Heat transfer in a microscale model using pulsed heating

Please login with a confirmed email address before reporting spam

Hello everyone,

I am new to COMSOL and I am trying to build a model to verify experimental results acquired in the following paper: https://www.science.org/doi/10.1126/sciadv.adj3825

For brief summary, the experiment used pulsed electron beam to heat a sapphire sample, thereby inducing thermal waves inside the sample. In the paper, the beam (heat source) was moved and the thermocouple (detector) was stabilized. However, in COMSOL, for simplicity, I kept the heat source constant in one position and placed point probes (1 to 5) along the length of the sample, as described in the attached "Schematic figure.jpg" file.

Theoretically, the thermal waves detected at further distances from the heat source are going to delay more compared the ones detected at nearer distances. The phase delay is determined by the following formula: Δθ=θ1-θ2=(√(πf/α))L, in which Δθ is the phase delay difference, f is the pulsed frequency, α is the thermal diffusivity, L is the distance between point 1 with a phase delay θ1 and point 2 with phase delay θ2.

My problem is the simulation results are not consistent with the experimental and theoretical results when comparing the thermal waves detected by point probes 1-5, as in the attached .mph file. So my question is:

Is it appropriate to build such a model using COMSOL (micro-submicron size, monitoring thermal wave to compare phase), since I guess this might not a job that people frequently do with COMSOL? Please kindly point out if I am doing wrong somewhere (which I suppose very possibly because I have little experience with COMSOL) or any ideas to solve this problem. All of your ideas or comments are highly appreciated. Thank you!



4 Replies Last Post Jun 25, 2024, 7:16 p.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jun 21, 2024, 12:14 p.m. EDT
  1. You must have several time steps (5-10, say) per period of a periodic excitation.
  2. Use strict time stepping.
  3. I don't see a good reason for boundary layer mesh in this problem.
1. You must have several time steps (5-10, say) per period of a periodic excitation. 2. Use strict time stepping. 3. I don't see a good reason for boundary layer mesh in this problem.

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jun 24, 2024, 9:17 p.m. EDT

Hi Dave Greve

Thank you very much for helping me.

I changed the [time stepping] to [strict] and let the mesh to be [physics-controlled mesh].

However, I don't really understand your 1st comment regarding several time step. I guess what you meant is I have to decrease the step, so I decreased it 10 times to be 0.001 ms rather than the original one 0.01 ms as in the attached image [time steps]. (The period of excitation is 0.1 ms). Still, the result was not consistent (phase delay between the largest (detected nearest the heat source) and the smallest thermal wave (furthest from the heat source) should be about 45 deg, but it wasn't as in the attached image [phase delay]).

I also attach the an update COMSOL file.

Please correct me if i am misunderstanding what you mean.

Hi Dave Greve Thank you very much for helping me. I changed the [time stepping] to [strict] and let the mesh to be [physics-controlled mesh]. However, I don't really understand your 1st comment regarding several time step. I guess what you meant is I have to decrease the step, so I decreased it 10 times to be 0.001 ms rather than the original one 0.01 ms as in the attached image [time steps]. (The period of excitation is 0.1 ms). Still, the result was not consistent (phase delay between the largest (detected nearest the heat source) and the smallest thermal wave (furthest from the heat source) should be about 45 deg, but it wasn't as in the attached image [phase delay]). I also attach the an update COMSOL file. Please correct me if i am misunderstanding what you mean.


Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jun 24, 2024, 11:51 p.m. EDT

I think I didn't notice your time units and I thought they were seconds. 0.1 period should be quite sufficient.

I am away from the computer and also I don't have any particular intuition about what your results should be. In your position I would look for any known analytic solutions.

I think I didn't notice your time units and I thought they were seconds. 0.1 period should be quite sufficient. I am away from the computer and also I don't have any particular intuition about what your results should be. In your position I would look for any known analytic solutions.

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jun 25, 2024, 7:16 p.m. EDT
Updated: 4 months ago Jun 25, 2024, 7:16 p.m. EDT

Thank you Dave Greve

I have an analytic solution for this problem. The solution and experimental results are consistent. However, I want to use COMSOL to simulate results for more complex models that analytic solution could also be too complex to solve. So I started with the simpliest model I have with COMSOL as I asked in the discussion.

Thank you Dave Greve I have an analytic solution for this problem. The solution and experimental results are consistent. However, I want to use COMSOL to simulate results for more complex models that analytic solution could also be too complex to solve. So I started with the simpliest model I have with COMSOL as I asked in the discussion.

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.