J. Chem. SOC., Faraday Trans. 1, 1982,18, 53-60 Microcatalytic Study of the Depolymerization of 2,4,6-Trimethyl- 1,3,5-trioxan (Paraldehyde) over Mordenite Surfaces BY PAUL JOE CHONG AND GEOFFREY CURTHOYS* Department of Chemistry, University of Newcastle, New South Wales, Australia 2308 Received 8th October, 1980 The catalytic depolymerization of paraldehyde (2,4,6-trimethyl- 1,3,5-trioxan) to acetaldehyde over Na+-, Ca2+- and La3+-exchanged mordenites has been investigated by means of conventional pulse-flow microcatalytic chromatography. The injection port of a gas chromatograph was used as a microcatalytic reactor chamber. The apparent activation energy was found to be 37.0 kJ mol-' for NaM and 16.0 kJ mol-l for CaM and LaM. The mordenite-catalysed depolymerization of paraldehyde fits the Langmuir- Hinshelwood kinetic model, which can be described by the following pathway (see Discussion for symbols) : k+i k+, k+3 (PA),+E + [PA...El* + [AA.. -El* + (AA)g+X k-1 slow fast fast Microcatalytic chromatography was first developed by Kokes et al.' and has been modified since by several investigator^.^.^ In the usual operation of this method, a pulse of reactant is injected into a stream of a carrier gas, passed through a microreactor containing catalyst and then through an analytical column. Bassett and Habgood5 proposed gas-solid microcatalytic chromatography for the kinetic study of first-order surface-catalysed reactions. They used the catalyst for dual functioning, both for the catalytic reaction and for the chromatographic separation. The pulsed version of this method has been theoretically analysed by Roginskii and RozentaP and several Ettre and Brenner8 proposed the attachment of a precolumn microreactor for the pulsed-flow method.Recently Schmiegel et aL9 have reported the use of a gas-chromatography injection port for the microcatalytic determination of the activation energy and Fowler et al.l0 for the quantitative recovery of adsorbed species such as pollutant molecules. In the present work the catalytic properties of the mono(Na+)-, di(Ca2+)- and tri( La3+)-valent cation exchanged mordeni tes have been investigated and the apparent activation energies (E,) were l 1 Since many surface-catalysed reactions are accompanied by side reactions, which complicate the kinetic analysis, depoly- merization of paraldehyde was chosen as a probe reaction.Paraldehyde undergoes clean conversion into a single product species, acetaldehyde. Langer et a1.12 stated that depolymerization of paraldehyde is suitable for testing acidic properties of solid surfaces. EXPERIMENTAL AND RESULTS AnalaR grade paraldehyde or 2,4,6-trimethyl- 1,3,5-trioxan, (CH,CHO),, ex. May and Baker, was used as reactant, the purity being confirmed chromatographically. 5354 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE Synthetic NaM (M = Mordenite), ex. Norton, was used as parent adsorbent. CaM and LaM were prepared from NaM by ion-exchange, using conventional exchange procedures.13 From the elemental analysis of the zeolites1* the unit-cell compositions were determined, and are shown in table 1.For use in the catalytic experiments, the cation-exchanged mordenite powders were compressed into pellets (35-45 mesh). An aliquot (10-12 mg) of the resulting catalyst was loaded into the microcatalytic reactor, and activated initially at 350 O C for 16 h in vacuo and then finally at 500 OC for 0.5 h in an oxygen atmosphere. The amount of catalyst chosen was small enough to prevent prolonged surface holding of reactant-product species, which otherwise causes the elution peaks to be diffuse and broadened? For evaluation of catalytic activity at different temperatures, conventional pulsed-flow microcatalytic chromatography was used with an injection port as a microcatalytic chamber as shown in fig. 1 . 5 t 9* l1 A Packard model 427 gas chromatograph having a Gow-Mac katharometer was used TABLE 1 .-UNIT-CELL COMPOSITIONS OF MORDENITE CATALYSTS13* l4 cation type structural composition ~ ~~~ % cation exchange (from) NaM (Na20),(A1203)4(Si02)4~.3023.2H20 100 (1 mol dm-3 NaC1) CaM (Ca0)3.4dNa2O)O. 54(A1203)4(Si02)41. 3023-2H20 86.6 (0.02 mol dm-3 CaC1,) LaM (La203)1.05(Na20)0.85(A1203~4~si02~41,3023~2H20 78.4 (0.1 mol dmP3 LaCl,) with an analytical column (length 196 cm; i.d. 0.23 cm; 0.d. 0.63 cm) of Carbowax 20 M. At a column temperature of 110 OC and flow rate of Fo = 12.0 cm3 min-l there was no measurable decomposition of paraldehyde during elution. The catalytic effect of the reactor wall was found to be negligible by comparison with blank runs, which were carried out using glass beads. The apparent activation energy, E,, was determined using the following equation :5-7 where Fo is the corrected flow rate of He carrier gas at 273 K (cm min-l), m is the mass of catalyst (g), Xis the fraction of reactant converted, Q is the fraction of reactant remaining unchanged, A' is a pseudo-frequency factor (constant), which includes the Arrhenius pre-exponential factor ( A ) , catalyst void volume ( v ) and integral constant of the van't Hoff equation ( I ) , and the other symbols have their usual meaning. The value of E, can be determined by the measurement of X / Q as a function of the reaction temperature.Fig. 2 shows typical chromatograms of the decomposition of paraldehyde on NaM over the temperature range 413-533 K. Fig. 3 shows the plots of The left-hand side term of eqn (1) is a measure of catalyticP. J .CHONG AND G . CURTHOYS 55 FIG. 1 .-Schematic diagram of pre-column microcatalytic reactor: A, A1 injection port; B, silicone septum (high-temperature durable); C, s/s ring (grooved); D, thermocouple; E, glass-wool plug (preheating zone); F, cartridge heater; G, microcatalytic reactor (etched exterior wall); H, catalyst bed; I, glass-wool plug; J, asbestos insulation; K, reactor-level adjuster; L, ferrule; M, Swagelok; N, g.c. analytical column; P, flow regulator (micro); Q, carrier gas (He); R, wall of g.c. oven cabinet; S, g.c. thermostat bath. In [(F*/m) In (1 + X / Q ) ] against 1 / 7’ for NaM, CaM and LaM. From the slopes and the Y-intercepts the values of Ea and A’ were obtained, as shown in table 2. DISCUSSION First-order surface-catalysed reactions can be investigated quantitatively by pulsed- flow microcatalytic chromatography, provided that the fractional conversion is independent of the input pulse or partial pressure of The chromatograms obtained in the manner as described are quite symmetrical, their integrated peak areas varying with the depolymerization temperature, which permits a quantitative evaluation of the catalytic effect by analysis of the resulting elution peaks.4* l 1 9 l5 As shown in fig.3, the Arrhenius plots were found to be linear over the range of temperature investigated. The Ea value derived for NaM was much higher than those obtained for CaM and LaM, the latter two cases being identical. The values of A’ varied between the mordenite samples but no simple correlation was found.In general,56 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE I II 90 60 30 0 90 60 30 0 retention time/s FIG. 2.-Typical chromatograms of catalytic decomposition of paraldehyde over sodium mordenite surface (10.0 mg) at (A) 240 and (B) 150 O C : (a) paraldehyde, (b) acetaldehyde and (c) air. FO(He) = 12.0 cm3 min-', sample size = 0.5 mm3, chart speed = 10 cm min-', TCD = 350 OC, column temp. = 110 OC. the true activation energy for surface reactions following first-order kinetics can be found from the measurement of retention v01umes.~~ l6 Since a relatively small packing of the catalysts was required for developing satisfactory chromatograms, the present method is not amenable to the determination of the true activation energy.NaM is known to be catalytically inactive for acid-catalysed reactions, while CaM and LaM are active due to the greater polarizing power of the respective exchanged cations.17* l8 The evidence indicates that the depolymerization of paraldehyde is an acid-catalysed reaction, the magnitude of E, being related to the strength of the surface acidity. As the role of a catalyst is to lower E,, it is clear that the polyvalent cation exchanged mordenites are catalytically more effective than the monovalent cation exchanged mordenites.16t In studying the catalytic depolymerization of paraldehyde in static media, Walvekar and Halgerilg stated that only silica/alumina surfaces having acid strength of pK, < -3.0 are catalytically effective and found that the catalytic activity of silica-based oxides was much stronger than that of alumina-based oxides.In this respect the highly siliceous nature of mordenites would be a contributing factor in the decomposition of paraldehyde. For a kinetic analysis in terms of classical models,20-22 the surface-catalysed depolymerization of paraldehyde is assumed to occur in the sequence adsorption- surface-reaction-desorption, tacitly ignoring the influence of diffusion to and from the surface. According to the absolute rate theory the behaviour of adsorbate molecules may be defined either by the Rideal mechanism, which refers to the surface interaction between the adsorbed and the unadsorbed species, or by the Langmuir- Hinshelwood model, which refers to the catalytic interaction proceeding only between adsorbed species.Both hypotheses may be operative for the depolymerization of paraldehyde in static media, but their validity in zeolite cavities is subjected to certain geometric restrictions.17* 239.0 - h M % + 8 . 0 1 - v E: - h E 1 5 Y E: - 7.0 6.C P. J. CHONG AND G. CURTHOYS 1 I 1 I I I a 57 2.00 2.20 2.40 2.60 2.80 3.00 lo3 KIT FIG. 3.-Plots of In [(Fo/rn) In (1 +X/Q)] against reciprocal of absolute temperature: 0, LaM; 0, CaM and (>, NaM. TABLE 2.-KINETIC PARAMETERS DETERMINED FOR THE DEPOLYMERIZATION OF PARALDEHYDE OVER DIFFERENT MORDENITE CATALYSTS cation apparent activation energy, pseudo-frequency type EJkJ mol-' factor, A' NaM CaM LaM 37.0 16.0 16.0 3 . 5 ~ 107 3 . 2 ~ 105 4.5 x 105 Mordenite has main channels parallel to the c-axis and side-pockets to the a-axis (fig.4).17 The main channels consist of elliptically distorted 12-membered rings, having a free aperture of 0.581 nm x 0.695 nm. They may admit paraldehyde (kinetic diameter 0.8 nm)24 or even larger molecules, owing to framework distortion and/or molecular def0rmabi1ity.l~ However, the main channel cannot pass more than one paraldehyde molecule at a time,23 although repeated interactions with the surface sites along the channel walls would be quite possible. Topologically the free rotation of a paraldehyde molecule is forbidden in the mordenite cavities. The side-pocket (0.387 nm x 0.472 nm) cannot accommodate paraldehyde 23 In view of the above facts, the Langmuir-Hinshelwood kinetic model is regarded as the preferred reaction 2 5 7 26 The depolymerization of paraldehyde over mordenite surfaces proceeds with the formation of carbocations as surface intermediate^.^^^ 27 3 FAR 158 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE [3 &m coplanar faces:.,carbon: 0 , h y d r o g e n : O oxygen.FIG. 4.-Framework unit cell structure of mordenite (cross-sectional planar view from c-axis; schematic encasing of paraldehyde molecule in main pore shown, not to scale).l79 24 Units in A. Undoubtedly the surface-catalysed depolymerization is initiated through the oxygen atom of paraldehyde molecule, possibly via H-bonding with surface acid sites. As a result the C-0 bond in the heterocyclic linkage is ruptured, creating a carbocation. Through subsequent rearrangement of the vicinal-bond electrons a series of chain scissions occurs, each with formation of monomeric fragments. A hydride ion is abstracted from the terminal monomeric unit and transferred to the surface, whereupon the acid site is restored.This will continue, as long as the surface sites remain catalytically active (fig. 5). All the evidence supports the fact that the ring-opening process is surface-catalysed and is rate-determining.21y 22$ 27 From the surface reaction occurring in the mordenite pores, as schematically depicted in fig. 6, the general stoichiometric relation may be expressed in the following way: k+l k*2 k+3 (PA),+X $ [PA. - *XI* [AA. - *XI* + (AA),+C k-1 k-2 k--3 where PA refers to paraldehyde, I= the vacant surface sites, [ ]* the activated surface intermediate, ( ), the molecules in a gas phase and k + , - to k , , the rate constants for the respective steps. As detailed by Hutchinson et aZ.,O several rate equations can be derived from the above relations, depending upon the state of equilibrium and the rate-determining steps and their derivation may be quite complicated.For the depolymerization under chromatographic conditions, however, simplification is possible through the following rationalization:7* 11, 22 (i) the reversal of the surface reaction step is negligible, viz. k-, LLI 0, (ii) the product species is removed from the surface immediately upon formation, uiz. k-, N 0, and (iii) the surface reaction is regarded as rate-determining,P. J. CHONG A N D G. CURTHOYS 0 1; C H 3 A 0 ' L C H , 0 0 0 0 'si/ \iL/ \ / \ ./ \JO / \ / \ ,si\ 7'\ Y\ c l I M (OH)',"_, H 0 0 0 0 0 Al Si Si A1 \si/ \-/ \ / \ / \-/ / \ / \ / \ / \ I\ 59 0 0 0 0 0 \ ./ \-/ \ ./ \ ./ \-/ SI Al SI SI AL / ' / \ / \ / \ / \ FIG.5.-Scheme for mechanism of depolymerization of paraldehyde over zeolite surface: (M = exchange cation with valence of n). a 11 a -+ free molecules (vapour phase) mass transfer 11 (diffusion) surface- bound ,,B pore-mou t h area mordenite cavity @g@--- s u r f a c e active centres methyl groups - - - _ _ FIG. 6.-Pictorial representation of depolymerization of paraldehyde in mordenite cage (hypothetical). 3-260 DEPOLYMERIZATION OF PARALDEHYDE OVER MORDENITE viz. d[PA+AA]*/dt 2: 0. This reduces the rate expression to an equation of the Langmuir-type, uiz. d(PA)g - k w w g - -- dt 1 +K(PA), where k is the surface reaction rate constant and K is the Langmuir adsorption constant.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 25 24 25 26 27 R. J. Kokes, H. Tobin and P. H. Emmett, J . Am. Chem. Soc., 1955, 77, 5860. W. K. Hall and P. H. Emmett. J . Am. Chem. Soc., 1957, 78, 2091. D. M. Nace, Ind. Eng. Chem., Prod. Res. Deu., 1969, 8, 1. W. A. Blaton Jr, C. H. Byers and R. P. Merrill, Ind. Eng. Chem. Fundam., 1968, 7, 61 1 ; 154th Natl Meeting of the Am. Chem. SOC. Chicago, 1967. D. W. Bassett and H. W. Habgood, J . Phys. Chem., 1960, 64, 769. S. Z. Roginskii and A. L. Rozental, Dokl. Akad. Nauk SSSR, 1962, 146, 152; Kinet. Katal., 1964, 5, 104. ( a ) G. A. Gaziev, V. Yu. Filinovskii and M. I. Yanovskii, Kinet. Katal., 1963, 4, 688; ( 6 ) T.Hattori and Y. J. Murakami, J . Catal., 1968, 10, 114. L. S. Ettre and N. Brenner, J . Chromatogr., 1960, 3, 524. W. W. Schmiegel, F. A. Litt and D. 0. Cowan, J. Org. Chem., 1968, 33, 3334. W. K. Fowler, C. H. Duffey and M. C. Miller, Anal. Chem., 1979, 15, 14, 2333. (a) J. R. Conder and C. L. Young, Physicochemical Measurement by Gas Chromatography (J. Wiley, New York, 1979), p. 550; (b) p. 531. S. H. Langer, J. Y. Yurchak and C. M. Shaughnessy, Anal. Chem., 1968, 40, 1747. T. A. Egerton and F. S . Stone, J . Chem. Soc., Faraday Trans. I , 1973, 69, 22. W. W. Marsh Jr, R. M. Heistand and J. 0. Rice, Reprints-General Papers, A-Petroleum Chemistry (Am. Chem. SOC., Petrol. Chem. Inc., Chicago, 1970), vol. 15(3), A-123. J. C . Bartlet and D. M. Smith, Can. J . Chem., 1960, 38, 2057. D. Atkinson and G. Curthoys, J. Chem. Ed., 1978, 55, 564; 1979, 56, 802. D. W. Breck, Zeolite Molecular Sieves (J. Wiley, New York, 1974), p. 29. D. Atkinson and G. Curthoys, J . Phys. Chem., 1980, 84, 1358. S. P. Walvekar and A. B. Halgeri, J . Res. Inst. Catal., Hokkaido Uniu., 1972, 20, 219. H. L. Hutchinson, P. L. Barrick and L. F. Brown, Chem. Eng. Prog. Symp. Ser., 1967, 63, 18. A. Clark, The Theory of Adsorption and Catalysis (Academic Press, New York, 1970), p. 239. J. W. Hightower and W. K. Hall, J . Phys. Chem., 1968, 72, 4555. A. H. Keough and L. B. Sand, J . Am. Chem. Soc., 1961, 83, 3536. D. C. Carpenter and L. 0. Brockway, J . Am. Chem. Soc., 1936, 58, 1270. P. B. Venuto and P. S. Landis, Adu. Catal., 1968, 18, 259. K. A. Becker, H. G. Karge and W. D. Streubel, J . Catal., 1973, 28, 403. M. L. Poutsma, Zeolite Chemistry and Catalysis, ed. J. Rabo. (Am. Chem. SOC., Washington, 1976), p. 529. (PAPER O/ 1540)