J . Chem. SOC., Faraday Trans. 1, 1984, 80, 237-248 X-Ray Photoelectron Spectroscopic Study of Montmorillonite Containing Exchangeable Divalent Cations BY HARUHIKO SEYAMA* AND MITSUYUKI SOMA National Institute for Environmental Studies, Yatabe, Tsukuba, Ibaraki 305, Japan Received 23rd June, 1983 Montmorillonite containing divalent Mg, Ca, Sr, Ba or Cd as an exchangeable cation has been studied by X-ray photoelectron spectroscopy and X-ray-induced Auger electron spectroscopy. The relative atomic abundances of the exchangeable cations in the montmorillonite samples, obtained by X.P.S. measurements, are consistent with the cation-exchange capacity of the montmorillonite. A comparison of the photoelectron binding energies and Auger electron kinetic energies of the exchangeable cations with those of corresponding halides and oxides reveals that they fall among those of corresponding cations in typical ionic compounds such as fluorides and chlorides.On the other hand, the bonding state of the non-exchangeable Mg ion in montmorillonite lattice, as indicated by the Mg 1s and KL,, 3L2, Auger electron energies, is similar to that of magnesium oxide and distinguishable from that of the exchangeable Mg ion. Clay minerals are important constituents of soil and sediment and greatly influence the geochemical behaviour of various cations by their cation-exchange capabilities. Many investigations on the cation-exchange reactions of clay minerals have been undertaken and many data, e.g. concerning cation-exchange capacities, have been accumulated.l* However, there has been a lack of systematic knowledge concerning the bonding states of exchanged cations in clay minerals simply because experimental methods applicable to probing the bonding states of various elements in solids such as minerals have not been available.X-ray photoelectron spectroscopy (X.P.S.), which has advanced rapidly in recent years, is a suitable method for this purpose. It detects all elements except H and He and determines the abundance of elements from the intensity of core-electron emission. Detailed spectral features such as the chemical shift in the electron binding energy, satellite structure etc. give information about the bonding states of elements. Adams et al. attempted the quantitative analyses of clay minerals and other aluminosilicate minerals by X .P . S . ~ ~ ~ The valence state of lead adsorbed on mont- morillonite was studied by Counts et al. based on the binding energies of Pb 4f electrons5 The spectral changes of Fe 2p3/2 and 0 Is photoelectrons of nontronite and biotite with oxidation and reduction of these clay minerals were investigated by Stucki et al.' More recently, Koppelman et al. systematically investigated the adsorption of transition-metal ions and their complex ions on chlorite, illite and kaolinite.'-ll The origins of the negative charge on the particles of these three clay minerals are considered to be due to a small amount of isomorphic substitution, to lattice imperfection and to broken bonds at the edges of the particles. Therefore the cation-exchange capacities (c.e.c.) of them are low (1040 mequiv.per 100 g for chlorite and illite and 3-15 mequiv. per 100 g for kaolinite).2 In this paper we report an X.P.S. study of montmorillonite, a typical smectite clay 237238 X.P.S. STUDY OF MONTMORILLONITE mineral containing a series of divalent exchangeable cations (Mg, Ca, Sr, Ba and Cd). In contrast to the above-mentioned chlorite, illite and kaolinite, sufficient isomorphic substitution occurs in smectite clay minerals. In order to compensate the negative charge originating from isomorphic substitution, exchangeable cations are held between the aluminosilicate layers, resulting in high c.e.c. values (80-1 50 mequiv. per 100 g).2 The choice of exchangeable cations was based not only on systematic considerations (the alkaline-earth elements) but also on geochemical, agricultural and environmental importance.The emphasis is on demonstrating the capability of X.p.s. for the non-destructive quantitative analysis of clay minerals and for characterizing the bonding states of cations by core-electron binding energies and the kinetic energies of X-ray-induced Auger electrons. A preliminary account of Mg-montmorillonite has been reported previously.12 EXPERIMENTAL MATERIALS The montmorillonite used in this study was Kunipia F, a processed natural bentonite mined in the Tsukinuno mine, Yamagata, Japan and obtained from Kunimine Industries. The nominal composition of Kunipia F in wt% was as follows: SiO,, 57.96; A1203, 21.87; Fe,O,, 1.92; MgO, 3.44; CaO, 0.54; Na,O, 2.98; K,O, 0.14.It was transformed to Na-montmorillonite containing only Na as an exchangeable cation in accordance with the method used by Posner and Quirk.I3 The original powdered clay was washed with 1 mol dm-, NaCl aqueous solution and separated by decantation, this treatment being repeated six times. Then the montmorillonite was suspended in 1 mol dmP3 NaCl aqueous solution at pH 3 adjusted with hydrochloric acid for ca. 1 h, followed by separation from the solution by centrifuging. After this procedure was repeated four times, the montmorillonite was resuspended in 1 mol dm-, NaCl aqueous solution at pH 3, stirred for cu. 36 h in order to replace completely the exchangeable cations in the montmorillonite with Na cation, separated again by centrifuging and washed twice with distilled water.It was then dialysed against deionized water for ca. 2 days and freeze-dried. M-montmorillonite containing an exchangeable divalent cation of M (M = Mg, Ca, Sr, Ba or Cd) was prepared by cation exchange of Na-montmorillonite. Na-montmorillonite was suspended in 7 x lo-, mol dmP3 M(NO,), aqueous solution for > 24 h. The total amount of M2+ in the solution was about four times as large as the c.e.c. of montmorillonite. The initial and final concentrations of M2+ in solution were determined by EDTA chelatometry. The c.e.c. of montmorillonite for each M cation was calculated from the concentration change of M cation in the solution. After the cation-exchange reaction, M-montmorillonite was separated from the solution by centrifuging, washed by resuspending in distilled water, separated again by centrifuging and dried in a vacuum desiccator.Metal salts used for X.P.S. measurements were commercial materials of guaranteed purity and used without further purification. X. P.S. MEASUREMENTS Photoelectron and Auger electron spectra were recorded on a Vacuum Generators ESCALAB 5 instrument with magnesium and aluminium X-ray sources. Typical measuring conditions were: X-ray power, 13 kV x 10 mA; electron pass energy, 50 eV; width of the entrance slit of the analyser, 4 mm. The instrumental resolution determined by the pass energy and the slit width was 1.25 eV.14 In order to improve the signal to noise ratio, the signals for weak intensity lines were accumulated on a Nicolet 1070 signal averager. All samples except metal oxides were ground to powders with an agate mortar and fixed onto stainless-steel sample holders of 10 mm diameter using double-sided sticky tape.Powder samples of metal oxides were directly deposited onto sample holders from acetone suspension. The sample of CdO was heated (600 OC, 1 day) in the analyser chamber of the instrument for the purpose of dehydration and decarbonation of Cd(OH), and CdCO, contaminants prior to making X.P.S. measurements according to the procedure previously described by Hammond et aI.l5 The samples of alkaline-earth oxides wereH. SEYAMA AND M. SOMA 239 exposed to 5 keV argon-ion bombardment at 10 pA for several tens of minutes in the analyser chamber of the instrument in order to remove hydroxide and carbonate contaminants on the sample surface.The photoelectron binding energies and the Auger electron kinetic energies of the alkaline-earth metal ions decreased and increased, respectively, as a result of the bombard- ment, which was continued until the metal-ion spectra became stationary. Before bombardment, the 0 1s spectrum of the alkaline-earth oxides consisted of only one line derived from the oxygen atom(s) of the surface contaminants. After bombardment, the same spectrum was broad or accompanied by a shoulder or other peak on the high-binding-energy side due to contributions from oxygen atoms in the residual hydroxide and/or carbonate ion contaminants, perhaps in the bulk. Therefore, the spectra of alkaline-earth oxide samples treated in this way were still not free from impurities.Photoelectron binding energies and Auger electron kinetic energies were determined relative to the Au 4f,,, binding energy (83.8 eV) of a gold film vacuum evaporated onto the sample as a primary standard. For the samples of montmorillonite, the Si 2s binding energy (1 53.4 eV) determined by this method was used as an internal standard for calibration. All the photoelectron binding energies and Auger electron kinetic energies, except for the Ca 2p,/, binding energy, were determined by A1 Ka excitation. The Ca 2p3/, binding energy was determined by Mg Ka excitation to avoid the overlapping of Ca 2p and Mg Auger KL,L,,, lines that occurs when using A1 Ka. RESULTS AND DISCUSSION ATOMIC COMPOSITION The c.e.c. values of montmorillonite for Mg, Ca, Sr, Ba and Cd cations are given in table 1.They were all identical (99 mequiv. per 100 g) within experimental error, and hence it was confirmed that there was no difference in the exchangeable-cation content of montmorillonite for the various ions. The wide-scan X-ray photoelectron spectrum of Ba-montmorillonite excited by A1 Ka radiation is shown in fig. 1. Photoelectron and Auger electron peaks due to the constituent elements of the montmorillonite lattice (Si, Al, Mg and 0) and to carbon present as a surface contaminant were found in the wide-scan X-ray photoelectron spectra of all the montmorillonite samples. In addition, characteristic peaks due to the exchangeable cation, such as the Ba 4d, 3d and M4,5N4,5N4,5 Auger lines in fig. 1, were observed for each montmorillonite sample.Table 1. Cation-exchange capacities (c.e.c.) of montmorillonite for divalent cations c.e.c. cation /mequiv. per 100 g Mg 96 Ca 101 Sr 100 Ba 102 Cd 96 mean 99 The relative atomic abundances of Al, Si, Mg and of the exchangeable cation in each montmorillonite sample were calculated from area intensities of the photoelectron and Auger electron spectra on the basis of the relative sensitivity of each line. The relative sensitivities used in the calculation were experimentally determined from the240 X.P.S. STUDY OF MONTMORILLONITE 12s 1 I 500 1000 binding energy/eV Fig. 1. Wide-scan X-ray photoelectron spectrum of Ba-montmorillonite excited by A1 Ka radiation. relative area intensities of the photoelectron and Auger electron spectra of compounds of known chemical composition, i.e.zeolites, halloysite, fluorides, chlorides and sulphates. The atomic abundances in the montmorillonite samples relative to A1 (atomic ratios) determined in this way are given in table 2. The calculated average atomic abundances of Si and Mg in the aluminosilicate layers of montmorillonite relative to A1 were 2.17 and 0.22, respectively. The bulk atomic abundances of Si and Mg relative to A1 calculated from the nominal composition of Kunipia F were 2.25 and 0.20, respectively, which are within 10% of the average atomic abundances obtained by X.P.S. This result suggests that the bulk chemical composition is maintained at the surface of the particle. Consequently, X.P.S. is a useful bulk quantitative analytical technique for aluminosilicate minerals, as pointed out by Adams et aL3 Since Mg-montmorillonite contains exchangeable Mg ions between the aluminosilicate layers, in addition to the non-exchangeable Mg ions as constituents of the aluminosilicate layers, the atomic abundance (0.34) of Mg relative to A1 in Mg-montmorillonite is larger than that (0.22) in other montmorillonites containing only non-exchangeable Mg ions.Thus the difference (0.12) between the relative atomic abundances of Mg in Mg-montmorillonite and in other montmorillonites is assigned to the relative atomic abundance of the exchangeable Mg ion in Mg-montmorillonite. The atomic abundances thus determined of exchangeable divalent cations relative to A1 in the montmorillonite samples range from 0.1 1 to 0.15 (mean = 0.12) and are equal to approximately half the amount of Na ion in Na-montmorillonite.The same quantity calculated from the c.e.c. (99 mequiv. per 100 g) and the nominal composition of Kunipia F is 0.12, which is in good agreement with that determined by X.P.S. The relative atomic abundance (0.15) found for Ba ions in Ba-montmorillonite isH. SEYAMA AND M. SOMA 24 1 Table 2. Atomic abundances in M-montmorillonite samples relative to A1 (atomic ratios) determined by X.P.S. M X-ray element line source Na Mg Ca Sr Ba Cd mean A1 Si Mg Na Ca Sr Ba Cd 2P 2s Auger Auger 2P 3d (KL'2,3L2,3) ( K L 2 , 3L2, 3) 3d 3d A1 Ka Mg Ka A1 Ka Mg Ka A1 Ka A1 Ka Mg Ka Mg Ka A1 Ka Mg Ka A1 Ka Mg Ka A1 Ka Mg Ka 1 1 2.20 2.13 0.2 1 0.25 0.27 1 1 2.26 1.99 0.34 (0.12)a - 1 1 1 1 1 1 2.15 2.22 2.00 2.06 2.33 2.28 0.22 0.21 0.23 0.23 0.22b a Relative atomic abundance of exchangeable Mg [0.34 (total Mg) - 0.22 (non-exchangeable Mg) = 0.121; Mean value for montmorillonite samples except for Mg-montmorillonite.25% larger than the average value for the exchangeable divalent cations in the montmorillonite, and requires comment. Adams and Evans4 used X.P.S. to estimate the atomic abundances of exchangeable cations relative to Si in beidellite and reported that the relative atomic abundance of Ba ions in Ba-beidellite was ca. 50% greater than that of Ca ions in Ca-beidellite. Preferential external surface adsorption of Ba ions was suggested as an explanation for the excess Ba detected by X.P.S. The same phenomena were observed for K- and Pb-beidellites.Although both beidellite and montmorillonite are smectite clay minerals, the deviation of the amount of Ba found by X.P.S. from that calculated from the c.e.c. in our case is not so pronounced as in their result. We should also note the shallow escape depths of Ba 3d photoelectrons, which have relatively low kinetic energies (ca. 700 and 470 eV for A1 Ka and Mg Ka excitation, respectively), as compared with the other cations. Thus the intensity of the Ba 3d line would be rather sensitive to the amount of surface contaminant, which makes the accurate determination of Ba concentration based on that line difficult. It may therefore be presumed that in the present case the preferential external surface adsorption of exchangeable divalent cations (Mg, Ca, Sr, Ba and Cd ions) does not occur and that they are held exclusively between the aluminosilicate layers of mon tmorilloni te.Mg IS AND KL,,,L,,, AUGER SPECTRA OF Mg-MONTMORILLONITE Fig. 2 shows the Mg Is and KL,,,L,,, Auger spectra of Mg-montmorillonite excited by A1 Ka X-rays. They are broad in comparison with those of montmorillonite containing only non-exchangeable Mg ions. In Mg-montmorillonite the exchangeable and non-exchangeable Mg ions have different locations, i.e. the former exist between the aluminosilicate layers while the latter form one of the constituents of the aluminosilicate layers. Accordingly, the broad character of the spectra of Mg-242 X.P.S. STUDY OF MONTMORILLONITE .-... . . ..-. .. .. . . (a ) non-exchangeable Mg . . .. . 1300 1305 1310 binding energy/eV .......I. -, .. _. ( b 1 exchangeable Mg I I I 1185 1180 1175 kinetic energy/eV Fig. 2. Resolution of (a) Mg 1s and (b) Mg KL,,,L,,, Auger spectra of Mg- montmorillonite. montmorillonite can be ascribed to the existence of two kinds of Mg ions. The observed Mg 1s and KL2,&2,3 Auger spectra were each resolved into two components as follows. The Mg KL,,,L,,, Auger kinetic energy for non-exchangeable Mg ion and its intensity relative to the A1 2p line due to skeletal A1 were determined for the montmorillonite samples which did not contain Mg as an exchangeable cation. This component was then subtracted from the Mg KL,,,L,,, Auger spectrum of Mg- montmorillonite and the residual spectrum was assigned to the exchangeable Mg ion. The result of the resolution is shown in fig.2. The relative intensities of the two components were consistent with the relative abundances of the non-exchangeable skeletal Mg ion and the exchangeable cation as determined with other montmorillonite samples. The Mg 1s line was also resolved similarly but with readjustment of the intensity factor in order to make the relative intensity of the resolved spectra consistent with the Mg KL2,,L2,, Auger result, as the Mg Is intensity relative to the A1 2p line varied from sample to sample. This variation is attributed to the shallow escape depth of Mg 1s photoelectrons, which have fairly small kinetic energies (ca. 180 eV) so thatH. SEYAMA AND M. SOMA 243 their intensity is very sensitive to the amount of surface contaminant. The resolved Mg 1s spectra are also included in fig.2. The Mg 1s binding energy and KL2,3L2,3 Auger kinetic energy resolved for the exchangeable Mg ion are 1305.1 and 1179.2 eV, respectively, while those for the non-exchangeable Mg ion are 1303.4 and 1181.2 eV, respectively. Thus it is apparent that there are differences of > 1 eV in both electron energies between the exchangeable and non-exchangeable Mg ion in montmorillonite. PHOTOELECTRON BINDING ENERGIES AND AUGER ELECTRON KINETIC ENERGIES The measured photoelectron binding energies and Auger electron kinetic energies of the Mg, Ca, Sr, Ba and Cd compounds are shown in tables 3-4, where also the same energies of exchangeable cations in Mg-, Ba- and Cd-montmorillonites are compared with those of the corresponding halides and oxides.A comparison of Mg Is binding energies and KL,,&2,3 Auger kinetic energies of the exchangeable and non-exchangeable Mg ions in montmorillonite with those in the magnesium halides and magnesium oxide is shown in fig. 3. Such a two-dimensional plot of photoelectron binding energy and Auger electron kinetic energy was proposed by Wagner et al. and called a chemical-state plot .16 The measured starting materials for magnesium chloride and bromide were the hexahydrates. However, the observed 0 Is peak heights for the compounds were lower than those expected for the hexahydrates, so it was considered that the water of crystallisation of these samples was partly lost in the high vacuum of the instrument. The same phenomena were also observed for X.P.S.measurements of other halides containing water of crystallisation. - non-exchangeable Mg in montmorillonite Ms 0 Mg Br2-6H20 0 i( 0 exchangeable Mg in montmorillonite I I 1307 1305 1303 Mg 1s binding energy/eV Fig. 3. Chemical-state plot for Mg. As shown in fig. 3, the position in the chemical-state plot of exchangeable Mg in Mg-montmorillonite falls between those of magnesium chloride and fluoride. There are only small differences (ca. 1 eV or less) in the Mg Is binding energy between the 9 FAR 1244 X.P.S. STUDY OF MONTMORILLONITE Table 3. Photoelectron binding energies of Ca and Sr compounds fluoride (CaF,, SrF,) 349.0 270.2 exchangeable cation 348.6 270.3 in montmorillonite a chloride (CaC1,) 348.5 - oxide (CaO, SrO) 345.9 268.8 a This value was not able to be determined owing to over- lapping of Sr 3p3/2 and Cl 2s lines.Table 4. Mg 1s binding energies, Mg KL,,,L,,, Auger kinetic energies and AEr values of Mg compounds Mg KL293L2,3 compound Mg ls/eV Auger /eV AEr/eVa MgFz 1306.3 exchangeable Mg 1305.1 non-exchangeable Mg 1303.6 MgCl, * 6H,O 1304.6 MgO 1303.7 MgBr, 6H,O 1305.1 in montmorillonite in montmorillonite 1177.0 1 179.2 1181.2 1 180.4 1181.5 1180.9 0.0 1 .o 1.5 1.7 1.9 2.7 Table 5. Cd 3d5/, binding energies, Cd M4N4,5N4,5 Auger kinetic energies and AEr values of Cd compounds Cd M4N4,5N4,5 compound Cd 3d5,,/eV Auger /eV AEr/eVa CdF, 405.8 exchangeable Cd 406.2 CdC1, * 2iH20 405.5 CdBr, * 4H,O 405.5 CdI, 405.6 CdO 404.0 in montmorillonite 378.4 378.5 380.0 380.2 380.7 382.7 0.0 0.5 1.3 1.5 2.1 2.5 exchangeable Mg ion in Mg-montmorillonite and each magnesium halide.The difference in the Mg KL,,3L2,3 Auger kinetic energy between the exchangeable Mg ion in Mg-montmorillonite and each magnesium halide is larger than the corresponding difference in the Mg 1s binding energy. Such larger differences in the Mg KL2,,L2,3 Auger kinetic energy may be attributed to the effect of extra-atomic relaxation,H. SEYAMA AND M. SOMA 245 Table 6. Ba 3d5/, binding energies, Ba M4N4,5N4,5 Auger kinetic energies and AEr values of Ba compounds Ba M4N4,5N4,5 compound Ba 3d5,2/eV Auger/eV AEJeV" BaF, 781.5 exchangeable Ba 781.0 BaC1, 2H20 781.4 BaO 779.5 in montmorillonite 595.1 0.0 595.4 - 0.2 595.1 -0.1 598.0 0.9 i.e. a screening of the final-state ion in the Auger transition by electrons from neigh- bouring atoms (see below).In contrast to the position of exchangeable Mg, the position in the chemical-state plot of non-exchangeable Mg, which is surrounded by four oxygen ions and two hydroxide ions (in octahedral coordination) in the aluminosilicate layers of mont- morillonite, is located close to that of magnesium oxide. The Mg 1s binding and KL,,,L,,, Auger kinetic energy of the non-exchangeable Mg ions are 1.5 smaller and 2.0 eV larger, respectively, than those of the exchangeable ions. These large differences may be attributed to the effect of neighbouring atoms, i.e. the flow of electronic charge to the Mg ion and the extra-atomic relaxation from surrounding oxygen atoms are larger for the non-exchangeable Mg ion, indicating its stronger interaction with oxygen.The success in differentiating between the exchangeable and non-exchangeable Mg ions by locating them among the reference compounds in the chemical-state plot demonstrates the usefulness of the chemical-state plot in characterizing the bonding states of cations in minerals. Chemical-state plots for Ba and Cd (3d5,, binding energy and M4N4,5N4,5 Auger kinetic energy) are shown in fig. 4 and 5. The M4N4,5N4,5 Auger line is weaker than the M5N4,,N4,, Auger line but is chosen for both chemical-state plots because the M5N4,5N4,5 Auger line is not sharp and therefore less suitable for a chemical-state plot, as pointed out previ0us1y.l~ Shifts in the Auger lines among barium compounds are smaller than those found among magnesium compounds, and the proximity of exchangable Ba ion in Ba-montmorillonite to barium halides is more pronounced.The differences in both Ba 3d5/, binding energy and M4N4,5N4,5 Auger kinetic energy between the exchangeable Ba ion and each barium halide are 0.5 eV or less, while there are differences of > 1 eV in both electron energies between the exchangeable Ba ion and barium oxide, as in the case of the exchangeable Mg ion. The chemical-state plot for Cd (fig. 5) shows a tendency similar to that for Mg (fig. 3). Thus the differences in Cd 3d5/2 binding energy between several cadmium halides and the exchangeable Cd ion in Cd-montmorillonite are small, whereas the differences in Cd M4N4,5N4,5 Auger kinetic energy between them are larger. The differences in both Cd 3d5,, binding energy and M4N4,5N4,5 Auger kinetic energy between the exchangeable Cd ion and cadmium fluoride are -= 0.5 eV, while the Cd 3d5,, binding energy of exchangeable Cd ion is significantly higher (2.2 eV) than that of cadmium oxide and the Cd M4N4,5N4,5 Auger kinetic energy of exchangeable Cd ion is significantly lower (4.2 eV).In the X.P.S. measurements of calcium compounds, the Ca L2,3M2,3M,,3 Auger line is observed but is not very intense. Further, it is difficult to determine its precise peak 9-2246 X.P.S. STUDY OF MONTMORILLONITE exchangeable Ba in montmorillonite 1 I I I 783 781 779 Ba 3d512 binding energy/eV Fig. 4. Chemical-state plot for Ba. exchangeable Cd in montmorillonite 407 405 403 Cd 3dS12 binding energy/eV Fig. 5. Chemical-state plot for Cd.H.SEYAMA AND M. SOMA 247 position (kinetic energy) because the Ca L2,3kf&,,, Auger spectrum is complex and its shape varies from compound to compound. Neither is an intense Sr Auger line observed in the X.P.S. measurements of strontium compounds. Therefore only photoelectron (Ca 2p3/, and Sr 3pSl2) binding energies of exchangeable Ca and Sr ions in montmorillonite were compared with those of calcium and strontium compounds. The measured Ca 2p3/, and Sr 3p3/2 binding energies are shown in table 3. In those cases also, the differences in the photoelectron binding energies between the exchangeable cations in montmorillonite and the halides are small, whereas the differences in the photoelectron binding energies between the exchangeable cations and oxides are large.EXTRA-ATOMIC RELAXATION The sum of the photoelectron binding energy (Eb) and the Auger electron kinetic energy (Ek) of a certain atom has been defined as the modified Auger parameter (a’) by Wagner et a1.l6* l7 Kowalczyk et al. have examined the relative effect of extra-atomic relaxation on Auger electron kinetic-energy and photoelectron binding-energy shifts in metals and salts.’* Extra-atomic relaxation occurs through a flow of electronic charge toward the final-state ion in the Auger transition from neighbouring atoms, and is one of the important factors in determining the Auger electron kinetic energy. The difference in the value of a’ (Act‘) for the same element between two compounds is a good approximation to the difference in the extra-atomic relaxation energy (AE,) for the element between these compounds.’* Thus x AE,.( 2 ) By use of eqn ( 2 ) the differences in the extra-atomic relaxation energies for Mg, Ba and Cd between fluorides and other compounds were calculated. The calculated extra-atomic relaxation energy differences (AE,) are included in tables 4-6. The AE, values of exchangeable Mg and Cd ions in montmorillonite are intermediate between those of the corresponding fluorides and chlorides, as shown in tables 4 and 5. On the other hand, the AE, value of exchangeable Ba ions in Ba-montmorillonite is approximately the same as those of barium fluoride and chloride (AE, x 0), although the overall variation of AE, for Ba among barium compounds is smaller than those for Mg and Cd (see table 6).In addition, AE, values of exchangeable Mg, Cd and Ba ions are smaller (ca. 1 eV or more) than those of the corresponding oxides. Thus the effects of extra-atomic relaxation for the exchangeable Mg, Cd and Ba ions in montmorillonite are similar to those for the corresponding cations in fluorides and chlorides but not as strong as those for the same cations in oxides. This result suggests that these exchangeable cations in montmorillonite are not subject to the strong extra-atomic relaxation from the 0 ions which constitute the aluminosilicate layers of montmorillonite. On the other hand, it is considered that the non-exchangeable Mg ion in the aluminosilicate layers is more subject to the extra-atomic relaxation than the exchangeable Mg ion, as indicated by AE, of non-exchangeable Mg ion, which is 0.5 eV larger than that of the exchangeable one.The effect of extra-atomic relaxation for the non-exchangeable Mg ion, however, is not as strong as that for the Mg ion in magnesium oxide because AE, of the non-exchangeable Mg ion is smaller by 0.4 eV than that of the Mg ion in magnesium oxide.248 X.P.S. STUDY OF MONTMORILLONITE CONCLUSIONS The bonding states of exchangeable alkaline-earth and cadmium divalent cations between the aluminosilicate layers of montmorillonite are similar to those in the typically ionic chlorides and/or fluorides, as deduced from the chemical-state plots. These results indicate that the exchangeable cation is not strongly influenced by the negative charge of the 0 ion which is part of the montmorillonite lattice.The present result can be compared with X.P.S. measurements on zeolite, an aluminosilicate mineral which also holds exchangeable cations to compensate the negative charge originating from the replacement of silicon by aluminium in the lattice. It does not have a layered structure such as possessed by clay minerals but rather the three- dimensional network structure of an aluminosilicate. It has been shown by X.P.S. that the exchangeable metal cations in zeolite also have a highly ionic bonding character.l9? 2o Thus it is suggested that the exchangeable cations which compensate the negative charge originating from isomorphic substitution in aluminosilicate minerals form nearly pure ionic bonds with the minerals. X.P.S. is very informative in characterizing the bonding state of cations in these minerals, especially when combined with Auger electron spectroscopy. S. L. Swartzen-Allen and E. Matijevic, Chem. Rev., 1974, 74, 385. J. M. Adams, S. Evans, P. I. Reid, J. M. Thomas and M. J. Walters, Anal. Chem., 1977, 49, 2001. J. M. Adams and S. Evans, Clays Clay Miner., 1979, 27, 248. M. E. Counts, J. S. C. Jen and J. P. Wightman, J. Phys. Chem., 1973,77, 1924. J. W. Stucki, C. B. Roth and W. E. Baitinger, Clays Clay Miner., 1976, 24, 289. M. H. Koppelman and J. G. Dillard, ACS Symp. Ser., 1975, 18, 186. M. H. Koppelman and J. G. Dillard, Clays Clay Miner., 1977, 25, 457. M. H. Koppelman and J. G. Dillard, J. Colloid Interface Sci., 1978, 66, 345. * R. E. Grim, Clay Mineralogy (McGraw-Hill, New York, 2nd edn, 1968). lo M. H. Koppelman, A. B. Emerson and J. G. Dillard, Clays Clay Miner., 1980, 28, 119. l1 M. H. Koppelman and J. G. Dillard, Clays Clay Miner., 1980, 28, 21 1. l2 H. Seyama and M. Soma, Chem. Lett., 1981, 1009. l3 A. M. Posner and J. P. Quirk, Proc. R. SOC. London, Ser. A, 1964, 278, 35. I4 M. P. Seah, Surf. Interface Anal., 1980, 2, 222. l5 J. S. Hammond, S. W. Gaarenstroom and N. Winograd, Anal. Chem., 1975,47, 2193. l6 C. D. Wagner, L. H. Gale and R. H. Raymond, Anal. Chem., 1979,51,466. l7 C. D. Wagner, J. Electron Spectrosc. Relat. Phenom., 1977, 10, 305. S. P. Kowalczyk, L. Ley, F. R. McFeely, R. A. Pollak and D. A. Shirley, Phys. Rev. B, 1974,9, 381. Kh. M. Minachev, G. V. Antoshin, E. S. Shpiro and Yu. A. Yusifov, Proc. 6th Int. Congr. Catal. (The Chemical Society, London, 1976), p. 621. 2o H. Vinek, H. Noller, M. Ebel and K. Schwarz, J. Chem. Soc., Faraday Trans. 2, 1977, 73, 734. (PAPER 3/ 1074)