Abstract
It is shown that the unitary transformation used as the basis of a t/U expansion for the Hubbard model in an earlier Brief Report is not equivalent to the canonical perturbation expansion (CPE) discussed in the preceding Comment. The CPE is valid only to leading order in t/U. More evidence is offered in support of our suggestion that higher-order terms in the t/U expansion tend to favor spin-liquid states over antiferromagnetically ordered states.
- Received 28 June 1989
DOI:https://doi.org/10.1103/PhysRevB.41.2565
©1990 American Physical Society