Greetings Linfield,
Apart from some microbiological parameters (eg like Salmonella in Europe which has to be tested in a 25g sample according to legislation - EC 2073/2005) there are no specific limits for everything. Generally the sample quantities depend on how many parameters you are going to check, as each one will require its own quantity from the sample, or if you have double samples in case something goes wrong with the1st sample. However, what you should consider when collecting the sample is how representative it is of the material you are testing. This means that if you take a 50g sample from a specific spot of one ton of product then it doesn't show you the actual condition of the bulk. However if you omogenize it or take smaller samples form different spots/depths/areas and put it all together to form a 50g sample then its more representative of the bulk and you will get more accurate data.
You should also consider sending the samples to an accredited lab (if it is not already the one you used), cause in my experience from the lab I work for and others we never ask for the smallest possible quantity. We always go for at least double the needed quantity or for double samples and making sure that they are as representative as possible of the whole bulk/production etc.
Of course the client has the last word if he wishes to send a 0.5g sample 
but still we inform them about the merits of the above mentioned practice!
PS If you are within limits and the process afterwards involves a kill step, then you shouldn't worry too much, unless you systematically receive positive out of limits results from your agglomerated powders. However you are bound to have positive results -within set limits-, as if the process resembles let's say beer brewing then the powders will have a small microbial content, which is being eliminated during the preparation steps for the yeast inoculation.
Hope it helps!
