So you are taking the product of b^(1/n)-c^(1/n) over all even squares n in addition to n=1? For starters, that doesn't make sense when either of b,c is negative. Secondly for positive b,c this should converge to 0...since b^1/n and c^1/n both approach 1.
Again, being a member of n^2 doesnt...