Abstract We advance an original assumption whereby a good state of the environment positively affects labor productivity in R&D such that deteriorating environmental quality negatively impacts R&D. We study the implications of this assumption for the optimal solution in an R&D-based model of growth, where the use of a non-renewable resource generates pollution. We show that in such a case, it is socially optimal to postpone extraction, as opposed to the situation in which the environment has no effect on productivity in R&D. Furthermore, insofar as environmental quality declines and subsequently recovers, we find that it is optimal to re-allocate employment to R&D in line with productivity changes. If environmental quality recovers only partially from pollution, R&D effort optimally begins above its long-run level, then progressively declines to a minimum and eventually increases to its steady-state level.