Introduction

Studies have been reported recently of the adsorption of sulfur dioxide on transition metal oxides supported on silica gel (1), on silica gels (2), and on y-aluminas (3). Calorimetric heats of adsorption and adsorption isotherms were determined at 423 ^оK. The adsorbents were characterized by measurement of surface area, pore size distributions, and hydroxyl group concentrations. These data were used to interpret the nature of the adsorbate-adsorbent interactions (1-3). The results obtained for the sorption of sulfur dioxide on silica gels (2), largely covalent solids, and y-aluminas (3), covalent solids with some ionic sites, indicated that a study of the sorption of sulfur dioxide on an oxide adsorbent in which ionic bonds predominate would help to elucidate any general features regarding the r81e of the solids in the surface reactions. Magnesia was chosen for this purpose with the added incentive that little or no work has been reported on the sorption of sulfur dioxide on this material. Useful previous studies of the sorption of gases on magnesia of some relevance to this work include those on water vapor (4) and carbon dioxide (5, 6). Results from the latter have been used to estimate the number (5) and type (6) of basic sites present on the surface of the oxide.

Experimental Materials

Magnesia was prepared by decomposing "Baker Analyzed" basic magnesium carbonate in air at 550 ^оC for 1 h. Samples of this material were then heat-treated in air at 600, 700, and 800 ºC for 24 h. X-Ray powder diffraction photographs showed the characteristic lines of magnesium oxide (7) for all the samples. However, the lines for the 600 ºC sample were more diffuse than for the others. Immediately after heat-treatment, samples were cooled over phosphoric anhydride, transferred to the appropriate apparatus and then outgassed at 150 ºC overnight at <10^-4 mm. The quality of the nitrogen and sulfur dioxide has been described (1).

Apparatus

Heats of adsorption were determined calorimetrically in a similar apparatus to that described earlier (I). Nitrogen adsorption-desorption isotherms (77 ºK) were measured in a standard apparatus (8) and pore size distributions (9) calculated with the aid of an IBM 360/50 computer. Mercury "lump" densities were measured as described previously (10). The "water content" of the magnesias was determined gravimetrically (8) to constant weight at 1200 ^оC in a static air atmosphere using a recording thermobalance (Stanton-Redcroft Model HTSA). This method gives the total amount of water released by the sample at 1200 ^оC and was not used to distinguish among the various states of water or its dissociated forms that may be present in the bulk or surface of the magnesias. Electron micrographs were taken with a Philips EM 300G instrument and crystallite structures examined by the direct transmission technique described earlier (3).

Results

Heats of Adsorption and Adsorption Isotherms for Sulfur Dioxide

Heats were determined at 423 ^оK for surface coverages from 0.2 to 7.7 mmol/m² for the 600 ^оC sample, and between 0.2 and 4.9 µmol/m2 for the 700 and 800 ^оC samples. High heats of adsorption were observed on all the adsorbents even at coverages approaching a monolayer (7.0 mmol/m² (l)), Fig. 1.

The heat curves are of similar general shape although the curve for the 600 ^оC sample exhibits a sharper heat drop than the others at a coverage of about 2.0 µmol/m2. In addition, although the heats evolved could be measured up to coverages of 11.4 µmol/m² for all the samples, the volume of gas adsorbed was too small (< 0.1 cc) to be measured accurately. A minimum estimate of these heats could be made by taking the volume of gas adsorbed as 0.1 cc for heats at coverages greater than a monolayer. On this basis these heats are of the order of 45 ± 10 kcal/mol. Figure 2 shows the adsorption isotherms for sulfur dioxide at 423 ºK. With the 600 ^оC sample a definite change in the shape of the isotherm occurs at a coverage of approximately 2.0 mmol/m². Adsorption equilibrium times for all the adsorbents were of the order of 15 to 45 min for coverages up to 2.0 to 5.0 mmol/m² respectively, and up to 24 h for coverages greater than 7.0 mmol/m².

Discussion

The heat values for all the samples are similar in magnitude at all points on the curves, Fig. 1. Heats greater than 35.0 kcal/mol were exhibited over the entire range of surface coverage studied and values of this order of magnitude clearly indicate that chemical adsorption and/or reaction of the sulfur dioxide has occurred (12). The principal sulfur compound detected in the residues was magnesium sulfite which has a heat of formation (13) of -27.7 kcal/mol at 423 ºK according to the reaction Thus, in the absence of any signficant quantities of other residues, substantial surface phenomena other than purely chemical combination have taken place. Studies of the effect of small water vapor pressures (< 5mm) on the sintering of MgO (4) suggest that the presence of H₂O provides a mode of surface transport for the O^2-- ion and possibly, also, for electrostatic reasons, induces cation migration. The postulated mechanism involves dissociative chemisorption of H₂O on the Mg²⁺ and O^2-- ions, and desorption of water from adjacent hydroxyl groups. Since water vapor pressures of this order of magnitude are present on calcining in air at temperatures up to 800 ºC, similar active centers are probably generated on the surfaces of the heat-treated magnesias used in the present study. In addition such active anion sites may be created purely by the calcination process which involves a dehydration mechanism among adjacent hydroxyl groups. Figure 1 shows that the fairly steady high heats for the 800 ºC sample do not occur at higher coverages than those obtained with the 600 and 700 ºC samples. This may be explained by the greater degree of agglomeration of the 800 ºC material with the formation of larger crystallites which act to reduce the number of active sites and lattice defects (14). The heat values of around 75 and 80 kcal/mol for the 800 ºC sample at coverages up to 2 mmol/m² indicate the possibility of the sulfur dioxide reacting directly with the active anion centers. Similar observations can be made for the 600 and 700 ºC samples. Thus sulfite formation probably occurs at these centers with a consequent readjustment of the position of lattice ions in the magnesia matrix.

Above a surface coverage of 4.0 mmol/m² the levelling off of the heat values, particularly noticeable with the 600 ºC sample, is paralleled by a large increase in the amount of sulfur dioxide adsorbed, Fig. 2. The major weight loss observed on heating the residues above 350 ºC, the decomposition temperature of magnesium sulfite (IS), together with the observation that the amounts of gas lost are approximately equivalent to one monolayer coverage of sulfur dioxide, strongly suggest that predominantly magnesium sulfite is formed in this region of the heat curves. However, the possibility of some reaction of sulfur dioxide with the magnesium sulfite has to be considered. Thus the difference between the theoretical heat of formation of magnesium sulfite and the observed heat, around 40 kcal/mol, could be explained if some sulfur dioxide was preferentially adsorbed on the magnesium sulfite prior to the apparent completion of the monolayer:

The rapid fall in the heats for all samples between 2.0 and 4.0 mmol/m² can be explained by a gradual increase in the contribution of reaction 2 to the overall process. Above monolayer coverage, the estimated heats of 45 ± 10 kcal/mol suggest that reaction 2 may predominate. These suggestions are supported by the presence of small amounts of sulfate ions detected on all the residues. Although disproportionation of magnesium sulfite is thermodynamically possible at 423 ºK (15):

only about 5% of the change would occur (15). It would not account for the very slow, but significant, adsorption of sulfur dioxide on the samples. Further, no magnesium sulfide could be detected in the residues. It is expected that some interaction of the sulfur dioxide with hydroxyl groups on the magnesia surfaces occurs but it is difficult to ascertain the influence of such interactions on the measured heats since the numbers of surface hydroxyl groups are unknown. However, infrared evidence (5) suggests that these numbers are relatively small at the temperatures of heat treatment used, and thus the "water content" must largely consist of bulk water/hydroxyl groups. The nitrogen adsorption-desorption isotherms are similar to those observed earlier (16), although no hysteresis loops were formed with the 700 and 800 ºC samples. However, this is not unexpected (17) since the present 700 and 800 ºC samples are essentially nonporous, having only a few pores in the region of 20 to 25 A diameter. The difference between these observations and those made earlier (16) is explained simply by differences in the methods of preparation. At 600 ºC, significant sintering has not occurred which agrees with earlier postulates that sintering is not expected below 640 ºC (14). At 700 and 800 ºC, particles have clearly agglomerated with cubic crystallites formed in the 800 ºC sample. This sintering has clearly occurred within rather than between the aggregates of magnesia crystals in accord with earlier observations (4). It is particularly significant that the electron micrographs of the 700 and 800 ºC residues, the edges of the magnesia crystallites have been attacked by the sulfur dioxide as indicated by the irregularity and smoothness of the edges. The presence of the small dark areas in the large, sintered crystallites may suggest that small crystallites of some other material, probably magnesium salts, may be present. When these heat curves of sulfur dioxide on magnesias at 423 ºK are compared with those determined earlier on heat-treated silica gels (2) and y-aluminas (3), a general trend is apparent. The heats of adsorption at a given coverage increase with an increase in the degree of ionicity of the solid, thus: magnesia > y-alumina > silica. A feature which is clearly related to the known ability of these materials to form compounds with sulfur dioxide.

1. R. W. GLASS and R. A. Ross. Can. J. Chem. 49,2832 (1971).

2. R. W. GLASS and A. Ross. Can. J. Chem. 50, 1241 (1972).

3. R. W. GLASS and R. A. ROSS. Can. J. Chem. In press.

4. P. J. ANDERSO and P. L. MORGAN. Trans. Faraday SOC. 60, 930 (1964).

5. ST. MALINOWSKSI., SZCZEPANSKA and, J. SLOCZYNSKI. J. Catal. 7, 67 (1967).

6. 0. V. KRYLOVZ, . A. MARKOVAI. , I. TRETIAKOV and E. A. FOKINA. Kinetika i Kataliz. 6, 128 (1965).

7. X-Ray powder data file. Inorganic volume PDIS 5iRB. 4-0829. A.S.T.M. Philadelphia, Pa. 1966.

8. L. G. JOYNER. Scientific and industrial glass blowing and laboratory techniques. Edited by W. E. Barr and V. J. Anhorn. Instruments Publishing Company. Pittsburg. 1949.

9. A. H. TAYLOR and R. A. Ross. J. Chem. Eng. Data, 14 (3). 347 (1969).

10. W. B. JEPSON. J. Sci. Inst. 36, 319 (1959).

11. F. FEIGL. Spot tests in inorganic analysis. Elsevier, New York. 1958.

12. S. J. GREGG. Surface chemistry of solids. Chapman and Hall, London. 1961.

13. Handbook of chemistry and physics. Chemical Rubber Publishing Co., Cleveland. 1970.

14. R. M. DELL and S. W. WELLER. Trans. Faraday. Soc. 55, 2203 (1959).

15. P. S. LOWELLK, . SCHWIZGEBETL., B. PARSONS and K. J. SLADEK. Ind. Eng. Chem. Proc. Res. Develop. 10, 384 (1971).

16. W. F. N. M. DE VLEESSCHAUWEP Rhysical and chemical aspects of adsorption and catalysis. Edited by B. G. Linsen. Academic Press, London. 1970.

17. D. H. EVERETT. The solid-gas interface. Vol. 2. Edited by E. A. Flood. Marcel Dekker, Inc., New York. 1967.