Solid Powder Raman Spectroscopy

It has been already three years now that I published the first results on [»] Raman spectroscopy. I then released the [»] OpenRAMAN Starter Edition one year after and the Performance Edition followed the year after that. All these (breadboard, starter and performance editions) were however only operating on liquids. This year I am glad to push the boundary of affordable Raman spectroscopy even further with the analysis of solid samples!

In this post, I will show some experimental results obtained on solids. All the CAD files required to reproduce this experiment are available on the [∞] companion website as usual. Don’t miss also our alignment video for the solid cuvette (here-below) which complements the already available [∞] assembly & alignment tutorial video for the starter edition.

All the data have been acquired using the Starter Edition. I unfortunately fried my laser of the Performance Edition and I don’t have the money right now to replace it… Also, I spotted some extra peaks in some spectra (e.g. paracetamol) which I cannot explain yet. I am conducting additional experiments to understand where this effect is coming from as it may reduce the spectral resolution.

A picture of the device required to measure solid samples is given in Figure 1. It consists of a right-angle mirror with adjustable orientation, a 19 mm doublet achromat to focalize the beam on the sample, a tray to place the sample on and a micrometre adjuster for adjust the focus on the sample surface.

The powder samples are placed in a small aluminium cup that can be bought from ChemLab Analytical (ex-Fiers if you are familiar with our standard suppliers). These cups are solid by 250 pcs and are guaranteed to have no residue that could contaminate your sample. The cup itself is hold in a 3D holder which can be machined as well in plastic like POM. The holder is mounted on Thorlabs tray using a glued SM05RR retainer ring. The cups shall ideally be replaced for every measurement although, like for the test tubes in the liquid device, you can reuse them about 10 times or so if you don’t mind a bit of contamination. The 3D holder itself can be used many more times as, in theory, it never comes in contact with the chemicals. In practice I see that it tends to stain and should also be replaced at the occasion. A schematic of the holder is shown in Figure 2.

Despite the Starter Edition limited laser power of 4.5 mW, I had no issue with signal strength. All spectra were acquired with exposure time between 5 and 30 seconds depending on the chemical sensitivity to Raman emission.

I was able to measure successfully 14 different chemicals I had on hand. Only EDTA and Zinc Chloride showed no signal. Concerning Zinc Chloride, I suppose it is because the chloride ion is too low in the wavenumber region and is clipped by Thorlabs dichroic mirror which is clearly not optimized for Raman spectroscopy. I have put all the spectra in .spc and .csv format in file that you can download [∞] here.

Among these, I was able to correlate four of them with the [∞] SDBS database : benzoic acid, paracetamol, salicylic acid and urea. The spectra are given in Figure 3 to Figure 6.

Generally speaking, we observe a slight shift of the wavenumbers of about 50 cm^-1 which can be explained by an incorrect laser frequency or an invalid wavelength calibration. Also, no baseline removal was performed. All the large peaks were identified and most of the medium/small peaks could be observed as well. The resolution of the Starter Edition is not sufficient to discriminate all nearby peaks however.

Concerning the resolution, I was able to evaluate it to 34 cm^-1 using the narrow 1000 cm^-1 peaks of urea, as shown in Figure 7. This is the very same resolution that was obtained with the nitrobenzene peak around the same wavelength with the liquid cuvette. This is also confirmed by the data of Figure 3 as we can detect the two peaks at 1604 and 1634 cm^-1. After subtracting a linear baseline from the 850-910 cm^-1 region, the signal to noise ratio of the peak in Figure 7 was evaluated to 730:1 (single acquisition – no averaging). I did not keep track of the exposure time used but I believe it was about 5 seconds here.

I must say I’m quite satisfied with the results obtained here but I should also comment on a few aspects that were (or still are) more challenging:

Alignment is not as easy as with the liquid samples. In the liquid sample device, once aligned, you can basically throw any liquid and it will work. Here, I noticed that some realignment is sometimes necessary. I guess this might be coming from the orientation of the crystals in the powder that influence the amount of signal received. I initially had some difficulties to align the system when using the Performance Edition which might be due to the smaller slit size although I don’t have enough information yet to conclude.

Spatial resolution is also different from the one I was expecting. The next planned step was to acquire 2D spectra by scanning XY the sample but I noticed that powders (at least some of them) show very strong subsurface scattering and I don’t know exactly how this is going to affect the spatial resolution of 2D spectra. Basically, light enters white solid powder and diffuse inside the powder very far from the initial emission spot. Instead of measuring a tiny spot of Raman activity, you average over a much larger volume of the sample.

Some peaks are doubled – or even tripled! I have no idea where this is coming from as it does not happen with all peaks. I will investigate both the laser for multiple modes and eventual reflections in the system. I cannot exclude that this may be coming from the sample itself because I essentially noticed it with paracetamol which was coming from painkiller pill. All others chemicals used were certified chemicals and I don’t notice anything suspicious in their spectra based on Figure 3 to Figure 6.

The mechanics itself is not 100% as I would like it to be. In my device, the lens cage is not sliding smoothly on the rods due to a slight bending introduced by incorrect tolerance in the right-angle mirror. As a result, the micrometre driver itself is not enough and you have to use force of your fingers to move the cage plate. I don’t know if all right-angle mirror holder have the same problem or if it just mine but you may need to enlarge slightly the holes of the lens cage plate to let it move more freely. Don’t overbore them because it might also block if tilted! Maybe some bushing might be required… but then I would replace the whole cage plate by a custom 3D part as well.

Apart from that, the system operates quite well and give really satisfactory results even at low exposure times. The device itself cost about 495€ but requires a larger initial investment due to the fact that the cups are sold by 250 pcs. Users who are on a budget can try to replace these by parafilm although it will not give any guarantee that it will not contaminate your sample (this is most probably ok for lab with students, though). Similarly, it is possible to drop the micrometre part of the assembly (including the custom 3D parts) and adjust focus only with your fingers. I would not recommend this latter solution however as it will make alignment much more difficult to achieve.

The most important thing for me, at the moment, is the question of alignment and the peak doubling observed in paracetamol. I will however have to wait to have enough money to buy a new narrow linewidth laser since I fried the current one. Upgrades such as 2D scanning and rework of the design is not foreseen at short term but may happen in the future.

Don’t hesitate to [∞] contact me if you would like to propose future work direction for OpenRAMAN! I’m especially keen to focus first on anything that can promote teaching of chemistry and a more widespread use of OpenRAMAN as an aid for students. I’m also looking for more financial support to develop new applications so if your employer, company or university, would like to help us they are more than welcome!

I would like to give a big thanks to James, Lilith, Samuel, Themulticaster, Sivaraman, Vaclav, Arif and Jesse who have supported this post through [∞] Patreon. I remind you to take the occasion to invite you to donate through Patreon, even as little as $1. I cannot stress it more, you can really help me to post more content and make more experiments! The device presented in this post was paid 100% through the money collected on Patreon!