Stage 2: Modified Microchannel Test Design

Stage 2: Modified Microchannel Test Design

Fabrication Steps: SU-8 Molds

Using the masks created by the microchannels team, SU-8 molds were created for fabrication of the PDMS microchannels. To create the SU-8 molds, a bare 3" silicon wafer was placed on the spinner and 2/3 of the wafer was covered with the SU-8 liquid. Next, the spinner was programmed using a recipe formulated to create a layer of SU-8 220 microns in height.

The spinning was complete within 30 seconds, and the SU-8 was then left to pre-bake on a hotplate at 95 C for 100 minutes. Once the pre-bake was complete the wafer was left to cool at room temperature for 30 minutes. The wafer was cooled slowly to stop the formation of cracks and defects in the SU-8 mold. At this point in the experiment, we were working under the assumption that cracks in the SU-8 mold would be detrimental to the final PDMS product.

Once the wafer was cooled to room temperature, it was placed in the aligner to be aligned with the mask. The exposure dose used was 900 mJ/cm2 and using an intensity meter, we measured the intensity of the light to be 26.6 mW/cm2. Dividing the dose by the intensity, we calculated the exposure time to be 33.7 seconds.

After the SU-8 had been exposed for 33.7 seconds, the wafer was placed on the hotplate to bake for 30 minutes at 95 C and then left to cool for an additional 30 minutes. Next, the wafer was put into a beaker filled with SU-8 developer and placed on a rocking table to develop. The wafer was developed for 22 minutes and then rinsed in fresh SU-8 developer and left to dry. This same process was repeated for all of the molds until a mold for each PDMS layer was fabricated.

Analysis of the SU-8 molds using the optical microscope revealed a substantial amount of cracks that can be seen in Figure 1. However, we continued with fabrication of the device confident that the cracks were small enough so that the effect on the final device would be negligible.

Figure 1: Cracks in an SU-8 mold. This is a reservoir region.

Fabrication Steps: PDMS Microchannels

Fabrication of the PDMS microchannel layers began by placing the previously fabricated SU-8 molds in an SDS (soap) solution and drying. The PDMS was then mixed with curing agent at a 10:1 weight ratio and poured over the SU-8 mold. The recipe used for PDMS was intended to create layers of 130 microns in height. However, similar to the SU-8 molds, time constraints prevented an accurate measurement from being obtained. Once the spinning was complete, the wafer was placed in the furnace to bake for 2 hours at 70 C.

Two sets of identical PDMS layers were fabricated, and assembly of the micro channel layers took place on two separate occasions. In both assembly trials, removing the PDMS microchannel layers from the SU-8 mold and aligning the layers in the proper sequence proved to be the most difficult part of the device fabrication. The microchannels were released from the mold by hand with the aid of razor blades and tweezers. Methanol was used as a release agent to allow the PDMS channel layers to slide easily off the mold. Aligning the layers was extremely difficult because the PDMS layers had the tendency to stick to each other when they were not coated with methanol. Alignment was further complicated because the layers became extremely slick whenever the methanol was used. Another problem arising from the fabrication of the PDMS device was the formation wrinkles and air pockets between the layers. Eventually, the layers were crudely assembled although it was easy to observe that some of the features were not properly aligned. In addition, the interconnect layer did not provide a connection between the top and bottom layers.

During the second trial, fewer defects were observed within the layers and there were no significant problems with air bubbles or delamination. This is due to the fact that during the second assembly trial, addition of each PDMS layer to the previous layer was followed by compression of the layers with a metal rolling pin. This rolling process removed excess moisture and air from the layers, resulting in fewer defects.

Experimental Trials: Phase 2

This project involves the development of a microfluidics device which would function to meet our goals, and which could be fabricated by us. It is very important to consider the manufacturing process when working towards a final proposed design. This is important because manufacturing constraints are the largest limitation to our design. We were able to design, fabricate, and test two prototypes of the phase two design. Both prototypes met some of our goals, and fell short of achieving others.

Our testing set up consisted of a (size?) syringe to inject liquid into the inputs of our device, and water colored with food coloring. We found that lighter shades of orange, red, and green showed up the best against the silicon wafer. For both prototypes, we injected liquid into each of the 5 inputs, and recorded the result. We also tried injecting liquid into the outputs and observed the results.

The first experimental prototype had several large problems, making it very difficult to test. The PDMS layers were thicker than we had anticipated, and therefore the interconnects did not transfer from the mold to the PDMS layer. The result was that our layers were not connected to each other, and the top layer sealed the channels from the outside. This limited us to testing the channels that were oriented only in the horizontal direction.

This problem also made it difficult to inject liquid into the inputs, since the top layer of PDMS sealed them off. We solved this problem by poking through the top layer with the syringe, and injecting the liquid into the input, under the top layer. This technique worked, but had limitations. We observed no capillary action in the channels, meaning that the liquid would only move through them with applied pressure. This required a seal between the syringe and the top layer of PDMS, otherwise the injected liquid would flow out around the needle and not into the channels. This problem was corrected by sealing the channel with the needle in it using applied pressure from a finger.

Once we were able to inject liquid into the channels we observed some success in moving the liquid through from input to output. Unfortunately, there were many air bubbles between the PDMS layers, causing the liquid to spread out and fill the air bubble instead of staying in the channel. Many of the problems we encountered during fabrication and experimentation were corrected for the second prototype.

The second prototype had many improvements over the first trial. Each layer was the correct thickness, allowing for interconnects between layers. The layers were aligned with good accuracy, meaning the interconnects connected the channels on both the top and bottom, and the inputs and outputs were open on the top layer. There were no air bubbles between PDMS layers.

With the more accurate fabrication of the design we were able to achieve several of our goals in testing. We successfully got liquid to flow in all five channels using applied pressure from the syringe. We were able to push liquid all the way through two of the five channels. We also observed two colors of liquid one on top of the other, as designed, proving that our channels were accurately fabricated.

We observed some of the same problems that we had encountered with the testing of the first prototype. We observed no capillary action, so we had to jam the needle into the end of each channel to obtain a seal. In doing this, the layers sometimes delaminated near the end of the channel. We once again corrected this problem by applying pressure behind the needle opening with a finger. Our biggest problem with getting fluid to flow through the channels occurred at the interconnects. We could not get fluid to flow vertically in any of the trials. In the channels that included vertical interconnects, the liquid would stop flowing when it reached the interconnect. To deal with this we applied more pressure to the fluid and the layers delaminated around the interconnect. This problem could have been caused by either our design or the fabrication of our prototype.

Future Work

The majority of the processing difficulties were associated with PDMS fluidic channel alignment and thickness of the PDMS and SU-8 layers. Accurate measurements of the PDMS channel layer and SU-8 heights would be extremely useful not only for device fabrication, but also to verify the spin coating recipes and determine if modifications to the recipe are required. A new alignment technique, perhaps making use of the mask aligner was suggested and is highly desirable to achieve greater layer alignment accuracy. However, testing of such a technique was not possible due to time constraints associated with the project.

Stage 3: Pressure Actuated Valve Test Design

The entire pressure actuated valve device was intended to be fabricated on a Pyrex wafer. However, the Pyrex wafer was very thin (500 μm) and when we attempted to drill holes for inputs into the microchannel using a diamond tip blade, a substantial amount of cracks resulted around these holes. Consequently, the Pyrex wafer completely cracked during the subsequent fabrication stages, making design of the final device impossible. To overcome the problem of the Pyrex wafer, the pressure actuated valve device was fabricated using a bare Pyrex wafer with no holes drilled for channel inputs. While the lack of inputs prevents any testing of the device, it provides a reference for future work.

[INCLUDE STAGE 3 PICTURES HERE]