Luminescence and Other Spectroscopic Studies of the Reaction of Pyridine and Oxygen with Thermally Activated SrO BY SALVATORE COLUCCIA,~ JEAN F. HEMIDY $ AND ANTHONY J. TENCH* Chemistry Division, A.E.R.E., Hanvell, Oxfordshire OX1 1 ORA Received 16th March, 1978 The reaction of pyridine and O2 on SrQ powder has been studied using several different spectro- scopic techniques to characterise the intermediate and the final products. Reflectance spectroscopy has been used to demonstrate the formation of the 4,4'-bipyridyl negative ion when pyridine is adsorbed on SrO. Subsequent adsorption of oxygen leads to the formation of the 0; ion by a one-to-one electron transfer process from the organic ion and photoluminescence studies confirm that the 4,4'-bipyridyl molecule is also formed on the surface when the electron transfer occurs.During recent years there has been considerable interest in electron transfer at oxide surfaces because of its close relation with various aspects of catalysis. Under normal conditions, there is little evidence that the alkaline earth oxides, when prepared in vacuu, will transfer electrons to oxygen without preirradiation. However, Iizuka 1 and Che et aL3 have reported formation of 0 2 on MgO and CaO when oxygen has been adsorbed on a surface which has previously been pretreated with pyridine (Py) or bipyridyls (Bpy) and recently it has been shown that, under similar conditions, 0; can be formed on Sr0.4 Up to now attention has been devoted mainly to the e.s.r. signal of the oxygen species. In this work we have extended the study of the reaction on SrO using optical absorption and luminescence techniques to characterise the intermediate organic species both in the anionic (before contact with 0,) and neutral (after contact with 0,) forms.Using e.s.r. we have shown that there is a one-to-one electron transfer from the organic anion to the adsorbed oxygen. EXPERIMENTAL Strontium oxide was prepared by thermal decomposition of strontium carbonate (Hopkins and Williams G.P.R. or Johnson Matthey Spec-pure) to give a polycrystalline sample with a specific surface area of N 10 m2 g-1.5 During the decomposition a zeolite trap at 77K was used to remove water vapour and carbon dioxide. The samples were sealed off after evacuation at a temperature of 1200 K to a pressure of < 1.5 x Pa Torr).The samples were prepared in a silica bulb with a break seal ; a side arm and a rectangular silica cell allowed the same sample to be studied by electron spin resonance, reflectance spectroscopy and photoluminescence. Pyridine (B.D.H.) was purified by the freeze-pump-thaw technique and was adsorbed on the sample under its room temperature vapour pressure. Spec-pure oxygen was used without further purification and adsorption was carried out under a pressure of 0.75 kPa (5 Torr). 4-4' Bpy (B.D.H.) and 2-2' Bpy (Hopkins and Williams) were used without further purification and adsorption on the sample was carried out by exposure to the sublimation pressure at 300 K ; the adsorption process was considered complete when the whole sample presented a homogeneous blue colour. t On leave from University of Turin.§ On leave from University of Caen. 27632764 E.s.r. spectra were recorded at 300 and 77 K using a Varian V4502 spectrometer with 100 kHz field modulation. Spin concentrations were calculated by direct double integration using an on-line computer system 6* and g values were measured relative to Cr3+ in MgO (g = 1.9797). The optical reflectance spectra were recorded on a Cary 14 spectrophotometer using magnesium carbonate as a reference.8 The photoluminescence spectra (at 300 and 77 K) were obtained using a 250 W Xenon lamp as an excitation source and the excitation wavelengths were selected using a double monochromator. An Ortec photon counting system was used to compare the photon pulse rates from two photomultipliers and the spectra were corrected for variations in excitation intensity with wavelength and time.A Corning 3-74 filter (cut-off at -400 pm) was used before the emission monochromator to eliminate scattered exciting light. REACTION OF PYRIDINE AND O2 WITH SrO RESULTS PHOTOLUMINESCENCE SPECTRA OF THE SYSTEM SrO+Py+O, The SrO samples prepared in vacuo were white and their luminescent spectra at 300 K showed a broad emission centred at about 475 nm [fig. l(u)J and a correspond- ing excitation spectrum with two maxima at 280 and 315 nm [fig. I@)]. The samples turned blue-violet on adsorption of pyridine and only very weak photoluminescence nm FIG. 1 .-Photoluminescence spectra of the SrO + Py + O2 system at 300 K : (Q) emission spectrum of SrO, Zx 1 ; (b) excitation spectrum of SrO, I x 1 ; (c) emission spectrum after adsorption of Py, Zx 250 ; ( d ) excitation spectrum after adsorption of Py, Zx 250.spectra were detected at 300 and 77 K with maxima for the emission at 475 nm [fig. l(c)J and for the excitation spectra at 280 and 315 nrn, [fig. l(d)]. Outgassing the excess Py at 300 K did not modify the spectra nor the colour of the sample. Admission of O2 at 300 K gave a colour change from blue-violet to yellow-pink and the emission spectrum excited at 280 nm showed a new maximum at 550 nm and [compare fig. 2 ( 4 and (b)] a slight increase of the intensity in the 400-450 nm region. The emission after the addition of oxygen [fig. 2(c)] measured at 77 K was much more intense than at 300 K and showed a structure of three main peaks at 429, 457 and 476 nm withS .COLUCCIA, J . F . HEMIDY AND A . J . TENCH 2765 shoulders at 450,490 and 515 nm ; the corresponding excitation spectrum [fig. 2(d)] had a maximum at 275 nm. Outgassing the excess of O2 at 300 K did not modify the spectra significantly. All the spectra of fig. 1 and 2 refer to the same sample and have been run in succession without moving the cell. The relative intensities of the spectra can be compared with reasonable confidence although the absolute intensities are not known. I------- nm nm FIG. 2.-Photoluminescence spectra of the SrO + Py+ O2 system. The sample is the same as for fig. 1 : (a) emission spectrum excited at 280 nm at 300 K after adsorption of Py, I x 250 ; (b) emission spectrum excited at 280 nm at 300 K after adsorption of O2 on preadsorbed Py, Zx 250 ; (c) as (b), but at 77 M, I x 12.5 ; (d) excitation spectrum of (c) Zx 12.5.nm FIG. 3.-Photoluminescence spectra of the SrO + Bpy + O2 system at 77 K : (a) emission spectra after adsorption of O2 on preadsorbed 4,4‘-Bpy; (b) excitation spectrum of (a); (c) emission spectrum after adsorption of O2 on preadsorbed 2,2’-Bpy ; (d) excitation spectrum of (c).2766 REACTION OF PYRIDINE AND Q2 WITH SrO PHOTOLUMINESCENCE SPECTRA OF THE SYSTEM SrOf BIPYR1DYLSfo2 Adsorption of either 2,2‘-Bpy or 4,4’-Bpy on the surface of SrO turned the samples blue and quenched the original emission. The blue colour was destroyed by exposure to O2 (5 Torr) and the samples turned yellow-pink. The photoluminescent spectra at 77 K were recorded after contact with 0, (fig. 3), for comparison with the equivalent spectra obtained when Py reacts with 0, on SrO.The emission spectrum excited at 276 nm when preadsorbed 4,4’-Bpy was contacted with O2 [fig. 3(a)] showed peaks at 440, 470 and 490 nm and shoulders at 459 and 51 5 nm, while the corresponding excitation spectrum has a maximum at 276nm. The emission spectra excited at 280 nm when preadsorbed 2,2’-Bpy was contacted with 0, on SrO [fig. 3(c)] showed a sharp peak at 490nm and a broad complex peak at 520nm with a shoulder at 560 nm. In addition it had a peak at 418 nm, whose real maximum could be at shorter wavelengths ; it is likely that we observed only the tail of this peak because of the filter. The excitation spectrum showed a maximum at 280 nin [fig 3jd)J.m Y .- El ,.-- aoo 600 400 nm FIG. 4.--Kubelka-Munk calculation on the reflectance spectra of SrO : (a) adsorption of pyridine ; (b) adsorption of pyridine followed by adsorption of 02. The photoluminescent spectra of the bipyridyls as mol dm-3 solutions in C2H50H have been run at 77 K for comparison with the earlier data. Under these conditions the 2,2‘-Bpy has an excitation maximum at 280 nm and an emission spectrum with bands at 427, 456 and 476 nm together with shoulders at 450, 490 and 515 nm. The 4,4’-Bpy has an excitation maximum at 275 nm and an emission spectrum with bands at 432 and 456 nm together with shoulders at 475 and 490 nm. REFLECTANCE SPECTRA OF THE SYSTEM SrO+Py+O, admission of O2 [fig. 4@)] have been transformed using the Kubelka-Munk function The reflectance spectra of SrO contacted with Py both before [fig.4(a)] and afterS . COLUCCIA, J . F . HEMIDY AND A . J . TENCH 2767 to give an absorbance scale. The spectrum before oxygen admission showed maxima at 254, 395 and 595 nm. After oxygen admission the two bands at 395 and 595 nm disappeared, the intensity of the band at shortest wavelength increased and the maxi- mum moved to 260 nm ; in addition a new band at 475 nm was evident. E . S . R . SPECTRA OF THE SYSTEM SrO+Py+O, The e.s.r. spectrum of SrO after adsorption of pyridine was a single isotropic line at g = 2.003 with a line-width which varied from 9-14 G depending on the sample. After adsorption of oxygen, the e.s.r. line disappeared and was replaced by a new spectrum (fig.4). The line shape of this new spectrum is characteristic of a species in an orthorhombic crystal field and at 77 K the principal g values were g, = 2.002, g, = 2.007 and g3 = 2.100. FIG. 5.43.s.r. spectrum at 77 K of a SrO sample treated with pyridine and then exposed to contact with 0 2 . Measurements of the spin concentration before and after the adsorption of oxygen indicated that the two paramagnetic species were present at essentially identical concentrations ("7 x 1017 spins g-I). From this evidence it is likely that an electron transfer reaction has occurred between the pyridine-type radicals and the oxygen molecules. DISCUSSION ADSORBED SPECIES The absorption and subsequent emission of light from SrO powders is very similar to that observed for high surface area MgO where it has been interpreted as arising from lattice ions on the surface in positions of unusually low coordination.A more extensive investigation has confirmed that this interpretation applies to SrO.ll Adsorption of Py does not significantly change the shape of the emission and excitation spectra of SrO but the intensities are much decreased. The remaining emission probably originates from a very small fraction of unreacted surface sites, but there is also the possibility that it originates from new surface species formed upon adsorption of Py. The emission spectrum of the Py sample is very weak and structureless even at 77 K. After admission of O2 on preadsorbed Py, the emission spectrum at 300 K is still very weak, but more complex. In contrast to the behaviour of the Py sample, lowering the temperature to 77 K affects both the intensity and the shape of the emission spectrum, increasing the intensity in the 450 nm region by 20 times.The very well defined structure, when compared with the data for bipyridyls2768 REACTION OF PYRIDINE AND O2 WITH SrO in frozen solutions and also adsorbed on the surface (fig. 3), strongly suggests that this emission is due to bipyridyl species formed on the surface. The emission spectra of 2,2’- and 4,4‘-Bpy in solution agree well with those reported by Gondo l2 for the free bipyridyls in different solvents and by other authors both for the free l 3 and the chelated 1 3 9 l4 2,2’-Bpy molecule. The clear similarities of the spectrum of fig.2 with the spectra of the free bipyridyls indicate that the product of the interaction of Py with O2 and SrO surface is a bipyridyl species. However, as the emission spectra of the solutions are very similar it would be difficult, on this basis alone, to decide which of the two bipyridyls is formed on the surface. But the emission spectra of the two bipyridyls are quite different when O2 is admitted to the molecules adsorbed on the surface of SrO, (fig. 3). The spectrum of 4,4‘- bipyridyl is very similar to the Py + 0, + SrO spectrum, both in the overall shape and in the spacings of the vibronic structure; in addition the excitation maximum gives better agreement with the maximum found for adsorbed 4,4’-Bpy (275 nm) than for 2,2’-Bpy (280 nm). This evidence leads us to conclude that the product of the inter- action of Py with 0, on SrO is 4,4’-Bpy.A full discussion of the emission spectra of the adsorbed bipyridyls is beyond the scope of this paper, but the difference in the spectra of the two molecules when adsorbed needs some comment. From fig. 3 it is clear that the spectrum of 4,4’-Bpy adsorbed on the surface is very similar to the spectrum of the frozen molecule; only a slight red shift and some variation in the relative intensities of the peaks is evident. In contrast, the spectrum of the adsorbed 2,2’-Bpy is quite different from that of the frozen molecules. This difference could be due to the different way in which the two molecules are likely t o interact with the surface sites. 2,2’-Bpy should be able to form a chelate species with both the nitrogen atoms of the two aromatic nuclei interacting with the coordinatively unsaturated surface cations, whilst the 4,4’-Bpy molecule can probably interact only with one nitrogen atom.The much stronger interaction of 2,2’-bipyridyl accounts for the large variation in the vibronic freedom of the molecule with respect to the free molecules, as has been observed in the i.r. spectra of complexe~.~’ The formation of 4,4’-Bpy by reduction of Py with alkali metals and subsequent oxidation has already been described 16-19 as well as the occurrence of the same reaction on the surfaces of alkaline earth oxides,2 but this is the first direct observation of the final product (neutral 4,4’-Bpy) on the surface. The formation of (4,4’-Bpy}- intermediate product has been suggested ; l 7 this negative species is characterised by two absorption bands at 380 and 580 nm.’ ‘ 9 For pyridine adsorbed on SrO we have found very similar values of 395 and 595 nm (fig.4) and this together with the initial strong e.s.r. signal indicates that the bipyridyl anion radical is formed on the surface of SrO. The e.s.r. spectrum does not show any hyperfine splitting and this can be explained by the large number of lines expected from so many nuclei with spin. The he-width variation from one sample to another probably indicates local concentration effects. In addition to the two bands of the anionic species, the reflectance spectrum of the Py+SrO samples show a weaker absorption of 254 nm that could well be due to an excess of Py still present on the surface after the short outgassing at 295 K.The two bands assigned to (4,4‘-Bpy}- disappear when O2 is admitted to the sample showing that an electron transfer process is taking pIace; the band at short wavelengths (260 nm) shows an enhanced intensity and this could be explained by the presence of the newly formed species on the surface, such as neutral 4,4’-Bpy and 0; which absorb in this region.2o* 21 After the adsorption of oxygen a reflectance peak at 475 nm is observed which is weak compared to the intensities of the 395 and 595 nm bands. It does not have the same thermal stability as the e.s.r. signal attributed toS . COLUCCIA, J . F. HEMIDY AND A . J . TENCH 2769 0 2 and may be associated with the formation of small amounts of polypyridyl species since similar bands have been observed in pyridine reduced by alkali metals.’ The e.s.r.spectra show that on adsorption of oxygen the organic radical species disappears and a new signal develops (fig. 5) with a g tensor very similar to that reported in literature for 0 2 22 and in particular for the alkaline earth oxides + Py + 0, systems, where I7O has been used to confirm the presence of OF, 3 p indicating that an electron transfer process takes place between the organic radical and O2 to form 0 2 adsorbed on the surface. 0 2 has been characterised in alkali halides 2o and in sodalite 21 systems by its emission spectra. In all these systems at 77 K a series of bands is observed centred at N 550 nm and extending from 400 to 650 nm, with a spacing be-tween the main peaks of - 1000 cm-l ; excitation maxima are quoted in the range 250 2o and 300 nm.21 The multiple band structure has been interpreted in terms of coupling with the vibration levels of the molecular ion.21 It is possible that there is some contribution from 0 2 species to the emission spectra seen in the O2 +Py+ SrO system, in particular the spectra at 300 and 77 K of fig.2. The spectrum at 77 K is clearly dominated by the emission in the 400-500 nm region and this has been assigned to 4,4’-Bpy ; no fine strlncture is observed at longest wavelengths where bands from the 0; spectrum would be expected. In contrast, the main feature of the spectrum at 300 K is a broad and complex emission peak centred at 550 nm and excited at 280 nm, which develops after admission of 02.A very similar spectrum at 300 K has been described for 0; in bromosodalite.21 It is possible that any contribution due to 0; adsorbed on SrO in the spectrum at 77 K [fig. 2(c)J, is hidden by the strongly temperature dependent emission of 4,4’-Bpy co-adsorbed on the surface. In addition energy transfer between different adsorbed species could strongly decrease the emission efficiency of OF. SURFACE SITES AND REACTION MECHANISM In the previous section it has been shown that adsorption of Py on SrO leads to the formation of the 4,4’-Bpy anion which can transfer an electron when 0, is adsorbed to give neutral 4,4’-Bpy and OF. The oxidation reaction to give the neutral molecule is straightforward. The formation of {2,2’-Bpy*}- and (4,4’-Bpy*}- when the bipyridyls are directly absorbed on the surface of SrO is indicated by the appearance of a blue colour which has been shown to occur on MgO as well using e.s.r.This demonstrates that the formation of the negative radical anions is not necessarily linked to a dehydrogenation stage and that electron donor sites exist on the surface. The intrinsic emission of the SrO is quenched when the Bpy and pyridine are adsorbed to form the blue anions, suggesting that the surface sites responsible for photoluminescence are involved in the electron transfer process. These sites have been described as ions on the surface in situations of unusually low coordina- tion.lo9 11* 23 We suggest that 02- ions on the surface in sites of low coordination have higher reactivity which is strongly enhanced by the lowering of the Madelung constant and are likely to be the electron donor sites on the surface.It must be stressed that the spin concentration quoted before shows that about 0.1 radical species are formed per 100A2, indicating that only the least stable 02- ions are able to take part in the electron transfer. Both the negative ions and the neutral molecules of Bpy are likely to be adsorbed on positive charged centres, i.e. on the cation Sr*+.2770 J. F. Hemidy acknowledges financial support from a European Exchange Fellow- ship of the Royal Society and provision of facilities by A.E.R.E., Harwell. S. Coluccia and A. J. Tench acknowledge financial support by NATO. The authors thank A. M. Deane for advice on experimental techniques and Prof.F. S. Stone for discussions. REACTION OF PYRIDINE AND O2 WITH SrO ' T. lizuka, Chem. Letters (Jupan), 1973, 891. T. Iizuka and K. Tanabe, Bull. Chem. Sue. Jupan, 1975,48,2527. M . Che, A. J. Tench, S. Coluccia and A. Zecchina, J.C.S. Faraduy I, 1976, 72, 1553. M. Che, S. Coluccia and A. Zecchina, J.C.S. Faraduy I, 1978, 74, 1324. R. L. Nelson, A. J. Tench and B. J. Harmsworth, Truns. Faraduy SOC., 1967, 63, 1427. A. M. Deane, C. Kenward and A. J. Tench, A.E.R.E. Report 7020. ' J. F. J. Kibblewhite and A. J. Tench, J.C.S. Faraduy I, 1974, '90, 72. * R. L. Nelson, J. W. Hale and B. J. Harmsworth, Trans. Furuduy Suc., 1971, 67,1164. G. Kortum, Reflexions Spektroscopie (Springer, Berlin, 1969). vol. 1, p. 171 ; A. J. Tench and G. T. Pott, Chem. Phys. Letters, 1974, 26, 590. lo S. Coluccia, A. M. Deane and A. J. Tench, Proc. Sixth Int. Congr. Cutulysis (London, 1976), ' ' S. Coluccia, A. M. Deane and A. J. Tench, J.C.S. Faraaizy I, in press. l 2 Y. Gondo and Y. Kaude, Bull. Chem. Soc. Japan, 1965,38,1187. l3 M . D. De Armond and J. E. Hillis, J. Chem. Phys., 1971,54,2247. l4 D. H. W. Carsteurs and G. A. Crosby, J. Mol. Spectr., 1970,34, 113. l5 J. S. Struckle and J. L. Walter, Spectrochim. Acta A, 1971, 27, 223. l6 R. L. Ward, J . Amer. Chem. Soc., 1961, 83, 3623. l7 A. Carrington and J. Dos. Santos Viega, Mol. Phys., 1962, 5,21. l 8 (a) J. W. Dodd, F. J. Hopton and N. S. Hush,Proc. Chem. SOC., 1962,61; (6) C. D. Schmulback, l9 V. Kalyanaraman, C. N. R. Rao and M. V. George, J. Chem. SOC. B, 1971, 2406. 2o J. Rolfe, F. R. Lipsett and W. J. King, Phys. Reu., 1961, 123, 447. 21 S. K. Deb and J. B. Gallivan, J. Luminescence, 1972,5, 348. 2 2 J. H. Lunsford, Catalysis Rev,., 1973, 8, 135. 23 A. Zecchina and F. S. Stone, J.C.S. Faraduy I, 1976,72,2364. C. C. Hinckley and D. Wasmund, J. Amer. Chern. Sac., 1968, 90,668. (PAPER 8/485)