@Condor are they for sure absent from captivity or are there facilities that have them, and if so which places are they ?
I have not received new information on the possible captive status since my earlier comment. Quote of the part dealing with captives, as it made my position clear (first four words):
Purely speculative: My guess is that the captives in Europe and North America are pure T.kanchil but probably also some T.kanchil X T.javanicus hybrids. Since relatively few T.javanicus are likely to have been imported compared to T.kanchil and few breeders cared much about the origin of their animals until the publication by Groves and Meijaard in 2005, it seems likely that the T.javanicus eventually were intermixed with the T.kanchil. They're presumably very close relatives and the differences upon which they were declared separate species are quite small. I wouldn't be surprised if they can hybridise relatively easily and the offspring is fertile.
The exact same can be said about T. versicolor and T. williamsoni. It is certainly possible that some were in captivity. From the limited available evidence, they appear to be found in Vietnam, northern Thailand and southern China - all places with a long history of animal collection. No one considered them distinct from the widespread greater/lesser chevrotain until 8 years ago. Without accurate collection data, it would actually be quite difficult for a zoo to be sure, even today. The slight silvery sheen of T. versicolor is not particularly distinct, and identifying T. williamsoni from external features would be even harder. That leaves skull measurements. How many zoos have skulls of their dead chevrotains cleaned and then measure them? How many zoos have people that actually know how to make these measurements? I suspect few.
However, because of their limited distributions and apparent rarity, I think it is entirely reasonable to assume that no pure T. javanicus/versicolor/williamsoni are in captivity today (at least outside their native range)
until:
1) Someone provides a photo of a captive that looks like the distinct grey-necked type of T. javanicus, or
2) Matching skull measurements of specimens from a captive lineage, or
3) Captives with known origin (not just from a local animal market; especially Javan markets are notorious for having animals from all around Indonesia).
Genetics were not provided by Groves and Meijaard 2005. Until someone makes a well-sampled genetic comparison of specimens of known origin, DNA samples from captives are of no use.
I might add that chevrotains in Java (regardless of their taxonomic status) based on limited evidence appear to be quite rare. They may well be seriously endangered. This can be compared to the endangered Javan slow loris, long overlooked because it was merged with several other more widespread species.