J. Chem. SOC., Faraday Trans. I, 1987, 83, 1667-1684 Isothermal Titration of Supported Platinum Part 2.-Alkene Titration using Cyclohexene Mariana S. W. Vong and Paul A. Sermon* Department of Chemistry, Brunel University, Uxbridge UB8 3PH Isothermal alkene titrations (AT) of Pt/SiO, have been studied on a continuous basis with simultaneous measurement of hydrogen, cyclohexene, cyclohexane and benzene concentrations. Cyclohexene either donates H atoms to the catalyst at 423 K or extracts them from the catalyst at 293-353 K, the temperature being determined by the kinetics and thermo- dynamics of the titration. After optimisation of titration parameters, standard temperatures for the dehydrogenation and hydrogenation steps in the titration were chosen to be 423 K and 310 K.At these temperatures rates of donation of H atoms (dehydrogenation) increased with increasing cyclohexene partial pressure, but subsequent rates of extraction of H atoms passed through a maximum at 500-630 Pa; the optimum partial pressure of cyclohexene selected for standard titrations was, therefore, ca. 460 Pa. Titration mechanisms are considered in detail and suggest that Pt surfaces areas can be estimated from the first phase of H-atom donation by cyclohexene; indeed these (51.8 m2 per g Pt) were in close agreement with those estimated by hydrogen chemisorption (53 m2 per g Pt). However, alkene adsorption can affect results from numerous repeated titrations. The alkene titration when operated in a dehydrogenation mode (AT,) with hydrogen donation to an oxygen-covered catalyst appeared far more satisfactory than previously used titrations involving alkene hydrogenation (AT,) by preabsorbed hydrogen.Since a complete differentiation of Pt-held from support-held hydrogen could not be achieved under isothermal conditions a temperature-programmed titration is proposed. Hydrogen can spill over or migrate from a supported metal to the adjacent support phase.l It may subsequently become directly involved in hydrogenation reaction,2 may be desorbed thermall~,~ reduce the support4 or migrate back to the surface of the supported metal for r e a c t i ~ n . ~ ? ~ The second and fourth processes are the most likely courses of action for hydrogen spilt over in Pt/Si02. Alkene titrations of hydrogen preadsorbed (AT,) on heterogeneous catalysts were first7 carried out isothermally at an arbitrary temperature and titrant partial pressure using pent-1-ene (and later ethene) with the aim of estimating the extent of metal-held hydrogen (and hence the metal dispersion) and also the extent to which spillover to the support was reversible.The first isothermal separations were based upon the kinetics with which these different surface hydrogen species were likely to be available for titration by an alkene and was not complete and hence not entirely successful. In an attempt to improve separation and resolution, other recent works has concentrated on titration of preadsorbed (and spilt over) hydrogen on catalysts with pulses of alkenes. The total quantities of hydrogen so extracted have been 5 to 25 times greater than the quantities thought to be held on the supported Pt alone in monolayer capacity, and presumably involved incomplete separation of metal-held hydrogen and support-held hydrogen.A suggestion8 for but-1-ene titration that the yield of butane was proportional to the number of corner Pt sites absorbing two H atoms, while the yield of but-2-enes was proportional t o the number of metal sites absorbing only one H atom was interesting 16671668 Cyclohexene Titration of Pt and potentially useful if alkene titration techniques can be optimised. One problem with pulse methods is that preadsorbed H, can be desorbed between pulses and re- equilibrationg between the support and Pt sites may disturb the initial hydrogen population of these surface sites and hence the ultimate separation.In addition, relatively high alkene concentrations constitute the pulses, and the adsorption characteristics for the alkene under these conditions is not well known. There is increasing need for such techniques of in situ catalyst characterisation. In the work presented here the new alkene titrant of cyclohexene was chosen to titrate silica-supported Pt catalysts since it readily donates and accepts hydrogen to and from the catalyst surface during dehydrogenation (AT,) and hydrogenation (AT,) phases of the titration. This is a radical but improved departure from past practice. To allow a good understanding of the kinetics of H-atom donation and extraction, this titrant was to be added continuously at constant concentration and temperature.The equilibrium position of the above reactions depends on the temperature and pressure, but temper- atures of 300 and 600 K favour hydrogenation and dehydrogenation, respectively ; relevant equilibrium constants for hydrogenation and dehydrogenation are : InK, = 30.31 at 298 K 18.08 at 400 K 5.83 at 600 K In Kd = -9.20 at 298 K 0.01 at 400 K 9.34 at 600 K. For this reason cyclohexene is an intriguing titrant for probing supported catalyst surfaces, but beside hydrogenation-dehydrogenation, cyclohexene can simultaneously undergo disproportionation to cyclohexane and benzene. However, this is not the dominant reaction under conditions of titration selected here. The extent of adsorption of cyclohexene and the extent of H, desorption has been measured on catalysts used here under conditions relevant to the titrationlo The continuous titration mode with simultaneous measurement of concentrations of hydro- carbons and H, was selected because of the ease with which kineticscould be interpreted and compared with these known absorptive and desorptive properties of the catalyst under the same reactant partial pressures and temperatures.Thus the continuous titration using cyclohexene has been optimised and the results subsequently obtained are now reported. Titration Model Fig. 1 represents possible mechanistic routes by which cyclohexene might interact with a Pt surface. Here * denotes a vacant site on a ‘clean’ metal surface, O,, an oxygen atom adsorbed in a bridged site and H, an absorbed hydrogen atom. In the new dehydrogenative alkene titration mode (with cyclohexene donating H atoms to the catalyst) the initial surface of the Pt may be ‘clean’ with a high concentration of vacant Pt surface sites and a high value of coverage 6, (in AT, titration) or may be partially covered by preadsorbed oxygen and have a moderate value of coverage (in AT, titration).In the older hydrogenative mode of alkene titration using cyclohexene the Pt surface on which the titration is initiated is one on which hydrogen has been preadsorbed and where 6,. % 8, (i.e. in AT, titration). Initially, in the dehydrogenative mode either k’ > k” or k’ < k” (depending upon the initial oxygen coverage of the supported Pt) and in the hydrogenative mode k > k’. It is important to note that surface coverages will change as the titration proceeds and that these inequalities will not be maintained.Under conditions of continuous hydrogenation on Pt, it has been suggestedll that half-hydrogenated C6H11* is the dominant hydrocarbon surface species. It is important to note here the effects upon the rate of titration of initial values of 8,, Bo,p or OH,.M . S . W. Vong and P . A . Sermon 1669 Fig. 1. Mechanisms of cyclohexene titration of vacant (*), oxygen-covered (Oe2) and hydrogen- covered (H,) sites on oxide-supported platinum where relevant primary forward adsorptive rate constants are k’, k” and k, respectively, and the first surface rate constants are k’, k . k ” is one of the rate constants defining conversion of 2H, into H,(g) and 2,. Two OH, may convert to H20(g) and O*2.Surface hydrocarbons may be mono-n, di-n, tri-n, mono-a, di-a or tri-a, adsorbed to * sites. In the hydrogenation mode Boudart et a1.12 have suggested that the half-hydrogenated species CGHll, dominates the surface hydrocarbon species. Experimental Materials Silica-supported Pt (E) prepared and characterised as reported previously12 was studied under isothermal titration conditions. Its main characteristics are given in table 1. Liquid cyclohexene (BDH research grade, purity > 99.8%) was purified by shaking with activated alumina before use to remove traces of peroxide which might be present. The purified cyclohexene contained no impurity detectable by gas chromatography. The hydrogen (BOC, 99.9% purity) and nitrogen (BOC, white spot, 99.9% purity) were used and were further purified by passage through beds of pre-reduced 1 % Pd-H, WO, (0.4 g) and MnO,-celite (7.0 g) powders, respectively, and cold traps at 195 K for the removal of traces of oxygen (< 1 ppb oxygen) and the vapour pressure of water to 0.9 ppm.1670 Cyclohexene Titration of Pt Table 1.Characteristics" of Pt/SiO, sample E 5 wt% Pt (3.9 wt% Pt determined by a.a.) Davison 70 support 3.35 x 1019 H atoms chemisorbed per g catalyst at monolayer capacity (qHC) 2.951 x 1019 ethene molecules per g catalyst at monolayer capacity (qCH) 16.83 x 1019 cyclohexene per g catalyst at monolayer capacity (qCH,,) 2.80 x 1019 H atoms desorbed per g catalyst in a, peak (380 K) 2.30 x 1019 H atoms desorbed per g catalyst in /?, peak (598 K) 7.68 x 1019 H atoms desorbed per g catalyst in y, peak (707 K) 12.78 x 1019 H atoms desorbed per g catalyst in all peaks of t.p.d.52.60 m2 per g Pt average surface area indicated by all chemisorptions 53.60 m2 per g Pt average surface area indicated by H, chemisorption 81.60 m2 per g Pt surface area indicated by a+/? t.p.d. peaks 5.33 nm average Pt particle size indicated by all chemisorptions 5.03-9.45 nm average Pt particle size indicated by TEM 5.30-5.90 nm average Pt particle size indicated by XRD ~ a Methods and results of chemisorption, X-ray diffraction, transmission electron microscopy and temperature-programmed desorption of H, have been described l2 Titration Apparatus A conventional flow system was used for cyclohexene titration on catalysts at atmos- pheric pressure.The cyclohexene saturator consisted of a 250 cm3 round-bottom flask (containing 25 cm3 of purified cyclohexene) and two empty bubblers, one which allowed the purified N, to be cooled to 243 K before entering the saturator flask in order to avoid any temperature gradient, which would affect the saturation vapour pressure of cyclohexene, and the second was to ensure gerfect cyclohexene saturation in N, before entering the catalyst reactor. Both bubblers and the flask were immersed in a thermostat bath (Grant LB8) filled with paraffin at 294 to 238 K with a sensitivity of kO.1 K and uniformity of f 0.3 K (measured using a precision mercury thermometer). Gas reactants entered a Pyrex microcatalytic reactor, passed through the sinter and reacted with the catalyst thereon.Samples of the resulting gas mixture were then taken by a gas-tight syringe for g.c. analysis. A mercury thermometer measured the temperature of the catalyst bed to f0.05 K. The temperature of the reactor was controlled by a stanton Redcroft linear temperature programmer (mark I11 LVP/CA4/R) and a Stanton Redcroft low-mass vertical furnace (LMVS 100). The flow rates of the gas were measured by a soap-film flowmeter (4 0.05 cm3 min-l). Gaseous hydrocarbons were analysed by gas chromatography (Perkin-Elmer gas chromatograph model F11 with a flame ionisation detector and a steel column packed with 5% polyethene glycol 400M coated on 100-200 mesh celite) and the peak areas were integrated (Minigrator, Spectra-Physics 2/947). Concentrations of H, in the gaseous product were analysed by a Perkin-Elmer gas chromatograph model F17) with a hot-wire conductivity detector and a glass column filled with molecular-sieve column packing material.Procedure Isothermal cyclohexene titrations were studied. Samples (0.4 g) of the catalyst with 0-covered surfaces in the reactor were purged with purified N, at room temperature. The sample was then heated to the titration temperature in N, and the dehydrogenation titration (AT,) commenced. Purified N, saturated with cyclohexene [p,(alkene partialM . S. W. Vong and P . A . Sermon 1671 pressure) 460 Pa] left the saturator and entered the reactor at a flow rate of 7 cm3 min-l. The concentrations of cyclohexane, cyclohexene, benzene and hydrogen in the reactor were monitored as a function of time by gas chromatography for 1 h.Then the cyclohexene-nitrogen reactant gas flow was stopped and the sample was cooled rapidly to room temperature before being purged with nitrogen (100 cm3 min-l). Titration AT, of the chemisorbed hydrogen at a low temperature was then recommenced by passing cyclohexene-nitrogen over the catalyst (7 cm3 min-l, p A = 460 Pa) and the reaction was followed until equilibrium was reached. Calculation Within certain temperature ranges, cyclohexene underwent slight disproportionation simultaneously with either hydrogenation or dehydrogenation over the catalyst. How- ever, the percentage dehydrogenation is given by (([C,H,] - [C,Hl,]/2) x 100/[inlet C,H,,,]} and the percentage hydrogenation is given by (([C,H,,] - 2[C,H,]) x lOO/[inlet C6Hlo]).Hence, the percentage conversion of the titrations in both hydrogenative and dehydrogenative directions was determined. From this the net rate of H-atom donation or retrieval from the catalyst could be determined by monitoring the gaseous H, concentration. Subsequently, these were converted to Pt surface areas assuming the number of Pt atoms preadsorbing an oxygen molecule in AT,(n,,) or a hydrogen molecule in AT,(H,,) was 4 and 2, respectively, irrespective of any adsorbate dis- sociation. Under conditions of dehydrogenative titration at 423 K, water produced was assumed to be released to the gas phase and not to interfere with the titration. Results for Optimisation of Isothermal Titration Thermal Reversibility First an attempt was made to study the thermal reversibility of the interaction of cyclohexene with a sample of Pt/SiO, (E), after this had been reduced in H, at 375 K for 1 h and stored in air at 295 K.Cyclohexene (460 Pa) in N, (101 kPa, predried at 195 K) was passed continuously over the catalyst at a rate of 7 cm3 min-l. As the temperature of the reaction was raised in a stepped manner from 273 to 495 K catalytic measurements at each step after 5 min equilibration revealed benzene produced by dehydrogenation at a rate which increased with temperature (see fig. 2). As the temperature was then reduced in steps after reaching 495 K, the rate of dehydrogenation decreased (fig. 2), reaching almost zero at 295 K, but the rate of dehydrogenation curve did not follow the same path when the temperature increased and decreased and in this context was hysteretic.It is reasonable to suppose that the Pt surface of the sample was initially covered by oxygen (as in an H,-0 titration) with the surface coverage of bridged oxygen on * sites, O0*,, being high; upon this the cyclohexene absorbs possible heterolytically and dehydrogenates (see fig. 1). Subse- quently surface OH groups release water and O0,, decreases. Thus B0*, decreases as temperature increases in fig. 2(a). In addition, as oxygen is removed, vacant platinum surface sites * are then occupied by further chemisorbed hydrogen (H*), also produced from cyclohexene dehydrogenatim (and some adsorbed cyclohexene, benzene or cyclohexane). As the reaction proceeds, the surface coverage of hydrogen 0,.(and hydrocarbon) increases until it reached an equilibrium value at the given temperature. Thus, when the experiment was continued with decreasing temperature on the same metal surface onto which hydrogen has been produced in the first run with increasing temperature, dehydrogenation of cyclohexene was slightly less favourable. Fig. 2 (b) shows the rate of desorption of hydrogen produced from the cyclohexene dehydrogena- tion in fig. 2(a), in addition to which some must be held by the Pt surface. Plots of1672 Cycluhexene Titratiun of Pt 37.5 - ( a ) c( E 3 c, CI k 27.5 - +.’ 0 M a % v) d 8 17.5- c1 z 0 - rl 2 7.5 - I 2 73 433 513 T/K T/K Fig. 2. Effect of increasing (0) and decreasing (a) temperature on the rate of cyclohexene dehydrogenation (a) and on the rate of H, desorption during cyclohexene dehydrogenation (b) on Pt/SiO, (E) with preadsorbed oxygen.p A = 460 Pa in N,; total flow rate 7 cm3 min-l. hydrogen desorption during the same experiment also show a hysteresis loop in which the rate of hydrogen released by cyclohexene and desorbed increased as the temperature increased, and decreased less slowly when the experiment was repeated with decreasing temperature on the same metal surface, even though dehydrogenation was less favourable. Thus the surface ultimately became saturated with hydrogen (or cyclohexene) during the dehydrogenation of cyclohexene, and subsequent hydrogen uptake by the catalyst is relatively low. Comparison of fig. 2(a) and (b) shows that the net rate of hydrogen retention by the catalyst during titration is low, but significant; at 433 K ca.21 x lo1’M . S. W. Vong and P . A . Sermon 1673 t/min Fig. 3. Rate of production of H atoms as a function of time during cyclohexene dehydrogenation (0, 0) and rate of their release from preoxidised Pt/SiO, (E) surface (@, a) at 474 K (0, @) and 423 K (0, m). Conditions as in fig. 2. Shaded areas relate to hydrogen retained by the catalyst. cyclohexene molecules are dehydrogenated per g catalyst min-l. This must produce 84 x 1017 H atoms per g catalyst min-l, but only 53 x 1017 H atoms per g catalyst were released to the gas phase during heating and 75 x lo1' H atoms during cooling [see fig. 2(b)] at this temperature. Thus a small but significant fraction of hydrogen is retained by the Pt/SiO, catalyst and this varies with the dominant surface coverages on Pt.The initial extents of hydrogen retention by the catalyst during dehydrogenation titration of cyclohexene might, therefore, be used to deduce the Pt surface area if adsorption stoichio- metries are known. However, one complication is the fact that hydrogen retention continued at a very low rate (see fig. 3), even at long titration times at 423 K and more so at 474 K, despite the titration reaching a state of dynamic equilibrium. This suggests that a slow continuing uptake of hydrogen by the catalyst is involved, which could be explained by hydrogen migration from the Pt to the silica support. This continued to occur at all titration times. More evidence will be discussed on these points wher, consideration is given to the two phases of titration.Surprisingly, cyclohexene hydro- genation, which was expected to be favourable at low temperature and high OH., was not observed. Fig. 4 shows t.he effect of temperature on the rate of cyclohexene disproportionation under the conditions in fig. 2; this passed through a maximum at ca. 373 K before falling to zero at ca. 503 K. At low temperature hydrogenation prevails, as does dehydrogen- ation at high temperature. Hence 310 K was chosen for the hydrogenative titration phases and 423 or 473 K for the dehydrogenative titration phases (see table 2). Under these titration conditions disproportionation was not dominant and correction could be made for the extent of its occurrence. Temperature of Titration The effect of the temperature of dehydrogenation titration was studied on fresh samples of Pt/SiO, (E) at 423 and 473 K.Results obtained are recorded in table 2 and fig. 3. The1674 Cycluhexene Titration of Pt 40 0 E 20- v) I; 2 10- 0 I 513 Fig. 4. Rate of disproportionation of cyclohexene over Pt/SiO, (E). Conditions as in fig. 2. Table 2. Effect of temperature of dehydrogenation titration on repeated hydrogenation titration cycles dehydrogenation titration hydrogenation ti tra tion no. H atoms donated* no. H atoms extracted run T/K ( x 1019 per g catalyst) T / K ( x 1019 per g catalyst) 423 5 6 473 6.85 8.10 3 10 310 310 310 3 10 3 10 310 310 3 10 3 10 5.39 2.64 2.37 2.02 I .36 0.76 5.25. 2.40 0.77 0.27 a Up to 60 min titration. amount of hydrogen abstracted by the first hydrogenative titration at 310 K with an H-covered catalyst in both runs 1 and 7 was in reasonable agreement.However, the amount of hydrogen retained by the catalyst in dehydrogenation up to 60 min titration was slightly higher at 473 K (as revealed by the shaded areas in fig. 3); possibly as a result of an increase in the rate of hydrogen spillover from the platinum to its silica support.M . S . W. Vong and P. A . Sermon 1675 '0 tlmin Fig. 5. Rate of production of H atoms as a function of time during cyclohexene dehydrogenation (0) at 423 K (and other conditions as in fig. 2) and rate of H-atom desorption from the Pt/SiO, surface (a). Data as in fig. 3. Hatching indicates H held on Pt and the unshaded area between the profiles suggests H, retained by the catalyst, but spilt over onto the silica support.However, it must be noted that donation of H to the catalyst has not finished within this experimental time and total uptakes would have been much higher. Repeated HydrogenatiowDehydrogenation Titration Cycles Repeated dehydrogenation-hydrogenation cycles were carried out as described above. Results of the repeated titration cycles on catalyst E at two dehydrogenation tempera- tures and constant hydrogenation temperatures are recorded in table 2. This shows that the amount of hydrogen titrated and removed by cyclohexene hydrogenation at 310 K decreased with the number of titration cycles and the rate of decrease was more rapid at the higher temperature of intermediate dehydrogenation. During 0,-H, titration cycles on Pt catalyst surfaces, enhancement of hydrogen adsorption and retardation of oxygen adsorption has been reported;13 this effect was partly attributed to surface restructuring on exposure to oxygen.From the dehydrogenation results in runs 1 and 7 in table 2 there was no significant enhancement on repetition. However, in hydrogenation the catalyst has become deactivated by the accumulation of long-lived adsorbed cyclohexene or a carbonaceous overlayer of Pt particularly at higher temperatures.14 Hence, repeated characterisation of a sample by cyclohexene titration would induce deactivation and cannot be advisable. Effect of Cyclohexene Partial Pressure (pA) The effect of cyclohexene partial pressure on the hydrogenative titration at 310 K was studied over Pt/SiO, (E) after catalyst reduction by cyclohexene dehydrogenative titration at 423 K for 1 h.It was possible to estimate the extent of hydrogen donation to and held on the catalyst during dehydrogenative titration (see fig. 5 ) from the shaded areas and unshaded areas (relating to Pt and support sites, respectively) between rates1676 Cyclohexene Titration of Pt X I., I 1 I lo 0.4 0.8 1.2 1.6 P A IkPa Fig. 6. Effect of cyclohexene partial pressure on the rate of cyclohexene dehydrogenation at 423 K. Conditions as in fig. 2. Table 3. Effect of cyclohexene partial pressure on repeated hydrogenation titration cycles dehydrogenation titration hydrogenation titration cyclo hexene partial rH pressure rate H atoms donated no. H atoms extracted /Pa T/K ( x lOlg per g catalyst min-l) T/K ( x lOlS per g catalyst) 289 423 340 423 459 423 629 423 723 423 913 423 1080 423 1372 423 0.602 0.764 0.927 1.050 1.380 1.522 1.490 2.046 310 310 310 310 310 3 10 3 10 3 10 2.22 2.79 3.66 6.76 2.86 4.0 1 3.10 2.60 of hydrogen liberation and desorption as a function of time at 423 K (fig.5 ) or as a function of different reactant cyclohexene partial pressures (see fig. 6). The rate of cyclohexene dehydrogenation was found to increase with the partial pressure of cyclohexenep, (fig. 6), resulting in an increasing rate of hydrogen liberated by the reaction (table 3). However, the partial pressure of hydrogen in the system also increased as more product hydrogen was desorbed. Since the extent of chemisorption of hydrogen on Pt and the rate of hydrogen spillover from Pt to SiO, are directly proportional to the partial pressure of hydrogen,15 the total amount and rate of hydrogen retained by the catalyst was found to increase with pA (see fig.7 and table 3). However, with the hydrogenative titration it became more difficult to differentiate at higher alkene partial pressures the metal surface area owing to the influence of reverse hydrogen spillover. In hydrogenative titration, the total amount of hydrogen extracted from the catalyst first increased rapidly with p,, passed through a maximum at ca. 533 Pa and then fell slowly at higher p A (see table 3 and fig. 8). As discussed above, the surface coverage ofM . S. W. Vong and P. A . Sermon 1677 I I I 20 40 t/mh Fig. 7. Rate of H-atom liberation during cyclohexene dehydrogenation at 423 K at different partial pressures of cyclohexene: 0, 1370; A, 913; V, 629; 0, 459 Pa.I I 1 0.5 1 .o 1 .! PAlkPa 120 30 k 3 6 M a PI \ 40 0 Fig. 8. Quantities of H-atom consumption during cyclohexene hydrogenation at 3 10 K at different partial pressures of cyclohexene. The surface areas for Pt were deduced assuming Pt, : H = 1 : 1 (or nH* = 2). Conditions as in fig. 2.1678 Cyclohexene Titration of Pt hydrogen OH, resulting from cyclohexene dehydrogenation (and cyclohexene coverage) increased with pA. Hence in subsequent hydrogenation more hydrogen would be abstracted. However, at even higher p A , the concentration of surface carbonaceous species 8, would be substantial, modifying the behaviour of the catalyst for the hydrogenation reaction from that of the clean surface.16 Consequently, a decrease in the amount of recoverable hydrogen from the titration surface was observed at high alkene pressure.Thus it is possible to see the separate effects of increasing hydrogen and hydrocarbon coverages. Results of Isothermal Cyclohexene Titration under Optimum Conditions Dehydrogenative Titration Benzene and hydrogen were produced in the cyclohexene dehydrogenative titration as a result of dehydrogenation. Fig. 5 has already shown the rate of hydrogen liberated at 423 K as estimated from dehydrogenation and the actual amount of hydrogen desorbed into the gas phase. The area between two curves is the amount of hydrogen consumed by and remaining on the catalyst surface either on Pt or spillover onto the silica support.Fig. 9 shows in greater detail the rate of hydrogen donation to the catalyst during titration at 423 K. Initially the rate of hydrogen donated (i.e. reacting with the siirface oxygen and the subsequently chemisorbed on the Pt sites *) increased rapidly, passed through a maximum and then declined to a low value. All rates produced compositions consistent with thermodynamics equilibrium (i.e. favouring dehydrogenation at 423 K). It is proposed that in the first phase the area below the curve is a measure of hydrogen consumed in titration and chemisorbed on the Pt, while the area in the second phase measures the amount of hydrogen spillover from the Pt to the support (provided the rate of titration is allowed to reach zero).The average surface areas of Pt so estimated by cyclohexene titration from the shaded area in fig. 9 and from equivalent plots on this catalyst are given in table 4 together with assumptions made (i.e. no* = 4 on the initial surface in AT, and nH2 = 2 after AT,). Areas (51.8 10 m2 g-l) so calculated are comparable with those from hydrogen chemisorption at room temperature (53 m2 per g Pt). The subsequent continuing retention of hydrogen may be related to spillover and the displacement at say 40 min (where the rate of H-atom donation is changing only slowly) titration time in fig. 9 is presumably a measure of the spillover rate. Therefore both rates of spillover and metal surface areas in Pt/SiO, samples may be measured using cyclohexene dehydrogenation titration at 423 K and 460 Pa cyclohexene.Hydrogenative Titration A series of experiments was carried out with samples (0.2 g) Pt/Si02 (E) which were reduced using cyclohexene dehydrogenation at 423 K for 1 h before cooling to 295 K and then purged with N,. Cyclohexene titrations were then carried out at various temperatures and the results obtained are recorded in table 4 and fig. 10. The rate of cyclohexene hydrogenation also increased with time, passed through a maximum and then fell to zero and a small amount of dehydrogenation occurred. No hydrogen was detected in the gas phase during hydrogenation. Fig. 11 shows the variations of the total amount of hydrogen abstracted by cyclohexene at various temperatures in terms of metal-held hydrogen in the primary titration peak; this was a maximum at ca.313 K. Hydrogenation was not observed below 293 K (where the cyclohexene hydrogenation is probably kinetically controlled and fails to reach completion and therefore results in underestimation in surface-area measurement) or above 353 K (where although reaction is more kinetically favourable and it is also accelerated by reverse spillover of hydrogenM . S. W. Vong and P. A . Sermon I I I I I679 I Fig. 9. Rate of H-atom liberation and donation to catalyst E in cyclohexene dehydrogenation at 423 K. The shaded area is assumed to relate to Pt-held hydrogen. Conditions are as given in fig. 2. from the SiO, causing an overestimation of metal surface areas; dehydrogenation is also favoured by a lower coverage of hydrogen on Pt and higher Kd than Kh).The total amount of hydrogen extracted from the catalyst determined from the amount of cyclohexane produced from plots such as that in fig. 10 (corrected for cyclohexene disproportionation) is recorded in table 4. Integration of the area under the main peak gave the total amount of hydrogen titrated. This was found to be much in excess of the amount measured by hydrogen chemisorption on Pt alone and was also somewhat variable (see table 4). The results can in part be explained if adsorbed hydrogen atoms are extracted not only from the Pt but also as a consequence of reverse spillover hydrogen from the SiO, support. It is unlikely that an alkene can be hydrogenated directly by hydrogen spilt over on SiO,, even though this occurs on A1,0,., Therefore, it is believed that during titration cyclohexene reacts first with chemisorbed hydrogen held on the Pt until OH* becomes very low, then hydrogen spilt over on to the silica migrates back to the platinum to sustain the titration further.Summary The dehydrogenation mode of the alkene titration seems to give better estimates of supported Pt areas. This is a significant advance on earlier titrations which have used a hydrogenative mode, It is interesting to note that in titrations at 328 K a small second peak was observed1680 Cyclohexene Titration of Pt Table 4. Titration of catalyst E with cyclohexenea dehydrogenation titration hydrogenation titration no. H atoms donated no. H atoms extracted run T/K ( x 10lQ per g catalyst) T/K ( x 1019 per g catalyst) 1 423 2 423 3 423 4 423 5 423 6 423 7 423 8 423 9 423 6.06 309 5.64 273 5.995 295 7.87 306.5 6.34 328 6.93 345 10.06 373 3 10 - - 299.5 6.8 1 0 1.33 3.83 4.57 1.31 0 7.06 2.66 a The stoichiometries assumed for determining the Pt surface area from cyclohexene dehydro- genative titration of a surface with preadsorbed oxygen was defined by nO2 being 4 before the titration and nH2 = 2 after titration (i.e.Pt,: 0: H = 2: 1 : 2) and product water was assumed to be desorbed at the titration temperature (423 K). This titration was denoted AT,. The stoichiometry assumed for determining the Pt surface area from the cyclohexene hydrogenative titration of a surface with preadsorbed hydrogen (denoted AT,) was given by nHz = 2 (i.e. Pt,: H = 1 : 1). Other adsorption stoichiometries might need to be assumed at very high Pt dispersions.The average Pt surface area so determined by the dehydrogenation titration AT, in its first six runs was 51.8 m2 per g Pt; precise areas increased only a little 48.5 to 55.4 m2 per g Pt, but the 7th value was higher. Surface areas deduced from the hydrogenation titration AT, were much more variable. 40i tlmin Fig. 10, Rate of H-atom consumption-xtraction in cyclohexene hydrogenation at 309 K. Conditions as in fig. 2.M . S . W. Vong and P. A . Sermon 1681 I I 1 I I I 1 293 313 333 3 53 T/K Fig. 11. Total quantities of H atoms extracted during cyclohexene hydrogenation (conditions as in fig. 2) over Pt/SiO, (E). '2 30c I I I 1 0 20 40 60 80 t/min Fig. 12. Rate of extraction of H atoms during cyclohexene hydrogenation over Pt/SiO, (E) at 328 K.Conditions as in fig. 2. 56 FAR 11682 Cyclohexene Titration of Pt at the tail of the major titration peak (e.g. fig. 12). The appearance of this second peak at high temperature may support the assumption of the existence of spilt over hydrogen. Unfortunately, at high titration temperature, for reasons mentioned above, the total amount of hydrogen abstracted by cyclohexene was relatively small. Discussion The temperatures used here for the dehydrogenative and hydrogenative modes of the alkene titration involving cyclohexene are consistent with those used previously for continuous catalysis of cyclohexene dehydrogenation and hydrogenation (423 and 310-295 K). Thus cyclohexene has been observed12 to dehydrogenate over Pt(100) at 373 K and to hydrogenate over silica-supported Pt at 273-3 13 K or over stepped Pt(223) surfaces down to 298 K or over Pt black at 295 K.Such thermal conditions and the cyclohexene partial pressures used here have previously suggested structure-insensitivi ty. Earlier hydrogenative alkene titrations7 with pent- 1 -ene and ethene found that the rate of H-atom extraction from silica-supported Pt samples at 373 or 473 K was a maximum at times close to zero. Surprisingly, therefore, with cyclohexene titrant here maximum rates of H-atom extraction are only seen after 10-20 min. The causes of the induction period are worthy of further consideration. First, the extent of cyclohexene adsorption qCHfr measuredlo on Pt/SiO, (E) at 295 K was greater than that of ethene (see table 1) and so the extent of alkene adsorption could not be said to cause the different rates of attainment of maximum titration rates.However, qCH” did appear to be dependent on the average Pt crystallite size dpt, possibly as a result of differences in their abilities to strongly chemisorb hydrocarbons in multiply-bonded states.17 Cyclohexene is adsorbed more weakly on Pt than benzene.ls Nevertheless, on Pt(ll1) at 293 K it has been suggested that z-bound cyclohexene may be a long-lived surface species. l4 Certainly, dehydrogenation activity on Pt( 1 1 1) and higher-index planes has been inhibited by strong hydrogen and benzene retention.lg Bearing in mind the small effect of Oo2, and 0, on the kinetics of the dehydrogenative titration mode (see fig.2), it is interesting that preadsorbed 0 on Pt surfaces has been noted to increase dehydrogenation and hydrogenation activity of Pt at 423 K.20 Hence, it is unlikely that preadsorbed oxygen on the surface of supported Pt used here could reasonably be expected to have inhibited the dehydrogenative titration here. The hydrogenative titration model (see fig. 1) is based upon that of Horiuti and Polanyi.21 In both hydrogenative and dehydrogenative modes let =II be the surface hydrocarbon species whose hydrogenation (or dehydrogenation) is rate determining. =I+ will therefore differ from mode to mode and with pretreatment conditions. It is important to note that Oo2* and 0,. have little effect upon the maximum rate and induction period of cyclohexene titrations (see fig.2) where rates on a fully oxidised surface are only slightly different from those when OH* is high). The induction period before reaching the maximum rate of dehydrogenation titration (8 min) is less than that (16 min) for the hydrogenative titration, but then the latter was at a significantly lower temperature. The rate of hydrogenative titration [k,f(O,* 0,) where the orders with respect to OH, may be very low in order to explain its small effect on the rate in fig. 21 will be a maximum at intermediate times since at low t , 0, is zero and 0,. is a maximum, while at long t, 0, is high and 0,. near zero. Equally, the rate of dehydrogenative titration [k,f(Oo2* 0,) in AT, or kdflO, 0,) in AT,, where again orders with respect to 0, and OO2* may be very small] will also be a maximum at intermediate titration times for the same reason.The observation of maxima in titrations using cyclohexene (but not ethene or pen-1-ene) may at first sight be considered a reflection on its slower adsorption either as a result of diffusion-limitation into the porous silica or an inherently slower adsorption step. However, fig. 7 shows no effect upon the induction period of increasing the cyclohexene partial pressure and this mi tigates against diffusional control or adsorptive control.M. S . W. Vong and P. A . Sermon 1683 Therefore the stoichiometry of alkene titrations using cyclohexene may be consistent with the assumptions in table 4: 3(Pt,-O) + 2 0 --+ 2(Pt2-),- 0 + 3H,O+ 2(Pt-H) where the average numbers of surface Pt atoms * which adsorb one oxygen (n0J and one hydrogen molecule (n,,), irrespective of whether dissociative chemisorption is involved, are 4 and 2, respectively, and the average numbers of surface Pt atoms n-bonding product benzene (nCHf) and reactant cyclohexene (ncHrt) may also be 2.This recognises the importance of n-bound surface hydrocarbons,22 but in practice ncH8 # nCH” and both are likely to be larger than in the above equations. However, at present it is not possible to define nCH, or nCH” or the nature of this carbonaceous adsorbate very precisely. Nevertheless, this does not detract from the value of AT, and AT, titrations since these merely count the numbers of H atoms added to or extracted from the catalysts using measurements of the rates of formation of gaseous cyclohexane and benzene, without the need to know the precise number or mode of cyclohexene adsorption (important though this is in catalysis and studies of the involvement of carbonaceous residues therein), provided self-hydrogenation is not involved.It has been noted that cyclohexene adsorption on hydrogen precovered Ni/SiO, is more extensive at 307 K than either benzene or cy~lohexane.~~ This means that product adsorption is unlikely to be more important (and interfere with these titrations) than titrant adsorption and that the titration assumptions used here are valid. Conclusions The optimum conditions for the preferred cyclohexene dehydrogenation titration on Pt catalysts studied were 423 K, p A = 0.4 kPa and 7 cm3 min-l cyclohexene-nitrogen.The surface areas of Pt determined under the above conditions were comparable to the surface areas measured by hydrogen chemisorption at room temperature. At high cyclohexene pressure and high titration temperature, metal surface-area measurements would be complicated by the extent of hydrogen spillover from the Pt to the SiO, support and also hydrocarbon retention on the metal and its support. The techniques will be a factor or two more sensitive on an oxidised surface (AT,) than on a clean surface (AT*). Cyclohexene-hydrogenative titration (AT,) at 3 10 K will involve cyclohexene molec- ules reacting with preadsorbed hydrogen on the metal producing cyclohexane. When 8,. on Pt becomes very low, reverse spillover of hydrogen atoms retained on the silica support will extend the titration in this mode.However, the isothermal cyclohexene- hydrogenative titration technique fails to differentiate metal-held hydrogen from support- held hydrogen and is very temperature-sensitive. It seems that this might be turned to advantage if it could be undertaken in a temperature-programmed mode. The use of this approach is being investigated. Therefore, at present the isothermal cyclohexene titration is promising for in situ catalyst characterisation; for Pt/SiO, there are merits of the cyclohexene dehydrogen- ative titration over previously used hydrogenation modes of titration. It is interesting that it can be used to characterise metal catalysts without the requirement of a ‘clean’ surface ; pretreatments to achieve this may elsewhere have involved severe catalyst surface restructuring.Further studies of the importance of (i) surface hydrogen migration, (ii) sites of alkene titration and (iii) carbanaceous species in isothermal alkene titrations will be de~cribed.,~ The provision of a research studentship to M. S. W.V. by the S.E.R.C. is gratefully acknowledged. 56-21684 Cyclohexene Titration of Pt References 1 P. A. Sermon and G. C. Bond, Catal. Retl., 1973, 8, 21 1 . 2 D. Bianchi, G. E. E. Gardes, G. M. Pajonk and S. J. Teichner, J . Catal., 1975, 38, 135; D. Bianchi, M. Lacroix, G. Pajonk and S. J. Teichner, J . Catal., 1979,59,467; G. E. E. Gardes, G. M. Pajonk and S. J. Teichner, J. Catal., 1974, 33, 145; S. J. Teichner, A. R. Mazabrard, G. Pajonk, G. E. E. Gardes and C. Hoang-Van, J .Colloid. Interface Sci., 1977, 58, 88; P. A. Compagnon, C. Hoang-Van and S. J. Teichner, Proc. 6th Znt. Congr. Catal. (Chem. SOC., London, 1979), vol. 1, p. 117; J. C. Schlatter and M. Boudart, J . Catal., 1972, 24, 482; J. L. Carter, P. J. Lucchesi, J. H. Sinfelt and D. J. C. Yates, Proc. 3rdZnt. Congr. Catal. (North Holland, Amsterdam, 1965), vol. 1, p. 644; A. J. Moffat, J. Catal., 1972, 27, 456. 3 R. Kramer and M. Andre, J. Catal., 1979,58,287; J. R. A. Anderson, K. Foger and R. J. Breakspere, J . Catal., 1979,57,458; J. P. Candy, P. Fouilloux and A. J. Renouprez, J . Chem. Soc., Faraday Trans. I , 1980, 76, 616; Y. Amenomiya, J . Catal., 1971, 22, 109. 4 S. Khoobiar, J . Phys. Chem., 1964, 68, 41 1 ; P. A. Sermon and G. C. Bond, J. Chem. SOC., Faraday Trans.I , 1976, 72, 730. 5 P. A. Sermon and G. C. Bond, J. Chem. SOC., Faraday Trans. 1, 1980, 76, 889. 6 P. B. Wells and G. R. Wilson, J. Catal., 1967, 9, 70. 7 P. A. Sermon and G. C. Bond, J. Chem. SOC., Faraday Trans. 1, 1976,72,745; G. Leclercq, J. Barbier, C. Betizeau, R. Maurel, H. Charcosset. R. Frety and L. Tournayan, J . Catal., 1977, 47. 389; G. C. Bond and P. A. Sermon, React. Kinet. Catal. Lett., 1974, 1, 3. 8 R. L. Augustine, K. P. Kelly and R. W. Warner, J. Chem. Soc., Faraday Trans. I , 1983, 79, 2639. 9 N. Hoyle, P. Newbatt, K. Rollins, P. A. Sermon and A. T. Wurie, J . Chem. Soc., Faraday Trans. I , 1985,81,2605. 10 M. S. W. Vong and P. A. Sermon, to be published; M. S. W. Vong and P. A. Sermon, J. Chem. Soc., Faraday Trans. I , 1987, 83, 1651. 1 1 J. W. A. Sachtler, M. A. van Hove, J. P. Biberian and G. A. Somorjai, Phys. Rev. Lett., 1980,45,1601; G. Leclercq and M. Boudart, J . Catal., 1981,71, 127; E. Segal, R. J. Madon and M. Boudart, J . Caral., 1978, 52, 45; S. M. Davis and G. A. Somorjai, J . Catal., 1980,65, 78; D. J. O’Rea, D. G. Loffler and M. Boudart, J. Catal., 1985, 94, 225. 12 A. R. Berzins, M. S. W. Vong, P. A. Sermon and A. T. Wurie, Ads. Sci. Tech., 1984, 1, 51. 13 E. Kikuchi, P. C. Flynn and S. E. Wanke, J. Catal., 1974, 34, 132. 14 J. L. Gland, K. Baron and G. A. Somorjai, f. Catal., 1975,36,305; D. W. Blakely and G. A. Somorjai, 15 P. A. Sermon and G. C. Bond, J . Chem. SOC., Faraday Trans. I , 1976, 72, 745. 16 S. M. Davis and G. A. Somorjai, J . Catal., 1980, 65, 78. 17 V. Ponec, Adv. Catal., 1983,32, 149; B. Van Keulen, W. R. Wichers and V. Ponec, React. Kinet. Catal. 18 Z. Paal and P. Tetenyi, in Surface and Defect Properties of Solids, ed. G. C. Bond and G. Webb 19 R. K. Herz, W. D. Gillespie, E. E. Petersen and G. A. Somorjai, J . Catal., 1981, 67, 371. 20 S. M. Davis and G. A. Somorjai, Surf. Sci., 1980, 91, 73. 21 J. Horiuti and M. Polanyi, Trans. Faraday SOC., 1934, 30, 1164. 22 J. J. Rooney and G. Webb, J . Catal., 1964. 3, 488. 23 R. 2. C. van Meerten, A. C. M. Verhaak and J. W. E. Coenen, J. Catal., 1976, 44, 217. 24 M. S. W. Vong, P. A. Sermon, G. Georgiades and R. L. Augustine, J . Chem. Soc., Faraday Trans. I , J. Catal., 1976, 42, 181. Lett., 1979, 12, 125. (Specialist Periodical Reports, The Chemical Society, London, 1976), vol. 5, p. 8 1 . submitted for publication. Paper 512127; Received 4th December, 1985