J . Chem. SOC., Faraday Trans. 1, 1985, 81, 3 7 4 8 Electron Spin Resonance Spectroscopy of Surface Species formed upon Adsorption of Nitrogen Oxides and Oxygen on High-surface-area NiO-MgO and Coo-MgO Solid Solutions BY VALERIO INDOVINA, * DANTE CORDISCHI, STEFANO FEBBRARO AND MANLIO OCCHIUZZI Centro di Studio del CNR su ‘Struttura ed Attivita Catalitica di Sistemi di Ossidi’, Department of Chemistry, University of Rome, Rome, Italy Received 27th February, 1984 Solid solutions of high-surface-area (h.s.a.) NiO-MgO (Ni content 0.1-5 atoms per 100 Mg atoms) were activated by heating under vacuum at 1173 K for 5 h before exposure to 0,, NO,, NO or N,O at temperatures in the range 298-815 K. With all gases, depending on the temperature of adsorption, two species were observed by electron spin resonance spectroscopy.Both are axial signals of nickel species on the surface of MgO, as shown by broadening experiments with 0,. The e.s.r. analysis, the chemical conditions under which the species are formed and their thermal stability allow the assignment of the two species to (i) a coordinatively unsaturated Ni3+ (Ni:+) and (ii) Ni3+ ions in a distorted octahedral surface complex (Ni:+ * * L). Adsorption of NO at 178-298 K produces a (Ni . . . NO),+ adduct, NO:- and Mg2+. * * NO. Solid solutions of h.s.a. Coo-MgO (Co content 0.1-5 atoms per 100 Mg atoms) activated as for NiO-MgO were exposed to NO, or NO. The labile cobalt nitrosyl adduct formed is tentatively assigned to Co2+. .(NO),; NO;- species and Mg2+- * .NO are also formed. Some total and reversible adsorptions (removed by evacuation at the temperature of adsorption) were determined volumetrically.Reflectance spectra were taken under the same conditions for the formation of the paramagnetic species. Surface processes by which the various species were formed are discussed. Solid solutions of transition-metal ions in oxide matrices are useful model catalysts for the investigation of fundamental aspects of heterogeneous catalysis. Simple reactions such as the decomposition of N,O, the CO+O, reaction and H,-D, equilibration have been studied on high- and low-surface-area Coo-MgO ~atalysts.l-~ NiO-MgO samples have also been For an understanding of the catalytic behaviour, we need to know (i) the electronicconfiguration of the active transition-metal ion and (ii) the nature and the reactivity of the species formed on the surface of the catalysts upon adsorption of reagents and products.With this aim in mind, we have previously studied the adsorption of N,0,4 0,8 and on h.s.a. Coo-MgO solid solutions. In this paper we present the data for the adsorption of NO,, NO, N,O and 0, on h.s.a. NiO-MgO and for the adsorption of NO, and NO on h.s.a. Coo-MgO. The results for the two systems are complementary, since the surface species which cannot be detected by electron spin resonance spectroscopy in the case of cobalt can be detected in the case of nickel and vice versa, because of the electronic structure of the two ions. Evidence is presented for the 3 + oxidation state, which has often been assumed to be formed in the elementary steps of oxidation reactions on catalysts containing Co or Ni.3738 E.S.R. STUDY OF NO, AND 0, ON NiO-MgO AND COO-MgO Investigation of the adsorption by volumetric techniques allows evaluation of the significance of the e.s.r. observations with respect to the total amount of chemisorbed gas. EXPERIMENTAL MATERIALS The h.s.a. NiO-MgO solid solutions were prepared by impregnation of magnesium hydroxide with a solution of nickel nitrate (Erba, RP). The magnesium was the same as used before to prepare high-purity MgO.* The samples were decomposed slowly by raising the temperature to 800 K before carrying out the final activation at 1173 K for 5 h under a pressure of < lo-, N m-,. This procedure gives a good dispersion of nickel ions in solid solution of MgO.” The h.s.a.COO-MgO solid solutions were prepared as described for NiO-MgO. Details of the preparation and characterization are given in ref. (8). All specimens are designated as MN or MC followed by a number giving the transition-metal content as atoms per 100 Mg atoms. B.E.T. surface areas (rn, g-l), determined by N, adsorption at 77 K, are indicated in parentheses after each catalyst: MN 0.1 (253), MN 1 (215) and MN 5 (193); MC 0.1 (213), MC 1 (200) and MC 5 (170). NO, N,O, NH, and CO, were purified by double distillation under vacuum. High-purity 0, (Air Liquide, 99.95%) was used. NO, was prepared before each experiment by reacting NO with an excess of 0, and by subsequent distillation under vacuum. All gases were dried before admission to the adsorption chamber.APPARATUS AND PROCEDURE A weighed amount of the solid solution (ca. 0.1 g) was placed in a silica bulb equipped with a side e.s.r. tube and activated under vacuum at 1173 K, as specified above, before exposure to the various gases. For the adsorption measurements, the sample was contacted with the gas at a pressure chosen so as to have in all cases a final pressure of ca. 600 N m-,. Pressure readings were made with a pressure transducer (MKS, Baratron) capable of detecting variations of 0.01 N m-,. The adsorption was considered complete when two successive readings at 5 min intervals did not differ by > 0.5 N rn-, (hereafter called ‘total adsorption’). The samples were subsequently evacuated for 10 min at the same temperature and the amount adsorbed (always at the same temperature) gave a measure of the amount of gas which could be easily desorbed (‘reversible adsorption’).Occasionally, after NO or N,O adsorption, the gas phase was analysed by mass spectrometry (Micromass 601, VG). The e.s.r. spectra at 77 or 298 K were recorded at X-band frequencies on a Varian E-9 spectrometer. The absolute number of spins was determined from electronically integrated spectra using Varian ‘strong pitch’ as standard [(3 & 1) x lo1’ spin m-l]. The u.v.-visible reflectance spectra were recorded at room temperature with a Beckman DK I-A spectrometer using a silica activation chamber with a side cell equipped with an optical window. Activations and adsorptions were carried out in situ. RESULTS NiO-MgO SOLID SOLUTIONS E.S.R.SPECTRA The e.s.r. spectrum observed for activated MN samples consists of an isotropic signal at g = 2.21 due to Ni2+ ions in octahedral sites.12 Because a large fraction of Ni2+ ions depart from exact cubic symmetry, in these high-surface-area materials the linewidth of the e.s.r. signal is larger (AH,, = 1500 G) than that measured for low-surface-area MN solid solutions (80-1 60 G). l2 Upon adsorption of NO,, NO, N,O or O,, the formation of various paramagnetic species is observed (see table 1). For each species table 1 also reports the g values. UponV. INDOVINA, D. CORDISCHI, S. FEBBRARO AND M. OCCHIUZZI 39 Table 1. Surface species on high-surface-area NiO-MgO samples, as detected by e.s.r. spectroscopy gas adsorbed species g values NO, NO N,O 0, adsorption of a given gas, the formation or not of a species is specified in table 1 by the symbol + or -, respectively. For each gas the temperature of adsorption and pressure under which the various e.s.r.signals were observed will be specified in the next section, where the concentration of the radicals will also be reported. E.s.r. signals of the species NO;- and Mg2+. -NO (table 1) were observed by Lunsford13 upon adsorption of NO on thermally activated MgO. Since the intensities of the e.s.r. signals of these two species are the same on MN samples as on pure MgO, no specific role can be envisaged for the nickel ions in their formation. The e.s.r. signal of the 0; species on the surface of pre-irradiated MgO (Mg2+ - - * O;, table 1) has also been reported previ0us1y.l~ However, in this case a specific role exists for the transition-metal ion since the 0; species are not formed on thermally activated Mg0.15916 The e.s.r.spectra of the species designated in table 1 as A, B and C are reported in fig. 1. Spectra (a) and (c) correspond to species A and C, respectively. Spectrum (b) consists of two signals, namely species A and B. Species A and B, when occurring together, can be distinguished since the signal of species A undergoes reversible and complete broadening upon exposure to oxygen whereas that of species B is only slightly broadened. The g values of the three axial signals are reported in the stick diagram of fig. 1 . FORMATION, CONCENTRATION AND STABILITY OF SPECIES A, B AND C Species C is considered first.This is formed by adsorption of NO on MN samples at 298 K or lower temperatures (down to 178 K). The intensity of the signal, which depends on the pressure of NO, reaches a maximum of 40 x 10l5 spin rn-, on MN 1 and 6.5 x 1015 on MN 0.1 at a pressure of 2 kN m-,. Species C is very labile and its e.s.r. signal disappears on evacuation for 10 min at 298 K. It is useful to compare this amount with the amount adsorbed as determined by volumetry on the MN 1 sample. The total adsorption at 298 K is 790 x 1015 molecule rn-, and includes both irreversible (NO;-) and reversible forms, the latter amounting to 180 x 1 OI5. Thus, theconcentration of the radical is less than the amount of NO reversibly chemisorbed at 298 K. Species A and B are formed with NO,, NO, N,O and 0, at the temperatures specified in table 2.For each gas, table 2 also reports for MN 1 and MN 0.1 the concentration of radicals, species A and €3, as a function of the temperature. The results show that species A and B are formed with NO, even at 298 K. Progressively higher temperatures are required for NO, N,O and O,, in that order. Both species are stable at 298 K, i.e. the intensities of their e.s.r. signals are unaffected by evacuation40 E,S.R. STUDY OF NO, AND 0, ON NiO-MgO AND COO-MgO .-i 2.36 2 . 0 2 species A species B 2.25 2.11 2.28 2.15 species C - - Fig. 1. E.s.r. spectra at 77 K on an MN 1 sample activated by heating under vacuum at 1173 K and exposed to: (a) 0, at 723 K, (b) NO, at 298 K and (c) NO at 298 K. The group of lines marked with an asterisk belongs to the Mg2+- * .NO and NO:- species.Table 2. Concentration of species A and B obtained upon adsorption of NO,, NO, N,O and 0, on NiO-MgO at various temperatures concentration/ spin m-2 NO2 NO N2O 0 2 T / K A B A B A B A B sample MN 1 298 70 50 466 572 717 - - 815 MN 0.1 298 6.6 8.4 572 - - - - - - - - 2.2 0.0 0.0 9.0 2.5 3.0 2.0 2.6 25.0 14.0 0.0 5.0 - - - - - - - - - 0.0 0.0 0.0 0.8 0.0 0.0 0.0 0.2 0.0 4.9 3.2 0.9 4.0 0.9 0.0 - - 4.0 - - - at this temperature. The concentration of species A + B is 4-8 times larger on MN 1 than on MN 0.1. The thermal stability of species A and B was further investigated as follows. The two species were formed on an MN 1 sample in two different ways: (i) by exposure to NO, at 298 K and (ii) by exposure to N,O at 572 K.The samples were thenV. INDOVINA, D . CORDISCHI, S. FEBBRARO AND M. OCCHIUZZI 41 0-4P 100 o--o--- t \ 8ol 2o t O\ o+o 1 , , , , , Q \ 0 0= T/K 273 473 673 873 Fig. 2. Thermal stability of species A and B, obtained either by exposure to N,O at 573 K (0 and 0, respectively) or NO, at 298 K (A and V, respectively). The samples were subsequently heated under vacuum at progressively higher temperature. Intensities are reported in arbitrary units as a percentage of the initial intensity (= 100). Table 3. Concentration of species A and B after adsorption of CO, or NH, concentration/ spin mP2 treatmenta A B A+B N,O at 573 K 25 10 35 N,O at 573 K 22 1 1 33 NH, at 298 Kb 18 16 34 CO, at 298 Kb 12 28 40 a Before exposure to N,O the MN 1 sample was activated by heating under vacuum at 1173 K.Before exposure to CO, or NH, the sample was outgassed at 298 K. evacuated at progressively higher temperatures, recording the e.s.r. spectra after each step. In fig. 2, the intensities of the signals for species A and B, reported as a percentage of the initial intensity (= loo), are plotted as a function of the evacuation temperature. Both species possess the same thermal stability and are stable up to ca. 573 K. At 673 K the original intensities are reduced by one-half and at 780 K the signals are destroyed. In other experiments, MN 1 samples containing a known amount of species A and B were exposed to CO, or NH, at a pressure of 13.5 kN m-2. The treatment provokes a decrease of the concentration of species A and an increase of that of species B, while the total concentration of radicals remains constant (table 3).OXYGEN ADSORPTION The total adsorption of 0, on MN I , activated by heating under vacuum at 1 173 K, was as indicated in parentheses at the various temperatures: 298 K (7 x 1015 molecule m-2), 428 K (19 x 1015), 563 K (22 x and 717 K (30 x 1015). At 298 K, the42 E.S.R. STUDY OF NO, AND 0, ON NiO-MgO AND COO-MgO amount reversibly adsorbed was 5 x 1015 molecule m-2 and the amount irreversibly adsorbed was 2 x 1015. Species A and B were not formed at 298 K (table 2). The 0; species was observed at low concentration (0.9 x 1015 spin rn-,). Thus, oxygen is mostly chemisorbed in diamagnetic forms : probably 0;- for the reversible adsorption and 02- for the irreversible adsorption.The MgO matrix does not chemisorb oxygen at 298 K (< 0.5 x 1015 molecule m-2). REFLECTANCE SPECTRA The spectrum obtained for an MN 5 sample after activation under vacuum at 1 173 K consists of five bands. These have been observed by other authors and assigned to bulk Ni2+ ions (8300, 14000 and 23000 cm-l) and surface Ni2+ probably in C4v symmetry (5000 and 23000 cm-l).ll* l7 After adsorption of N,O, NO, and 0, a broad and intense band is observed. This disappears upon evacuation at ca. 870 K. A similar band was observed by Hagan et al.ll on h.s.a. MN samples following adsorption of 0,. Note that the broad adsorption is observed under the same conditions for the formation of species A and B and, moreover, the temperature at which the absorption band disappears is not far from that at which species A and B are destroyed.COO-MgO SOLID SOLUTIONS E.S.R. SPECTRA Upon adsorption of NO (6 kN m-,) on MC 1 or MC 5 samples activated by heating under vacuum at 1173 K for 5 h, intense e.s.r. signals are observed (fig. 3). These consist of a broad band at 7 = 2.12 showing a poorly resolved hyperfine structure arising from interaction v ;ch the cobalt nucleus [MC 1, spectrum (a)]. In addition, as in the case of NO adsorption on MN, the e.s.r. signals of species adsorbed on sites of the MgO matrix can be recognized: NO:- and Mg2+. * .NO. The species is very labile and is destroyed by a few minutes evacuation at 298 K. Thus, the stability is the same as for species C of the MN system. The intensity of the signal for the MC 1 sample corresponds to 23 x 1015 spin rn-, at 298 K and 58 x 1015 at 178 K.The total adsorption of NO is 930 molecule m-, at 298 K and the reversible adsorption 370 molecule md2 at the same temperature (MC 1). The reasons for the NO radical concentration being smaller than the reversible adsorption of NO are as reported above for the MN system. The maximum radical concentration on the MC 5 sample is 1.0 x 1017 spin m-,. ADSORPTION OF 0, AND N,O The adsorption of 0,8 and N,O has been previously investigated in our laboratory on h.s.a. MC samples. Upon adsorption of O,, the most significant results, recalled here in order to underline some important similarities with the MN system, concern the formation of Co3+ - * 0; (for this species, up to 1 .O x 10" spin m-, were detected) and that of 0; and 0; species chemisorbed on sites of the matrix.The total adsorption of oxygen, as determined volumetrically, was larger than the concentration of radicals. Therefore, 0;- and 0,- are also thought to be present on the surface.8 Adsorption of N,O leads to the formation of 0-, 0; and 0; chemisorbed on sites of the matrix and to the formation of Co3+ - - -0;. The formation of 0- species suggests an intermediate species Co3+. - -O-, which was not detected as it is very labile .4V. INDOVINA, D . CORDISCHI, S. FEBBRARO AND M. OCCHIUZZI 500 G 1 43 -I- 2.007 - 1.996 1.89 c o 1 2 12 A = 94G Fig. 3. E.s.r. spectrum at 77 K of (a) MC 1 and (b) MC 5 heated under vacuum at 1173 K and exposed to NO. The lines at g, = 1.996 and g,, = 1.89 are due to the Mg2+.* . NO species and that at g, = 2.007 to NO:-. REFLECTANCE SPECTRA The reflectance spectrum of an MC 5 sample activated under vacuum at 1173 K was nearly identical to that previously reported for similar samples by Hagan et all1 The broad background absorption, obtained by exposure to 0, at 298 K,ll is also detected upon adsorption of N,O and NO, both at 298 K. Evacuation at 298 K does not affect the spectra. Evacuation at progressively higher temperatures shows that the broad absorption disappears at a temperature in the range 723-850 K. DISCUSSION THE NATURE OF SPECIES A AND B The e.s.r. axial signals of species A and B are due to different forms of nickel with spin 1/2. The two species are on the surface of the MgO matrix, as shown by the fact that their e.s.r.signals are reversibly broadened upon addition of 0,. In principle, two oxidation states are possible: Ni+ or Ni3+. In the following it will be shown that species A and B are surface Ni3+ ions. The assignment relies on both spectroscopic and chemical evidence. Spectroscopic evidence will be considered first. In most cases, relying on g values alone it is rather difficult to distinguish between the low-spin d7 configuration (Ni3+) and the d9 configuration (Ni+). In fact, the orbitally degenerate ground state (2E2s)44 E.S.R. STUDY OF NO, AND 0, ON NiO-MgO AND COO-MgO is not split in a cubic or trigonal crystal field, but the degeneracy is removed by the Jahn-Teller effect (either static or dynamic). Thus, for both configurations an axial spectrum with g values in the range 2-2.4 is expected.However, the effect of the tetragonal distortion on the principal values of the g tensors is opposite for the two configurations. In particular, for an elongated octahedron gll < gl is expected for a low-spin d7 ion and gI1 > g , for a d9 ion; the reverse is predicted for a compressed octahedron. Thus, given that in the present case gll < g,, the only alternative left is between Ni3+ ions in an elongated octahedron and Ni+ in a compressed octahedron. The latter configuration can be ruled out since we have substantiated that species A and B are on the surface where a compressed octahedral form appears to be very unlikely. Chemical arguments in favour of Ni3+ rather than Ni+ arise from an inspection of (a) the conditions in which species A and B are formed and (b) their thermal stability.In particular, oxidation of the surface Ni2+ ions of the MN solid solutions is expected upon exposure to NO,, NO, N20 and 0,. More specifically, oxygen atoms appear to be involved in the oxidation of Ni2+ since mass-spectrometric analysis shows the formation of N, when NO or N,O are contacted with the MN samples. In other words, NO and N,O decompose during the formation of species A and B. Analogously, with O,, oxygen atoms arise from the cleavage of the 0-0 bond during the activation of molecular oxygen. As far as the thermal stability of species A and B is concerned, this is found to be nearly identical to that previously observed for Ni3+ ions present in the bulk of MgO-NiO-Li,O solid solutions, in which the Ni3+ oxidation state was induced by the presence of lithium?, The significance of the comparison relies on the fact that for both systems (h.s.a.MN and MgO-NiO-Li,O) the decay of the Ni3+ signal (either bulk or surface) is related to desorption of oxygen from the surface of the samples. Having established that species A and B are two different Ni3+-containing species, the next step toward the identification of species A and B is as follows. The anisotropy of the g values of species A and B cannot be fully explained by using the equations given by Lacroix et al. for low-spin d7 ions undergoing a Jahn-Teller effect.’* The equations have been successfully used to assign isotropic or axial signals observed in several matrices for bulk Ni3+ ions undergoing either static or dynamic Jahn-Teller In the case of species A and B, Ni3+ ions are in an octahedral field with strong axial distortion, the cubic field being near to the cross-over region between the ,EB and the 4qg states (fig.4). A similar situation has been discussed previously for other octahedral Ni3+ complexes by Reinen et aL21 The analysis, which applies to any symmetry with a quaternary axis, givesgil and g,values as functions of S,/r and S,, ?/<, in which r is the spin-orbit coupling constant (fig. 4). By following this analysis it is found that the g values of species A are fitted by S2/t = 6 and = 2.45 and those of species B by S,/t = 0.7 and S2.4/t = 3.30. Thus, the S, parameter, which increases with increasing axial field, is higher for species A than for species B, suggesting that species A consists of coordinatively unsaturated Ni3+ (Ni;+) and species B of surface Ni3+ in an octahedral distorted complex (NiE+ - * - L).Again, the assignment is consistent with the chemical behaviour observed upon exposure to NH, or CO, ; i.e. NiE+ is transformed into Ni;+ - * L by adsorption of either NH, or CO, schematically, Nig+ + L --+ NiE+ - - L). In the case of N,O, NO or 0, adsorption, the ligand L is thought to be 02- (Ni3+- * -02-). The fact that the e.s.r. parameters of species B are largely independent of the particular L implies that the axial component of the crystal field is of the same order of magnitude for the various ligands. This is possible since ligands such as those used here have been shown to be linked to the surface by bonds whose strength is comparable to that of the 02-ions of the lattice.,,V.INDOVINA, D. CORDISCHI, S. FEBBRARO AND M. OCCHIUZZI 45 2 / Eg / Fig. 4. Splitting of the lowest terms, zEg(r&,ei) and 4Tg (rig,ei), in a strong tetragonal field near to the cross-over region. Finally, evidence has been given for the previous assignment of the broad band in the reflectance spectra of NiO-MgO solid so1utions.l1Vl7 In fact, species A and B, whose behaviour parallels that of this reflectance band, have been shown to consist of Ni3+ ions. NITROSYL ADDUCTS OF Ni AND Co Surface species formed on MN (species C) and MC (broad e.s.r. signal at g = 2.12) upon adsorption of NO have the typical behaviour of surface nitrosyl complexes, such as high lability. For species C, gll > gl.The situation is the opposite to that observed for species A and B and, therefore, by using the same arguments given above for species A and B, species C can be assigned to a surface complex of a nickel ion in the dg configuration, i.e. to a Ni+ complex. Accordingly, the simplest formulation of the complex would be Nil * -NO+. This formulation has been adopted previously for mononitrosyl Ni adducts in zeolites, with g values similar to those found here for species C.23- 24 The assignment also relies on i.r. spectra which showed the typical bands of nitrosyl ligands (ca. 1800 cm-l). On the other hand, the e.s.r. spectra of Ni adducts in zeolites have been analysed using the equations: These apply to a dg ion in a tetragonal elongated octahedron. On the basis of simple crystal-field theory a ratio Ail/Al < 1 is expected, as found in the case of Ni+ in various However, as shown in table 4, the All/A1 ratios calculated from the g values of various nitrosyl adducts are greater than unity.The discrepancy can be attributed to the fact that the Ni-NO bond is, to some extent, covalent, thus accounting for the apparent failure of crystal-field theory. Thus, in the absence of hyperfine structure in the e.s.r. signal of species C and of a more detailed description of the structure of the nickel adduct, it is better to formulate species C as (Ni. * Coming now to the broad band showing hyperfine interaction with the Co nucleus, 26 and also for several complexes of rather than as Ni+.* .NO+.46 E.S.R. STUDY OF NO, AND 0, ON NiO-MgO AND COO-MgO Table 4. Spectroscopic parameters of Ni nitrosyl adductsa system 811 g, 4 / A l b ref. NP-Y zeolite 2.33 2.169 2.04 24 NiII-Y zeolite 2.34 2.171 2.00 23 NiO-MgO 2.28 2.15 2.14 this work Ni+ in NaF 2.766 2.1 14 0.59 25 Ni+ in SrF, 2.597 2.092 0.60 26 a The data are compared with those obtained for Ni+ in the matrix of NaF and SrF,. A/\/AL = 4kl-ge)/kl/ -ge)* Table 5. Surface species observed by e.s.r. on h.s.a. NiO-MgO and COO-MgO solid solutions on adsorption of 0,, N,O and NO surface species gas adsorbed NiO-MgO CoO-MgOa a The data with 0, and N,O for the COO-MgO system are taken from ref. (4) and (8). this cannot be due to the analogous mononitrosyl adduct since this, in the case of Co, is diamagnetic or of even spin.Possible paramagnetic complexes are : Co+ - * NO, Co3+ - - - NO and Co2+ - - - (NO),. A tentative assignment to Co2+ - - * (NO), is suggested by (i) the known tendency of Co to coordinate more than one molecule1* and (ii) the fact that the adsorption of NO is higher on MC than on MN (by a factor of ca. 2). Moreover, no evidence for a Co3+ - - - NO species was found in a special experiment performed by adsorbing NO on an h.s.a. MC doped with Li in order to induce the formation of Co3+: the intensity of the e.s.r. signal was the same as obtained after NO adsorption on a normal MC sample. SURFACE REACTIONS The mechanisms of 0, adsorption and N20 decomposition have been discussed previously in the case of COO-MgO solid ~ o l u t i o n s .~ ~ ~ The data presented here for the related NiO-MgO system, along with the additional information for COO-MgO, provide further evidence for the mechanisms reported previ0usly.~9 In particular, the T3+ and T3+ * - 0,- which could not be observed when T was cobalt have now been detected with the nickel system. The reverse is true for the T3+. -0; species which was observed in the case of cobalt but not in that of nickel. Table 5 lists the surface species formed on the MN and MC systems upon adsorption of 0,, N,O and NO. From a qualitative view point, the nature of the surface species points to a strict analogy between the two systems. However, differences exist as regards the temperature at which the surface species are formedV. INDOVINA, D.CORDISCHI, S. FEBBRARO AND M. OCCHIUZZI 47 and their stability. These differences can be used to explain the higher catalytic activity of COO-MgO as compared with NiO-MgO for the decomposition of N2028 and the CO+O, reaction.' This aspect will be discussed elsewhere. Note, however, that the absence of 0- and 0, species on the MN surface (0, only being observed, table 5) can be accounted for by the higher temperature at which the N,O decomposition is carried out on the MN catalysts as compared with the MC catalysts. In previous work from our group the thermal stability of 0- and 0; species was found to be lower than that of O;(O- < 0; < O T ) . ~ ~ ~ Coming now to the adsorption of NO, the surface species observed can be accounted for by the following reactions. In particular, whereas the elementary steps for the formation of Nii+ and Ni:+ * .* 02- cannot be given, the overall process can be shown by: 2 NO(g) + 2 NiE+ + 0:- + Ni:+ + Ni;+ - - * 0,- + O& + N,(g). The formation of NO;-, (Ni - - - NO),+ and Mg2+ - - NO is given by : NO(g)+Oi-+NO;- (1) NO(g) +NiE+ + (Ni * - - NO),+ (2) NO(g)+Mgi+ + Mg2+- - *NO. (3) The analogous surface reactions (1)-(3) are also observed for Coo-MgO. Note that oxygen species (0-, 0; and 0;) are not seen upon adsorption of NO on both systems. A possible cause is that once formed these species react with gas-phase NO, yielding NO; and NO;. In agreement with this hypothesis, the concentration of radicals is found to be smaller than the total adsorption of NO. CONCLUSIONS Two main conclusions may be drawn from these results: (i) Ni3+ and Co3+ ions are formed on the surface of NiO-MgO and COO-MgO solid solutions upon adsorption of the appropriate gases and (ii) nearly all the Ni2+ and Co2+ ions on the surface can be oxidised to Ni3+ and Co3+, respectively.The first conclusion follows from the e.s.r. evidence and the second can be substantiated as follows. A surface analysis by X.P.S. of the MN 29 and MC 30 systems has shown that the surface composition is only slightly different from that of the bulk. Therefore, the surface concentration of Ni2+ or Co2+ is 1.2 x 1017 atom m-, in MN 1 and MC 1 samples and 0.12 x 1017 in the MN 0.1 sample. Comparison of these surface compositions with the concentrations of paramagnetic species leads to conclusion (ii).In fact, the maximum concentration of Ni:+ + Nii+ - - L was 1.2 x 1017 spin m-, on MN 1 and 0.15 x 1017 on MN 0.1 ; the concentration of Co3+ - - - 0; was 1 .O x 10'' spin rnP2 on MC 1. A. Cimino and F. Pepe, J. Catal., 1972, 25, 362. V. Indovina, A, Cimino and F. Pepe, Gazz. Chim. Ztal., 1980, 110, 13. V. Indovina, A. Cimino, M. Inversi and F. Pepe, J . Catal., 1979,56, 396. 75, 21 77. A. Cimino, R. Bosco, V. Indovina and M. Schiavello, J . Catal., 1966, 5, 271. A. Cimino, V. Indovina, F. Pepe and M. Schiavello, J . Catal., 1969, 14, 49. V. Indovina, A. Cimino and F. Pepe, 9th. Iberoamerican Symposium on Catalysis, 1984, Lisbon, Portugal. * V. Indovina, D. Cordischi, M. Occhiuzzi and A. Arieti, J. Chem. SOC., Faraday Trans. 1 , 1979,48 E.S.R. STUDY OF NO, AND 0, ON NiO-MgO AND COO-MgO D.Cordischi, V. Indovina, M. Occhiuzzi and A. Arieti, J. Chem. SOC., Faraday Trans. I , 1979, 75, 533. D. Cordischi, V. Indovina and M. Occhiuzzi, J. Chem. Soc., Faraday Trans. I , 1980, 76, 1147. lo V. Indovina, D. Cordischi and M. Occhiuzzi, J. Chem. Soc., Faraday Trans. I , 1981, 77, 81 1. l1 A. P. Hagan, M. G. Lofthouse, F. S. Stone and M. A. Trevethan, Preparation of Catalysts II, ed. B. Delmon, P. Grange, P. Jacobs and G. Poncelet (Elsevier, Amsterdam, 1975), p. 417. l2 A. Cimino, D. Cordischi, G. Guarino and A. Micheli, Trans. Faraday Soc., 1971,67, 1776. l3 J. H. Lunsford, J. Chem. Phys., 1967, 46, 4347. l4 A. J. Tench and P. Holroyd, Chem. Commun., 1968,471. l5 V . Indovina and D. Cordischi, Chem. Phys. Lett., 1976, 43, 485. l6 D. Cordischi, V. Indovina and M. Occhiuzzi, J. Chem. Soc., Faraday Trans. I , 1978, 74, 456. l7 M. Lo Jacono, A. Sgamelotti and A. Cimino, Z. Phys. Chem. N.F., 1970,70, 179. Is J. Tanaka, I. Shindo and M. Tsukioka, J. Phys. Soc. Jpn, 1980, 49, 120. 2o D. M. Hannon, Phys. Rev., 1967, 144, 366. 21 D. Reinen, C. Friebel and V. Propach, Z. Anorg. Allg. Chem., 1974,408, 187. 22 D. Shopov, A. Andreev, P. Pumpalov, J. Catal., 1970, 9, 398. 23 P. H. Kasai, R. T. Bishop Jr and D. McLeod Jr, J. Phys. Chem., 1978,82-83, 279. 24 C. Naccache and Y. Ben Taarit, J. Chem. Soc., Faraday Trans. I , 1973, 69, 1475. 25 J. W. Orton, Rep. Prog. Phys., 1959, 22, 204. 26 P. J. Alonso, C. J. Gonzales, H. W. Den Hartog and R. Alcala, Phys. Rev. B, 1983, 27, 2722. 27 P. A. Ayscough, Electron Spin Resonance in Chemistry (Methuen, London, 1967) p. 191, table 6.8. 28 A. Cimino, Chim. Ind. (Milan), 1974, 56, 27. 2s A. Cimino, B. A. De Angelis and G. Minelli, J. Electron Spectrosc. Relat. Phenom., 1978, 13, 291. 30 A. Cimino, B. A. De Angelis and G. Minelli, Surf. Interface Anal., 1983, 5, 150. R. Lacroix, V. Hochli and K. A. Miiller, Ber. Tag. Schweiz. Phys. Ges., 1964, 37, 327. (PAPER 4/334)