J. Chem. SOC., Faraday Trans. I, 1985, 81, 2027-2041 Comparison of the Surface Reactivity and Spectroscopy of Alkaline-earth-metal Oxides Part 2.-Dependences upon Temperature of Pre-activation for SrO BY JOHN NUNAN, JOHN A. CRONIN AND JOSEPH CUNNINGHAM* Chemistry Department, University College, Cork, Ireland Received 12th September, 1984 High-surface-area samples of strontium oxide, prepared by thermal decomposition of high-purity SrCO, in vacuo, have been examined for their reactivity and/or catalytic activity towards molecular gas probes and their luminescence spectra have been recorded. Temperature profiles are reported, as a function of prior outgassing temperature, for the development of room-temperature excitation and emission spectra. These are compared with temperature profiles for the development of room-temperature catalytic activity for &-type oxygen isotope exchange, 1 6 0 2 + leOp e 21s0180, and for the development of activity for nitrous oxide decomposition at various temperatures. Consideration is given to interpretations of the results in terms of ions at coordinatively unsaturated surface locations, to the extent to which these are exposed after thermal activation at various temperatures and to the involvement of point defects in the surface locations responsible for the various surface processes.Absorption features lying to the long-wavelength side of their absorption edges, and attributable to the photogeneration of correlated electron-hole pairs within the lattice (i.e. bulk excitons), have been reported for single crystals of high-purity alkali-metal halides1* and for some alkaline-earth-metal 0xides.~9 In terms of the tight binding approximation (generally considered applicable to these groups of highly ionic solids) such bulk-exciton transitions may be thought of as promoting an electron from an anion site enjoying full octahedral coordination onto a similarly coordinated adjacent cation site.Thus within the bulk of a perfect single crystal composed of Mi,+ and 0:; ions, the upward transition may be approximated by hv [M6c(nS0)2+, 0 6 C ( 2 p 6 ) 2 - ] [M6c(ns1)f, 06C(2p5)-l*' (1 a) Excitons thereby photogenerated, whilst highly mobile throughout perfect lattices at very low temperatures, are subject to radiative and non-radiative decay in real systems at normal temperatures, usually with a nett overall effect equivalent to the reverse of reaction (1 a).Radiative decay can occur via the intermediacy of self-trapped,2b defect- trapped5a or imp~rity-trapped~~ excitons and studies of such photoluminescence (emission) can yield information on the nature of the luminescence Spectroscopic observations reported in the present series of papers for the isostruc- tural alkaline-earth-metal oxides, MgO, CaO, SrO and BaO relate, however, to other electronic transitions, which are displaced to longer wavelength by 2-4 eV relative to bulk-exciton transitions. These are most readily observable in samples of high surface area, such as powdered oxide samples prepared by thermal decomposition in vacuo of the corresponding carbonates or hydroxides at high temperatures.Careful reflectance measurements, mainly by Stone et al., have revealed three partially resolved absorption 20272028 SURFACE REACTIVITY AND LUMINESCENCE OF SrO features, which they denote by I, I1 and 111 and which they interpret in terms of surface excitons, i.e. transitions analogous to reaction (1 a), except that the anion and/or cation sites involved are at surface locations and do not enjoy full octahedral c~ordination.~? * According to the interpretation espoused by Stone et al., features I1 and 111 in the reflectance spectra are to be associated with photogeneration of surface excitons involving 02- having only four- or three-fold coordination, e.g. hVl I [M~,(W~+, 0 , , ( 2 ~ 6 ) 2 - 1 ~ [M,,(w+, 0,,(2~5)-1* (1 b) where the subscript lc denotes low but unspecified coordination.Variation of these transition energies for the different alkaline-earth-metal oxides in accordance with the Mollwo-Ivey relation for localised excitons8 lent support to the interpretation of features I1 and I11 in terms of localised surface excitons. However, the transition energies of feature I were observed to vary in a similar fashion to freely diffusing excitons of the bulk, and this has been assigned to photogeneration of surface excitons at surface locations involving oxide ions having five-fold coordination upon flat { loo} terraces of the metal oxide: hVI [M~c(ns~)~+, 05c(2P6)2-I --+ [Mlc(nsl)+, 05c(2p5)-]* (14 Interpretations favouring steps, edges or corners as the surface topographical features mainly responsible for the existence of anions (and to lesser degree cations) with four- or three-fold coordination have been espoused not only by Stone and coworkers on the basis of reflectance measurements but also by Tench and coworkersg on the basis of luminescence spectroscopy .Recently, Duley has advanced an interesting variation upon the idea that coordin- atively unsaturated oxide ions are important in interpreting luminescence and absorption features of high-surface-area alkaline-earth-metal oxide Accor- ding to Duley's model, luminescence emission from MgO and CaO involves a charge-transfer-type transition from an electronically excited state of 0:; to the ground state of an adjacent F,+ centre: *Of;* * F , + d OFc * * F,* + hv. (2) An essential difference between this representation and those in reactions (1 a-d) is the recognition given to an important role of point defects at the surface [the defects of importance in reaction (2) being F,+ centres, consisting of vacant anion sites at the surface upon each of which a single electron has been localized].In another recent paper Duley has considered the involvement of coordinatively unsaturated hydroxide ions in luminescence from MgO and Mg(OH), powders at various stages of hydration deh~dration.~~ A preliminary account has recently been givenlo of the marked effects of Ba2+ dopant upon surface processes on MgO and of a probable influence of surface reconstruction and/or relaxation in modifying the degree of coordination of surface ions. A brief account was also given of problems attached to an interpretation of surface processes mainly in terms of ions at unrelaxed step, edge or corner locations.In view of these additional influences upon surface luminescence from alkaline-earth-metal oxide powders, and particularly in view of the very differing emphases possible for surface defects and surface impurities, we have initiated a programme to examine the extentJ. NUNAN, J . A. CRONIN AND J. CUNNINGHAM 2029 of correlation between surface luminescence and other surface properties likely to be dependent upon surface concentrations of point defects or impurities.'O Surface reactivity and catalytic activity for gaseous reactions mediated by charge transfer were selected as appropriate defect-sensitive processes.Furthermore, in view of the reported dependences of the surface reactivity and catalytic activity of alkaline- earth-metal oxide powders11-15 upon the temperature of prior activation, an important element of the strategy adopted in this second paper of the series is a careful comparison, made throughout with the same SrO material, between the dependences exhibited by the surface luminescence and the surface catalytic activity upon the temperature of prior activation. EXPERIMENTAL VACUUM PROCEDURES Both the thermal decomposition of strontium carbonate to SrO and its subsequent activity for N,O decomposition, or for the homomolecular oxygen isotope exchange (i.e. Oi6 + Ois + 2016018) were studied in a static quartz reactor using conventional vacuum procedures (base pressure 5 x Torrf) with mass-spectrometric analysis (base pressure Torr).In order to follow the progress of the thermal decomposition of high-purity SrCO, (Spex spectroscopically pure grade) in the quartz reactor, a powdered sample was slowly heated to 1273 K over a period of 10 h, during which the total pressure was monitored and the composition of gaseous decomposition products determined by leaking samples to a VG Micro Mass 6 mass spectrometer via a leak valve. The progress of thermal decomposition during activation in uacuo at temperatures up to 1273 K was also monitored by taking i.r. spectra at 300 K, after periods of activation at high temperature, of thin self-supporting discs, initially fabricated from SrCO, but converted progressively to SrO. Nitrous oxide was supplied by B.D.H.Chemicals and had a purity of > 99%. Before contacting PNzO z 3.5 Torr with the catalyst, the gas was further purified by a series of freeze-pumpthaw cycles, after which the purity was further checked by leaking the gas into the spectrometer through a by-pass valve. With the SrO catalyst at temperatures in the range 673-873 K, the reaction was initiated by admitting purified N,O and periodically obtaining the mass spectrum of gas samples leaked into the spectrometer. Reaction temperature and catalyst mass were chosen such that the reaction time for 50% conversion was > 10 min. Under these conditions the reaction rate was found not to be limited by diffusional effects. For oxygen-isotopic-exchange studies, an isotopically non-equilibrated mixture consisting of 50% Oi6+50% Oi8 was employed (Norsk Hydro). This gas was contacted with the catalyst at room temperature.The mass-spectrometric procedure was the same as for N,O decom- position, but oxygen pressures in the range 0.1-0.01 Torr were used. In general SrO was activated prior to reaction by heating under vacuum to the required activation temperature and maintaining this temperature for 1 h before cooling to room temperature in uucuo. The base pressure in the vacuum system employed was 5 x Torr. GAS-CHROMATOGRAPHIC PROCEDURE Nitrous oxide decomposition was also studied using g.1.c. and microcatalytic, flow-reactor systems operating at atmospheric pressure. A feature of these studies was the use of both the continuous-flow and pulsed-reactant procedures and full details have been given elsewhere of both the methods of operation.16 Strontium carbonate samples (200 mg) were decomposed in situ in the quartz microcatalytic reactor by heating at different temperatures, up to a maximum of 1273 K, for 1 h in a flow of dry helium at a flow rate of 40 cm3 min-l.Prior to the admission of nitrous oxide as pulses or as a continuous flow, the catalyst temperature was lowered to the reaction temperature in a flow of pure dry helium. The reactor was then isolated whle N,O + He mixtures were prepared such that the pressure of N,O was in the range 5&250 Torr. The total flow rate was then restored to 40 cm3 min-' t 1 Torr z 133.3 Pa2030 SURFACE REACTIVITY AND LUMINESCENCE OF SrO LUMINESCENCE PROCEDURES Instrumentation used in obtaining the excitation and emission spectra has previously been described in detail.Q Excitation was provided by a 250 W xenon lamp, from which appropriate wavelengths were sequentially selected using two Spex f4 monochromators whilst recording the excitation spectra.Another Spex monochromator, placed between the sample and the detector, was held at a fixed wavelength and augmented by appropriate cut-off filters when recording emission spectra. Spectra were automatically corrected for variations in excitation intensity with time, or changing wavelength, via instrumental comparisons (with an Ortec photon-counting system) of the pulse rates from two photomultipliers : one monitoring luminescence intensity and the other the intensity of the source at the same wavelength.Excitation spectra at acceptable signal-to-noise ratios were only obtained at wavelengths > 230 nm because of low source intensity at shorter wavelengths. When recording emission spectra under excitation at fixed wavelength, sharp cut-off filters (Corning 3-74 and 0-52) were placed between the sample and the emission monochromator to minimise scattered wavelengths from the xenon source reaching the photomultiplier. A band-pass of 5 nm was used for both the excitation and emission results. RESULTS SAMPLE ACTIVATION Decomposition profiles of SrCO, as a function of the outgassing temperature are shown in fig. I. Use of total pressure of gaseous products as the monitor indicated that the carbonate started to decompose at 873 K, reached a maximum at 1023 K and was almost complete at 1073 K.Mass-spectrometric analysis shows that the major decomposition products were CO, and CO, but some water vapour was also present. The ratio of CO to CO, changed markedly as the activation temperature was increased above 1073 K, with CO becoming the dominant decomposition product at the higher temperatures, presumably because of the onset of a dissociation of residual surface CO, upon sites exposed at the higher temperatures. Progress of the thermal decomposition of SrCO, and its conversion to SrO is also shown in fig. 2, which presents a summary of i.r. spectra obtained at 300 K from a self-supporting disc, following activation in vacuo at the indicated temperatures. Clearly evident is the progressive diminution of i.r.bands (at 2400, 1780 and 1090 cm-l) characteristic of bulk Cog- and the appearance of bands (at 1000-800 cm-l) characteristic of bulk Particular interest attaches, however, to the temperature needed for total removal of bands at 1090 and 1460 cm-l, which previous workers have identified with Cog- in the aragonite Our observation that these were not totally removed until vacuum outgassing reached 1173 K make it necessary to take into account the persistence of some SrC0,-like regions to considerably higher temperatures than might be suggested (ca. 1073 K) by the mass-spectrometric data in fig. 1. Another significant feature of the i.r. studies, which is not depicted in fig. 2, was the lack of evidence to indicate extensive hydroxylation at any of the stages represented in fig.2. Effects of prior outgassing at different temperatures on the emission and excitation spectra of SrO at room temperature are shown in fig. 3. The excitation and emission spectra appeared only on outgassing above 1073 K, which fig. 1 showed to correspond to decomposition of the bulk carbonate, Further increases in outgassing temperature brought about a progressive increase in the emission-peak intensity and an accom- panying increase in the emission-peak intensity, which fig. 2 would suggest to coincide with removal of the last traces of Cog-. Increasing the outgassing temperature also brought about significant changes in the peak shape and position. Thus in the excitation spectra, both a peak at 280 nm and an accompanying shoulder at 3 15 nm were observed to grow and experience changes in relative intensity on going from 10731 0 0 n E 80 8 + 60 LO 20 J.NUNAN, J. A. CRONIN AND J. CUNNINGHAM 3.0 I- 203 1 Fig. 1. Mass-spectrometric observations on the SrCO, -+ SrO conversion in uucuo as a function of outgassing temperature: A, total pressure of gaseous decomposition products; , combined peak heights of (CO+CO,); 0, ratio of CO to CO, peak heights. I I I I I I I I I I :OOO 2500 2000 1800 1600 1400 1200 1000 800 600 v/cm-' Fig. 2. Infrared spectra (at 300 K) of SrCO, + SrO coqversion after vacuum outgassing for 1 h at (a) 293, (b) 973, (c) 1073 ahd (4 1173 K.2032 SURFACE REACTIVITY AND LUMINESCENCE OF SrO 1 1 I 2 5 0 3 0 0 350 X/nm Fig. 3. (a) Excitation spectra and (b) emission spectra of high-surface-area strontium oxide obtained by outgassing SrCO, at various temperatures from 1073 to 1293 K: (i) 1 h at 1073 K, (ii) 1 h at 1123 K, (iii) 1 h at 1123 K, (iv) 1 h at 1173/1273 K and (v) 4 h at 1293 K.to 1273 K [cf. fig. 3(a), plots (i)-(iv)]. Upon outgassing for 4 h at 1293 K the relative intensity in the latter shoulder was observed to decrease again [cf. fig. 3(a), plot (v)]. Parallel observations upon the emission spectra of the same sample showed that outgassing at 1 123 K led to two broad peaks centred at 400 and 455 nm [cf. fig. 3 (b), plot ($1. Outgassing at the higher temperatures of 1173, 1223 and 1273 K resulted in a single peak at 465 nm, as shown in fig. 3 (b), plots (iii) and (iv). However, extensive outgassing at 1293 K for 4 h led to a shift in the emission maximum back to 450 nm and a reduction in its intensity of ca.40% [fig. 3(b), plot (v)]. Fig. 4 illustrates the effects of exciting the emission at different excitation wavelengths for a SrO saqple first outgassed at 1123 K and then at 1293 K. It is clear that for the sample pre-treated in this manner the emission-peak shape and position remained independent of the exciting wavelength. This was also found to be the case for samples outgassed at 1173, 1223 and 1273 K. These excitation and emission results agree well with those reported by Coluccia et aZ.9b In excitation the band having a maximum at 280 nm, similar to that previously attributed by Coluccia et aZ. to five-coordinate surface positions, appeared first. The other weaker band with a maximum at 3 15 nm, similar to that attributed to Coluccia et al.to four-coordinate surfaces locations, only developed as the outgassing temperature increased towards 1273 K. In emission, the bands shown in fig. 3(b) as the first to develop in our spectra, with maxima at 400 and 450 nm, are very similar to those assigned by Coluccia et al. to emission from surface locations involving five- and four-fold coordination, respectively. The development of the emission with a maximum at 465 nm only after outgassing at the higher temperatures, as illustrated by fig. 3(b), plot (iii), had previously been interpreted in terms of the involvement ofJ. NUNAN, J. A. CRONIN AND J. CUNNINGHAM 2033 ( a ) (iii) f : I . , . ... 1 1 1 1 1 350 400 450 500 550 Xlnm ( b ) (iv) ,.#-*\ i '\.i '. ; \ I 1 1 1 1 I 350 400 450 500 50 600 X/nm Fig. 4. Emission spectra of SrCO, after outgassing at (a) 1073 and (b) 1293 K, as excited by different wavelengths of exciting light: (i) R = 280 nm, (ii) R = 299 nm, (iii) ;Z = 315 nm and (iv) ;Z = 325 nm. surface ions with three-fold coordination. Furthermore, our observation that, regardless af whether excitation of SrO outgassed at the higher temperatures was made at 280 or 315 nm, emission was dominated by the band at 450 nm reproduces observations made earlier by Coluccia et al. and attributed by them to energy transfer from sites of five-fold coordination (excited by photons having R = 280 nm but emitting at 400 nm) to sites of four-fold coordination (excited by photons having R = 315 nm but emitting at 450 nm).It was encouraging that our spectral parameters for luminescence are qualitatively so similar to those of Colluccia et al., despite differences in starting materials and steps for its conversion to SrO. However, details of the dependence of luminescence upon temperature of sample outgassing, which are also contained within fig. 2-4, will, when compared with the temperature dependence of the surface reactivity/catalytic activity (see below), serve to call into question the adequacy of relying upon varying relative exposures of O;;, 0:; and 0:; as the sole arbiter of surface properties. SURFACE REACTIVITY The stoichiometry of the nitrous oxide decomposition at reaction temperatures of 673-873 K over SrO samples pretreated as outlined in the experimental section agreed within experimental error with the reaction N2O(g) + Ndg) + P2(g)- (3) The reaction exhibited first-order kinetics, whether studied by the mass-spectrometric (m.s.) procedure or by the gas-chromatographic (g.c.) procedure.In fig. 5 (a) are shown the first-order plots of N20 decomposition at 627,656 and 683 K obtained using mass spectrometry. Linear regression analysis gave correlation coefficients ranging from 0.992 to 0.997 for the three plots. In the gas-chromatographic continuous-flow studies, 67 FAR 12034 SURFACE REACTIVITY AND LUMINESCENCE OF SrO 0.74 I I I I 20 40 60 80 tlmin -13.8 - 1 3 . 6 - 1 3 . 4 Q) +d $ - 13.2 - -13.0 - 1 2 . 8 - 12.6 '\ 4.3 L.5 4 . 7 4 . 9 5.1 5.3 5 . 5 5.7 In (PN201TOrr) Fig. 5. Evidence from kinetic analysis for first-order character of N,O decomposition at 700 +_ 50 K following preactivation of the SrO material at higher temperatures.(a) First-order plots of data obtained by mass-spectrometric monitoring of N,O decomposition in a static reactor over SrO preactivated at 1133 K; PNZ0 (initial) = 2.5 Torr, reaction temperatures: 0, 627; A, 656 and U, 673 K. (b) Plot of In rate against In PNzO for steady-state decomposition in a continuous-flow reactor operated in the differential mode at 758 K over SrO preactivated at 1273 K for 6 hr in an argon flow. Inset shows plot of steady-state conversion against reciprocal space velocity at 740 K for the same sample.J. NUNAN, J. A. CRONIN AND J. CUNNINGHAM 2035 kinetic data were obtained under conditions where the reactor was operating in the differential mode.This was achieved by selecting the reaction temperature, catalyst mass and flow conditions such that the conversions were < 4%. Plots of the reciprocal space velocity against conversion within this range were found to be linear, as shown by the insert on fig. 5(b). Consequently the percentage conversion gives a direct measure of the rate of reaction. Applying the method of plotting In rate against In PNPO, the reaction was shown to be first order with respect to N,O pressure in the pressure range 50-250 Torr, as illustrated in fig. 5(b). In this manner first-order kinetics in N,O decomposition were shown to be obeyed over SrO (ex SrCO,) preactivated at various temperatures in the temperatures range 773-1 273 K, indepen- dent of whether the reaction was studied using m.s.or g.c. procedures. The effect of outgassing temperature on the reaction rate is shown in fig. 6, where the first-order rate constants obtained using g.c. are plotted against outgassing temperature. It is evident from fig. 6 that, while some steady-state activity for N,O decomposition became apparent after outgassing at 900 K, the rate only began to increase rapidly after outgassing at temperatures > 1073 K, i.e. in the range where fig. 2 indicated removal of residual traces of COi-. The use of the gas-chromatographic procedure in its pulsed-reactant mode allowed investigatior s to be made of any progressive changes brought about in sample activity by contacting a preactivated SrO sample with a succession ofindividual pulses of nitrous oxide.In this way low-exposure activity profiles (1.e.a.p.) of surface activity could be developed. Fig. 7 shows how such 1.e.a.p. plots varied with increasing pulse number (each pulse ca. 4 x mol of N,O) for different temperatures. Comparison of the three 1.e.a.p. profiles, all measured at an identical reaction temperature, demonstrate two important points: firstly, the activity maximum attained in a profile was lowest after preactivation in vacuo at 993 K and was increased by increased preactivation temperature, and secondly, although the drop-off in activity after the maximum was quite abrupt for samples preactivated at 993 or 1083 K (so that activity declined rapidly towards zero after a small number of pulses), no such abrupt drop-off in activity was found for the sample preactivated at 1273 K, but rather the activity for N,O decomposition continued at a high level up to large pulse numbers.The first of these points agrees well with existing hypotheses that the level of surface activity depends upon the extent to which more highly coordinatively unsaturated surface ions are exposed and/or developed at progressively higher preactivation temperatures. An adequate explanation of the second point requires some extension and modification of existing hypotheses, which is attempted in the Discussion. Such extension and modifications will also be relevant to observations made upon the ratio of N, to 0, products from N,O pulses delivered in the experiments depicted in fig.7. These showed that, whilst at any of the selected temperatures the N, to 0, product ratio did not vary significantly with pulse number, the ratio took the very different values of 2.0, 3.0 and 3.8 during sequences of pulses introduced for samples previously outgassed at 993, 1083 and 1273 K, respectively. Oxygen isotope exchange was studied using mass spectrometry.la After preactivation of the SrO (ex SrCO,) samples at adequate temperatures the reaction occurred readily at room temperature. This is illustrated in fig. 8, where the mole fractions of 1602, 160180 and 180, are plotted as a function of time. During the course of the reaction the atom fractions of l60 and l80 remained constant in the gas phase, thus indicating that an &-type oxygen-isotope-exchange process was occurring, i.e.l6O,(g) + l80,(g) e 21601SO(g) (4) The effect of temperature of prior activation upon the level of activity for &-type without significant accompaniment of an R,- or R,-type exchange with lattice l60. 67-22036 8 7 6 n 5 E S f A 4.4 .& 3 2 1 2, i ./. .-.--.--.- 723 823 923 1023 1123 1223 1323 Fig. 6. Influence of temperature of prior activation of SrO samples on the steady-state rate of N,O decomposition at 758 K in the continuous-flow g.c. procedure with PNzO = 380 Torr. 40 t 4 2 0 - \ \ 7 ‘.L1 10 - 0 1 0 2 4 6 8 10 N20/10-6 mol Fig. 7. Plots of low-exposure activity profiles (1.e.a.p.) of SrO samples, obtained by measurement in the pulsed-reactant g.c. procedure. These show the extent of non-steady state N,O decompo- sition at 758 K for each of a succession of individual pulses; PNs0 = 40 Torr.The three profiles were obtained for SrO samples preactivated at 0, 993; ., 1083 and A, 1273 K.J. NUNAN, J. A. CRONIN AND J. CUNNINGHAM - - - - - 2037 0.5 h E 0 0.L 2 It: 0.3 2 0 2 0 v 0.2 5 0.1 h c 0 . 0 I 1 1 1 1 1 1 1 t l s 0 10 20 30 10 50 60 Fig. 8. Progress of the &-type oxygen-isotope-exchange equilibration process at room temperature on admitting an isotopically non-equilibrated equimolar mixture of 1 6 0 2 + I S 0 2 into contact with an SrO sample preactivated in vacuo at 1223 K. Plots show the mole fraction of 1 6 0 2 and 1802 decreasing with time (a) and the mole fraction of 1601S0 increasing with time (m), in agreement with reaction (4). oxygen-isotope exchange subsequently measured at room temperature is clearly shown in fig.9. There the reciprocal of z, the time required for surface-assisted movement by 50% towards total equilibration, is used as a measure of rate of the isotopic equilibration process and is plotted as a function of prior outgassing temperature. For a constant preactivation period of 1 h at each temperature, fig. 9 indicates a rather sharp onset of activity after outgassing at 1023 K, followed by an essentially linear increase for higher outgassing temperatures. Fig. 9 represents for purposes of comparison the dependence upon temperature of preactivation observed in a related set of experiments where the magnitude of the first-order rate constant for N,O decomposition at 758 K was determined after 1 h preactivation of the SrO at various temperatures.Although the same static-reactor and mass-spectrometric method of analysis were employed for both sets of experiments, it is clear that the onset of activity is more sharply defined for the oxygen-isotope-exchange plot and that its initial slope is much steeper than for N,O decomposition. DISCUSSION The above results show the considerable success that there has been in the search for similarities between dependences upon temperature of prior activation in the development of surface photoluminescence and of surface reactivity. Such a correlation is most direct in the case of the &-type oxygen-isotope-exchange reaction, since surface luminescence and surface activity were both measured under similar conditions, i.e. at room temperature after cooling down from the higher preactivation temperatures. Furthermore, it is well established from previous r e p o r t ~ ~ - ~ of quenching of the surface luminescence by molecular oxygen that 0, species interact reversibly at 300 K with2038 SURFACE REACTIVITY AND LUMINESCENCE OF SrO 35 30 25 - I 0 v) 20 - 1 .y 15 10 5 800 1000 1200 TIK Fig.9. Influence of temperature of prior activation of SrO samples upon the rates of N,O decomposition (m) and of &-type oxygen isotope exchange (0) as monitored by mass spectrometry. The kinetic parameters utilised as measures of the rates of the processes were, respectively, the slopes of plots similar to those in fig. 5(a) and the time to 50% equilibration in plots similar to those in fig.8. the centres responsible for luminescence, even at the low pressures ( 10-1-10-2 Torr) employed in obtaining the surface reactivity data shown in fig. 8. The fact that the temperature profile for development of oxygen-isotope-exchange (cf. fig. 9) rises rapidly across the same temperature range (1073-1273 K) as was required for rapid enhancement of surface luminescence provides good support for the idea that a relationship exists between sites involved in surface photoluminescence and those conferring catalytic activity for oxygen-isotope exchange in accordance with reaction (4). It has been deduced from previous studies1*919 of the latter process upon ZnO surfaces that (a) the catalytically active sites involve surface cations at locations of high coordinative unsaturation and (b) the ready availability of at least one electron from a shallow energy level or trap in the immediate vicinity of the M& locations is also necessary to trigger oxygen isotope exchange (via an electron-transfer-catalysed chain reaction requiring negligible energy of activation).Both of these requirements seem likely to be satisfied by the Of;. . . F,+ surface locations proposed by D ~ l e y ~ ~ as being responsible for surface luminescence from high-surface-area MgO and CaO, since (a) the Sr2+ cations involved in the F,+ point defect necessarily have at least two degrees of coordination unsaturation and (b) relatively low energies can be expected for transfer of an electron from a ground-state 0;; or the adjacent F:. According to D ~ l e y , ~ ~ photoinduced transfer of an electron between these is responsible for a broad absorption centred upon 2 eV and detectable in MgO previously exposed to high-energy radiations in vacuo.Still lower energy could be expected for thermal,J. NUNAN, J. A. CRONIN AND J. CUNNINGHAM 2039 rather than optical, promotion of an electron from Of; to an adjacent F: for MgO. Further reductions in energy could be anticipated for SrO relative to MgO, and for Ot; . . . F,+, , locations relative to Of;. . . F,+ locations. * Thus it is reasonable to envisage oxygen-isotope-exchange upon SrO proceeding, in analogous fashion to that previously detected for ZnO surfaces, via activation of oxygen under the combined influence of a readily available electron and coordinatively unsaturated cation(s).Viewed in this context, the experimentally observed dependence of the extent of room- temperature oxygen-isotope-exchange activity of the SrO (ex SrCO,) surfaces upon temperature of prior activation (cf. fig. 9 ) should reflect the increase in number of 0:;. . . F,+, , surface locations with preactivation temperature (n = 1 or 2). The exact implications of the good agreement with first-order kinetics observed here [cf. fig. 5(a) and (b)] for decomposition of N,O at 758 K clearly depend upon the mechanism of the metal-oxide-assisted decomposition, and in particular upon which step is the rate-determining process. Electron transfer from active sites on the metal oxide surface to N,O adsorbate has been proposed20* 21 by many workers as the initial step.However, the rate-determining role has often been assigned to a step by which the necessary reverse electron transfer is accomplished.22 We may denote these electron-transfer processes by N,O(ads) + e-D, -+ N,O-(ads) -+ N2(g) + 0-(ads) D, ( 5 a) followed by O-D, + N,O(g) -+ N,(g) + 0, + e-D,. Note the adoption at this stage of the non-commital notation D, for surface locations with which an electron or an oxygen fragment can be loosely associated (as in e-D, and 0-D, respectively) and which are envisaged to play important roles in N,O decomposition. However, the ability of such a two-step mechanism to account for the pseudo-steady-state rate of N,O decomposition in the continuous-flow mode is largely independent of the nature of D,. Thus the operation of reaction (5b) as the slow rate-determining step in conditions such that available D, sites were effectively saturated by 0- would account for first-order dependence upon PNZO.Furthermore, the experimentally observed dependence of the pseudo-steady-state rate of decom- position on the temperature of preactivation (cf. fig. 6) may arise because the available surface concentration of the entity 0-D, in reaction (5b) was predetermined by the surface concentration, ODs, of active sites developed by prior activation. In the context of this two-step mechanism and of the similarities between the experimentally observed profiles for development of activity for N,O dissociation and oxygen- isotope-exchange (cf. fig. 9) it is of interest to consider whether the e-D, sites envisaged in reaction (5 a) may be similar to the 0;;.. . FZ, , locations favoured earlier as the active sites for oxygen-isotope-exchange on SrO. At first glance the fact that the profiles in fig. 9 do not reveal three distinguishable segments, corresponding to temperature regimes dominated by the growth of Ot;, 0;; and 0;; as envisaged by others in temperature profiles for development of photoluminescence features, might be construed as an argument against equating e-D, with the 0:;. . .F,+, 0;;. . . F,+ and 0:;. . . F: locations envisaged by Duley. However, an alternative interpretation which cannot yet be overruled and which would permit equivalence between e-D, and 0:;. . . F,+, sites, is that the electron transfer required to initiate N20 dissociation [cf.reaction (5 a)] or oxygen-isotope-exchange [cf. ref. (18)-(20)] can occur to adsorbed reactant with comparable facility from 0:;. . . Fi, s , 0;;. . . FA., , and 0;;. . . F:, , locations, thereby rendering molecular probes incapable of distinguishing between * Note added inproof: OkadaZ3 has recently associated the absorption at 2.17 eV in MgO with Ff centres.2040 SURFACE REACTIVITY AND LUMINESCENCE OF SrO such locations. In such,a situation, some differences between the degree of detail revealed by temperature profiles for development of surface activity and of surface photoluminescence could be expected. The value observed experimentally for the ratio of N, to 0, (between gaseous N, and 0, products detected as well formed peaks in the g.c. technique) represents a further criterion for assessing the mechanism involving reactions (5a) and ( 5 b).Thus the value of 2.0 observed for this ratio in the continuous-flow conditions under PNPO from 50 to 380 Torr agreed well with the mechanism. Note, however, the very significant upward deviation of this ratio observed in the pulsed-reactant mode whenever the temperature of preactivation of SrO was progressively increased (values of 2.0, 3.0 and 3.8 after preactivation at 993, 1083 and 1273 K, respectively). Such deviation would be consistent with enhanced competition, under the conditions prevailing in the pulsed-reactant mode, between the above two-step mechanism with reaction (5b) as the slow oxygen-producing step and another process which instead incorporated fragments from N,O decomposition into the oxide surface.Likely processes of the latter type are N,O(g) + mSrfz -+ N,(g) + OrnSrt: (6 b) which involve incorporation of an oxygen fragment from N,O into an anion vacancy doubly occupied by electrons, 2e-0,, or by surface groupings of m low-coordinate cations. Some approach towards equilibrium concentrations of point defects at the surface is thermodynamically expected during the preactivation treatment, greater con'centrations being expected during preactivation at high temperatures. A ' freezing- in' of some fraction of such defect concentrations during the rapid cooling to 758 K would result in the probability of reactions (6a) and/or (6b) occurring on contact with the initial pulses of N,O admitted in the pulse-reactant mode.The extent to which this could drive the N, to 0, ratio of the 1.e.a.p. plots away from the value of 2:l expected if only reactions (5a) and (5b) were significant would, as observed in the pulsed-reactant experiments, be greater after higher temperature of preactivation because of freezing-in of a greater number of defects. For the continuous-flow conditions only a low survival probability would be expected for such defects after the much higher integrated exposures (typically 2 x lo6 Torr s) to N,O which preceded measurement of steady-state activity for N,O decomposition. This would account for the negligible deviation of the N, to 0, ratio from 2.0 in the continuous-flow mode regardless of the sample temperature during preactivation. Deviation of the ratio from 2.0 during the determination of the 1.e.a.p.plots, with the extent of deviation observed to increase with the temperature of prior activation, contrasts with the steady state results and implies that surface defects at the SrO surface have a major role at low exposures. This work was supported by the U.S. Air Force Office of Scientific Research under grants AFOSR 82-0023 and 83-0074. We also thank the Department of Education of the Irish Government and University College, Cork for support (J.A.C.). (a) J. E. Ely, K. J. Tergarden and D. B. Dutton, Phys. Rev., 1959,116, 1099; (b) W. H. Hamill, Phys. Rev., 1969, 185, 1182. * (a) D. L. Dexter and R. S. Knox, Excitons (Wiiey Interscience, New York, 1965); (b) T. Higashimura, Y. Nukaoka and T. Iida, J .Phys. C, 1984, 17, 4127. B. Henderson and J. E. Wertz, Ado. Phys., 1978, 17, 749. V. M. Bermudez, Prog. SurJ Sci., 1981, 11, 1 .J. NUNAN, J. A. CRONIN AND J. CUNNINGHAM 204 1 (a) W. W. Duley, Philos. Mag. B, 1984,49,159; (b) K. Teegarden, in Luminescence of Inorganic Solids, ed. P. Goldberg, (Academic Press, New York, 1966). R. C. White1 and W. C. Walker, Phys. Reu., 1969, 188, 1380. A. Zecchina, M. G. Lofthouse and F. S. Stone, J. Chem. SOC., Faraday Trans. I , 1975,71, 1476. E. Garrone, A. Zecchina and F. S. Stone, Philos. Mag. B, 1980, 42, 683. (a) A. J . Tench and G. T. Pott, Chem. Phys. Lett., 1974, 26, 59; (b) S. Collucia, A. M. Deane and A. J. Tench, J. Chem. Soc., Faraday Trans, 1978,74,2913; (c) S. Collucia and A. J. Tench, J. Chem. SOC., Faraday Trans 1, 1983, 79, 1881. lo J. Nunan, J. Cunningham, A. M. Deane, E. A. Colbourn and W. C. Mackrodt, in Adsorption and Catalysis on Oxide Surfaces, ed. M. Che and G. Bond (Elsevier, Amsterdam, 1984). l 1 (a) D. Cordischi, V . Indovina and M. Occhiuzzi, J . Chem. SOC., Faraday Trans. I , 1978, 74, 456; (b) V . Indovina and D. Cordischi, J. Chem. SOC., Faraday Trans. I, 1982,78, 1705. l2 (a) M. J. Baird and J. H. Lunsford, J. Catal., 1972,26,440; (6) Y. Tanaka, H. Hattori and K. Tanaka, Chem. Lett., 1976, 37. l3 (a) H. Hattori and A. Satoh, J. Catal., 1976, 45, 32; (b) M. Mohri, K. Tanabe and H. Hattori, J. Catal., 1974, 32, 144. l4 (a) H. Hattori, K . Marayama and K. Tanabe, J. Catal., 1976,44, 50; (b) M. Mohri, H. Hattori and K. Tanabe, J. Catal., 1974,32, 144. l5 (a) L. Parrot, J. W. Rogers Jr and J. M. White, Appl. Surf. Sci., 1978, 1, 443; (b) H. Noeller and K. Thomke, J. Mol. Catal., 1979, 6, 375. l6 J. Cunningham, B. K. Hodnett, M. Ilyas, J. P. Tobin and E. L. Leahy, Faraday Discuss. Chem. SOC., 1981, 72, 283. l7 K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, (Wiley Interscience, New York, 3rd edn, 1978); (b) S. Pinchas and I. Ilaalicht, Infrared Spectra of Labelled Compounds (Academic Press, New York, 1971). (a) J. Cunningham and E. L. Goold, J. Chem. SOC., Faraday Trans. I , 1981, 77, 837; (b) J. Cunningham, E. L. Goold and J. L. G. Fierro, J. Chem. SOC., Faraday Trans. I , 1982, 78, 785. lD J. Cunningham, in Comprehensive Chemical Kinetics, ed. C. F. Tipper and C. H. Bamford (Elsevier, Amsterdam, 1984), vol. 19, p. 362. 2o J. Cunningham, J. J. Kelly and A. L. Penny, J. Chem. SOC., Faraday Trans. I , (a) 1970, 74, 1992, (b) 1971, 75, 617. 21 N. B. Wong, Ben Taarit and J. H. Lunsford, J. Chem. Phys., 1974, 60, 2149. 22 (a) F. S . Sione, J. Solid State Chem., 1975,12,271; (6) G. Blyholder, J. Phys. Chem., 1971,75, 1037. 23 M. Okada, Cryst. Lattice Defects and Amorphous Materials, 1985, 11, 73. (PAPER 4/ 1579)