Published: 2019-05-18 | Categories: [»] Engineering, [»] Opticsand[»] Chemistry.

One of the improvement that was planned for the [»] DIY Raman spectrometer was the correction of the astigmatism in the initial cuvette holder.

The astigmatism in the initial holder originates from the cylindrical shape of the test tube which acts like a cylindrical lens. When the laser is focused, the vertical and horizontal rays have different focus positions as represented in Figure 1. In practice, the light is focalised into a large volume which a waist that lies in-between the two focal plane and was about 180 µm large (RMS spot size). By the way, I am using here test tubes because glass can easily adsorb fluorescent compounds that quickly ruin your spectra and that they are much cheaper to replace than quartz cuvette (0.02€/pcs vs. 30€+).

Figure 1 – Astigmatism in initial cuvette

Astigmatism is a problem for Raman spectroscopy because the molecules in the vial scatter light back in all directions which results in a large blurry spot imaged onto the spectrometer slit. This can be directly confirmed by the experimental images where the slit is visible as ~50×4 pixels square image instead of a small disk. About 90% of the scattered light collected by the optics is lost that way in the original spectrometer setup.

An improvement of the setup would be to remove the astigmatism in the cuvette focusing mechanism such as to provide a tighter spot on the slit and so a higher yield of photons actually injected into the spectrometer. This is exactly what this upgrade is about.

To compensate the astigmatism of the test tube, I inserted a negative cylindrical lens between the vial and the doublet lens. The distances were adjusted to yield the tightest possible spot. My optical design software predicted a 9 µm RMS spot size which would result in a 9-fold improvement of the signal intensity at best since the slit integrates the signal vertically.

To hold everything I designed a new 3D-printed cuvette holder. I also took the occasion to add further improvements such as a foldable cap, the possibility to add an interlock to switch off the laser when the cap is open and a beam-dump behind the sample. The new cuvette is shown in Figure 2 and can be downloaded [∞] here. It is mandatory to print them using SLS technology. I paid 41.35€ for the printed parts to which you will need to add a few bucks for the helicoils, set screws and so on.

Figure 2 – New cuvette holder

The lenses are placed on mechanical tolerance and no adjustment is foreseen. They are held in place by M4 nylon-tap screws. The cage rods are fixed using M3 DIN913 set screws. A DIN7 3x50 m6 pin provides the axis for the cap folding mechanism.

The laser was stirred using a Rhodamine solution as done previously and the performances were compared with the initial cuvette holder using ethanol. Exposure time could be either reduced from 1 sec to 250 ms or the gain decreased by 13 dB. Both results suggest a 4/4.5-fold improvement in signal. Obviously, one will always want to use the longest exposure time such as to reduce the noise effect that appears in high gain mode.

This is about half of the theoretical improvement predicted by the optical design software and can probably be explained by several factors such as the beam quality (M² factor) of the laser, the diffraction effects or the fact that the re-imaging lens is currently placed with the wrong orientation due to mechanical constraints.

A spectrum of ethanol is given in Figure 3.

Figure 3 – Ethanol spectrum, 4.5 mW 1 min. exposure

The noise was estimated from the 1800-2000 cm-1 part of the spectrum and the SNR of the peak near 1500 cm-1 computed as 1:1035. The FWHM of the peak near 900 cm-1 is 20 cm-1 and the three peaks at 3000 cm-1 distant by 46 and 51 cm-1 are clearly defined. The spectrometer range was not changed by this upgrade and still covers the 500-3600 cm-1 region.

Concerning the other upgrades of the spectrometer setup, I am still working on the laser. I was actually expecting to release the laser upgrade first and work on the cuvette last but I’ve had a minor set back with my laser driver which actually fries the lasers instantaneously instead of driving them nicely. I’m still expecting to release the upgrade by the end of the summer but I will probably first start using a laboratory laser driver and then try to do the cost reduction process when final resolution is validated with the lab driver. This way of doing will allow me to validate the breadboard and freeze the specifications once and for all. Unfortunately, the lab driver is going to cost me a lot of money and I am therefore thinking about opening a patreon page or something similar… I’ll see in the following weeks the best way to give you an opportunity to participate into the development of the setup :)

Stay tuned for updates! Big stuff are coming!

[⇈] Top of Page

You may also like:

[»] Raman Imaging Lens Upgrade

[»] Confocal Spectroscopy

[»] Slit Distortion

[»] Complete 400-800 nm Spectrometer Design from A to Z

[»] Predicting Thorlabs Achromats RMS Spots Size