Week of 04/16/2018 - Progress Report 16

During the whole duration of this project, our main concern was stability of our system and getting rid of the fringes before taking data and analyzing it in the software that we wrote in MatLab. Even though the waveforms looked smooth on the spectrometer, when looked at in the MatLab analyzer, we would always deal with lots of noise, fringes and inconsistencies with distilled water and various glucose concentrations. We took a lot of data earlier in the week trying to maximize the power by adjusting the collimating setup, but the results were inconclusive. We would see similar trends, but nothing was 100% sure that we actually did it. However, after extensive conversations with Dr. Asghari, we decided that instead of trying to eliminate the fringes, we should use them to our advantage. By taking data at a different resolution, sensitivity and average waveforms than we have been in the previous months, we would maximize the power of the system, but also get a lot of fringes. This week’s settings were High 1, 1 Average waveform, 0.2nm resolution. However, fringes on the absorption spectra weren’t our concern anymore because we decided to write another code and to look at the Fourier Transform of the absorption spectra. This would give us to look at the reflection of the laser when it hit the glass, at the edges and right in the middle. The results from the code gave us a spectrum analysis and the Fourier Transform:

The spectrum was then transformed into the Fourier Transform and the results looked like the following:

When zoomed in at the peak in the middle, which should mean the middle of the cuvette, the result was the following:

In the top right corner, you can see the legend. The blue line, the highest one, is the Distilled Water one. This is good, as there is no glucose in that solution, and it should be the highest one. Next one is the 15mg/dL glucose, the one after that is 45mg/dL and the one after that is 330 mg/dL. This is a trend we were looking for!! And this is really good, basically proving that our project is possible to be integrated – which was the end goal of this research project.

Another proof of this would be to zoom in on one of the smaller peaks to the left or right of the middle one. The results were the same, indicating a repeating trend, which is good.

The waves further to the left/right are the reflection off the cuvette, and those shouldn’t have a power difference. Which they don’t, as can be seen from the picture below:

Another code was written to look at the refractive index of the cuvette. This code gave a time domain result, which at first glance looks like the Fourier transform graph. However, when zoomed in at the time domain one, we can still see the same trend with the DI water and the concentrations, exponentially dropping, shown below.

You can still see the legend and verify for yourself that the concentrations of glucose are indeed dropping, even in time domain.

As mentioned previously, there was also the refraction index proof. The plot generated for that looked like this:

And when zoomed in at the lower peak, it looked like this:

Plans for next week: keep taking lots and lots and lots of data and be able to replicate the data demonstrated in this report.