Comment on ``Roughening Transition of Interfaces in Disordered Media''

Emig and Nattermann (Phys. Rev. Lett. 81, 1469 (1998)) have recently investigated the competition between lattice pinning and impurity pinning using a Renormalisation Group (RG) approach. For elastic objects of internal dimensions $2 < D < 4$, they find, at zero temperature, an interesting second order phase transition between a flat phase for small disorder and a rough phase for large disorder. These results contrast with those obtained using the replica variational approach for the same problem, where a first order transition between flat and rough phases was predicted. In this comment, we show that these results can be reconciled by analysing the RG flow for an arbitrary dimension $N$ for the displacement field.