I often have to use sto-3g to converge the scf for metal-containing systems, and then project the wfn onto a larger basis. But this one segfaults in computing the projection.
molecule {
0 5
FE 0.000000000000 0.000000000000 0.123274779640
F 1.565379000000 0.000000000000 -0.647320000000
F -0.000000000000 -1.700830000000 0.639825000000
F 0.000000000000 1.700830000000 0.639825000000
F -1.565379000000 -0.000000000000 -0.647320000000
}
set {
basis sto-3g
guess sad
scf_type direct
reference uhf
}
energy('hf')
basis {
assign 6-31++G**
assign Fe 6-31G**
}
set guess read
optimize('hf')
I am using the 1.1 release, Git: Rev {HEAD} add49b9, on Linux x86_64.