In Raman spectroscopy the experimental challenge is to be able to measure a very weak signal (usually several orders of magnitude weaker than the laser intensity) which is also very close (in wavelength) to the laser wavelength. It is best to start with a sample with known strong Raman signals when building a new setup. I have not worked with gases/liquids, but you could try Benzene which has a strong signal at ~992cm-1 shift. With a 532nm laser this Raman signal will arise

at ~561.7nm. While building our own setup for solid samples (in back scatter geometry) we used a Si sample for tests, where the signal was at ~520cm-1 shift (~547nm with a 532nm laser).

Since a Raman signal is very sharp you should scan with a step size 1/5 of the band pass setting of your monochromator. Also a CCD instead of a PMT+Lockin, is a much better (much faster, equally if not more sensitive, safe from Det damage) option for Raman measurements.

Do you try to align directly to an (unknown) spectrum with the notch filter in place? This is very difficult.

However, the Raman signal is spectrally very close to the pump laser, it is a useful practice to align to the reflected/Rayleigh scattered laser light, which is much much more intense. This makes it in particular more easy to find the point, when you actually couple light into the detector path, i.e. the fiber. Depending on laser power you can do this visually with your eye by following the green light on a buisness card. You might need to lower the pump light intensity to protect your eyes or detectors. Although, you can often align with laser glasses protecting from green light, as the green light often causes weak (yellowish) flourescence on a buisnesscard, which is visible through glasses.

What is the diameter of your fiber core? Aligning even to a 100 or 200 µm core is challenging, espescially, if one has no experience in this task.

Be aware, not to burn the buisnesscard. Focussing a few 10 mW can already burn the card, which can cause dust on your optics. Further, be sure not to blind your detector.

Further: Do you focus your excitation beam? or are you trying to collect from the whole sample volume?

Once, you have maximized the laser light, you can insert the filter again and try to see if you get a spectrum and optimize on that. Again, be aware that Raman is very very weak, especially from gases or certain liquids. Try to find a strong scatterer for the alignment.

At the start remove monochromator to detect integral Raman signal and after reliable monitoring this signal return monochro mator back tuning it on on red shift line near 600-1000 cm-1 from 532nm or use standard Raman substance like C6H6 .

Sandip Ghosh Eugene A. Tikhonov Michael Rüsing Thank you very much for the reply. I am using benzene as the sample since it has strong raman signal. The sample cell is about 10cm long and has a window of 1 inch diameter at right angle. I've set the monochromator to ~561.8nm and in order to detect the signal I have connected the PMT output to an oscilloscope. I have focused the excitation beam at the center of the sample cell (close to the window). The fiber core diameter is 400 µm. It is my first time working with optical alignment. We had the monochromator and PMT in our lab and so we wanted to work with what we have and make use of it.

One of the issues I can think of is that the scattered light may not be focused onto the optic fiber core. I can visually see yellowish scattering when I look through the notch filter. Is there any way I can direct the scattered light onto the optic fiber with high efficiency?

As suggested above, I will try aligning the optics to get the Rayleigh scattering and use that setup to get the raman signal. Any additions suggestions ?

Again, I am appreciative of all the time and effort.

I am not sure if I understand what you are doing on the detection end. A PMT is a current source and directly connecting it to a Oscilloscope is not the best way to look for a weak signal. I thought you mentioned a Lockin earlier ? A PMT can be run in a 1) Geiger mode (this is for very low light levels, needs high PMT bias voltage and typically uses a 50ohm terminator on the PMT + a fast voltage amplifier + Photon Counter) or in 2) Proportional mode (this is for moderate to low light levels, needs lesser PMT bias voltage and typically uses a current to voltage amplifier (CVA) + a Lock-in amplifier, periodic laser excitation with ideally 50% duty cycle). The current input of a typical Lock-in can be used when not using a CVA.

So you should look for the weak Raman signal on a Photon Counter or a Lock-in. Also dont expect to get a signal just sitting at ~561.8nm, because wavelength calibration error in a Monochromator (MC) can result in the signal arising at a different wavelength. You should do a scan around ~561.8nm with step size 1/5 of the MC bandpass.

Before this ensure proper optical alignment as suggested by Eugene A. Tikhonov and Michael Rüsing.

Dear Rawunus Hserus , in our discussion nothing was told about intensity of excitation. Pay attention that if intensity is more than about 50MW/cm2 along laser beam direction could be excited stimulated Raman scattering (case with C6H6). And note that You operate with the nanosecond laser pulse duration and your system of signal digitation also has to be enough quick to treat nanosecond signal or even microsecond after PMT.

Greeting Eugene A. Tikhonov Sandip Ghosh ,

After re-aligning the optics I was finally able to get the raman signal. In addition to the re-alignment I also increased the beam energy which made the signal even stronger. Sandip Ghosh my apologies, the the oscilloscope is connected through PMT + lock-in. I also, took the MC calibration error and scanned a wider ranger around 561.8nm. I was able to get the max signal wavelength.

Eugene A. Tikhonov I had a confusion, is it the direction of the scattered beam or the energy of the excitation beam, or both, that determines whether it is stimulated of spontaneous RS? If I'm detecting the scattered signals at right angles to the excitation beam is it still spontaneous RS? Right now I am using around ~25 J/pulse at 10Hz excitation beam in order to detect the signal spontaneousRS at right angles.

Once again thank you for the suggestions.

-Hserus.