george, 2007 usy-tpd

8/13/2019 George, 2007 Usy-tpd

1/7

Acid Site Characterization of Coked USHY Zeolite Using Temperature

Programmed Desorption with a Component-Nonspecific Detector

Baodong Wang and George Manos*

Department of Chemical Engineering, UniVersity College London, Torrington Place, London WC1E 7JE, U.K.

In this paper we describe a novel NH3-TPD methodology to study the amount as well as strength of acid sitesof coked catalysts. Conventional temperature programmed desorption (TPD) of ammonia cannot be appliedto coked catalyst as coke precursors removed during the temperature program falsify the ammonia signal.The effects of coke formation from different reactants, time-on-stream, and reaction temperatures on acidsite deactivation were investigated. Besides acid site deactivation, characterization of coke precursors canalso be revealed. The initial deactivation preferentially on strong acid sites is very fast. The concentration offree acid sites is inversely correlated well with the total concentration of coke rather than individual cokegroups. Coke precursors tend to be more stable at higher reaction temperatures.

1. Introduction

The catalytic conversion of hydrocarbons over zeolites isapplied in large-scale petroleum refining and petrochemicalprocesses, among which catalytic cracking is the most important.

It is generally recognized that the cracking performance ofzeolites as catalysts strongly depends on their aciditysnumberand strength of acid sites.1 An undesirable side reaction is theformation of coke that drastically reduces the catalyst activity.Deactivation of zeolites by coke deposition is a very importantproblem due to their technological and economic relevance. Thecoke formation, its effect on acidity, and the influence of thepresence of coke on the catalytic activity have received muchresearch attention over the years.2 Coke can poison the acidsites or block their access because these carbonaceous com-pounds generally are trapped in the zeolite pores.3 On the otherhand, the removal of coke, which is generally carried out byoxidative treatment at high temperatures, often has detrimentaleffects on the catalytic activity.4

A coke classification into coke precursors and hard cokecategories was reported in our previous work.5 Coke precursorsare removed from the catalyst sample simply through volatiliza-tion in inert nitrogen. Hard coke remains on the catalyst evenat high temperature (873 K) and is removed by burning.

The coke formation on zeolites and the effect of coke onacidity of catalyst were studied by several methods, such asIR, NMR, UV-vis, and temperature programmed desorption(TPD).3,6-8 Among them NH3-TPD seems to be one of the mostapplicable and efficient ones, providing not only the amount ofacid sites but also their strength distribution. However, theregularly used thermal conductivity detector (TCD) cannotdifferentiate the signal of ammonia and other gas-phasecomponents such as desorbed coke precursors. Hence, in thecase of nonavailability of a component-specific detector, thenormal NH3-TPD method cannot be applied to coked catalystsbecause of falsification of the TPD signal by coke precursorsremoved at high temperatures. In this research, we adopted anindirect TPD method with mild temperature sample pretreatmentto study the acidity of coked zeolites.

The objective of the present research is to apply this novelNH3-TPD method in order to investigate the changes in strength

and number of the acid sites of coked USHY zeolites duringacid-catalyzed reactions with reactants of 1-pentene,n-heptane,and ethylbenzene. The effects of time-on-stream and reactiontemperature during 1-pentene reactions were also examined.Furthermore, this method allows the characterization and

classification of coke precursors.

2. Experimental Section

2.1. Materials. The catalyst used in this study was anultrastable Y zeolite in acidic form (USHY) provided by GraceGmbh in powder form with an average crystallite size of 1 m,a framework Si/Al ratio of 5.7, and a bulk Si/Al ratio of 2.5.The micropore area was 532.4 m2/g, and the micropore volumewas 0.26 cm3/g. N2 BET surface area was 590 ( 23.5 m2/g.The catalyst was pressed into pellets, crushed, and sieved,producing particles in the size range of 1.0-1.7 mm. Beforeeach reaction, the catalyst was calcined in an oven at 873 K for12 h.

1-Pentene (99% purity) and ethylbenzene (99% purity) weresupplied by Sigma-Aldrich Chemicals.n-Heptane (99.5% purity)was from BDH Chemicals Ltd. Nitrogen (CP grade) as carriergas was supplied by BOC group.

2.2. Experimental Procedure.Coking reactions with 1-pen-tene,n-heptane, and ethylbenzene as reactants were carried outon USHY zeolite in the temperature range of 523-623 K andatmospheric pressure, in a stainless steel tubular fixed-bedreactor, with an inner diameter of 15 mm. The amount ofcalcined catalyst used in each experiment was 0.65 g (1-cm-long catalyst bed). The catalyst bed was placed between twometal meshes to ensure isothermicity, and it was indeed checkedto be isothermal using a thermocouple inserted in a small metalprotection tube that was placed in the center of the reactor. Thecarrier gas, nitrogen, passed through a saturator containing theparticular reactant placed in a heated water bath at specifiedtemperature (for 1-pentene, PN2 ) 0.2 bar, Preactant ) 0.8 bar,for n-heptane, PN2 ) 0.65 bar, Preactant ) 0.35 bar, forethylbenzene, PN2 ) 0.88 bar, Preactant ) 0.12 bar) separately,and then into the fixed-bed reactor. Unfortunately, it was notpossible to have exactly the same experimental conditions withall three reactants due to the huge volatility difference of thesecomponents. After a specified time-on-stream that denoted theend of the experimental run, the saturator was bypassed andthe reactor cooled for 10 min under nitrogen atmosphere. Then

* To whom correspondence should be addressed. Fax: +44 (0)207383 2348. E-mail: [email protected].

7977Ind. Eng. Chem. Res. 2007, 46, 7977-7983

10.1021/ie0708733 CCC: $37.00 2007 American Chemical SocietyPublished on Web 10/20/2007


2/7

the reactor was disconnected from the rig and put in ice forquick cooling to ambient temperature. Coked samples wereobtained at 1, 2, 3, 7, and 20 min of time-on-stream.

2.3. Temperature Programmed Desorption Procedure. Thecoked samples were investigated by a novel method with atemperature programmed desorption (TPD) apparatus, a Mi-cromeritics AutoChem 2910. The common method of acidityestimation using TPD needs sample pretreatment at relativelyhigh temperatures. Furthermore, the temperature program itself

causes coke precursors to volatilize and/or decompose intosmaller volatile fragments.9 This decomposition might also leavenonvolatile fragments on the catalyst surface. These nonvolatilefragments are accounted for as hard coke. Current work islooking further into the exact mechanism of removal of cokeprecursors through thermal treatment. Experiments with differentheating rates are carried out for estimating the apparentactivation energy of the process. Relatively high apparentactivation energy values should indicate decomposition as achemically activated process, while low values should indicatediffusion limitation possibly through volatilization. In both casesthese coke components cause a falsification of the TPD signalthat does not represent the ammonia amount desorbed. To avoidthis falsification due to the chemically active character of coke

precursors, we adopt indirect TPD methods with mild temper-ature sample pretreatment to study the acidity of coked zeolites.With this method the free acid sites of coked zeolites as wellas acid sites inhabited by hard coke can be quantitativelydetermined. Furthermore, the character of coke precursors canbe studied.

The TGA (thermogravimetric analysis) temperature programfor classification of coke precursors and hard coke used inprevious work5 was applied to the TPD analysis. A blank TPDanalysis was carried out in the absence of catalyst in order toverify if ammonia retained in the apparatus setup was significant.As the TCD (thermal conductivity detector) signal during thisanalysis did not increase above the baseline, one can concludethat the amount of ammonia retained by the system itself isnegligible. An around 50 mg coked catalyst sample was placedin a U-shaped quartz cell. It was then preheated at 10 K/min to473 K, where it stayed for 1 h. During the process, reagentsadsorbed on the catalyst surface and most of water moleculeswere removed. After cooling to 353 K, adsorption of ammoniawas carried out in a He stream (10% NH3, 20 mL/min). Afterthe catalyst surface became saturated, the loaded sample washeated to 383 K at 10 K/min for physisorbed NH3 to bedesorbed. The linear temperature program (10 K/min) was thenstarted from 383 to 873 K and remained at 873 K for 30 min.The desorbed ammonia and coke precursors were monitoredcontinuously with a thermal conductivity detector. This is calledthe first TPD. The fresh zeolite was also analyzed by this method

to measure the total free acid sites. The total number of acidsites could be found quantitatively by driving the desorbedammonia from NH3-TPD experiments to a standard HCl solutionof a specified concentration, e.g., 0.1 M, and subsequent titration.

The same TPD was carried out initially without precedingammonia adsorbed. This was called TPD without ammonia.Furthermore, availability of GC/MS or HPLC/MS instrumentscoupled to the TPD rig could characterize the molecularcomposition of the coke components removed. A second TPDwith ammonia was carried out with 873 K for preheating 30min instead of 473 K. During this period of preheating, cokeprecursors as well as water adsorbed and reaction mixturecomponents were removed. The first TPD contains bothammonia and coke precursors adsorbed on the zeolite. However,

the TPD without ammonia contains only coke precursors.Comparing the signal of TPD without ammonia with that of

the first TPD, it is obvious that they overlap in the high-temperature zone for any coking system due to removal of cokeprecursors and the strong acid sites, which contribute at the hightemperature (above 600 K) ammonia signal, being deactivated.By subtracting the signal of TPD without ammonia from thesignal of the first TPD, the free acid sites of coked sampleinhabited by both coke precursors and hard coke can becalculated as presented in Figure 1. The first TPD of fresh zeolitedetermines directly the total acid sites of fresh zeolite, which isalso shown in Figure 1. Since coke precursors have beenremoved after the pretreatment at 873 K in He stream duringthe second TPD, only hard coke remained deposited on thecoked catalyst. Therefore, the signal of the second TPD detectedby TCD is due exclusively to desorbed ammonia and reveals

the free acid sites of coked zeolite samples inhabited only byhard coke, after complete removal of coke precursors. All theseTCD signals were normalized by the zeolite weight excludingcoke components. These calculation processes can be clearlyillustrated in Figure 2a. With these methods, not only the amountof acid sites but also the acid site strength distribution of cokedzeolites can be determined. In order to make the temperaturetreatment very clear, we also present the whole temperatureprogram in Figure 2b.

Since the samples used were exposed to high-temperaturecycles, to test their thermal stability the acidity of a regeneratedcatalyst was also measured. The NH3-TPD curve of theregenerated catalyst after removal of all coke components at873 K in air was exactly the same as this of the fresh sample.

The structure of the fresh catalyst samples being ultrastabilizedzeolite and having been calcined at 873 K stays intact atexposure at high temperatures.

3. Results and Discussion

3.1. Effect of Different Reactants.The indirect TPD methodwas applied to deactivated catalysts coked by different reac-tants: an alkane (n-heptane, 35% in N2, residence time ) 0.178s, weight-hourly space velocity (WHSV) ) 59.579 h-1), analkene (1-pentene, 80% in N2, residence time ) 0.055 s, WHSV) 86.211 h-1), and an aromatic hydrocarbon (ethylbenzene, 12%in N2, residence time ) 0.239 s, WHSV ) 25.886 h-1). Becauseof the huge volatility differences of these reactants, it was not

Figure 1. Description of free acid sites by the difference of first TPD andTPD without NH3 (1-pentene reaction, T) 623 K, TOS ) 1 min). Forcomparison is shown the acidity of fresh USHY zeolite.

7978 Ind. Eng. Chem. Res., Vol. 46, No. 24, 2007


3/7

possible to have the same experimental conditions with them.With all reactants the reaction temperature (623 K) and TOS(20 min) of analyzed samples were the same. With these threedifferent reaction systems, different coking mechanisms takeplace, resulting in differences in deactivation of acid sites andcomposition of coke precursors.10 TPDs without ammonia ofdeactivated USHY zeolite coked by 1-pentene, n-heptane, andethylbenzene at the reaction temperature of 623 K and 20 minof TOS are displayed in Figure 3. The peaks in the TCD outputare due to coke precursors in the carrier gas. The area belowthe TPD thermograms would be proportional to the amount ofcoke precursors if the composition of different coke precursors

had the same TCD signal response factor. These three TPDintegrals and the corresponding amounts of coke precursor

content measured by thermogravimetric analysis are presentedin Table 1. The fact that the ratios of the TPD area to thecorresponding coke TGA weight are similar indicates that theTCD signal response factors are not profoundly different. Fromboth methods, TPD and TGA, the order of formation of cokeprecursors is 1-pentene > ethylbenzene > n-heptane.

From Figure 3, it can be clearly seen that most of the cokeprecursors from 1-pentene reactions were removed at hightemperatures. There are two coke precursor desorption peaksfrom 1-pentene reactions: a small one located at 540 K and alarge one located at 760 K. This suggests that a small part ofthe coke precursors can be removed at low temperature whilemost of the coke precursors are more stable and can be removedat higher temperatures. As for n-heptane, a small peak at 540

K is attributed to the easy removal of coke precursors and arelatively large peak at 650 K is attributed to stable cokeprecursors. There is only one coke precursor peak for ethyl-benzene residing at 627 K, indicating that it is much more easilyremovable than in the case of 1-pentene. Coke precursorsproduced by 1-pentene are more difficult to remove than thoseby then-heptane/ethylbenzene systems at low temperature andneed much higher temperatures for that. This is possibly due tothe faster progress of coking during 1-pentene reactions. Cokingis not a static process, and transformations between cokecomponents take place continuously. Coke precursors transformto more stable ones, which convert further to hard coke. Notonly the amount of coke precursors with 1-pentene is a lot higherthan with the other reactants, but also the hard coke with

1-pentene is an order of magnitude higher than the hard cokewith ethylbenzene or n-heptane.

The estimated, by the method described in section 2.3, acidsites of deactivated USHY zeolites coked during 1-pentene,n-heptane, and ethylbenzene reactions respectively are presentedin parts a, b, and c of Figure 4. For 1-pentene and n-heptanesystems, the signals of TPD without ammonia and the first TPDoverlap from 590 to 873 K, while for ethylbenzene system thetwo signals overlap from 650 to 873 K. This suggests that lessstrong acid sites have been poisoned during ethylbenzenereactions compared to 1-pentene/n-heptane systems. The areain the low-temperature range obtained from the difference inTPD curves between the TPD without ammonia and the firstTPD is due to the ammonia adsorption. In all cases strong acid

sites are occupied by coke first and deactivated. Hence, ammoniacan only be adsorbed at weak acid sites (showing maximaaround 500 K) left after coking. The phenomenon confirms thatcoke preferentially deactivates the strongest acid sites.11

Figure 5 shows the second TPD for coked catalyst samplesfrom 1-pentene, n-heptane, and ethylbenzene reactions. Sincecoke precursors had been removed through pretreatment at 873K in inert flow, only hard coke remained on the catalyst beforethe second TPD. The second TPD signal presents the free acidsites not occupied by hard coke. They include acid sites whichwere occupied by coke precursors but have been freed throughthe removal of coke precursors during the pretreatment.Although hard coke contents from thermogravimetric analysis(TGA) results are quite different, 0.78 gcoke/100 gzeolite for

Figure 2. (a) Procedure of mild temperature pretreatment and indirect TPD.(b) TPD temperature program.

Figure 3. TPD without ammonia of deactivated USHY zeolite coked during

different reactant systems (T) 623 K, TOS ) 20 min).

Table 1. Coke Precursors Content from Different Reactants

Measured by TPD and TGA (T) 623 K, TOS ) 20 min)

1-pentene ethylbenzene n-heptane

TPD area [auK] 333 205 121TGA weight [mg/gcat] 65.1 38.2 19.8TPD area/TGA weight 5.1 5.4 6.1

Ind. Eng. Chem. Res., Vol. 46, No. 24, 2007 7979


4/7

n-heptane, 1.03 gcoke/100 gzeolite for ethylbenzene, and 15.40gcoke/100 gzeolite for 1-pentene, respectively, the free acid sitesnot coked by hard coke for these three different reactants arenearly the same. This also means that the acid sites blockedonly by hard coke are almost the same. The number of acid

sites occupied or blocked by hard coke are not proportional tothe content of hard coke. If we set

then R1-pentene (1.45) , Rethylbenzene (23.09) Rn-heptane (26.01).Hard coke is formed on strong acid sites where the adsorbates

are strongly adsorbed (chemisorbed) and possess intense acidsite catalytic properties for cracking reactions. For 1-pentenesystem, coke is formed during cracking reaction through asequence of reaction steps, such as protonation, alkylation,isomerization, hydride transfer, deprotonation, and ring closure.12

This kind of hard coke molecule on each strong acid site seemsto be much larger and heavier than that from ethylbenzene andn-heptane systems. Another possible explanation for the very

different R-values might be the very different amounts of hardcoke concentrations: around 1% forn-heptane and ethylbenzenecompared to 15% for 1-pentene. The initially formed cokemolecules might be able to spread further apart, deactivatingproportionally a larger number of acid sites. Larger cokeamounts lead to a denser packing in the narrow zeolitic poreswhich decreases the number of acid sites deactivated per coke.

3.2. Effect of Time-On-Stream (TOS). The TPDs withoutammonia of coked USHY zeolite during 1-pentene cracking at623 K at various TOS are presented in Figure 6. The integralarea of the TPD curve increases with TOS, indicating that theamount of coke precursors increases with increasing TOSalthough slightly stronger than TGA results indicate (Figure 7).This might be due to differences in TCD response factors

especially as the composition of the coke precursors change asshown below. The corresponding TPD area/TGA weight valuesare nonetheless inside the (10% error indicated in Table 1. Itcan be also observed that there are two peaks in the TPD signalresulting in two types of coke precursors. The first small peaklocated at relatively low temperature represents the cokeprecursors which can be removed more easily, while the secondstands for more stable coke precursors. The first peak becomessmaller with TOS while the second peak becomes larger,indicating a transformation of coke precursors from one typeto another. Through aromatization, coke deposits become largerand more aromatic with TOS and coke content.13,14 Since cokeformation is an extremely fast process at the beginning ofcatalyst exposure to the reaction mixture, most coke precursors

Figure 4. (a) First TPD, TPD without NH3, and free acid sites ofdeactivated USHY zeolite coked during 1-pentene reactions (T) 623 K,TOS ) 20 min). (b) First TPD, TPD without NH 3, and free acid sites ofdeactivated USHY zeolite coked during n-heptane reactions (T) 623 K,TOS ) 20 min). (c) First TPD, TPD without NH3, and free acid sites ofdeactivated USHY zeolite coked during ethylbenzene reactions (T) 623K, TOS ) 20 min).

Figure 5. Second TPD of deactivated USHY zeolite coked during differentreactant systems (T) 623 K, TOS ) 20 min).

R )no. of acid sitesoccupied by hard coke

concn of hard coke

area of fresh -area of 2nd TPD

ghard coke/100 gzeolite



5/7

are formed in the first minute TOS. Although the amount ofcoke precursors increases slowly, their composition still changesvia solid surface reactions of coke components. The unstablecoke precursors convert to stable coke precursors with TOS,resulting in the decrease of the amount of unstable cokeprecursors and the corresponding increase of the amount ofstable coke precursors.

Figure 8 shows the first TPDs of deactivated USHY zeolitesamples coked at 623 K reaction temperature at different TOS.The first peaks are larger at each TOS than the corresponding

ones in Figure 6 due to the additional adsorption of ammoniaon the weak acid sites. The corresponding second peaks atdifferent TOS in both figures overlap due to saturation of cokeon strong acid sites leading to complete poisoning of strongacid sites.

The free acid sites of coked USHY zeolites from 1-pentenereaction at 623 K from different TOS are shown in Figure 9.Compared with the fresh catalyst, the amount of free acid sitesof coked catalyst decreases with TOS. It is obvious that thefirst minute of TOS sample suffers a very fast strong initialacid site deactivation with a relative slow acidity deactivationafterward. After 7 min, the acid sites almost do not decreaseany more. Furthermore, the acid site deactivation is correlatedwith the content of total coke. Also, acid site distribution can

be illustrated with Figure 9. The loss of acid sites is morepronounced at strong acid sites than at weak acid sites, whichconfirms the higher contribution of strong acid sites on cokedeposition.

From Figure 10 we can see that the acid sites deactivated byhard coke at different TOS periods are not profoundly differentas the content of hard coke increases slightly with increasingTOS (Figure 7). This could be explained by the rapid initialcoking on strong acid sites causing their full deactivation. Theamount of strong acid sites decreases rapidly at the start ofexposure of catalyst to the reaction mixture. After the strong

acid sites have been deactivated in a very short time, cokecontinues to deposit on weak acid sites with a much lower rate.At the same time, hard coke still continues to accumulate andgrows on these strong acid sites at a lower rate.

3.3. Effect of Reaction Temperature. As shown in Figure11 (data obtained by TGA measurements), the amount of hardcoke increases with increasing reaction temperature while theamount of coke precursors decreases. This can be explained bythe fact that coke precursors can transform to hard coke fasterat high temperatures. However, the amount of total coke almostdoes not change with reaction temperature. Figure 12 displaysthe TPD without ammonia at various reaction temperatures(TOS ) 20 min) against TPD temperature. The integral area ofthe TPD without ammonia curve decreases with increasing

Figure 6. TPD without ammonia of deactivated USHY zeolite coked during1-pentene reactions at different TOS (T) 623 K).

Figure 7. TGA-measured coke content of deactivated USHY zeolite cokedduring acid catalytic cracking reaction of 1-pentene at different TOS (T)623 K).

Figure 8. First TPD of deactivated USHY zeolite coked during 1-pentenereactions at different TOS (T) 623 K).

Figure 9. Free acid sites of deactivated USHY zeolite coked during1-pentene reactions at different TOS (T) 623 K).



6/7

reaction temperature; that is correlated well with the amount ofcoke precursors from TGA results. Taking into account thethermal conductivity detector (TCD) working principle, it seemsreasonable to assume that coke precursor molecules are removed

from the catalyst without decomposition. It also can be seenthat with increasing reaction temperature the peaks derived fromcoke precursors shift from low TPD temperature to high TPDtemperature. Coke precursors formed at high reaction temper-atures are more difficult to remove than those formed at lowreaction temperatures. Coke precursors become more stable andcontribute to hard coke with increasing reaction temperature,resulting in less integral area located at high TPD temperatures.

The free acid sites of the coked catalyst reacted at differenttemperatures (Figure 14) are calculated from the difference of

the corresponding first TPD (Figure 13) and TPD without NH3(Figure 12). There is not much difference among the three freeacid site curves. The amount of free acid sites agrees well withthe amount of total coke (Figure 11). Moreover, the free acidsite distribution is very similar. Both coke precursors and hardcoke contribute to acid site deactivation. The effect of the slightincrease of the concentration of hard coke with temperature iscompensated by the slight decrease of the concentration of cokeprecursor. Hence, reaction temperature does not have a distincteffect on the amount of total coke and acid site deactivation.

Figure 14 compares the TPD of the acid sites of fresh catalyst,the TPD of free acid sites of coked catalyst (first TPD minusTPD without NH3, i.e. free acid sites not blocked by cokeprecursors and hard coke), and the second TPD (free acid sites

Figure 10. Second TPD of deactivated USHY zeolite coked during1-pentene reactions at different TOS (T) 623 K).

Figure 11. TGA-measured coke content of deactivated USHY zeolite cokedduring acid catalytic cracking reaction of 1-pentene at different reactiontemperatures (TOS ) 20 min).

Figure 12. TPD without ammonia of deactivated USHY zeolite cokedduring 1-pentene reactions at different reaction temperatures (TOS ) 20min).

Figure 13. First TPD of deactivated USHY zeolite coked during 1-pentenereactions at different reaction temperatures (TOS ) 20 min).

Figure 14. TPD of fresh catalyst, TPD of free acid sites, and second TPDof deactivated USHY zeolite coked during 1-pentene reactions at differentreaction temperatures (TOS ) 20 min).



7/7

not blocked by hard coke). The gap between the fresh and thesecond TPD, i.e., the acid sites inhabited only by hard coke, ismainly located at high TPD temperature area, suggesting strongacid site deactivation. On the other hand, the gap between thesecond TPD and free acid sites, i.e., that of acid sites onlyinhabited by coke precursors, lies preferentially at relativelystrong acid sites, too. Furthermore, the area of acid sites blockedby coke precursors is larger than that of hard coke, in contrastto the corresponding TGA results (Figure 11), where the weight

of hard coke is higher than that of coke precursors. This meansthat the number of acid sites blocked per mass unit of cokecomponents is higher for coke precursors than for hard coke.This is in good agreement with the fact that hard coke moleculesare more aromatic and bigger than coke precursors. Thus, eventhough the molar amount of hard coke is lower than that ofcoke precursors, its weight is larger than that of coke precursors.

4. Conclusions

A novel temperature programmed desorption (TPD) methodusing mild temperature pretreatment and combining of TPDwithout and with ammonia is introduced in order to investigatethe deactivation of acid sites by coking. The method allows usto quantitatively measure the free acid sites of coked and freshsolid acid catalyst as well as their strength distribution. Themethod also provides information on acid site deactivationcaused by hard coke only. Since coking is of great interest tothe petroleum refineries that use commercial FCC catalysts inlarge quantities, the application of the method on industrialcatalysts would give a better understanding on the effect ofcoking on catalyst deactivation and will help the design of tailor-made catalysts with fewer coking problems.

A series of TPDs with and without NH3 are carried out onthe coked sample: a first TPD (with NH3), a TPD without NH3,and a second TPD (with NH3). The difference of the first TPDand the TPD without NH3represents the free acid sites of cokedcatalyst. The second TPD provides information on acid sites

inhabited by hard coke. Furthermore, the amount of cokeprecursors as well as coke precursors stability can be determinedby TPD without ammonia.

The USHY zeolite suffered a strong reduction of free acidsites especially at the initial stage during 1-pentene reaction,while it slowed considerably afterward. The concentration offree acid sites is inversely correlated well with the totalconcentration of coke. Coke preferentially deactivates thestrongest acid sites. Coke precursors become more stable withtime-on-stream and increasing reaction temperatures.

Acknowledgment

The financial support of the Overseas Research StudentsAwards Scheme and K. C. Wong Scholarship is gratefullyacknowledged.

Literature Cited

(1) Borges, P.; Pinto, R.; Lemos, M.; Lemos, F. Activity-acidityrelationship for alkane cracking over zeolites: n-hexane cracking overHZSM-5. J. Mol. Catal. A: Chem. 2005, 229, 127.

(2) Cumming, K. A.; Wojciechowski, B. W. Hydrogen transfer, cokeformation, and catalyst decay and their role in the chain mechanism ofcatalytic cracking. Catal. ReV.sSci. Eng. 1996, 38, 101.

(3) Hopkins, P. D.; Miller, J. T.; Meyers, B. L.; Ray, G. J.; Roginski,R. T.; Kuehne, M. A.; Kung, H. H. Acidity and cracking activity changesduring coke deactivation of ultrastable Y zeolite. Appl. Catal., A: Gen.1996, 136, 29.

(4) Biswas, J.; Maxwell, J. E. Recent process- and catalyst-relateddevelopments in fluid catalytic cracking. Appl. Catal. 1990, 63, 197.

(5) Chen, S.; Manos, G. Study of coke and coke precursors duringcatalytic cracking ofn-hexane and 1-hexene over ultrastable Y zeolite. Catal.

Lett. 2004, 96, 195.(6) Gil, B.; Mierzynska, K.; Datka, J. In situ IR and catalytic studies of

the effect of coke on acid properties of steamed zeolite Y. MicroporousMesoporous Mater. 2007, 99, 328.

(7) Derouane, E. G.; Gilson, J. P.; Nagy, J. B. In situ characterizationof carbonaceous residues from zeolite-catalysed reactions using highresolution solid state 13C-NMR spectroscopy. Zeolites 1982, 2, 242.

(8) Dedecek, J.; Capek, L.; Wichterlova, B. Nature of active sites indecane-SCR-NOx and NO decomposition over Cu-ZSM-5 zeolites. Appl.Catal., A: Gen. 2006, 307, 156.

(9) Cerqueira, H. S.; Ayrault, P.; Datka, J.; Guisnet, M. Influence ofcoke on the acid properties of a USHY zeolite. Microporous Mesoporous

Mater. 2000, 38, 197.(10) Cerqueira, H. S.; Ayrault, P.; Datka, J.; Magnoux, P.; Guisnet, M.

m-Xylene Transformation over a USHY Zeolite at 523 and 723 K: Influenceof Coke Deposits on Activity, Acidity, and Porosity. J. Catal. 2000, 196,149.

(11) Magnoux, P.; Cartraud, P.; Mignard, S.; Guisnet, M. Coking, aging,and regeneration of zeolites: II. Deactivation of HY zeolite duringn-heptanecracking. J. Catal. 1987, 106, 235.

(12) Guisnet, M.; Magnoux, P. Organic chemistry of coke formation.Appl. Catal., A: Gen. 2001, 212, 83.

(13) Matsushita, K.; Hauser, A.; Marafi, A.; Koide, R.; Stanislaus, A.Initial coke deposition on hydrotreating catalysts. Part 1. Changes in cokeproperties as a function of time on stream. Fuel 2004, 83, 1031.

(14) Holmes, S. M.; Garforth, A.; Maunders, B.; Dwyer, J. A solventextraction method to study the location and concentration of coke formedon zeolite catalysts. Appl. Catal., A: Gen. 1997, 151, 355.

ReceiVed for reView June 26, 2007ReVised manuscript receiVed August 28, 2007

Accepted August 30, 2007

IE0708733


george, 2007 usy-tpd

Documents