THE INFLUENCE OF CATALYST AND FEEDSTOCKON PRODUCT DISTRIBUTIONS

FOR CRACKING OF c. PARAFFINS ON HY ANDHZSM-5 ZEOLITES

Zainurlls ZainuddlnChemistry Department, University of Tasmania

GPO Box 252C, Hobart 7001, Tasmania, Australia

ABSTRACTCatalytic cracking of n-octane and 2-methylheptane, as well

as mixtures of these paraffins, has been studied on HY andHZSAI-5 zeolites, and also on combinations of the catalysts. Forcracking on individual zeolites, both feedstock and catalystinfluence the resulting product distributions. For all feedstocks,product distributions are shifted towards smaller fragments forreaction on HZSM-5 compared to flY.

Ratios of branched to linear paraffins are much more stronglyinfluenced by catalyst type and feedstock than the correspondingratios for olefinic products. For reactions on catalyst mixtures,distributions of the Iota I products by carbon number correspondwell to a summation of contributions on the individual catalysts.However, a greater departure from prediction is seen for in-dividual distributions of paraffins, olefins and aromatics, as wellas for ratios of branched 10 lineal' paraffins, showing that hydro-gen transfer processes and isomerization must occur. The addi-tion of pentasil has also resulted in enrichment of the linearsaturates at 101-,'ercarbon number which is due to preferentialcracking of linear paraffins over the branched isomers.

INTISARIReaksi pemecahan n-okt ana, 2-metilheptana, dan campuran

kedua parafin ini pada zeolit flY, zeolit lIZSM-5, dan campurankedua katalis tela]t dipelajari. Dalam reaksi pemecahan padamaslng-masing zeolit, baik pereaksi maupun: katalis keduanyamempen garuhi distribusi hasil reaksl, Dlstribusi total hasil reaksipada zeolit HZSM-5 bergeser ke [ragmen yang lebili kecil apabiladibandingkan dengan distribusi husil reaksi pada zeolit JlY, danini berlaku pada kedua paraJin.

Perubalian ratio parafin bercabanglrant ai lurus pada keduakatalls lebili peka dibandingkan perubalian ratio olefin. Dalamreaksi-reaksi pada campuran katalls, distrlbusi total luisil reaksibcrdasarkan nomor karbon dapat dlsamakan dengan pen-[umlahan dlstribusi hasil reakst dari keduo kat alls secara ter-pisali. Namun demikian, terdapat devlasi yang menonjol padadlstribusi parafin, olefin, don aromatik. Penyimpangan jugaterdapat pada ratio parafin bercabangjrantai lurus, yang ke-semuanya menunjukkan terjadinya proses pemindalian hidrogendan lsomerisasi. Penambahun pentasil juga menyebabkan pening-kat an jumlalt parafin berantai pendek don lurus sebagai hasilreaksi, ini disebohkan karena reaksi pemecalian parafin berantailurus lcblh mudali berlangsung daripada reaksi pemecalianparafin dengan rantai bercabang.

30

INTRODUCTIONThe petroleum industry is one of the most important

industry in the world today in term of the value of its finalproducts. Many studies on the reactions of hydrocarbons oncracking catalysts have been undertaken since thecommercial introduction of the catalysts. Most of theseinvestigations have been concentrating on the crackingbehaviour of hydrocarbons on single component catalysts.And to a lesser extent on combinations of catalysts.

Combinations of catalysts have been used duringcommercial cracking processes since the last decade. Nu-merous patents describe many different formulations ofcracking catalysts, but the scientific literature containsmuch less information relating cracking phenomena ofcatalyst combinations to those of their components.

There has been recent interest in the effects of usingcombinations of zeolites as cracking catalysts. In particular,it has been found that addition of small amounts of HZSM-5 to HY zeolite can enhance the octane number of thegasoline produced [1-11]. Several explanations have beenproposed to account for the effects of the pentasil additiveduring cracking of gas oils. Some investigators [1, 6] havesuggested that there is a relative increase in the cracking ofgasoline range linear paraffins, leading to higher propor-tions of the branched unsaturated isomers, contributing to ahigher octane rating. Other investigator [8] reported thatenhanced olefin cracking by the addition of HZSM-5, inpreference to hydrogen transfer, would reduce formation ofadditional saturated products on the Iaujasitc.

In spite of these studies, some of the principles arc stillunder ongoing debate and are not clearly understood. Thisaccumulated body of research shows that catalytic reactionof hydrocarbons with dual-zeolite function is worthy offurther investigations. In this study, the cracking cha-racteristics of Cg paraffins on HY and HZSM-5, and on arandom mixture of these zcol ites have been examined.

EXPERIMENTALThe feedstocks, n-octane (99.7%) and 2-methylheptane

(99.8%) were obtained from Aldrich and used withoutfurther purification. Impurities present in l1-octane were n-

JKTI, VOL 4 - No.1, Junl, 1994

hexane (0.2%), 2-methylheptane (0.08%) and isopentane(0.02%) while the impurities in 2-methylheptane were 3-methyl heptane (0.12%) and n-octane (0.08%).

HY zeolite (97.3% exchanged) was prepared from NaY(Linde Lot No 45912, SK40; Si/Al = 2.4 ) by repeatedexchange with O.5M ammonium nitrate solution. HZSM-5(Si/AJ = 105) was provided by SNAM Progretti S.P.A.Milan, Italy. Catalysts of mesh size 80/100 were calcined at500°C before use.

Experiments were performed on catalysts of mesh size80/100 (-175 micron) which was diluted in a sand matrixthat had been thoroughly washed with concentrated HCIand neutralized with concentrated NH3. As the sand actedas an inert support material, it was crucial that it wascatalytically inactive. For this reason, the sand matrix usedwas soaked in concentrated HCI for several days to removeimpurities present in the sand. The liquor was replacedeveryday to maintain the cleaning activity of the acid. Aftersoaking, the sand was washed thoroughly with distilledwater followed by neutralization with concentrated am-monia. Finally, the sand was filtered and rinsed withdistilled water several times before it was oven dried at500°C overnight. The activity of the sand was tested byperforming the experiments in the absence of catalyst(thermal cracking).

Preheating Zone

Gaseous productsto water buretterr===~

Liquid products werecollected in the. flask

Ice bath

Figure 1: Experimental Apparatus

JKTI, VOL 4 - No.1, Junl, 1994

The experimental apparatus, as shown in Figure 1,consisted of a quartz reactor about 60 em long and 2 em i.d.The reactor was enclosed in an aluminium alloy furnacewith three thermocouple probes attached to it. The reactorconsisted of two zones: a preheating zone at the top sectionand a catalyst zone at the bottom. The preheating zone wasfilled with inert material, beryl saddles (Be~12Si6018)' Thecatalyst used for a particular run was placed in the catalystzone diluted with the sand matrix. Glass wool was used toseparate the zones and to hold the catalyst-sand matrix inplace.

For a set of experiments, a known amount of catalystwas placed in the catalyst zone, and a known amount offeed was fed in from the top of the reactor, so that therewould be a defined catalyst to feed ratio (CIF). The feedwas pumped by a motor driven syringe that could be set togive different rates, resulting in different time on stream(T.O.S.), that is the time required to deliver a fixed amountof feed into the reactor.

Prior to a catalytic run, the catalyst was conditioned bypurging the reactor with dry, C02-free nitrogen at 0.50L/min for 5 minutes. The feed was then injected into thetop end of the reactor at a constant rate by a motor drivensyringe pump. The liquid products exiting the bottom endof the reactor were collected in a round bottom flask im-mersed in an ice bath, while the gaseous products weretrapped in a 3 L gas burette by displacement of water. Afterthe run, the reactor was once again purged with nitrogen atthe same flow rate for the same period of time as before inorder to flush down all remaining liquid and gases left inthe reactor,

The liquid products were analysed using a HewlettPackard 5890 II gas chromatograph with an SGE fusedsilica capillary column (50 m x 0.22 mm i.d.) and flameionization detector. Gaseous products were also ana lysedusing a Hewlett Packard gas chromatograph of the sametype with a Chrompak capillary column (25 m x 0.32 mmi.d.). Data handling was facilitated using a DAPA softwarepackage. Identification of hydrocarbon products was assis-ted by use of a Hewlett Packard 5890 gas chromatographcoupled to a 5970 mass selective detector.

The GC temperature program for analysing the gaseousproducts was isothermal at 30°C for 3 minutes, followed byfirst ramping at 6°C/min to 60°C then at 30°C/min up to200°C at which it was held for 8 minutes before it thencooled down to the initial temperature. The carrier gas usedwas a mixture of helium and nitrogen set at 27 kf'a headpressure. The GC program for ana lysing the liquid productswas isothermal at 30°C for 6 minutes, followed by firstramping at 5°C/min to 150°C and 25°C/min to 200°C atwhich it was held for 12 minutes, then cooling down to theinitial temperature. The carrier gas used was the same asthat for analysing the gaseous products, set at 165 kPa headpressure.

After a run, the catalyst became poisoned by thc car-bonaceous material, regarded as coke, left in its pores. This

31

was indicated by the black appearence of the catalyst. Inorder to achieve catalyst regeneration, the reactor tem-perature was set to 500°C with dry, COr free air waspassing through. The gas stream coming off the bottom endof the reactor contained water vapour, CO and smallamounts of CO2 that were formed from decomposition ofthe coke. This gas stream then passed through a drieriteabsorbant, in which the water vapour was removed, then analumina supported platinum catalyst (Pt/Al203) held at350°C, in which the carbon monoxide was converted tocarbon dioxide. The stream was finally passed through anascarite absorbant, where all the carbon dioxide wasabsorbed. It was assumed that the weight increases of theseabsorbants provides a reliable quantitative measure of thecoke produced. The duration of the catalyst regenerationvaried between 1 to 3 hours depending on which and howmuch zeolite was used, as well as times on stream. Thisregeneration time was determined to be sufficient byperforming duplicate runs at the same time on stream.Regeneration of HY zeolite normally took place in about1.5 to 3 hours, and within 1to 2 hours for HZSM-5 zeolite.With this procedure, it was possible to perform repetitiveexperiments with reproducible results.

RESULTS AND DISCUSSIONFor individual paraffins and the mixed feedstocks,

experiments were carried out over a range of times onstream, generally between 500 and 1500 seconds. It has

been shown that both HY and HZSM-5 undergo ageingduring cracking reactions of n-octane at 400°C. Applica-tion of the time on stream theory to these systems [14]shows that for reaction of n-octane on HY between 75%and 95% of active sites would be poisoned with these limitsof run duration. On HZSM-5 under similar conditionsbetween 35% and 62% of active sites are poisoned. Ingeneral, the observations reported here were found to beindependent of time within the experimental run durationsstudied, unless otherwise stated.

A. Reactions of Individual Feedstocks

Distribution of Products by Carbon Number

Results in Figures 2 and 3 compare the total productdistributions from cracking of 2-methylheptane and n-octane on HY and HZSM-5 respectively, at similarconversion levels. Product distributions from reactions onHY show that the tendency to produce C1 to C3 fragmentsis higher for the linear paraffin. In contrast, for reactions onHZSM-5, the proportions of C1 and C2 cracking fragmentsare higher for reaction of the isoparaffin, However, patternsof overall fragmentation are similar on both catalysts(Figures 2a and 3a), showing much less variation thanchanging the catalyst using a particular feedstock (Figure4). Greater variations can be seen by examination ofindividual distributions of paraffins, olefins and aromatics(Figures 2b-d and 3b-d).

rocarbons

( • = 2-Methylheptane; ~ = n-Octane )

1.0

2 4 6 7 8 9

c) efinsE-U:::> 0.3Qo0::Q., 0.2

~...:Io 0.1::;

3 6 7

CARBON NUMBER

7

Figure 2 Product distributions from reaction of 2-methylhepUlIle (cat/feed = 0.0620. 25.9% conversion)and n-octane (cat/feed = 0.0579. 17.0% conversion) on BY at 400°C

2 4 5

32

0.6

0.5

0.4

0.3

0.2

0.1

0.02 3 4 5 6 7 8

0.03

0.Q2

0.Ql

6 8 9

JKTI, VOL 4 - No.1, Junl, 1994

( • = 2-Methylheptane; ~ = n-Octane )

Paraffins1.2

0~E-<CI::~;>Z0U0~~~~....'l0~ 0.8

E-<U

0.6;:J00CI:: 0.4~~....'l0 0.2

~0.0

2

2 3 4 5 6 7 8 9

3 4 5 6

CARBON NUMBER

7

2 3 6 7 84 5

As can be seen in Figure 2d, there is little difference inthe total amount or distribution of aromatics produced fromeither paraffin feedstock on HY. It has been suggested [15]that aromatics are formed during cracking of paraffinsma inl y through secondary processes involving olefins.Figure 2c shows that the distribution and amounts ofpotential olefin precursor are similar for both feedstocks.For reaction on HZSM-5, there is a significantly highertendency to produce aromatics from reaction of theisoparaffin (Figure 3d). For example, the amount ofbenzene produced is ten times greater for reaction of 2-methyl heptane at similar conversion levels. However, ifolefin precursors are aga in considered, Figure. 2c showsthat amounts and distributions are similar, suggesting that,on HZSM-S aromatic products are formed directly from thefeedstock, as well as through olefins in secondary pro-cesses.

0.02 d) Aromatics

0.Ql

0.006 8 97 •

Figure 3 Product distributions from reaction of 2-methylheptane (cat/feed = 0.0416, 22.4% conversion)and n-octane (cat/feed = 0.0179.18.0% conversion) on HZSM-5 at 400"C

Ratios of Branched to Linear IsomersThe relative amounts of branched and linear isomers

present in gasoline for both olefins and paraffins isimportant in determining the octane rating. A pcntasiladdition to an HY zeolite can influence the ratio ofbranched to linear (B/L) isomers produced during crackingof gas oils [3, 7, 8]. Figure 5a gives ra tios of B/L isomersfor C4 and Cs olefin products, showing the influence ofboth catalyst and feedstock for reaction of 2-111ethylheptaneand n-octane. In general, higher ratios arc observed for both

JKTI, VOL. 4 - No.1, Junl, 1994

( • = HY ; fili.I = HZSM·5)

·3

Figure 4 Product distributions from reaction of2-111cthylheptane 011HY (cat/!eed = O.O~20.25.9% conversion) ,u1(1HZSM·5 (car/feed= 0.0416. 22.4% conversion) at 400 C

CARBON NUMBER

33

(a) Branched to Linear Olefins ( • = C4 ;. ~ = Cs )

o....E-<<~r.l..:l,~ (b) Branched to Linear Paraffins ( • = C4 ; D= Cs ; ~ = C6)

30

" " iO' E~ ~ uS S,. ,.

0: ~ 0: :;:N ~0: 0:

" " iO' E:;: :;: u

'" '"S S,.

i,.

~0: 0:

N ~0: 0:

CATALYST (FEED)

Figure 5 Ratios of B/l isomers from reaction of Zcmethylbeptane (2MH) and .r-ocmne (OCT)on HY and HZSM-5 at 400°CFor reaction of Zcruethylheptnne on HY: cat/feed = 0.0620, 25.9% conversion

on HZSM-5: cat/feed = 0.0416, 22.4% conversionFnr reaction of »-octnne 011HY : cat/feed = 0.0579, 17.0% conversion

on I-IZSM-5: cat/feed = 0.0179,18.0% conversion

C4 and C5 olefins for reactions of either feedstock on thepentasil. The effect of changing either the catalyst or thefeed on the ratios of BIL olefins is more noticeable for theC4 species.

Variations in the ratios of BIL paraffin products aremuch more significant than observed above for the oIefins(Figure 5b). In contrast to the olefinic products, bothcatalyst and feedstock exert a major influence on theobserved degree of branching of the saturated crackingfragments. For both feedstocks, the ratios of BIL paraffinproducts is significantly higher for reaction on HY com-pared to HZSM-5._This effect is seen particularly for the C,paraffins, where, for both feedstocks the observed BILratios are 12 - 24 times higber on the faujasite, Figure 5balso shows that, for either catalyst the ratios of BILsaturated isomers are significantly higher for cracking ofthe branched feedstock. This result shows that theintermediates formed are not identical for cracking of thelinear and branched Cg paraffins. The dominant crackingpathway on either catalyst cannot result from equilibriumof the various possible branched and linear Cs carboniumions as previously suggested [16].

Branched C6 paraffins can result from cracking of 2-methylheptane by direct cleavage of a C-C bond whichwould produce exclusively 2-methylpentane. Alternatively,bimolecular processes leading to a C6 carbonium ion afterB-scission could give rise to either 2-methylpentane or 3-

34

methylpentane after rearrangement of the residual C6carbenium ion. Rearrangement of a C6 carbcnium ion togive branched isomers occurs on both HY and HZSM-5 toproduce similar product distributions [17, 18]. For skeletalisomerization of linear hexenes, the ratios of 2-methyl-pentenes to 3-methylpentenes formed were 0.76 and 0.62on HY and HZSM-5 respectively [17, 18]. Figure 6 showsa marked preference for formation of 2-methylpentanerather than 3-methylpentane for cracking of 2-methyl-heptane on both HY and HZSM-5. The ratios of the C6

isomers formed suggest that on both catalysts, directcracking via protolysis is favoured over processes in-volving an intermediate C6 carbenium ion. Results obtainedover a range of times on stream as illustrated in Figure 6show that the reported phenomena are not stronglyinfluenced by catalyst decay.

(a) 2-Methylpcntanc (0) and 3-Mcthylpentane (e) on HY

on 18

~<:>~ 16...CJ= 14'00... 12A......

j[)0

'"'"'0 ------------:E400 600 800 WOO 1200 1400 1600

(b) 2-Methylpentane (0) and 3-Methylpentane (.) on HZSM-S

801~---------------- _

ow=;=::;:::;==~:=;::=~--l400 600 800 WOO 1200 1400 1600

TIME ON STREAM (seconds)

Figure 6 Formation of2-methylpentane and 3-methylpentane from reaction ofz-rrerhylheptnne on HY (cat/feed = 0.0620,25.9% conversion)and HZSM-5 (cat/feed = 0.0416,22.4% conversion) at 400°C

Reaction on a Random Mil/lire of HY and HZSM-5

Product distributions for reactions of n-octanc and 2-mcthylhcpranc respectively on a random mixture of HYand HZSM-5 are shown in Figures 7 and 8. Calculateddistributions are also presented, using the product distri-butions for reactions on the individual catalysts, assumingthat the relative contribution is proportional to activities onthe isolated zeolite. Although the overall distributions onthe random mixture of HY and HZSM-5 are well

JKTI, VOL. 4 - No.1, Juni, 1994

represented by the calculated values (Figures 7a and 8a),greater deviations are seen considering the observed andpredicted distributions for individual product classes. These

deviations can be attributed to additional processes whereproducts formed on one catalyst interact with the secondcatalyst, particularly in secondary processes involving

(.~ OBSERVED ; ~= PREDICTED)

0.8

0~ 0.6~~~>-0.4Z

0U0 0.2~~'""~ 0.0...J

2 3 4 5 6 7 8 9 2 3 4 5 6 80::;;- 0.5 c) Olefins 0.03 d) A~U;:> 0.4Q0 0.02~ 0.3Q..

~...J 0.20 0.01::;;

0.1

0.0 0.002 3 4 5 6 7 6 7 8 9

CARBON NUMBERFigure 7 Product distributions from reaction of 2-methylheptane on a random mixture of HY and HZSM-5

(HYIHZSM-5 = 1 : 0.67, 34.6% conversion) at 400°C

a) Total Hydrocarbons

(.= OBSERVED; ~= PREDICTED)

0.02 3 4 5 6 7 8 9

c) Olefins 0.05

0.04

0.03

0.02

0.01

0.003 4 5 6 7 6

0~ 0.8~~~ 0.6>-Z0 0.4UQ~ 0.2~'""~ 0.0~0::;;- 0.4~U;:> 0.300~Q..

0.2~...J0::;; 0.1

0.02

0.6

0.4

0.2

2 3 4 5 6 7 8

d) Aromatics

7 8 9

CARBON NUMBERFigure 8 Product distributions from reaction of n-octane on a random mixture of HY and HZSM-5

(HYIHZSM-5 = 1 : 0.31. 32.1 % conversion) at 400°C

JKTI, VOL. 4 - No.1, Junl, 1994 35

hydrogen transfer and isomerization. Figure 9 shows acomparison between observed and predicted BIL ratios forboth olefins and paraffins. Although observed ratios for theolefins are close to predicted values, addition of HZSM-5to the HY zeolite leads to a cracking product distributionwith a somewhat higher degree of branching of the productolefins. Predicted ratios for the C4 - C6 paraffins showgreater deviation from the observed values (Figure 9b) withmore branching observed than expected for reaction of thebranched isomers, while the reverse is true for reaction of.the linear Cg paraffin.

(.= OllSERVEJ); ~= PREDICTED)

a) Olefins

c. <5RANDOM (2MH)

c. <5RANDOM (OCT)

RANDOM (2M II) RANDOM (OCT)

CATALYST MIX (FEED)

Figure 9 Ratios of BIL isomers from reaction of 2-methylheptane (2MH) andn-octanefOCI) on a random mixture of HY and HZSM-5 at 400°CFor reaction of2-methylheptane: HY/HZSM-5 = 1:0.67,34,6% conversionFor reaction of n-octane: HY/HZSM-5 = 1: 0.31, 321% cceversion

B. Reactions of Mixed Feedstocks

The relative rates of conversion of Cs isomers in amixed feedstock (1 : 1 molar ratio), shown in Figure 10,agree with those of Chen and Garwood [19]. A randommixture of the catalysts produces preferential cracking ofthe linear isomer, in agreement with previous studies [2J oncomplex feedstocks. Figure 11 shows a comparison ofproduct' distributions from cracking the paraffin mixtureshifted to lower molecular size on the pentasil. Ra tios ofBIL paraffin isomers are significantly higher for reaction onHY compared to HZSM-5 (Figure 12). In contrast, thecorresponding ratios for C4 and Cs olefins are increased by

36

the pentasil. Both effects were previously observed forreaction of the individual feedstocks. Figure 12 also showsobserved ratios for olefinic products corresponding closelyto calculated values, with ratios for paraffins significantlyhigher than calculated.

(0= 2-Mcthylhcptailc ;. = /I-Octanc)

5()·-r-..,.--::--,----:-:-:-:-------~a) Reaction on II Y

40

30

r--.500 1000 1500 2000

~50E-< b) Reaction on HZSM-S

~ 40'--' ......--------Z0 30~ifJ

20p.:;~ __ 0---0-

~10

0U 500 1000 1500 2000

60c) Reaction 011 Random Mixture

50

40

30

20()(JSIX) I(XX) 15(X)

T.O.S. (seconds)

Figure 10 Conversion of 2-methylheptane and n-octane in a mixed feed fromreaction on HY, HZSM-5 and a random mixture of the catalysts at 400°C

Product distributions from reaction of the Cs paraffinmixture on a random mixture of the zeolites is given inFigure 13. As for reactions on the individual catalysts, thedistribution of total cracking products is well representedby the predicted distribution. The BIL ratios for C4 olefins(Figure 14a) lie between those for reaction on theindividual catalysts whereas for Cs olefins this ratio ishigher than cracking on either HY or HZSM-5 (Figure12a). Production of the branched isomers is higher thanpredicted. Figure 14b shows that the ratios for paraffins liebetween those obtained on the individual catalysts (Figure12b), but aresomewhat lower than predicted.

Cracking of a mixture of paraffins on a random mixtureof zeolites can be considered in the context of reportedstudies for cracking of complex gas oil mixtures concerningthe impact of addition of HZSM-5 to a Y zeolite. It has

JKTI, VOL 4 - No.1, Juni, 1994

(. = HY; ~ = HZSM-5 )

1.2 a) Total Hydrocarbons 0.7

1.0 0.60~ 0.5E- 0.8~ 0.4~~ 0.6Z 0.3

0 0.4U 0.2

0 0.2 0.1~~~ 0.0 0.0~2 3 4 5 6 8 9~ 7 2

.....:l0 0.6 ns 0.03::;--.E-U;:J 0.4 0.0200~~~ 0.2 O.oI.....:l0:;;

b) Paraffins

3 4 6 85 7

d) Aromatics

0.00

2 3 6 8 94 6 7 75

CARBON NUMBER

Figure 11 Product distributions from reaction of a mixture of 2-methylheptane and n-octaneon HY and HZSM-5 at 400°C (TOS = 1000 seconds)

(b) Branched to Linear Paraffins

(a) Branched to Linear OIefins ( • = e.r;m = c 5)

320to Linear Olefins b) Branched to Linear Paraffins

2

1

oHY HZSM-5 HY* HZSM-5* HY HZSM-5. HY* HZSM-5*

CATALYST

Figure 12 Ratios of BIL isomers from reaction of a mixture of 2-methylheptane and II-octaneon HY and HZSM-5 at 400°C (TOS = 1000 seconds)Calculated values are denoted by asterix (*)

JKTI, VOL. 4 - No.1, Junl, 1994 37

(. = HY; !Zi:l = HZSM-5)

1.0 a ons 0.6 b) Paraffins

Q0.8WE-<

=::: 0.4W 0.6;>Z0 0.4U 0.2

Q 0.2WW~

0.0 0.0W 2 3 4 5 6 7 8 9 2 3 4 5 6 7 8•••0~ 0.6 0.Q3 d)---E-<U;:JQ 0.4 0.020~Q..W 0.2 0.01....:l0-~

0.0 0.002 3 4 5 6 7 6 7 8 9

CARBON NUMBER

Figure 13 Product distributions from reaction of a mixture of 2-methylheptane and n-octaneon a random mixture of HY and HZSM-5 at 400°C (TOS = 1000 seconds)

(b) Branched to Linear Paraffins

(a) Branched to Linear Olefins ( • = C4; ~= C 5)

a) Olefins b) Paraffins

OBSERVED PREDICTED" OBSERVED PREDICTED

Figure 14 Ratios of BIL isomers from reaction of a mixture of 2-lI1cthylheptanc and II-octaneon a random mixture of HY and HZSM-5 at 400°C (TOS = 1000 seconds)

38 JKTI, VOL. 4 - No.1, Juni, 1994

been suggested that pentasil addition produces selectiveremoval of low octane linear paraffins, and can contributetowards enhancement of gasoline quality. However, thepresent studies consistently show that cracking of eitherlinear or branched paraffins results in a greater proportionof linear paraffin isomers at lower molecular weight. Incontrast, reported cracking studies on gas oils indicatehigher B/L ratios of C4 and C5 paraffins on pentasiladdition [3, 6, 8]. This suggests that enhanced selectivecracking of linear paraffin isomers by HZSM-5 is not thedominant mechanism for octane rating enhancement.Higher ratios for B/L isomers of light olefins have alsobeen reported for studies on gas oils [2, 5, 6], in agreementwith the present studies. The presence of HZSM-5,however, contributes to the isomerization of olefins,increasing the proportion of branched isomers, therebyenhancing the octane rating of the gasoline produced.

CONCLUSIONThis present study shows that both catalyst and

feedstock are important factors in determining productdistribution for cracking of paraffins, particularly on ratiosof branched to linear paraffins produced. Preferentialcracking of linear paraffins rather than branched isomers ina complex mixture by introduction of a pentasil additivemay be expected to result in enrichment of the linearsaturates at lower carbon number. As this has not beengenerally observed for gas-oil cracking, selective crackingof linear paraffins by the pentasil additive may not be thedominant mechanism leading to octane enhancement of thegasoline produced.

ACKNOWLEDGEMENTThe author is indebted to the Research and Develop-

ment Centre for Applied Chemistry of the IndonesianInstitute of Sciences for the financial support.

REFERENCES

..

1. K. Rajagopalan and G.W. Young. HydrocarbonCracking Selectivities with Dual-Function ZeoliteCatalysts.A.C.S. Symp. Ser. 375: 34-48 (1988).

2. J. Biswas and I.E. Maxwell. Octane Enhancement inFluid Catalytic Cracking: I. Role of ZSM-5 Additionand Reactor Temperature. Appl. Catal. 58: 1-18 (1990) .

3. M.F. Elia, E. Iglesias, A. Martinez and M.A. PerezPascual. Effect of Operation Conditions on theBehaviour of ZSM-5 Addition to a RE-USY FCCCatalyst. Appl. Catal. 73: 195-216 (1991).

4. K. Rajagopalan and G.W. Young. HydrocarbonCracking Selectivities with Dual Zeolite Fluid CrackingCatalyst Containing REY and ZSM-5. Prepr.: Am.Clicm. Soc., Div. Pet. Cliem. 32(3,4): 627-631 (1987).

JKTI, VOL. 4 - No.1, Junl, 1994

5. C.D. Anderson, F.G. Dwyer, G. Koch and P. Niiranen,A New Cracking Catalyst for Higher Octane UsingZSM-5. Proc. Ninth Iberoamerican Symp. on Catalysis,Lisbon, 1984, pp247-258.

6. D.A. Pappal and P.H. Schipper. ZSM-5 in CatalyticCracking: RISER Pilot Plant Gasoline CompositionAnalyses. Prepr. -Am. Chem. Soc., Div. Pet. Chem.35(4): 678-683 (1990).

7. SJ. Miller and C.R. Hsieh. Octane Enhancement inCatalytic Cracking Using High-Silica Zeolites. Prepr. -Am. Cltem. Soc., Div. Pet. Cltem. 3(4): 685-693 (1990).

8. RJ. Madon. Role of ZSM-5 and Ultrastable Y Zeolitesfor Increasing Gasoline Octane Number. J. Catal. 129:275-287 (1991).

9. G. Anders, I. Burkhardt, U. IlIgen, I.W. Schulz and J.Scheve. The Influence of HZSM-5 Zeolite on theProduct Composition after Cracking of High BoilingHydrocarbons Fractions. Appl. Catal. 62: 271-280(1991).

10. J.S, Buchanan. Reactions of Model Compounds overSteamed ZSMc5 at Simulated FCC Reactions Con-ditions. Appl. Catal. 74: 83-94 (1991).

11. Z. Zainuddin, F. N. Guerzoni and J. Abbot. Crackingand Isomerization of Hydrocarbons on Combinations ofHY and HZSM-5 Zeolites. J. Catal. 140: 150-167(1993).

12. J. Abbot and B. W. Wojciechowski. Hydrogen TransferReactions in the Catalytic Cracking of Paraffins. J.Catal. 107: 451-462 (1987).

13. J. Abbot and B. W. Wojciechowski. Catalytic Reactionsof n-Dodecane on Aluminosilicates. J. Catal. 115: 521-531(1988).

14. J. Abbot and F.N. Guerzoni. Role of Bronsted andLewis Sites During Cracking of n-Octane on H-Mordenite. Appl. Catal. 85: 173-188 (1992)

15. J. Scherzer. Octane-Enhancing, Zeolitic FCC Catalysts:Scientific and Technical Aspects. J. Catal. Rev. -Sci.Eng. 31(3): 215-354 (1989).

16. A. Conna, J. PlaneIles, J. Sanchez-Marin and F. Tomas.TIle Role of Different Types of Acid Sites in theCracking of Alkanes on Zeolite Catalysts. J. Catal. 93:30-37 (1985).

17. J. Abbot and B. W. Wojciechowski. Catalytic Crackingand Skeletal Isomerization of n-Hcxcnes on HZSM-5Zeolite. Canad.J. Chern. Eng. 63: 451-461 (1985).

18. J. Abbot and B. W. Wojciechowski. Catalytic Crackingand Skeletal Isomerization of n-Hexcucs on HY Zeolite.Canad.J, Chern. Eng. 63: 278-287 (1985).

19. N.Y. Chen and W.E. Garwood. Some CatalyticProperties of ZSM-5: A New Shape Selective Zeolite. 1.Catal. 52: 453-458 (1978).

39