ED-XRF spectra - recurrant artifact below 0 keV?

  • Context: Graduate 
  • Thread starter Thread starter zappacake
  • Start date Start date
  • Tags Tags
    artifact Spectra
Click For Summary

Discussion Overview

The discussion revolves around the phenomenon of a recurring peak observed below the 0 keV mark in XRF spectra graphs. Participants explore potential explanations for this peak, including its relation to electronic noise and detector behavior, while addressing the complexities of measuring photon energy and interpreting the resulting data.

Discussion Character

  • Exploratory
  • Technical explanation
  • Debate/contested

Main Points Raised

  • One participant describes a high intensity peak below 0 keV in XRF spectra, questioning its nature beyond vague descriptions of electronic noise.
  • Another participant inquires about the measurement of photon energy and the possibility of negative values in the context of detector output.
  • A participant explains the process of photon detection and charge conversion, expressing skepticism about the interpretation of negative values in this context.
  • Discussion includes the type of detector used (Si(Li)) and the non-linear calibration that could lead to misinterpretations of energy readings.
  • Fluctuations in background signals are noted, with some suggesting that these could manifest as both positive and negative signals, potentially leading to artifacts.
  • One participant proposes that the observed events may not be related to significant energy photons, attributing them instead to electronic noise from the detector.
  • Another participant agrees that the noise is likely from the detector itself and suggests testing the detector's response by covering it with lead.

Areas of Agreement / Disagreement

Participants express differing views on the origin of the 0 keV peak, with some attributing it to electronic noise and others questioning the validity of negative energy readings. The discussion remains unresolved, with multiple competing explanations presented.

Contextual Notes

Participants mention limitations in understanding the calibration of the detector and the influence of thermal noise, indicating that these factors may contribute to the observed phenomena without reaching a consensus on their implications.

zappacake
Messages
4
Reaction score
0
Hello,

I can't seem to find a comprehensive explanation for this recurring phenomenon on my XRF spectra graphs. It's a high intensity peak that occurs seemingly below the 0 keV energy mark.

And it seem to be common to many graphs, yet nowhere seems to say what it actually is, aside from a rather vague 'electronic noise peak' or 'pulses', neither of which make sense, as noise is created via the fluorescence of materials in the X-ray tube, in the vicinity of the sample and in the detector, and other interactions in the sample. Noise creates actual transitions peaks as it results from the fluorecsence of actual elements - just not the sample elements!

So what the hell IS that damn 0 keV peak?!
 
Physics news on Phys.org
How do you measure the photon energy? Which measurement where could be interpreted as negative value?
 
When photons enter the detector unit of the XRF instrument, they are converted to a series of electrical charges until the photon absorption occurs. The charge are moved to the sides of the detector, in turn creating a deltaV that can be amplified. In this process, detector output voltage moves from background levels, to some maximum peak voltage, before returning to its original level, creating a pulse. A higher energy photon leads to more chargecharges, and creates a larger overall voltage pulse in the detector.
None of this seems to me as if it could be interpreted as a negative value in any way. I've managed to find some amptek manuels that cites series and parallel electronic noise from input capacitance and thermal leakage current, respectively. But it's says this 'noise corner' occupies the lower energy end of the resulting spectrum, rather than the...minus end! But I suppose I should dig a bit deeper with the above, when time allows...
 
What kind of detector? Silicon drift diode (SDD), Si(Li), Ge or something like this?

The energy calibration is not completely linear. It is quite possible that an integrated charge (peak height) of zero does not correspond to zero keV. Thermal noise would create tiny peaks with almost zero charge, and the non-linear calibration would then give a fake energy below zero. In any case, try to adjust the threshold for minimum peak height, that should solve the problem.
 
Fluctuations in your background could occur in both directions - more electrons or less than average. These fluctuations could be interpreted as small signals, positive and negative. Your software should remove these artefacts.
 
Thanks M Quack and mfb.

Yes, it's an Si(Li) detector. It does seem plausible that it's something to do with thermal leakage noise from the X-ray tube.

Mfb - what do you mean? If we consider only bremsstrahlung radiation, then surely the resulting spectral continuum curve (I love using the word 'continuum' in a sentence...!) results directly from the non-characteristic photons that reach the detector, rather than on an averaged value. This will be a constant presence that will fluctuate of course, but how would that result in a negative signal? I'm still very much in the process of learning about the instrument, in practical terms, and yeah, further work with the software would help immensely, but I'm afraid my time with the machine is up! I got some very clear spectra from mixed metal samples, but comparing glass samples was very much more difficult. I suspect that this is a direct result of lower atomic number elements, lower energy k-alpha transitions, and thus less intense peaks due to reduced probability of characteristic x-rays being able to escape the surface of the sample. Thanks for your help!
Lee.
 
I think that your events are not related to any photons (of significant energy) at all.
"Electronic noise" is noise in your detector, due to thermal noise or other effects.

Consider this situation:

hipmtchk.jpg


While you get clean signals from high-energetic photons, your readout has some noise, which gives small positive and negative fluctuations, compared to the average.
 
I agree with mfb. this is most probably noise from the detector itself. Try to cover the detector with a piece of lead and start anxacquisition. Is the ghost peak still there?

Usually such noise is screened out in the analog electronics, before the pulse height digitizer. That avoids saturating the counting chain.
 

Similar threads

Mathematica This Week's Finds in Mathematical Physics (Week 228)
  • · Replies 2 ·
Replies
2
Views
3K
Mathematica This Week's Finds in Mathematical Physics (Week 228)
  • · Replies 2 ·
Replies
2
Views
4K