Thanks Charles, but as mentioned I cant find supporting documentation stating that the EC limits is = 0, or anything otherwise. A lot of sources and legislations state the allergens but does not mention when it should be declared. Maybe I missed it, will go through them again.
That's because there are no limits well, almost, (I will explain that statement further in a mo.) The reason being that different allergenic consumers have different sensitivity and currently there is insufficient research to prove a limit can be "ok". It could also be argued that if you detect some allergens, albeit at a low level, you cannot be certain of the homogeneity of the batch so if they are detectable at a low level, it's possible they could be detectable at a higher level elsewhere.
Next, it also assumes complete recovery from your food matrix when the product is tested. I've had experience with sulphites where this is very poor and I know someone else has had very big problems with egg when cooked with wheat so your lab result of, say, "10ppm" may actually be out by a vast margin. (I kid you not, I had a sample I knew to be c. 1000ppm SO2 reported as <10 and I know someone else who did a controlled trial on egg at similar levels and had results out by factors of 100s.)
How EU manufacturers approach this is by declaring all ingredients which go in which contain allergens (with a few minor exceptions e.g. glucose derived from wheat) even if that allergen is in all likelihood destroyed by the process. So you will always see "barley" declared on vinegar but you'll never find gluten if you test it.
Then you validate clean downs between products. How we do it is possibly OTT vs. what others do but we test positive controls (which checks the lab can actually recover the allergen, which, as I explained is not a given) and "dirty" swabs then test clean swabs and the first three products off the line to ensure the clean is effective. We then use visual cleanliness as our ongoing verification (but we use minimal allergens at my site). Other sites using a lot of allergens may use regular rapid swab tests for verification.
As for limits, there are two, kinda. SO2 or sulphites if <10ppm does not have to be declared. Cereals containing gluten doesn't officially have a limit but you can declare something is "gluten free" if there is <20ppm.
So in summary, if you can detect it for EU sale in most cases, that's a problem. As a result, and this does seem disingenuous, be very careful about what you test. I recommend only testing product or processes where you can hold the product until you know the results or you are very confident of your processes, otherwise you could be having a public recall.