Monte Carlo simulations of the magnetic structure factor of the two-dimensional Hubbard model are in qualitative agreement with the slave-boson approach and, in the low-temperature intermediate-coupling limit, they are in even better quantitative agreement with the random-phase approximation, as long as a renormalized repulsion U is used. This renormalization comes from maximally crossed diagrams, which account for two-body short-range correlations. One of the consequences is that Stoner ferromagnetism is not a generic property of the two-dimensional one-band Hubbard model.

1 aChen, L1 aBourbonnais, C1 aLi, T1 aTremblay, A.-M., S. uhttps://www.physique.usherbrooke.ca/pages/node/7237