The surface hydroxylation of silica gels and powders has been studied by determining the stoichiometry of the reactions of SiMezClz, Tic&, and BCI, with the surface hydroxyl groups. The results demonstrate that the fully hydroxylated silica surfaces studied carry two distinct types of surface hydroxyl sites, which are distributed as follows: 1.4 f 0.1 single surface hydroxyls (type A sites), and about 3.2 f 0.1 interacting hydroxyl groups arranged in pairs (1.6 f 0.1 type B sites). On heating the silicas above ambient temperatures, in uucuo, the type B sites are progressively removed. By correlating this result with earlier studies, it is concluded that at temperatures of 500 50", in uucuo, virtually all the type B sites are eliminated. However, the type A site concentration remains constant at evacuation temperatures up to 600 f : 50" at around 1.4 f 0.1 A sites/100 A2.These conclusions suggest that the silica surface corresponds to an array of different crystal planes, some of which contain OH groups at relatively large interhydroxyl spacings and others containing OH groups held in such a way as to promote interhydroxyl hydrogen bonding.(1)

Methyl chlorosilane vapours react with the surface hydroxyls of Aerosil silicas at temperatures above about 250°C. Infra-red spectroscopic studies on deuterium exchanged samples show that only the external hydroxyl-groups are affected. The trimethyl-and dimethyl-substituted chlorosilmes react selectively but completely with the isolated or single surface hydroxyls, whereas the monomethyl and tetrachloro compounds also react with some of the interacting hydrogen-bonded surface groups. Analyses of the solid reaction products from the dimethyl and trimethyl chlorosilane+silica reactions suggest that these compounds react largely 1 : 1 with the single surface hydroxyls although 1 : 2 bridging reactions occur on preheating the solid to temperatures of the order of 500°C. By comparing these results with those reported earlier it is suggested that the selectivity of reactions of this type is controlled largely by the stereochemistry of the activated transition-state structure of the organic molecule.

The selective adsorption isotherms of the n-C6, Cs, C12: CI4 and CI6 fatty acids at the silica/ benzene and silica/n-hexane interface have been determined at 27°C. The adsorbents used were a fully hydroxylated silica carrying a small concentration of micropores on its surface (R.A.3) and an annealed but still fully hydroxylated silica (R.R.A.3/700). With the exception of the C 6 and C s acids the adsorption isotherms with the benzene solutions were independent of the chain length of the fatty acid. Adsorption from the n-hexane solutions showed that lengthening the hydrocarbon chain of the adsorbate decreased the surface coverage at fixed equilibrium solution concentration. The adsorption isotherms obtained from benzene solutions showed a limiting surface coverage equivalent to about 0.5 molecule of fatty acid per 100 A' of adsorbent surface for the Cs-C1 6 acids. From the n-hexane solutions the limiting adsorption ranged from about 1.85 solute molecules per 100 A' of solid surface for the c 6 acid down to 1.40 for the C1 6 species. Partly dehydroxylating the adsorbent so as to yield a silica surface carrying only single hydroxyls (R.A.3/700) resulted in this material showing virtually no selectivity for fatty acid adsorption from benzene solutions.Although the adsorption from benzene and n-hexane solutions of n-fatty acidsespecially lauric and stearic 1* 2-is currently used as a means of estimating the specific surface areas of oxide adsorbents and fillers, relatively little is known about the governing features of the adsorption process. The currently accepted view describes the adsorption of n-fatty acids acids on to silica surfaces as being selective from both benzene and n-hexane solutions, with the limiting adsorption which is observed corresponding to the close packing of dimeric fatty acid species oriented with their hydrocarbon chains lying parallel to the surface.This particular model for the surface phase is unlikely to be correct since it requires the hydrocarboii chain to displace the benzene from the surface. Adsorption studies at the gas/solid interface of the relative affinities of silicas for saturated and aromatic hydrocarbons show that this is improbable. The object of the present work was to determine those factors which exert a controlling influence on selective adsorption in such systems. E X P E R I M E N T A L MATERIALSThe n-fatty acids were obtained from Fluka A.G. and stated to be at least 99 % pure. G.L.C. analysis confirmed this, the only impurities being trace quantities (<1 %) of the two nearest even-numbered homologous acids. The acids were stored in dark bottles under dry nitrogen and opened only for the minimum time required for the necessary transfer present address : Procter and Gamble Ltd., Longbenton, Newcastle-on-Tyne.

The heats of (selective) adsorption of the n-C6, C,, CIz, CI4 and c 1 6 fatty acids at tlic silica/ benzene interface have been determined from measurements of the heats of immersion of " fully " hydroxylated silicas in benzene + n-fatty acid solutions. Within experimental error, the molar heat of adsorption is independent both of the surface coverage of the fatty acids and of their hydrocarbon chain length. The heats of solution of the fatty acids in benzene at 27°C have also been measured and indicate that the acids exist as monomeric species only at low solution concentrations. At concentrations above about m the species tend to associate due to hydrogen bonding between thecarboxylic head groups as shown by infra-red spectroscopic studies. The present results and those in the preceding paper indicate that the limiting adsorption is observed because only the monomeric fatty acid species is surface active, causing the surface coverage to tend to a limiting value as the concentration of the monomeric species in the equilibrium solution phase approaches an upper limit set by the aggregation process.The thermodynamics of adsorption at the solid/liquid interface with reference to the determination of the differential heat of adsorption of one component from a binary-solution has been considered by Crisp,' Everett and Corkill et aL3 These workers have shown that AH1 = ( a h I / a n S ) P , T , Z , n ; ,(1) where is the differential heat of adsorption at the solid/liquid interface of component 1 from the liquid solution phase, h, is the heat of immersion of the evacuated solid adsorbent into the binary liquid solution, ni and n; are the number of moles of components 1 and 2 respectively in the interfacial phase, and E is the area of the interfacial phase. Thus, to evaluate ml as a function of n: it is necessary to measure the heat of immersion of the adsorbent in a range of binary liquid solutions and also to know the adsorption isotherm of component 1.In practice, when one immerses an evacuated solid in a finite quantity of a binary solution from which one component (1) is selectively adsorbed the wetting process causes a dilution of the solution phase. In this case, where hT is the total heat change measured by the calorimeter during the immersion process, h, is the heat of immersion defined above and hD is the heat change caused by the dilution of the binary liquid phase. In many systems, hD is sufficiently small hT = h1-I-h~(2)

No abstract