4

I would like to see if the idea I have for an experiment is even worth the effort and money to perform. I would like to use electrical impedance tomography to image a developed thin layer chromatography plate with a silica gel stationary phase. The operating principle here is that the spots of interest on the silica gel containing any organic analyte on the plate could have a different electrical response to an applied current as compared to the tomogram acquired of the plate prior to development. I am a chemist by education and training and having little electrical experience I need some outside advice before I spend any more time or money on this idea.

My main questions around the feasibility are:

  1. Would the high resistance of the silica gel (10^15 ohm cm per American Elements) prevent any electrical measurements?
  2. Assuming 1 is not a non-starter, how could I determine the correct system specifications? I.e. voltage of the signal generator, minimum current needed, etc. (I am happy to read any literature myself- which literature/chapters would be useful).
  3. Would having the electrodes arranged in a rectangle as opposed to a circle around the edges of a typical 7.5 x 2.5cm plate pose an issue for image reconstruction algorithms?

If this is a hopeless endeavor then I would like to know before I commit any more time to it besides thoughts and reading.

Dallon Penney
  • 41
  • 3
  • 1
    How thin is the layer of silica? Will the plate be tested in air or vacuum? It seems like you might need a test voltage so high it will ionize air. It will probably be difficult to connect the electrical source to your experiment with commercially available wire insulation. – Theodore Mar 31 '23 at 21:27
  • For analytical thin layer chromatography, the layer can be between 0.1-0.25mm thick. I would rather not have to image in vacuum, but I could seal the plate off from any air by placing silica side down on top of a flat surface with the electrodes built-in and sealing off the top. Fabricating something like this wouldn't be a big deal. – Dallon Penney Mar 31 '23 at 22:05
  • I thought at first glance that _ohm cm per American Elements_ is a crazy mixture of imperial and SI units. – D Duck Mar 31 '23 at 22:38
  • 1
    I have read a paper that showed a very thin layer of silica can be used as a relative humidity sensor, with the sensor's impedance measured as a function of relative humidity using an LCR meter. They showed that applying a 100 kHz AC wave at 1 V, the resistance dropped significantly from the nominal at 24x10^6 ohm. Link to the paper here: [link](https://www.sciencedirect.com/science/article/pii/S0925400504007270) – Dallon Penney Mar 31 '23 at 22:39
  • What kind of geometry would that look like? Perhaps a grid of electrodes like a capacitive touch screen, but "touching" the TLC plate instead? Yeah, I suppose there might be some differences, especially if you can measure impedance vs. frequency, but it might be pretty slight in many cases, except when a highly polar or ionic molecule is present. Maybe that's useful enough to extend the analytical range of other methods (UV, staining, etc.)? – Tim Williams Mar 31 '23 at 23:06
  • @TimWilliams I was thinking 16 electrodes arranged evenly around the edges of the 2.5 by 7.5cm TLC plate, which is alluded to in my question about if that arrangement could even work. The resolution of EIT is about 15% the distance between electrodes, so that would put the resolution at about 2 millimeters. Considering the dielectric relaxation of silica I am leaning towards it being possible if the frequency of AC current applied is high enough. – Dallon Penney Mar 31 '23 at 23:23
  • Ah yeah, tomography huh... I know very little about it algorithmically speaking, but that sounds reasonable. But you get so very little cross-section edge-on: the dielectric constant isn't terribly high (uh, I don't have a figure, but I would guess 4 or less for pure/dry/desorbed material, up to... 100s maybe but only when water-soaked?), it's thin and wide, so most of the electric field is in space around rather than across the plate, and the signal is very weak indeed. Not impossible, but a challenge to be sure. – Tim Williams Mar 31 '23 at 23:33
  • Resistivity seems even more a challenge, as it's basically an insulator to start with, and any moisture or other ionic materials will affect that profoundly. How well does tomography deal with dynamic range? If a point is 10^6 times more conductive, or 10^9 or 10^12, than the base material, how does that affect the measurement (accuracy and resolution)? – Tim Williams Mar 31 '23 at 23:35
  • Just for sake of thought experiment -- a more localized field, like the toroidal "fringe" field at the end of a coaxial section (outer ring = GND, inner dot = signal, annulus = insulator), might be swept over the plate, plotting impedance (capacitance and resistance) at a given frequency, or at multiple frequencies; some of which will correlate with DC resistance, but interesting things may be found about the relaxation rate of adsorbents by doing this multi-spectral. I'm not sure how well that kind of scheme translates to tomography, though. – Tim Williams Mar 31 '23 at 23:38
  • @TimWilliams the moisture would definitely have a large effect on the resistivity of the silica. The thought there is what if prior to imaging the plate was immersed in water so that the results can be the same roughly each time irrespective of air humidity? If non-polar organic molecules are adsorbed to the plate I don't think they would be affected by being in contact with water. – Dallon Penney Apr 01 '23 at 00:11
  • 1
    @TimWilliams biomedical EIT uses frequency response impedance to produce a signal, which removes the need for a "background" image to subtract any changes from, as different tissues have different frequency responses. The paper that I linked earlier showed that the resistance is dominated by relative humidity at low frequencies but showed little change in response to relative humidity at high frequency, at least at the scale shown in the paper. – Dallon Penney Apr 01 '23 at 00:20

0 Answers0