I agree that $$\hat a|0\rangle=0$$ But then, based on the above, the following should hold $$\hat a_k |N_1,...,N_{k-1},0,N_{k+1},...\rangle=|N_1\rangle\oplus\cdots\oplus |N_{k-1}\rangle\oplus \hat a_k|0\rangle \oplus|N_{k+1}\rangle\cdots$$ $$~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~=|N_1\rangle\oplus\cdots\oplus |N_{k-1}\rangle\oplus 0 \oplus|N_{k+1}\rangle\cdots$$
However, I've seen QFT textbooks claim that the calculation should result in $0$. $$\hat a_k |N_1,...N_{k-1},0,N_{k+1},...\rangle=0$$ Then does this mean that $$|N_1\rangle\oplus\cdots\oplus |N_{k-1}\rangle\oplus 0 \oplus|N_{k+1}\rangle\cdots=0\tag{*}$$ I don't see how what we've derived makes sense. Perhaps someone could explain why (*) makes sense mathematically?