Inconsistent charge density wave in the triangular lattice Hubbard model

Hello,
I have some unexpected results for the charge density in the triangular lattice using the Hubbard model at 1/4 filling.

I’m basically calculating this quantity:

expect(psi, “Ntot”)

For a system of 24x2 sites with PBC I get kind of the following densities for sites away from the boundaries:

0.48915028610176636
0.48822772815756676
0.5171531499656885

and repeats as long as I’m away from the boundaries.

I’ve been playing with very small U (0.05 or 0.1), even U = 0, and I always get the same, even though I would expect a homogenous charge density of 0.5 (since I’m at 1/4 filling), but I get a sort of CDW

I’ve also been playing with the size of the system and nothing changed.

I usually get 15 sweeps with maxlinkdim=6574 and maxerr=1.00E-15

I’m actually interested in a different system with triangular symmetry and still get that CDW (even more inhomogeneous) for a very small U, even U = 0.

What would be the source of this problem?
Highly degenerate states for small U?
Is hard for DMRG to do calculations away from half-filling?

Thanks!