Thanks for the valuable suggestions.

If you are not limited by volume of sample, the best way to measure liquids is in a larger volume (even a cuvette) and focus into the liquid with a x10 objective.

If you are limited by volume. then a short/tight focus is best, but you could also use a coverslip rather than a slide.

Let me summarize the answers and add some tips

A) If you want to work with small volumes

            Regardless experimental design you will have some contribution from substrate. As your signal is dominated by glass spectra, I can guess you are using 785 nm (quite common) or close to that excitation. So, there are two main ways to use: to optimize excitation and to optimize substrate. For substrate you can use metallic surface. For example, with 785 nm standard aluminum foil gives much weaker contribution. We used it successfully for wide range of organic solutions. Effective, simple and cheap method. Just take regular glass slide, wrap it using aluminum foil and deposit the drop. Take into account that small drops will evaporate – absolute and relative concentrations of constituents can/will change really fast. It is usually fine (for example, for comparative-analysis) if you use absolutely the same timing for all measurements. evaporation can be minimized by using metallic plate with small indentation. If it is not enough, you can also cover the drop using slides made of fused silica. Good quartz/glass is more expensive but not too much. Or you can place drop between two such slides. Another method – you can use capillaries. NMR capillaries can work here (take in to account they can be made from slightly different types of glass). Just fill it, fix on the microscope stage and focus inside. However, most probably, you will not eliminate substrate contribution completely. Direct substrate subtraction will be required.  If so, you can try to do it with spectra you already collected. Some samples can give relatively good own signal even if the raw spectra are totally dominated by substrate. Look at your spectra after substrate contribution removal. Keep in mind that you sample may have own fluorescent contribution to Raman spectra. You will not remove it with any substrate or cuvette design.

            A more radical solution is to change excitation. For example, excitation closer to 400 nm will reduce substantially even regular glass contribution. If you have this opportunity, you may try it to reduce substrate contribution and fluoresce of sample itself.

B) If you can afford to work with big volumes

            A cuvette made of good quartz and low magnification objective is a good choice. Advantage of a big volume is that you can reduce effect of sample degradation under laser beam. You can even use continuous agitation for highly sensitive samples or low concentration samples which required long accumulation times.  

Thanks Vitali for your valuable comments...Yes the excitation source is 785 nm. If I want to use cuvette then what is the arrangement of recording spectra using  microraman..

As I suggested before, use a deep vessel, such as a cuvette and focus from the top, through the meniscus, into the liquid with a x10 objective. Once the focus is inside in the liquid, you get maximum signal.

You may find that there is not enough room for a convention cuvette under the microscope, so what we often do is take a plastic cuvette and cut it in half to reduce the height.

Thanks Hugh for your valuable suggestion..However, there may be background spectra for plastic, it is  necessary to subtract, isn't it?

The depth of focus  - and therefore the volume from which 99.9% of the Raman signal comes from -  for the x10 objective is maximum a few millimetres, and therefore, it the focus is inside the liquid, but above the bottom of the cuvette, you get no signal from the cuvette.

Nothing to add here. Dr Byrne is right. You can use any small vial if you have enough sample.  Only if your solution should be isolated from air (highly flammable or reactive one), cuvette with stopper should be used. In this case try (as start point) standard quartz cuvette, fill it completely avoiding bubbles and place on microscope stage having working surface pointed toward objective (it is ok to place cuvette on side surface if it capped). There are micro cuvettes and special design cuvettes with thin walls. More expensive and shipping can take many days.

But even for volatile fluids try to record thought the meniscus first – to check if you have the potential to get good Raman signal.

sorry- this is my last pot on this subject....

"Improved Protocols for Vibrational spectroscopic analysis of body fluids",

Franck Bonnier, François Petitjean, Matthew J. Baker, Hugh J. Byrne,

Journal of Biophotonics, 7, 167-179 (2014)

Perhaps my answer is late ... But yes it is possible, you need to have some ability to adjust the focus and preferably use slides thinner microscope. The signal may show weaker and with a bit of noise, but it is possible. You can also use a quartz cuvette with thin walls used for UV-visible analysis if it has a larger amount of sample. If you can make a good focus will not need to subtract the background spectrum.

i am currently using a quartz cuvette for my spectrum. I am able to get a neat spectrum of toluene but unable to get a spectrum of my protein (hemoglobin). I use a 633nm laser. Please help

Is there anyone using a RomanStation instrument who used to be manufactured by a UK company but now part of PerkinElemer?

With best regards,

Yun-feng

Hugh J Byrne and Vitali Sikirzhytski if the concentration of my sample is too low like in micromolars, the mostly peaks of solvent comes, not of my solute. then how should I record the raman spectra??

If the spectrum of your solution is good, but the spectrum of the solute cannot be detected amongst the noise of your solvent signal, then;

Try to improve your signal to noise (e.g. longer accumulation)

Change the solvent if possible?

More likely, you have to go to higher concentration- if possible

Thank you for your valuable answer

Hii Aurelio Sanz Arranz .. yes I can do that but for control I need the Raman signal of low concentration of solute also.

@Gagandeep you should know your limit of detection by Raman.

Jp Wu yes, I know the limit of detection.

Hugh J Byrne Is it possible to record the time dependent SERS spectra in liquid.

Gagandeep Kaur - I have not done a huge amount with SERS, but my first thought are that it depends on what you mean by time dependent. The interaction of your analyte with the SERS substrate, be it nanoparticles or a 2D coating, is a dynamic process, and the adsorption may be to a greater or lesser extent irreversible. Depending on how strong your Raman signal is, and the sensitivity of your spectrometer, you could for example measure the build up of the SERS signal when you first add your analyte, on a time scale of ~seconds. If the analyte substrate interaction has come to an equilibrium, I imagine it then depends on what you want to measure in a time dependent fashion. If the analyte is tightly bound to the substrate, and the coverage of the substrate has saturated, the SERS may not identify changes to unbound molecules in the silution.

Hugh J Byrne thank you so much.