Hi Miles,

I am trying to calculate the magnetization susceptibility for a quantum ferromagnet and in the process, I need the expectation value of @@M_N^2@@, where $$M_N \equiv \frac{1}{N} \sum_i S_i.$$

This implies that one requires the expectation value of @@S_i S_j@@ for all @@i, j@@. Although this scales terribly (@@\sim L^4@@) for a 2D system, in principle, I understand how to compute this using the same methods documented for computing two-point correlation functions.

However, I was wondering if there is a more efficient way of computing this. In particular, suppose I define an operator $$ O \equiv \frac{1}{N^2} \sum_{i,j} S_i S_j, $$ which is nothing but @@M_N^2@@. Then, is it possible to simply construct this full operator as an MPO and contract it with the MPS ground-state yielded by DMRG? If so, I am not quite sure of how to gauge the MPS correctly.

Thanks a lot for your time.

Best,

RS