Team:EPFL/Hardware
Espress'EAU - Hardware
Our goal is to create a user-friendly machine that enables the staff of small water facilities (for example the fontainiers) to perform a general water-quality assessment at the lowest possible cost. This would increase the testing frequency of drinkable water in small communities, which usually do not have the resources to conduct as many tests as the bigger ones. The Espress’EAU machine was designed to combine those aspects in a single low-cost, open-source, modular, and DIY device.
Inspiration
The Espress’EAU machine takes its inspiration from the Nespresso coffee machine capsule concept. Wouldn’t it be great to just pour water and pop a capsule in a device which will indicate to you whether this water is drinkable or not?
Our idea is to make water quality check-ups as simple as making a coffee. Yet, this requires a little bit more work backstage. Indeed, the Espress’EAU machine relies on optical density and fluorescence measurements of genetically modified yeast and we aim at users that are not trained to deal with such organisms in their daily life.
Design
The Espress’EAU machine can be subdivided into components:
Capsule
Electronic components
3D-printed components
Figure 1: Espress'EAU prototype
Capsule Design
The capsule contains the freeze-dried yeast and thus constitutes one of the main parts of the device in terms of design. By opting for a ready-to-use type of capsule, we make the handling of modified organisms accessible to anybody. The capsule is made to guarantee the user safety by providing a leak-proof and sealed environment for yeast growth so that no contact between the organism and the user is possible.
Figure 2: Espress'EAU capsule scheme
To conceive the capsule, we made extensive paper research to find which type of materials to use and choose a design compatible with optical density detection and fluorescence measurements. We got our inspiration from the eVOLVER 1 and the Klavin’s lab turbidostat 2. The capsule prototype we imagined is made from commercially available components for a total price of ~7CHF/capsule (See Table 1). It can be assembled in less than 5 minutes and does not require fancy drilling or cutting tools (we used a hole puncher from the supermarket!).
Table 1: Capsule components and prices
We demonstrated that our capsules are viable for the growth of freeze-dried yeast.
Commercial dry baking yeast was used to ensure that the capsule was a suitable environment for the growth of our organisms. For this, 0.0128 g of dry yeast and 5 mL of SC medium were placed in a capsule prototype. A shaker was used overnight and the Optical Density (OD600) was measured after 16 hours using a plate reader. These 16 hours were also the average length of the experiments that were carried out to test the different strains.
4 samples of 100μL were taken before and after the 16 hours and the average OD600 was measured. At the beginning of the experiment, the average OD600 was 0.224. At the end, the average OD600 was 1.28. This experiment showed a 5-fold increase for the yeast growth, showing that the capsule that we designed is a viable environment for dry yeast.
Table 2: Measurements of OD600 in the capsule
Figures 3 and 4: Growth of commercial dry yeast inside the capsule. On the left before, on the right after 16 hours.
Despite the great benefit of a capsule system in terms of handling, it can rapidly become a source of waste. Thus, we privileged durable and autoclavable materials such as glass so capsules can be re-used several times and a waste-treatment plan can be implemented.
Detector Design
Background and overview
Espress'EAU is a user-friendly, low-cost and modular water quality testing device. In comparison to similar devices which are not portable, Espress'EAU is optimized to be compact, aesthetic in design, and at the same time not compromising on cost. It is equipped with a self constructed magnetic stirrer to maintain the homogeneity of the medium and a temperature check, controller to regulate the temperature of the environment. This allows us to create a compatible environment for water testing. Testing is done by measuring the optical density (to monitor cell growth) and fluorescence of the engineered yeast. These two variables are measured using light-to-frequency converter (TSL235R) and a RGB light sensor (ISL9125) respectively. Then,the collected data is stored on the host machine for further data processing and analysis. The Espress'EAU external structure is made using 3D printed parts which allows it to be personalized and hence can be used for a wide range of applications.
3D components
To integrate the different components of the device together, we decided to 3D print the structures. This was done using a Prusa i3 MK3S 3D printer. Since our goal is to make the device sustainable, we chose Polylactic acid (PLA) as the filament, which is one of the most environmentally-friendly 3D printing materials and, unlike Acrylonitrile butadiene styrene (ABS), is biodegradable 3. The measurement of all the 3D pieces and corresponding components was taken using a Caliper scale. The design of these 3D components was inspired by the work done by the ETH iGEM 2019 team
We primarily constructed three 3D components whose STL files can all be found on Github:
Vial and sensors (OD, Fluorescence) holder
Magnetic stirrer and temperature control holder
A box to cover the whole device so that sunlight cannot interfere with the readings
Figures 5 and 6: 3D model of the vial holder and the cover box
Magnetic stirrer
We use a DC motor which is powered by a 9V, 1.5A power supply and a 3D printed structure holding 2 magnets of the stirrer. These magnets are cube shaped and are balanced on a plastic plate on the axle of the DC motor.
The voltage to the circuit is controlled using a potentiometer which is of 10,000Ω and a N channel MOSFET. The potentiometer adjustment creates a varying potential difference across the gate of the MOSFET, and the source pin of the MOSFET simply follows this value of the potential difference and adjusts the voltage across the motor accordingly.
The magnets are placed in such a way so that half of the structure is north and the other half is south. This increases the area of magnetism and holds the stirrer firmly into its magnetic pull, creating a vortex so that the medium in each vial is homogenous.
Click on the figure for the Github link of the 3D design!
Figure 7: 3D model of magnetic stirrer
Temperature controller
For the yeast to have an optimal growth, we need to calibrate the environment to a suitable and constant temperature.
For this we need to achieve a temperature of reference of 30 degrees. If the reading exceeds 31 degrees, the Peltier element is shut down, otherwise it heats in order to only allow slight deviations on either side of the reference temperature.
We did use a thermoelectric Peltier element which is then regulated using a TMP36 temperature sensor. As temperature increases, the voltage across a diode TMP36 increases at a known rate. By precisely amplifying the voltage change, it is easy to generate an analog signal that is directly proportional to temperature which is captured by the Arduino.
Since the Peltier consumes a lot of current, we needed to use a MOSFET as the safety circuit. We could have selected either a P channel or an N channel transistor, but the N-channel MOSFET has several advantages. For example, the N-channel majority carriers (electrons) have a higher mobility than the P-channel majority carriers (holes). Because of this, the N-channel transistor has lower RDS(on) and gate capacitance for the same die area. Thus, for high current applications the N-channel transistor is preferred and creates more safety for the temperature regulator of the circuit 4.
Optical density
We expect the growth of the genetically modified yeast to be influenced by the contact with polluting agents. To determine the amount of the pollutants in a sample, we thus want to measure the yeast growth when it is in contact with the water analyte. To do so, an optical-density measurement of the yeast growing in the water sample can be done.
Measuring optical density here, means measuring the light absorbance of the sample at 600nm: the more yeast cells there are, the more light will be absorbed (the more dense in cells the sample is).
Figure 8: Scheme of the optical density sensor
To do this measurement, only two components are needed: a light-to-frequency converter (TSL 235r) and a white 2-pin LED. They are put in the device on both sides of the vial separated by an angle of 180°: this way the sensor captures the light from the LED that goes straight through the vial. The idea of using the TSL235r came from EPFL professor Luc Patiny that created a DIY spectrophotometer using the same component 5.
One note about the TSL235r: before it saturates, it takes a long time to measure high frequencies. Since we are interested in the evolution of the frequency over time and the sensor took too much time to measure the optical density of a clear sample a 10kΩ resistor was used, instead of the 10Ω planned originally.
Tests were done to verify if this OD600 sensor works:
Comparison to the values from a spectrophotometer using standard yeast dilutions
Measurement of growth curve
Figure 9: Linear relation between the OD600 measured with 11 standard samples in a spectrophotometer (x axis) and the transmitted light measured by our device (y axis).
For the first test, 11 samples were prepared by dilution from an overnight yeast culture and their OD600 was measured on a spectrophotometer to get a reference calibration curve. The samples were from OD600 0 to OD600 1 (with a .1 step). The frequency captured by the sensor from the Espress’EAU device was then measured for 16 hours and compared the results obtained from the calibration. A linear relationship between the reference and the measurements done with our sensor was observed.
Figure 10: Light transmittance of yeast cells measured by our sensor
Figure 11: Growth curve of yeast cells measured by our sensor
Since our device measures the transmitted light intensity but we want to display the optical density, which is more adapted to monitor cell growth, a conversion had to be done to obtain the final results. Indeed, Figure 10 shows a more rough data while Figure 11 displays data that has been treated in order to make the results read-out easier.
Figure 11 shows that it was possible to measure the OD600 using our device in order to monitor yeast growth for more that 16 hours.
We conclude from the linear relationship between the frequency measurements and the OD600 measurements, as well as the growth curved obtain by measuring a yeast culture, that the OD600 sensor of our device does give convincing results, and can be used to measure optical density.
Fluorescence sensor
When the yeast reporter strains come in contact with pollutants, we expect the cells to express a red fluorescent protein in response to a stressing environment. This red fluorescence can be excited by yellow light.
To measure the fluorescence we used two components: an RGB light sensor ISL29125 and a yellow 2-pin LED (530-560 nm). The light source and the sensor are put around the vial at a 90° angle. This is because we want the sensor to measure light coming from the fluorescence, and not from the light source. With such an angle, the yellow light coming from the source that does not encounter a cell goes through the vial without being detected by the sensor, while the yellow light absorbed by the yeast cells will be replaced with a red fluorescence emission at a different angle, This will guarantee a higher chance of being captured by the sensor and reduce background noise.
Figure 12: Scheme of the fluorescence sensor.
Tests:
Measurement of fluorescence using an inducible yeast strain
Future plans: measurement of the fluorescent signal from genetically modified yeast
The first test was carried out and its results are showed in the Figure 12 below. An red fluorescent inducible strain was activated using β-estradiol. This strain can be used as a positive control to detect red fluorescence signals. It was compared to the signal detected at a blue and a green wavelength. Water was used as a negative control as no fluorescence should be detected. As expected, only the inducible strain signal from the red fluorescence measurement shows a visible peak for the fluorescence intensity. This confirms that the fluorescence sensor is functional and can be implemented in the Espress'EAU design.
The preliminary experiment we conducted on the fluorescent sensor indicate that we can indeed measure fluorescence coming from an induced strain. We could confirm our result by changing the baseline from the current measurements we made (from water to wild-type yeast).
Figure 13: Fluorescence measurements with inducible strain for 3 different wavelengths.
For our purposes a yellow LED was selected specifically to excite the mScarlet-i fluorescent protein, but depending on the wavelength of the fluorescence that needs to be measured, another LED can be used in the same design.
Integration of Espress'EAU
Cost of components
Table 3: Hardware components and prices
Open source project and Main Circuit diagram
Figure 14, 15 and 16: 3D schematics and printed prototype
We use the Arduino IDE for writing the C++ code for interfacing and an Arduino UNO board as the microcontroller. The complete code, circuit diagrams and the 3D schematics can be found on our Github.
Click on the figure for the Tinkerpad link!
Figure 17: Complete circuit schematics
User feedback
User feedback was an essential part for the design of Espress'EAU. During the last phases of our project we met the with our main stakeholders in order to get a feedback on the first prototype of Espress'EAU:
-
Quentin Morezzi, president of the AFSR, confirmed that Espress'EAU met the needs of a fontainier in terms of size, price and simplicity of usage and that the minimalist design would fit perfectly into an office environment. He suggested that the cover and the system support could be merged together by using insulating foam. Additionally, an automated processing of the data as well as the use of an SD card to store it would greatly simplify its usage by a fontainier.
-
Endre Horvath, CEO of Swoxid, complimented the design of Espress'EAU and the choice of 3D printed components. Indeed, he affirms that it is the best choice if one wants to share the project around the world and the easiest way to prototype and experiment. He draw attention on the fact that our product would have to pass a safety certification (because of the integration of water and electrical components), but the case ensures safety in case of explosions. He suggested that we could use UV light in order to kill yeast after the recording of measurements.
References
Special thanks to our sponsors!
Follow us on our social media.
@EPFL iGEM 2020