Detection of Bacteria on Metal Microarrays

Shaun Williams, PhD

Research Goals

Determine how far the detection limit could be pushed

how small of a bacterial number density can be detected.

Record absorption spectra for a variety of bacteria

Begin building a bacteria spectral library

First Mesh of the Summer

Transmission spectrum of a mesh with and without M. luteus.

Transmission spectrum of a mesh with and without M. luteus zoomed in on the region from 500 - 1200 wavenumbers.

Reproducibility Test

A Big Problem

Comparison of bare mesh spectra.

Same original mesh dimensions
Same deposition time

Time was measures with wristwatch
Accurate to at best a half-second

Better control of deposition time is clearly needed

A Big Problem

Transmission spectrum of a mesh with and without M. luteus.

Same original mesh dimensions
Same deposition time

Time was measures with wristwatch
Accurate to at best a half-second

Better control of deposition time is clearly needed

A Big Problem

Transmission spectrum of a mesh with and without M. luteus zoomed in on the region from 500 - 1200 wavenumbers.

Same original mesh dimensions
Same deposition time

Time was measures with wristwatch
Accurate to at best a half-second

Better control of deposition time is clearly needed

Design of a Timer

Designed electronic timer control.

Arduino electronic control board (bottom in image)

Hand coded
Time accuracy of 1 millisecond

Control breadboard (middle in image)
Power controller (top in image)

Program triggers this on an off as appropriate

Comparison of Mesh with New Timer

Comparison of bare mesh spectra with which the new timer system was used.

Both deposited for 6 seconds (6000 ms)
How can this difference be explained?

How Big is this Difference

This difference isn't as big as it seems compared to the original nickel mesh
The overall dominant peaks at \(\sim 750\,\mathrm{cm}^{-1}\) has a %T of ~70% in the original nickel mesh.

This amounts to percent changes of 99.1% and 99.9% respectively in %T.

The dominate peak at \(\sim 1450\,\mathrm{cm}^{-1}\) has a %T of ~7% in the original nickel mesh.

This amounts to percent changes of 84% and 90% in %T.

A spectrum of the original nickel mesh. The percent open area of the original nickel mesh is 26.2%.

Comparison of bare mesh spectra with which the new timer system was used.

Explanation of the Differences

Let's compare the surface area of the mesh in these two case.

Green spectra: \(0.0195\,\mathrm{in}^2 = 0.126\,\mathrm{cm}^2\)
Purple spectra: \(0.0272\,\mathrm{in}^2=0.175\,\mathrm{cm}^2\)

Comparison of bare mesh spectra with which the new timer system was used.

The distance between the anode (the mesh) and the cathode is approximately \(0.102\,\mathrm{in}=2.59\,\mathrm{mm}\) during deposition and the potential difference between the two is \(10.0\,\mathrm{V}\).
This yields an electric field strength between the anode and cathode of \(3860\,\bfrac{\mathrm{V}}{\mathrm{m}}\) or \(3.86\,\bfrac{\mathrm{kV}}{\mathrm{m}}\).
Hypothesis: Larger surface area means that the maximum electric field strength is in a larger volume of the solution. Therefore, more deposition.

First Bacteria Run with Mesh Created with New Timer

First controlled mesh spectrum bare and with M. luteus.

First controlled mesh spectrum bare and with M. luteus zoomed in on the region from 500 - 1200 wavenumbers.

Spectrum of Micrococcus luteus

Extracted absorption spectrum of M. luteus.

A program was written that allows for the automated removal of the background spectrum
The program eliminated the need to extract the background point-by-point, spectrum-by-spectrum.

Future Work

Further improve the transmission-to-absorption program to further eliminate artifacts.
Push the detection limits to see how small an amount of bacteria can be detected.
Analyze as many bacteria as possible.