This paper presents a brief review of recent results from the theoretical modelling of ammonia synthesis, based upon kinetic data from single crystals and catalysts. The conclusion from this work is that correct predictions of industrial yields can only be made by including a highly activated step of nitrogen dissociation from the gas phase. This fits with recent experimental evidence that nitrogen dissociation on Fe(ll1) is thermally activated, a finding supported by theoretical calculations and older work on real catalysts. Models which do not inchme this activation barrier predict low yields of ammonia. Experimental work on single crystals indicates that the "active site" is intimately connected with the two major promoters (alumina and potash) and that these play a multiplicity of roles. The catalyst for the industrial synthesis of ammonia is a classic example of an empirical development made by technologists (applied scientists) which has tared the minds of fur&mental researchers for the last 70 years, These researchers have been attempting to describe (alter) the mechanism and kinetics of tbe reaction at the molecular level; good reviews of this work are available elsewhere [l-S]. At the centre of this research is the nature of the "active site" at which the rate determining step occurs. It is not clear, as will be described in more detail below, that we know what this active site is, but an attempt to define it will be given (section 6). The surest sign of ignorance of the way in which current catalysts work is the lack of new commercial materials for this reaction, most current versions being slightly modified forms of the original 'Swedish magnetite" [3], and consisting of approximately 92% Fe oxide, 0.5% KsO, 2% Alz03 and 2% CaO and other ingredients [3]. One new material is promoted Ru [6], which will be utilised iu a full plant in the near future.
Having indicated our ignorance on this subject, there has nevertheless continued to be a good deal of work in this area, because it is also of great academic interest. This probably stems from the following features (i) simple reagents and products, (ii) 100% selectivity to the