Our project is to design a biophotovoltaic cell using the photosynthetic abilities of cyanobacteria. This was accomplished utilizing a screen printing technique. The biophotovoltaic cell is composed of two electrodes, both printed with a carbon conductive ink. These electrodes are at different concentrations from one another creating a concentration cell to promote the flow of electrons. A layer of biological ink containing our cyanobacteria cells was placed on top of the conductive ink anode creating a “bioanode” that is the source of free electrons as the cyanobacteria cells go through photosynthesis. Finally, a hydrogel that has been soaked in a 2% NaCl BG-11 media solution is placed on top of the printed biophotovoltaic cells to act as both a source of nutrients for the cells, and a salt bridge for protons.
Screen printing is the process in which ink is spread through a mesh screen consisting of a design onto a surface, leaving behind just the design1. Mesh screens are created by tightly stretching silk mesh across a frame, and a coat of emulsion is applied to make the screen impervious to the printing ink1. To create the desired screen printed design the pattern is printed onto transparency film, which can be hand drawn or for more precision printed using a standard inkjet printer, and the film is cured on the emulsion using UV light. The emulsion that was blocked from the UV by the printed stencil gets washed away from the screen and what is left is the screen with the now printable desired design.
The advantage of screen printing comes down to its ability to be accomplished using very simple and cheap materials, and its small learning curve. Along with this, screen printing allows for a large quantity to be printed at a time, decreasing the time and cost of prints compared to expensive 3D printers that would only be printing one voltaic cell at a time. For this project specifically, screen printing made the most sense for printing due to the high viscosity of the carbon conductive ink used, which contributes to its low resistivity, that would make it very difficult to print using traditional inkjet methods. In terms of future applications screen printing is also an advantageous avenue due to the versatility of materials you can screen print on, so you are not limited to just a paper substrate4.
One downside to screen printing is the time it takes to set up everything and carry out the entire printing process, along with the amount of manual labor required during the printing process. To combat this our team decided to make a prototype of an automated screen printer, inspired by automated screen printers found in large scale commercial screen printing facilities, that would decrease the amount of time and manual labor required to print our biophotovoltaic cells in large quantities. In doing so another advantage of an automated design is the ability to standardize the printed cells by decreasing the effect of human inconsistencies. With this design the only manual steps include adding and removing the paper substrate from the platform, placing hydrogels to the printed cells, and attaching wires to the printed cells. The entire printing process, however, is automated including the drying step. The automated design includes three stations that are able to rotate in a timed fashion on their own. Each of the three stations has its own arm that is able to raise and lower using a pulley in a timed fashion, facilitating the creation of the biophotovoltaic cells. Two of these arms contribute to the actual printing of the biophotovoltaic cells through being equipped with the necessary linear motion components to facilitate the movement of a roller. The third arm acts as the drying station and is equipped with two small fans that have been coded using an arduino. By speeding up the drying process the printer is able to output product at a faster rate compared to having to wait for the inks to dry on their own.
Our screen printer system addresses the need for a 2D printer in synthetic biology, especially in circumstances where large quantities need to be printed quickly and accurately with a high amount of replicability. While 3D printing has become popular for many applications including cancer therapy, tissue engineering, bone regeneration, and wound healing among others, 2D printing in synthetic biology is just as important for applications such as studying microbial interactions, DNA printing, and circuit assembly. While manual screen printing could be used to accomplish this there are many steps involved which not only requires the need for training, but also provides several areas in which human variation will have a negative impact. Factors such as printing speed, pressure applied, amount of ink used, and alignment between layers all become standardized with the addition of automation to create ideal prints every time. Along with this the automated screen printer requires far less training than having to learn the individual steps of the screen printing process, leaving it the preferred option in terms of practicality in labs. Another driving factor for our printer is its inexpensive cost and ability to be scaled up. There is not a small scale automated screen printer that would be able to fit in a standard lab on the market. Most manual multi-station screen printers typically range from $100 to close to $1,000 USD, and industrial sized automated screen printers range from about $1,000 to $10,000 USD.
Our automated screen printer provides a template that can be customized to be scaled up or down depending on the number of stations needed, provided the structural skeleton is able to configure this number. This allows for a variety of applications beyond just screen printing biophotovoltaic cells, as we anticipate users will adapt their printers to their own purposes. To help aid in this process a detailed manual is provided that contains detailed documentation on the engineering process.
In biophotovoltaic cells, photosynthetic organisms are placed upon an anode. These photosynthetic bacteria capture solar energy and convert it to usable energy, while simultaneously splitting water molecules in photosystem II. The splitting of water molecules releases free electrons that are encouraged to flow towards the printed cathode due to a difference in concentration between the two electrodes, the cathode has a higher concentration. This design is known as a concentration cell. Through this flow of electrons, a current is created and a voltage can be measured across the cell entirely produced by the bacteria. In order to accomplish this biophotovoltaic cell setup, each aspect of the setup was carefully considered including the substrate, the composition of the biological ink, the conductive ink, and the hydrogels used.
Figure 1: Diagram of biophotovoltaic cell
A paper substrate was chosen to act as an inexpensive, widely available, and practical substrate to be printed on. The paper was microwaved for three seconds prior to coming in contact with cyanobacteria cells. Paper was also a strategic choice in substrate when it comes to future applications of scaling up our design as it is easily upscaled but still affordable.
Our printed biophotovoltaic cells were set up using carbon conductive ink as the electrodes. A carbon conductive ink was chosen specifically because it does not contain any harsh chemicals or elements that would interact negatively with our cyanobacteria, like a common silver based conductive ink. Because the anode and the cathode are composed of the same material, literature suggests that a concentration cell needs to be formed to encourage the flow of electrons. A concentration cell is formed when the electrodes are made up of the same material, but the concentration of the cathode is greater than the concentration of the anode. This causes a transfer of electrons from the anode with the lower concentration, to the cathode with the higher concentration, and this transfer of electrons is what produces a voltage output. Three different carbon conductive inks Novacentrix, Nagase ChemteX 2001, and Nagase ChemteX 2042, were tested against each other to see which conductive ink performed the best overall in terms of screen printability and facilitating the most voltage output for a biophotovoltaic cell. The data found from performing these tests was used by the modeling team in the electricity model to provide a better understanding of the role of the conductive ink in the overall biophotovoltaic cell. The variables tested were composition, consistency, resistivity, current, and electrode design.
As previously mentioned it was important to select inks that were all carbon based to ensure they would not harm our cyanobacteria cells. The Nagase ChemteX 2042 ink formula had extra additives aside from just carbon and graphite within its composition that we predicted would not react well with our cyanobacteria, and may kill them quicker than the other inks. These additives include ethylene glycol monobutyl ether acetate, isopropanol, and nitrocellulose.
The viscosity of the conductive ink was looked at optically. An ideal ink would have a high viscosity, to prevent it seeping outside of the design when being printed, similar to that of normal screen printing ink. Through observation, it was found that the Novacentrix ink was the least viscous compared to the Nagase ChemteX inks, and did not screen print cleanly as seen by the smeared edges in the image below. The two Nagase ChemteX inks both had about the same consistencies and both were able to be screen printed cleanly.
Figure 2: Screen printed Novacentrix ink vs Nagase ChemteX-2042 ink
Resistance was measured across individual printed electrodes of the same size composed of the different inks to find the most conductive ink. Resistance is a measurement of the opposition to flow of current in a circuit. This opposition of flow results in a decrease in the voltage potential across the conductive ink. Therefore, the higher the resistance the more opposition to the flow of current, so less voltage output is able to occur. This relationship can be seen by using the formula from Ohm’s Law, V=IR, where V represents voltage, I represents current, and R represents resistance. Resistance is inversely proportional to conductance, and we wanted the most conductive ink to prevent the most amount of voltage loss across the electrodes as possible.Therefore, we wanted to find the ink with the least amount of resistance. Through experimentation it was found that printing more layers of the conductive ink on top of each other and allowing them to dry fully had a substantial effect in decreasing the resistance, so we made sure to implement this into our final biophotovoltaic cell design. Our final design consists of electrodes that are printed with three layers of the conductive ink. In printing more layers, the cross sectional area is being increased by a very small amount while the resistance decreases by a much larger amount, so overall the resistivity is being decreased as seen in the equation for resistivity: ρ =
Figure 3: Setup for testing resistance across Novacentrix ink, Nagase ChemteX-2042 ink, and Nagase ChemteX-2001 ink
Figure 4: Printed conductive inks
The printing in the image consisted of printing a single layer of each conductive ink. The resistance across the Nagase ChemteX 2042 ink was by far the lowest measuring 109.11 ohms, making it the best option in terms of resistivity.
Current was measured across the electrodes composed of the different inks by using an arduino to supply 5 volts to the anode and cathode. The most effective design was determined by the one that we measured to have the most current throughout using a multimeter. Results from these measurements can be found on the modeling page.
Different printed electrode designs were tested for the biophotovoltaic cell, and results from these experiments were used by the modeling team to model the electricity output based on the differing factors in each design. Design variations tested included the length of the electrodes, the concentration ratio between the anode and cathode, the distance between the anode and cathode, the shape of the electrodes, and amount of interactions between the anode and cathode. Comparison between these designs was tested by measuring the resistance and current across them. Results from these experiments can be found on the modeling page.
Figure 5: Labeled sheet of electrode designs tested
In order to screen print our cyanobacteria, they needed to be suspended in a substance that had a high enough viscosity to prevent the biological ink from spreading outside of the intended pattern. Based on literature found on how to spatially control cyanobacteria when screen printing, three different additives, xanthan gum, carboxymethyl cellulose (CMC), and gum arabic, were tested3. These additives were combined with liquid BG-11 media so that the biological ink would support cyanobacteria life. These inks were tested for four different variables: visual clarity, E.coli viability, consistency, and cyanobacteria viability.
Figure 6: Xanthan gum, carboxymethyl cellulose, and gum arabic additives
Biological Ink Preparation Protocol
The visual clarity of the biological inks was observed optically. There was a clear difference in the transparency of the gum arabic biological ink compared to the CMC and xanthan gum. This was due to the ratio of gum arabic added being much higher than the CMC and xanthan gum ratios. Based on observation we concluded the CMC was most transparent, and thus was the best choice to allow the most light through to our cyanobacteria cells for them to optimally perform photosynthesis.
Figure 7: Xanthan gum, carboxymethyl cellulose, and gum arabic biological inks
The consistency was important to consider to ensure the biological ink was not too viscous to where it could not be pushed and screen printed through the screen, but also did not spread outside of the design when being screen printed by not being viscous enough. After looking at the results from both the visual clarity and E. coli viability tests CMC was chosen as our additive. Different percentages of CMC were added to BG-11 and tested to see how viscous of a biological ink we could achieve that was still able to be screen printed through our mesh. The viscosity of the biological ink with the different CMC concentrations was observed optically, with the viscosity of the ink increasing with each increased concentration as expected. After screen printing each biological ink with a different CMC concentration through a screen, the screen printing results were analyzed using ImageJ. ImageJ is an image processing program that allows users to quantify and validate scientific data. The goal was to find which ink had a high enough viscosity, similar to regular screen printing ink, that would prevent the ink from spreading outside of the design on the screen. Based on the ImageJ results, a 10% weight-to-volume ratio of the CMC performed the best; however, viscosity also needed to be taken into account with our automated screen printer design. A future advancement of our design would use peristaltic pumps that need to be able to push the ink through tubing to be dispensed. Peristaltic pumps operate by compressing an inner tube with rotating rollers in order to create a vacuum that draws fluid in and also dispenses it. If the ink is too viscous, the pump will not be strong enough to push the ink through the tubing, preventing it from being dispensed. Because of this, we decided that lowering the concentration slightly to an 8% weight-to-volume ratio would be optimal.
Figure 8: Screen printed biological inks processed through ImageJ
Figure 9: Screen printed biological inks
E. coli viability in each biological ink was tested to see how well each biological ink was able to sustain life. The E. coli used was transformed to contain Green Fluorescent Protein (GFP) and was added to each biological ink. The inks were spread in a thin layer on paper substrate in a Petri dish, and a hydrogel soaked in LB medium was placed on top. These samples were placed in an incubator and imaged using Saphire Software with the assumption that if the E. coli cells die they will not fluoresce. From this experiment it was concluded that all biological inks were able to sustain life, with the CMC ink producing results most similar to our positive control again making it our top candidate.
Figure 11: Results from Sapphire Software imaging
Protocol for E. Coli Viability in Biological Inks
Based on the data found from our E. coli. viability experiment, we are making the assumption that these inks are able to sustain bacterial life. For future direction, we would like to transform cyanobacteria with Yellow Fluorescent Protein and perform the same kind of analysis that we did with the E. coli.
The hydrogels we selected were the Spenco 2nd Skin Squares based off of the hydrogels used in the Sawa et al. paper2. They were selected for their size, clarity, and absorbability. These factors corresponded to the size of the biophotovoltaic cells we were aiming to make, maximizing the amount of light that would be able to reach our cyanobacteria, and ensuring the hydrogels would be able to absorb the media to offer nutrients to the cyanobacteria, as well as act as a salt bridge to facilitate the hydrogen gradient2. In order to optimize them for our biophotovoltaic design needs, several experiments were performed to test different variables surrounding our hydrogels including the amount of time it took them to dry out comparing different medias and different locations, as well as their change in resistance while drying out.
Figure 12: Graph of how the media the hydrogels are soaked in affects their drying time
Figure 13: Graph of how different locations affect the drying time of hydrogels
Fig. 12 shows the trend in drying times of hydrogels that have been soaked in different media. From this graph we can conclude that BG-11 media by itself, along with 2% NaCl BG-11 media have the fastest drying times. A slower drying time is favorable for the biophotovoltaic cells because they will be able to sustain life of our cyanobacteria cells longer.
Fig. 13 shows the trend in how different locations affect the drying times of hydrogels. As expected the samples placed in the incubator dried out faster, which indicates that our prototypes should not be placed in an incubator to keep the hydrogel hydrated for as long as possible.
Figure 16: Hydrogels soaking in media
In the design process of our biophotovoltaic cells, many prototypes were created to work towards maximizing the amount of voltage we were able to measure easily.
Our first attempt at creating a functional biophotovoltaic cell was a very inspiring experience. This prototype was created using the components we had spent time testing and optimizing including our conductive ink, biological ink, and hydrogels soaked in 1% salt BG-11 media, as well as materials we already had in the lab including copper wire, tape, a multimeter, paper, and a pipette tip box to help maintain sterility. The protocol we followed was created by us and based on the design found in the published Sawa et al journal2. Here, three layers of the conductive ink were printed using the screen printing technique while three layers of the biological ink were pipetted on top of the anode. This was because the concentration of CMC in our biological ink was too low, and the ink was not viscous enough to also be screen printed without leaking outside of the pattern. Each layer of ink was given time to dry before the next was printed to achieve a layering effect. Once the final layer of biological ink was dry, squares of hydrogels were placed on top of each anode and cathode pairing to facilitate a hydrogen gradient in our biophotovoltaic cells. Here, the hydrogels were sized so that they would be touching each other. Finally, two small lengths of copper wire were cut, stripped, and one was taped to the first anode while the second was taped to the last cathode in order to read the voltage across five of our cells. Our GW Instek GDM-8245 multimeter was used to probe the wires to show the voltage drop across our cells. Because the wires were connected to the conductive ink using a small piece of tape, only a small portion of the copper wire was making contact with the ink. Along with this, the connection was not very strong and the wires would often shift, especially when the probe would make contact with the wires, leading to very unstable readings. With these connection considerations in mind, the output of this first prototype was reading between 0.12-0.25 mV. To ensure the voltage being read was because of the cyanobacteria cells, we measured across a control setup that consisted of just the conductive ink without any cyanobacteria cells. When we measured this voltage it came back as zero volts, showing that the voltage we were measuring in our prototype was directly caused by the cyanobacteria. Another note is that we were not seeing an increase in voltage drop across biophotovoltaic cells in series. One reason for this could have been due to the cyanobacteria spreading from the anode to the cathode, creating one large anode across the paper and removing the flow of electrons that would be seen in a concentration cell. This spreading of the cyanobacteria was caused by the biological not having a high enough concentration of CMC, and not being viscous enough to remain strictly on the anode. A future improvement for subsequent prototypes is creating a more viscous biological ink with a CMC concentration of 8%.
Figure 18: Measuring the voltage across our first prototype
Figure 19: First prototype
After learning about the shortcomings of measurement of the first design, more supplies were purchased because it was clear that we needed to better the connections of the wires to the electrodes to obtain more consistent readings. The first change in materials included hydrogels that were cut down smaller to only cover the area of the anode and cathode of one biophotovoltaic cell at a time. Our first prototype was fluctuating between both positive and negative voltages, showing that the electrons were not strictly flowing from the anode to the cathode which we believed was caused by all of the hydrogels being in contact with one another. Along with this the hydrogels were soaked in a higher salt concentration of BG-11 media at 2% salt to encourage a stronger hydrogen gradient. Copper tape was used in place of the copper wires due to its high conductivity, flexibility, and increased contact surface area with the electrodes. By increasing the contact surface area a more stable reading was able to be measured. A new 87V Max multimeter was donated to our lab from Fluke and used in place of our old multimeter. With these changes in our new prototype, the multimeter was reading a consistent 17 mV across one biophotovoltaic cell, but also gave the same reading of 17 mV across four biophotovoltaic cells connected in series. In this prototype, we still were not able to measure an increase in voltage drop across multiple biophotovoltaic cells in series.
Figure 20: Second prototype
For the third prototype, the goal was to be able to measure an additive voltage drop in voltaic cells in series. To do this, one consideration we had was printing the adjacent anodes and cathodes of different voltaic cells closer together so the oxygen diffusion from the anode to the cathode is across a shorter distance. Having a shorter distance is both faster for oxygen diffusion because the molecules do not have to travel as far, and more favorable for the oxygen molecules compared to having to diffuse across a longer distance so they would diffuse at a greater rate. When the oxygen molecules reach the cathode, they participate in the reduction half-reaction that contributes to the formation of an electrical current, so allowing more oxygen molecules to travel to the cathode results in a greater contribution to the electrical current formation. Another observation we had in our second prototype was that when we measured the voltage across the cells using the multimeter, the voltage would begin to drop. To address this, we attached a 10 microfarad capacitor across four of the voltaic cells in series. Capacitors act to temporarily store built up voltage in a circuit, and release this energy when needed. With this addition we were able to measure 454 millivolts across the four cells, compared to 174.5 millivolts across the four cells without a capacitor. Measuring across one set of electrodes with the capacitor resulted in 92 millivolts, therefore we concluded that the voltage drop from the four cells in series was additive, and our improvements worked.
Figure 21: Third prototype with capacitor connected across a single pair of electrodes
Figure 22: Third prototype with capacitor connected across a series of four pairs of electrodes
Figure 23: Graph of the average voltage measured from each prototype
For subsequent prototypes, one goal is to be able to light up a diode. Having proven that we were able to build up voltage across multiple biophotovoltaic cells in our third prototype, a next step would be to print more biophotovoltaic cells in series, and measure if their voltage drop would continue to be additive. One consideration for this setup is that the more biophotovoltaic cells that are printed, the larger the value of the resistance will be between the first and last cells. Because of this, naturally some voltage output will be lost due to the resistance. To combat this another design we could test is connecting multiple series of biophotovoltaic cells in parallel. There is no voltage drop over cells in parallel, but the distance between them will decrease which will decrease the resistance experienced by the current. Another way we would be able to build up more voltage would be by adding more capacitors to the individual biophotovoltaic cells connected in series. This would not only allow the biophotovoltaic cells to charge up their voltage before releasing it, but would also give a more stable reading to the multimeter when measuring voltage across the circuit.
Another future direction would be testing these biophotovoltaic cell prototypes using transformed cyanobacteria. Currently, we have only tested these prototypes using wild-type cyanobacteria, Synechocystis sp. PCC 6803. A part of our wet lab team’s efforts was put into genetically modifying cyanobacteria with psbE and psbF. The alpha and beta subunits respectively of cytochrome b559 play a role in the generation of electrons from water on photosystem II 5. Transforming cyanobacteria with these has been proven by the wet lab team to work, so the next step would be growing enough transformed cyanobacteria to test our biophotovoltaic cells. The assumption would be that these transformed cells would be able to produce more voltage than the wild-type strain.
A third future direction would be looking into incorporating ethanol distillation into our prototypes. The ethanol production by our cyanobacteria is one of our modules in the wet lab aspect of our project. With this, we have found that ethanol can be distilled from the hydrogels sitting on top of the cyanobacteria cells in our biophotovoltaic cell design. This distillation process creates a cleaner way to produce ethanol and would be a strong addition to an upscaled version of our biophotovoltaic cell systems.
The manual screen printer setup includes the screen, transparency film with the desired design printed, a scoop coater, emulsion, the conductive ink, the biological ink, paper substrate, a squeegee, UV light, and a hairdryer.
Figure 1: Screen printing supplies
For a more affordable and accessible design, we built our own screen printing silk screens using the wooden frames found in canvases for painting, and using a staple gun to attach silk screen mesh to the border of the frame.
Figure 2: Finished screen
Building Screen Printing Screen Protocol:
To prevent there being air bubbles and other disruptions in the emulsion layer that would decrease the detail of our stencil, each screen was first removed of dust and other debris using tape. Once the screen was clean, the emulsion was applied in a dark room with the lights turned off to prevent the photoreactive emulsion from curing prematurely. The emulsion was applied with a scoop coater constructed from cardboard and duct tape, and smoothed out with a squeegee. Once the emulsion was applied, four thumbtacks were placed in the corners of the screen and the screen was placed upside down in a black stage container to prevent any light exposure while the emulsion dried for about 24 hours. Our desired design for the screen was created in Procreate and printed onto UOKHO transparency film paper using an inkjet printer. To get the design onto the screen, the screen was pulled out of the black container and placed in a dark room. The transparent film was laid over the screen, and a large glass panel was placed on top to keep the transparent film flush against the screen. The UV light was placed above the setup and was found experimentally to work the best at 3:30 seconds. After curing the screen, it was removed from the setup and placed in the sink to wash out any uncured emulsion, leaving only our stencil. A t-shirt was placed at the bottom of the sink to act as a filter and prevent any emulsion from going down the drain and into the waterways. Once the screen was washed out it was left for 24 hours to dry and after that was ready to use for printing.
The emulsion we chose, Photocure BLU from Murakami Screen USA Inc, was recommended to us by Eileen Bushnell, a professor in screen printing at the College of Art and Design at the Rochester Institute of Technology. It is applied to the screen using a scoop coater, a small trough used to apply an even layer of emulsion on the screen. In order to figure out how long our screens needed to be exposed under UV light for using our UV light setup, two different emulsion Step Tests were performed to gauge the optimal time. A Step Test is a method used in screen printing for finding the optimal amount of time needed to expose emulsion. It is accomplished by applying the emulsion to the screen using the protocol listed below, and once it is dry sectioning off the screen into areas. A design gets placed over the screen and each area gets exposed to UV light for a different amount of time. Then, when the emulsion is washed off of the screen it is observed which areas are able to be washed off the fastest with the least amount of effort, and the least amount of residual emulsion left on the screen, while maintaining the detail of the design. The results from the second Step Test showed that the best resolution occurred at a UV cure time of 3:30 seconds with our setup.
Figure 3: Emulsion being applied with a scoop coater
Emulstion Test 1: Step Test
Figure 4: UV light setup
Figure 5: Washed out screen after first step
Emulsion Test 2: Resolution Step Test
The same protocol was followed as for the first Step Test, the only differences being this test used a design with finer detail and the time intervals were shorter to obtain a more precise UV cure time. After the screen was cured it was washed in the sink, again filtering any emulsion with a t-shirt to prevent it from going down the drain. Once the screen was dry it was printed using regular screen printing ink. A picture was taken of the print and this image was put through ImageJ to quantify how much surface area was printed at each time interval compared to the expected surface area taken from the pattern on the transparency film, and the results are shown in Fig. 11. From this it was concluded that 3:30 seconds performed the best because it had the lowest difference in area compared to the intended design.
Figure 6: Resolution Step Test transparency curing under UV light on screen
Figure 7: Washed out screen after Resolution Step Test
Figure 8: Screen printing ink through Resolution Step Test screen onto paper
Figure 9: Printed Resolution Step Test screen
Figure 10: Print ran through ImageJ software
Figure 11: Resolution Step Test Results
Figure 12: Participant reading screen printing protocol
Figure 13: Participant measuring resistance across conductive ink print
Figure 14: Prints of screen printing ink, conductive ink, and CMC ink from participant
To gain insight into different perspectives on the screen-printing process, we invited individuals with limited experience in screen printing to try out our protocol. We provided a detailed instruction manual and all of the necessary materials for the process. After completing the process, participants filled out response forms to share their feedback.
While the participants believed the instructions were straightforward, they found the process labor-intensive due to the time commitment. Regarding the different inks used throughout the testing, participants all agreed the inks were generally easy to spread but had mixed feelings about which ink was the easiest to use. For instance, while some participants felt CMC ink was harder to lay on the screen due to its thick consistency, others felt that this actually made it easier to spread and print.
Throughout the process, applying two layers of ink proved to be especially challenging due to having to realign the screen for the second layer and waiting for the first layer to dry. This required precision and created room for error. Participants also remarked that the process felt standardized and repetitive after a while, with some even not looking at the instructions at certain points. One suggestion for improvement was to use a wider squeegee that fits the width of the screen, which would make the process more efficient. In the end, all participants agreed that the screen-printing hardware is useful in real-world applications.
Reflecting on the participants' comments and answers to our questions, we feel that creating an automated screen printer would be beneficial and address the issues that our participants faced. With automated screen printing, the amount of labor done to efficiently produce a product would be eliminated as the process could be done by a device that would just need to be started. The printer would release only the amount of ink that is needed, allowing for an even layer and improving precision. Additionally, the printer would align the screens perfectly each time, eliminating the struggle our participants had when manually printing. By including a drying station on our automated screen printer, we can also decrease the drying time of the screens and accelerate the process of printing each layer.
We built upon a manual screen printing setup, the Vevor 4 Color 4 Station Silk Screening Screenprint Press Screen Printing Machine, to create a small scale automated screen printer. For the creation of this printer there were 3 main areas of movement we needed to consider:
Rotational Motion of the Arms Around the Base
The first area of movement we chose to focus on was the central rotational movement of the arms. The biggest challenge we had to work around was that the rotational axle of the screen printer we purchased was attached internally to the base of the printer, and unable to be removed. To power this motion, a Nema 17 stepper motor was chosen. Stepper motors operate off of electrical impulses that get inputted to the desired rate of rotation. Each impulse facilitates a partial rotation of the shaft of the motor known as a “step” allowing for precise control over the desired amount of rotation. We had to create a way to attach the shaft of our stepper motor to the existing axle, while also connecting the shaft to the base to allow it to rotate as the shaft rotates. To accomplish this three pieces were manufactured. The first piece was a machined square piece of steel that was welded onto the existing bolt in the center of the base that the spindle rotates around. Along with being welded to the base, a universal mounting hub was also welded to the square providing a secure way to fasten the stepper motor shaft to the axle. All welding was done using tungsten inert gas (TIG) welding. TIG welding was the preferred method due its ability for more control and precision with working in tighter spaces, such as in between the bolt and the mounting hub.
To allow the base to rotate around the motor the second piece was manufactured, a 3D printed structure was created using computer-aided design (CAD) software. The overall design was created to sit on top of the stepper motor and screw into pre-existing holes in the screen printer base. This structure was designed to not only connect the stepper motor with the base, but also serve as the structural support for the aluminum extrusion that was used to hold the stepper motors for the reciprocative motion of the screen printer arms. Finally, the third piece designed was a small support collar created using CAD software that went around the shaft, and sat in between the top of the universal mounting hub and the top of the stepper motor. The main function of this piece was to add additional support to the stepper motor when taking into consideration the weight of the aluminum extrusion tower, and the motors that would be attached to it, sitting on top. Without this support piece there would be excess force on the motor shaft that could damage the motor and decrease performance.
Figure 15: Labeled diagram of automated screen printer base
Pivoting Motion of the Arms
The second area of movement we chose to focus on was the reciprocating motion of the arms. When the arms are in the downward position the base is unable to rotate because two arms always get restrained by a channel when lowered. This meant we needed to come up with a way to raise and lower the arms at the appropriate times, depending on what station was necessary. To accomplish this a pulley system was created, and the pulleys were designed using CAD. The holes in the pulley were designed so the central hole fit over the shaft of the stepper motor, and the four surrounding holes were used to screw into a universal mounting hub that was attached to the shaft as well. This secured the pulley to the shaft and made sure it would spin with the motor. Kevlar cord was used to attach the top of the arm to the pulley because of its high tensile strength. One end of the string was attached to the arm by wrapping it around a pre-existing bolt and securing it with a nut, the other end of the string was fastened to the pulley and wrapped tightly until it was taught and could support the weight of the arm. Due to the arms being composed of solid steel and the addition of the linear motion setup with the roller, the force required to lift one from a completely lowered position was too much for one of our stepper motors. The best solution we were able to come up with was connecting two of the steppers motors to one of the arms. In doing so, the motors still struggled to lift the screen from its completely lowered position, but had no issue in supporting the weight of the arm and raising it past about an angle of about thirty degrees. In future designs we hope to solve the problem of being able to raise the arm by using a stronger motor, creating a different pulley system, or adding springs to the arm to help reduce the loads.
Figure 16: Labeled diagram of automated screen printer
Linear Motion of the Ink Roller
To achieve the rolling motion of our printing roller, a linear motion setup was created and attached on the frames of the screens placed into the screen printer. A linear motion system was chosen because of the amount of control it offers for factors including the speed and the path of the roller, and the simplicity in the materials necessary. To accomplish this movement a stepper motor with a lead screw that ran along the entire length of the screen was chosen to be used as the motor for this motion. To keep the movement of the roller running parallel along the screen, a long steel shaft about the length of the screen was placed on the side of the screen opposite to the motor. The roller was custom made using a 3 inch diameter PVC pipe that was wrapped in a layer of thin rubber and adhered using spray adhesive. Two end caps were created for the PVC pipe that fit a bearing and allowed a horizontal steel shaft to fit through the center of the roller. Using the bearings allowed the roller to spin freely around the steel shaft. There were four 3D printed components designed that were necessary for the linear motion. The first was the end support pieces that hold the lead screw and the parallel steel shaft in place. They also serve to keep these two elements parallel to maintain the ability for the roller to move. The second was a holder for the stepper motor to hold it in place even with the reciprocative motion of the arm. This piece was designed to screw into a pre-existing hole already in the screen printer arm. The final two pieces were created to be able to connect the horizontal shaft that the roller rests on to the lead screw and the other steel shaft. They needed to be designed slightly different from each other due to the way they each needed to be attached. For the lead screw, the 3D printed piece was necessary to screw into the mounting hub that came with the motor to facilitate the motion of the roller along the lead screw. Similarly, the 3D printed piece used to attach the horizontal shaft to the other steel shaft was designed to fit over a linear ball bearing on the parallel steel shaft. This allows for the horizontal shaft to move smoothly along the parallel shaft while the lead screw drives the force.
Figure 17: Labeled diagram of automated screen printer linear motion design
Round 1:
The team had Eileen Bushnell, a professor in screen printing from the Rochester Institute of Technology, come into the lab to provide feedback on the functionality of the screen printing elements themselves of the printer, as that is where her expertise lies. Overall she was very excited to see an automated screen printer like this because she has only ever seen the large industrial sized ones. What stuck out to her the most was that we were using a roller instead of a squeegee which is traditionally used in screen printing. We explained to her that for our design purposes a roller served better functionality than attempting to automate a squeegee to operate in the same way using one manually would operate. After hearing our concerns with using a squeegee she agreed and told us that using a roller was a very clever way to get around this. When we brought up our concerns about the ink drying on the screen and interfering with the pattern, she circled back to the roller and mentioned how a large reason ink ends up drying on the screen is because the friction applied while screen printing with a squeegee results in heating up the ink slightly and lessens the drying time. Because we are not using squeegees, she predicts that there will be less friction on the ink and that will help prevent the ink from drying as quickly. As we were modeling the automated movement, she noticed that the downward motion of the arms was generating enough wind to potentially blow a piece of paper out of position on the platform it is laying on. Our original solution for this was to just tape the paper to the platform; however, Eileen recommended we get registration tabs. These are little plastic tabs that attach to the platform and hook the edges of the paper to keep it in place. This not only solves the risk of the paper being accidentally knocked out of place, but they also ensure every piece of paper will be printed in the exact same orientation. Along with this, they are reusable and remove the need of having to use large amounts of tape. For the fans attached to the drying station arm on our screen printer, Eileen recommended we use a hair dryer or heat gun instead because the heating aspect greatly reduces the amount of time necessary to dry the ink.
Round 2:
For a second round of feedback the team had Ines Drissi Qeytoni, the Lab & Safety Manager of the 2023 University of Rochester iGEM Team. During her time on the iGEM Team, Ines had put a lot of her efforts into the hardware aspect of their team’s project, a dual-channel 3D bioprinter. Ines had the most input on the reciprocating motion of the arms as this is the part of the automation we struggled with the most. Her biggest piece of advice was thinking about our weight distribution from the base of the arm to the end of the screen, because this plays a major role in how easy it will be for the motor to generate enough force to be able to lift the arm. Along with this, her recommendation was to think about switching the current steel arm out for something of a lighter material, for example by 3D printing it. This way, we would be able to customize the arms in CAD to sit exactly like the current ones, with the advantage that they would be made of a much lighter material. Currently, the roller is also contributing a large portion of the weight on the arm. Some future improvements for this could include using a lighter weight plastic other than PVC, as well as using a thinner piece of rubber to wrap the PVC in.
Future Improvements
Currently the base, the arms, and the roller are automated. For future improvements we would like to automate additional aspects of the screen printer as listed below, and continue optimizing the current automated aspects.
Software
