Hello,
Could you check if the convention of the literature and in TB2J are the same?
From Eq. 1 of Hutchings M. T., Samuelsen E. J., Phys. Rev. B 6, 9, 1972, 3447, the summation runs over distinct ij pairs, whereas in TB2J , it runs over both ij and ji (so counted twice).
Therefore, the J in TB2J should be 1/2 of that in the reference paper.
Best regards,
HeXu