search

dschick / udkm1DsimML

A MATLAB Simulation Toolkit for 1D Ultrafast Dynamics in Condensed Matter
Other
7 stars 4 forks source link

Unwanted Pulse echos from the end of the substrate #5

Open AvonReppert opened 6 years ago

AvonReppert commented 6 years ago

udkm1Dsim is capable of simulating strainwaves in samples that a launched after ultrfast laser-heating. Such coherent phonon oscillations can for example be observed in ultrafast X-ray diffraction experiments and as their precise shape provides relevant information on the layer thickness and laser penetration depths.

However when one wants to "quickly optimize" the sample structure by matching the simulated strain pattern to the observed experiments one always needs simulate a substrate that is so thick that the initial shockwaves generated at the sample-air interface are reflected back from the end of the substrate into the thin film.

The longer the simulation runs the thicker the substrate thus needs to be, which makes the simulation slow. Thus it would be very useful if one could set the boundary condition at the end of the substrate such that no strain-waves are reflected. Maybe one could introduce a layer at the end that has per definition zero strain in each timestep and thus works as black hole for sound waves.

Of course this should be an optional feature, which from my perspective would increase the simulation speed since it reduces the substrate thickness that needs to be taken into account.

dschick commented 6 years ago

This would be a nice feature, but I fear there is physically no way of realizing this! Since one cannot change the way the MATLAB ODE solver works, one has to define the correct boundary conditions in order to implement such a perfect absorber. I am not aware of a way of how to implement such kind of perfect sound absorber, do you?

AvonReppert commented 6 years ago

Couldn't we by default include 10 layers of the substrate material at the end with a phonon damping that is close to infinity? This would be more a workaround maybe or do you see any problems with that?

dschick commented 6 years ago

the problem with that is: a very high damping also increases the acoustical impedance leading again to bad matching and high reflections at this interaface. One needs a layer with the same acoustical impedance AND with high damping. I fear this is not possible.