Abstract: The microbalance in chemisorption and catalysis

Abstract

The microbalance of primary concern here is the commercial beam balance and not the quartz spiral‐spring balance or the quartz resonating‐crystal balance. The restriction to commercial balances reflects a frequent choice to avoid the problems and time consuming process of microbalance construction. With a quartz spiral spring, the minimum surface area on which monolayer adsorption can be measured with 1% precision, is approximately 10 m2/g. For the commercial beam microbalance the corresponding value is closer to 0.1 m2/g. For surface areas significantly below this, the quartz‐crystal resonator microbalance must be used. In effect this means that for high‐surface‐area granular or powder materials, either the spring balance or the beam microbalance must be used. Where low cost and simplicity of operation are important, the former may be desirable. When low surface area (<10 m2/g) materials are under study the beam microbalance or (for evaporated films) the quartz resonator crystal should be used.Beam microbalances are not without problems, and these usually fall into one of three classes: electrostatic, thermomolecular, and buoyancy or convection. Electrostatic problems may be eliminated by carefully grounding the sample holder and hangdown wire. Thermomolecular forces give rise to spurious weight changes in the pressure range 0.13–2660 Pa. These may be reduced by maintaining maximum symmetry in the balance chamber, the residual effects then being calibrated and allowed for. At higher pressures (typically ≳133 Pa) buoyancy and subsequently (13.3–40 kPa) convection difficulties arise. For precise measurements this necessitates the use of a tare material of near identical density and may involve (helium) density measurements of the sample under investigation. Convection effects may be reduced by careful hangdown tube design.One of the most important applications of the microbalance is the precise measurement of surface areas. These may be total areas, as measured by nitrogen adsorption at liquid‐nitrogen temperatures, or partial surface areas of supported metal crystallites, as measured by (say) hydrogen chemisorption. The gravimetric approach avoids the possibility of cumulative error during gas inlets in a volumetric adsorption system and can measure lower surface areas (0.1 m2/g as opposed to 0.5 m2/g). Care, however, must be taken to avoid small differences in temperature between the sample and the walls of the hangdown tube. In relation to the measurement of partial surface areas, it is often wise to check that the chemisorbing adsorbate does not spillover onto the ’’inert’’ support material. It is also interesting to observe the effects of high temperature outgassing on these supports. Many complex silicates release physisorbed water, and chemisorbed water, then undergo reversible and subsequently irreversible dehydroxylation as the outgassing temperature is gradually increased. Such effects may be monitored directly by measuring the weight loss and analyzing the gas phase with a residual‐gas analyzer. Alternatively, the effects may be partially or completely backtitrated with water vapor. Water adsorption in this case would be a mixture of physisorption and chemisorption. Other interesting examples of weak chemisorption that have been studied include, for example, the hydrogen‐bonded adsorption of long‐chain fatty alcohols, an adsorption which produces a new hydrophobic surface. In all these cases long term microbalance stability is essential.Microbalance systems lend themselves to the in situ preparation of catalysts in an oxidized or reduced form and the nature and amount of gas involved or consumed may be continuously monitored. We have had particular success with the low temperature formation of Ni, Cu, and Cu–Ni alloy powder materials.1 The ability of the balance to detect minute weight changes was used to define the onset of the reduction process of the corresponding oxides. Virtually the entire reduction was then carried out at a low temperature with only the residual final step being carried out at high temperatures. The resultant catalysts were found to have not only a higher surface area, but also a higher specific activity as well as a greater homogeneity.In studying catalytic reactions themselves the obvious requirement of some kind of sample weight change would seem to eliminate the ’’flow microbalance reactor’’ for many systems. In many processes, however, catalyst weight changes occur though they may not always be apparent. Thus, gravimetric catalytic study of the oxidation of isopropyl alcohol to acetone on cupric oxide becomes feasible because of the strong adsorption of the acetone.2 Catalytic polymerization is similarly amenable to study in this way.3 In addition, the undesirable, and sometimes unsuspected, process of carbon deposition can be examined. Systems examined in this way have included the decomposition of formic acid on boron phosphate4; of ethylene, cis‐2‐butene and acetylene on nickel5; and of benzene on iron, cobalt, nickel, and their alloys.6 Other apparently stable hydrocarbons (e.g., propane) have given indication of such decomposition,7 but have yet to be thoroughly examined in a flow‐microbalance reactor. No doubt many other new applications are yet to be found particularly where the microbalance is used in conjunction with some other technique.