Hi,

1. ) If you are using two lj potentials for the two different interactions ( one between group 3 type and group 4 type ; and another between group 2 type and group 4 type) , then each of the lj potentials must be distinguished from the other by using unique integer id as suffix to it (Note that there is a space between lj/cut and 1) :

############################################

pair_coeff 3 4 lj/cut 1 0.04 0.97 NULL NULL W C

pair_coeff 2 4 lj/cut 2 0.02 0.5 NULL Cr NULL C

###############################################

Now lj/cut 1 and lj/cut 2 represent the corresponding lj potentials for interactions between two different sets of pairs, namely 3-4 and 2-4. The integer ids 1 and 2 distinguish these two lj potentials from each other. In the absence of the integer IDs , the software will read the two lj potentials as lj/cut 0.04 and lj/cut 0.02, and in context of not finding integer in the ID suffix part, it will pass the message that it is expecting an integer there. In summary, when a given interatomic potential is used more than once, then its name must be followed by the integer ID to render each pair_coeff lines unique .

The two eam/fs potentials can also be made distinguishable/unique by using the integer ID suffixes:

###################################################

pair_coeff * * eam/fs 1 FeCrW_s.eam.fs Fe Cr W NULL

pair_coeff * * eam/fs 2 Fe-C_Hepburn_Ackland.eam.fs Fe NULL NULL C

#######################################################

2. ) Moreover, in the pair _style line, the two lj potentials with cutoff (r_c) values must be included.

###################################################

pair_style hybrid/overlay eam/alloy eam/fs eam/fs lj/cut 2.5 lj/cut 2.5

###################################################