Indian Journal of Engineering & Materials Sciences Vol. 5, October 1998, pp. 285-290 Surface analysis of supported platinum catalysts-encapsulation model for metal-support interaction: An XPS study R MaIathi, P Madhusudhan Rao, B Viswanathan & R P Viswanath Department of Chemistry, Indian Institute of Technology, Chennai 600 036, India Received 14 January 1998; accepted 29 June 1998 The surface properties of supported platinum catalysts have been investigated using hydrogen chemisorption and XPS. A mixed oxide system, PtlfiOrAh03 has been compared with titania-supported systems. A metal- support interaction after high temperature reduction (HTR) has been observed in these systems. The chemisorption measurements indicate a weak suppression in hydrogen uptake after HTR for the mixed oxide system. The encapsulation of the metal particles by the reduced support after reduction at 773K has been elucidated based on the difference in the binding energies between oxygen atom and the support metal. This has been found to be more marked for Ti0 2 prepared from gel and for mixed oxide samples than for the commercial Ti0 2 . The role of the SMSI state in catalytic activity towards reduction of cinnamaIdehyde has also been investigated. Supported metal catalysts have been used extensively in heterogeneous catalysis. Many investigations have been inspired by the possibility that an interaction between the metal crystallite and the support can alter the chemisorptive properties of the catalyst. Evidence for this concept has been provided by Sinfelt et al.l Earlier, Schwab" and Solymosi and Szabo" have attempted to investigate metal-support interactions induced by supporting the metal with reducible transition metal oxides. This effect has been studied using noble metals supported on Ti0 2 , Ah03. Si0 2 and other oxides. The interaction of the noble metals with Ti0 2 support has been deduced from H2 and CO chemisorption measurements at room temperature by Tauster and Fung'' and it has been shown that these systems exhibit the so called "strong metal-support interaction" (SMSI). The various reasons attributed to metal-support interaction, especially SMSI state, are (i) an ordered or arranged dispersion of the metal in a particular geometry, (ii) topology, (iii) the formation of compounds, (iv) phase transformation of the support as well as (v) electron transfer from support to metaI 6 • 7 • Madhusudhan Rao et al. s , have postulated the formation of a ternary oxide phase and its subsequent re-organization under reduction at 773 K in hydrogen as the cause for the induction of SMSI state in the Feffi0 2 system. The aim of this paper is to examine this postulate, especially with respect to Ptffi02 system, which is known to exhibit pronounced SMSI characteristics, employing Ti0 2 support (prepared in a similar manner to that reported by Madhusudhan Ra0 9 ) and to compare these results with those obtained from a mixed oxide support (TiOrAh03). Mixed' oxide supports have been employed and their role in inducing SMSI state has been examined in the past. McVicker and Ziemiack lO considered the effect of Ti0 2 -Ah03 support for group vm metals and reported on the basis of hydrogen chemisorption measurements that the effect is in between that of Ti0 2 (an SMSI support) and of Ah03 (considered to be a non-SMSI support).Other reports available with mixed oxides of TiOrAh03, Nb 2 0 s -Si0 2 and Ti02-Si0211-'4 have addressed to the question of the relative dispersions of the two oxides as well as to the question of the chemical state of the supports. The objective of the present investigation includes the study of the mixed oxides in comparison to simple oxides by different characterization techniques, especially X-ray Photo Electron Spectroscopy (XPS). Platinum was supported on titania (commercial), titania (gel) and a mixed oxide containing 22% titania and 78% alumina. The catalysts were characterized by X-ray diffraction (XRD), hydrogen chemisorption measurements, sorptometry (BET surface area) and XPS techniques.
Transcript
Indian Journal of Engineering & Materials SciencesVol. 5, October 1998, pp. 285-290
Surface analysis of supported platinum catalysts-encapsulation model formetal-support interaction: An XPS study
R MaIathi, P Madhusudhan Rao, B Viswanathan & R P Viswanath
Department of Chemistry, Indian Institute of Technology, Chennai 600 036, India
Received 14 January 1998; accepted 29 June 1998
The surface properties of supported platinum catalysts have been investigated using hydrogen chemisorptionand XPS. A mixed oxide system, PtlfiOrAh03 has been compared with titania-supported systems. A metal-support interaction after high temperature reduction (HTR) has been observed in these systems. The chemisorptionmeasurements indicate a weak suppression in hydrogen uptake after HTR for the mixed oxide system. Theencapsulation of the metal particles by the reduced support after reduction at 773K has been elucidated based on thedifference in the binding energies between oxygen atom and the support metal. This has been found to be moremarked for Ti02 prepared from gel and for mixed oxide samples than for the commercial Ti02. The role of theSMSI state in catalytic activity towards reduction of cinnamaIdehyde has also been investigated.
Supported metal catalysts have been usedextensively in heterogeneous catalysis. Manyinvestigations have been inspired by the possibilitythat an interaction between the metal crystalliteand the support can alter the chemisorptiveproperties of the catalyst. Evidence for this concepthas been provided by Sinfelt et al.l Earlier,Schwab" and Solymosi and Szabo" have attemptedto investigate metal-support interactions inducedby supporting the metal with reducible transitionmetal oxides. This effect has been studied usingnoble metals supported on Ti02, Ah03. Si02 andother oxides. The interaction of the noble metalswith Ti02 support has been deduced from H2 andCO chemisorption measurements at roomtemperature by Tauster and Fung'' and it has beenshown that these systems exhibit the so called"strong metal-support interaction" (SMSI).
The various reasons attributed to metal-supportinteraction, especially SMSI state, are (i) anordered or arranged dispersion of the metal in aparticular geometry, (ii) topology, (iii) theformation of compounds, (iv) phase transformationof the support as well as (v) electron transfer fromsupport to metaI6•7• Madhusudhan Rao et al.s, havepostulated the formation of a ternary oxide phaseand its subsequent re-organization under reductionat 773 K in hydrogen as the cause for the inductionof SMSI state in the Feffi02 system. The aim ofthis paper is to examine this postulate, especially
with respect to Ptffi02 system, which is known toexhibit pronounced SMSI characteristics,employing Ti02 support (prepared in a similarmanner to that reported by Madhusudhan Ra09
)
and to compare these results with those obtainedfrom a mixed oxide support (TiOrAh03).
Mixed' oxide supports have been employed andtheir role in inducing SMSI state has beenexamined in the past. McVicker and ZiemiacklO
considered the effect of Ti02-Ah03 support forgroup vm metals and reported on the basis ofhydrogen chemisorption measurements that theeffect is in between that of Ti02 (an SMSI support)and of Ah03 (considered to be a non-SMSIsupport).Other reports available with mixed oxidesof TiOrAh03, Nb20s-Si02 and Ti02-Si0211-'4 haveaddressed to the question of the relativedispersions of the two oxides as well as to thequestion of the chemical state of the supports.
The objective of the present investigationincludes the study of the mixed oxides incomparison to simple oxides by differentcharacterization techniques, especially X-ray PhotoElectron Spectroscopy (XPS). Platinum wassupported on titania (commercial), titania (gel) anda mixed oxide containing 22% titania and 78%alumina. The catalysts were characterized by X-raydiffraction (XRD), hydrogen chemisorptionmeasurements, sorptometry (BET surface area) andXPS techniques.
286 INDIAN 1. ENG. MATER. scr., OCTOBER 1998
Experimental ProcedureThe platinum supported on titania catalysts were
prepared by the conventional wet impregnationmethod using an aqueous solution of chloroplatinicacid. The loading of platinum was 5%(w/w) whichwas estimated quantitatively by spectrophotometricanalysis using the stannous chloride method'<. Acommercially available titania (Baker, UK, 99%pure anatase) designated as Ti02(C) and alaboratory prepared titania gel, prepared by thehydrolysis of titanium tetrachloride", indicated asTi02(G), were used as simple oxide supports. Amixed oxide support, TiOrAh03 indicated as T-Awas prepared for comparison.
The mixed oxide was prepared by the co-hydrolysis of titanium isopropoxide and aluminiumisopropoxide in isopropanol solution'", Themixture was stirred continuously for one hour atroom temperature in dry N2 to avoid contact withmoisture and prevent hydrolysis. The solution wasthen added dropwise to distilled water(totalvolume of the solution being 300 mL) to form theprecipitate which was then aged for one hour. Theprecipitate was filtered and washed 5 times withhot distilled water. The product was dried for 16 hat 383 K, then calcined in air at 573 K for 2 hfollowed by final calcination at 773 K for 2 h. Acomposition of 78% Ah03 and 22% Ti02 wasestimated quantitatively by gravimetric analysis ofaluminium by the oxine method". In all the cases,the catalysts were reduced in hydrogen flow afterpurification of the gas by copper and moisturetraps. The flow rate of hydrogen was 30 mL/min.The reduction was carried out at two differenttemperatures namely, 573 K for low temperaturereduction (LTR) and 773 K for high temperaturereduction (HTR). The reduction time wasmaintained at 24 h at the respective temperatures.The characterization of these samples were carriedout using these freshly reduced samples.
The X-ray diffractograms of all the freshlyreduced catalyst-s were recorded using a RigakuMiniflex diffractometer using Co-Kn radiation.
The measurements of BET surface area of thesamples were carried out by adsorbing nitrogen at77 K using a Carlo Erba Sorptometer(1800). Thesamples were degassed at 393 K prior to all theexperiments.
The room temperature hydrogen chemisorptionmeasurements were carried out in a conventional
all glass static vacuum system. The volume ofhydrogen adsorbed by the catalyst was determinedin the system after evacuating the catalyst for 6 h ata pressure of 10-6 Torr at 573 and 773 K (LTR andHTR) , respectively. It is assumed that each metalatom (Pt) adsorbs one hydrogen atom and hencethe ratio, HlPt, is taken to be one as reported in theliterature". The particle size was calculated usingthe expression ,ds(nm) = 108 / %D , where D is thedispersion given by20
D = Number of surface metal atomsTotal number of metal atoms
The X-ray photoelectron (XP) spectrum wasrecorded using ESCALAB MARK II of VGScientific (UK) .Limited. This instrument consistsof a twin anode (AlIMg anode). In this, Mg anodewas selected (1253.6 eV) because of low line-width [0.8 eV at full width half maximum(FWHM)]. Charging effects were corrected byadjusting the main CIs peak to 285.0 eV bindingenergy. The pressure in the analytical chamberduring measurements was nearly IxlO-9 Torr forrecording the XP spectra. The samples were takenin the form of pellets. 01s, CIs, Pt4f, Ti2p andA12s, Al2p regions were scanned.
Preliminary investigation of the activity of thesecatalysts were studied by the liquid phasehydrogenation of cinnamaldehyde. The hydroge-nation was carried out in liquid phase usingethanol as the solvent. The kinetics of the reactionwas monitored on a time-on stream basis for 5 h atintervals of 30 min in each case.
Results and DiscussionThe hydrogen chemisorption results have been
used for calculating the hydrogen adsorptioncapacity, metal particle size and dispersion and aregiven in Table 1. The BET surface area is alsoexhibited in Table 1.
It is observed from XRD that the titaniasupported samples exhibit the pure anatase form inall the cases. Thus, anatase to rutile transformationis not predominant even after high temperaturereduction (773 K) of the catalyst. This indicatesthat the phase transformation of the support neednot be the cause for SMSI state nor for themigration of the support on the metal particles. Ithas been postulated'" that the metal dispersion will
MALATHI et a/.: SURFACE ANALYSIS OF SUPPORTED PLATINUM CATALYSTS
be governed by the procedures employed for thepreparation of the support and dissolution of themetal precursors in the support. This latterphenomenon can be expected to be limited in thecase of mixed oxide support as compared to titaniasupport. The individual oxide phases could beidentified from the XRD of mixed oxide. Thesystems which resist dissolution of metalprecursors into the support phase should show
higher dispersion of the metal on the support ascompared to systems which are amenable for thedissolution of metal precursor into the supportphase, namely Ti02(G).
The hydrogen chemisorption measurementsshow that platinum supported on titania samplesexhibit the characteristic SMSI behaviour, 'i.e.,complete suppression of hydrogen uptake afterHTR. But the mixed oxide catalyst exhibits only a
287
Table I-Hydrogen chemisorption and surface area results of supported platinum samples
Catalyst Temperature of H2 uptake Dispersion BET surface areawt% reduction, K J.1 mol.lgcat % m2/g5% Pt / Ti02 (C) 573 4.4 4 13
773 0.0 105% Pt/ Ti02 (G) 573 7.5 6 100
773 M W5% Pt / T-A 573 5.3 4 211
773 3.9 3 200
'"T·A
HO.O 465.0Binding _.gy, eV
450.0
Fig.I-XP spectra of Ti2p regions for platinum supportedsamples reduced at (a) 573K and (b) 773 K.
"/T·A
~1161
.~.64.0 ".0
PUf
"IT·A
71.' 77.0 no8indi"ll._gy.eV
".0
Fig.2-XP spectra of Pt4 f regions for platinumsupported samples reduced at (a) 573K and (b) 773 K.
IU
".0
288 INDIAN J. ENG. MATER. SCI., OCTOBER 1998
moderate suppression even after HTR probablydue to the presence of y-Ah03 which has beenproved earlier to be a non-interacting support".Moreover, the pretreatment conditions used for thesupport seem to have little effect on chemisorptionbehaviour except for the fact that the dispersion ofplatinum is higher for the gel and mixed oxidesystems. This is due to the surface area of themixed oxide being higher than that of the gelwhich in turn is higher than that of the commercialcatalyst as shown in Table 1.
Figs 1 and 2 show the XPS patterns of the Pt4fand Ti2p levels, for all the catalysts reduced at 573K (LTR) and 773 K (HTR) respectively. Thesurface concentration was calculated using the
. 22equation
NI Al 0"2 ~K.EIK.E2 ... (1)
N2 A20"1
where NI and N2 are the surface concentrations ofelements I and 2 while Al and A2 are area underthe peaks of I and 2, (}I and (}2 are thephotoionisation cross-sections " of the levelsprobed for these elements and K.EI and K.E2 arethe kinetic energy values corresponding to thepeaks of the elements 1 and 2, respectively. Thekinetic energy was obtained from the conventionalequation,
... (2)
where E is the energy of the MgKa source.
Fritsch and Legare" have proposed, on the basisof their observations, a new feature in terms ofchanges in the valence band spectrum of aluminaon platinum depositions and its manifestation inthe OKVV Auger transition. According to theirproposal the interaction between the metal and thesubstrate involves the oxygen atoms. This proposalshould be examined in the context of the fact that,for SMSI state, there is considerable reduction ofthe support as shown" by the presence of Ti2
+ orTi3
+ species (when TiOz support has been used).Moreover, the reduced species is located in theoutermost region on the surface with the metalparticles". The point of contention here is, whetherthe rehybridisation of the metal orbitals and theoxide ion orbitals, as proposed by Fritsch and
Legare, is connected with such defect species orwhether it corresponds to the core hole generatedin the oxide ion of an idealized oxide matrix (asituation more favorable with Ah03 support).
Since it is difficult to find, at present, specificanswers to these questions, an attempt has beenmade to obtain partial answers from the XPS studyof the three catalyst systems. The XP spectra ofTi2p and Pt4f for these supported systems afterreduction at 573 and 773 K are given in Figs 1 and2. The relevant data extracted from these figuresare given in Table 2. The following observationscan be made from these data:
(i) The extent of reduction of the support is leastfor Ti02(C) while a considerable extent ofreduction is discernible in the case of Ti02(G) aswell as the mixed oxide support. The separation inbinding energy between 01 s level and the Ti2plevels has been deterrnined'" and the data are givenin Table 2. This difference in binding energyvalues should be about 71.00 eV for Ti02 while itis 73.4 eV for Ti20}, 75.0 eV for TiO and 77.0 eVfor metallic Ti from the literature values. The datagiven in Table 2 show that it is around 71.0 eV forPtffi02 systems reduced at 573 K while itincreased when these systems were reduced at773 K. The increase was more for Ti02(G) sampleshowing that reduction of Ti4+ is more marked onthis support. In the mixed oxide system, themagnitude of this separation cannot be used as thecriterion, since 01s level could have arisen fromthe oxide ions of alumina support as well.However, it is seen that this separation increaseswith increase in reduction temperature showingthat reduction of Ti4+could have taken place in thissample as well. In all the cases, the binding energyof Ti(2p) was found to be around 458 ± 0.5 eV, asshown in Table 228
.3°.
(ii) The extent of dispersion of the metaldecreased as a result of high temperature reductionfor the Ti02(G) and mixed oxide supportedsystems (a result deduced from the values of Pt/Tiratio given in Table 2).
(iii) Munuera et al." have devised a method fordeducing the encapsulation of the metal particlesby the reduced support. This is based on theobservation of increase in the values of h fhs;where h, is the height of the peak due to emissionfrom the metal core level and hb' that of thebackground at higher binding energy with respect
MALATHI et at. : SURFACE ANALYSIS OF SUPPORTED PLATINUM CATALYSTS 289
Table 2-XPS results of all the supported platinum samples
Catalyst Temperature Binding energy, eV [ FWHM ] Ratios Difference in
wt% of reduction O(ls) Pt(4j) AI(2s) Ti(2p) AI(2p) O/Ti Pt/Ti Ptl Al 01 AI B. E. BetweenK o and Ti
to that of the core level. The logic behind thisproposal is that the metal core level electrons arelikely to undergo inelastic collisions and, therefore,will contribute to the background tail below theelastic photoelectron peak. Accordingly, thesevalues have been evaluated and are given inTable 2; they show that encapsulation is moremarked in the case of Ti02(G) and mixed oxidesupports as compared to Ti02(C).
(iv) The FWHM of 0 Is peak for Ti02(G) isquite broad (2.1eV)32 thus showing that oxide ionsare in I1]ore than one environment. It is proposedthat the metal particles are incorporated into thereduced .support matrix and the oxygen orbitals arerehybridised so as to coordinate with platinumparticles as well as with reduced Ti species. Thissituation may also hold good in the case of mixedoxide support. However, this could not beunambiguously established since there are alreadymultiple types of oxide ions, namely, oxide ions ofalumina and titania in addition to oxide speciesthat coordinate to the metallic particles. This resultis also corroborated with the binding energy valueof Pt4f state which increases with increase inreduction temperature for TiOlG) as well asmixed oxide support. Such an increase is notobserved for Ti02(C) showing incorporation andinteraction of platinum orbitals with oxide ionorbitals are the least in this case.
The ~B .E. values for the a and Ti indicate thepresence of partially reduced support species(namely Te+) near the surface of the catalyst. Thisresult can be explained on the basis that the Ti3+
Table 3-Percentage conversions on titania and mixed oxidesupported platinum catalysts for cinnamaldehydehydrogenation.
The values in parentheses indicate the % selectivity towardscinnamyl alcohol.
near the platinum surface sites will act as a sourceof electrons which could affect the catalyticactivity and selectivity for the hydrogenation ofa,~-unsaturated aldehydes. Our preliminaryinvestigations.. the results of which are given inTable 3, corroborate the earlier conclusion derivedbased on the M.E. values.
It is observed from Table 3 that the selectivitytowards cinnamyl alcohol is exhibited only in thecase of HTR catalysts or in SMSI state 'and it isgreater for platinum on titania(gel) sample than onTi02(C).
ConclusionsThe results indicate a more pronounced SMSI
condition for Ti02(G) sample than for the mixedoxide system. From the ratio of the intensity of themetal peak to that of the background in XPS, itmay be concluded that in the case of Ti02 supports
290 INDIAN J. ENG. MATER. SCI., OCTOBER 1998
the encapsulation of the metal by the support is thecause of reduction in the uptake of hydrogen afterHTR treatment. However, such an unambiguousstatement is not possible for TiOrAh03 systemsfrom XPS study in view of multiplicity ofinteractions with different metal species. The highselectivity for cinnamyl alcohol in the case ofPtffi02 shows that SMSI state is prevalent in thissystem.
References1 Sinfelt J H , Taylor W F & Yates D J C, J Phys Chern, 69
(1965) 3857.2 Schwab G M, Adv Catal, 27 (1978) 1.3 Solymosi F & Szabo Z G, Actes Congr Int CataI 2nd. 2
(1961) 1627.4 Solymosi F, Catal Rev, 1 (1967) 233.5 Tauster S J & Fung S C, J Catal, 55 (1978) 29.6 Tauster S J, Fung S C & Garten R L, J Arn Chern Soc,
100 (1978) 170.7 Madhusudhan Rao P, Sudharshan V & Viswanath R P, in
Catalysis: Present and future-International Series onChernical Engineering. edited by Kanta Rao P andBeniwal R S (Wiley Eastern and Publications andInformation Directorate, New Delhi). 1995,218-224.
8 Madhusudhan Rao P, Viswanathan B & Viswanath R p, JMater Sci, 30 (1995) 4890.
9 Madhusudhan Rao P, Metal-support Interactions inFeffi02 and PtlTi02 Systems- Characterization andConsequences, Ph.D Thesis. Indian Institute ofTechnology, Madras, 1995.
10 McVicker G B, & Ziemiack J J, J Catal, 95 (1985) 473.11 Stranick M A, Houalla M & Hercules D M, J Catal, 106
(1987) 362.
12 Ko E L, Bafrali R, Nuhfer N T & Wagner N J, J Catal, 95(1985) 260.
13 Murrell L L. & Yates D J C, Stud Surf Sci Catal, 7 (1981)J470.
14 Ko E L & Wagner N J, J Chern Soc Chern Commun, 19(1984) 1274.
15 Milner 0 1& Shipman G F, Anal Chern. 27 (1955) 1477.16 Sankar G, Kannan K R & Rao C N R, Catal Leu, 8
(1991) 27.17 Lahousse C. Aboulyt A. Mauge F. Bachelier J & Lavalley
J C. J Mol Catal, 84 (1993) 283.18 Claassen A & Bastings L. Analyst. 92 (1967) 614.19 Bogdanchikova N E. Buyanova N E. Zaikonskii V L.
Zapreeva 0 F & Kamaukhov A p. Kinet Katal. 30 (1989)999.
20 Fung S C. J Catal, 76 (1982) 225.21 Anderson S L T & Scurrell M S. J Catal, 71 (1981) 233.22 Viswanathan B. Mary U D & Viswanath R p. Indian J
Eng Mater Sci; 1 (1994) 335.23 Conesa J C & Soria J. J Phys Chern, 86 (1982) 1392.24 Fritsch A & Legare p. Surf Sci, 184 (1987) L355.25 Tang S. Xiong G & Wang H. J Catal, III (1988) 136.26 Tamura K. Bardi U & Nihei Y. Surf Sci, 197 (1988)
L281.27 Sexton B A. Hughes A E & Foger K. J Catal. 77 (1982)
85.28 Gopel W. Anderson J A. Frankel D. Jaehnig M. Phillips
K, Schafer J A & Rocker G, Surf Sci, 139 (1984) 333.29 Idriss H, Pierce K G & Barteau M A. J Arn Chern Soc.
116 (1994) 3063.30 Raupp G B & Dumesic J A, J Phys Chern, 89 (!985)
5240.31 Munuera G, Gonzalez-Elipe A R & Espinos J P, Surf Sci,
2111212 (1989) 1113.32 Sayers C N & Armstrong N R, Surf sa, 77 (1978) 301.
