Self Consistent Expansion for the Molecular Beam Epitaxy Equation

Motivated by a controversy over the correct results derived from the dynamic renormalization group (DRG) analysis of the non linear molecular beam epitaxy (MBE) equation, a self-consistent expansion (SCE) for the non linear MBE theory is considered. The scaling exponents are obtained for spatially correlated noise of the general form $D({\vec r - \vec r',t - t'}) = 2D_0 | {\vec r - \vec r'} |^{2\rho - d} \delta ({t - t'})$. I find a lower critical dimension $d_c (\rho) = 4 + 2\rho $, above, which the linear MBE solution appears. Below the lower critical dimension a r-dependent strong-coupling solution is found. These results help to resolve the controversy over the correct exponents that describe non linear MBE, using a reliable method that proved itself in the past by predicting reasonable results for the Kardar-Parisi-Zhang (KPZ) system, where DRG failed to do so.