Structural study by energy dispersive X-ray diffraction of amorphous mixed hydroxycarbonates containing Co, Cu, Zn, A1 Marilena Carbone,' Ruggero Caminiti*yb and C. Sadud 'Dipartimento di Chimica, Universita' di Roma La Sapienza, P.le A. Moro, 5, 00185 Roma, Italy bIstituto Nazionale per la Fisica della Materia, Universita di Roma 'La Sapienza', P.le A. Moro, 5, 00185 Roma, Italy A non commercial 8-8 energy dispersive X-ray diffractometer equipped with a solid state detector has been used for the determination of the amorphous structures of nine mixed hydroxycarbonates containing Co, Cu, Zn and Al. Energy dispersive X-ray diffraction (EDXD) provides information about the local order around the divalent metal ions and Al"'. An octahedral configuration around all the cations was found.Debye functions were calculated and compared with the experimental curves. Large angle X-ray scattering (LAXS) is a powerful technique for determining the structural parameters of liquid and amorphous systems, since it can provide information about the short-range order.'.' In particular, energy dispersive X-ray diffraction (EDXD)394 has been found to be a suitable tool for the investigation of such systems owing to its speed and reliability compared to a traditional angular scanning diffractometer. We have used the EDXD technique to determine the struc- tural properties of amorphous mixed Cu-Zn-Co-A1 hydroxy-carbonates, and the parameters which can influence the degree of crystallinity of such samples. It is important to establish which parameters, during the preparation, influence the struc- tural characteristics of copper mixed hydroxycarbonates, since these compounds are precursors for alcohol synthesis catalysts as well as for CO oxidation, namely the corresponding mixed oxides generally obtained by thermal a~fivation.~-'' A great deal of work has been devoted to the structural characterization of crystalline mixed hydroxycarbonates con- taining copper and one or more metal ions such as Zn, Co and Al,l2-I5 but no attempt to characterise the corresponding amorphous samples has been reported.Providing an investi- gation method for amorphous materials is even more important because of the growing interest in copper mixed amorphous materials for use as methanol synthesis catalysts.Wright et al." investigated the structural properties and the catalytic activity for CO oxidation of amorphous Cu-Co-Mn mixed oxides obtained by keeping the activation temperature of the corresponding carbonates low. Coteron and HayhurstI6 performed structural characteriz- ation and catalytic activity tests in methanol synthesis of Cu-Zn, Cu-Zr and Cu-Zn-A1 amorphous systems obtained by the spark erosion technique. Because of the important role played by the precursor structures in the catalytic activity, we were interested in determining the critical cation concentrations in some ternary and quaternary mixed hydroxycarbonates which lead to the formation of amorphous mixed hydroxycarbonates. It is also important to determine the structural parameters of the synthe- sized samples, namely the coordination number and the first coordination shell of the cations, and the influence of some preparation parameters, such as the drying method, on the sample structures.Coprecipitation from nitrate or acetate mixed-cation solu- tions with NaHCO, is a common synthesis method for crystal- line copper-based binary hydroxycarbonate~.~~.~~Binary systems such as Cu-Zn, Cu-Co and Cu-Mn have been obtained by coprecipitation and well charaterized from a structural point of view.12914 Different coprecipitation pro-cedures are reported as well, which lead to different crystalline phases. One important parameter is the addition sequence; different phases are obtained if the nitrate (or acetate) solution is added to the hydroxycarbonate or vice versa. If we refer to a Cu-Zn binary system, a malachite-like phase [Cu2(OH),C03] is obtained in the first case, while gherardite [Cu2(0H),N03] is obtained in the second." The cation ratio is also a critical parameter.If we consider the same Cu-Zn system, a Cu/Zn atomic ratio of 5.7 is the critical value above which a malachite-like phase forms. At higher Zn concen-trations an aurichalcite phase [(Cu,Zn), (OH),( CO,),] forms together with the malachite-like phase." The presence of a cation with higher oxidation number can also influence the crystalline phase of the sample. A hydrotal-cite-like phase [M116M111Z(OH)16C03.4Hz0]forms'3 when the ratio of divalent and trivalent cations is 3 :1.However, the preparation procedures reported in the litera- ture" do not always lead to the formation of crystalline samples. Marchi et aLZ0 studied the parameters that can influence the crystallinity of Cu-Co-A1 samples obtained by coprecipitation. Among the important parameters they quoted the ratio between divalent and trivalent cations, the pH of the solution and the supersaturation ratio. The structural properties of the mixed oxides, their homogen- eity and cation interdispersion, which contribute to the deter- mination of the efficiency of the catalytic process, are greatly affected by the structural properties of the corresponding precursor^.^-^ We considered two series, one of ternary and one of quatern- ary samples, with varying Co concentrations (3-20%), and the effects of this variation on the physical state of the samples was investigated.Moreover, A1 was present in the series of quaternary samples at a constant percentage (9%0), a value lower than that at which crystalline hydrotalcite forms (25%0)." Materials The compounds studied were prepared by coprecipitation, according to the procedure reported in refs. 12-15. Atomic ratios are reported in Table 1. Surprisingly, although we adopted the same preparation procedure for the synthesis of crystalline hydroxycarbonates, we obtained almost always amorphous samples. In a typical preparation 0.6 1 of an 0.7 mol I-' aqueous solution of the cation nitrates in suitable proportions was added to 0.9 1 of NaHCO, (1.1 moll-') at 323 K under vigorous stirring.The slurry was digested for 4 h under the same conditions. The pH just after the precipitation was 6.5-7.0 and it increased to 8.5-9.0 at the end. The precipitates J. Muter. Chem., 1996, 6(lo), 1709-1716 1709 Table 1 Atomic ratios in the studied compounds sample cu Zn co A1 czco 68 29 3 --CZCl 63 27 10 NOAL 63 18 19 NOAL1 56 24 20 CZCAO 63 26 3 9 CZCAl 60 25 6 9 CZCA2 57 25 9 9 CZCA4 55 23 13 9 CZCA3 51 22 18 9 MACA 50 22 19 9 MARUL 50 22 19 9 were pale blue; they turned green after 2 h. The samples were washed with 8 1 of distilled cold water (to eliminate Na' and NO,-ions), dried at 363 K and finally ground in an agate mortar.In order to determine the effect of the preparation conditions, the coprecipitation of the sample containing Cu :Zn :Co :A1 in atomic ratio 51 :22 :18:9 (CZCA3) was carried out at a higher temperature, (373 K). Moreover the sample containing Cu :Zn :Co :A1 in atomic ratio 50 :22 :19:9 was prepared twice. In the first preparation the conventional procedure was followed (MARUL sample), in the second the drying of the washed precipitate was performed over CaC1, (MACA sample). Techniques Elemental Cu, Zn, Co, A1 analyses were performed by atomic absorption with a Varian Spectra AA30 instrument and Table 2 reports the analytical results referring to the nine amorphous samples studied. The anion contents of the hydroxycarbonates were determined by evaluating the amount of H,O and CO, produced in a thermal decomposition in a flowing system connected to a traditional BET vacuum line.The X-ray diffraction experiments described here were car- ried out by employing a non-commercial X-ray energy scanning diffractometer,' equipped with an X-ray generator (water cooled, W target with 3.0 kW maximum power), a solid-state detector (SSD) connected to a multichannel by means of an electronic chain collimator system, step motors and a sample holder. The X-ray source is a standard Seifert tube, used at 50kV and 40mA or 45kV and 35mA, whose white Bremsstrahlung component was used. The primary beam inten- sity Io(E)was measured directly using the same voltage (50 or 45 kV), by reducing the tube current to 2 mA at the zero scattering angle without the sample.The transmission of the samples was measured under the same conditions; from the following equation we obtain the experimental value exp[-p(E)]t that it is used in eqn. (6) and (7) for the absorption corrections.22 The fluorescence lines present in the 5-11 keV range due to Table 2 Element concentration in the prepared hydroxycarbonates ~~~ sample Co/moll-' Cu/mol 1-1 Zn/moll-' Zn/Cu Al/moll-' CZCl 3.61 1 17.399 6.894 0.396 -NOAL 3.978 13.290 3.978 0.288 -NOAL1 5.593 14.221 5.622 0.395 -CZCAl 1.604 12.03 1 5.615 0.467 1.872 CZCA2 2.466 12.054 4.657 0.386 2.192 CZCA4 3.397 11.237 4.181 0.372 2.091 MACA 3.647 9.899 4.428 0.447 2.084 MARUL 3.647 9.899 4.428 0.447 2.084 CZCA3 5.590 11.740 4.752 0.405 2.516 1710 J. Muter.Chew., 1996,6( lo), 1709-1716 Fig. 1 The energy scanning diffractometer used for the measurements Table 3 Scattering parameters associated with the minimum and maxi- mum values of the used energy for each measurement angle 17-38 keV 17-42 keV angle/degrees smm Smax Snun Srnax 21.0 6.18 13.80 6.18 15.26 15.5 4 61 10.30 4.61 11.38 10.5 3 14 7.02 3 14 7.39 8.0 2.40 5 36 2.40 5 92 5.0 1.50 3.36 1.50 3.71 3.5 1.05 2.35 1.05 2.60 3.0 0.90 2.01 0.90 2.22 2.0 0.60 1.34 0.60 1.48 1.5 0.45 1.oo 0.45 1.11 1.0 0.30 0 67 0 30 0.74 W, Co, Cu, Zn and, to a lesser extent, Al, do not disturb our measurements, since they are outside the region of our interest.A Seifert and Rich high-voltage power supply (stability -=0.1%) was used. The detecting system consisted of a EG&G liquid-nitrogen cooled ultrapure Ge SSD (ORTEC, model 92X) connected to a PC 286 via ADCAM hardware. The current pulse from the detector was converted into a digital signal, which was visualized on a computer screen through a multichannel analyser (MCA). The collimating system was composed of four adjustable-width W slits purposely placed to reduce the X-ray beam angular divergence. The X-ray tube and detector holding arms rotated in the vertical plane around a common centre in order to reach the desired 26 scattering angle; the movement was accomplished by step motors which allowed reproducibility within 0.001 O for the scattering angles.The diffractometer is shown in Fig. 1. The working conditions used were as follows. Supply: high voltage =50 kV, current 40 mA, total power =2000 W for the samples CZC1, CZCA1, CZCA2, CZCA3 and CZCA4, or 45 kV, 35 mA, total power =1575 W for the samples MACA, MARUL, NOAL and NOAL1. Measurement angles and the energy ranges used are reported in Table 3. Using the formula s= 1.014 E sin I!? it is possible calculate the s range of the angles used. The powder samples were formed into pellets in order to perform the measurements. The total intensity 1scattered by a sample and observed by an energy dispersive detector in approximation of single scat- tering and transmission geometry4,,, can be expressed as: with where 8 is the scattering angle, E is the photon energy revealed by the detector and E‘ is the initial energy of a photon scattered inelastically at the observed energy E.From Compton’s incoherent scattering we have the expression r 1 1E’=E K is the scale factor between the intensity reaching the detector and the intensity scattered by a stoichiometric unit of the sample. Io(E) is the energy spectrum of the primary beam measured at 8= 0” (as described previously). P(E,8) is the polarization factor by a scattering of a primary radiation with polarization @(E)that is: P(E,8)= ( 1+ COS’ 28)/2+ sin228 @ (E)/2 (4) with @ (E)= C(lp,n(E) -Ip.p(E )l/CIp,n(E)+ Ip,p(E)l ( 5) where lp,nand Ip,pare the intensities of the normal and parallel polarization components, respectively, with respect to the scattering plane.Acoh(E,8) is the X-ray elastic absorption coefficient Ac0h(E,8)= exp [ -p(E) t sece] (6) A,,,(E,E’,B) is the X-ray inelastic absorption: exp [ -p(E) t sece] -exp [ -p(E’) t sed?]A,nc(E,E’,o) = -p (E) t sec8-p (E’)t sec8 (7) where p is the absorption coefficient and t the thickness of the pellets. Icoh(E,8) is the total elastic scattered intensity: 1Coh(E*8)=c cnfn2(s)+i(s) (8) n IInc(E,8)is the incoherent contribution to the total scattered intensity.1c,f: (s) is the self scattering intensity; c, is the concen- n tration of the different species, i(s) is the intensity of interfering waves scattered by atom pairs, s is the scattering parameter and is defined by 471 sin 8 s= A = 1.014 E sin 8 (9)~ when E is expressed in keV and s in A-’.Data treatment After correction of the experimental data for the escape peak suppression, the intensity data were handled by means of our DIFl program written in FORTRAN IV. This program also made the necessary absorption corrections to combine the -----21.r-159 10s * 8.W---5.0. 3.5’ 3.0’ . 2.w .. . . . . . -1.5’ . 1.0’ 0 2 4 6 8 10 12 14 s1A-l Fig. 2 Picture of the scattered intensity (e.u.) for the consecutive measurements angles various angular data sets, as described in the paper by Nishikawa and Iijima.” The rescaled intensity, in electron units (e.u.), for each scattering angle used is shown in Fig. 2 for the sample NOAL.Normalization to a stoichiometric unit of volume containing one Co atom was performed. Table 4 reports the elemental normalized concentrations. Radial distribution functions, D(r), were calculated from the static structure functions i(s): i(s)= ICoh(E,8)-1CL2(s) ( 10) n according to the expression: D(r)= 4nr2p0+ 2rn-’ s-i(s)-M(s)sin(rs)ds (11) In this equation po= [1nif,(0)]2V-’,where V is the stoichio- 1 metric unit of the chosen volume, niis the number of atoms i per unit volume, andfi is the scattering factor per atom i. M(s)= ~f2~o~~~~zco~~~~~~~P~-0.010 s2) ( 12) is the sharpe$ng factor. For the upper integration limit s,, we used 14.5 A-’. Theoretical peaks were calculated by a corresponding Fourier transform of the theoretical intensities for pairs (p, q) of interactions (Debye functions): using the same sharpening factor and the same s,,, value as for the experimental data and assuming the rms variation in the interatomic distance to be opq.Data analysis The observed structure functions, in the form s-i(s)-M(s),for both series of samples are reported in Fig. 3(u)and (b).In the figures, the Al-containing samples are ordered by decreasing Table 4 Element concentrations, density (d) and stoichiometric volume (V)[the concentrations are given as number (n,)per stoichiometric unit of volume. I/ was chosen, in all cases, to correspond to a value containing one Co atom] sample co cu Zn A1 Zn/Cu 0 C H d/g cm-3 v/A3 CZCl 1.o 4.8 183 1.9092 -0.369 22.3653 3.6367 11.4556 3.2830 495.8541 NOAL 1.o 3.4667 1.oOOo -0.299 15.5334 2.3333 8.5333 3.3000 335.5612 NOALl 1.o 2.3889 0.9444 -0.395 14.7778 2.3333 7.7778 3.3072 278.9758 CZCAl 1.o 7.500 3.500 1.1667 0.467 45.500 1 7.1667 24.000 1 2.6737 1035.1 160 CZCA2 1.o 4.888 1.8889 0.8889 0.386 3 1.7778 4.8889 17.1 1 11 2.7397 673.4508 CZCA4 1.o 3.3077 1.2308 0.6154 0.372 21.9231 3.4615 11.5385 2.6134 488.7652 MACA 1.0 2.7143 1.2 143 0.5714 0.447 18.2143 2.5714 10.5000 2.6051 455.3012 MARUL 1.0 2.7143 1.2143 0.5714 0.447 18.2143 2.5714 10.5000 2.4848 477.3485 CZCA3 1.o 2.1000 0.8500 0.4500 0.405 12.0500 1.2500 8.300 2.7953 297.0258 J.Muter. Chew., 1996,6(10), 1709-1716 1711 2 x lo4 1.5 x lo4 5000+ ? J0.tNOAL 1 0 3 6 91215 0 3 6 91215 1.5 x lo42x1044 ikZCA 1 CZCA 4 -1.5 x 1O4 ' f ' ! ' 1 0 3 6 91215 0 3 6 91215 0 3 6 91215 t -. -l""i.-. J,,, 'b' ,,, '4',, ,~~,,,l 0 MARUL CZCA 3 I,,,, ,,*I, I,,,. It,,, I,,,,-1.5 x lo4 I I I I 0 3 6 91215 sIA-' Fig. 3 Observed structure functions s.i(s).M(s)(in em. A-')for (a) samples not containing A1 and for (b) samples containing A1 A1 content calculated per unit-cell volume of Co. The samples without A1 are ordered, instead, by increasing Co content. The radial distribution functions in the DIFF(r) =[D(r)-4zr2p,] form are reported in Fig. 4(u) and (b). The main feature which emerges from the s-i(s)M(s)analysis of the samples containing no A1 is the loss of structure of the oscillations at increasing Co content.The amorphous sample at low Co concentration (CZC1) shows structured oscillations over the whole s range examined, while these structures gradually decrease in intensity in the samples with higher Co contents. The quaternary samples show a similar behaviour, namely a loss of structure is displayed in the si(s)-M(s)functions at increasing Co content for samples CZCA1, CZCA2 and CZCA4. This behaviour is not followed by the MACA, MARUL and CZCA3 samples, since the s.i(s)-M(s)functions of these samples seem to be more structured. We assume, therefore, that other factors can influence the local structure, as will be discussed in the following section.1712 J. Muter. Chem., 1996, 6(lo), 1709-1716 The radial distribution functions in the DIFF(r) form, obtained from the scattered intensities with the previously described proFedure, show the prFsence in all the samples of a peak at 2.0A and one at 3.2A. The first peak has been attributed to the metal-oxygen interaction, the second one to the interaction between metal ions. Peaks at higher r values are expected to be produced in the DIFF curve by the interaction between cations bridged by an oxygen atom. Therefore the peaks at 5.4, 6.2 and 8.4A, that appear in the DIFF curves have been mainly attributed to interaction of metal atoms belonging to adjacent coordination polyhedrons. More detailed information on the studied systems can be obtained by the analysis of the radial distribution function D(r), which is the area of each peak which is proportional to the number of scattering atoms and to their scattering factors.We have focused our attention on the peak at 2.0A, which provides information about the cations first coordination shell. Therefore, by comparing the experimental peak with the theoretically calculated peak shape, the coordination numbers 401-i20 10 -1 0 -3 0 0 2 -4 6 3 8 1 1 0 -20/NoA;ll I ' I I ~ -30 0 2 4 6 I ' ~, ' I 8 1 (b) 403 40: 40)30 30120 I h L Y Q 10 .'20 CZCA ? -30; 0 2 4 6 8 1 0 0 2 4 6 8 1 0 0 2 4 6 8 10 40 c 301 1 40q30 4 ..401 20'Ol : * *5: -* 0 .... f : ' .* */ :: ... 0 . .. ,. ... .* .. 0.... .* r' 0... i -30 0246810 0246810 024681rlA-' Fig. 4 Radial distribution functions, in the form D(r)-4471p2p, (in e2 k3x lop3)referring to (a)the series of samples containing A1 and (b) the series of samples not containing A1 and the bond distances of all the cations have been derived for each sample. The experimental D(r)peak shapes compared to the theoretical peak shapes are reported in Fig. 5(u) and (b). Results and Discussion The first interesting feature emerging from the analysis carried out on the prepared hydroxycarbonates is the possibility of obtaining either crystalline or amorphous samples under similar preparation conditions. The conventional X-ray powder diffraction patterns for all prepared samples showed that some samples had showed that some samples have an amorphous or low crystallinity pattern, without evident Bragg peaks, while other samples have a crystalline structure (CZCO, CZCAO).All the diffraction pat- terns, obtained with Cu-Ka radiation, are reported in the thesis of M. Carb~ne;~~we have studied by EDXD only those with no evident Bragg peaks. If the samples do not contain Co, or if they contain Co at low concentrations (3%, CZCO, CZCAO), their structures are completely crystalline. In samples without Al, a malachite-like and an aurichalcite phase are formed. The presence of A1 in these hydroxycarbonates favours the formation of hydrotalcite together with the usual malachite-like phase.The degree of crystallinity changes when a higher concentration of Co is present in the samples. We varied the Co concentration in the samples in a discrete way, so we do not know exactly the critical Co concentration. However, the samples containing 8 ?lo Co (series without Al) or 6% Co (series with Al) are completely amorphous, suggesting that Co concentration is a fundamental parameter affecting the crystallinity of the samples. The Co critical value must be in the range 3-6% when both Zn and Cu are present in the hydroxycarbonates. This effect is even more surprising if we consider that binary Cu-Zn or Cu-Co as well as quaternary Cu-Zn-Co-A1 hydroxycarbonates with similar cation proportions are ~rystalline.l~-~~ J.Muter. Chem., 1996,6(lo), 1709-1716 1713 60. 60 . ... 60 . 50-r 50-7 50-r ..-40-7 40-r 40-r 30-1 30-: L 20-I 20-: lo-: lo-: :NOAL : NOAL1 0-0. -1 0 I1I1IItl II I II III I a ILILlII I II IIt -10 lI1lI1lllI1lllI1llII1lllIII1fIL1ll I I1II I I I I "I I -1 0--I II I ' I I' I 'III I I ; I 40:.i; 30-: 30f 20-r lo-: CZCA2 CZCA 4 -1 0~ -1 0i 60. 60 60. ...* SO-: '* SO-: 2.... 50-r .. 40-r 40-: 40-: 30-r 30-: 30-: * * 20-: 20-: lo-; MACA lo--CZCA 3 0. 0. -1 0 lI1l~ILIIILI1l~llllIfI1lI1lll;llllo-=' I I f ' I" I' I I I II1' I " I---10-,' I I j fI I 1 I I I I I I I I I II I i t I I I I-1 rlA Fig. 5 Theoretical (solid line) and experimental (dotted line) radial distribution functions D(r) (in e A-3 x of (a) samples containing A1 and (b)samples not containing A1 A further increase of the Co content favours an increase of MACA and MARUL, which have similar cation concen-the amorphous character as shown by the loss of structure trations but were dried by different methods, show similar in the si(s).M(s)functions for the samples without Al.Therefore s-i(s).M(s)functions, suggesting that this step of the preparation the Co content is a parameter that controls not only the procedure does not influence the final sample structures. formation of amorphous samples but also the extent of their Quite different is the case of the CZCA3 sample, in which amorphous character. the relatively high temperature of preparation might have Even A1 seems to have an effect on the amorphous state of favoured the formation of a more crystalline sample.the samples. In comparing the structure functions of two The metal-oxygen distances, the coordination numbers and samples with similar cation proportions but different A1 con- the standard deviations used in reproducing the first pfak of tents [CZCl and CZCA2 in Fig. 3(u) and (b)respectively] we the experimental radial distribution function (at 2.0 A) are see that the former, which shows less structured oscillation, is reported in Table 5 for all the samples, and the values for therefore more amorphous. sample CZCA3 are given in Table 6. The chosen distances are LongestAnother parameter which can influence the degree of crystal- averaged over the values reported in the literat~re.~~-~* linity of the mixed hydroxycarbonates is the Zn/Cu ratio.We and shortest literature values are collected in Table 7. can see this effect, by comparing the s.i(s)-M(s)of two similar We assumed a coordination number of 6 for both Co and samples, in this case, two samples with similar Co concen- Cu on the basis of diffuse reflectance spectra, whose absorption trations: CZCA4 and MARUL. The second one, with a higher bands clearly indicate the presence of the cations in octahedral Zn/Cu ratio, shows a more structured s.i(s).M(s).The samples and distorted octahedral symmetry, respectively. 1714 J. Muter. Chem., 1996, 6(lo), 1709-1716 Table 5 Metal-oxygen distances (r), number of nearest neighbours (n) and standard deviation (a)used for all the samples r/A n CIA 1.98 0.1 2.50 0.1 2.07 0.1 2.12 0.1 1.88 0.1 Table 6 Metal-oxygen distances (r),number of nearest neighbours (n) and standard deviation (a)used for sample CZCA3 r/A n CUP 1.95 4 0.1 CUa 2.80 2 0.1 Table 7 Lowest (rmln) and highest (rmax) metal-oxygen distances reported in the literature 1.88 2.04 2.23 3.97 2.07 2.15 2.10 2.20 1.85 1.98 We performed two different peak shape calculations, using 4 and 6 as the coordination number of the Zn cation.The peak area is, in all the cases, lower when coordination 4 is used [Fig. 5(a) and (b)],thus indicating that 6 is the actual coordination number for Zn.Fig. 6(a) and (b) compare the experimental curves and the theoretical peak shapes using 4 and 6 as the coordination number for zinc for the samples CZCAl and CZCl respectively. Obviously the coordination 50 -1 A -1 ot-s 0 0.5 1 1.5 2 2.5 3 3.5 Q (blL I 4050f -1o>l 0 0.5 1 1.5 2 2.5 3 3.5 r /A Fig. 6 Radial distribution functions D(r) (in e2 k3x of (a) the sample CZCAl and of (b) sample CZC1: experimental curve (dotted line); theoretical curve using 6 as coordination number for the zinc (solid line); theoretical curve using 4 as coordination number for zinc (dashed line). numbers 4 and 6 used are given with an error between 5 and lo%, as reported in the literature for this technique. It is worthwhile noting that the cation coordination could be deduced even without the aid of the reflectance spectra.A good agreement between theoretical and experimental peaks is achieved only when 6 is used as the coordination number for all of the cations. The use of coordination number 4 for any of them always gives a theoretical peak shape lower than the experimental one. When calculating the theoretical peak shapes, we obtained the best agreement with the experimental curves using two different copEer-oxygen bond lengths. Indeed, the experimental peak at 2.0A shows a small asymmetry at higher r values, which can be easily attributed to a greater Cu-0 distance due, in six-fold coordination of Cu, to two oxygen atoms Jahn-Teller distorted away from the central metal.The coordination distances we found for Co and Cu are very close to those reported by Wright et al." referring to Cu-Mn and Co-Mn crystalline carbonates with the rho- docrosite (MnCO,) structure, investigated by EXAFS. The two different Cu-0 distances, both planar and apical, reported for malachite12 are not detected in the amorphous mixed hydroxycarbonates. Two different sites are present in malachite,oone having planar and apical bond distances of 1.98 and 2.71 A re$pectively, the other having bond distances of 2.01 and 2.41 A. The presence of Co or Zn in a solid solution with malachite, although varying the bond distances does not remove the duplicity of the ~ite.'~,'~ Only one kind of site is present, instead, in the amorphous samples studied, with a bond length which is nearly the average of the malachite sites.We checked that only one site contributes to the Cu-0 coordination, by calculating a peak shape with the Cu-0 distances reported for malachite, but the agreement with the experimental D(r) was worse. Different values of Cu-0 coordination distances had to be used to reproduce the peak of the CZCA3 sample, which, as explained before, was prepared at a higher temperature. In particular, this sample has shorter planar and longer apical distances compared to the other samples. This might indicate the presence of small copper oxide particles, whose formation was favoured by the high temperature during the preparation. This hypothesis is confirmed by the reflectance spectrum, which shows the typical CuO absorption at 12500cm-l, due to the transition between the valence and conduction bands.Finally, it is interesting to note that identical parameter values were used for all the samples, irrespective of the cation ratios. This implies that the first coordination shell of the amorphous mixed hydroxycarbonates is not significantly affec- ted by the cation concentrations. On the basis of literature findings2' and of the experiments carried out, we can conclude that the coprecipitation method for hydroxycarbonate preparation can be used in the synthesis of amorphous samples as well, the determining parameter being the cation ratios. EDXD has been employed in order to study some Cu-Co-Zn-A1 mixed hydroxycarbonates with fixed A1 content and variable Co concentration, prepared by coprecipitation. Co has a primary role in determining the physical state of the samples. The presence of Co in the ternary and quaternary mixed hydroxycarbonates can reduce long-range order.The critical content of Co inducing the formation of amorphous samples lies in the range 3-6%. Moreover, increasing Co content together with the presence of A1 contribute to a lowering of the degree of order in the amorphous samples. References 1 R. Caminiti, C. Munoz Roca, D. Beltran Porter, and A. Rossi, Z. Naturforsch. Ted A, 1988,43, 591. J. Mater. Chem., 1996, 6(lo), 1709-1716 1715 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 D Atzei, R Caminiti, C Sadun, R Bucci and A Corrias, Phosphorus, Sulfur, Silicon Relat Elem, 1993,79,13 T Egami, H J Guenterhodt and H Beck Glassy Metals I, Springer Verlag, Berlin, 1981, p 25 G Fritsch and D A Keimel, J Muter Sci Eng A, 1991,134,888 K Klier, Adv Catal, 1982,31,243 K Klier, Appl Surf Sci ,1984,19,267 K Klier, Inorg Chem ,1989,28,3868 A Suger and E Freund, US Put, 4 122 110,1978 A Sugier and E Freund, US Pat, 4 291 126,1981 J E Baker, R Burch and S E Golunski, Appl Cat, 1984,53,279 P A Wright, S Natarajan, J M Thomas and P L Gai-Boyes, Chem Muter, 1992,41053 P Porta, S De Rossi, G Ferrans, M Lo Jacono, G Minelli and G Moretti, J Catal, 1988,109,367 P Porta, R Dragone, G Fierro, M Inversi, M Lo Jacono and G Moretti, J Muter Chem, 1991, 1, 531 P Porta, G Moretti, M Lo Jacono, M Musicanti and A Nardella, J Muter Chem, 1991,1, 129 S Morpurgo, M Lo Jacono and P Porta, J Muter Chem, 1994, 4,197 A Coteron and A N Hayhurst, Appl Catal A, 1993,101,151 P Courty and C Marcilly, Preparation of Catalysts I, Elsevier, Amsterdam, 1976,p 119 18 P Courty and C Marally, Preparation of Catalysts IZI, Elsevier, Amsterdam, 1983, p 485 19 J S Campbell, Ind Eng Chem Process Des Dev, 1970,9,588 20 A J Marchi, A G Sedran and C R Apesteguia, Proc IVth Int Symp on The Scienttjic Bases for the Preparation of Heterogeneous Catalysts Louvazne La Neuve 1986, ed B Delmon, Elsevier, Amsterdam, 1987, p 529 21 R Caminiti, C Sadun, V Rossi, F Cilloco and R Fehci, XXVth Italian Congress of Physical Chemistry, Cagliari, 1991, p 4, 138, It Pat, RM/93 A000410,1993 22 K Nishikawa and T Iijima, Bull Chew SOC Jpn ,1984,57,1750 23 M Carbone, Graduational dissertation, Univ ‘La Sapienza’, Rome, 1992 24 J N Van Niekerk and F R L Schoening, Acta Crystallogr, 1953, 6,609 25 H Von Riffel, F Zettler, H Bokern and H Hess, Z Anorg Allg Chem, 1979,454,175 26 G B Johansson and 0 Lundquist, Acta Crystallogr Sect B, 1976,32,407 27 D T M Cromer, J Kay and A C Larson, Acta Crystallogr, 1966, 21,383 28 R F Zarhrobsky and W H Baur, Acta Crystallogr Sect B, 1968, 24,508 Paper 6/02815B, Received 23rd April, 1996 1716 J Muter Chem, 1996,6(10), 1709-1716