Well, RP reacts with iodine to form HI (hydroiodic acid) which is a reducing agent. BUT it is also able to reduce RP to phosphine - a very toxic gas. Phosphorous triiodide is also a product and that is able to react with norcodeine, normorphine, nordesomorphine-C and nordesomorphine-D (nordesoxymorphine) to form phosphoramides - a toxic class of chemical.
So as it stands, korkodil WILL contain at least some of the above.
A step up is to make the HI, remove excess RP and then go on to use it - I've seen that done. But you still have the phosphine risk most people are unaware of.
I would use calcium hydroxide to form the calcium salt of the product (which is insoluble so drops out of solution) and go from there to make the hydrochloride. Of course, mechanical losses only make it practical at scale, or at least not the handful of doses a batch normally provides.
Overall, I think it better to steer clear of the stuff. There were serveral unidentified compounds so it's not possible to say that the above would remove them. I learnt of krokodil long ago and made a concerted effort at HR but what's better? Telling people how to make it safer and risk more 'cookery' or not tell people and let the damage occur? I don't have the answer to that one. I just know desomorphine was only use briefly in one nation and was then taken out of use, so maybe even in it's pure form it has adverse qualities?
