1
$\begingroup$

I am sorry if this question is too vague, but what is peak identification? I am working with a chemistry professor to program an autosampler and he says he would like to "program the robot for a peak identification feature that he would use on his project." Can someone please explain in detail what peak identification is?

Many thanks in advance for any help

$\endgroup$
7
  • 1
    $\begingroup$ One can discuss this topic if it is related to analytical chemistry but you have to add more details to your question. What are you doing, what type of measurements are you making. Above all, did you ask your professor if are not sure about it, what does he want and what is that "project"? $\endgroup$
    – AChem
    Jul 11, 2020 at 5:10
  • 1
    $\begingroup$ Autosampler (in analytical chem work) has nothing to do with peak identification. $\endgroup$
    – AChem
    Jul 11, 2020 at 5:11
  • 1
    $\begingroup$ Sorry, I don't know what you are doing. You have to describe your work and the nature of measurements so that others can suggest a better tag. Yes peak detection is indeed part of analytical chemistry. $\endgroup$
    – AChem
    Jul 11, 2020 at 5:13
  • 1
    $\begingroup$ Haversine, You really have to ask your professor because he can tell you his requirements. I really have no idea what are you doing and you seem to be unclear as well. My friend has a very good website on all you would ever need to know on signal processing. Here is the link to Peak Finding. terpconnect.umd.edu/~toh/spectrum/PeakFindingandMeasurement.htm $\endgroup$
    – AChem
    Jul 11, 2020 at 5:23
  • 2
    $\begingroup$ Haversine should at least know and tell us if the peak is obtained in mass spectrometery or in gas chromatography or in another technique. Without this information, we cannot help him. $\endgroup$
    – Maurice
    Jul 11, 2020 at 9:42

1 Answer 1

5
$\begingroup$

Autosampler - is a piece of equipment that aspirates the sample from a vial and introduces it into a chromatographer or a detector. It has nothing to do with peaks or their identification - that comes later.

After the sample is injected into the detector and the measurements are taken you get back data describing physical properties of the sample. If you inject directly into the detector you can plot it as signal intensity vs. physical property, here's an example of spectra from Mass Spectrometer: enter image description here You may need to find peaks on such graph in order I don't know.. to identify how many impurities there are.

But often before introducing the sample directly to the detector chemists would run chromatography - it'd separate different components of the mixture and make them elute at different time. Now we'd measure physical properties not once - but do it at some rate. If we measure (and sum up or average) the signal at different time we may get peaks when the compound is eluted from chromatographer: enter image description here Chemists would then have to define the boundaries of the peaks and integrate their area in order to find how many substances there are in the mixture, in some cases calculate how much of each there are. Software can automate some of this work - identify the peak boundaries for you. This may be what your professor is asking for. Or maybe he wants you to write algorithms determining what substance is represented by this or that peak.

Note that this problem is very challenging and not universally solved - finding peaks can be problematic because:

  • different chromatograms will have different amount of high-frequency noise
  • there could be a baseline drift (low-frequency noise)
  • peaks often are skewed
  • compounds can elute very closely to each other, so you get merged peaks (rider peaks, shoulders, round peaks)

Due to these reasons you may not always be able to find the peaks reliably e.g. enter image description here So for autodetection you may want to go over several algorithms. And the person who uses them may have to understand and tune the parameters for it to work successfully on a particular chromatogram.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.