Bill,
         When I have tried this the biggest problem is that the
samples corrode their way out of the tin cups in a matter of minutes,
the resulting mess in the autosampler carousel is quite impressive. I
also find that it attacks the quartz tubes of the combustion furnace
and shortens their life considerably.
 Procedure here now is to use the old steam distillation method to
get the NH4 into a more manageable solution. This method is now out of
favour in many labs as it is complicated & time consuming but it works
fine.  There are diffusion techniques which could also be adapted but,
again, it will entail further sample preparation.

   Darren Sleep
   Stable Isotope Facility
   ITE Merlewood
   Cumbria UK.