Hi Pichaty, 

I would suggest that you stick with the same methodology that you use for Functional requirements - just note them as non-functional. In the 'good old days' of BRDs and the like, there would be a different section for Non-Functional requirements, but it was still in the same document. This should be carried forward for todays world where you may not be using a BRD (i.e. Agile). Keep to the same methodology otherwise the people who have to read your Non-functional requirements and carry them out may not understand what it is you are asking for. 

Sticking to the same methodology should remove this danger. A requirement us still a requirement - whether it is a functional or non-functional one. So why change methodology?

In an ideal world, a functional requirement should be linked to a non-functional one, and vice-versa. So splitting them out would probably make yours and other peoples lives,  working on your project harder.