Optical Signal Simulation in CADENCE

I am currently studying various pixel structures and would like to simulate some parameters of the image sensor. I have a photodiode model in CADENCE however I do not know how to simulate the optical input. Is there some ready-made source in CADENCE or some other way around it? I would put a current source in series with a capacitor however this approach will be very inaccurate with respect to using the photodiode itself so I would rather know if an optical input is possible.

Any response will be greatly appreciated.
Thanks in advance,
Alessandro

You can simulate the photocurrent by a simple parallel current
source across the collecting junction; it's calibration that's
the main problem here. Much of low-light sensitivity is about
the "waste" terms - recombination, photons that overtravel the
collection regions, fill factor and so on. You probably need to
pull data or find data for responsivity in an example structure.
How you'd determine the effects of your variations of layout,
I don't know. I've seen more than a few pixel design papers
and they tend to be "cut and try and report", without a lot of
theory that could be applied to another technology (how do
you expect to get surface state density, surface recombination
velocity and bulk trap density for a new-to-you process flow
or pixel layout, let alone map that to a flux:current model?).

You might find some photodiode models or scholarly papers
about them, out in the wild. But how applicable they will be
to your present effort, other than pointing the way, is to be
determined.

Thanks for the response, I'm not sure whether I understood you completely but all the parameters you mentioned have to do with the photo-diode model right? That I managed to obtain one which should serve my purposes (have to design an image sensor and see what type of pixel is most viable with respect to noise, dynamic range, QE etc..). The problem I am having is that this photodiode has an optical input which I can not figure out how to give using CADENCE, I only see voltage sources, current sources etc.. I am a complete beginner with CADENCE so I do not if such an optical source exists. Otherwise as you mentioned I have to use a high level model with a current source for now just to analyse operation and noise introduced by the external circuitry of the pixel.

I don't think Spectre can do more than simulate electrical
behaviors (although you might consider veriloga models to
implement other things, which have an electrical output
and call the code guts a "black box").

Quantum efficiency is an outcome and the sum of many
actors (reflection losses, absorbtion efficiency, recombination
losses, avalanche gain and its statistical fog, etc.) and I
suspect that Cadence may be the wrong tool for the job.
For designing the readout IC, or "everything but the
photodiode" in a CMOS image sensor, it's a well suited IC
design tool. But deep semiconductor electro-optical physics,
that's not really their thing. So you have to feed it the first
electrical stimulus in the chain and go from there. And the
difficulty I see is, you don't have a good tool set (and data)
identified that would let you analyze all of the what-if
photodiode degrees of freedom. Not sure if there are even
sensible tools for that, and if there were, you'd need to
populate them with relevant device physics type details.

If you can constrain the problem to "electrical in, electrical
out" then the Cadence tools would be capable enough. It
just doesn't sound like this is the plan. But maybe you
renegotiate the plan so that success is more likely (like,
assume some photonic:electrical transfer function and
move forward with the front end electronics design).

For the photodiode alone, you might be able to obtain
a transfer function of optical power input, to current
output, at some wavelength(s) of interest and bias
voltage and temperature and so on.