some question about the responsivity of the PD from the paper


Dear everyone
recently I have read a paper about the photodetector,which contains some results of the FDTD siulation.
I want to repeat the same results under the same condition with the FDTD but I failed.nextly I want to tell the difference about the result between the paper’s and mine.

  1. firstly the result I obtained of the top view image of the electric field of the one centered top contact and two off-centered top contact is different from the paper’s.

    2.In the paper,the results show top centered Al contact reduces the responsivity by 19%,but In my simulation the value is much lower than 19%,which is just about 7%.

    3.As the picture show below,the responsivity will be higher than 1A/W when the PD length is just 10um in the paper.In fact,I feel uncertain about the result of the paper from the result about my own simulation ,from which I got the value of the responsivity is just about 0.84A/W while the PD dimension is 8umX25um.

Can you give me some help and suggestion about the simulation of mine.I look forward to getting some comment from you and making the results of my simulation more accurate.
Thanks a lot.

one centered top contact.fsp (695.5 KB)
two off-centered top metal.fsp (696.5 KB)
no_metal.fsp (685.0 KB)


Hi @jbwei

I guess we need to break down this problem into smaller steps and hopefully we can replicate the results.

In your current setting it looks like the waveguide mode is truncated with PML layer. You will need to increase the FDTD simulation region to avoid evanescent fields interfering with PML layer. Having PML layer close to geometry might cause even severe problem when you use a metal as some plasmonic modes might interfere with PML layer.

Also, it will be a good idea to adjust the PML setting to absorb radiated light. This will help to lower simulation times as well.

Can you please do the modifications and let me know of the results?



dear bkhanaliloo
Thanks for your feedback.In your opinions,firstly I should increase the FDTD simulation region to avoid evanescent fields interfering with PML layer,which I understood.And you also said that I should adjust the PML setting to absorb ratiated light.I have confused about this.What do you mean and how to adjust the PML setting.Hope to get your feedback.Thanks.


Hi @jbwei

Sorry for the confusion.

I meant to say that after you adjusted your simulation, if you encountered divergence problem or if your PML doesn’t absorb light properly, you can employ PML and dispersive materials divergence testing. We want to make sure the results for no_metal.fsp file are reasonable before moving into simulating geometry with metal layer on top.



dear bkhanaliloo
Thanks for your feedback.I will run the simulation again under your guidance.If I have some results,I will share with you. Thanks again for your help.:slight_smile: