Craig L Christianson and Dr. Aaron Hawkins, Electrical and Computer Engineering

Purpose

The purpose of this project is to build a microfluidic channel using the planar thin film method while increasing the cross-sectional area of the channel by extending the base of the channel down into the substrate.

Procedure

The base enlarged channel procedure is an adaptation of a method currently used by the Electrical Engineering and Chemistry departments at Brigham Young University. The substrate is a four inch borofloat glass wafer. These wafers are then processed in a cleanroom environment. The wafers must be thoroughly cleaned before they can be processed because they come packaged in brown paper, which is a very dirty material from a cleanroom standpoint.

First we put the wafer in water mixed with a critical-cleaning detergent and scrub the wafer with a foam swab to remove the larger particles from off the wafer. To remove the smaller size particles, we spin the wafer at about 1500 rpm and spray the wafer three times with acetone followed by isopropyl alcohol. To ensure that there is nothing else left on the wafer, we soak the wafer in sulfuric acid and hydrogen peroxide.

Once the wafer is clean, we can begin processing it. The wafer is covered with AZ nLOF 2020 negative photoresist, exposed, and developed so that places on the wafers where the microfluidic channels will be built are left uncovered. The exposed glass is then etched to the desired depth enlarging what will be the base of the channel. Silicon Dioxide, aluminum, and photoresist layers are then added to the top of the wafer using plasma enhanced chemical vapor deposition, thermal evaporation, and spinning respectively. The photoresist is exposed and developed, and the aluminum is etched leaving aluminum and photoresist as sacrificial layers over the etched bases of the channels. Plasma enhanced chemical vapor deposition is again used to deposit an additional 3μm of silicon dioxide on top of the wafer. The ends of the channel are then etched away to expose the sacrificial layers. These layers are then etched away leaving a hollow channel.

Problems

Several problems were encountered during this project, but the two most critical problems were that the substrate did not etch as planned and the top oxide layer over the channels broke off while the sacrificial layers were being etched away.

Substrate Etching

Two methods were tried to etch the substrate. The first method used a buffered oxide etchant to etch into the substrate. The second method used Tetrafluoromethane gas (CF4) in a reactive ion etch machine to etch into the substrate. Neither method was able to etch to the desired depth of 3μm. Before the depth would reach 3μm, parts of the photoresist would also etch away causing the wafer to etch in the wrong places. Isotropic etching was also a problem in both cases. This caused the base of the channels to be significantly wider than the top of the channels.

Top Oxide Layer

The top oxide layer fell away during the etching of the sacrificial cores. We have found three possible reasons why this happened.

First, the oxide layers were most likely thinner than the expected 3μm. During the project, our lab installed a new chemical vapor deposition machine. Though the throttle pressure, percent power, and gas flow rates were the same, the deposition rates throughout the chamber were not as anticipated. This caused the oxide layer to be much thinner than our target thickness.

The second problem was poor conformality. For the same reasons as the thin top oxide layer, the conformality of the oxide layer was much worse than anticipated. This probably caused cracking where the top oxide layer met the bottom oxide layer which could have caused the tops of the channels to break off.

Third, due to the isotropic etching discussed above, the top oxide layer did not connect with the bottom oxide layer where planned. This, combined with the poor conformality was certainly a cause for the top oxide layer coming off.

Possible Solutions

The substrate etching problem could possibly be solved by finding a way to improve the adhesion between the photoresist and the substrate making the substrate etch more anisotropic. The photoresist could also be spun on thicker or applied several times to ensure that only the intended areas of the wafer are etched.

The problems with the top oxide layer could be solved by performing a better study of the deposition rates throughout the chamber of the chemical vapor deposition machine and finding a recipe with good conformality.

Conclusion

Although this project did not have the results we hoped for, the problems we have come up against led us to the discovery of possible ways to overcome weaknesses not only in base enlarged channels, but for all channels currently being researched at Brigham Young University. This will lead to better channel yields on a broader spectrum of microfluidics. With more research, the problems will surely be resolved and this project will be able to continue.