J . Chem. SOC., Faraday Trans. 1, 1982, 78, 3537-3544 Thermal Analysis of the Decomposition Mechanism of Platinum and Palladium Tetrammine Faujasite X BY DIETER EXNER, NILS JAEGER, KARIN MOLLER AND GUNTER SCHULZ-EKLOFF* Applied Catalysis Research Group, Department of Chemistry, University of Bremen, D-2800 Bremen 33, Federal Republic of Germany Received 4th March, 1982 The decomposition of platinum and palladium tetrammine faujasite X has been studied by temperature- programmed desorption (t .p.d.) spectroscopy, differential thermal analysis and derivative thermogravimetry . Well resolved t.p.d. spectra allowed the deduction of a detailed mechanism with respect to autoreduction, ammonia evolution and the reaction of ammonia with acid sites. Characteristic differences observed between the palladium and platinum samples are related to the lower stability of the palladium tetrammine complexes within the faujasite matrix.Noblz-metal-loaded faujasite catalysts find wide applications in petrochemical processes.’? Since the activity depends on the metal dispersion, the factors influencing the dispersion process have been investigated in several studies of noble-metal-loaded fa~jasites.~-ll In a recent study Reagan et a1.12 discussed the decomposition of platinum and palladium ammine complexes exchanged into zeolite Y, and the catalytic properties of the resulting reduced platinum and palladium zeolites. The present study applies temperat ure-programmed desorp tion (t . p. d .) spectroscopy, differential thermal analysis and derivative thermogravimetry, and focusses on the evaluation of the decomposition mechanism in a vacuum, or in argon or oxygen.The use of zeolite X was of interest because of previous observations that metal agglomerates far exceeding supercage dimensions can be grown and stabilised within the matrix under special experimental conditions. EXPERIMENTAL SAMPLES Zeolite NaX (Na~6~(A~02)~6(slo2)l~6} . nH20) was prepared by hydrothermal cry~tallisation.~~~ l4 The composition (Si/Al ratio) was determined by chemical analysis and X- ray fluorescence spectroscopy. The preparation of iron-free samples could be checked by flame atomic absorption spectroscopy and by electron spin resonance spectroscopy. Iron is claimed to influence the platinum dispersion in faujasites.15*16 Ion exchange was carried out with solutions of 0.01-0.05 mol dm-3 Pt (NH,),Cl, and Pd(NHJ4C12 at room temperature.The loaded zeolites were washed until chloride-free. The crystallinity of the PtNaX and PdNaX zeolites was largely maintained, as has been confirmed by X-ray diffraction, electron microscopy and a dynamic method” involving the nitrogen physisorption capacity being normalised to the standard NaX. The samples PtX 13, PtX 25, PtX 42, PtX 52, PdX 7, PdX 13, PdX 25, PdX 43, where the numbers denote the degree of exchange, have been used in the following investigations. 35373538 DECOMPOSITION OF FAUJASITE x METHODS TEMPERATURE-PROGRAMMED DESORPTION The decomposition of platinum and palladium ammine complexes in faujasite X was studied in a vacuum (pumping time constant < min; residual gas pressure < lo-' mbar) on the one hand, and in the presence of additional media (argon or oxygen, 1 bar) on the other.A linear heating rate of 5OCmin-l was used in all cases. The ion currents of the evolved compounds were recorded by a computer (PDP 11) coupled to a quadrupole mass spectrometer (Q 200, Leybold-Heraeus, Koln). The relevant mass-spectrum region could be recorded in steps of 5 O C . The desorption rates of ammonia and nitrogen are given in the figures below. For studies in additional media (argon or oxygen) a thin layer of the sample (50 mg cm-2) was deposited onto a glass frit in an external reactor, and the dried-gas medium was streamed with a flow rate of 70 cm3 min-' during both the pretreatment (room temperature, 12 h) and the temperature programmes.To record the compounds evolved during the ammine complex decompositiy process, the exit of the reactor was connected to the quadrupole mass spectrometer recipient uia a leak valve. DIFFERENTIAL THERMAL ANALYSIS AND DERIVATIVE THERMOGRAVIMETRY The measurements were carried out with a thermobalance (L 81, Linseis, Selb), using tempered (1 100 "C) alumina (Merck, Darmstadt) as a reference in flowing dried-gas media during the pretreatments (room temperature, 12 h) and the decomposition studies. Flow rates of 70 cm3 min-' and heating rates of 5 O C min-' were used. Sample charges of 40 mg cm-2 in Pt crucibles were applied. Additional studies of the decomposition of platinum and palladium tetrammine chlorides confirmed the results reported by Wendlandt et al.l*, l9 RESULTS DECOMPOSITION IN A VACUUM Fig.1 shows rates of desorption of ammonia and nitrogen evolved in the temperature-programmed decomposition of platinum and palladium ammine com- plexes in faujasite X which have been degassed and dehydrated in a vacuum at room temperature overnight. The decomposition of platinum tetrammine faujasite X starts at ca. 50 "C; it reaches the first significant maximum in the rate of ammonia desorption at ca. 200 "C and the last maximum between 300 and 350 "C. The decomposition of the corresponding palladium compound, measured by ammonia desorption, starts immediately above room temperature and shows its first maximum at ca. 75 O C . The ammonia desorption is largely complete by ca. 300 O C .The following qualitative features can be drawn from the t.p.d. spectra. (1) The pattern of the spectrum depends on the degree of exchange. (2) Ammonia and nitrogen are liberated in several steps. (3) The ammine complexes of palladium are less stable than those of platinum. (4) In the case of palladium complexes ammonia desorption is never accompanied by nitrogen desorption. The following values for the ratio NH,/N, have been estimated for the totally evolved amounts of the two gases: 5 (PtX 13, vacuum), 6 (PtX 25, vacuum), 8 (PtX 42, vacuum), 9 (PtX 52, vacuum) and 6 (PtX 52, Ar). For all palladium samples ratios of ca. 6 were found. Overall H,/N, ratios of ca. 1 could be determined for all samples. The platinum samples PtX 42 and PtX 52 exhibit maxima of hydrogen evolution shifted to higher temperatures in comparison with those of nitrogen evolution, indicating higher adsorption energies for H, than for N, at the reduced platinum.Activation energies for ammonia evolution have been estimated using procedures given by McCarty and Madix20 and Chan et aL21 and were 144 kJ mol-1 (PtX 13, 240 "C; PtX 25,240 "C), 108 kJ mol-l (PtX 42,310 "C), 88 kJ mol-1 (PtX 42,210 "C), 80 kJ mol-1 (PtX 52, Ar, 310 "C) and 52 kJ mol-1 (PdX 43, 100 "C).3539 D. EXNER, N. JAEGER, K. MOLLER AND G. SCHULZ-EKLOFF d (a 1 PtX 13 PtX 25 I I 1 I I I b o 100 200 300 400 500 r 1 I 1 I I b 0 100 200 300 400 500 T/"C 0 x Y .- E +.I C .d A ( a ) I PtX 42 \ PdX 4 3 0 r 1bO 200 I 300 1 400 I 500 I # T/"C FIG. 1 .-Rates of evolution (arbitrary mass-spectral intensities) of ammonia (-) and nitrogen (---) in the temperature-programmed decomposition ( 5 K min-I), of (a) platinum and (6) palladium tetrammine faujasite X in a vacuum.3 540 DECOMPOSITION OF FAUJASITE x The rates of desorption of water are more than two orders of magnitude lower than those of ammonia up to 450 "C.Above 500 "C a marked evolution of water from the partial destruction of the protonated zeolite lattice is observed. The maximum rate of water desorption is shifted to lower temperatures with increasing degree of exchange. DECOMPOSITION IN ARGON Desorption spectra for temperature-programmed decomposition in streaming argon are given in fig. 2. Ammonia is desorbed at higher temperatures than in a >r E c 4- .- Y .- I \ x +- 0 c W c .I Y .- I I I I I I b r 1 I I I I + 0 100 200 300 400 500 0 100 200 300 400 500 FIG.2.-Rates of evolution (arbitrary mass-spectral intensities) of ammonia (-1 and nitrogen (----), d.t.a. curve, and rate of weight loss in the temperature-programmed decomposition ( 5 K min-') of (u) platinum tetrammine faujasite X (PtX 52) and (h) palladium tetrammine faujasite X (PdX 43) in argon. TPC T/"C vacuum, with a maximum rate at ca. 300 O C , corresponding to the results of Reagan et a1.I2 The maximum rate of ammonia desorption from the decomposing palladium complex is accompanied by the maximum rate of nitrogen desorption, unlike the vacuum case. The maximum rate of ammonia desorption from the platinum complex at ca. 300 "C corresponds to an endothermic peak in the differential thermal analysis [fig.2(a)] and to a maximum rate of weight loss in the derivative thermogravimetry [fig. 2(a)]. A broad endothermic peak at ca. 250 "C in the differential thermal analysis has no corresponding maximum rate of weight loss in the derivative thermogravimetry. No marked corresponding effects in differential thermal analysis and derivative thermogravimetry are observed in the decomposition of the palladium complex [fig. 2(b)], because of the broader temperature region of ammonia evolution. Water is only partially removed during the pretreatment at room temperature with the dried medium prior to thermal analysis. This was indicated by rates of desorptionD. EXNER, N. JAEGER, K. MOLLER AND G. SCHULZ-EKLOFF 3541 of water ranging in the same order of magnitude as the desorption rates of ammonia.Large amounts of water could be removed if the pretreatment temperature was raised to 150 O C . However, the spectral patterns were not changed by this thorough dehydration, indicating that water has no additional effect in the presence of a medium. However, the d.t.a. and d.t.g. spectra suffer from the fact that the ammonia desorption is overlapped by the water desorption. They are nevertheless helpful for supporting conclusions which will be mainly drawn from the t.p.d. spectra. T I \ I \ ' \ ! \ r I 1 I I I b 0 100 200 300 400 500 FIG. 3.-Rates of evolution (arbitrary mass-spectral intensities) of ammonia (-) and nitrc,:en (---), d.t.a. curve, and rate of weight loss in the temperature-programmed decomposition ( 5 K min - I ) of platinum tetrammine faujasite X (PtX 52) in oxygen.Tf"C DECOMPOSITION IN OXYGEN The use of oxygen instead of argon as a medium in the decomposition of the platinum ammine complex strongly increases the N,/NH, ratio evolved at ca. 300 "C with a maximum rate (fig. 3). A simultaneous, strong desorption of water is observed. Derivative thermogravimetry gives a corresponding maximum rate of weight loss (fig. 3), and differential thermal analysis exhibits a strong exothermic peak at 300 O C (fig. 3). DISCUSSION DECOMPOSITION IN A VACUUM The resolution of the desorption spectra obtained for the temperature-programmed decomposition of dehydrated platinum and palladium tetrammine faujasite X samples in a vacuum can be interpreted by the following reactions.3 542 DECOMPOSITION OF FAUJASITE x Pt SAMPLES Below 2OOOC the evolution of ammonia is due to the decomposition of the tetrammine complex [Pt (NH3),l2+ -+ [Pt(NH,),]2i + (4 - x)NH,.(1) Above 200 O C further decomposition of the platinum ammine complex includes the autoreduction of Pt2+ ions [Pt (NH3),]2i + Pto - (NH) + 2Hi + (X - l)NH, 2PtoNH + 2Pt0 + N, + H,. (2 a) (2 b) Reactions (2 a) and (2 b) occur simultaneously. reduction process (2 a). They react with additionally liberated ammonia At this point Brsnsted-acid sites ZOH (Z:Zeolite) are formed during the auto- Z-OH + NH, + Z-O-NH;. (3) Since overall ratios H,/N, z 1 could be estimated, and ratios NHJN, z 6 were found, the proposed autoreduction mechanism [reaction (2 a)] can be supported, in the course of which a complete reduction of the metal ions should have taken place.Ratios NHJN, > 6 are indicative of incomplete reduction. This is observed for the samples PtX 42 and PtX 52, which contain more than 4 platinum tetrammine complexes per supercage. It might be postulated that up to 4 platinum tetrammine complex ions can be located in the tetrahedral supercages by forming partial bonds with the zeolite lattice and which can be decomposed more easily under autoreduction. In these cases ammonia evolution should always be accompanied by nitrogen evolution, as is observed for PtX 13 and PtX 25. If the supercage is loaded with more than 4 platinum tetrammine complex ions, then some of the complex ions might be able to liberate ammonia ligands without simultaneous autoreduction, leading to an incomplete overall reduction and to ammonia evolution not accompanied by the simultaneous evolution of nitrogen (PtX 42 and PtX 52).The reason for the easier autoreduction of those platinum complexes which are partially coordinated to the zeolite lattice might be found in the relatively low bond symmetry of this coordination, which will be more favourable to the splitting of N-H bonds which is necessary in the autoreduction process. In the temperature range between 300 and 400 O C the liberation of ammonia ions from the acid sites can be observed: Z-O-NH,+ -+ ZOH + NH,. (4) Above 450 O C nitrogen and hydrogen are evolved from the decomposition of NH, desorbed from sites of high acidity. Pd SAMPLES Reactions (1) and (2a) can also be postulated for the decomposition of the palladium tetrammine faujasite X.Reaction (2 b) is shifted to temperatures above 400 O C . This hypothesis includes the assumption of palladium imide species being more stable than platinum imide species. Such species are found to be relatively stable surface compounds in the catalytic decomposition of ammonia on 23 and palladium.24 An insufficient amount of hydrogen is evolved to support an alternative assumption of NH, decomposition evolved from acid sites. In the case of palladium no significant desorption of NH, from Brsnsted-acid sites is observed. This could be due to the fact that most of the ammonia has left the zeoliteD. EXNER, N. JAEGER, K. MOLLER AND G. SCHULZ-EKLOFF 3 543 structure at rather low temperatures and is not available to neutralize acid sites formed in the autoreduction step.A comparison between platinum and palladium tetrammine faujasite X on the one hand (fig. l), and platinum and palladium chloridel8V l9 on the other, with respect to the temperature at which ammonia evolution due to the decomposition process begins, leads to the conclusion that the cation complexes are less stable in the faujasite matrix than in the chloride compound. The more pronounced destabilisation of the palladium cation complex could be due to the higher tendency of palladium, as compared with platinum, to form covalent bonds with oxygen, which can be deduced from the relatively weak shift of the Pd(3d5,,)X.p.s. signal in Pd025 as compared with Pd, and the relative strong shift of the Pt(4f,,,)X.p.s.signal in Pt026 as compared with Pt. Thus the expected stronger covalent bond between palladium and the faujasite lattice oxygen should facilitate the liberation of the ammonia ligands. The estimated activation energies for the evolution of ammonia from the decom- posing complexes (52-144 kJ mol-l) are within the range of values found for the desorption of ammonia from p l a t i n ~ m ~ ~ ~ 27 on one hand, and from Brlansted-acid sites28 on the other. DECOMPOSITION I N ARGON Less resolved temperature-programmed desorption spectra are obtained if the decomposition is carried out in argon (fig. 2). In the competitive reaction paths of ammonia evolution and reaction with acid sites, the former will be favoured in a vacuum and the latter should be favoured in the presence of a medium, owing to the slower diffusion of the ammonia molecules out of the zeolite framework and, consequently, the increased probability of their reaction with an acid site.The endothermic peak aroung 250 O C in d.t.a. indicates the liberation of ammonia from the platinum at the same temperature as in a vacuum, while no increased rate of weight loss can be observed in d.t.g. For palladium samples the simultaneous desorption of NH, and N, can now be observed due to the fact that NH, is available at higher temperatures as compared with the vacuum process and can now be catalytically decomposed. DECOMPOSITION I N OXYGEN A strong exothermic peak is found in the differential thermal analysis of the platinum tetrammine decomposition (fig.3) at a temperature corresponding to the maximum rate of ammonia desorption on one hand, and the maximum rate of catalytic ammonia decomposition on the other; the latter can be deduced from the maximum rate of nitrogen evolution. Since the catalytic decomposition of ammonia produces hydrogen, the exothermic peak should be due to the formation of water from this hydrogen and the medium oxygen. This is supported by a simultaneous maximum for the desorption of water observed in the t.p.d. spectrum. The heat produced in the water-formation reaction should be responsible for the enhanced ammonia decom- position, which can be seen from the large N2/NH, ratio. There was no indication in the mass spectra of the formation of nitrogen oxides. CONCLUSIONS The study of the well resolved t.p.d.spectra of the decomposition of platinum and palladium tetrammine complexes within a faujasite X matrix allows the deduction of a detailed mechanism with respect to autoreduction, ammonia evolution and reaction of ammonia with acid sites.3 544 DECOMPOSITION OF FAUJASITE x Characteristic differences observed between the palladium and platinum samples may be related to the lower stability of [Pd(NH3)J2+ complexes within the faujasite matrix. Financial support by the Senator fur Wissenschaft und Kunst der Freien Hansestadt Bremen is acknowledged. A. P. Bolton, Am. Chem. SOC. Monogr., 1976, 714. E. Gallei, Chem. Ing. Tech., 1980, 52, 99. P. Gallezot, in Catalysis by Zeolites, ed. B. Imelik et al. (Elsevier, Amsterdam, 1980), p.227. J. A. Rabo, V. Schomaker and P. E. Pickert, Proc. 3rdInt. Congr. Catal., (North Holland, Amsterdam, 1965), vol. 2, p. 1264. J. A. Rabo, P. E. Pickert and R. L. Mays, Znd. Eng. Chem., 1961, 53, 733. vol. 2, p. 1329. E. Czaran, K-H. Schnabel and M. Selenina, 2. Anorg. Allg. Chem., 1974, 410, 225. J. C. Vedrine, M. Dufaux, C. Naccache and B. Imelik, J. Chem. Soc., Faraday Trans. I , 1978,74,440. C . Naccache, N. Kaufherr, M. Dufaux, J. Bandiera and B. Imelik, Am. Chem. SOC. Symp. Ser., 1977, 40, 538. lo P. Gallezot, A. Alarcon-Diaz, J. A. Dalmon, A. J. Renouprez and B. Imelik, J. Catal., 1975,39, 334. l1 (a) D. Exner, N. Jaeger and G. Schulz-Ekloff, Chem. Zng. Techn., 1980, 52, 734; (b) D. Exner, N. Jaeger, R. Nowak, H. Schriibbers and G. Schulz-Ekloff, J. Catal., 1982, 74, 188; (c) G. Schulz- Ekloff, D. Wright and M. Grunze, Zeolites, 1982, 2, 70. ti R. A. Dalla Betta and M. Boudart, Proc. 5th Int. Congr. Catal. (North Holland, Amsterdam 1973), l2 W. J. Reagan, A. W. Chester and G. T. Kerr, J. Catal., 1981, 69, 89. l3 D. W. Breck and E. M. Flanigan, Molecular Sieves (Society for Chemical Industry, London, 1968), l4 H. Kacirek and H. Lechert, J. Phys. Chem., 1975, 79, 1589. l5 D. Geschke, H. Winkler and D. Wendt, Z. Phys. Chem. (Leipzig), 1973, 252, 235. l6 J. B. Uytterhoeven, Acta Phys. Chem., Szeged, 1978, 24, 53. l7 S. Briese-Gulban, H. Kompa, H. Schrubbers and G. Schulz-Ekloff, React. Kinet. Catal. Lett., in press. W. W. Wendlandt and J. P. Smith, The Thermal Properties of Transition-metal Ammine Complexes (Elsevier, Amsterdam, 1967), p. 179. W. W. Wendlandt and L. A. Funes, J. Inorg. Nucl. Chem., 1964, 26, 1879. p. 47. 2o J. G. McCarty and R. J. Madix, Surf. Sci., 1976, 54, 121. 21 C. N. Chan, R. Aris and W. H. Weinberg, Appl. Surf. Sci., 1978, 1, 360. 22 C. E. Melton and P. H. Emmett, J. Phys. Chem., 1964, 68, 3318. 23 D. G. Loffler and L. D. Schmidt, J. Catal., 1976, 41, 440. 24 K. Kunimori, T. Kawai, T. Kondow, T. Onishi and K. Tamaru, Surf. Sci., 1976, 59, 302. 25 K. S. Kim, A. F. Gossmann and N. Winograd, Anal. Chem., 1974, 46, 197. 26 K. S. Kim, N. Winograd and R. E. Davis, J. Am. Chem. Soc., 1971,93, 6296. 27 J. L. Gland, Surf. Sci., 1978, 71, 327. N. Y. Topsse, K. Peddersen and G. Derouane, J. Catal., 1981, 70, 41. (PAPER 2/384)