Abstract
The ordering dynamics of a system with a non-conserved order parameter is considered following a quench into the ordered phase from high temperature. Newman and Bray (1990) have set up an expansion in powers of 1/n for the O(n) model to obtain correlation functions and the two-time exponent lambda , but their calculation contains a simplifying assumption which is incorrect. Tight upper and lower bounds for lambda are obtained as a function of the space dimension d. These bounds exclude the result of Newman and Bray, although the dependence of lambda on d is qualitatively very similar. Comparison with simulations shows that the first-order 1/n calculation does not agree with numerical results as well as previously thought.