There are some recent papers on doing it, and they even seem to have gotten decent efficiency. The lab I was in was doing this in 2000, and I suspect they got

There are some recent papers on doing it, and they even seem to have gotten decent efficiency. The lab I was in was doing this in 2000, and I suspect they got their frequency doubler from some other lab. It worked, but it certainly didn’t work well. The 1064nm laser was decently large (a couple J per pulse, from vague memory), and the expected UV energy was quite low. The laser was being used for some form of imaging at short range (fluorescence or absorption in burning gasses? Maybe Raman spectroscopy if everything got lucky?).