1

I have a CW laser which is modulated by an optical chopper rotating at 1kHz. The modulated CW beam struck a sample holder and the absorption signal is recorder via an Infrared Associates MCT detector. The MCT detector, in turn, is connected to its own preamplifier, which in turn is connected via a BNC to a lock-in amplifier (SRS830 model). The optical chopper control provides a reference signal that is fed to the lock-in amplifier.

My data is collected via an NI USB 6210 DAQ (data acquisition) system.

The SRS 830 is able to provide instantly the magnitude and phase of the signal. However, I am absolutely lost on how to use this data to reproduce a vibrational spectra. The data I collect in the time domain corresponds to the absorption signal of the sample in the frequency domain.

I have been trying desperately to find literature on the subject of spectrometer building, but it seems I almost need a degree in physics to do this (I am a chemist, btw).

A depiction of the system I am trying to recreate can be shown here:

https://pubs.acs.org/cms/10.1021/ja200539d/asset/images/medium/ja-2011-00539d_0002.gif

A full description of the system can be found here:

https://pubs.acs.org/doi/10.1021/ja200539d

Can anyone provide me with some useful information and tips on spectrometer building, on how to understand signal processing from a IR detector and what should I do to get a spectra from the collected data I have?

Strelok
  • 111
  • 2
  • Pretty much, I am building a IR spectrometer. I am trying to replicate the instrument from this paper: https://pubmed.ncbi.nlm.nih.gov/21446655/ I am having problems with understanding how to generate an absorption spectra from the lock-in amplifier data. – Strelok Aug 07 '23 at 11:40
  • Essentially, this is what I am building: https://pubs.acs.org/cms/10.1021/ja200539d/asset/images/medium/ja-2011-00539d_0002.gif I just want to understand how to interpret the data obtained from the lock-in amplifier, and how to transform it into a spectra. Perhaps the fact that I am not able to phrase the question that seems "helpful to be helped" may indeed clarify that I am swimming a bit here. – Strelok Aug 07 '23 at 11:56
  • Thanks. I very much appreciate the first link and the newly added 2nd image. That's the kind of thing that stimulates discussion and movement in the right direction. Please ignore old memories of old times. Sorry about that. My apologies. I'll delete those comments as unhelpful. What you've added **is** helpful. I'd recommend adding them directly into your question rather than requiring others to read the comments below your question. (Just a thought.) – periblepsis Aug 07 '23 at 12:11
  • So this document: [Quantum-Cascade Laser-Based Vibrational Circular Dichroism](https://daylightsolutions.com/wp-content/uploads/2017/09/Lu%E2%95%A0%C3%AAdeke_JACS2011_QCL-VCD.pdf)? – periblepsis Aug 07 '23 at 12:22
  • Yes, that is indeed the correct article that is the basis of what I am currently trying to do. – Strelok Aug 07 '23 at 12:32
  • Perfect. This goes a long way in understanding what you are on about. Thanks. Do you have any details that deviate from the document? Or are you trying to directly replicate their results without adding anything new? Replication or something novel? Just noting that they state: *"Two lock-in amplifiers were used to extract the VCD and IR signals"*. Is that also where you are headed? – periblepsis Aug 07 '23 at 12:38
  • I´ll eventually get there. For now, I am just trying to get the IR signal. The only thing new that I added was the NI-DAQ thingy to the front output of the LIA´s, to make the data acquisition more streamlined. So, no photoelastic module is installed at the moment. The setup presently corresponds only to the IR-only part. – Strelok Aug 07 '23 at 12:58
  • 1
    Okay. Thanks. I think you've done a yeoman's job in responding well. I'll add +1 to the question. But please, where you feel appropriate, add as many details to the question itself as you feel like. Leaving them in comments requires others to read those comments. And that can be a bit of pain for some. Regardless, thanks so much for the dialog so far. I really appreciate your efforts so far. It matters. I'm heading to bed. So left for others, I suspect. – periblepsis Aug 07 '23 at 13:02
  • I'm not clear what you don't understand. You say "The data I collect in the time domain corresponds to the absorption signal of the sample in the frequency domain." So plot the frequency on the x-axis of a chart and the absorption (or transmission) on the y-axis, and you have your spectrum. If that isn't enough information, can you say more specifically where you get stuck? – The Photon Aug 07 '23 at 15:28
  • @ThePhoton, I am stuck in a couple of things: a) what is the meaningful data I can get from the Lock-in amplifier (LIA)? Do I only need the amplitude (R) value? Do I need also the phase? What do I do to the data collected from the LIA in order to get a spectra? Also, how do I record the data from a NI USB 6210 using Python? The nidaqmx documentation is not very helpful for someone not acquainted with LabView. – Strelok Aug 08 '23 at 11:28
  • You say "what should I do to get a spectra from the collected data I have?" but we can't answer this because you never told us what data you have. Please edit your question post to clarify exactly what question you want an answer to, and to provide the necessary background information to answer the question. If you have multiple questions (like how to use specific pieces of equipment) you can make multiple posts or consult the documentation for that equipment. – The Photon Aug 08 '23 at 14:36

1 Answers1

0

what should I do to get a spectra from the collected data I have?

First, spectra is the plural of spectrum. You can have several spectra, but if you only have one, it's a spectrum.

In this case it sounds like you specifically want an absorption spectrum. That is, you want a table or chart showing, for a range of optical wavelengths or frequencies (I'll assume wavelengths from here on) what the absorption coefficient of some sample is.

A depiction of the system I am trying to recreate can be shown here: ...

I don't have access to the full article, but I believe what you have here is a tunable external-cavity laser. When you adjust the PEM voltage, you adjust the optical length of the laser cavity, and the optical wavelength changes proportionally.

The SRS 830 is able to provide instantly the magnitude and phase of the signal.

There are two LIA's indicated in your system diagram so it's not clear what you mean by the SRS 830.

In any case, your goal is to figure out, based on the data you have, at each moment of time, what was the wavelength of the laser, and how much of the light was absorbed by the sample.

To get the wavelength of the laser, you need to know what voltage was applied to the PEM cell. Possibly you need to start by measuring the wavelength with a fixed PEM voltage using a spectrometer. Or you might use a reference sample with a known absorption spectrum and locate a well-known absorption line or two to find a couple of points on the V-λ curve.

To get the fraction of the light absorbed by the sample, you need to know how much light would be delivered in the absence of the sample. If the laser output varies when the wavelength is adjusted, you would want to do a measurement with the sample removed entirely (or if it's something like a gas cell, replaced with a vacuum cell to include the reflections off the cell walls). Then measure with the sample in place and compare the two measurements to determine the absorption coefficient.

Once you have the wavelength of the laser at each PEM voltage, and the absorption coefficient for each PEM voltage, you simply re-arrange your data to align one with the other, and plot the relationship absorption vs wavelength.

The Photon
  • 126,425
  • 3
  • 159
  • 304