Page 1 of 1

Beginner in spectroscopy

Posted: Sun Jan 21, 2024 7:34 pm
by rama
Hello all,

I've been playing with the idea of getting a spectrometer for a long time.
I don't really need it, but I like to research things, I have scientist mind and it's a very interesting device.

Professional ones are insanely expensive, so DYI is probably the only option.
But there are also various commercial spectrometers that appear to be reasonably priced,
but it is not clear to me whether what is measured can be trusted and what they can measure actually.

Let me ask you a few questions:

1. What can spectrometry practically detect?
* Can it detect all possible substances or only some of them?
* Is it possible to find out the quantity of the measured substances and then reliably conclude something from it?
For example, to recognize quality, concentration, impurity, ...
* Can unwanted additives be detected? Pesticides in food, poor quality food,
quality of tea, coffee, substances in the soil, contamination of clothing from China, occurrence of PFAS/PFCA (polyfluoroalkyl/Perfluorinated carboxylic acids) in jackets, distinguish types of plastics, etc.
What can the spectrometry actually be used for, specifically raman?

2. How does it work?
I understand that a sample is irradiated with a laser and the scattering of the beam is captured by the spectrometer.
If I have a measured spectrum, what I need to do next? how can I interpret it.
Is there a translation table with frequencies and their elements? Or an app that would tell me what I measured?

I somehow understand the individual wavelengths represent elements or bonds between them
and their measured combination will then represent some substance or compound.
The intensity how much of the substance is in the sample - this is probably very difficult to calibrate.

Wouldn't it depend a lot on what place on the sample is measured - where the laser hits the sample?
Different places may contain various substances. If I put there apple for example.
I don't understand how it can seriously work. Except that only substances with 100% purity could be measured.

What if the sample contains several substances, how can you tell which ones they are from the spectrum -
when their spectra are mixed together?

Are there some freely available databases or applications that would interpret
what does spectrum actually mean? What elements/substances/compounds does the spectrum describe
in what concentrations, what kind of material is it, is it dangerous, etc.?
Is the interpretation easy or after the measurement I need to spent month researching what I actually measured.
What if I had wrong calibration - can it give me totally wrong result?

3. Is it safe?
I read that OpenRaman uses Laser 3b which is dangerous. It can damage eyes if it hits retina.
Do you use safety glasses during every measurement? Or it enough that the sample is covered with a cap
and there is a fuse to prevent the laser from starting with the cover open?

4. Are all spectrometers the same and only differ in which spectrum span they work?
or you can have a universal spectrometer with a wide spectrum for everything UV+VIS+NDIR
and then it only matters which laser is used for irradiation.
That is, is a raman spectrometer different from a normal spectrometer if it can sample the same spectrum?

5. What are the limits for minimum usable quality?
E.g. what is the minimum usable resolution in nm that is useful - so that it isn't a toy for children.
OpenRaman - is it a solid device, are there better ones?

6. What do you think about these cheap spectrometers?

thunderoptics

$150 - Mini USB Spectrometer
$440 - R-Spectrometer Research Grade
$450 - M-Spectrometer
$1149 - Mono Spectroscopy Kit
$3999 - Gem Raman System TO-GRS-532/TO-GRS-532-E

others
€2317 - High Resolution Spectrometer 200-1200nm
€1834 - Broadcom AFBR-S20M2VN Qmini USB Spectrometer

7. R-Spectrometer Research Grade
If I were to purchase R-Spectrometer Research Grade from ThunderOptics, could I then just buy a laser and shine it on the sample, or is it all more complicated, and I would need some lenses, calibration samples, etc.

8. What can OpenRaman do?
What can be practically measured with it. Is it only for very limited use in a chemical lab than for home substance exploration?

What are actually your use cases with spectrometers?

What I would like to do with a spectrometer device (not just raman, generally):
* detection of substances - I have a volatile substance in the bottle (alcohol, gasoline, diesel, alcohol, technical gasoline, etc.) and I would like to find out what kind of substance it is
* I have food (apple, pear, banana, ...) and I would like to find out if it is of good quality and does not contain pesticides
* I have clothes and I would like to find out if they contain dangerous substances
* I would like to distinguish between different types of plastic
* I would like to find out the composition of alloys (aluminum, steel, tin, if the solder is lead or lead-free, what it is made of, etc.)
* I would like to find out the composition of the water - if it does not contain dangerous substances for health
* I would like to recognize color, if my monitor emits blue light, how much etc.
* I would like to see spectrum of room light
* I would like to see spectrum of sun - if it doesn't contain too much of UV
etc.

...and if possible I apologize if my questions are too stupid, I don't fully understand this area but I am very interested.

Re: Beginner in spectroscopy

Posted: Mon Jan 22, 2024 10:40 am
by Luc
Hello rama,

That's a lot of questions -- I will try to answer them one by one. No questions are stupid btw and we're all here to help :)

Generally speaking, professional Raman spectrometers are indeed expensive -- reason for which I started the OpenRAMAN project. On the other hand, you should not neglect that there will be a gap between a $35,000 instrument and a $3,500 one! For highly demanding applications, such as very low dilutions, you will need a professional instrument. Nonetheless, educational ones can already offer competitive results for standard laboratory work.

1 & 8/ Typically, almost all pure organic compounds can be detected and many pure inorganic compounds can be detected as well. There are a few exception to that such as halides because their Raman peaks are too low for OpenRAMAN (below 400 cm-1). It's possible to detect these as well but you need more expensive filter set which will typically double the price of the instrument.

Mixed compounds can be analyzed as well. On the website, I show for instance how you can trace methanol in ethanol down to ~1% or how to track a esterification reaction with the reactants transforming into products. I usually call a "mix" something were the compounds are in a non-neglectable amounts (e.g. 10% or above).

Below that, you have the very diluted compounds. It's still possible to track them but it becomes more difficult. In controlled conditions, it's possible to detect down to ~100 ppm of nitrate or isopropanol in water. "Measuring down to" is really the limit and I would recommend staying 10x above that in real conditions.

Traces elements (<1 ppm) are very difficult to detect with Raman as-is. At that concentration you either need pre-concentration techniques (some even work in-situ, like ultrasonic focusing) or use more advanced techniques like antibodies fixed to metalic particles (typ. gold or silver) with a dye. I have no personal experience with these techniques however.

2/ In vibrational spectroscopy you have two regions in your spectra : one region where you can assign peaks to functions (-CH3, -C=C-, -C-OH etc.) and another region where it's usually more crowded and specific to the molecule under analysis. Even in the first region, assigning peaks to function requires experience (that I don't have btw) as peaks can shift depending on other proximity groups.

Another approach is to record spectra of pure compounds and make a regression analysis. That's how I tracked methanol in ethanol using a least square decomposition into the basis of pure ethanol and pure methanol spectra.

That's for homogeneous samples. For heterogeneous samples, like your apple, the spectra will change depending on the local composition. One advantage is that you can do 2D mapping of samples to have an image (a "data cube") of chemical composition at every point.

3/ While the system is generally safe, it's indeed using a Class 3b laser in its more powerful version. In Class 3b, the blink reflex of your eye is not enough anymore to prevent local damage on your retina if the beam hit your eye. I recommend basic laser safety training and wearing certified laser goggles. The dangerous part is mostly in the alignment phase when you might put your hand into the beam. The risk is to get unwanted reflection on jewlery or watches which can redict a lot of energy in your eye... Now, think about all the people using 20W infrared lasers for DIY engraving without any enclosure! That's a million time more dangerous.

4/ You indeed have different type of spectroscopy and the wavelength range change depending on the type. A Raman spectrometer is just a regular spectrometer but with a limited bandwidth (typically ~100 nm) with very good light collection capabilities and a highly sensitive detector. Raman is a very weak effect and every photon is important.

I'll soon start a series on UV/Vis and IR spectroscopy so I invite you to stay tuned for explanations on these later.

5-7/ The total effective resolution, laser included, shall be better than 1 nm. In OpenRAMAN, it can go to ~0.25 nm. The most difficult step is not to have a high resolution spectromter (that's easy actually) but one that is efficient in light sensitivity. Then you must pair it with a very good laser as most cheap laser will limit your resolution to 1-2 nm.

If you're serious about Raman spectroscopy, I always recommend Wasatch spectrometers. I have never worked with the ones of ThunderOptics. At OceanOptics, you should avoid entry level (~$2000) spectrometers because you won't be able to measure Raman at all with these -- I discussed this with some folks from Liege University who tried the USB2000 and didn't get any signal at all even for strong Raman emitters.