general discussion ofstrategies to maximize LCO in theFCCU. We will present laboratory

results which quantify the effects ofvarious recycle streams. A residfeedstock was cracked over a lowZ/M MIDAS® catalyst in GraceDavison’s circulating riser pilot plant(DCR). The product bottoms wasdistilled to five recycle fractions(650-750˚F, 650-800˚F, 650-850˚F,650˚F+ and 750˚F+), blended backat various levels with the originalfeedstock and cracked over aMIDAS® catalyst in the ACE unit.Laboratory testing results wereused to model a commercial opera-tion to demonstrate the yield advan-tage of selecting the appropriaterecycle stream, recycle ratio andcatalyst technology.

Recycle Strategiesand MIDAS®-300for MaximizingFCC Light Cycle Oil

M

www.e-catalysts.com2

Figure 1LCO and Bottoms vs. Conversion

1.0 Introduction

The Energy Information Administra-tion (EIA) expects higher dieselprices relative to gasoline to contin-ue through 2010. An average dieselprice of 2.27 $/gallon is expected in2009, increasing to an average 2.54$/gallon in 2010. Gasoline isexpected to be an average 1.87$/gallon in 2009, reaching an aver-age 2.18 $/gallon in 2010.[2]

New Corporate Average FuelEconomy (CAFE) standards arealso expected to increase dieseldemand relative to gasoline inupcoming years. This new standardrequires auto manufacturers toboost fuel mileage to 35 mpg by2020 and applies to all passengerautomobiles, including light trucks.To meet this challenging new stan-dard, more efficient vehicles pow-ered by hybrid and diesel enginesare expected.

Refiners are increasing the produc-tion of LCO from their FCCU’s totake advantage of the significantlyhigher value of diesel relative togasoline. Figure 1 shows how LCOand bottoms shift versus conversion

for a high and low zeolite/matrix ratiocatalyst. LCO, like gasoline, is anintermediate product increasing withconversion at very low conversion lev-els, eventually reaching an over-crack-ing point. Past the over-crackingpoint, LCO yield declines with increas-ing conversion. This high conversionregime represents the traditional

FCCU operating point. A low Z/Mcatalyst generally produces higherLCO at the expense of bottoms for agiven conversion levels, as suggest-ed by Figure 1.

Refiners tend to focus on the follow-ing strategies to maximize FCC LCOproduction:

1. Reduced gasoline end pointa. Increased gasoline endpoint

b. Higher LCO endpoint

2. Operating conditionsa. Lower reactor temperatureb. Higher feed temperaturec. Lower equilibrium catalystactivity

3. Feedstocka. Removal of diesel rangematerial from the FCC feed-stock

b. FCC feed hydrotreating sev-erity optimization

c. Residual feedstock optimiza-tion

4. Catalyst Optimizationa. Increasing bottoms conver-sion

b. Lower zeolite to matrixsurface area

Conversion, wt.%

90

80

70

60

50

40

30

20

10

0

24

22

20

18

16

14

12

10

Bottoms, wt.%

806040200 806040200

LCO, wt.%

High Z/M Low Z/M

Conversion, wt.% 54 58 68 75Rx Exit Temp, ˚F 950 950 971 970Regenerator Temp, ˚F 1350 1350 1270 1270Feed Temp, ˚F 701 574 700 299C/O Ratio 4.3 5.0 5.9 9.4Dry Gas, wt.% 2.0 1.9 2.6 2.2LPG, wt.% 8.2 8.9 11.4 13.3Gasoline, wt.% 38.4 42.0 48.0 51.9LCO, wt.% 22.2 21.7 19.2 16.7Bottoms, wt.% 24.0 20.0 12.8 8.6Coke, wt.% 5.2 5.3 5.9 7.1

Boiling Point Distribution of 650˚F+ Bottoms650-700˚F 5.3 4.8 3.5 2.5700-750˚F 4.8 4.2 2.9 2.0750-800˚F 4.3 3.6 2.2 1.4800-850˚F 3.6 2.9 1.6 1.1850-900˚F 2.5 2.0 1.2 0.7900-950˚F 1.6 1.2 0.8 0.5950˚F+ 1.9 1.4 0.8 0.4650-750˚F 10.1 6.3 4.4650-800˚F 14.4 12.6 8.5 5.9650-850˚F 18.0 15.5 10.1 6.9650˚F+ 24.0 20.0 12.9 8.6750˚F+ 13.9 11.0 6.5 4.1

9.0

Table IDCR Runs to Generate Recycle Feedstock


c. Maintaining C3+ liquid yieldand gasoline octane

5. Recycle streamsa. Heavy Cycle Oil (HCO) orbottoms

Reducing gasoline end point is asimple and effective way to increaseLCO production. The LCO flashpoint specification and main frac-tionator salting often determine howlow a refiner can reduce the LCO ini-tial boiling point (reduce the gaso-line end point). The LCO end pointshould be increased within the max-imum main fractionator bottomstemperature, slurry exchanger foul-ing and diesel hydrotreater con-straints.

Maximizing LCO in the FCCU atreduced conversion without produc-ing incremental bottoms oil is thetrue challenge. Shifts in operatingconditions to reduce conversionincrease LCO, but also increase bot-toms oil yield, as seen in Figure 1.In order to produce increased LCOwithout producing incremental bot-toms, refiners will often reformulatetheir FCC catalyst to a lower activitywith lower zeolite-to-matrix forimproved bottoms conversion and tominimize LCO over-cracking.

Recycle is eventually required tominimize bottoms production as therefinery reduces conversion toreach an optimal LCO yield, as sug-

gested by Figure 1. Almost all FCCunits operated with large recyclestreams prior to the introduction ofzeolite catalyst in the 1960’s and1970’s. However, since the 1970’s,recycle was generally removed fromFCC’s as the catalyst and equipmenttechnology improved and the FCCoperated at high conversion and fee-drates to produce gasoline, C4’s andC3’s. As a result, our industry has lim-ited recent experience with theserecycle streams and their effect onFCC products and coke.

To better understand these recycleeffects with the state-of-the-art maxi-mum bottoms cracking catalystMIDAS®, Grace has conducted aseries of riser pilot plant (DavisonCirculating Riser or DCR), ACE, andcomputer simulation studies.

2.0 Experiments

2.1. DCR Pilot Plant Runs andPreparation of Recycle Streams

A commercially available MIDAS® cat-alyst was deactivated, without Ni or V,at 1465°F for 20 hours, using theAdvanced Cyclic Propylene Steamprotocol described by Wallenstein.[3]

After deactivation, the catalyst had 94m2/g zeolite surface area, 83 m2/gmatrix surface area, and a unit cellsize of 24.30Å. The deactivated cata-lyst was charged in our DCR pilotplant[4], where cracking of a resid

feedstock was conducted.Reaction severity was varied byadjusting the temperature set pointsof riser top, regenerator, and feedpre-heat. We obtained four bal-anced runs with conversion levels of54, 58, 68, and 75 wt.%. The DCRconditions and product yields arelisted in Table I. The C4- productswere analyzed by gas chromato-graph, while C5+ liquid products(syncrude) were analyzed by simu-lated distillation and expressed asgasoline (C5-430˚F), LCO (430-650˚F) and bottoms (650˚F+), asshown in Table I. The detailed boil-ing point distribution of the bottomsfraction is also provided in Table I.These results provide the amount ofhydrocarbon in a given boilingrange when an ideal distillation isachieved. These results were usedas a basis to determine the maxi-mum quantity of each recyclestream.

C5+ liquid products from each DCRrun was first separated by atmos-pheric distillation on a modifiedHempel still (ASTM D295) to obtainthe 650˚F+ fraction. Each 650˚F+fraction was further separated byvacuum distillation (ASTM 1160) toobtain the desired boiling fractions.The properties of the various boilingfractions are shown in Figures 2 and3.

2.2. ACE of Recycle Blends

To simulate HCO and bottoms recy-cle, we prepared feed samples byblending various boiling range frac-tions back into the starting residfeedstock. These feed blends, listedin Table II, can be separated intotwo groups. One group consists ofrecycle fractions with various boilingranges obtained at 54 wt.% conver-sion, while the other group consistsof recycles with one boiling range,650˚F-750˚F, but obtained at variousconversion levels from the DCRruns. The percentage of recycle ineach blend was selected based onsimulated distillation listed in Table Iand the strategy to keep the recyclefraction low enough so that one canuse two-pass cracking to simulate

Conversion, wt.% Recycle Stream, ˚F Blend Ratio wt.% Original Feed wt.% ˚API100.0 20.60

54 650-750 8.3 91.7 20.4254 650-750 6.3 93.7 20.3954 650-800 11.7 88.3 20.3754 650-800 9.7 90.3 20.3854 650-850 13.4 86.6 20.2954 650-850 11.4 88.6 20.3054 650+ 14.7 85.3 19.8354 750+ 7.1 92.9 19.8758 650-750 91.7 20.2958 650-750 6.3 93.7 20.2968 7.3 92.7 19.9568 650-750 5.3 94.7 20.0375 650-750 5.4 94.6 19.7275 650-750 3.4 96.6 19.93

8.3

650-750

Table IICombined Feeds Used in ACE Study


steady state operation. The steadystate approximation will be dis-cussed further in the Data Analysissection below. The HCO streamswere blended at two recycle ratiosto demonstrate the sensitivity andreproducibility of yield changes dueto recycle.

The ACE runs[5] were conducted usingthe same laboratory deactivatedMIDAS® catalyst as above and thesame commercial FCC resid feed asthe base feedstock. All ACE runswere conducted at a reactor tempera-ture of 950˚F using the same amountof feed of 1.5g and a constant feed

delivery rate of 3.0g per minute. Inorder to achieve desired conver-sion, catalyst to oil ratio was variedby changing the amount of catalystcharged in the reactor in each run.As in the above DCR study, gas andliquid products were analyzed by

Figure 2Properties of Recycle Feedstocks Obtained from DCR Run at 54% Conversion

˚API Gravity

20

15

10

5

Feed ID

Fresh

Feed

650-

750˚

F

650-

800˚

F

650-

850˚

F

650˚

F+

750˚

F+

800˚

F+

850˚

F+

1000

900

800

700

600

50% vol.% ˚F

Conradson Carbon, wt.%

Hydrogen, wt.%

16

12

8

4

12

11

10

9

0

Fresh

Feed

650-

750˚

F

650-

800˚

F

650-

850˚

F

650˚

F+

750˚

F+

800˚

F+

850˚

F+

Dat

a

Figure 3Properties of 650-750˚F Recycle Fraction Obtained from DCR Runs

API Gravity @ 60% ˚F

20

15

10

5

Feed ID

Fresh

Feed

650-

750˚

F@

54%

Con

50% vol.% ˚F

Conradson Carbon, wt.%

Hydrogen, wt.%

4.8

3.6

2.4

1.2

0.0

Dat

a

12

11

10

9 600

900

800

700

8 500

650-

750˚

F@

58%

Con

650-

750˚

F@

68%

Con

650-

750˚

F@

75%

Con

Fresh

Feed

650-

750˚

F@

54%

Con

650-

750˚

F@

58%

Con

650-

750˚

F@

68%

Con

650-

750˚

F@

75%

Con


gas chromatography and simulateddistillation. Coke on catalyst wasmeasured using a LECO analyzer.

2.3. Data Analysis

In steady operation with HCO recy-cle, it is conceivable that somehydrocarbon molecules could gothrough the riser multiple times. Inour DCR-ACE experiment, we areapproximating steady state yieldswith yields from two-pass cracking.We will use the Feed-Element-Tracking-Approach to discuss thevalidity of this approximation.Consider 100 grams of oil, which isfed into the FCC unit and crackedinto various products, of which thebottoms are partially recycled. Forexample, 10 grams of bottoms arerecycled and fed into the unit againto crack further. Additional prod-ucts are obtained, and some of theresulting bottoms, e.g., 1 gram, arerecycled, and so on. By collectingthe products along the route of this100 gram crude oil, we can get theproduct yields on the fresh feedbasis. The process is shown inFigure 4. RR is the recycle ratio,defined as the fraction of the recy-cled HCO stream in the total feedinto the unit. RR is equal to 0.1 inthe following example.

Using the Element-Tracking-Approach, we can calculate theLCO yield as follows:

LCO = LCO1 + RR x LCO2 + RR2 xLCO3 + • • • + RRi-1 x LCOi

(1)

where LCOi is the LCO yield of i-thpass cracking of the recycled por-tion of (i-1)-th pass bottoms. Since

the quality of the feed becomes worseand worse when recycling further,LCOi < LCOi-1. Obviously, if the recy-cling ratio RR is small, the secondorder and above term of RR can besafely ignored. In this work, the maxi-mum RR is 0.15; so, the third term onthe right-hand-side of Equation 1 isonly about 2.25% of the first term.

Therefore, if we can get LCO2, we willhave a reasonable estimate for LCO.Looking at the second pass cracking,the total feed in this pass consists of(1-RR) fresh feed and RR recycledHCO from the first pass cracking. Thissecond pass corresponds to the ACEstudy in our DCR+ACE experiment.Denoting the LCO yield in the ACEstudy as LCO’, LCO1 as the LCO yieldin the cracking of the base feed(which corresponds to the first pass),and noting that the portion of the freshfeed in the combined feed has theLCO yield of LCO1, we have:

LCO’ = (1-RR) x LCO1 + RR x LCO2(2)

which can be rearranged to:

LCO2 = LCO1 + (LCO’ – LCO1)/RR(3)

The yields on a fresh feed basis aredetermined as follows:

LCOFF = LCO’/(1 – RR)(4)

BotFF = (Bot’ – RR)/(1 – RR)(5)

The Element-Tracking-Approach pre-dicts that at low (<15%) recycling

ratio, two-pass cracking is veryclose to the steady state operationbecause the higher order term inthe yield expression is negligible.

3.0 Results and Discussion

3.1. Effect of Recycle Streams

Table III shows the interpolatedyields of the original feed at 70 and55% conversion, as well as theyields of the combined feeds at55% conversion. The yields areexpressed as wt.% of the total feed(fresh + recycle). To better illustratethe contribution of each recyclestream, the yields of LCO, bottoms,coke, and gasoline, as a function ofthe recycle ratio, are plotted inFigure 5. With the exception of the750˚F+ recycle feed, all recycle-containing feeds made higher LCOand lower bottoms than the originalfeed. With the exception of the 650-750˚F recycle feed, all recycle-con-taining feeds made higher coke andlower gasoline than the originalfeed. The data quality confirms thatthe ACE testing has the sensitivity tomeasure the yield contribution ofthe recycle streams at the desiredrange of recycle ratios.

Using the Element-Tracking-Approach described earlier, we cancalculate the theoretical yieldsderived from the second-passcracking of each of the recyclestreams. This is shown in Figure 6.The recycled streams are lesscrackable than the base feed, asindicated by the much higher cat tooil ratios required to achieve thesame conversion. This is expected,as the easy to crack material of therecycle streams has been crackedin the first pass. The crackability ofthe recycle streams increases withthe API gravity (Figure 7). Asexpected, the 650-750˚F streammade the most LCO and gasolineand the lowest coke for a given con-version, when comparing yieldsamong the recycle streams. Thetrends in LCO and gasoline yieldsfrom the lightest stream (650-750˚F)to the heaviest stream (750˚F+)appear to be continuous and con-

Figure 4Schematic Diagram of Feed-Element-Tracking-Approach

Feed RR*Feed RR2*Feed

FCCUnit

FCCUnit

FCCUnit


Table IIIInterpolated Yields of Base and Combined Feeds at 70% and 55% Conversion

Recycle Boiling Range None None 650-750˚F 650-750˚F 650-800˚F 650-800˚F 650-850˚F 650-850˚F 650˚F+ 750˚F+Recycle Ratio 0.0% 0.0% 8.3% 6.3% 11.7% 9.7% 13.4% 11.4% 14.7% 7.1%Wt.% Conversion 70 55 55 55 55 55 55 55 55 55Cat-to-Oil Ratio 6.0 3.4 3.5 3.5 3.6 3.6 3.7 3.6 3.7 3.6Hydrogen, wt.% 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1Total C1's & C2's, wt.% 1.4 1.0 1.0 1.0 1.1 1.1 1.1 1.1 1.0 1.1Propylene, wt.% 3.3 2.1 2.2 2.1 2.2 2.2 2.2 2.2 2.1 2.0Total C3's, wt.% 3.9 2.4 2.5 2.4 2.5 2.5 2.6 2.5 2.4 2.4Total C4='s, wt.% 5.1 3.9 3.9 4.2 3.8 3.9 4.0 4.0 3.9 3.8Total C4's, wt.% 8.5 5.6 5.7 6.0 5.6 5.7 5.9 5.8 5.8 5.5C5+ Gasoline, wt.% 49.4 40.6 40.8 40.4 40.2 40.1 39.8 39.8 39.9 40.0RON 89.6 89.5 89.3 89.6 89.5 89.5 89.4 89.5 89.5 89.3MON 78.7 77.6 77.6 77.7 77.7 77.7 77.8 77.7 77.8 77.6LCO, wt.% 20.5 24.7 25.8 25.5 25.6 25.5 25.1 25.0 25.0 24.7Bottoms, wt.% 9.5 20.3 19.2 19.5 19.4 19.5 19.9 20.0 20.0 20.3Coke, wt.% 6.7 5.6 5.6 5.6 5.7 5.6 5.7 5.7 6.1 6.1

Coke Burn Limited 1.00 1.20 1.20 1.21 1.18 1.20 1.18 1.19 1.11 1.10Wet Gas Limited 1.00 1.53 1.49 1.46 1.49 1.48 1.44 1.45 1.48 1.54Catalyst Circulation Limited 1.00 1.76 1.72 1.71 1.70 1.68 1.62 1.69 1.63 1.68Fresh Feed RateCoke Burn Limited 1.00 1.20 1.10 1.13 1.04 1.08 1.02 1.05 0.95 1.03Wet Gas Limited 1.00 1.53 1.36 1.37 1.32 1.33 1.25 1.29 1.26 1.43Catalyst Circulation Limited 1.00 1.76 1.58 1.60 1.50 1.52 1.40 1.50 1.39 1.56

Relative Feed Rate

Fresh Feed Rate

Figure 5Plots of Interpolated Yields at 55% Conversion vs. Recycle Ratio

LCO, wt.%25.8

25.5

25.2

24.9

24.6

6.00

5.75

5.500.00 0.04 0.08 0.12 0.16 0.00 0.04 0.08 0.12 0.16

20.4

20.1

19.8

19.5

19.2

40.8

40.4

40.0

Recycle Ratio

Bottoms, wt.%

Coke, wt.% C5+ Gasoline, wt.%

650-750˚F 650-800˚F 650-850˚F 650+ 750+


sistent with the trend in the APIgravity (Figure 2). However, theincrease of coke appears to be verygradual up to the 650-850˚F streamand becomes stepwise higher forthe 650˚F+ and 750˚F+ streams. Thecoke yield trends very closely with theConradson Carbon (Figure 7), whichis concentrated in the 850˚F+ range(Figure 2). These results suggest thatduring first-pass cracking, coke pre-cursors in the boiling range of850˚F+ are formed. These molecules

Figure 6Theoretical Yields of Second Pass Cracking at 55% Conversion of Recycle Streams

Cat-to-Oil Ratio6

5

4

3

Feed ID

Coke, wt.%

C5+ Gasoline, wt.%

LCO, wt.%

Dat

a 2

40

30

20

10

Base

650-

750˚

F

650-

800˚

F

650-

850˚

F

650˚

F+

750˚

F+

40

35

30

25

12

8

10

20

Base

650-

750˚

F

650-

800˚

F

650-

850˚

F

650˚

F+

750˚

F+

6

4

Figure 7Effect of API Gravity and Conradson Carbon on Catalyst

to Oil Ratio and Coke Yield at 55% Conversion

Cat-to-Oil Ratio vs. ˚API Gravity

0.0 1.5 3.0 4.5 6.014 16 18 20 22

6.0

5.5

5.0

4.5

4.0

3.5

7

6

˚API Gravity

Cat

-to

-Oil

Rat

io

Coke vs. Conradson Carbon

9

8

11

10

13

12

5

Conradson Concarbon

Co

ke

Base Feed

Recycle Streams

are responsible for coke productionduring second-pass cracking.

While the 750˚F+ stream is not a prac-tical recycle stream, it does providevaluable insight on the negativeimpact of recycling heavy bottoms.This stream made more than doublethe coke yield of the base feed. Aclose examination of the hydrocarboncompounds by GC Mass Spec (TableIV) shows that the 750˚F+ fractioncontains higher aromatic compounds,

and in particular tetra-aromaticcompounds, than the lighter 650-750˚F+ fraction. It is likely that thecoke precursors formed duringfirst-pass cracking are indeed thetetra-aromatic compounds. Wenoticed that Ye and Wang[6] report-ed slightly less coke formation(0.6%) with recycling of highly aro-matics bottoms in FCC unit.However, their recycling ratio wasmuch lower, only 1.5%.

3.2. Modeling Overall Yields

Table III also lists the interpolatedyields for max gasoline operation at70 wt.% conversion of the basefeed. Compared to the yields at70% conversion, the LCO yield at55% conversion is higher while theyields of wet gas and coke aremuch lower and the C/O ratio islower. If the unit changes from maxgasoline (70% conversion) to maxLCO (55% conversion) operation,one should be able to increase totalfeed rate until the unit reaches cokeburn, wet gas compressor or cata-lyst circulation constraint, assumingthere are no other limitations. Theresults of Table III suggest that thecoke burn constraint will bereached much sooner than the wetgas or catalyst circulation con-straint. (Catalyst circulation couldbe a limit at reduced catalyst activi-

ty.) At a coke burn limitation, thecombined feed rate of the max LCOoperation is 10 to 20% higher thanat the max gasoline operation.

The data analyses so far have beenconfined to yields on the combinedfeed basis with the selected recycleratios. The following examplesdemonstrate how to use this data todetermine the recycle stream andrecycle ratio to optimize LCO pro-duction. We will examine a maxi-mum recycle case and a constantbottoms case. A modeling of acommercial unit (heat balanced) willbe presented later.

Case 1. Maximum Recycle

The goal of the calculation was tomaximize recycle ratio of each recy-cle stream until the coke yield of thebase feed at 70% conversion wasreached. The hydrocarbon yields,on the fresh feed basis, calculatedusing the Element-Tracking-Approach are shown in Table V. Inthe cases of the 650-750˚F and650-800˚F streams, the maximumavailable recycle levels, based onSIMDIST (Table I), were reachedbefore the coke yield limit was

reached; therefore, the maximumrecycle ratio was used.

The highest LCO yield of 30.2% wasachieved with maximum recycle(14.4%) of the 650-800˚F HCOstream. The next highest LCO yield of29.9% was achieved with 15.6% recy-cle of 650-850˚F HCO stream. Eventhough the 650-750˚F stream had thebest yields by the Element-Tracking-Approach, because it was limited to amaximum recycle ratio of 10.1%, thecombined feed with 650-750˚F streammade only 28.9% LCO and muchhigher bottoms. In the case of the650˚F+ bottoms stream, due to cokelimitation, only 15% out of the avail-able 24% recycle stream could berecycled. The operation recycling650˚F bottoms made lower LCOthroughput and higher bottomsthroughput than the operation withrecycling HCO streams of 650-800˚Fand 650-850˚F. Thus, it is advanta-geous to recycle HCO rather than bot-toms. Gasoline yields on fresh feedbase for all the recycling streams areabout 4-6% higher than that of thecase without recycling, which corrob-orates the results reported byFernandez et al.[7]

Case 2. Constant Bottoms

The goal of this calculation was toadjust the recycle ratio of eachrecycle stream until the bottomsyield of the base feed at 70% con-version was reached. The hydro-carbon yields, on the fresh feedbasis, are shown in Table VI. In thiscase, all the combined feeds withHCO recycle had higher LCO selec-tivity than bottoms (650˚F+) recycle.The difference also comes from thecoke yield differences, which allowthe feeds with HCO recycle to beprocessed at higher feed rates thanthe feed with bottoms recycle.Again, this example shows it isadvantageous to recycle HCOrather than bottoms.

3.3. Effect of Conversion Level

The objectives of this work were todetermine how the composition ofthe HCO stream changed with con-version and how recycling HCO,obtained at varying conversion lev-els, affected the LCO yield. Asdescribed earlier, DCR syncrudesamples obtained at 54, 58, 68 and75 wt.% conversion were distilledand the 650-750˚F fraction of each

Table IVGC Mass Spec Analysis of Bottoms Fractions

from Resid Cracking in the DCR

54 wt.% Conversion 68 wt.% Conversion650-750˚F 750˚F+ 650-750˚F 750˚F+

SATURATES Avg., wt.% Avg., wt.% Avg., wt.% Avg., wt.%C(N)H(2N+2) Paraffins 5.2 4.3 3.4 1.2C(N)H(2N) Monocycloparaffins 7.2 8.7 3.6 3.1C(N)H(2N-2) Dicycloparaffins 4.9 5.4 2.4 2.7C(N)H(2N-4) Tricycloparaffins 3.1 3.9 1.5 2.0C(N)H(2N-6) Tetracycloparaffins 0.0 0.2 0.0 0.1

TOTAL SATURATES 20.4 22.4 11.0 9.1MONOAROMATICS

C(N)H(2N-6) Alkylbenzenes 4.3 6.1 2.6 2.6C(N)H(2N-8) Benzocycloparaffins 2.4 2.3 0.6 0.6C(N)H(2N-10) Benzodicycloparaffins 1.6 1.1 0.3 0.1

DIAROMATICSC(N)H(2N-12) Naphthalenes 4.3 2.9 3.4 1.8C(N)H(2N-14) 14.5 3.2 15.6 1.9C(N)H(2N-16) 21.2 7.7 24.4 6.4

TRIAROMATICSC(N)H(2N-18) 13.9 9.3 18.9 9.8C(N)H(2N-22) 3.3 25.0 4.2 38.5

TETRAAROMATICSC(N)H(2N-24) 0.0 8.7 0.0 13.6C(N)H(2N-28) 0.0 1.1 0.0 2.2

TOTAL AROMATICS 65.3 67.4 70.1 77.5


syncrude was collected and ana-lyzed (Figure 3). The 650-750˚Ffractions were blended with theoriginal feed and tested in the ACE.

The theoretical yields of the sec-ond-pass cracking on a unit recyclefeed basis are calculated using theElement-Tracking-Approach discus-

Table VCase 1 (Max Recycle)--Yields on Fresh Feed Basis

(Constant Coke Yield Relative to Max Gasoline Case)

Max. GasolineBase

Base NoRecycle 650-750˚F 650-800˚F 650-850˚F 650˚F+

Conversion, wt.% 70.0 55.0 61.2 64.2 65.2 64.7Recycle Ratio 0 0 0.10 0.14 0.16 0.15Maximum recycle available 0.10 0.14 0.18 0.24Cat-to-Oil Ratio 6.05 3.43 3.48 3.56 3.59 3.60Hydrogen, wt.% 0.11 0.09 0.10 0.11 0.12 0.12

1.4 1.0 1.1 1.3 1.4 1.4C3=, wt.% 3.3 2.1 2.4 2.6 2.7 2.7Total C3's, wt.% 3.9 2.4 2.7 2.9 3.1 3.1Total C4='s, wt.% 5.1 3.9 4.5 4.5 4.7 4.8Total C4's, wt.% 8.5 5.6 6.6 6.6 6.9 7.0C5+ Gasoline, wt.% 49.4 40.5 44.6 46.8 47.0 46.4RON 89.6 89.2 89.4 89.5 89.5 89.7MON 78.6 77.3 77.7 77.8 77.7 77.9LCO, wt.% 20.5 24.7 28.9 30.2 29.9 29.3Bottoms, wt.% 9.5 20.2 9.9 5.6 5.0 6.0Coke, wt.% 6.7 5.6 6.1 6.5 6.7 6.7Relative Combined Feed Rate Const Coke 1.00 1.20 1.23 1.21 1.18 1.18Relative Fresh Feed Rate 1.00 1.20 1.10 1.04 1.00 1.00Relative Coke Production Rate 6.7 6.7 6.7 6.7 6.7 6.7

sed in Section 2.3 above. The differ-ence in the yields of gasoline, LCO,and coke between the second-passcracking of the recycle stream andcracking of the fresh feed is shown inFigure 8. The maximum recycle ratioat each conversion, calculated basedon simulated distillation, is also plottedin Figure 8. At lower conversion, there

is more 650-750˚F fraction availablefor recycle. The low-conversionrecycle stream made much higherLCO than the fresh feed, while mak-ing about the same gasoline andcoke. However, at higher conver-sion there is less 650-750˚F streamavailable. The high-conversionrecycle stream made much lower

Figure 8Yields of Second Pass Cracking Minus Fresh Feed Cracking vs. Conversion

LCO, wt.%

0

-10

-20

-30

-40

15

10

5

Conversion

Max Recycle Ratio

Coke, wt.%

C5+ Gasoline, wt.%

55 60 65 70 75 55 60 65 70 75

10

8

6

4

16

12

8

4

0 0

20


gasoline, similar LCO and muchhigher coke. These results can beexplained by examining the proper-ties of the recycle steams in Figure3. Generally speaking, highercracking severity in FCC units leadsto more gasoline, but a much higherconcentration of condensed aro-matics in the bottoms.[8] Althoughthe 50 vol.% boiling points areabout the same for each stream, theAPI gravity and hydrogen contentdecrease with increasing conver-sion. This is consistent with themass spectrometry data in Table IV,which shows the tri-aromatics of the

650-750˚F stream, obtained at 68%conversion, is much higher than that at54% conversion.

Figure 9 shows the yields of gasoline,LCO, bottoms, and coke as a functionof conversion for cracking of only thebase feed (first cracking). The samefigure also shows the correspondingyields, normalized to the fresh feedbasis, for cracking of the combinedfeed (base feed + maximum recycle ofthe 650-750˚F stream at each conver-sion level). At a given conversion, byrecycling the 650-750˚F fraction, onecan lower bottoms and increase LCO

without sacrificing gasoline and withonly a minor penalty in coke. This isachieved for two reasons. First,recycling lowers bottoms becausepart of the bottoms is crackedtwice. Secondly, the second-passcracking of the 650-750˚F stream isvery selective toward producingLCO, and while coke increases, sodoes conversion; consequently,coke selectivity is not much higherthan that of the first-pass crackingof the base feed.

The results from Section 3.2, Case 1(Table 5) of 650-800˚F and 650-850˚F recycle are also plotted onFigure 8. Compared to recyclingwith the 650-750˚F stream, the LCOgain from recycling these heavierstreams is higher; however, the cokepenalty is greater. These resultssuggest that one can achieve thedesirable yield shift by loweringconversion on the combined feedbasis and selecting the proper recy-cle stream.

4.0 MIDAS® CatalystTechnology

Application of the correct catalysttechnology is critical to ensure highLCO yield and minimal bottoms andcoke yield. A balanced approach is

Figure 9Yields vs Conversion of Base Feed and Combined Feed Normalized to Fresh Feed Basis

LCO, wt.%

52

48

44

40

9

7

6

Conversion

Bottoms, wt.%

Coke, wt.%

C5+ Gasoline, wt.%

55 60 65 70 75 55 60 65 70 75

20

15

10

5

30

25

20

8

No Recycle Recycle 650-800/850 ˚F Recycle all 650-750 ˚F

Figure 10Bottoms Cracking Fundamentals

R R

Catalytic

Coke MIDAS® Catalyst is themost effective catalystfor Type I and III

Type I Precracking and Feed VaporizationType II Dealkylation of alkyl aromaticsType III Conversion of naphthenaromatics

Feed

Type II Type III

Type I

Thermal/Catalytic


required to achieve maximum bot-toms upgrading to LCO and othervaluable products. Grace DavisonMIDAS® catalysts have been provento minimize bottoms production[9].As a result, a MIDAS® catalyst wasused for the experiments describedabove.

Commercialized in 2008, MIDAS®-300 catalyst series is speciallydesigned for today’s distillate-drivenmarket. MIDAS® catalyst has beenused in 52 units since its first intro-duction. Currently in three commer-cial units, MIDAS®-300 catalyst isthe latest result of Grace’s longcommitment to developing industry-leading bottoms cracking catalysts.MIDAS®-100 series of catalyst wereintroduced in 2005 and in 2007 theMIDAS®-200 series was commer-cialized.

MIDAS®-300 catalysts offer higheractivity matrix surface area, bal-anced with optimized zeolite con-tent. The majority of matrix porosityin MIDAS® technology is found in thecrucial 100-600 Å pore size diame-ter range, ensuring high LCO selec-tivity.

LCO selectivity is maximized via thethree-step bottoms cracking mecha-nism originally described by Zhao[10]

as shown in Figure 10. Each of thesebottoms cracking mechanisms is par-ticularly critical in lower conversionoperations to ensure high LCO yieldand low bottoms.

It is critical that feed vaporization bemaintained when the FCC operates atreduced reactor temperature to maxi-mize LCO. At low operating reactortemperature, optimization of Type Icracking becomes more critical due tothe reduction in the riser mix zonetemperature. Catalyst design plays acritical role in maintaining the rightconditions. Since resid feeds containa high percentage of molecules boil-ing above the mix zone temperature,pre-cracking is necessary to achievecomplete vaporization. Porosity in the100-600Å range is essential for thepre-cracking reactions that facilitatevaporization. MIDAS®-300 catalystshave the highest porosity in this criticalrange of any cracking catalyst.MIDAS®-300 ensures that feed is prop-erly vaporized even at low severity.

Most of the LPG and gasoline pro-duced in an FCC comes from dealky-

lation of aromatics or Type II crack-ing. Zeolite is much more effectivethan matrix in cracking long chainalkyl aromatics. Type II cracking isimportant to reduce the molecularsize and promote eventual conver-sion of bottoms; however, we mustprevent any LCO that is producedfrom being over-converted to lightercomponents. The zeolite level inMIDAS®-300 has been optimized toprovide sufficient dealkylation activ-ity, yet maintain the product yield asLCO rather than LPG and gasoline.

Finally, Type III cracking destroysnaphthene rings in naphthenoaro-matic compounds. The size of typi-cal naphthenoaromatic molecules istoo large to easily fit into the zeolite.The cracking of these moleculeswill occur on the matrix sites or onthe external surface of the zeolite.The selective cracking of this typeof molecule requires the properdesign of matrix activity and theinteraction of matrix and zeolite.The high mesoporosity of MIDAS®-300 catalysts improves LCO selec-tivity by converting coke precursorsinto valuable liquid product.

Table VIYields on Fresh Basis (Constant Bottoms Yield Relative to Max Gasoline Case)

Max GasolineBase

Base NoRecycle 650-750˚F 650-800˚F 650-850˚F 650˚F+

Conversion 70.0 55.0 61.2 62.0 62.4 62.5Recycle Ratio 0.00 0.00 0.101 0.112 0.118 0.120Maximun recycle available 0.10 0.14 0.18 0.24Cat-to-Oil Ratio 6.0 3.4 3.48 3.53 3.56 3.56Hydrogen, wt.% 0.1 0.1 0.10 0.10 0.11 0.11Total C1's & C2's, wt.% 1.4 1.0 1.1 1.2 1.3 1.3Propylene, wt.% 3.3 2.1 2.4 2.5 2.5 2.6Total C3's, wt.% 3.9 2.4 2.7 2.8 2.9 3.0Total C4='s, wt.% 5.1 3.9 4.5 4.4 4.5 4.6Total C4's, wt.% 8.5 5.6 6.6 6.4 6.6 6.8C5+ Gasoline, wt.% 49.4 40.5 44.6 45.1 45.1 44.9RON 89.6 89.2 89.4 89.5 89.5 89.6MON 78.6 77.3 77.7 77.7 77.7 77.8LCO, wt.% 20.5 24.7 28.9 28.6 28.2 28.0Bottoms, wt.% 9.5 20.2 9.9 9.5 9.4 9.5Coke, wt.% 6.7 5.6 6.1 6.2 6.4 6.4Relative Combined Feed Rate Const Coke 1.00 1.20 1.23 1.21 1.19 1.19Relative Fresh Feed Rate 1.00 1.20 1.10 1.08 1.05 1.04Relative Coke Production Rate 6.7 6.7 6.7 6.7 6.7 6.7


CaseOperating ModeCatalyst/AdditiveRecycle, % FF, vol.%Ecat ActivityFeedstock Properties˚APIConradson Carbon, wt.%1050˚F+, vol.%Operating ConditionsReactor Temperature, ˚FFeed Temperature, ˚FRegenerator Temperature, ˚FCat/OilAir BlowerWet Gas Compressor Rate

Product Yields/PropertiesDry Gas, scfbLPG, vol.% FFGasoline, vol.% FFRON/MONLCO, vol.% FFBottoms, vol.% FFCoke, wt.% FF

Base 1 2Max Gasoline Max LCO Fully Optimized Max

MIDAS®-1000 11 1168 64 64

21.6

20

21.63.0

21.63.02020

3.0

995

13507.2

950405

12907.8

950405

12907.8

33123.956.7

92.6/80.622.96.85.2

23519.351.9

90.0/79.532.06.05.2

23530.044.0

92.9/80.733.45.05.2

MIDAS®-100 MIDAS®-300 & OlefinsUltra®

400

Base Base BaseBase Base0.75 Base

C3+, vol.% 110.3 109.2 112.4

Incremental ProductValue, $/B

Base +0.10 +1.40

Grace Davison can also deliverenhanced LCO selectivity in an addi-tive form. BX™-450 is Grace’snewest catalytic additive offering andis the first of its kind designed specif-ically for maximum distillate yield.BX™-450 is based on MIDAS®-300technology and offers high activitymatrix surface area balanced with anoptimized zeolite level to maximizeLCO selectivity. The optimized zeolitelevel in BX™-450 provides sufficientcatalytic activity, enabling 1:1replacement of fresh catalyst.

5.0 Commercial MaximumOperation Using Recycleand MIDAS® Technology

A commercial operation was mod-eled using results from the workdescribed above together with ournew MIDAS®-300 technology. TableVII shows a maximum gaso-line/conversion FCC operation witha residual feedstock. Two cases

are modeled off this base operation.Case 1 is a reduced conversion oper-ation together with a recycle of 11vol.% HCO (650-800ºF). Case 2 rep-resents a fully optimized maximumLCO operation. Cases 1 and 2 weremodeled assuming the base maxi-mum gasoline/conversion operationwas operating at an air blower and wetgas compressor constraint.

January 2009 FCC product valueswere used to assess the relative prod-uct values for Cases 1 and 2. LCO isvalued at 8.00 $/bbl greater thangasoline. Those product values areshown in Table VIII.

LCO is increased in Case 1 by reduc-ing conversion via lower reactor tem-perature, lower equilibrium catalyst(Ecat) activity, higher feed tempera-ture and the incorporation of 11 vol.%recycle. Coke yield is the same as thebase operation. Lower reactor tem-perature and higher feedstock tem-

perature reduce the unit cokedemand, allowing the incorporationof recycle at the same air blowerdemand as the base operation. Asdescribed above, coke yield or airblower rate will be the primary con-straint reached. The FCC catalyst inthe Base and Case 1 operation isMIDAS®-100 catalyst.

Case 1 has a higher cat to oil ratiodue to the introduction of recycleand lower catalyst activity andresults in increased bottoms con-version to LCO. Wet gas rate islower in Case 1 as a result of lowerLPG and dry gas.

Despite a 9 vol.% increase in LCOyield, the Case 1 product value isonly a modest 0.10 $/bbl greaterthan the base operation. Thereduced C3+ vol.% and gasolineoctane compared to the base oper-ation hinder the total product value.

Table VIICommercial Maximum Gasoline and Maximum Conversion Modeling Results


Case 2 represents a fully optimizedoperation using MIDAS®-300 cata-lyst and OlefinsUltra® ZSM-5 addi-tive. LCO is increased at theexpense of bottoms due toimproved bottoms cracking withMIDAS®-300 catalyst. OlefinsUltra®

additive is used to increase thegasoline octane, which droppedsharply from the base operation dueto lower reactor temperature. Wetgas rate increases compared to thebase operation due to additionalLPG production. Air blower and wetgas compressor rates in Case 2 areidentical to the base operation.

The optimized Case 2 product valuehas increased 1.40 $/bbl relative tothe base operation. These productvalues are achieved despite a lowvalue of propylene and confirm theneed to maintain liquid yield andgasoline octane while operating in amaximum LCO mode.

6.0 Conclusions

Maximizing LCO yield is largely abottoms management process.Operating at reduced conversionsby shifting operating conditions andreducing Ecat activity increasesLCO, but also increases bottoms.To fully maximize LCO from the FCC,the refinery should consider severalstrategies in parallel.

Recycle can be employed to fully max-imize LCO at reduced conversion,while maintaining bottoms equal to atraditional maximum gasoline opera-tion. The crackability and LCO yieldproduced by a particular recyclestream are consistent with its APIgravity and hydrogen content. Due tothe reduced first-pass cracking con-version, the potential of recyclestreams to produce LCO with minimalcoke production are improved overhigh conversion operations.

The 650-750˚F stream, when recycled,produces the most LCO and gasolineand the lowest coke for a given con-version. However, it is not produced atsufficient quantities to fully maximizeLCO. High Conradson Carbon levelsconsistent with higher tetra-aromaticand heavier compounds limit the yieldof LCO when 650+˚F or 750+˚Fstreams are recycled. The 650- 800˚For 650-850˚F recycle stream producesthe highest LCO when recycledagainst a coke burn and bottoms con-straint.

MIDAS®-300 is the latest developmentin Grace’s successful MIDAS® seriesof catalysts. MIDAS®-300 catalystimproves bottoms conversion via highactivity matrix surface area balancedwith an optimized zeolite level.OlefinsUltra® and OlefinsMax® ZSM-5additives should be considered topreserve C3+ liquid yield and gasolineoctane.

References

1. R. E. Ritter, J. E. Creighton, “ProducingLight Cycle Oil in the Cat Cracker”,Catalagram® 69 (1984) 5.

2. Energy Information Administration (EIA)short-term outlook January 2009.

3. D. Wallenstein, R.H. Harding, J.R.D.Nee, L.T. Boock, “Recent Advances in theDeactivation of FCC Catalysts by CyclicPropylene Steaming (CPS) in the Presenceand Absence of Metals,” Appl.Catal. A:General 204 (2000) 89.

4. G. W. Young, G.D. Weatherbee, “FCCUStudies with an Adiabatic Circulating PilotUnit,” AIChE Annual Meeting, November,1989.

5. J.C. Kayser, Versatile Fluidized BedReactor, U.S. Patent 6,069,012.

6. A. Ye, W. Wang, “Cracking performanceimprovement of FCC feedstock by addingrecycle stock or slurry”, Lianyou Jishu YuGongcheng (2004), 34(6), 5-6.

7. M. L. Fernandez, A. Lacalle, J. Bilbao, J.M. Arandes, G. de la Puente, U. Sedran,“Recycling Hydrocarbon Cuts into FCCUnits”, Energy & Fuels 16 (2002) 615.

8. R. Venugopal, V. Selvavathy, M.Lavanya, K. Balu, “Additional Feedstock forFluid Catalytic Cracking Unit”, PetroleumScience and Technology 26 (2008) 436.

9. Schiller, R. et al “The Genesis™CatalystSystem” Catalagram® 102, Fall 2007.

10. Zhao, X., et al, “FCC Bottoms CrackingMechanisms and Implications for CatalystDesign for Resid Applications” NPRA AM-02-53.

11. A. Corma, L. Sauvanaud, “How can weincrease the LCO yield and quality in theFCC: cracking pathways analysis”, Preprints- American Chemical Society, Division ofPetroleum Chemistry (2006), 51(2), 447-451.

Table VIIIYields on Fresh Feed Basis (Constant Bottoms Yield

Relative to Max Gasoline Case)

34.045.447.955.929.00.49

C3=, $/bblC4=, $/bbl

Gasoline, $/bblLCO, $/bblBottoms, $/bblGasoline Octane, Base 86.9 (R+M)/2


Hydroprocessing Catalysts fromThe Chevron and Grace Joint Venture

Advanced Refining TechnologiesColumbia, MD USA +1.410.531.8282Houston, TX USA +1.281.449.9949Singapore +65.6831.4132Richmond, CA, USA +1.510.242.1312Worms, Germany +49.6241.4030Toda, Japan +81.48.431.1952

ART is the only catalyst companyexperienced at minimizing sulfurand heavy metals in the full boilingrange of products.

More refiners use ART’s highperformance catalyst systemswhile processing difficult feedin hydroprocessing units thanany other catalysts.

Upgrading the Bottom of the Barrel

Why Make ART Part ofYour HPC Process?

s the distillate market demandhas increased over the last fewyears, the production of ultra

low sulfur diesel (ULSD) has prompt-ed refiners to look for ways to maxi-mize their diesel pool. One way toaccomplish this has been to increasethe use of opportunity feedstockssuch as additional LCO, dieselstreams from other hydroprocessingunits, and feeds from various synthet-ic crudes. Some of these opportunityfeedstocks, having already beenprocessed through conventional refin-ery processes, may pose unexpectedchallenges to refiners wishing to incor-porate them into the distillate pool.Some of these streams have proven tobe significantly more difficult toprocess, underscoring the fact that itis important to understand the poten-

Brian WatkinsHydrotreating Technical ServicesEngineer

Charles OlsenWorldwide Technical ServicesManager

Advanced Refining TechnologiesChicago, IL USA

A tial impact of processing new feedstreams in order to avoid unpleas-ant surprises. This paper highlightsa few examples demonstrating sig-nificant differences in feed reactivityfor a variety of different feed com-ponents which are not necessarilyanticipated from the usual bulk feedanalyses.

FCC LCO and coker diesels havelong been used as feed compo-nents combined with a straight run(SR) feed source to produce ULSDproducts. The quality of the LCOvaries with distillation range, anddepends on the severity of the pre-treatment of the FCC feed as well ason the conditions in the FCC andthe FCC catalyst employed. A com-mon element in LCO is a very high

Distillate Pool Maximization byExploiting the Use of OpportunityFeedstocks Such as LCO and Syncrude


concentration of polynuclear aro-matic compounds relative to otherfeeds.

Synthetic diesel material is often ini-tially processed by either a coker orebullating bed residue hydrocrack-ing unit, and then processedthrough a hydrotreater orhydrotreater/hydrocracker combina-tion. These hydrocracking unitstend to operate at severe conditionsin conjunction with high hydrogenpartial pressures. At these condi-tions, the removal of all the easy,less refractory sulfur is readilyachieved, and the majority of themulti-ring aromatics are saturated.This leaves a product which is rela-tively low in sulfur and PNA’s and,when added to the feed to a ULSDunit, gives rise to a surprisingly diffi-cult feedstock to process.

Likewise, the use of diesel rangeproducts from an H-Oil®, LC-FIN-INGTM unit or fixed bed resid desul-furizer can also have a significant

impact on downstream diesel catalystactivity for similar reasons. The gener-al properties of these types of dieselfeeds often indicate that they may berelatively easy to hydrotreat due totheir low sulfur content and API gravi-ty which is often similar to SR materi-als. Table I lists the properties for sev-eral diesel feeds including the dieselproduct fractions from an ebullatingbed resid (EB) unit, a fixed bed resid(FB) unit, and a diesel fraction from aCanadian synthetic crude.

ART conducted pilot plant testing toinvestigate the impact of variousdiesel feed components on catalystactivity. The pilot work utilized the SRdiesel shown in Table I as the basecase feed. The other componentsshown in Table I were blended into thebase feed at 20% by volume to showthe effects on catalyst performance.The pilot plant work involved severaltailored catalyst systems as well aschanges to operating pressure andhydrogen rate in order to cover abroad range of operation.

The base case testing was done ata hydrogen pressure of about 700psia, a LHSV of 0.7 and 1300scf/bbl hydrogen/oil ratio. The cata-lyst system was a stacked system ofhigh activity CoMo and NiMo cata-lysts containing >80% CoMo cata-lyst. This system was chosen dueto limited hydrogen availability anda desire to minimize hydrogen con-sumption Additional information onthe theory, design and use of thistype of staged catalyst loading canbe found in references 1-5.

Table II shows the analysis of thedifferent feed blends. The 20% LCOhas 1600 ppm lower sulfur, a onenumber lower API, and 20 ppmhigher nitrogen content comparedto the SR feed. The total aromaticcontent in the blend is also higherby 10 volume percent absolute.

Compare this to the feed blendscontaining the EB diesel, FB diesel,or the synthetic diesel where allthree have even lower sulfur com-

Table IDiesel Feedstock Analysis

Light SRGas Oil LCO EB Diesel

SyntheticDiesel FB Diesel

Sulfur, wt.% 1.11 0.17 0.017 0.07 0.006

Nitrogen, wppm 138 203 135 261 71

API Gravity 32.37 24.09 31.24 31.96 31.42

Aromatics, vol.%

Total 24.18 64.61 41.26 36.58 44.83

Mono 14.47 31.97 36.55 32.52 41.95

Poly 9.71 32.64 4.71 4.06 2.88

Distillation, D2887, ˚F

0.5 286 219 303 206 285

10 492 367 407 349 379

50 601 470 589 517 515

90 731 578 707 662 652

99.5 799 644 759 738 775

Thiophenes 8 0 0 0 0

Benzothiophenes (BT) 2 69 0 0 0

Substituted BT’s 2793 1083 0 26 0

Di Benzothiophene (DBT) 222 110 0 7 0

Substituted DBT’s 3453 436 72 266 154,6 DiMethylDiBenzo

199 0 78 29 43Thiophene

C3-DBT 4410 0 17 370 24


pared to the SR, and slightly higherAPI gravities, despite having a high-er total aromatic content. Otherchanges to note are the fact thefeed nitrogen content stays fairlyconstant, and the mono aromaticcontent is higher and PNA contentlower for these blends compared tothe SR feed.

Figure 1 summarizes some of thepilot plant data comparing the SRand LCO feed blends. It shows thatthe SR diesel requires a 43°Fincrease in temperature to go fromabout 100 ppm sulfur down to 10ppm sulfur at base LHSV and pres-sure. The LCO blend requiresalmost 20°F higher temperature toachieve the same product sulfur rel-ative to the SR feed. The productfrom the LCO blend has a two tothree number lower API comparedto the SR product, and hydrogenconsumption increases significantlyfor the LCO blend due to saturationof additional polyaromatic com-pounds found in the LCO. These lat-

Table IIBlended Diesel Feedstock Analysis

Light SRGas Oil 20% LCO

20%20% Syncrude

20%EB Diesel FB Diesel

Sulfur, wt.% 1.11 0.95 0.88 0.92 0.92

Nitrogen, wppm 138 158 144 179 131

API Gravity 32.37 31.75 33.17 33.24 33.18

Aromatics, vol.%

Total 24.18 34.58 30.1 29.09 30.12

Mono 14.47 19.7 21.2 20.33 21.94

Poly 9.71 14.9 8.8 8.76 8.18

Distillation, D2887, ˚F

0.5 286 257 288 265 302

10 492 428 470 442 460

50 601 574 598 586 592

90 731 715 722 718 723

99.5 799 795 794 796 801

Thiophenes 8 12 8 10 8

Benzothiophenes (BT) 2 13 1 1 0

Substituted BT’s 2793 2279 2161 2105 1861

Di Benzothiophene (DBT) 222 206 35 148 133

Substituted DBT’s 3453 2497 2324 2809 2464

4,6 DM-DBT 199 119 116 148 130

C3-DBT 4410 3598 3508 4061 3583

Figure 1Activity Comparison on SR and Blended SR/LCO

0

10

20

30

40

50

60

70

0 20 40 60 80 100 120

LCO

SR

Req

uir

edTe

mp

erat

ure

,˚F

Product Sulfur, ppm


ter consequences set limits on theamount of LCO which can beprocessed and still meet productcetane specifications and hydrogenavailability constraints.

Results indicate that both feedshave similar API number upgrades(i.e. product – feed) as the reactorincreases in temperature; however,the actual product API is different atequal product sulfur. Even thoughthe LCO blended feedstock requiresa higher temperature to achieve thesame product sulfur, the productAPI is still about a full number loweras shown in Figure 2.

The LCO also has the additionalissue of increasing hydrogen con-sumption when added to the ULSDoperation. Figure 3 compares thearomatic saturation achieved on theblended LCO feedstock as com-pared to the SR. The majority of thearomatic saturation occurs with thepoly aromatic compounds and, asshown in Table II, the LCO blendcontains significantly more PNA’scompared to the SR feed. Thehydrogen consumption is estimatedto be 125-150 scf/bbl higher for theLCO blend at 10 ppm product sulfur.The figure above shows that pro-cessing LCO is significantly moredifficult than processing the SRfeed. One option to gain back someof the lost activity is to change the

end point of the LCO in the feed. ARTcompleted pilot plant testing on anLCO stock as received and the sameLCO with a 40°F end point reduction tosimulate how this can affect catalystperformance. Table III lists the majorcomponent analysis of the two LCOfeeds. The decrease in endpoint low-ers the total sulfur by almost 1000ppm and total nitrogen decreases by129 ppm.

The impact this degree of LCO end-point reduction has on ULSD perform-ance is over 30°F difference in activitywhich corresponds to additional life in

the hydrotreater. A comparison ofthe two LCO feeds blended at 30%into SR feed is shown in Figure 4.

The addition of LCO has a majorimpact on activity for both the lowand high endpoint LCO materials.The required temperature increasefor ULSD in going from 0 to 30%LCO for the lower endpoint materialis about 1.2°F per percent LCO.Processing the higher endpointLCO increases the required temper-ature to about 1.4°F per percentLCO. Figure 5 demonstrates thismore clearly in the form of a plot ofthe required temperature increaseas a function of LCO content.Notice from the chart that the activ-ity effects are not exactly linear withincreasing LCO content. The first15% LCO has a larger impact onactivity than the next 15%.

The diesel products from an EB unit,a FB unit and the synthetic crudediesel provide very different sulfurdistribution patterns compared tothe SR feed and LCO shown inTable I. Almost all of the sulfurspecies in those feeds are multi-substituted dibenzothiophenes, theso-called hard sulfur species. Thespecies groupings from sulfur spe-ciation using a GC-AED technique,however, indicate little about whatthe actual molecular structure is

Figure 2Comparison of Product API for the SR and LCO Blend

34.0

34.5

35.0

35.5

36.0

36.5

37.0

37.5

38.0

38.5

0 20 40 60 80 100 120

Increasing WABT, ˚F

Pro

du

ct˚A

PI

LCO

SR

Figure 3Comparison of Aromatic Saturation


Ch

ang

ein

Tota

lAro

mat

ics,

vol.%

-9

-8

-7

-6

-5

-4

-3

-2

-1

0

10 20 40 60 80 100 120

LCO

SR


since the basic technique sepa-rates out the sulfur based on boilingpoint distribution. The sulfur mole-cules left in these previously treatedfeeds have already been processedonce in a high temperature, highpressure hydrotreating application.Those conditions easily remove themajority of sulfur molecules andleave only those sulfur species thatare multi-ring, sterically hinderedmolecules and other aromatic nitro-gen compounds. It is these speciesthat require a greater level of satu-ration or ring opening before thenitrogen or sulfur can be removed.It is likely for there to be very lowconcentrations of multiple-ring, par-tially saturated compounds thatneed to be more fully saturated inorder to remove the sulfur. This isenough to make it more difficult toproduce 10 ppm sulfur productfrom such feeds.

An understanding of the upstreamprocessing is important when con-sidering the use of syntheticcrudes. Production of syntheticfuels involves a combination of sev-eral processes in order to accom-modate downstream processing.These upstream processes includecoking or an ebullating bed residoperation, followed by a hydrotreat-ing or hydrocracking operation inorder to produce a lighter gradematerial. These products are thenblended in with other heavier mate-rials as a diluting or cutting stockand sent downstream as syntheticcrude. The synthetic diesel used inthis work is taken from a productdiesel cut from a synthetic VGOhydrocracker. Figure 6 shows theactivity difference between the SRand the blended SR/syntheticdiesel. Note that at higher productsulfur, the two feedstocks respondfairly similarly to each other. As theapplication becomes more de-manding, the required reactor tem-perature increases dramatically forthe synthetic diesel feed as com-pared to the SR feed. The blendedfeed requires more than 25°F highertemperature relative to the SR toachieve ULSD sulfur levels.

Table IIIComparison of Boiling Point Reduction on LCO

Type

˚APISulfur, wt.%Nitrogen, ppmAromatics, lv.%

Mono-, lv.%Poly-, lv.%

Dist., D2887, ˚FIBP10%50%70%90%FBP

LCO(Low FBP)

18.310.948708

66.8622.6544.21

249425531600677772

LCO(High FBP)

15.311.041837

68.8118.4450.37

256432550620699812

Figure 4Impact of Endpoint Reductionon Hydrotreating Performance

Product Sulfur, ppm

Req

uir

edTe

mp

erat

ure

Incr

ease

,˚F

0

20

40

60

80

100

120

140

160

0 100 200 300 400 500 600

30% Hi FBP LCO

SR

30% Lo FBP LCO

Figure 5Activity Comparisons at Different LCO

FBP and Concentration

% LCO

WA

BT

Incr

ease

toA

chie

veP

rod

uct

Su

lfu

r

0

5

10

15

20

25

30

35

40

45

50

0 5 10 15 20 25 30 35

Hi EP LCO

Lo EP LCO


It is reasonable to expect that theupstream hydroprocessing of thesynthetic diesel material results in afeed which behaves similarly toother previously hydrotreated feed-stocks like those from the EB andFB resid applications. Two feed-stocks from these sources areshown in Figure 7. These two feed-stocks have a remarkably similarresponse as that observed for thesynthetic diesel feedstock. Thefixed bed diesel fraction, which hassignificantly lower sulfur and nitro-gen than the other two feedstocks,shows over 40°F higher SOR thaneither the EB or synthetic diesels at10 ppm sulfur. These data showthat upstream processing prior totreating in a ULSD unit can have adramatic effect on the activity of theunit and consequently decreasecycle length.

Figure 8 examines how the productAPI is changed during processingfor the synthetic diesel blend. Ascan be seen, there is only a onenumber increase in product APIover an almost 100°F WABT changecompared to >two number increasefor the SR feed over a similar tem-perature span.

Aromatic saturation in the ULSD unitis also a concern in order to meetrequired cetane and aromatic tar-gets. The higher temperaturerequired to process these previous-ly processed streams may make itdifficult to achieve much aromaticssaturation because of the approachto the thermodynamic equilibriumlimit for aromatic saturation. Figure9 compares the aromatic saturationachieved for the SR diesel and thesynthetic diesel blend. The synthet-ic diesel has a low level of poly aro-matic compounds, and the blendactually has a slightly lower concen-tration of PNA’s compared to the SRfeed. Less saturation is achievedon the synthetic blend, probably areflection of the fact that mono aro-matic molecules are the predomi-nant species, and these are quitedifficult to saturate. The equilibriumlimit on conversion is readily appar-ent in the figure. The syntheticdiesel provides a two number

Figure 6Activity Comparison of the SR and Synthetic Diesel Blend

Product Sulfur, ppm

Req

uir

edTe

mp

erat

ure

,˚F

0

10

20

30

40

50

60

70

80

90

0 20 40 60 80 100 120

SR

Synthetic

Figure 7Activity Comparison of Previously Hydrotreated Streams

Product Sulfur, ppm

Req

uir

edTe

mp

erat

ure

Incr

ease

,˚F

0

20

40

60

80

100

120

0 20 40 60 80 100 120

FB

Synthetic

EB

Figure 8Comparison in Product API


Pro

du

ctA

PI

35.5

36.0

36.5

37.0

37.5

38.0

38.5

0 20 40 60 80 100 120 140 160

Syn Diesel

SR


decrease in total aromatics whilethe SR diesel has an almost 10 num-ber decrease. This can be prob-lematic if trying to meet an aromatictarget in the diesel pool.

When evaluating opportunities tohydrotreat previously treatedstreams for ULSD, the need toexamine the catalytic effects ofthese feeds is important. ART com-pleted a series of pilot tests usingthe synthetic diesel feed blend cov-ering a wide range in operating con-ditions. Figure 10 summarizessome of the results for one of ART’shigh activity CoMo catalysts. Thebase case condition was again 700psia, 1300 scf/bbl hydrogen/oil ratioand 0.7 LHSV. As you can see fromthe figure, the base case conditionsresult in the highest start of run tem-perature for ULSD. Increasing theH2/Oil ratio to 2600 scfb results in adecrease in the SOR WABT of over15°F. The third data set showsresults for higher hydrogen partialpressure, 1300 psi, and 2600 scfbhydrogen to oil ratio. This results ina gain of 25°F lower SOR comparedto the base case conditions or anincremental 10°F lower temperaturedue to the increase in hydrogenpressure. Finally, increasing thehydrogen rate to 4000 scf/bbl athigh pressure provides over 30°Flower SOR as compared to the basecase system, but only an additional5°F relative to the high pressurelower gas rate case. These dataclearly demonstrate the significantbenefits of increasing hydrogenpartial pressure when treating thesetypes of difficult feeds.

Making the switch to using a NiMocatalyst in this application has muchmore significant effect on unit per-formance. In Figure 11 the basecase conditions of low pressureand low hydrogen/oil ratio actuallyresult in activity which is similar towhat was observed at these condi-tions for the CoMo catalyst. Afterincreasing the hydrogen rate at lowpressure, the all NiMo system gainsover 25°F relative to the base condi-tions, a larger activity gain thanobserved for the CoMo catalyst.

Figure 9Change in Total Aromatics on SR and Synthetic Diesel

Tota

lAro

mat

ics,

vol.%

15.0

17.0

19.0

21.0

23.0

25.0

27.0

29.0

31.0

33.0

0 20 40 60 80 100 120 140 160

Increasing WABT, °F

Syn Diesel

SR

Figure 10CoMo Catalyst Activity on Synthetic Diesel

Req

uir

edTe

mp

erat

ure

Incr

ease

,˚F

Product Sulfur, ppm

0

20

40

60

80

100

120

0 20 40 60 80 100 120

700psia & 1300 H2/Oil

700psia & 2600 H2/Oil

1300psia & 2600 H2/Oil

1300psia & 4000 H2/Oil

Figure 11NiMo Catalyst Activity on Synthetic Diesel

Req

uir

edTe

mp

erat

ure

Incr

ease

,˚F

Product Sulfur, ppm

0

20

40

60

80

100

120

0 20 40 60 80 100 120

700psia & 1300 H2/Oil

700psia & 2600 H2/Oil

1300psia & 2600 H2/Oil

1300psia & 4000 H2/Oil


The activity benefit increases toover 40°F lower required tempera-ture at 1300 psia, and increasingthe hydrogen rate to 4000 scf/bblresults in over 50°F lower tempera-ture compared to the base condi-tions. The NiMo catalyst is at least20°F more active than the CoMocatalyst at this last set of conditions.

Clearly there is a need to maximizesaturation when treating these pre-processed types of feeds. Anyincrease in hydrogen partial pres-sure helps this and, as the data justdiscussed indicates, the catalystselection has a significant impact.High activity NiMo catalysts are bet-ter saturation catalysts compared tohigh activity CoMo catalysts, andthis appears critical to removing the

harder sulfur species present in thesepreprocessed feeds. In units that areconstrained by limited hydrogen orlower hydrogen pressures, the use ofeven a small amount of NiMo catalystwill prove to be beneficial in order toremove the remaining difficult sulfurs.

Advanced Refining Technologies canwork closely with refining technicalstaff to help plan for processingopportunity feeds such as those dis-cussed above. One of the keys isbeing aware of the potential impactprocessing certain feeds will have onunit performance. Feeds which havebeen previously processed presentunique challenges and ART is wellpositioned with its experience at pro-viding customized catalyst systemsfor ULSD applications. Opportunity

feeds provide yet another objectiveto consider when designing theappropriate catalyst system to max-imize unit performance.

References

1. Olsen, C., Krenzke, L.D., Watkins, B.,AICHE Spring National Meeting, NewOrleans, March 2002.

2. Krenzke, D., Armstrong, M., 2001 ERTCMeeting, Madrid, Spain.

3. Olsen, C., Krenzke, L.D., 2005 NPRAAnnual Meeting, Paper AM-05-17.

4. Olsen, C., D’Angelo, G., 2006 NPRAAnnual Meeting, Paper AM-06-06.

5. Olsen, C., Watkins, B., Shiflett, W., ERTCMeeting, Barcelona, Spain, November 2007


he production of syntheticcrude from oil sands is expect-ed to increase dramatically in

the next decade. The effect of thesefeedstocks on the yields from the FCCand the selection of the appropriateFCC catalyst are the subject of thispaper. A hydrotreated synthetic crudesample was obtained from a refinerand characterized by standard meth-ods, as well as by High ResolutionMass Spectroscopy (HRMS). Thesynthetic crude was distilled to650˚F+, blended at various levels witha conventional paraffinic VGO, andtested in an ACE unit over varioustypes of FCC catalysts, in order todetermine the effect of the feed prop-erties and catalyst type on productyields. A range of catalysts with zeo-lite/matrix ratios varying between 1.3and 3.8 were tested. Analysis of thesulfur species in the feed and thecracked products is also presented.

Characterization and CatalyticCracking of Synthetic Crude

Feedstocks

Michael ZiebarthManager, Synthesis Research

Rosann SchillerProduct Manager, FCC Catalysts

Grace DavisonRefining TechnologiesColumbia, MD USA

T Introduction

Synthetic crude produced fromCanadian oil sands is a growingfeedstock source that is being uti-lized by an increasing number ofrefiners. Canada currently pro-duces about 1.2 million b/d of oilfrom oil sands and it is projected toincrease to 3 million b/d by 2015.The recent drop in crude prices hasforced delays in several develop-ment projects. However, withCanadian reserves of 170+ billionbarrels of viable oil, economic fore-casts predict that oil sands will con-tinue to be a significant crudesource for the foreseeable future.[1]

Synthetic crude from oil sands hassignificantly different characteristicscompared to traditional VGO feeds.Oil sands contain bitumen which isseparated out and then further


processed to yield synthetic crude.The unprocessed bitumen is highlyaromatic, has low hydrogen con-tent, low API (high specific gravity),and high levels of sulfur, nitrogenand metals. The final characteris-tics of the synthetic crude varydepending on the amount of pro-

cessing the bitumen has undergone.Typically, after the bitumen is separat-ed from the oil sands, it is upgraded ina fluid coker or an ebullating bedresidue upgrader. The resulting prod-uct quality can be further improved byhydrotreating to remove additionalmetals, sulfur and nitrogen.

This paper will discuss the impactthat synthetic crude has on hydro-carbon yields in the FCC unit. Wealso examine the type of FCC cata-lyst that is most effective at crackingthese feeds. Analyses of the sulfurspecies in the feed and the crack-ing products are also included.

Table IFeed Analysis

Distilled 650F+ Fraction Synthetic Crude 8

Standard VGO Synthetic Crude Hydrotreated Coker VGO˚API 25.50 19.2 17.3

K Factor 11.94 11.4

Hydrogen Content 13.10 11.8 11.5

Sulfur, wt.% 0.37 0.43 0.43

Total Nitrogen, wt.% 0.1 0.15 0.22

Basic Nitrogen, wt.% 0.1 0.0 0.0

Conradson Carbon, wt.% 0.7 0.7 0.5

Specific Gravity 0.9 0.9 1.0

Refractive Index 1.5 1.5 1.5

Paraffinic Carbons Cp, wt.% 63.6 49.3

Naphthenic Ring Carbons Cn, wt.% 17.4 24.4

Aromatic Ring Carbons Ca, wt.% 18.9 26.2 24.7

Sim Dist

% 1000+°F 14.0 9.4 8.3

Table IIHRMS Analysis of Feeds

Standard VGO Synthetic crude

SATURATES Avg., wt.% Avg., wt.%C(N)H(2N+2) Paraffins 12 0.1C(N)H(2N) Monocycloparaffins 20.7 12.6C(N)H(2N-2) Dicycloparaffins 10.9 12.2C(N)H(2N-4) Tricycloparaffins 6.3 8.7

C(N)H(2N-6) Tetracycloparaffins 0.9 1.3

TOTAL SATURATES 50.9 34.9MONOAROMATICSC(N)H(2N-6) Alkylbenzenes 7.8 11.4C(N)H(2N-8) Benzocycloparaffins 4.4 7C(N)H(2N-10) Benzodicycloparaffins 3.5 6.5

DIAROMATICSC(N)H(2N-12) Naphthalenes 4.9 6.4C(N)H(2N-14) 5.6 7.5

C(N)H(2N-16) 9.4 11.9

TRIAROMATICSC(N)H(2N-18) 6 6.4

C(N)H(2N-22) 1.8 1.6

TOTAL AROMATICS 43.4 58.7THIOPHENIC COMPOUNDSC(N)H(2N-4)S Thiophenes 0 0

C(N)H(2N-10)S Benzothiophenes 2.7 2.4C(N)H(2N-16)S Dibenzothiophenes 3 4.1C(N)H(2N-22)S Naphthobenzothiophenes 0 0

TOTAL THIOPHENIC COMPOUNDS 5.7 6.5


Properties of Synthetic Crude

A synthetic Oil Sand B grade crudesample was obtained from a refinerand distilled to remove the fractionthat boiled below 650°F. The dis-tilled feed was characterized bystandard methods as well as byHRMS. The refiner described thesample as hydrotreated syntheticcrude that contains some resid andnonhydrotreated material.

A standard VGO sample, used as abase case comparison, was ana-lyzed by the same methods. Table Ishows the properties of the twofeeds. The key difference betweenthe two is that the synthetic crude

contains high levels of aromatic andnaphthenic molecules, while the VGOis more paraffinic. This results in asignificantly lower API (higher specificgravity) and lower hydrogen contentfor the synthetic crude. The VGO isslightly heavier with 14% boiling above1000°F versus 9% for the syntheticcrude. The sulfur and ConradsonCarbon levels are similar for the twofeeds. The nitrogen is directionallyhigher for the synthetic crude.

Table II shows the results of an HRMS22-Component Hydrocarbon TypesAnalysis (carried out by PetroMass).A breakdown of the hydrocarbontypes further highlights the aromaticand naphthenic character of the syn-

thetic crude. The synthetic feed isessentially devoid of paraffinic mol-ecules and has higher monoaromat-ics, polynuclear aromatics (PNAs),and naphthenoaromatics whencompared to the VGO feed [Figure1]. All the data point to less hydro-gen and less crackable material inthe synthetic crude feedstock. Thedistribution of sulfur is also slightlydifferent. The synthetic crude has ahigher proportion of sulfur in theheavier molecules. This will affectthe distribution of the sulfur in thecracked products.

Experimental

The standard VGO and twoVGO/synthetic crude blends werecracked over four different fluid crack-ing catalysts with a wide range of zeo-lite to matrix surface area ratio (Z/M).

Feed

Two blends containing 40 wt.% and80 wt.% synthetic crude were test-ed versus the standard VGO basecase. The properties of the feedblends are shown in Table III. Thedata show that the measured blendproperties are close to what wouldbe expected if calculated from theindividual VGO and synthetic crudeproperties.

Figure 1Aromatics and Naphthenoaromatics in Feed

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

Wt.

%in

Fee

d

Std Gas Oil Synthetic Crude

Alkylbenzenes NaphthenoMono-aromatics

NaphthenoDi-aromatics

Naphthalenes

Table IIIFeed Blend Properties

Feed Name Standard VGO60% Standard VGO

40% Synthetic Crude20% Standard VGO

80% Synthetic Crude˚API 25.5 22.9 20.5Sulfur, wt.% 0.37 0.40 0.42Total Nitrogen, ppm 0.12 0.13 0.14Basic Nitrogen, ppm 0.05 0.03 0.03Conradson Carbon, wt.% 0.68 0.6 0.7

K Factor 11.94 11.73 11.52Specific Gravity 0.90 0.92 0.93Paraffinic Carbons Cp, wt.% 63.6 57.5 52.3Naphthenic Ring Carbons Cn, wt.% 17.4 21.1 22.8Aromatic Ring Carbons Ca, wt.% 18.9 21.4 24.9

Distillation, 10% 607 624 634Distillation, 50% 818 807 791Distillation, 90% 1034 1025 997


Catalyst

Four Grace Davison alumina cata-lysts were used in the study. AURO-RA® is a high zeolite alumina sol cat-alyst that is one of our most flexiblecatalyst technologies. AURORA®

catalyst can be formulated with mul-tiple activity and hydrogen transferlevels, and metals trapping function-ality.[2]

ADVANTA® is a catalyst technologyjointly developed by Grace with

ExxonMobil.[3] ADVANTA® catalyst ischaracterized by high activity with lowto moderate surface area and a mod-erate Z/M ratio.

GENESIS™ catalysts are a blend oftwo catalyst types in which one com-ponent is a MIDAS® catalyst.[4] TheMIDAS® catalyst component providesa high matrix input, maximizing con-version of bottoms and increasingLCO yield. The resultant GENESIS™blend has a moderate to low Z/M ratio.

Each catalyst technology was deac-tivated in the laboratory via cyclicpropylene steaming (CPS)[5] at1450°F without metals. After deac-tivation, the Z/M ratio ranged from3.8 for the AURORA® catalyst to 1.3for the GENESIS™-2 catalyst.Properties of the deactivated cata-lysts are shown in Table IV. Thematrix surface area (MSA) and the100-600Å pore volume correlatewith the Z/M ratio. As the Z/M ratiodecreases, the MSA and pore vol-ume both increase. Total surfacearea was essentially equal for eachcatalyst.

Apparatus and test conditions

The testing was done in an ACE unit at980°F with C/O ratio varied from four toeight. Naphtha octanes were deter-mined by a gas chromatograph (GC)compositional analysis model knownas GCON.[6]

Results and DiscussionFeed Effects

In order to compare the ACE yieldand product quality data for the dif-

Table IVCatalyst Properties

AURORA® ADVANTA® GENESIS™-1 GENESIS™-2Al2O3 wt.% 45.9 53.0 48.7 48.5

RE2O3 wt.% 3.2 2.9 2.7 2.4

Total Pore Volume cc/g 0.39 0.41 0.42 0.48

CPS No-Metals Deactivation

Unit Cell Size Å 24.33 24.35 24.32 24.33

Surface Area m2/g 192 184 183 188

Zeolite Surface Area m2/g 152 131 122 106

Matrix Surface Area m2/g 40 53 61 82

Z/M 3.8 2.5 2.0 1.3

Bottoms, wt.%

Coke, wt.%

3

4

5

6

7

2.5

3.0

3.5

0 20 40 60 80

6.0

6.5

7.0

0 20 40 60 80

AURORA® (Z/M=3.8)

ADVANTA® (Z/M=2.5)

GENESISTM-1 (Z/M=2.0)

GENESISTM-2 (Z/M=1.3)

Cat-to-Oil Ratio

% Synthetic Crude

Figure 2Constant Conversion Interpolated Yields


Table VConstant Conversion Interpolated Yields

AURORA® AURORA® ADVANTA® ADVANTA® GENESIS™-1 GENESIS™-1 GENESIS™-2 GENESIS™-2% Synthetic crude 0 80 0 80 0 80 0 80

Cat-to-Oil Ratio 3.6 5.4 4.0 6.5 4.9 7.3 4.1 6.7

Dry Gas, wt.% 1.3 1.9 1.2 1.8 1.4 1.8 1.3 1.8

Wet Gas, wt.% 16.6 18.1 16.0 17.4 16.8 17.6 15.9 17.4

C5+ Gasoline, wt.% 54.1 51.4 54.7 52.1 53.6 51.8 54.6 51.8

RON 89.2 91.7 89.1 91.3 89.7 91.5 89.7 91.5

MON 79.0 81.6 78.6 81.1 79.0 81.1 78.9 81.2

Paraffins, wt.% 4.1 3.2 4.1 3.2 3.9 3.2 3.9 3.1

Isoparaffins, wt.% 34.2 33.5 32.7 32.8 32.7 32.5 31.9 32.7

Aromatics, wt.% 28.7 38.3 28.7 37.8 29.7 37.0 28.4 37.8

Naphthenes, wt.% 11.9 10.4 12.0 11.1 11.4 10.9 12.0 10.9

Olefins, wt.% 21.1 14.6 22.5 15.1 22.2 16.4 23.8 15.6

LCO, wt.% 20.1 20.6 20.5 20.9 20.8 21.0 21.2 21.3

Bottoms, wt.% 6.9 6.4 6.5 6.1 6.2 6.0 5.8 5.7

Coke, wt.% 2.3 3.6 2.4 3.5 2.6 3.6 2.4 3.7

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 40 80

%B

enze

ne

inG

aso

line

% Synthetic Crude in the Feed

AURORA ADVANTA GENESIS-1 GENESIS-2

Figure 3Benzene Concentration in Gasoline

ferent feeds, the data were interpo-lated to constant conversion and theyields plotted as a function of thepercent synthetic crude in theblend. Figure 2 shows graphs forcat to oil, bottoms (650°F+) andcoke yield. For all catalysts, as thepercentage of synthetic crude inthe blend is increased the feed ismore difficult to crack, requiring ahigher cat/oil ratio and producinghigher coke. Both of these effectscan be attributed to the hydrogen

deficient nature of the aromatic syn-thetic crude. The bottoms yield actual-ly decreases at higher synthetic crudelevels. This is a result of the highercoke yield (conversion of polyaromat-ics in the bottoms to coke rather thanlighter products) and the lower level ofmaterial in the feedstock boiling above1000°F.

The constant conversion yields for thefour catalysts on the VGO and 80%synthetic crude/VGO blend are shown

in Table V. In addition to the highercat/oil and coke yield, the interpolat-ed yields show significant increasesin wet and dry gas as well as adecrease in gasoline with increas-ing amounts of synthetic crude inthe feed blend.

The gasoline composition alsochanges notably with the feed. Theoctane of the gasoline increaseswith more synthetic crude due tothe large increase in aromatics con-


LCO Sulfur, ppm

ADVANTA®

% Synthetic Crude

80

90

100

110

66

70

0 20 40 60 80

3500

4000

4500

0 20 40 60 80

Gasoline Sulfur (thru 428˚F), ppm Benzothiophene, ppm

Figure 4Interpolated Sulfur Yields at Constant Conversion

0 20 40 60

2.1

2.0

% Feed Sulfur to Gasoline

2.3

2.2

2.5

2.4

2.7

2.6

1.9

ADVANTA®

% Feed Sulfur to LCO

1.8

17

16

19

18

21

20

23

22

80 0 20 40 60 80

Figure 5Interpolated Sulfur Yields at Constant Conversion

tent. Along with the higher aromat-ics content also comes an increasein benzene [Figure 3], which can beundesirable.

The distribution of sulfur in the liquidproducts also changes even though

the VGO and synthetic crude havesimilar overall sulfur contents. Thegasoline range thiophenes decreasewith synthetic crude addition while theheavier aromatic sulfur speciesincrease. This results in lower gaso-line sulfur and dramatically higher

LCO sulfur, where the alkylbenzo-and dibenzothiophenes are found(Figures 4 and 5). This correlateswith the HRMS data on the feeds,where the synthetic crude had ahigher proportion of heavy sulfurmolecules. The heavier sulfur mole-cules in the feed appear to becracked to the LCO range ratherthan the gasoline.

When the catalysts are compared atconstant coke, the major effect ofadding synthetic crude is a sub-stantial loss in conversion. The lossin conversion is reflected in higherbottoms and LCO yields as well as aloss in gasoline. To demonstrate theyield shifts, Table VI shows the con-stant coke yields for ADVANTA® cat-alyst for the different feed blends.

Catalyst Effects

An analysis of the constant conver-sion data in Table V indicates thatthe use of an appropriate catalystcan significantly improve cracking


% Synthetic Crude 0 40 80

Conversion, wt.% 75.9 73.3 71.0

Catalyst to Oil Ratio 5.3 5.5 5.5

LPG, wt.% 16.4 15.4 14.7

Wet Gas, wt.% 17.9 17.0 16.3

Gasoline, wt.% 55.2 53.8 52.0

RON 89.5 89.9 90.9

MON 79.3 79.9 80.6

LCO, wt.% 18.6 20.4 21.9

Bottoms, wt.% 5.3 6.1 6.9

Table VIConstant Coke Interpolated Yields for ADVANTA® Catalysts

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

-30-24-18-12-60

Wt.

%

Z in CnH(2n+Z)

Active/Hi SA Matrix Inert/Low SA Matrix

Figure 6Naphthenoaromatic Conversion

yields for feeds containing syntheticcrude. The four catalysts in thestudy show significant differencesin bottoms yield at essentially con-stant coke. For the 80% syntheticcrude blend, the GENESIS™-2 cat-alyst shows a 0.7 wt.% decrease inbottoms yield, about 11% on a rela-tive basis, as compared to theAURORA® catalyst. The generaltrend is that the lower Z/M ofADVANTA® and GENESIS™ cata-lysts improves bottoms cracking.The lower Z/M catalysts also showdirectionally lower dry gas yield andimproved gasoline plus LCO yields.High matrix porosity is balancedagainst the acid strength of theactive sites to selectively minimizebottoms yield without the gas penal-ty often observed with other highmatrix catalysts. These are charac-teristics that are the hallmark ofDavison’s ADVANTA® and GENE-SIS™ catalyst technologies.

The key for improving product yieldswith a synthetic crude blend is thecracking and conversion of thenaphthenoaromatics. The syntheticcrude HRMS analysis indicated thatit is composed mainly of monoaro-matics, polynuclear aromatics(PNAs), and naphthenoaromatics.PNAs are difficult to crack and oftenconvert to coke under cracking con-ditions. However, the saturatednaphthenic rings attached to thearomatics can be cracked with theappropriate catalyst. The size oftypical naphthenoaromatics mole-cules is too large to easily fit into thezeolite pore structure. As a result,the cracking of these moleculesneeds to occur on matrix sites or theexternal surface of the zeolite. Theselective cracking of this type ofmolecule requires selective matrixactivity and adequate pore volumefor the diffusion of these large mol-ecules into, and the cracked prod-ucts out of, the catalyst. Literatureresults indicate that the pore sizeneeds to be in the 100-600Å rangefor the free diffusion of the mole-cules of this size.[7]

With adequate matrix activity andpore size the naphthenoaromatics

can be cracked, as is shown in Figure6. The literature example comparesthe cracked products from an activeselective matrix catalyst with productsfrom a catalyst with an inert matrix.Although both catalysts converted lit-tle of the PNAs without saturatedrings, the two catalysts showed signif-icant differences in converting thenaphthenoaromatics.[7] Low Z/M cat-alyst formulations also minimize hydro-gen transfer activity that could convertthe naphthenoaromatics to PNA’srather than to lighter products.

The importance of cracking andconverting the heavy bottoms mole-cules is shown by further analysesof the bottoms fraction of thecracked product. In Figure 7 weshow the results of distilling the bot-toms into boiling point fractions andthe conversion for each boiling pointrange. The difference in bottomscracking for the catalysts comesalmost exclusively in the highestboiling range. As the MSA and crit-


AURORA® ADVANTA® GENESISTM-1 GENESISTM-2

0

0.5

1

1.5

2

2.5

3

3.5

650-700˚F 700-750˚F 750-800˚F 800+˚FB

ott

om

s,w

t.%

Figure 7Bottoms Conversion

20 22-0.3

-0.2

-0.1

0

0.1

0.2

24 26 28 30 32 34 36 38 40

Rel

ativ

eS

lurr

yY

ield

,lv.

%

% Synthetic Crude in Feed

AURORA®

ADVANTA®

Figure 8Commercial Data from Synthetic Crude Operation

ical pore volume increase, the bot-toms yield decreases. In summary,the desirable catalyst features forcracking naphthenoaromatics andsynthetic crudes are low Z/M cata-lyst formulations that contain cokeselective matrix with high surfacearea and high pore volume in the100-600Å region.

Commercial Experience

A major refiner in North America hasbeen processing synthetic crudesfor many years. The synthetic crudehas been upgraded and is mixedinto their typical crude diet in ratiosranging from 20 to 40%. The yielddata presented in Figure 8 repre-sent monthly mass balance pointsover a period of several years.

When utilizing Grace Davison’sAURORA® alumina sol catalyst, asimilar trend in bottoms yield wasobserved as presented previouslyin Figure 2; as the amount of syn-thetic crude in the feed increases,bottoms yield decreases. Over ayear ago, this refiner switched to anADVANTA® catalyst. After turnoverto ADVANTA® catalyst, the refinerrealized a reduction in bottoms yieldon the order of 0.3 lv.% of feed,which is similar to the improvementobserved in the ACE testing forADVANTA® over AURORA® catalyst(Table V).

The processing of synthetic cruderemains an important operating strate-gy for this refiner who, based on thetest results presented here, will bereformulating to a GENESIS™ catalystsystem to further drive bottoms reduc-tion and unit profitability.

Conclusions

Lab data and extensive commercialexperience show synthetic crude fol-lows the same general cracking rulesas more typical VGO feeds. Syntheticcrude typically contains high levels ofaromatics and naphthenoaromatics.

Porosity in the 100-600Å range iscritical for the free diffusion of theselarge molecules into and out of thecatalyst pores. The associated lowfeedstock hydrogen level makes thefeed more difficult to crack andtends to increase coke levels. Theuse of an appropriately designedcatalyst can mitigate these negativeeffects. The proper design ofmatrix activity and the interaction ofmatrix with zeolite ensure conver-sion of aromatic coke precursorsinto valuable liquid products.Grace Davison ADVANTA® andGENESIS™ catalyst technologieswith low Z/M ratios demonstratereduced bottoms, lower dry gas,and higher gasoline + LCO yields atconstant conversion.

The selective cracking of naph-thenoaromatics requires GraceDavison’s matrix technology inADVANTA® and GENESIS™ cata-lysts. Both systems possess cokeselective matrix with high surfacearea and high pore volume in the100-600Å region. Their high matrixporosity translates into excellentcoke selectivity in commercial appli-cation and both are used success-fully in 70% of the units operatingwith synthetic crude feed blends.


References

1. Canada National Energy Board.“Canada’s Oil Sands Opportunities andChallenges to 2015: An Energy MarketAssessment.” June 2006, <www.neb.gc.ca>

2. Petti, N. et. al., “Recent CommercialExperience in Improving Refinery Profitabilitywith Grace Davison Alumina Sol Catalysts,”NPRA 2006, AM-06-68

3. R. W. Fowler and L. McDowell, NPRA 2003,“The AdVanta Edge: a Novel FCC CatalystDevelopment from ExxonMobil and DavisonCatalysts”; NPRA 2003, AM-03-27.

4. Schiller, R., et.al., “The Genesis CatalystSystem,” Catalagram® 102, Fall, 2007.

5. Wallenstein, D. et.al., “Recent Advances inthe Deactivation of FCC Catalysts by CyclicPropylene Steaming (CPS) in the Presence andAbsence of Contaminant Metals”; AppliedCatalysis A: General, 204, (2000) 89-106

6. Cotterman, R.L., et al., “Effects ofGasoline Composition on Octane Number”;Proceedings of the Symposium of the Div. ofPetroleum Chemistry, ACS Mtg., Sept. 10-15,1989.

7. Zhao, X., et.al., “FCC Bottoms CrackingMechanisms and Implications for CatalystDesign for Resid Applications”; NPRA 2002,AM-02-53.

8. Ng, S. H., et. al., Energy Fuels 2004, 18,160-171.


race Davison’s SuperDESOX® SOx reduction addi-tive is an effective way to

reduce wet gas scrubber causticconsumption and improve the over-all economics of SOx removal.

Sodium hydroxide or caustic soda isone of the most widely used com-modity chemicals. Refiners use it inthe FCCU wet gas scrubbers (WGS)to remove SO2. Caustic pricingtends to be cyclical, depending notonly on the demand for caustic butalso on the demand for the co-pro-duced chlorine. The slow down inthe housing market has drasticallyreduced the demand for chlorinederivatives, limiting caustic produc-tion and causing significant priceincrease in caustic soda. Despite arecent reduction in spot pricingfrom the high of $1000/ton in 2008,the average cost for caustic sodaremains over 100% higher thanaverage pricing just two years ago.This run up in price has promptedrefiners to consider use of SuperDESOX® to reduce the SO2 loadingon the WGS.

The high efficiency of Super DESOX®

at modest SOx reduction levels makesit an economically attractive option toreduce WGS caustic consumption.Using Super DESOX® to reduce theamount of SOx going to the WGS pro-vides refiners an opportunity to lowercaustic consumption and minimize theoverall cost of controlling SOx emis-sions. To illustrate the potential sav-ings, let’s consider an FCCU withuncontrolled SOx emissions of 500ppm. We assumed a moderate SuperDESOX® efficiency that yielded a pick-up factor (PUF) of 25 at 60% SOxreduction. We also assumed that for

every mole of SOx removed, 2.2moles of caustic could be eliminat-ed. Optimal savings occur between40 and 50% SOx reduction and canexceed $350,000 per year in justcaustic purchases, taking intoaccount the cost of the additive.Additional savings may result fromreduced water, utilities, and wastedisposal.

For more information on reducingWGS caustic consumption, pleasecontact your Grace Davison repre-sentative.

Improve Wet Gas Scrubber Economicswith Super DESOX®

Caustic = $800/ton

Est

imat

edA

nn

ual

Sav

ing

s

$0

$50,000

$100,000

$150,000

$200,000

$250,000

$300,000

$350,000

$400,000

$450,000

$500,000

20 30 40 50 60 70 80 90

% SOx Reduction via Super DESOX

Caustic = $700/ton Caustic = $600/ton

Annual Savings

G

Break the Bottom of Your Barrelwith Grace Davison’s MIDAS® FCC Catalyst

Grace Davison’s revolutionary MIDAS®-300 catalyst is commercially proven to upgrade

the bottom of your FCC barrel. Designed by Grace Davison specifically for max bottoms

destruction, MIDAS®-300 catalyst offers high activity matrix surface area, balanced with

an optimized zeolite level.

MIDAS®-300 FCC catalyst is just the latest result of our 60+ year commitment to

understanding bottoms cracking mechanisms and catalysts. Contact us for information

on how Grace Davison can provide catalytic solutions to your bottoms cracking needs.

Grace Davison Refining Technologies7500 Grace DriveColumbia, Maryland 21044 USA+1.410.531.4000 • +1.410.531.8245 (fax)www.e-catalysts.com

roper management of FCCfeed hydrotreater outages isbecoming increasingly

important as refiners continue torely heavily on hydrotreating tomeet their per gallon gasoline sul-fur limits. Continued implementa-tion of gasoline sulfur regulations,both globally and locally, are forc-ing some refiners to either expandexisting FCC feed hydrotreatingcapacity or increase severity ontheir existing hydrotreaters andhydrocrackers. Both options are agreater financial hardship intoday’s unstable economy, requir-ing additional capital expense orincreased operating costs from ahigher frequency of hydrotreaterand hydrocracker turnarounds.

The use of Grace Davison GSR®

technology in the FCC unit canreduce the sulfur reduction requiredby the FCC feed hydrotreater, result-ing in longer hydrotreater catalystlife and a lower severity operation.Refiners continue to use GSR® topreserve hydrotreater catalyst lifeand delay hydrotreater outages. Inaddition to delaying an outage, ithas been proven commercially costeffective to continue the use ofGSR® technology during ahydrotreater turnaround in order toprovide FCC feedstock flexibilitywhile maintaining gasoline pool sul-fur compliance. Coordinated effortswith Grace Davison have allowedrefiners to baseload their FCCUinventory and achieve sulfur reduc-tion targets in as little as 2 weeks.

Commercially available GraceDavison GSR® technologies that haveevolved from over 16 years of continu-ous R&D efforts include D-PriSM®,SuRCA®, GSR®-5, and NEPTUNETM.These products have been used inover 85 FCC units worldwide, with andwithout hydrotreating hardware, toprovide 20%-45% gasoline sulfurreduction from the FCCU. GraceDavison GSR® technologies create

economic advantages around feed-stock blending, operating flexibilityduring hydrotreater outages, gaso-line stream blending options andadvantages with naphtha post treat-ing. In-unit reduction of FCC gaso-line sulfur continues to create a vari-ety of opportunities and options forrefiners to drive profitability.

# of Applications

> 365Days

86 total applications worldwide including 17 current users

0 10 20 30 40 50 60

> 100Days

TIME

1.4Years

8.1Years

Longest Application

(ongoing)

Average Length for86 Sulfur ReductionApplications Worldwide

38

57

Grace Davison Experience in Sulfur Reduction Applications

GSR® Technologies Reduce Severity onthe FCC Feed Hydrotreater and ExtendCatalyst Life

P


roduct color is a common con-cern for refiners with a numberof petroleum products includ-

ing kerosene, jet fuel, diesel fuel andlube base oils. With the introduction ofultra low sulfur diesel (ULSD) the issueof diesel product color has becomemore of an issue as the typical ULSDunit cycle length may now be limitedby color degradation of the product.Refiners have been uncertain aboutend of run (EOR) reactor outlet tem-peratures with expectations in therange of 710-760°F. The typical ULSDunit has a deactivation rate in therange of 2-3°F/month so an increasein EOR temperature of 10-20°F has asignificant impact on a refiner’s plan-ning and economics.

Figure 1 summarizes data from a com-mercial ULSD unit using ART cata-lysts. The data shows that in this case

the product color exceeded 2.5ASTM, the pipeline color specifica-tion for diesel, at a reactor outlettemperature above 730°F. The feedto this unit contained 30% LCO andit was operated at 1.0 LHSV and850 psig inlet pressure.

It is well known that the color of distil-late products is affected by the reac-tion conditions in the hydrotreater,especially temperature and hydrogenpartial pressure. As (outlet) tempera-ture increases and/or hydrogen par-tial pressure decreases, the productcolor degrades. It is also generallyaccepted that the species responsi-ble for color formation in distillates arepolynuclear aromatic (PNA) mole-cules. Some of these PNA’s aregreen/blue and fluorescent in colorwhich is apparent even at very lowconcentrations of these species.Certain nitrogen (and other polar)

Greg RosinskiTechnical Services Engineer

Brian WatkinsTechnical Services Engineer

Charles OlsenWorldwide Technical ServicesManager

Advanced Refining TechnologiesChicago, IL USA

P

Factors Influencing ULSD Product Color


compounds have also been impli-cated as problems for distillateproduct color and product instabili-ty. These species can polymerize toform condensed aromatic struc-tures which tend to be green to yel-low/brown in color and can alsoform sediment via oxidation andfree radical reactions.(1) Work con-ducted by Ma et.al.(2) concluded thatthe specific species responsible forcolor degradation are anthracene,fluoranthene and their alkylatedderivatives. These are both threeringed aromatic structures and areshown in Figure 2.

PNA’s such as these are readily sat-urated to one and two ringed aro-matics under typical dieselhydrotreating conditions at start ofrun (SOR), but as the temperatureof the reactor increases towardsEOR, an equilibrium constraint maybe reached whereby the reversedehydrogenation reaction becomesmore favorable. At some combina-tion of low hydrogen partial pres-sure and high temperature thedehydrogenation reaction predomi-nates and PNA’s begin to formresulting in a degradation of thecolor of the diesel product. Otherwork completed by Takatsukaet.al.(3,4) showed that the color bod-

ies responsible for diesel productcolor degradation were concentratedin the higher boiling points in thediesel (>480°F). This suggests thatcolor can be improved by adjustingthe diesel endpoint. They also sug-gest that the color bodies responsiblefor color formation in desulfurizeddiesel are newly formed PNA struc-tures from desulfurized aromatic com-pounds.

To learn more about color degradationin ULSD, ART completed a pilot plantstudy which investigated diesel prod-

uct color over a wide range of oper-ating conditions. The study utilizedspent ART CDXi, a premium highactivity CoMo catalyst for ULSD.The sample of spent catalyst hadbeen in commercial diesel servicefor well over a year and had a car-bon content of 10.9 wt.%. The test-ing program included straight run(SR) diesels, a 30 vol.% LCO blendand a 30 vol.% light coker gas oil(LCGO) blend. The properties ofall the feeds are listed in Table I.The test was designed to examinethe effects of H2 partial pressure,H2/Oil ratio and temperature onULSD product color. H2 partialpressure varied from 300-1150 psiand the H2/Oil ratio covered therange of 700-2100 scfb.

Figure 3 shows how the diesel prod-uct color changes with temperatureand pressure for the straight runfeed (SR #1). Not surprisingly, pres-sure clearly has a significantimpact. At the lowest operatingpressure, which corresponds to300-350 psi H2 pressure, the prod-uct color exceeds 2.5 ASTM at atemperature greater than 740-750°F.Doubling the unit pressure to 800psig allows the temperature toincrease to 780°F before the prod-uct color reaches 2.5 ASTM, and ateven higher pressures the productcolor is well below 2.5 ASTM for allpractical temperatures encounteredin ULSD processing. At these con-

Pro

du

ctC

olo

r(A

ST

M)

Average Bed Temperature ˚F

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

580 600 620 640 660 680 700 720 740 760 780

Figure 1ULSD Product Color Using a SmART Catalyst SystemTM

anthracene fluoranthene

Figure 2Primary Fluorescence Species in

Hydrotreated Diesel (Ref 2)


ditions the H2 partial pressureincreases by a factor of about 3.4going from 400 psig up to 1200 psigtotal pressure for temperaturesaround 750°F.

The data in Figure 3 were generatedat the low end of H2/Oil ratios inves-tigated. Figure 4 shows the effect ofincreasing the H2/Oil ratio at 400psig on the SR #1 feed. At this lowpressure, the H2/Oil ratio has a sig-nificant impact on product color.The data show that the temperaturecan be increased from 745°F to wellabove 760°F before product colorexceeds 2.5 ASTM. AT 400 psigtotal pressure and 760°F, changingthe H2/Oil ratio from 700 to 2100scfb results in a 10% increase in H2partial pressure which appears tobe enough to keep the reactionenvironment on the favorable side ofthe hydrogenation-dehydrogenationequilibrium curve. At higher operat-ing pressures the impact of increas-ing the H2/Oil ratio is reduced whenprocessing the SR feed, but still hasa positive effect on suppressingproduct color.

As might be expected, adding LCOto the ULSD unit feed makes theproduct color situation worse.Figure 5 compares the productcolor for the SR feed and the 30%LCO feed at 2100 scfb H2/Oil ratioand two pressures. The SR feedresults in acceptable color over thewide range of temperatures for bothpressures shown. This compares

˚APISulfur, wt.%Nitrogen, wppmTotal Aromatics, vol.%PNA’s (2-ring+), vol.%ASTM ColorDistillation (D2887), ˚F

IBP10%50%90%FBP

SR #133.270.6812031.510.7L3.5

275454580600837

SR #1/LCO27.730.9329340.021.9L6.5

265439565703841

SR #234.441.1212727.39.5

L2.0

222477613681740

SR #2/LCGO33.631.3424930.110.8L5.5

238437579667754

Table IFeedstock Properties

AS

TM

Co

lor

Temperature, ˚F

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

680 700 720 740 760 780 800

Straight Run

800 Psig

400 Psig

1200 Psig

Figure 3Product Color Improves with Pressure

AS

TM

Co

lor

Temperature, ˚F

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

680 690 700 710 720 730 740 750 760 770

Straight Run

700 scfb H2

2100 scfb H2

Figure 4Product Color Improvement with Increased H2/Oil Ratio


with the 30% LCO feed which goesoff color at about 730-740°F at 800psig; and at 1200 psig the tempera-ture can exceed 760°F beforereaching 2.5 ASTM color. This datademonstrates the significant impactthat pressure has on diesel productcolor when processing feeds thatcontain LCO.

Figure 6 demonstrates the effects ofthe H2/Oil ratio on product colorwhen processing the 30% LCOfeed. It shows the temperature atwhich the color reaches 2.5 ASTMas a function of H2/Oil ratio for both800 and 1200 psig total pressure.The temperature increases by about25°F when the H2/Oil ratio isincreased from 700 to 2100 scfb.That range of H2 rates correspondsto an increase in hydrogen partialpressure of 5-10%

The pilot plant program also investi-gated the effects of a coker derivedmaterial on ULSD product color.Figure 7 compares the productcolor for the second SR feed and a30% LCGO/70% SR #2 blend at 800psig. The data indicates that thefeed containing LCGO behaves sim-ilarly to the SR feed. In both casesthe outlet temperature can exceed780°F before product colorapproaches the ASTM 2.5 level.This is not surprising when compar-ing the properties of the two feeds.The aromatics level, and in particu-lar the PNA concentrations, areessentially the same for the SR andthe coker blend. Compare this withthe LCO blend shown in Table Iwhere the PNA’s are twice that ofthe SR or LCGO feeds.

As mentioned previously, it is gen-erally accepted that product coloris related to PNA’s, and earlier workhas concluded that specific three-ringed aromatics are responsiblefor color degradation in diesel.Figure 8 shows a comparison ofthe product PNA’s (three-ring aro-matics) and diesel product colorfor all the feeds and conditions ofthe study. It is readily apparent thatthe PNA’s correlate reasonably wellwith product color..

AS

TM

Co

lor

Temperature, ˚F

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

640 660 680 700 720 740 760 780 800

LCO

SR

800 psig

1200 psig800 psig

1200 psig

Figure 5Comparison of Product Color for SR and 30% LCO

Tem

per

atu

refo

r2.

5A

ST

MC

olo

r,˚F

H2/Oil Ratio, scfb

700

710

720

730

740

750

760

770

780

500 1000 1500 2000 2500

800 Psig

1200 Psig

Figure 6Effects of H2/Oil Ratio on Product Color for 30% LCO

Temperature, ˚F

AS

TM

Co

lor

0.0

0.5

1.0

1.5

2.0

640 660 680 700 720 740 760 780 800

SR #2

LCGO

Figure 7Product Color Comparison for LCGO and SR


From these data it is clear thathydrogenation of PNA’s is key tomaintaining acceptable productcolor in ULSD. This suggests a cou-ple of approaches that allow anincrease in EOR outlet temperaturesand thereby increase the ULSD unitcycle length.

One approach which has been putto commercial practice is toincrease quench to the bottom bedof the hydrotreater. This accom-plishes two things which are impor-tant to maintaining a good environ-ment for hydrogenation of PNA’s. Itreduces the outlet temperature andhelps to increase the outlet hydro-gen partial pressure relative to no orlower amounts of quench.

This, of course, requires that theupper beds of the hydrotreater berun at higher WABT’s in order tomaintain the required HDS conver-sion. This means that the furnacemust have sufficient capacity toachieve the higher inlet tempera-tures. Operating in this manneroffers the potential to add an addi-tional 10-20°F on to the cycle lengthdepending on the unit capabilities(furnace, quench capacity).

Another approach, which may beimplemented with the one just dis-cussed, involves adjusting the feedto the unit. The data from this workshows the significant impact LCO

has on diesel product color. Reducing(or eliminating) the amount of LCO inthe feed will help to suppress productcolor degradation as the unitapproaches EOR. There is also datashowing that the color bodies thatcause problems for ULSD tend to beconcentrated at the higher boilingpoints of the distillation on thefeed/product. Reducing the endpointof the LCO reduces the concentrationof these species which will help main-tain acceptable product color as theunit moves towards EOR.

References

1. J. Pedley et.al, ACS Division of FuelChemistry, 35 (4), 1100-1107 (1990).

2. X. Ma et. al., Energy and Fuels, 10, pp91-96 (1996).

3. T. Takatsuka et.al., 1991 NPRA AnnualMeeting, Paper AM-91-39.

4. T. Takatsuka et.al., Journal of the JapanPetroleum Institute, Vol. 23, No. 2, pp 179-184, 1992.

Product PNAs (3 rings+), vol.%

AS

TM

Co

lor

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0.0 1.0 2.0 3.0 4.0 5.0

SR

LCO

LCGO

Figure 83+ Ring Aromatics Correlate with Diesel Product Color


Grace Davison Refining TechnologiesGrace Davison Refining TechnologiesDelivering customized, environmentally friendly,

catalytic solutions for the dynamicpetroleum refining industry.

Surface ScienceMatrix ModificationsAcidity and Porosity

Zeolite SynthesisNanotechnologyActivity and Selectivity

ContaminationResistance

Clean FuelsEmissions Reduction

AnalyticalCapability

ExperiencedEngineers

ApplicationExpertise

StrongCustomer

Relationships

IndustryRecognizedTechnical

Service

ManufacturingFlexibility

Leading-edgeMaterialScience

Grace Davison Refining TechnologiesAdvanced Refining TechnologiesW.R. Grace & Co.-Conn.7500 Grace DriveColumbia, MD 21044410.531.4000

Grace Davison Refining TechnologiesAdvanced Refining TechnologiesAsia PacificW.R. Grace (Singapore) Pte. Ltd.#07-02 Wheelock PlaceSingapore 23888065.6737.5488

Grace Davison Refining TechnologiesAdvanced Refining TechnologiesEuropeGrace GmbH & Co. K.G.In der Hollerhecke 1Postfach 1445D-67545 Worms, Germany49.6241.40300

[email protected] [email protected]

®

Catalagram®, Grace®, Grace Davison®, MIDAS®, OlefinsMax®, GSR®, OlefinsUltra®, OlefinsMax®, AURORA®, ADVANTA®, Super DESOX®,D-PriSM®, SuRCA®, GSR®-5 and SmART Catalyst System® are registered trademarks in the United States and/or other countries,of W. R. Grace & Co.-Conn.

DXTM , NEPTUNETM and GENESISTM are trademarks of W. R. Grace & Co.-Conn.

This trademark list has been compiled using available published information as of the publication date of this brochure and may notaccurately reflect current trademark ownership.

© 2009 W. R. Grace & Co.-Conn.

Documents

Grace - High-Performance Specialty Chemicals and …...The GENESIS TM Catalyst System www. grace.com 7500 GraceDrive¥Columb ia,MD21044USA Mos trefiners need flexiblecatalystsystems