A priori design of catalysts is not yet possible. Such task would dema
nd unavailable scientific knowledge of the correlations among synthesi
s parameters and resulting solid state and surface structures, on the
one hand, and among those atomic-level structural details and their ca
talytic functions, on the other hand. To avoid testing every possible
combination, therefore, the applied chemist or chemical engineer must
identify empirical correlations underlying the existing experimental d
ata base. The ability of artificial neural networks to identify comple
x correlations and to predict the result of experiments has recently g
enerated considerable interest in various areas of science and enginee
ring. In this paper, neural networks are used to identify composition-
performance relationships in automobile exhaust catalysts. This work e
mploys an artificial neural network technique to do a sensitivity anal
ysis of the conversions of pollutant gases as a function of the cataly
st composition and the operating conditions. This approach converges o
n the optimum catalyst composition and operating condition in order to
produce specified conversions of carbon monoxide, hydrocarbons and ni
trogen oxides, to carbon dioxide, water and di-nitrogen respectively.