Choosing the proper laser for Raman spectroscopy can be a difficult task. You want one that is performant enough but without having to pay the high price for it. Another difficulty adds up when recommending specific laser brands because you also want the lasers to have reproducible performances from user to user. Cheap e-bay lasers are not an option in that regards because they are a lottery: very few ones will have correct performances but most of them will be crappy or even dangerous to operate.
I already investigated several approaches with the OpenRAMAN project: from the very cheap [∞] Thorlabs CPS532 all-in-one module to the high-end 3k$+ [∞] laser kit. I also proposed a [∞] custom laser and TEC driver to lower the price of the high-end kit by half. More recently, I proposed a [∞] fiber-coupled version for people already owning fiber-coupled lasers in their lab. And the work is not over because Im now investigating tuning options to increase the performance of cheap lasers!
Here, we investigate an off-the-shelve solution from [∞] CNI Laser which is a good fit between the expensive 3k$+ kit and the cheaper custom driver versions.
The first thing we should ask is what is a good laser for Raman spectroscopy? Assuming you already selected the laser wavelength for your setup, the driving factor Im immediately looking at is the linewidth of the laser. Lasers rarely consist of a unique emission peak unless you spend a lot on them and they will come with some broadness or number of longitudinal modes. In a [»] previous post, I already described how temperature and current can influence the broadness of your emission spectrum I strongly encourage you to read it now if you already havent.
Every bit of wavelength of the laser will participate in the Raman effect so you directly get a convolution of the Raman native signal of the molecule you study with the broadness of the laser emission spectrum. Assuming a Gaussian distribution (this is an approximation but it also give satisfactory results for [»] square distributions) with a FWHM of 0.15 nm, we expect at 532 nm to broaden the Raman peaks by about 5 cm-1. In practice the resolution will be affected by the size of the slit and the resolution of the spectrometer as well but if you target 5 cm-1 for all of them (in wavelength equivalents), you can expect an overall ~√3 factor so about 9 cm-1 which is the order of magnitude I obtain with the OpenRAMAN performance edition.
When placing the order for my CNI MGL-III-532-50mW I specifically asked to have a linewidth of 0.15 nm maximum. When the laser arrived, I measured the longitudinal modes distribution using my [»] high-performance spectroscopy setup. The results are shown in Figure 1.
Although the result complies with my request, Im not 100% sure they understood the specification correctly because in the QC document they mentioned a 15 pm linewidth. So, if you choose to order a laser yourself, I would instead recommend to ask for FWHM < 0.15 nm just to be sure you are talking about the same thing.
The performances translated directly to excellent peak widths in actual Raman spectra. Using a 15 µm slit, I obtained a peak width of 11.5 cm-1 for iso-propanol (major peak located at 800 cm-1 - uncalibrated). The results are shown in Figure 2.
Now that we have covered the first important parameter, lets talk about the power of the laser.
More emission power means more photons. More photons mean more signal on the camera and so higher signal-to-noise ratio and dynamic range at fixed exposure time. This is particularly true for weak Raman emitters or when using a narrow slit like we do here.
At some point however, more power is unnecessary. In some samples, fluorescence will prevent usage of high exposure time and in other samples its the Raman emission of the solvent (or matrix) that you are using that will limit the maximum exposure you can use. Also, remember that more power also means that the laser is potentially more dangerous to operate (for both the sample and your eyes always use laser safety goggles!), that it will be more expensive and that it may be difficult to get the requested FWHM which is the first parameter you should be looking at in Raman spectroscopy.
Here I asked for 50 mW and got about 75 mW output (remember that the driving request is the FWHM so youll get the power at which your laser diode gives the best performance in terms of longitudinal modes). At 75 mW I already saturate my camera at its maximum exposure time (32 seconds) with the Raman spectrum of water. I wouldnt therefore recommend more than 50 mW for the OpenRAMAN setup. Its also good to know that if you ask for more than 50 mW, the price of CNI will increase. Specific usages may require more power (e.g. faster acquisition for real-time monitoring) but thats a decision you have to take.
As an order of magnitude, at 1 second exposure I get plenty of signal for iso-propanol with a dynamic range of more than 10,000 for the main peak. You can see in Figure 3 that the spectrum is very clean despite the low exposure time used.
Now is the question: how much does the CNI laser cost? Here, I have to break down a bit the prices and what I paid.
The laser itself, at 50 mW, cost $1300 which is just a touch more expensive than the DJ532 version with the custom LD and TEC driver. To that, you however have to add shipment costs ($100 for me) and bank charges ($25). When placing the payment, I also paid 35 bank charges and USD/EUR conversion charges on my side. Added to that, I was finally charged 21% VAT import taxes by the custom plus a fixed fee by the transporter. In total, I paid a bit less than 1700 for the laser which puts it half-way to the Thorlabs full-laser kit. That is still a couple hundred euros less than if I went to a local distributor of CNI in Europe but at the expense of a lot of trouble to get the bank proceed to the international payment! Big thanks to CNI who was extremely comprehensive with the delays of my bank as the payment was cancelled twice by the bank and it took me 1 month and 2h of phone calls to get it done. Dont overlook these aspects as well and check with your bank before doing anything.
CNI also offers single-mode fiber coupled version for $2200 and super-narrow fiber-coupled versions for a bit more than 3k$ which should be both compatible with our [∞] fiber-coupled add-on. At that price, you may however want to consider upgrading to the singlemode version MGL-U-532 which can be a better investment for your lab. I wouldnt recommend their solution for a 100 µm fiber because Im relatively confident that much of the laser light will be lost.
Which laser you pick is ultimately a trade-off in your own requirements but I definitively keep the MGL-III-532-50mW as a viable option for the OpenRAMAN spectrometer, especially for people wanting performances but without the need to fine-tune lasers or solder PCB themselves.
Last-but-not-least, I have been working with CNI laser for a few years now and they are a trusted actor in the laser world. I never had any QC issues with them and they are always very helpful when you contact them. I would therefore strongly recommend them for your laser supply! Please note that this post is not sponsored by CNI so its just my experience working with them.
Finally, there is currently no modified baseplate to use the CNI laser. One will come in the near future (when I can find some time!) but until then you can simply fix the standard baseplate on a 300×300 mm breadboard like shown in Figure 4. While not optimal, it is stable enough to be used for experimentation.
I would like to give a big thanks to Naif, Young, Samuel, Eric, James, Lilith, Hitesh, Sebastian, Jesse, Sivaraman, Andrew, Jon, Cory, Karel, Alex, Tayyab, Stephen, Marcel and videohero who have supported this post through [∞] Patreon. I also 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![⇈] Top of Page
You may also like: