Discussion Forum

Apologies for the long post. The model is attached in the attachments.

A simple drawing of what I am trying to model is shown in attachment (1).

I am trying to simulate a type of Quantum Cascade Laser, I'm looking at the effect the temperature has by running a 10V voltage pulse through a terminal at one of the GaAs-copper boundaries where the 10V is applied vs ground at the under-side of the heat-sink alongside a repetition frequency of 10kHz.

The variable in order to produce the 10V voltage pulse is setup as follows: V_appl = V0*rect1((mod(t,t_repeat)-t_on)/t_pulse), where V0 is 10V.

With:

V0 = 10V
t_on = 1us
t_pulse = 2us
f_repeat = 10kHz
t_repeat = 1/f_repeat

I created a heat flux at each other surface to allow the heat to escape (convective losses). I realise I need a solid mechanics constraint also, but which one should I use? I used a fixed constraint on one boundary of my system (randomly chosen). When I try to compute a stationary study with a fine mesh I receive the error: "Relative error is greater than relative tolerance", or a value such as the one attached in attachment (2).

The geometry was adapted in order to make the model compute to the geometry in attachment (3), however this geometry is not optimal. This time, the temperature rise was far too small, being ~0.04K compare to the ~50K value I expected.

I realise the boundary conditions must be wrong, and the geometry setup is not optimal. However, I can't figure out the correct geometry or boundary conditions to use for my system.

Thanks for your help,

Jonathan.