I am trying to do stimulated Raman scattering microscopy. In this experiment two pulsed laser beams are used . The wavelength of these two laser beams is tuned to match a specific Raman frequency of the system under investigation. When this condition is fulfilled one of the laser beams loses a slight amount of energy and one of the laser beams gain a slight amount of energy. There is no other wavelength involved. The change in the intensity is the signal itself. This is known as stimulated Raman gain or stimulated Raman loss. Now The laser light itself is very strong and compared to that these fluctuations in the intensity are very very small and no detector can sense that. To overcome these problem , one of the laser beams is modulated at a particular frequency (using a Acousto optical or electro optical modulator). This means that the relative raman loss or Raman gain will also be modulated and you can use a lock in amplifier to detect those fluctuations.

Since, you are detecting the laser light directly , PMT and APD are automatically out of the contention. Usual standard protocol and state of the art right now is that people take a large area Si photodiode (peak wavelength somehwere in the range of near IR) and drive that to a large reverse bias (60-70V) and feed that output to a lock in amplifier. Currently the photodiode that people around the world use is thorlabs FDS 1010. The vendor rating is 25 V but research group using this use home made circuits apply 70V of reverse bias. So far I have not been able to find a commercial solution and am looking for building one myself.

Lastly, sensitive enough means under those above mentioned conditions the change in intensity / laser intensity < 10^-6 should be detectable when lock in amplifier is used.

Our wavelength of interest : 1031 nm , lock in amp is SR 865a, BNC or SMA connections.

I have a bare photodiode FDS1010, THORLABS. I need to convert this into a full fledged detection system equipped with variable reverse bias, BNC output for signal etc. What will be the best place to start with (in terms of literature, lessons etc). For instance, one confusion I have is that whether a reverse bias of 50 V means that the point marked as the reverse bias in the circuit diagram has to be connected to a external Dc source at 50 Volts?

  • \$\begingroup\$ You didn't mention modulator frequency. That big diode has 375pf @ 5V. Who knows how much less capacitance it has at 50V? In any case, connecting to your lock-in should be done with a short, low-capacitance cable. \$\endgroup\$
    – glen_geek
    Feb 1, 2019 at 21:26
  • \$\begingroup\$ we are modulating one of the beams at 1.7MHz \$\endgroup\$
    – user188062
    Feb 1, 2019 at 22:20

2 Answers 2


I'll answer your actual question. Yes, if you want a bias of 50V then you need to connect your 50V source as shown in the diagram for the "reverse bias". Note however that the maximum reverse bias for this diode is 25V so connecting 50V would be a very bad idea.

  • \$\begingroup\$ Agreed. This is why I contacted the authors of this protocol and got a one liner in response "that way too conservative value". However thorlabs have denied any help if maximum specs are overstepped. \$\endgroup\$
    – user188062
    Feb 1, 2019 at 20:41

I see this as an XY question where assumptions may be invalid for optimal Raman wave spectra, resolution and efficiency.

It is known that;

  • light energy is proportional to frequency or inverse to wavelength.
  • silicon detectors are broadband but proportional to wavelength up to 1050 nm then decline rapidly.
  • RAMAN efficiency is reported to be proportional to \$\zeta _r=\dfrac{k}{\lambda ^4} \$
  • shorter wavelengths give smaller resolution
  • therefore the optimal efficiency using a silicon diode is 525~550 nm (green)
  • other emitters and detectors such as green lasers and green LED arrays as detectors (InGaAs) or EMCCD arrays may be more effective with RGB for up to 600 frames per second instead of 1 second.

enter image description here

Here I used a k=1e-10


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.