I am working on Ti coating on top of diamond layer to figure out how the diamond in-plane thermal conductivity changes as the Ti coating thickness is changed. I am using LAMMPS for my NEMD simulation. I have defined my regions and grouped them according to the following script.
---------- Group Definitions ----------
Example: if atom type 1 = C (diamond) and atom type 2 = Ti
group diamond type 1
group titanium type 2
Define regions for heat layers
region hot_vacuum block INF INF 0 5 INF INF
region cold_vacuum block INF INF 184.971 189.971 INF INF
region fixed_heat_source_side block INF INF 5 15 INF INF
region fixed_heat_sink_side block INF INF 170 184.971 INF INF
region hot block INF INF 15 45 INF INF
region cold block INF INF 140 170 INF INF
Create groups from regions
group region_hot region hot
group region_cold region cold
group region_hot_vacuum region hot_vacuum
group region_cold_vacuum region cold_vacuum
group region_fixed_heat_source_side region fixed_heat_source_side
group region_fixed_heat_sink_side region fixed_heat_sink_side
Now form groups by intersecting with diamond
group hot intersect region_hot diamond
group cold intersect region_cold diamond
group hot_vacuum intersect region_hot_vacuum diamond
group cold_vacuum intersect region_cold_vacuum diamond
group fixed_heat_source_side intersect region_fixed_heat_source_side diamond
group fixed_heat_sink_side intersect region_fixed_heat_sink_side diamond
Form unions for end groups
group end_one union hot_vacuum cold_vacuum
group end_two union fixed_heat_source_side fixed_heat_sink_side
group end union end_one end_two
Now, my understanding was, if my region and group commands are correct, the heat will go from source side to the sink side. As Ti is bonded to C, during this heat progression, Ti atoms will also be affected as heat is travelling from source of the diamond to the sink of the diamond. Kindly, keep in mind that, I did not define source and sink region in the Ti since electron is the main carrier in Ti, it will not be feasible.
So my question is:
is my region and group definitions are correct or not?
Since I used “Group Intersect” command, is there any possibility that my Ti will NOT be affected at all by the heat?
My guess was since Ti is metallic, applying Ti coating on diamond will increase the in-plane thermal conductivity value. But I am getting completely opposite result. As Ti thickness is increased, in-plane thermal conductivity gradually decreased. Why is that?
Any suggestions, explanations, or idea will be highly appreciated. Thanks.
Questions 1 and 2 are something for you to figure out yourself. It is very simple to determine if a group covers the intended atoms by using the write_dump command before and after a delete_atoms command with that group. Then you can compare the two dump files and determine if the expected atoms were deleted. With a region you do the opposite and create atoms in that region and thus you can see which volume the region covers.
Question 3 is not about LAMMPS but about your research and thus something you need to discuss with your adviser, supervisor, or tutor. It is off-topic for this category where we focus on LAMMPS questions (and the fact that you are using LAMMPS for your research does not make questions about your research LAMMPS questions).
Thanks for your suggestion. Also, I apologize for posting research related question in LAMMPS thread. This is my first time posting so I did not understand.
I can offer some very general technical comments (bearing in mind that, as @akohlmey said, it is up to you and your group to be the primary experts in your subject matter; LAMMPS just does what you tell it to).
You must know how to “read out” “local quantities” from your simulation. Please learn how to use compute chunk/atom with fix ave/chunk to determine, for example, what the local temperature (or more precisely the local average particle kinetic energy) is in slices across the box, especially for only the “diamond” and only the “titanium” subsections.
Also determine the temperature profile across slices “parallel” to the heat flow. Different parts of the “titanium” are at different distances from the “diamond”.
Statements like “electrons are the main carrier in Ti” and “Ti is metallic” can be very misleading when you are doing modelling in general, and molecular dynamics in particular. There is no guarantee that your model captures all aspects of how the real world substance behaves. For example, your MD simulation has no explicit “electrons” running around in a “conduction band” of “titanium” (hence the quote marks); the MD “titanium” is a bunch of spheres exerting classical forces on each other (even if they are many-body) and there is no guarantee that those dynamics capture any aspect of the conductive “electron gas” inside real world titanium.
Indeed, you should probably start by benchmarking the thermal conductivity of your MD “titanium” and comparing that to real world titanium. If your “titanium” were less thermally conductive than your “diamond” then it would be entirely reasonable for the system’s average conductivity to decrease as the proportion of “titanium” increased.
and it is loaded with useful advice on posting in this forum section.