J. MATER. CHEM., 1991 1(4), 531-537 531 Copper-Cobalt Hydroxysalts and Oxysalts: Buik and Surface Characterization Piero Porta,* Roberto Dragone, Giuseppe Fierro, Marcello Inversi, Mariano Lo Jacono and Giuliano Moretti Centro CNR su 'Struttura e Attivita Catalitica di Sistemi di Ossidi' (SACSO), c/o Dipartimento di Chimica, Universita di Roma 'La Sapienza', Piazzale A. Moro 5,00185 Roma, Italy Copper-cobalt oxysalts of various Cu : Co atomic ratios (100 :0. 92:8,85: 15, 77 :23,67 :33,15: 85,0:100) have been prepared by coprecipitation, at constant pH, from solutions of copper and cobalt nitrates added to a solution of NaHCO,; the structure and chemical nature of the Cu-Co oxysalts have been investigated by several complementary techniques such as X-ray diffraction (XRD), magnetic susceptibility, reflectance spectroscopy and X-ray photoelectron spectroscopy (XPS) to obtain information both on their bulk and on their surface pro pert ies.In the range of Cu :Co atomic ratios from 100 :0 to 67 :33 the materials consist of Co-containing malachite, Cu,~,Co,CO,(OH),, while for Cu : Co equal to 0: 100 and 15:85 the products are CoCO,, spherocobaltite, and Cu-containing spherocobaltite Coo,,,Cuo~,,CO,, respectively. From both reflectance spectroscopy and magnetic susceptibility measurements the presence of only Cu2+ and Co2+ species has been ascertained in all materials. Antiferromagnetic interactions are strong at high copper content and become weaker at increasing cobalt loading. The observed decrease in volume of the malachite monoclinic unit cell with incorporation of Co2' ions, in spite of the larger Co2+ ion radius with respect to that of Cu2+, indicates that distortion of the MO, polyhedra at the cation site of the malachite lattice is less when Cu2' is replaced by Co2+.The presence of Cu2+ in the spherocobaltite trigonal lattice shows a small cell volume decrease as expected from ionic radii difference. XPS has confirmed the presence of only Cu2+ and Co2+ species at the surface of the materials. It is also shown that the local nature of the chemical bond is essentially determined by the nature of the first-neighbour ligands. At higher copper content, where the materials are cobalt-containing malachite, the surface is enriched in cobalt.Keywords: Copper-cobalt oxysalt; Mixed oxide precursor; Catalyst 1. Introduction It is essential that the initial precursor homogeneity is maintained during the transformation steps of the solid. One Mixed copper-cobalt oxides, especially those with the spinel of the most promising preparation methods for such multic- structure, are active catalysts for the oxidation of CO by 02.1 Cu-Co oxides (containing also A1 or Cr, Zn and an alkali omponent materials to satisfy the above requirements is the metal) are also used for the conversion of synthesis gas (CO/ coprecipitation technique, by which oxysalt precursors such C02/H2) to higher alcohol^.^-^ The composition of these as hydroxycarbonates, carbonates, oxalates, acetates, etc., are materials, developed mainly at the Institut Francais du produced and where all the metals are possibly in the same Petrole,s corresponds to that of conventional copper-based crystalline structure.methanol synthesis catalysts modified by the addition of This paper follows studies (both bulk and surface charac- cobalt. Mixed Cu-Co-Zn-Cr oxides have also been tested terization) performed in our laboratory on Cu-Zn hydroxy-as useful catalysts for the synthesis of hydrocarbons.6 carbonate precursors and on derived Cu-Zn mixed ~xides,~ It is well known that the chemical and physical properties as well as on Cu-Mn (hydr0xy)carbonate precursors and on of mixed oxide catalysts depend on the mode of preparation derived Cu-Mn mixed oxides." In this work we present the and on their thermal history since these effect the component results of the bulk and surface characterization of some Cu- interdispersion and definite m~rphology.~ Co precursors for mixed oxide systems.In a following paper To this end, many considerable efforts have been made in finding preparative we will report the results of the bulk and surface characteriz- methods that are capable of giving precursors where two or ation of the Cu-Co mixed oxides obtained by calcination of more metals are in the same or parent crystalline structure the precursors in air and in NZ, and of their behaviour under and which, after calcination at low temperature, may provide H2 reduction. well interdispersed mixed oxides of required specific surface area and particle size.Note that the solid-state reactions of interest are mainly 2. Experimental topotactic transformations, which lead to oxide products that 2.1 Sample Preparation and Analysis inherit certain structural and morphological features of the precursors; at low temperature the topotactic reactions require Compounds with Cu :Co atomic ratios of 100 :0,92 :8,85 :15, constrained mobility of species and limited structural 77 :23, 67 :33, 15 :85 and 0 :100 were prepared by coprecipi- rearrangement during calcination.* [For the definition of tation at a constant pH of 8 from Cu and Co nitrate solutions topotaxy see ref. 8(b).] As a consequence of their useful in suitable proportions with NaHCO, in excess by the same interdispersed state, the different oxide materials may interact, method described for Cu-Mn salts [see ref.10(a) for details]. favouring synergetic effects; they may act as multicomponent All the precipitates were repeatedly washed with cold water promotion compounds, manifesting a wide variety of electric, to eliminate the presence of nitrate ions and sodium. After electronic and structural properties, and they display several washing, the sodium content was analysed by atomic absorp- reactivity properties under subsequent treatments (under inert, tion and found to be 100 ppm. Once washed, the solids were reducing or oxidising atmosphere). heated in an oven at 383 K for 24 h and finely ground in an 532 agate mortar. The colour ranged from light green to grey- green for samples with Cu: Co atomic ratios of 100-67 :33, was pink for the 0: 100 composition, and light violet for the 15:85 compound. Elemental copper and cobalt analyses were performed by atomic absorption with a Varian SpectraAA-30 instrument.The percentage of water loss from the samples was also determined by weight loss after heating them at 443 K for 4 h. Table 1 lists the prepared compounds, the analytical data and other features. 2.2 X-Ray Diffraction The powder diffraction patterns were obtained with a Philips automated PW 1729 diffractometer equipped with an IBM PS2 computer for data acquisition and analysis (software APD-Philips) and an HP plotter. Scans were taken with a 20 step size of 0.01", using Co-Ka (iron-filtered) radiation.The intensities of the reflections were estimated by evaluation of the integrated peaks. Structural data for reference compounds were taken from the literatures.' 2.3 Reflectance Spectra Diffuse reflectance spectra were carried out by utilizing a Cary 2300 spectrometer equipped with a diffuse reflectance accessory and an IBM PS2 computer for data acquisition and analysis (software Spectra Ca1c.-Galactic Industries Corp.), in the wavelength range 200-2500 nm, covering the UV, visible and near-infrared regions. 2.4 Magnetic Susceptibility Magnetic susceptibilities were measured by the Gouy method over the temperature range 100-300 K. A check that the susceptibilities were independent of magnetic field strength was made. Correction was made for the diamagnetism of the samples.Then the mean magnetic susceptibility values, xat= xsp(x,,rncu+xcornco), where xsprepresents the magnetic suscep- tibility per gram of paramagnetic species, and x and rn are the relative molar fraction and atomic weight of Cu and Co, were evaluated for each precursor. An example of the linearity on the l/xat us. T plot is shown in Fig. 1 for the sample with atomic ratio Cu :Co =85 :15. From the intercepts of the line on the temperature axis the value of the Weiss temperature, 0 (which accounts for the strength of antiferromagnetic inter- actions among paramagnetic ions in the lattice), can be inferred. From the slope of the l/xat us. T plot the value of the Curie constant C=xat(T+0)can be obtained.The values of the magnetic moment ~~2.83 JC were calculated for all 4ooL 300-/* 1 -$ 200-8' -1 00 0 100 200 300 TIK Fig. 1 Reciprocal atomic susceptibility, 1/xat, us. temperature, T, for the Cu :Co =85 : 15 composition J. MATER. CHEM., 1991, VOL. 1 the samples studied. Table 1 reports the data obtained from the magnetic measurements. 2.5 X-ray Photoelectron Spectroscopy 2.5.1 Qualitative Analysis The X-ray photoelectron and X-ray excited Auger spectra were obtained using both Mg-Ka (hv= 1253.6 eV) and A1-Ka (hv = 1486.6 eV) radiations with a Leybold-Heraeus LHS-10 spectrometer operating at constant transmission energy (E, = 50 eV), and connected to an HP 21 13 computer for data acquisition and analysis. The spectrometer was calibrated by using the following photoemission lines (with reference to the Fermi level): EB Cu 2p,,, =932.8 eV; EB Ag 3d5,, =368.3 eV; and EB Au 4f7,, = 84.0 eV.The instrumental resolution expressed by the full- width at half maximum (FWHM) of the Au 4f,,, peak was 1.2 and 1.4 eV using Mg and A1 sources, respectively. The analysis chamber during the experiments was evacuated to better than 1 x10-* mbar. The spectra were recorded at room temperature and at low X-ray fluxes (anode operating at 10 kV and 20 mA with Mg-Ka, and at 12 kV and 10 mA with Al-Ka), and with an acquisition time of ca. 1 h (0.05 eV per channel, dwell time 50ms) to avoid X-ray induced reduction of Cu2 + species, as verified for experiments with longer acquisition time.The data analysis procedure generally involved smoothing, background substraction by a linear or 'S-type' integral profile and a curve-fitting procedure, using a mixed Gaussian-Lorentzian function, by a least-squares method (software DS4,X by Leybold Heraeus). The spectra were analysed in terms of the relative peak area intensity, the FWHM of the peaks, the chemical shifts of the 2p,,, and L3M45M45 tran- sitions of copper and cobalt, and the 0 1s transition. The experimental errors were estimated to be kO.2 eV for the photoelectron peak; and f0.2 and k0.5 eV for the Auger peaks of copper and cobalt, respectively. For cobalt the FWHM of the Auger peak is ca. 10 eV and the maximum is taken as the centre of the broad square line.The samples were pressed into a grooved tantalum sample holder. Charging effects were corrected using the C 1s peak relative to the carbonate group which was fixed for all the samples at EB =289.1 eV. Note that the use of A1-Ka radiation produces interference in the Co spectral region (at the lowest binding energies) by the Co L3M23M45 Auger transition, which causes the absolute Co 2p3,, integrated intensity to be overestimated by ca. 15%.12 On the other hand, when Mg-Kcr excitation is used the Auger peak interferes with the 0 1s photoelectron peak. Note also that, as reported by Dillard and K~ppelman'~ and verified by us, when the Mg source is used there is a partial overlap between the oxygen Auger KLL peaks and the low- energy region of the Co 2p,,, which modifies its background lineshape.The data reported in Table 2 were obtained with the Al-Kcr radiation. 2.5.2 Quantitative Surface Analysis The surface compositions of the compounds were obtained on the basis of the peak area intensities of the Cu 2p3,, and Co 2p3/, levels using the sensitivity-factor approach.gd The relative atomic sensitivity factor was determined from exper- imental measurements on Cu2C03(0H), and CoCO,, using the data obtained with the A1-Ka radiation. The following expression was applied to obtain the surface atomic ratio from the XPS intensity ratio (c~/co)xPs=0.76CI(CU 2P,,,)/I(CO 2P3/2)1 Where I represent the peak areas and 0.76 is the relative Table 1 Phases present detected by XRD" phases cu:co cu:co (XRD) CU% CO% H,O% ('4'4) (XPS) xcu Ccalc Cexp NPB O/K a/A blA CIA PI" 55.6 -2.9 --1.o -0.6 2.19 200 9.518 11.947 3.249 98"45' 50.9 3.9 2.4 11.9 5.4 0.93 0.86 0.8 2.53 110 9.471 12.019 3.220 97"55' 46.7 7.5 4.1 5.8 4.5 0.85 1.08 1.o 2.83 75 9.410 12.038 3.210 97" 10 43.4 11.1 5.6 3.6 2.5 0.78 1.30 1.28 3.20 80 9.400 12.082 3.198 97'05' 38.8 16.3 6.8 2.2 1.5 0.69 1.59 1.6 3.57 55 9.398 12.121 3.185 97" 8.2 37.3 5.2 0.2 0.2 0.16 3.25 3.7 5.40 70 4.660 14.949 -48.8 2.5 ----3.77 5.50 60 4.668 14.991 malachite-like, S =spherocobaltite-like.Copper and cobalt weight percent (by atomic absorption, accuracy within 2%); weight percent of water obtained by weight loss at 443 K (accuracy atomic ratios determined by atomic absorption (AA)and by XPS (accuracy within 15%); copper molar fractions, Xcu; nominal and experimental copper molar fractions, Xcu; Curie and experimental, Cexp),magnetic moments (accuracy within 2%); Weiss temperature (K) (accuracy within 10%); lattice parameters calculated for Co-malachite (monoclinic (trigonal cell).Table 2 Binding energies of 0 Is, Cu 2p,,,, Co 2p3/, and Auger kinetic energies for CU(L~M~~M~~, 'G)and Co (L3M45M45) transitions" Eb (0 Is)/eV E,(C Is; COi-531.2 (2.4) 933.8 (4.5) -916.9 -289.1 (2.5) 531.1 (2.7) 933.8 (4.8) 781.1 (4.8) 9 16.8 769.4 289.1 (2.4) 531.3 (2.8) 933.8 (4.5) 781.1 (4.8) 9 16.8 769.4 289.1 (2.4) 53 1.2 (2.7) 933.5 (4.3) 781.3 (4.8) 9 16.4 769.4 289.1 (2.4) 531.3 (3.0) 933.7 (4.5) 781.5 (4.8) 916.3 769.7 289.1 (2.9) 531.2 (2.8) 933.5 (4.4) 781.3 (4.8) 916.4 770.0 289.1 (2.4) 531.2 (2.7) 781.2 (4.8) 770.4 289.1 (2.4) ~~ the full width at half maximum of the photoelectron peaks are reported. For all the samples the charging was corrected with reference to the C Is peak of the COi-group fixed 534 atomic sensitivity factor obtained using the intensity ratios [(metal 2p3/,/(oxygen 1 s)] recorded on the pure precursors by the expression Sco/Sc, =3(Ic0/10)/~Icu/~~)=8.09/10.7=0.76 Note that the measured intensity for Cu 2p3/2 and Co 2p3/2 peaks takes into account both the satellite and the main peaks.The water content of the compounds (see Table 1) does not influence the sensitivity factor appreciably (0.75 instead of 0.76). The surface Cu:Co ratios are reported in Table 1, column 7.3. Results and Discussion 3.1 Phase Analysis The X-ray pattern of the sample with atomic ratio Cu :Co = 1OO:O shows exclusively the reflections of the mineral mala- chite, CU~CO~(OH)~, hereafter labelled M." The precursors with Cu :Co compositions in the range 92 :8-67 :33 also give XRD spectra similar to malachite, at least in the sequence of the X-ray lines. Fig.2 shows as an example the observed spectrum for the compound with Cu :Co =85 :15, and super- imposed the lines of malachite." This observation, together with the evidence that no extra lines of other phases are detectable, points out that solid solutions of cobalt-containing malachite of chemical formula Cu2 -xCo,C03(0H)2 (hereafter labelled CoM) are formed in the range of Cu:Co= 92 :8-67 :33.It should be recalled that cobalt-containing malachites have indeed been found to exist also as natural products.'" The samples with Cu :Co atomic ratios 0 : 100 and 15 :85 give XRD patterns corresponding to that of the mineral spherocobaltite, CoCO3," hereafter labelled S. The XRD pattern of the precursor with Cu :Co = 15:85 (Fig. 3) matches very well with that of S, no other phases being detectable. It is inferred that Coo~ssCuo, 15co3copper-containing spheroco- baltite, hereafter labelled CuS, is achieved. 3.2 Diffuse Reflectance Spectroscopy The reflectance spectrum performed on the Cu :Co = 100:0 sample shows [Fig. 4(a)] two absorption bands at ca.8000 (1250) and 12 500 cm-' (800 nm). These bands are character- istic of Cu2+ (3d9 electronic configuration) in a distorted 1600 1400 -1 1200 I 1000 B 3 800 0 600 400 200 0 20 40 60 80 2810 Fig. 2 X-Ray pattern (Co-Kcr, radiation) of the Cu :Co =85 :15 com-position. The X-ray lines (26' and relative intensities) of the reference spectrum for malachite (ref. 1I) are also shown J. MATER. CHEM., 1991, VOL. 1 I6O0 1400 I 1200 I000 (I).L C5 800 0 600 400 i II IIIII Ill 10 20 40 60 80 201" Fig. 3 X-Ray pattern (Co-Ka, radiation) of the Cu :Co = 15:85 com-position. The X-ray lines (26' and relative intensities) of the reference spectrum for spherocobaltite (ref. 11) are also shown 1 I I 40 000 30 000 20 000 10 000 wavenumbers/cm-' Fig.4 Reflectance spectra of Cu-Co precursors.(a) Spectrum of Cu: Co = 100:0; (b) spectrum of Cu :Co =85 : 15; (c) spectrum of cu:co=o: 100 octahedral co-ordinati~n~"~~~-'~and attributed to the +dX2-y2 and dxy,xz,yz-+dx2 transitions, respectively.d,~ -,,z Another strong absorption starting at ca. 25 000 cm-' (400 nm) is due to the charge-transfer transitions from primar- ily ligand orbitals (bonding and non-bonding) to the highest dX2-y~orbital of copper.17 The reflectance spectrum of the Cu :Co =0 : 100 compound shows [Fig. 4(c)] two bands at ca. 8300 (1200) and 20 000 cm-' (500 nm) typical of Co2+ (3d7 electronic con- figuration) in octahedral symmetry and assigned to the tran- sitions "T1,+"T2, and 4T1, ("F)+"T,, ("P), respectively.18 Large absorption is exhibited for wavelengths shorter than 300 nm (33 000 cm-'). For the intermediate Cu:Co compositions all d-d tran-+ +sition bands, corresponding to both Cu2 and Co2 species are visible in all the materials. The spectrum for the Cu :Co = 85: 15 compound is shown in Fig.qb), as an example. It should be pointed out that with the increase of cobalt content, especially up to Cu:Co=67:33, the band of copper at 8000 cm-' is shifted towards higher energy owing to the superposition of a band of Co2+ at 8300 cm- ',and the peak of Cu2+ at 12 500 cm-' is shifted towards lower energy. The latter observation seems to signify that the presence of cobalt J. MATER.CHEM., 1991, VOL. I in the same structure of malachite induces a certain lengthen- ing in the Cu-0 bonds. 3.3 Magnetic Properties The experimental values of the Curie constants, C, magnetic moments, p, and Weiss temperatures, 8, for each Cu:Co composition are reported in Table 1. The values of C for pure malachite (Cu :Co = 100:0) and spherocobaltite (Cu :Co-0: 100) are 0.6 and 3.8, respectively, as expected for Cu and Co in the oxidation state of +2 (Cu2+ being in distorted- octahedral symmetry'" and Co2 + in nearly octahedral co- ordination"). The above C values for Cu2 + and Co2 +,and the respective copper and cobalt analytical molar fractions (Xcuis reported in Table 1, column 8) were used to calculate the relative C values for each intermediate composition by applying the additivity law formula: C =Cc,X +Cco(1 -X).All the calcu- lated C values (except that for the 15 :85 composition) agree fairly well with the experimental ones (Table l).This assures the presence of Cu2+ and Co2+ species in the mixed precursors. Table 1 shows also that the Weiss temperature, 8, (which specifies antiferromagnetic interactions present within the solid and is directly proportional to the average number of nearest-neighbour magnetic ions surrounding each magnetic ion) has the highest negative value (-200 K) in the pure malachite, and becomes less negative at the increase of cobalt content. This implies that Cu-0-Cu exchange interactions are magnetically stronger than those of Cu-0-Co or co-0-co.3.4 Structural Bulk Properties It was observed from the XRD spectra that many X-ray lines of the CoM and CuS solid-solution phases, in addition to having different intensities with respect to pure M and S, were systematically displaced with increasing concentration of for- eign atoms within the host malachite or spherocobaltite lattices (see Fig. 2 and 3 as examples of displacement). The unit-cell parameters, reported in Table 1, were determined for all compounds using the observed d spacings and the relative hkl Miller indices for each given reflection. It is evident from Table 1 that (i) the incorporation of Co2+ ions in the malachite monoclinic lattice gives a decrease of the cell volume, I/. In particular, with increasing Co content, both a and c parameters decrease and b increases.Since the p angle progressively decreases this means that the lattice becomes less and less distorted at the increase of Co content, i.e. comes nearer to 90" and the lattice approaches an orthorhombic one. This behaviour has been found by us also for the Zn-containing'" and Mn-containing" malachites. (ii) The introduction of Cu2 + in the spherocobaltite trigonal lattice gives a slight decrease of both a and c parameters, which provides an overall slight decrease in cell volume. As regards point (i) recall that the monoclinic lattice of M contains two types of non-equivalent Cu2+ ions: Cu of type I (Cu,) is surrounded by four oxygens belonging to C03 groups and two OH ions, whereas Cu of type I1 (Cull) is surrounded by two oxygens of the carbonate anion and four OH ions2' Both CU, and CuII are in distorted octahedral co- ordination, the mean values of the four coplanar and two apical interatomic Cu-0 distances being for Cu, 1.98 and 2.71 A, and for Cull 2.01 and 2.41 A.Both CuI06 and CulI06 distorted octaedra form strings in the c direction linked by edges. Co2+ can be present in the mineral mala~hite,'~ as we found also for Zn2+ and for Mn2+.'"910 Our study has demonstrated that, up to 33% cobalt content, monophasic solid solutions with a malachite structure are formed. It must be added that beyond the above value other metastable phases (the characterization of which is in progress) are produced.The observed decrease of the cell volume of the CoM samples with increasing cobalt content seems in contrast with expectations based on the isomorphous substitution of Cu2 + by Co2+ ions within the malachite lattice, since the octahedral ionic radii for Cu2 + and Co2 + (high-spin configuration) are 0.73 and 0.745 A, respectively.21 The unexpected volume shrinkage of the malachite monoclinic unit cell with incorpor- ation of Co2 + ions (bigger than Cu2 +)indicates that distortion of the MeO, polyhedra at the cation site of the malachite lattice is less when Cu2+ is replaced by Co2+. The decrease of local octahedral distortion (with a consequent overall decrease in the cell volume) is linked, in our opinion, to the different electron configuration of Cu d9 and Co d7 ions, the latter having less tendency than copper to give elongation of the oxygen octahedra.22 Recall that in the corresponding oxides, CuO and COO, Cu2+ and Co2+ are co-ordinated differently to oxygens, having nearly square-planar and octa- hedral symmetry, re~pectively.~~ It should be noticed that similar behaviour has been observed in zinc malachites'" and manganese malachites"" (in this case we observed a lower increase in the cell volume than that expected on the basis of a pure ionic radii difference).The presence of Co2+ in CoM solid solutions should increase the structural and thermal stability of the malachite structure, as already observed for ZnM'" and MnM." Some preliminary thermogravimetric (TG, DTG and DTA) findings confirm this hypothesis; CoM materials show a displacement (ca.80" for the 92: 8 compound) of decomposition to higher temperature with respect to pure M.As for point (ii) the lower volume observed for the Coo~85Cuo,15C03compound with respect to spherocobaltite, C0C03 (whose structure is similar to calcite and contains cations in nearly octahedral ~o-ordination)~~ agrees with ionic radii considerations since the substitution of slightly larger Co2+ ions (ionic radius=0.745 A) by Cu2+ (0.73 A) induces an overall shrinkage of the lattice. 3.5 Surface Properties The presence of Cu2 + and Co2 + species has been ascertained by X-ray photoelectron spectroscopy (XPS) at the surface of the materials. Table 2 reports the binding energies and the kinetic energies of the 2p33,2 photoemission and L3M45M45 Auger transitions for both copper and cobalt.In the same table the binding energies of 0 1s and C 1s (carbon atom of the carbonate group) and the full width at half maximum, FWHM, for all photoelectron peaks (in parentheses) are reported. The 2p3,2 transitions of Cu2+ and Co2+ ions present, beside the main peak, a satellite peak at higher binding energy. The origin of such satellites has been the subject of much discussion in the literature. For Cu2+ the satellite and the main peaks are related, respectively, to the 2p53d9 and 2p53d"L electronic configurations in the final state, where L refers to a ligand hole [see ref. 9 (d) and references therein].For Co2+ the interpretation of both satellite and main peaks should be similar and related to the 2p53d7 and 2p53d8L electronic configurations in the final state, respe~tively.~~-~~ It has also been reported that for Cu2 +-containing compounds the satel- lite-main peak energy separation increases and the satellite- main peak intensity ratio decreases with increasing covalent character of the metal-ligand bond.24 The presence of the satellites (at ca. 9 and 5 eV from the main peaks, respectively, for copper and cobalt)27in the XP spectra of the pure malachite and spherocobaltite as well as of the solid solutions (both Cu, -xCoxC03(OH)z and co~~8~c~o~~~co3)thus becomes clear evidence of the oxi-dation state of +2 for cobalt and copper in all the studied materials.The 2p3,, satellite-main peak separation and their relative intensity ratio (ca.0.7 and 0.5 for Cu2+and Co2+,respectively) remain constant, within experimental error, for both Cu and Co in all compounds. Table2 shows that the 0 1s binding energy is constant within experimental error. These results suggest that the local nature of the metal-oxygen bond is essentially determined only by the nature of the first-neighbour ligands. In Fig. 5 the Cu:Co atomic ratios in the surface region (XPS) as well as in the bulk (chemical analysis by atomic absorption) are reported for the hydroxycarbonate-carbonate solid solutions. In the figure the straight line with slope equal to unity indicates that the surface composition reflects bulk composition for both cobalt and copper. The results reported in Table 1 and Fig.5 clearly show that up to Cu:Coz6 (sample 85:15)the surface region is slightly enriched in cobalt, whereas for the sample 92 :8 there is strong evidence for high cobalt enrichment at the surface. We interpret these results in terms of strain-energy effects.The Co2+ions are bigger than the Cu2+ and prefer less distorted octahedral symmetry (highly distorted MeOs polyhedra are present in the pure malachite). As a consequence, Co2+ ions tend to segregate at the surface, the driving force being the strain energy term of the segregation enthalpy. However, the hypothesis of higher reactivity of Co2+ ions with moisture cannot be excluded (chemical driving force).For the sample with atomic ratio Cu: Co =15:85 (Cu2+ incorporated into the spherocobaltite lattice) high homogen-eity between surface and bulk has been observed. 4. Conclusions The results obtained can be summarized as follows. (1) At high Cu:Co atomic ratios (up to 67:33) single phases of Cu, -xCoxC03(0H), malachite-like solid solutions are formed; the incorporation of Co2 ions into the malachite+ 12110 / 92:8 0 : p 3 , , , , I 15 :85 2 4 6 8 1012 Cu : Cu (AA) Fig. 5 Cu: Co compositions at the surface (XPS) and in the bulk (chemical analysis by atomic absorption) for hydroxycarbonate mala-chite-like compounds. The straight line (slope= 1) represents the homogeneous distribution between surface and bulk J.MATER. CHEM., 1991, VOL. 1 lattice produces less distortion in the MeOs polyhedra, higher structural and thermal stability, and Cu-0-Co exchange antiferromagnetic interactions weaker than the initial Cu-0-Cu ones. (2) Spherocobaltite, C0C03, and coO~~~c~o~~~co3copper-containing spherocobaltite are formed at high cobalt loading. (3) The surface is enriched in cobalt (much more at low Co content) in the malachite-like solid solutions owing to the strain-energy contribution, arising from size and electronic effects (Co2+ is bigger than Cu2+, and prefers less distorted octahedral symmetry). (4) Surface composition reflects bulk composition for both copper and cobalt in the Cuo.15Coo.s5compound. We gratefully acknowledge support from the Italian CNR 'Progetto Finalizzato Chimica Fine 11'. References 1 S.Angelov, D. Mehandjiev, B. Piperov, V. Zarkov, A. Terlecki-Baricevic, D. Jovanovic and Z. Jovanovic, Appl. Catal., 1985, 16, 431. 2 K. Fujimoto and T. Oba, Appl. Catul., 1985, 13, 289. 3 X. Xiaoding, E. B. M. Doesburg and J. J. F. Scholten, Catal. Today, 1987,2, 125. 4 J. E. Baker, R. Burch and S. E. Golunski, Appl. Catal., 1989, 53, 279, and references therein. 5 (a) A. Sugier and E. Freund, US.Pat., 4 122 110, 1987; (b) P. Courty, D. Durand, E. Freund and A. Sugier, J. Mol. Catal., 1982, 17, 241. G. Fornasari, S. Gusi, F. Trifiro and A. Vaccari, Znd. Eng. Chem. Res., 1987, 26, 1500, and references therein. D. Waller, D. Stirling, F.S. Stone and M. S. Spencer, Faraday Discuss. Chem. SOC.,1989, 87, 107. (a)L. Volpe and M. Boudart, Catal. Reu. Sci. Eng., 1985, 27, 515; (b) H. R. Oswald and J. R. Gunter, in Crystal Growth and Materials, ed. E. Kaldis and H. J. Scheel,North-Holland, Nether-lands, 1977, p. 415. 9 (a) P. Porta, S. De Rossi, G. Ferraris, M. Lo Jacono, G. Minelli and G. Morreti, J. Catal., 1988, 109, 367; (b) P. Porta, G. Fierro, M. Lo Jacono and G. Moretti, Catal. Today, 1988, 2, 675; (c) P. Porta, R. Dragone, M. Lo Jacono, G.Minelli and G. Moretti, Solid State Zonics, 1989, 32/33, 1019; (d) G. Moretti, G. Fierro, M. Lo Jacono and P. Porta, Surf. Interface Anal., 1989, 14, 325: (e) G. Moretti and P. Porta, J. Phys. Condens. Matter, 1989, 1, 193; (f)G.Moretti and P. Porta, Surf. Interface Anal., 1990, 15, 47. 10 (a) P. Porta, G. Moretti, M. Lo Jacono, M. Musicanti and A. Nardella, J. Mat. Chem., 1991, 1, 129; (b) P. Porta, G. Moretti,M. Musicanti and A. Nardella, Catal. Today, 1991, 2, 21 1. 11 X-ray Powder Data File, ASTM 10-399, malachite; ASTM 11-692, spherocobaltite. 12 N. G. Farr and H. J. Griess, J. Electr. Spectrosc. Relat. Phenom., 1989, 49, 293. 13 J. G. Dillard and M. H. Koppelman, J. Coll. Interface Sci., 1982, 87, 46. 14 M. Deliens, R. Oosterbosch and T. Verbeek, Bull. SOC. Fr. Mineral. Cristallogr., 1973, 96, 371. 15 V. F. Anufrienko, T. M. Yurieva, F. S. Hadzieva, T. P. Minyukova and S. Yu. Burylin, React. Kinet. Catal. Lett., 1985, 27-1, 201. 16 F. H. Chapple and F. S. Stone, Proc. Br. Ceram. SOC.,1964, 1, 45. 17 (a) R. D. Willet, L. Liles Jr. and C. Michelson, Znorg. Chem., 1967, 6, 1885: (b) P. Porta and A. Mazzarano, in Reactivity of Solids, ed. P. Barret and L. C. Dufour, Elsevier, Amsterdam, 1985, Part B, p. 1053. 18 F. Pepe, M. Schiavello, G. Minelli and M. Lo Jacono, 2.Physik. Chem. NF, 1979, 115, 7, and references therein. 19 A. Cimino, M. Lo Jacono, P. Porta and M. Valigi, 2. Physik. Chem. NF, 1970,70, 166. 20 A. F. Wells, Acta Crystallogr., 1951, 4, 200. 21 R. D. Shannon, Acta Crystallogr., Sect. A., 1976, 32, 751. 22 J. D. Dunitz and L. E. Orgel, Nature (London) 1957, 179, 462. 23 L. Bragg and G. F. Claringbull, in Crystal Structures ofMinerals, Bell, London, 1965. J. MATER. CHEM., 1991, VOL. 1 537 24 25 26 G. van der Laan, C. Westra, C. Haas and G. A. Sawatzky, Phys. Rev. B, 1984, 29,4401. B. W. Veal and A. P. Paulikas, Phys. Rev. B, 1985,31, 5399. T. A. Carlson, J. C. Carver, L. J. Saethre, F. G. Santiba’nez and 27 C. D. Wagner, in Practical Surface Analysis by Auger and Photoelectron Spectroscopy, ed. D. Briggs and M. P. Seah, Wiley, Chichester, 1983, pp. 477-509. G. A. Vernon, J. Electron. Spectrosc. Relat. Phenom., 1974, 5, 247. Paper 0/05767C; Received 27th December, 1990