Sciencemadness Discussion Board - Spectrometry, how to extend performance.

Spectrometry, how to extend performance.

semiconductive - 4-2-2022 at 19:14

HI All,
I built my first spectrometer ... and overlooked something important.
The fiber obtic cable I bought is only good in the visible to near infrared region, 300 - 1100nm. So, when I started sampling, I realized that the longer wavelength data is attenuated and useless.

I'd rather not redesign the spectrometer, so I'm trying to figure out if there are techniques I can use to infer deeper infra-red spectrum information out from near-infrared measurements.

There's several articles on the web about identifying 'functional' groups in chemicals from the near infrared region (800-2500nm). I only have about a quarter of that region, but I can see some of the spectra talked about in articles using a 1cm cuvette a silicon light detector, and common chemicals ( water, alcohols, ethers, etc. ) So, I'm encouraged!

That's when I got to thinking about overtones and overtone detection.
Every atom when it changes the dipole moment of a molecule, will absorb a particular mid infrared wavelength. It also tends to absorb at frequency multiples of that wavelength, which 'smears' out in the visible and near infrared region.

So, I was wondering .... maybe I could get a mid infrared light source, for example, and build some kind of simple filter or prism in order to make it so I can select a narrow region of IR light to shine into the sample .... would exciting a bond in a liquid also cause a slight effect on the visible spectrum the sample absorbs?

I'm thinking the change in the visible spectrum will be due to harmonics (fluorescence?), or perhaps thermal heating of the liquid, excitation of the molecule, etc.

So, if I took a absorbance spectrum with and without the mid-IR light source turned on, I could 'detect' the bonds reacting to the mid-IR light?

Anyone have any practical suggestions on how to go about this type of data analysis; and how I might most easily test the theory (using easy to get / low toxicity chemicals.)?

semiconductive - 4-2-2022 at 19:46

Some other ideas I have are changing the sample's termperature, applying a capactivie electric field of a 1000V or so across the sample vial -- which would tend to align dipoles in a certain direction. Or perhaps Radio frequency energy might be used to excite the molecules in a period fashion especially in the presence of a strong magnetic field.

There also might be some simple chemicals that are sensitive to the presence of a particular bond, like pH strips change color with ion concentration of hydroxide and hydrogen ions.

Each of these modifications might allow visible-nir spectroscopy to gather data about molecular bonds that are not directly detectable by the spectrometer; but I'm not sure what I should try first .... and what things will allow the most sensitive (wide range) measurements of molecular properties to be extracted.

neptunium - 4-2-2022 at 23:20

Do you have any pics to show your equipment and progress? which spectrometer and software are you using ?

Rainwater - 5-2-2022 at 03:17

Quote: Originally posted by semiconductive

The fiber obtic cable I bought is only good in the visible to near infrared region, 300 - 1100nm.

Remove the filter. Most of the prefab cables, I install, have a high pass filter located on one end of the cable.
If welding a cable termination, one must be installed for proper data transmission.

The filter is a thin flat coating on the end of the cable. It can be ground off with a polishing tool.
Its not really a visible component. Its installed by adding a drop of solution to the tip of the glass then polishing until dry.

JohnnyBuckminster - 5-2-2022 at 03:54

“.... maybe I could get a mid infrared light source, for example, and build some kind of simple filter or prism in order to make it so I can select a narrow region of IR light to shine into the sample .... would exciting a bond in a liquid also cause a slight effect on the visible spectrum the sample absorbs?”

You are looking at different things, a carbonyl group absorbs in the IR around 1700 cm-1 (~0.2 eV) due to bond-stretching, but a transition in the visible region, say 500 nm (~2.5 eV) is due to an electronic transition, electrons are shuffled around in the molecule. If you have an electronic transition, you most likely also will induce vibrational motions, but you are not likely to induce an electronic transition by illuminating in the IR region.

semiconductive - 9-2-2022 at 12:43

Quote: Originally posted by neptunium

Do you have any pics to show your equipment and progress? which spectrometer and software are you using ?

Sure, I'll take pics.
The optic cable and cuvette holder are standard, Ocean Insight brand. I didn't have the equipment to make a fiber cable -- which is why I just bought two of them. One is a cuvette, the other is a reflectence probe.

As far as softare ... I'm just using gnuplot, and open source stuff on linux.
I've ordered a Jaz spectrometer to do a comparison against my spectrometer with.
They should be very comparable, and here in a few days.

There's freeware for the O.I. hardware, but it's very limited. Their flagship software, is pretty good (I haven't installed it yet) -- but I've watched the how to videos. It doesn't look like it's powerful enough for me to do the kind of analysis I want to do ... so I'm probably going to have to write some python software to do the job.
Unfortunately, I don't have working USB -- but only TCP/IP -- and can't use the open software driver from ocean insight. I'm not sure how to work around that, yet.
I'm not supposed to reverse engineer their software, but I'm not sure if I'll be lucky and they use the same protocol with TCP/IP that they do with USB, just encapsulated. (Here's hoping.)

The only difference is that I took pyroelectric detectors out of motion sensors for automatic outdoor lights and cut the blocking window out of it so the chip is exposed. That way, I can scan the sensor across the spectrum landing in order to get a sample rather than just get a snapshot all at once. So, the PIR dector version is very slow. But it hooks up to the same SMA connectors as a Jaz does.. Unfortunately, it doesn't get any deeper IR scans than the silicon sensor does with the O.I. cable. I can detect deeper IR with the cable removed -- but then I can't get light through the cuvette.

I'm probably not going to photo that much, as I'm considering selling them if I can figure out a way to adapt the fiber optic assembly from ocean insight to give me deeper infrared samples. I'd be happy to buy their product and modify it, if it would give me the wider IR range.

I'm also looking carefully whether I can buy a cheap laser pointer, and a dichroic mirror, in order to make a termperature controlled laser source for a doppler shift spectrometer. It looks like there's mode hopping problems with cheap diodes, but that might be controllable if I thermally regulate the diode carefully.

semiconductive - 9-2-2022 at 12:57

Doppler ... for those of in Chemistry and not physics ... it's called Raman spectrocopy. Sorry, I have a bad habit of using words from different disciplines.

semiconductive - 9-2-2022 at 13:00

Quote: Originally posted by Rainwater

Quote: Originally posted by semiconductive

The fiber obtic cable I bought is only good in the visible to near infrared region, 300 - 1100nm.

Hmm. That makes sense. I can probably get some 1 micron diomond lapping paper ... but I don't have a lapping machine to pollish with. How sensitive is sanding to light transmission of the fiber ... if the end of the cable is a little rough, it might still work with a diffraction grating ?

semiconductive - 9-2-2022 at 13:13

Quote: Originally posted by JohnnyBuckminster

“.... maybe I could get a mid infrared light source, for example, and build some kind of simple filter or prism in order to make it so I can select a narrow region of IR light to shine into the sample .... would exciting a bond in a liquid also cause a slight effect on the visible spectrum the sample absorbs?”

You are looking at different things, a carbonyl group absorbs in the IR around 1700 cm-1 (~0.2 eV) due to bond-stretching, but a transition in the visible region, say 500 nm (~2.5 eV) is due to an electronic transition, electrons are shuffled around in the molecule. If you have an electronic transition, you most likely also will induce vibrational motions, but you are not likely to induce an electronic transition by illuminating in the IR region.

Because of E=hf
500nm / ( 0.2/2.5 ) = 6250nm. So, that's still in the detection range of a PIR (Pyroelectric Infrared Detector ... who go down to 10 microns. eg: 10,000nm ).

Yes, I get that they are going to be different effects. But, the liklihood of seeing a transition depends on things like temperature. So, if I shine a light that is absorbed by some molecules and not others ... it should have the effect of warming up certain molecules preferentially to others. eg: a gradient will form.

So ... why wouldn't shining light of the color to excite the bond stretching at 6250nm ... cause the molecule temperature to change in the vicinity of of molecules with carbonyl bonds?

I would think absorbing energy at the carbonyl range would cause the molecule's liklihood to shift in the transisitions of all bonds, electronic and otherwise, ... by some miniuscule amount.

Do you think I'm on the wrong track with that Idea , and why/why not?

[Edited on 9-2-2022 by semiconductive]

semiconductive - 9-2-2022 at 15:31

IN the photo:

Down on the table with the two orange fiber optic leads is the cuvette holder with an acrylic 1CM cuvette in it (empty). Off to the side, with the slightly blue liquid is a cuvette full of Ether. This is eucalyptis globulus, that I attempted to purify. It should be slightly polar, and so I hope it has deep IR bonds that can be stretched.

It was purified by taking KOH and dissolving it in methanol, and waiting until the solution becomes saturated. THen it was mixed with the eucalyptis globulus from ebay, "GreenHealth"(tm) brand.

The result, after shaking, is that the Eucalyptis will become very turbid and brownish. Most of the organic matter that is more polar than the Eucalyptis, will react with the KOH and settle to the bottom. See the separatory funnel in the background.

After letting it sit overnight, almost all the remaing KOH and methanol will settle to the bottom with crystals (probably carbonate) forming. Unfortunately, I originally did this in the blue bottle that the Eucalyptis comes in. Even though KOH and methanol don't attack the bottle ... when Eichalyptis is in there, it starts to dissolve the blue pigment from the bottle. Each day the Eucalyptis sits, more of the blue color crystallizes out and the liquid becomes clearer and less turbid. It's only a tiny amount of blue ... less than 0.1% ... but it's annoying.

The strong odor of Eucalyptis is reduced this way, which suggests that most of the odor is from an alcohol or other product dissolved in the Ether.

https://pubchem.ncbi.nlm.nih.gov/compound/Eucalyptol

I'm not sure why there's an 'ol' ending on the chemical as if it's an alcohol. From what I'm able to see, it's an ether with an oxygen off to one side. It's not as reactive as an alcohol and potassihm hydroxide will barely dissove in it, and will settle out after a while. Usually, in alcohols, KOH dissolves and stays in solution happily.

C10H18O.

So, I figured I'd try to use this as it's cheap, safe, doesn't evaporate fast, and should have at least two absorbtion bands in the visible near infrared region. I could use some help calculating at what frequencies I should see optical effects happening, as I'm not very good at calculating bond energies myself. I know how to do AMU masses for Oxygen, and other parts of the molecule, and calculate resonances using a weak spring model; But, I'm not sure what spring constant to use with the oxygen to carbon bonds to model this molecule.

The stuff in test tubes, is just running a variation of an electroplating experiment in an alcohol ether mix. Dimethyl ether seems to be pretty good at conducting ions, and Eucalyptis is much less active. I have vanilla, and other ethers around as well.

semiconductive - 24-2-2022 at 22:38

I got my Jaz, today. So, I'll begin experiments .... shortly. At first, I think I'm just going to save results onto an SD card or USB stick and transfer them to my computer to analyze.

Fun!