Adsorption at the Gas/Solid Interface

BY N. D. PARKYNS and K. S. W. SING

1 Introduction

The first survey of the literature on adsorption at the gas/solid interface, which appeared in Volume 1 of this series,1 was mainly concerned with the interpretation of the equilibrium data of physisorption (i.e. the adsorption energy and isotherm). In the present Report, which deals primarily with the literature of 1972 and the early part of 1973, emphasis is placed on the characterization of the gas/solid interface and the study of particular adsorption systems.

In recent years it has become increasingly clear that the distinctive shape of a physisorption isotherm is dependent not only on the texture of the adsorbent but also on the nature of the adsorbate-adsorbent and adsorbate-adsorbate interactions. In the region of low surface coverage, the isotherm character is directly related to the density and uniformity of the solid surface; if specific adsorbateadsorbent interactions are involved (e.g. hydrogen bonding between hydroxygroups), then the presence of certain functional groups on the surface becomes important. Considerable progress has been made in the characterization of oxide surfaces (particularly those of alumina, silica, and titania) by the use of i.r. spectroscopy. For example, the surface hydroxy-groups have been identified and the specific interactions with various polar molecules have been confirmed. The application of certain spectroscopic techniques (e.g. laser-Raman spectroscopy) is still at an exploratory stage, whereas n.m.r. spectrocopy is now firmly established for surface chemical studies.

Oxide adsorbents of high surface area are usually microporous or mesoporous (or both) and structurally ill-defined (amorphous or poorly crystalline). They readily undergo changes in texture and crystallinity as a result of sintering or lowtemperature ageing and therefore are not very suitable as adsorbents for fundamental physisorption studies. However, because of their industrial importance such materials have been studied in great detail. Various proposals have been made for the adoption of specially prepared oxides as nonorous reference adsorbents and tables of standard adsorption data (e.g. for argon and nitrogen adsorption) on these solids have been published.

The variation with coverage of the differential enthalpy of adsorption has revealed that the surface of a non-porous oxide is energetically heterogeneous. Of all the readily available high-area solids, graphitized carbon blacks probably provide the most uniform and stable type of surface. Moreover, the graphite basal plane, which forms each face of the polyhedron, interacts only non-specifically with a wide range of polar and non-polar adsorbate molecules. Furthermore, the overall adsorbate-adsorbent interaction energy may be reduced and the uniformity of the basal plane preserved if the graphitized carbon surface is coated by the preadsorption of non-volatile molecules. Such surface-modified graphitized carbon blacks, coated with polymers, have been used for gas chromatography, and those coated with pre-adsorbed xenon have been used for fundamental studies of the adsorption of nobleMgas atoms. These studies are extremely important in view of the difficulties encountered in attempts to improve the theoretical basis for the calculation of the intermolecular forces involved in physisorption.

The series of Specialist Periodical Reports3 'Surface and Defect Properties of Solids' deals inter alia with the spectroscopic properties of surfaces and adsorbed species and the identification of chemisorption sites. Some overlap with the subject matter of the present Report is therefore unavoidable, but the approach is quite different in the two cases. Here we are concerned with the relationship between the characteristic features of physisorption and the properties of the gas/solid interface, rather than with chernisorption and the defect properties of solids. The sorption properties of clays and zeolites, which have received a great deal of attention in the literature, will be reserved for a future Report. Diffusion and rate studies are also excluded from the present review.

2 Experimental Methods

Determination of Adsorption Isotherms. — Volumetric Methods. Recent developments in instrumentation have allowed several refinements to be introduced in the techniques used for the determination of adsorption isotherms. Pierotti and his co-workers,4 for example, have described in some detail a high-precision volumetric apparatus, which incorporates a null-capacitance manometer and a precision cryostat. This apparatus was designed for the investigation of the adsorption of Ne, Ar, Kr, and Xe at low coverage over the temperature range 65 — 300 K. Special precautions were taken in the calibration of the various parts of the dead-space volume, giving a precision in the adsorption measurement of 10 — 100 times greater than that usually obtained with a conventional volumetric apparatus.

The failure of the adsorbent to attain the same low temperature as the cryogenic bath has been investigated by Teichner and his co-workers. In the volumetric technique, good thermal contact exists between the adsorbent and the liquid nitrogen, but a slight difference in temperature (ca. 0.02°C) is found between the liquid nitrogen surface (the coldest region) and its bulk. This difference should be taken into account in accurate measurements.

In discussing the use of helium for dead-space determinations, Everett 6 has pointed out that the assumption usually made, that helium is neither adsorbed nor absorbed by the solid, is certainly not true for many microporous adsorbents, and that the only proof that the adsorption of helium is zero is that the apparent value of the adsorbent volume is independent of temperature (or if the data are sufficiently precise, varies with temperature in accordance with the coefficient of thermal expansion of the solid).

Krypton adsorption at 77 K is often used for the determination of low surface areas (e.g. of ceramics) and various modifications in the technique have been introduced to improve the accuracy and to assist in routine measurements. For this purpose, thermistor pressure gauges have been used, and a radiochemical method has been employed to measure the, β-activity of the adsorbate labelled with 85Kr.

Much attention has been given in recent years to the automation of the volumetric method for routine applications. Commercial equipment has been evaluated 11 and new designs have been introduced.12 A method of feeding results from a volumetric apparatus direct to a computer for determining BET (Brunauer-Emmett-Teller) areas has been developed by Deloye.13 Ciembroniwicz and Lason14 describe a semiautomatic manostat device which they have used for the determination of adsorption isotherms of Ar, N2, O2, and CO2.

Gravimetric Methods. Vacuum microbalance techniques are well established for the determination of physisorption and chemisorption isotherms. An example of the application of a recording microbalance for both physisorption and chemisorption measurements occurred in a study of the reduction of nickel oxide, which involved the determination of isotherms of argon and oxygen. The details of a microbalance designed for measurement of nitrogen adsorption have been given. Other microbalances have been adapted for the determination of adsorption isotherms over the temperature range 77 — 700 K. A short paper has described an extrapolation technique for reducing the weighing times in adsorption measurement. Robens has published the description of computer programs to handle the raw microbalance data and to print out a smoothed isotherm, together with other desiderata.

On the experimental side, a design of microbalance has been described based on the principle of alteration of the frequency of a quartz crystal on adsorption. A high sensitivity is claimed for this particular model. The resonating quartz crystal has also been used to study the adsorption of 4He at very low temperatures and the adsorption of water vapour at ambient temperatures. The papers presented at the Ninth and Tenth Conferences on Vacuum Microbalance Techniques have now been published. A number of these papers dealt with problems involved in the use of microbalances for adsorption studies, e.g. control and measurement of adsorbent temperature, the avoidance of mercury vapour adsorption, and the utilization of standard adsorption data.

Gas Chromatographic Techniques. Determination of adsorption isotherms by gas chromatography (g.c.) can be used as an alternative to the more conventional direct volumetric or gravimetric measurements. This approach is particularly useful where the adsorption is required to be carried out at higher than ambient temperatures or where the adsorptive is of low volatility. New developments of the technique have been reviewed by Gosselain, and by Hopfe and Marx. A comparative evaluation of methods available for determination of adsorption isotherms by g.c., and a description of a method for the calculation of the isotherms of both components of a binary mixture, have been given by Gerritse and Huber. In these papers the factors which influence accuracy (flow rates, pressure drop, an d rates of mass transfer in the column) are discussed; and a description is given of the determination of the isotherms of polar and non-polar adsorbates. Two new methods have been described 30 for determining adsorption isotherms of a vapour from very low concentrations up to saturation, and possible errors in the isotherm determination have been discussed. With some systems an apparent dependence of the slope of the derived isotherm on gas flow rate may be due to a failure to establish equilibrium. This problem could be overcome by increasing the temperature of the column.

Two other papers on chromatographic methods for determining adsorption isotherms have appeared. One deals with the use of pressure-drop columns and frontal analysis for the detection system, while the other proposes a means of eliminating the steep front associated with frontal analysis by replacing it with a steadily increasing signal and continuous injection of adsorbate.

G.c. is also used for the determination of chemisorption, very often involving high heats of adsorption, on to catalyst surfaces. Some examples, such as hydrogen on platinum and ethanol on silica-alumina, are given in a short review of the subject by de Mourgues. Topchieva and her co-workers have also described the evaluation of strongly acid sites by selective poisoning with pyridine and the subsequent determination by g.c. of the heats of adsorption of the weak bases benzene and toluene. A fairly pronounced dependence of the isosteric enthalpy on the degree of poisoning by pyridine enables conclusions to be drawn about the distribution of acid sites at the surface of zeolites. G.c. provides a rapid and sensitive technique for the determination of surface area. The adsorbent is made the stationary phase and the retention volume of a suitable probe molecule, such as n-octane or n-heptane, is measured. This arrangement is particularly suited to the study of organic polymers such as PVC, polystyrene, or polyethylene. A Russian patent has been granted for chromatographic apparatus for adsorption studies. An accurate apparatus for the determination of adsorption isotherms, of say nitrogen, at 77 K on various adsorbents has been described by Pommier, Juillet, and Teichner. This differs from other apparatus in having the gas flowing from the adsorption cell analysed by a separate g.c. column. The authors claim that the area of a solid can be determined from the resulting isotherm with an accuracy of ±5% with a total surface area of about 0.2 m2 in the cell.

Various modifications of the Nelsen-Eggertsen method 44 have appeared. A straightforward modification, using n-butane in a stream of C02 as carrier gas, has been described by Ruzicka. On desorption, the amount of butane evolved is determined by scrubbing out the CO2 with alkali and measuring the volume of residual gas. Other modifications of the Nelsen-Eggertsen method employ Ar-He and N2-He mixtures, the Ar and N2 being adsorbed at 77 K. Refinements have been introduced by Karp, Lowell, and Mustacciuolo to allow the continuous flow method to be used for the determination of adsorption and desorption isotherms.

The measurement of small specific surface areas from adsorption isotherms obtained by the dynamic method has been discussed by Olah, Gaspar, and Borocz with particular attention given to errors aris,ng from the non-linearity of the detector. The role of thermal diffusion in producing a spurious response was investigated by Kourilova and Krejci, with particular reference to the flow rate and column dimensions. The use of a simplified commercial version of the Nelsen Eggertsen technique — the Perkin Elmer Sorptometer — has been evaluated by Bartels, who concludes that although the technique can be employed for rapid comparison purposes, accurate measurements demand a set of standard samples of known surface area. The use of particular organic vapours for dynamic measurements has also been described.

All the methods described here which employ chromatographic analysis require calibration with a known gas mixture. Pure nitrogen can be obtained quantitatively by the decomposition of ammonium dichromate, and it has been suggested that calibration in this way enables specific surface areas of < 1 m2 g-1 to be determined.

Miscellaneous Techniques. The study of controlled thermal desorption is largely associated with the names of Cvetanovic and Amenomiya, who introduced the technique of Temperature Programmed Desorption (TPD). This technique has been applied in various investigations of particular systems, but no radically new changes in experimental procedure have appeared. Calculations of desorption energy and rate factors have been made by Czanderna and his co-workers, who have used a microbalance to produce the desorption data as a function of temperature. Similarly, stepwise temperature-desorption from molecular sieves was used to measure the rate of desorption as a function of initial coverage; the data were subsequently used in the analysis of binary gas mixtures.

Dollimore and co-workers have used a mass spectrometer for studying desorption from carbon surfaces. The evolution of surface oxides formed during oxidation can be detected in this way. A few papers have appeared dealing with the determination of very low specific surface areas using desorption techniques. Thus, the desorption of Xe from metal or carbon filaments can be used to measure their surface area by flash desorption at high temperature. Measurement of the time taken to pump out carbon blacks is alleged to give a good correlation with surface area and oxide content.

An unusual technique is that of torsion-effusion, which measures the rate of desorption as the change in torque of a suspended thread on which the adsorbent cell is mounted. This method was used to measure water adsorption on zeolites at equilibrium pressures of 10-2 — 10-4 Torr over the temperature range 280 — 350 K.

Although adsorption at above atmospheric pressure has not been very much studied because of experimental difficulties, there is considerable technological and theoretical interest in such studies. A suitable volumetric apparatus was designed by Zhukov and Serpinskii for the determination of CO2 adsorption on zeolites at pressures up to 72 bar and over the temperature range 273 — 423 K. Another apparatus has been developed for measuring adsorption of O2 and N2 from gas mixtures in He at 70 K and at high pressures by means of a recirculating loop system.

The effect of gas adsorption on the dimensions of the adsorbent has been studied with a limited number of systems. In one recent study65 the dimensional changes caused by the adsorption of water and other vapours on Saran charcoals were explained in terms of the bridging of micropores. Dubinin and his co-workers have also investigated the phenomenon in relation to water adsorption on zeolites and clays. Details are given of the optico-mechanical dilatometer used for this work. The elastic constants due to adsorption of water on compacted kaolin samples have been measured by means of a linear-displacement transducer. In another investigation, the changes in the dimensions of porous glass were related to the heat effects produced by the sorption and desorption of water vapour at temperatures between 230 and 280 K.