OpenRAMAN Spectrometer Efficiency

I recently had the occasion to measure the efficiency of the [»] spectrometer part of the OpenRAMAN setup using a calibrated powermeter (Thorlabs PM100D) at the office and I’m therefore taking the occasion to create a short post on the subject.

The spectrometer I tested was actually an adapted version of OpenRAMAN designed to work at slightly higher wavelengths, in the red region. The grating and lenses were optimized for the new wavelength ranges as well as the angle of the grating that is more pronounced than on OpenRAMAN. I experimentally measured an efficiency of 45% from the slit to the camera sensor plane on that modified setup. This includes the grating efficiency and the optic transmission. Although the setup that I tested is not exactly the same, I’m expecting that OpenRAMAN should have a very similar efficiency as I only made small changes between the two setups.

The measurement itself is relatively simple and you can easily reproduce it using an amplified photodiode such as the one [»] I presented on this website. The results will not be calibrated in mW but as you only need a ratio between input and output power, it should be fine enough. Also, since we are expecting efficiency on the order of 50%, linearity of the response of the amplifier circuit is not an issue as long as you use the same amplification settings on the driver circuit. You will have to put the photodiode just behind the slit (input power) and then at the sensor plane of the camera (output power - you will have to remove the camera from the objective). Efficiency will be given by the ratio between output and input power.

Concerning the light source, you will have to use a laser, preferably coupled to a small-core fiber, as you will need the source to be both powerful and monochromatic. Monochromatic light is preferred because if you use a broadband source it may spread beyond the diameter of the measurement probe and you will lose light. The small-core fiber is to avoid losing too much light on the slit as it is relatively narrow (about 10-15 µm in our setup). At the office I used an expensive red laser that felt just in the measurement range but we don’t have that kind of possibility here. Eventually you could use a red HeNe (632.8 nm) laser if you have one available but it is not common in an amateur lab. What you can still do is to use your green excitation laser (532 nm) and tilt the grating to slightly alter the wavelength range of the instrument. Ideally, the laser should be relatively centred on the camera. Note that it is extremely important that you remove the camera when doing this as exposing the camera sensor to focalized laser light may cause irreversible damage to the sensor!

One last thing that you should pay attention to is to reduce the effective numerical aperture of the input beam. If you simply connect your fiber to the spectrometer, you will measure a very low efficiency that is actually biased by your coupling efficiency from a very large-NA fiber (typically from 0.22 to 0.5) to your moderate-NA spectrometer (0.1 in our setup). Coupling efficiency will decrease as the square of the ratio between the spectrometer NA and the fiber NA. One way to avoid this is to use a 4f relay system with an aperture stop designed to reduce the numerical aperture well below the spectrometer acceptance angle. For example, if you use two 50 mm lenses for the relay system, you will have to use an iris of less than 10 mm diameter (half of that to be safe). I recommend performing the experiment at various decreasing apertures until the values settles within a few percent. This is exactly what I did at the office btw :)

This is all for today!

I would like to give a big thanks to James who has supported this post through Patreon. If you’d like to support me too and get credits as well as many other stuff, consider [∞] donating :) You can make the difference!