# Thermal conductivity of multi-component fluids

Is the formula in the given compute heat/flux not applicable to multi-component fluids? Recently, I have read some literature and found that some people have pointed out that the LAMMPS formula is incorrect for multi-component fluids and needs to consider the partial enthalpy. However, I have seen many articles that use LAMMPS directly to do it without mentioning the partial enthalpy value

Without knowing which articles you refer to, it is impossible to give detailed comments. I would say that I wouldn’t immediately trust a multicomponent fluid simulation done with non-equilibrium heat flow – thermophoresis (particles’ mobility coefficients in thermal gradients being unequal) would be an immediate concern.

https://www.sciencedirect.com/science/article/pii/S016773220600331X#section-cited-byhttps://www.sciencedirect.com/science/article/pii/S0378381204002274#aep-bibliography-id27
I have read articles about calculating the thermal conductivity of multi-component liquids, and most of them do not mention the partial enthalpy term. However, there are several articles that mention it, such as the one I provided above. I don’t quite understand the partial enthalpy term, which roughly means “this part of the energy is not transmitted, but silently carried, so it needs to be subtracted? Because it seems to be divided into the convective term.”. Of course, I don’t quite understand whether my explanation is correct, and I still haven’t figured out where the partial enthalpy value comes from. https://pubs.aip.org/aip/jap/article/112/5/054310/916898In the following article, his conclusion is that it can be disregarded for liquid mixtures. Has LAMMPS not implemented this type of problem? From the “compute heat/flux” page of LAMMPS, is what LAMMPS provides only applicable to calculating pure components? I hope to receive the help of professionals

In the vast majority of the articles I have read, it seems that the issue of multi-component partial enthalpy terms has not been mentioned, but rather the calculation is directly carried out using LAMMPS. Is this reasonable?