Week of 10/23/2017 - Progress Report 3

Topics covered: Calibrating the system tested from last week, testing using both 1550 nm NIR LED and the 980nm pulse laser in the Faraday Cage.

Materials used: Cuvettes, pipettes, beakers, 100mL graduated cylinder, powder glucose, stirring rods, DI water, scale, optical breadboard, a NIR LED 1550 nm, a InGaAs photo diode 1550nm, Agilent Voltmeter, Peak Meter Digital Multimeter, Tektronix Osciloscope, conducting wires, duct tape, 980nm pulse laser.

The objective for this week was to calibrate the system which was implemented last week, and to perform testing using the calibrated system in the Faraday cage. What needed to be confirmed is the fact that using the 1550nm wavelength and the Near-Infrared LED with a InGaAs photodiode is better than using a 980nm pulse laser with a SI-Photodiode of 350-1100nm wavelength.

Calibration meant setting up the optical system and making it permanent for testing. Last week, the system was tested in South Hall lab, and to ensure some accuracy, I used my hands to cover the system in order to shade it from the light. After having a conversation with our project advisor, dr. Asghari, I realized that the whole optical system needs to be set up in a way that would be sturdy and hard to move. Also, extra carefulness was applied to everything this week, to ensure the maximum possible precision of the results.

This was done by first, attaching the optical breadboard used to the heavy, metal optical table and stabilizing it with duct tape in multiple locations, and also conducting wires, which were found in the lab. The 3D printed cuvette stand was also stabilized with duct tape and by putting conducting wires all around it, making sure it wouldn’t move. How the optical set up looked this week is shown in the picture below. 5 Volts of power was applied by using the Agilent voltmeter.

Since last week’s solutions were just sitting in a lab for over 4 days, new solutions were made. The same process outlined in Progress Report 2 were followed for making the solutions. The same concentrations of glucose were used, except this week, I was even more careful with measuring glucose and adding the DI water to it. The glucose measurement for this week are as follows:

• 50 mg/dL – 54.2 mg

• 80 mg/dL – 88.6 mg

• 160 mg/dL – 161.6 mg

• 240 mg/dL – 248.8 mg

• 2000 mg/dL – 2017.3 mg

• 7000 mg/dL – 7064.4 mg

These different solutions of glucose and DI water were put in cuvettes using the disposable pipettes and the testing started.

Differently from last week, testing was done in the Faraday cage in the Electronics Lab at Loyola Marymount University. The Faraday cage is a space which blocks electromagnetic fields. All the testing supplies and materials were moved there. When testing, a cuvette with glucose was handled with care and precision, and the system was covered with a carton box. This is shown in the picture below. Then, the lights in the cage were turned off, as well as all other electronic devices that would emit light. A small ray of light was shined on the voltmeter, just to record the values obtained. Similarly to last week, five trials were done.

What I was expecting to see out of these results is the dropping trend between the glucose concentration and the voltage. As the concentration increased, the voltage was supposed to drop. This was confirmed and proven. At first, I was having trouble proving this for the 2000mg value of glucose concentration. As I was testing higher glucose concentration of glucose, the voltage was dropping, confirming the theory proposed. All except the 2000 mg one. So I replaced the solution in it with a new one. Still, the voltage reading was not consistent. Then, I thought that something might have interfered with the system in some way. I didn’t move the set up, but I re-tested all the previous solution and made sure that the values I got were correct and that the experiment can be repeated and confirmed every time. They were the same. At that moment, I noticed that there were some bubbles and smudges on the cuvette of the 2000mg glucose solution where the light was supposed to be emitted from the 1550nm NIR LED. I immediately disposed of that cuvette, grabbed a new one, sterilized it with DI water and placed the solution with 2000mg of glucose in it. Finally, I got the reading I was looking for. This is a good lesson to just re-iterate how important it is to be super careful with this system. It is extremely sensitive, and aside from the fact that it can’t be moved and that no external influences can be present, the cuvettes through which the light is being shined can get dirty, or there can simply be bubbles in the solution. The cuvettes were moved around by holding them for the ribbed sides of the cuvette, the side through which there wasn’t light to be shined through and this ensured extra accuracy.

The glucose measurements during five trials and their average for each concentration are shown in the table below. The plot of them is shown as well. In the previous progress report, the two plots, one of them being mine from a week ago and the other one being the one a student tested last year, were compared. The trend is the same here. As the concentration of glucose in the solution increased, the voltage decreased. This, once again, confirms that the theory, with a lot of work, can possibly be built into a system that can constantly monitor glucose non-invasively.

Something to take out of this testing in the Faraday cage is the accuracy of the results obtained. There was less fluctuation while taking measurements, whereas last week, there was a lot of fluctuation for the measurements. The resting voltage when there was no solution in the 3D stand was 183.9 mV at all times. When the cuvettes were being switched out, it would always go back to this voltage, regardless of the time taken between the measurements. This also confirms that doing testing in the Faraday cage was a good idea and ensured the calibration and the accuracy were at an all-time high.

The next thing done was observing how the wavelength of each of these concentrations was looking for each solution. The Tektronix oscilloscope was used for this. The waves at 50mg and 7000mg respectively are shown in pictures below. It seems like these signals are similar, but even at different set ups, a lot of what could be seen was just noise. Even though I was in the Faraday cage, I didn’t see very good waves. This needs to be cleaned up by using the GUI that is currently being developed by the software team. When we start looking at absorption lines of the glucose on the spectrometer, the GUI will eliminate all the noise and we’ll be able to look at pure absorption lines and analyze them. The absorption lines of each element in the world are different and unique, so this is a very innovative, and I hope, accurate way of progressing with our project.

Another theory was proved, using the 980nm pulse laser and the Si-Photodiode in this wavelength range, from 350 to 1100 nm. The set up for this system is shown in the picture below.

The laser was stabilized to the metal optical table with a screw. The laser was pointed at where the cuvette would go. The InGaAs photodiode was replaced with the SI-Photodiode. From the beginning, I realized, though being in the Faraday cage, the voltage across the resistor by the photodiode was fluctuating quite a bit. There was no resting voltage that was constant, like with the 1550nm NIR LED. So I started testing and measuring voltage for different glucose concentrations. The table with measurements taken over five trials for all glucose concentrations are shown below. The graph of the plotted average value and the voltages is also shown.

As you can see, the graph is not linear. There is no trend, or standard. I realized while taking measurements that the voltages were not constant for any concentration no matter what I did. Lights off or on, input voltage higher or lower, there was no standard that could be replicated and recreated to make conclusions. This system is unreliable, and this proves that glucose absorption levels will work better at the 1550nm wavelength, which is what we will be using for our project as soon as our components are delivered and we can start interfacing the built hardware with the GUI in Matlab using the 1550nm laser and the YOKOGAWA spectrometer.

This week’s progress was crucial for understanding that it is way better and precise to look at the behavior of glucose at the 1550 nm wavelength, which is what we will be doing. The GUI is being developed right now and it looks great, and it will be interfaced with the hardware as soon as it arrives.

For next week’s progress, thorough research about absorption lines and how to read them should be done. Extra literature should be consulted about the GUI and perfecting it and getting all the components on it we need to accurately read absorption lines. Also, the chemistry department should be contacted and we should get capsules of different gases to look at absorption lines on the spectrometer. These capsules will be connected to the spectrometer by using the single mode optical fiber. I will also have a full diagram of the system and what it will look like with the components that we ordered.