Dear ALPS community,
I am trying to check if DMFT(QMC) calculations are consistent with my
DMFT(NRG) calculations. I am working on Kondo lattice model (KLM), and I
guess to simulate it with QMC, I have to consider PAM with correct
parameters that translate to the KLM limit; the repulsion of f states
should be high, U_f=8, on site energy e_f = -4 and hopping on the c states
for example t=0.8. There is also spin-spin interaction J in KLM, here just
the interaction between c and f orbitals, in PAM just V.
Correct me if something is wrong please:
a) To simulate such lattice, I should set 4 flavors. The U matrix should
look like (c_up c_down, f_up, f_down):
[0 0 V 0 ]
[0 0 0 V ]
[V 0 eps_f 2*eps_f + U_f]
[0 V 2*eps_f + U_f eps_f]
How do I set such a matrix in parameters? (U1 = U_f, ...)
b) The hopping terms are t0=0.8 and t1=0 ? (ALPS automatically assumes 0, 1
belong to first site, 2, 3 to the second?)
c) I need antiferromagnetic solutions; is it sufficient to calculate using
ANTIFERROMAGNET=1 or should I use some sort of cluster schemes to really
get the correct spectral function? (I using a cluster scheme as simple as
setting SITES=2 for example?)
d) I need to simulate the model in a magnetic field, is this possible in
QMC? Is magnetic field possible in x direction (the spin symmetry is broken
here, there is no S_z symmetry anymore)? The solution when magnetic field
is applied also creates AF magnetization in x direction, is it possible to
break symmetry in a way to achieve this (small field in x direction, get it
to 0 ...)
Best regards,
Ziga