«by Jong Moon Park A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Electrical Engineering) ...»
A PIEZOELECTRICALLY ACTUATED CRYOGENIC MICROVALVE
WITH INTEGRATED SENSORS
Jong Moon Park
A dissertation submitted in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
in The University of Michigan
Professor Yogesh B. Gianchandani, Chair
Professor Khalil Najafi
Professor Kensall D. Wise Associate Professor Luis P. Bernal © Jong Moon Park All rights reserved 2009 To my parents ii
ACKNOWLEDGEMENTSI would like to give my sincere gratitude to my advisor, Professor Yogesh Gianchandani, for giving me the opportunity to work on the exciting project and also for his teaching, guidance, and support. I also thank my committee members, Prof. Wise, Prof. Najafi, and Prof. Bernal for their interest and valuable insights in various aspects of my work.
I also would like to express my thanks to Prof. Klein and Prof. Nellis and their students, Ryan Taylor, Tyler Brosten, Kristian Rasmussen at University of Wisconsin at Madison for their contribution in numerical modeling related to this work.
I am thankful to many former Ph.D. students in my group, Dr. Senol Mutlu, Dr.
Kabir Udeshi, Dr. Kenichi Takahata, Dr. Bhaskar Mitra and Dr. Amar Basu, who have helped me to get a good start on my research. I especially thank my colleague, Allan Evans for his efforts in carrying out the project together. I also thank all my friends and
colleagues in my group and SSEL for their companionship, support, and encouragement:
Weibin, Tao, Mark, Christine, Scott Green, Scott Wright, Naveen, Karthik, Jae Yoong, Tzeno, Andy, Razi, Sang-Hyun, Seow Yuen, Ruba, Jay, SangWon, Hanseup, Sangwoo, Junseok, Bercu, Gayatri, Kyusuk, and many other.
I would like to acknowledge Lurie Nanofabrication Facility (LNF) staff members for keeping the cleanroom safe and enjoyable place to work.
iii Finally, I thank my parents, sister and my girlfriend, Jenny, for their endless love and support.
This work was supported primarily by NASA under award number NNA05CP85G.
TABLE OF CONTENTSDEDICATION
LIST OF FIGURES
LIST OF TABLES
CHAPTER 1 INTRODUCTION
1.1 Distributed Cooling System
1.2 Operating Requirements for the Valve
1.3 Overview of Microvalves
1.3.1 Electromagnetic actuation
1.3.2 Electrostatic actuation
1.3.3 Piezoelectric actuation
1.3.4 Thermal actuation
CHAPTER 2 MICROVALVE FOR CRYOGENIC APPLICATIONS
2.1 Microvalve Design
2.1.1 Selection of valve actuation scheme
2.1.2 Device Concepts and Operation
2.1.3 PZT characterization
2.1.4 Perimeter augmentation
2.2 Modeling of the Microvalve
2.2.1 Flow through the grooves
2.2.2 Flow through the land
2.2.3 The microvalve flow model
2.3 Device Fabrication
2.3.1 SOI Wafer Fabrication
2.3.2 Glass Wafer Fabrication
v 2.3.3 Device Assembly
2.4 Experimental Results
2.4.1 Room Temperature Flow Test
2.4.2 Cryogenic Temperature Flow Test
2.4.3 High Temperature Flow Test
2.4.4 Dynamic Characteristics
2.5 Discussion and Conclusions
CHAPTER 3 MICROVALVE WITH INTEGRATED SENSORS
3.1 Microvalve and Sensor Design
3.2 Device Fabrication
3.3 Experimental Results
3.4.1 Flow Measurements
3.4.2 Pressure Sensor Measurements
3.4.3 RTD Measurements
3.4.4 Stress Effects
CHAPTER 4 VALVE APPLICATIONS
4.1 Liquid Flow Modulation
4.1.1 System Overview
4.1.2. Microvalve Liquid Flow Tests
4.2 Joule-Thomson Cooling System Test
4.2.1 Joule-Thomson Effect
4.2.2 Microvalve Test
4.2.3 MEMS Heat Exchanger
4.2.4 Heat exchanger self cooling test
4.2.5 Self Cooling Test with Heat Exchanger and Valve
CHAPTER 5 CONCLUSION AND FUTURE WORK
5.1 Summary of the Work
5.2 Recommendations for Future Work
Figure 1.1: Illustration of the Single Aperture Far-Infrared Observatory (SAFIR)  2 Figure 1.
2: Temperature dependence of SAFIR sensitivity. Lower vertical values represent better sensitivity. Low temperature is very important to achieve desired sensitivity. (Figure is adapted from .)
Figure 1.3: Illustration of a distributed cooling system concept.
An array of cooling elements, each consisting of a valve, heat exchanger and a sensor, is placed across the structure to be cooled
Figure 1.4: Proposed application of an actively controlled valve.
Microvalves can (a) modulate the flow of coolant in distributed heat exchangers, or (b) be used as expansion valves in Joule-Thomson cryocoolers.
Figure 1.5: Two main configurations using piezoelectric actuation: (a) Stacked PZT and (b) thin film PZT.
Figure 1.6: Flow rate vs.
power consumption for currently reported microvalves using electrostatic, piezoelectric, electromagnetic, bimetallic, shape memory alloy, and thermopneumatic actuations. The target for our valve is to provide flow rates that exceed 1 L/min but still maintain less than 1 mW of power.
Figure 2.1: Microvalve concepts: The valve consists of a ceramic-PZT-Si-glass structure.
A PZT stack actuator moves the silicon plate to open or close the valve.
Figure 2.2: (a) Schematic of the test apparatus for PZT characterization and (b) picture of the assembled test facility
Figure 2.3: Measured effective PZT stack actuation constant as a function of temperature.
Figure 2.4: Measured PZT stack coefficient of thermal expansion as a function of temperature.
Figure 2.5: Microvalve concepts: The valve cross-section of the ceramic-PZT-Si-glass structure is shown on the far left.
Micro groove patterns to increase flow area are depicted in the middle. The recessed glass along with the side view of the Si piece is illustrated on the right
Figure 2.7: Top and side view of the land and groove region with assigned symbols.
30 Figure 2.8: Groove friction factor as a function of the Reynolds number based on the groove hydraulic diameter.
Figure 2.9: Mass balance terms for the control volumes of the numerical solution.
.. 33 Figure 2.10: Simulation results showing the effect of perimeter augmentation............ 34 Figure 2.11: Valve flow model prediction: Room temperature nitrogen flow rate at various inlet pressures
Figure 2.12: Si-glass micromachining process: The buried oxide layer in the SOI wafer acts as an etch stop for DRIE.
A three-step DRIE process is illustrated for the SOI wafer. A glass wafer undergoes a wet etch process and a electrochemical discharge machining process for inlet and outlet hole creation. Finally, the Al metal layer is deposited and patterned in preparation for anodic bonding to prevent bonding of the valve seat to the glass substrate.
Figure 2.13: SEM photographs of grooves on a valve plate fabricated by DRIE. (a):
Valve plate. Part of a flexure is also seen on the left side, (b) and (c): A valve plate is diced to expose the cross-sectional view.
Figure 2.14: Schematics of the ECDM setup for machining glass wafers
Figure 2.15: SEM photograph of a hole in a Pyrex glass wafer formed by ECDM process: (a) Top view and (b) the cross-sectional view
Figure 2.16: Pre-stressed microvalve assembly procedure: (a) First, a PZT stack is attached to the Macor ceramic structure using epoxy.
(b) Then a Si-glass die is bonded at the end of the PZT and Macor structure by an epoxy joint.
During this procedure, the PZT stack is actuated until the epoxy is fully cured, which results in a normally-open valve (c)
Figure 2.17: Completely assembled valve structure.
(a) Two valves (front and back) are shown. (b) Another view of the valve with the Macor header and a US penny
Figure 2.18: Schematic of the room temperature flow measurement test apparatus.
.... 43 Figure 2.19: Flow rates as a function of voltage from experimental results and the analytic model. As the actuation voltage increases, the valve is closed, which results in a decrease in flow rate.
Figure 2.21: Flow rates as a function of actuation voltage measured at room temperature with an inlet pressure of 345 kPaG and an inlet to outlet pressure differential of 34 kPa.
Experimental data are shown with the numerical modeling results. Flow rate measurements present the hysteretic behavior of the PZT actuator
Figure 2.22: Schematic of the test setup for cryogenic flow measurement
Figure 2.23: (a) Measured flow rates as a function of voltage at liquid nitrogen temperature (80 K) compared with (b) the analytic model
Figure 2.24: Normalized flow rate vs.
actuation voltage obtained at cryogenic temperature.
Figure 2.25: Measured flow rates as a function of voltage at 110˚C and 26˚C with inlet pressure of 25 kPa
Figure 2.26: Transient response of the valve measured using a laser Doppler vibrometer with a sampling rate of 1.
28 MHz. The response time for 0 to 100 V actuation is measured to be 700 µs. Some oscillations are evident due to the mechanical properties of PZT.
Figure 2.27: The bandwidth of the valve at room temperature is determined by 100 V square waves applied from 100 Hz to 4.
5 kHz in 100 Hz increments. The displacements are measured and normalized to the DC valve. The dots represent data points with error bars. The best fit curve is also plotted.
The 3 dB bandwidth occurs at around 820 Hz
Figure 3.1: Microvalve concepts: (a) A valve plate suspended by a Si membrane with integrated sensors viewed from the bottom.
The RTD sensor is located on the backside of the wafer and is not visible. (b) A cut away view of an assembled ceramic-PZT-Si-glass valve structure.
Figure 3.2: Si-glass micromachining process: sensors are formed on the device layer of the SOI wafer by various surface micromachining techniques.
Then DRIE is performed to form membrane and groove patterns. A glass wafer undergoes two wet etch steps for a recess and through-hole formation.
Next, the two wafers are anodically bonded and diced.
Figure 3.3: (a) Photograph of the top of the SOI wafer showing the metal contact layer with (b) an expanded view of the platinum RTD.
(c) Wide and (d) expanded views of the circular serpentine groove patterns for perimeter augmentation from the bottom of the SOI wafer. This side bonds to the glass wafer.
Figure 3.5: An assembled valve looking from the glass side with a US penny.
The pressure sensor cavity is connected to the inlet through a passage. The RTD is positioned on the backside of the silicon die
Figure 3.6: Flow rates as a function of actuation voltage measured at room temperature.
Flow measurements of (a) normally-open valve, and (b) normally-closed valve measured at a differential inlet pressure of 52 kPa.
Figure 3.7: Flow characteristic of a normally-open valve at different temperatures.
The measurement was made at 34 kPa differential pressure with 160 kPa absolute inlet pressure. As the temperature decreases, actuation of the PZT has a lesser effect on flow modulation, thus resulting in a higher flow rate at a given actuation voltage. This is mainly due to a degraded piezoelectric performance at lower temperatures. Hysteresis of the PZT actuation is represented as error bars here.
Figure 3.8: Output voltage from the piezoresistive pressure sensor at various differential pressures and temperature.
The linearity of the pressure sensor was still good at low temperatures; however, the offset voltage and sensitivity of the sensor changes with varying temperatures
Figure 3.9: Measured RTD resistance versus temperature change with best fit lines (a) on bulk silicon, and (b) in a packaged device.
The RTD shows bilinear behavior, with a sensitivity of 0.29 %/K above 140k and 0.37 %/K below it.
Figure 3.10: The effect of temperature on the piezoresistive pressure sensor.
(a) The measured sensitivity of the pressure sensor decreases with increasing temperature with 356 ppm/kPa at room temperature. The sensitivity is plotted in both ppm/kPa and ppm/torr. The behavior is dominated by the temperature dependency of the piezoresistance factor. (b) Measured offset voltage also changes with temperature due to thermally induced stress on the membrane
Figure 3.11: Comparison of the coefficient of thermal expansion of Pt, Pyrex glass, and silicon
Figure 4.1: Proposed system schematic of the microvalve controlled portable drug delivery system.
Figure 4.3: Flow rate changes as a function of actuation voltage shown at two different inlet pressure values (Allan Evans ).
Figure 4.4: (a) Temperature-entropy diagram and (b) basic components of a JT cooler equipped with a recuperative heat exchanger.
Figure 4.5: Plot of flow rate as a function of actuation voltage while inlet and outlet pressures are held constant
Figure 4.6: The data in Figure 4.
5 is re-plotted with expected flow rates through a jewel orifice. The differential pressure is maintained at 229 kPa............ 85 Figure 4.7: Flow rates of a microvalve at various differential pressures compared with measured flow rates using 0.015” and 0.010” orifices.