The irregular shape of common commercial cuvettes made an analytical understanding of the light paths passing through it difficult. Thus it was challenging to design a sensing system around it. While there are straight-sided commercial cuvettes, they are so small that modifying them directly is also difficult, especially since we wished to modify the cuvette such that bacterial culture could flow through it continuously.

We thus decided to design our own cuvette (Figure 1) using 1.5 mm thick acrylic sheets. Apart from allowing us to ensure that the walls are flat, it also allowed us flexibility in the placement of the acrylic tubes channeling media through the cuvette.

Our cuvette was designed to have as few unique parts as possible. This was to facilitate easy assembly, which is an important design consideration since the cuvette is small. We would need to be able to quickly fabricate many cuvettes for testing.

This design choice proved useful when we were troubleshooting the cause of our cuvette’s leakiness. Acrylic glue could not form seamless bonds. Even if it appeared so from visual inspection, leaks sprung once we tested the cuvettes by pumping water through them at high pressure. After several rounds of trial and error with other adhesives, we discovered that Acrifix 1R 0912 UV adhesive formed a watertight seal and is clear when cured. We used it in our final design iteration for the cuvette.

The 3D-printed casing houses the sensing system, which comprises a TSL235R light-to-frequency converter, a 600 nm LED, a 535 nm LED, and a LEE filter with a peak transmission of 600 nm. (Figure 3). The 600 nm LED is in the slot opposite the TSL235R, and is used to measure optical density (OD). The 535 nm is in the other LED slot, and is used to measure fluorescence. The LEE filter needs to be cut to an appropriate size and attached over the TSL235R. We used masking tape to secure the filter.

A block diagram of our feedback control system is shown below (Figure 3).

1. Prepare 20 ml of E.coli cells harbouring Brep-RFP-YbaQ at OD = 3. These cells should have been previously been continuously exposed to blue light to repress RFP expression.
2. Dilute the cell medium to get 12 different samples with ODs ranging from 0.1 to 3. Label them from 1 - 12.
3. Measure the actual OD of sample 1 using the NanoDrop.
4. Transfer sample 1 into the cuvette with a syringe. Place the cuvette into the casing and close it.
5. Using PLX-DAQ, obtain approximately 10 readings of TSL235R’s output frequency when the 600 nm LED is lit.
6. Repeat steps 3-5 for the rest of the samples.
7. Plot a graph of output frequency against OD readings from the NanoDrop.

1. Prepare 20 ml of E.coli cells harbouring Brep-RFP-YbaQ at RFP reading = 7100.
2. Dilute the cell medium to get 12 different samples with RFP readings ranging from 100 to 7100. Label them from 1-12.
3. Measure the actual RFP reading of sample 1 using the NanoDrop.
4. Transfer sample 1 into the cuvette with a syringe. Place the cuvette into the casing and close it.
5. Using PLX-DAQ, obtain approximately 10 readings of TSL235R’s output frequency when the 535 nm LED is lit.This LED excites the RFP, causing it to fluoresce red with a wavelength of 600nm.
6. Repeat steps 3-5 for the rest of the samples.
7. Plot a graph of output frequency against RFP readings from the NanoDrop.

By following our experimental procedure, we were able to plot calibration curves for OD and fluorescence sensing. Please visit Results:Cell-Machine Interface for the results of our characterization of this sensor.

• Arduino Uno x 1
• 600 nm LED x 1
• 535 nm LED x 1
• Light-to-frequency converter TSL235R x 1
• LEE filter (peak transmission 600 nm)
• OD=6 mm ID=3 mm acrylic tubes x 1 m
• Acrifix 1R 0912 UV Adhesive