Design of Methods and Fabrication of MEMS Endoscope Mason Adams, Martin Feldman Ph.D, Sean Hou

School of Electrical Engineering and Computer Science, Division of Electrical Engineering, LSU Electronic Material and Device Laboratory, LSU, Baton Rouge, LA

ABSTRACT Micro-Electro-Mechanical Systems (MEMS) will play a major role in the future of surgery. An important area of research, and a primary use of MEMS probes, is MEMS endoscopy. The microfabrication of scanning mirrors and actuators in these probes is being used to fabricate a narrow 2mm diameter endoscope. An endoscope of this size allows improved navigability, cost effectiveness, and, more importantly, early detection of precancerous and cancerous conditions of the pancreatic duct and common bile duct. An important property of this particular MEMS endoscope is that it is forward-facing, allowing for a more natural field of vision for the surgeon. The objective of the research work was to design and engineer the MEMS endoscope with all of its functioning components. The steps in the MEMS endoscope building process are numerous, and each must be completed with the utmost precision. First, a large, 10mm2 MEMS chip was processed and patterned by a commercial semiconductor fabrication line, or “Fab line.” The large chip consisted of nine small, 2mm2 subchips, three of which were useable. Through the use of Potassium Hydroxide (KOH 30%) etching and circular sawing, these three useable subchips were separated. The KOH etching process dissolves silicon, but will not dissolve an overlaid nitride pattern. Once sawed or “diced,” the subchips measure approximately 1.84mm2. After sanding the chips to thin them, an optical fiber is inserted in a small groove created by the KOH etch. Before the fiber is laid, the chip must be etched with Hydrogen Fluoride (HF 48%) in order to release the mirror from the substrate. After stitch bonding gold wires to the subchip, the mirror may be lifted via probes. After assembly inside the stainless steel tubing, the endoscope is ready for testing. Many problems were encountered in each of the processing steps. The major accomplishments were the construction of an apparatus that would safely and successfully harbor the MEMS chip for KOH etching, the discovery of the correct method of sawing and sanding the MEMS chip, and acquiring an optimal technique for releasing the mirrors from their protective coating on the subchips. Every problem encountered thus far has been resolved. The work on the MEMS chip is ongoing, and I am confident that my work in designing processes and engineering parts will guide future endeavors and provide a clearer picture for work on MEMS endoscopes.

BACKGROUND

RESULTS

MATERIALS AND METHODS

ACKNOWLEDGEMENTS

CONCLUSIONS

KOH Etching the Chip: Etching creates a groove for the future placement of the optic fiber. 1. Heat KOH in beaker to 70°C. 2. Submerge apparatus in KOH. 3. Seal beaker with foil. 4. Insert thermometer. 5. Monitor temperature and etch for ~20 hours.

Sawing the Chip: Dicing is necessary to isolate the subchips of interest. 1. Hot glue whole chip to microscope slide. 2. Load microscope slide into micrometer arm

with rubber band to apply consistent force. 3. Use vernier scale of micrometer to precisely

measure distances for sawing. 4. Turn saw on and slowly lower the chip onto

blade. 5. Saw for ~1 hour, then measure and cut again. 6. After sawing is complete, use heat to separate

the subchips from the microscope slide.

OBJECTIVE The objective of the research work was to design and engineer the MEMS endoscope with all of its functioning components.

This work was orchestrated by Dr. Martin Feldman PhD of LSU, Dr. Mandi Lopez DVM, MS, PhD of LSUSVM, and Dr. Vinod Dasa MD of LSUHSC. A special thanks to Sean Hou and the rest of the staff at the EMDL . I also acknowledge the NIH and the Research and Development Program of the Louisiana Board of Regents which partially funded this project through grants 1R21CA1390993 and LEQSF(2006-09)-RD-A-05, respectively.

Releasing the Mirror with HF: HF dissolves the oxidized layer over the mirror allowing for operation. 1. Clean chip in acetone. 2. Place chip in curved apparatus and

secure with rubber bands. 3. Submerge chip in HF for 25

minutes. 4. Extract chip and place in CO2

dryer to remove stiction.

Sanding the Chip: The chip’s V-groove must be thinned in order to properly accommodate the optic fiber. 1. Hot glue subchip to a microscope slide

alongside two pieces of silicon wafer. 2. Mount slide onto a hinge of the sanding

machine. 3. Magnetically swing hinge toward the

spinning sand paper. 4. Sand chip to approximately the same

thickness of the silicon wafer (550μm).

The most efficient and durable design for a KOH etch apparatus is the clear PVC. To make further progress, more work needs to be done with the Teflon wrapped PVC. This design provides the most resistance to mirror indentation into the PVC. Also, more tests must performed to conclude the optimal amount of time necessary for the chip to be submerged in HF to produce a successful mirror release. Efforts to resolve these issues are ongoing.

Small scale mirrors are a common use of MEMS. The mirrors have many uses, but the one used in this research was for a 2mm in diameter endoscope. In most endoscopes of this size, the mirror is positioned parallel to the substrate of the endoscope. However, this particular endoscope has a mirror that, when actuated, stands vertically, or, perpendicular to the substrate of the endoscope. This allows for a forward-facing field of vision. An optical fiber is placed in a V-groove that has been etched on the backside of the MEMS chip by KOH.

Fig. 1: Drawing of assembled MEMS endoscope.

Fig. 2: Entire MEMS chip. Circled in red is one of the three useable subchips with a square mirror.

Fig. 3: KOH bath setup.

Fig. 4: Dimensions for saw cuts. Note: 1 mil = .001 inches = 25.4 μm.

Fig. 5: Sanded V-groove, 140 μm.

DESIGN OF KOH ETCHING APPARATUS Clear PVC was chosen as the material to use for the KOH etch apparatus. It allowed a clear view of the patterned side of the chip while etching, and also did not become malleable or fall apart at 70oC. Two layers of PVC cement were applied and cured in an oven sequentially. The first layer was built to the height of the chip, and the second created a seal around the chip.

Fig. 6: Test mirror released after HF.

Fig. 7: Visible pattern in KOH.

KOH Etch Design Outcome

Feldman Apparatus Warped at 70oC.

PVC Knockout Plug Floats / Warped at 70oC.

Rubber Gasket & Vaseline KOH dissolves Vaseline.

Quartered PVC Warped at 70oC.

Clear PVC Successful.

Fig. 9: Dr. Feldman’s original apparatus design. Fig. 10: Warped Feldman Apparatus.

HF Mirror Release Design Outcome

Curved PVC Mirror arm break and indentation.

Standing Clear Curved PVC Mirror arm break and indentation.

Teflon Wrapped PVC Mirror arm break.

Fig. 8: KOH Etch Design and Outcome table.

Fig. 11: HF Mirror Release Design and Outcome table.

Fig. 12: Mirror with broken arm after HF. Fig. 13: Intact mirror embedded in PVC.