Why the DFB laser needs RPCVD, page-12

Obviously you didn't read the paper it actually called their method "plasma-assisted molecular beam epitaxy (PAMBE)". If you read their rationale it was all about avoiding a deposition method which (such as MOCVD) which pacifies the p type GaN. Yes, MBE also does that but has significant "growth" and preparation issues compared to MOCVD (which doesn't work for this application). RPCVD also is an "active as grown" deposition method which has most of the advantages of MOCVD but without the productivity issues of MBE. Yes MBE could be used but if you do their are productivity (i.e cost) issues compared to RPCVD