BENCHTOP DIP COATING DEVICE

During the summer of 2024, I interned at a membrane engineering startup called Osmoses, Inc. By coating their hollow fiber membranes with their own unique polymer, the team at Osmoses were able to develop membrane modules: a unit of membrane fibers bundled together in a tube complete with an inlet and an outlet.  These membrane modules were made to perform gas separation. Gas separation involves feeding a mixture of gases such as hydrogen and methane into a membrane module. The modules serve as a filter for the mixture, allowing, for example, hydrogen to be collected through its outlet and dispensing methane as waste. The efficacy of these modules are measured through a parameter called selectivity, a unit-less quantity that measured how much hydrogen was separated from the hydrogen-methane gas mixture. In order to achieve this selectivity, the fibers had to be evenly coated with the unique polymer before bunching them into modules.

THE PROBLEM:
At the time, the goal for selectivity was 150. However, previous coating methods such as in-situ coating failed. There was another coating method that involved manual hand coating which reached a selectivity around 11. Throughout my time there, my main project was developing a benchtop dip coating device that could improve the membrane coating process. To be coated, the 20 cm long fiber would have to move in and out of a coating solution.


 

 

 

 

 

 

 

 

 

 

The goals of this device were:

• Evenly coat fibers to maximize selectivity.  This meant that throughout the entire coating process, the fiber had to be taut (remain in tension). This proved to be a difficult challenge later on as the fibers used were brittle.

• Minimize costs. Existing dip coating devices cost hundreds of dollars so we would ideally want to manufacture something cheaper than that.

• Coating process could be controlled electronically. This meant that someone could operate the movement of the fiber with the press of a button.

• Allows multiple fibers to be coated at once.
  
   

THE FIRST PROTOTYPE:

With this in mind, it was time to start designing. With the conceptual design, I decided to use a lead screw stepper motor system to drive the motion that would allow the fiber to move in and out of the solution. I also chose to use a motor controlling device to allow the user to control the motion. This consists of a potentiometer, buttons and a motor driver. This allowed the user to change the direction of the motion and the speed of the motion. The final aspect of the design was to design and manufacture custom parts that would secure the fiber and hold the coating solution, which was 20 ml initially. I initially decided to design an aluminum bent sheet metal part that would attach to the lead screw nut. This sheet metal part had a long thin platform where the fiber could be attached. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Before manufacturing, there were a couple of details to keep in mind. First of all, the materials had to be compatible with the coating solvents 1, 3 - dioxolane and diethyl ketone, which were mixed with the coating polymer. Therefore, I had to research materials that were corrosive-resistant to these solvents. Secondly, the weight on the bent sheet metal part had to be minimized. This is because the greater the weight, the more torque likely to be generated which would make it difficult for the membrane to be coated evenly. Most of the manufacturing also had to be done in-house in the Engine building to reduce costs. 

To manufacture this device, an 8020 aluminum extrusion was cut and mounted to two aluminum plates which were waterjet cut. One of these plates mounted the device while the other one mounted the stepper motor. In order for the lead screw mechanism to function, the nut rotating around the lead screw had to be constrained horizontally. This was achieved by 3D printing a nylon "connector" that would attach the nut to the aluminum extrusion. This nylon connector could freely slide along the aluminum extrusion, allowing the nut to translate vertically. The bent sheet metal part was manufactured through SendCutSend sheet metal bending services. The coating solution tray was manufactured by CNC Milling. Small aluminum clamps were also waterjet to secure the fiber onto the tray. 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

The device was manufactured and then tested by attaching a fiber to the tray and dipping it and out of the solution. In terms of what was accomplished, a functional device was achieved which would allow the user to control its motion. The fiber was also able to be attached and coated. Unfortunately, in terms of what needed to be improved on, while it was possible to attach the fiber, this was difficult and took a long time. Sometimes the brittle fibers would break while attaching it on the tray due to the sharp edges of the aluminum part. The fiber tray was a bit heavy, which lead to a jiggling motion. Furthermore, only one fiber could be coated at once so the device had to be scaled up so that 10 fibers could be coated at once (six fibers were necessary to create one membrane module) and the coating solution provided would increase to 100 ml as a result of that. 
​

THE SECOND PROTOTYPE:

After discussing with various coworkers, we determined that we were going to replace the bent sheet metal part with teflon which was lightweight and softer for the fibers to attach on. This was waterjet and then assembled to the nylon connector using spacers and countersink screws.  The teflon had thin slits to wrap the fiber around and attach it. The fiber could then be further secured using paper clips. However, the potential downside to teflon was its lack of rigidity compared to aluminum which means there could be bending due to uneven forces being applied to it. Since a 100 ml coating was used, a new tray had to be used. Fortunately, I was able to purchase a solution tray that was compatible with this. After a couple trips to the machine shop, the new assembly was completed.

 

 

 

 

 

 

 

 

In terms of what was accomplished, the teflon allowed up to ten fibers to be coated at once. Unfortunately, the bending predictions came true and the teflon bent at times which led to uneven coating. Since the middle of the teflon sheet was secured to the nylon connector, this led to an uneven weight distribution, causing the ends of the teflon tray to bend downwards. The device was also getting difficult to transport due to the amount of wires used so I also had to consider creating housing for the electronics.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

THE THIRD PROTOTYPE:

To resolve the bending issue, the nylon connector was enlarged horizontally to allow four countersink screws to attach to it instead of two. More holes were made in the teflon so that it the center was secured using four countersink screws. This reduced the uneven weight distribution on the teflon, reducing bending. Housing was also developed for the electronics through 3D printing. The photo of the final prototype could not be found so you will just have to take my word for it.

RESULTS:

To verify the efficacy of the device, testing was done by creating membrane modules from membranes that were dip coated using the benchtop device. By the end of the summer, the selectivity values had reached a peak of 36, which was about a 300% improvement from previous dip coating methods. To get to 150, other aspects of the coating process will have to be optimized such as the coating time, increasing coating layers, and sealing potential coating defects. For now, the benchtop device had shown to improve the coating process.

LESSONS LEARNED:

The first lesson learned came from overseeing the process of seeing a product being developed. From conceptual design to testing, I was involved in every step of the way. Even though it was nice to be able to see the final result, I gained a deeper understanding of the process of developing a product. Another lesson learned was how to work cross-functionally. Most of my coworkers had a chemical engineering background and prior to this experience, I had no knowledge in membrane engineering. I had to learn to speak their language by reading textbooks and asking questions whenever I had a hint of doubt. At the same time, I also learned to communicate my "mechanical engineering language" to my coworkers that allowed them to understand and provide useful feedback and ideas.