The Iby and Aladar Fleischman Faculty of Engineering

TEL AVIV UNIVERSITY

The Iby and Aladar Fleischman Faculty of Engineering

The Zandman-SlanerSchool of Graduate Studies

Development and Characterization of a Micro-machined Shear-Stress Sensor

A thesis submitted toward the degree of

Master of Science in Mechanical Engineering

by

Youry Borisenkov

This research was carried out in the Department of Fluid Mechanics and Heat Transfer

under the supervision of Prof. Avi Seifert

March 2008

Abstract

Real-time measurement of shear-stress on aerodynamic surfaces is a critical enabling technology in any feedback flow-control system. Shear-stress sensors should be able to provide valuable information about the boundary layer dynamics, spectral content and tendency to develop separations. The development of suitable sensor can lead to a breakthrough in stabilizing Miniature Aerial Vehicles (MAV's), where such sensor can be integrated in a closed-loop control system along with suitable actuators and controllers. Due to the small size of such vehicles, in addition to cost and weight considerations, the sensor should be miniature.

The current research objective is the development and characterization of a micro-machined shear-stress sensor,keeping in mind that the design and packaging should produce minimal disturbance to the flow. The proposed MEMS sensor is based on a direct measurement technique, i.e., a floating element with capacitive sensing by a comb drive and a back side contacts. The shear-stress generated by the flow over the sensor ideally causes a one-dimensional movement of the floating element. This movement causes a displacementof the interdigitated teeth of the comb drive bridges, attached to the floating element and to the anchor of the sensor.The resultingdisplacement is then measured by an electronic circuit designed especially for the sensor. The circuit is able to detect small capacitance changes of the comb-drive in the order of 0.01pF. The concept also allows implementing a closed feedback-loop on the floating element, i.e., balancing the aerodynamic shear-force by the electrostatic force of the comb-drive, thereby zeroing thedisplacement.This increases the sensors accuracy at the cost of a slightly reduced bandwidth.

A theoretical analysis and numerical models weredeveloped characterizing the sensor as an under-damped second-order system with onedegree-of-freedom. The sensor wasfabricated from Silicon in the clean room of TAU.The testing of the working prototypeshowed a good agreement with the theoretical and numerical model and a deflection that is proportional to the boundary layer shear. This novel concept of using comb-drive sensing technique proved to be efficient for measuring shear stress in low speed environment using optical and force rebalancing modes. However, at calibration stage, the sensor response to external flow was shown to be unsuitable due to poor aerodynamic and stiffness design and therefore further modeling is required.

Table of Contents

Abstract

Table of Contents

List of figures

List of tables

Chapter 1. INTRODUCTION

1.1 Outline

1.2 Origin of Wall Shear-Stress

1.3 Motivation For Shear-Stress Measurements and Sensor Specifications

1.4 Shear stress measurement techniques

1.4.1 Conventional Shear-Stress Measurement Techniques

1.4.1.1 Indirect Techniques

1.4.1.2 Direct Techniques

1.4.2 Silicon micromachined sensors

1.4.2.1 Overview on micromachining

1.4.2.2 Micromachined shear stress sensor

1.4.2.2.1 Indirect methods

1.4.2.2.2 Direct Methods

1.5 The proposed sensor:

1.6 Objectives

1.7 Scope

Chapter 2: SENSOR DESIGN

2.1 Detection Principle – Capacity Position Sensing

2.2 Capacity measurement

2.3 Sensor Dimensions

2.4 Electrical connection scheme

2.5 Static and Dynamic Mechanical Analysis

2.5.1. Theoretical Analysis

2.5.2. Numerical analysis

2.5.2.1 Modal analysis

2.5.2.2 Strain analysis

2.5.3 Numerical model

2.5.3.1 Dynamic model

2.5.3.2 Electrostatic model

2.5.3.3 Design of closed-loop

2.6 Theoretical Minimum and Maximum Shear Stress

2.7 Different Sensor Designs and Test Structures

2.7.1 Transverse direction

2.7.2 Sensor arrays

2.7.3 Two μm resolution teeth

2.7.4 Tether length variations

Chapter 3: SENSOR FABRICATION

3.1 Mask design

3.1.1 Mask layers

3.1.2 Alignment marks:

3.1.3 Wafer layout

3.2 Design variations

3.3 Process Flow

3.3.1 Starting Material

3.3.2 Handle Wafer Process

3.3.2.1 Back side electrode formation:

3.3.2.2 Back side release holes and isolation rings formation:

3.3.3 Device layer Process:

3.3.4 Dicing:

3.3.5 Over etch:

3.3.6 HF release:

3.4 Discussion of fabrication issues

Chapter 4: SENSOR TESTING

4.1 Electrical Functionality

4.2 Dynamic & Static Calibrations

4.2.1 Frequency Response

4.2.2 Stiffness

4.2.3 Pull-in Voltage

4.2.4 Discussion

4.3 Sensor calibration

4.4 Packaging

4.5 Other Performance Criteria

4.5.1 Repeatability

4.5.2 Drift

4.5.3 Cross-Axis Sensitivity

4.6 Experiment Setup for Measurements in a Laminar\Turbulent Boundary Layer in a Wind Tunnel

Chapter 5: CONCLUSIONS

Appendix A

First Run Process:

Second Run Process:

Appendix B- Detailed Drawings

Appendix C -Measurement circuit

Appendix D- Hot wire Shear-stress measurement

REFERENCES

List of figures

Fig 1.1: Origin of wall shear-stress. (p.2)

Fig 1.2: Classification of wall shear stress measurements techniques, Haritonidis [4] (p.5)

Fig 1.3: Pressure gradients on floating element (p.8)

Fig 1.4: (a) Schematic illustration of key step in a bulk micromachining process; (b) schematic illustration of the key steps in a surface micromachining process [32] (p.11)

Fig 1.5:Schematic illustration of electrostatic actuation (a) parallel plates; (b) comb drive (wikipedia) (p. 14)

Fig 1.6:Optical sensing principle. [49] (p.14)

Fig 1.7:Direct piezoelectric effect (wikipedia) (p.15)

Fig 1.8: Side-view schematic illustration of differential capacitance floating element sensor developed by Schmidt el al. [44] (p.16)

Fig 1.9: Top-view schematic illustration of an axial piezoresistive floating element sensor developed by Ng el al. and Goldberg et al. [45, 46] (p.16)

Fig 1.10: Schematic illustration of the MEMS-based surface fence [47].(p.17)

Fig. 1.11: Side-view schematic illustration of a differential capacitance surface micromachined floating element sensor developed by Pan et al. and Hyman et al [51].(p.18)

Fig 1.12: The proposed sensor. The cross-hatched regions mark "fixed-end" clamping. (p.19)

Fig 2.1: Scheme of the floating element, indicating the flow direction. (p.22)

Fig 2.2: Sketch that highlights some of the sensor diminutions. (p.25)

Fig 2.3: Sketch that highlights some of the sensor diminutions (Zoom-in area). (p.26)

Fig 2.4: Sketch showing the combdrive fringe effect.(p.26)

Fig 2.5: CAD model of the chip (back-side view).(p.27)

Fig 2.6: CAD model of the chip (front-side view).(p.28)

Fig 2.7:ANSYS presentation of the first mode displacement.(p.30)

Fig 2.8:ANSYS presentation of the second mode displacement.(p.30)

Fig 2.9: ANSYS presentation of the third mode displacement.(p.31)

Fig 2.10: ANSYS presentation of the fourth mode displacement.(p.31)

Fig 2.11: ANSYS presentation of the fifth mode displacement.(p.31)

Fig 2.12: Maximum stress presentation from the ANSYS analysis for the laminar sensor. (p.32)

Fig 2.13: Maximum stress presentation from the ANSYS analysis for the laminar sensor (zoom-in). (p.33)

Fig 2.14:Maximum stress presentation from the ANSYS analysis for the turbulent sensor. (p.33)

Fig 2.15:Simple Couette configuration using two infinite flat plates.(p.34)

Fig 2.16: Turbulent sensor step response, step force of 6.00E-06[N]. (p.36)

Fig 2.17: Laminar sensor step response, step force of 6.00E-06[N]. (p.36)

Fig 2.18: Frequency response(Y- direction movement). (p.37)

Fig 2.19: Guide User Interface for pull in simulation. (p.39)

Fig 2.20: Output figure of the pull in GUI simulation. (p.40)

Fig 2.21: Close loop block diagram.(p.41)

Fig 2.22: Transverse direction sensor. (p.42)

Fig 2.23: Sensor array, orthogonal pair. (p.43)

Fig 2.23:Sensor array, in-line pair. (p.44)

Fig. 3.1:Device layer mask (L-edit). (p.46)

Fig. 3.2:Nominal sensor (L-edit). (p.46)

Fig. 3.3: Electrical connection (L-edit). (p.47)

Fig. 3.4: Sensor structure (L-edit). (p.48)

Fig 3.5:Devise layer features (L-edit). (p.48)

Fig 3.6:Gold layer-front side (L-edit). (p.49)

Fig 3.7:Gold layer pads (L-edit). (p.50)

Fig 3.8:Back side layer (L-edit). (p.50)

Fig 3.9: Single chip back & device layer (L-edit). (p.51)

Fig 3.10:Single chip back side tunnel & device layer (L-edit). (p.52)

Fig 3.11:Single chip back side isolating rings & device layer (L-edit). (p.52)

Fig 3.11: Alignment marks, gold layer (L-edit). (p.53)

Fig 3.12: Alignment marks, device layer (L-edit). (p.54)

Fig 3.13: Alignment marks, back layer (L-edit). (p.54)

Fig. 3.14: Sensor array - Using such sensor will allow us to check the response of the sensor to changes in flow direction. (p.55)

Fig. 3.15: Sensor variation - This variation was designed as parallel plate capacitor sensing instead of comb-drive capacitor. (p.55)

Fig. 3.16: Sensor array - This variation was designed to compare the results and expand the sensitivity of the sensor. (p.55)

Fig. 3.17: Starting material: handle, box and devise layers unprocessed. (p.56)

Fig. 3.18: Die sketch informing cut locations. (p.57)

Fig. 3.19: (i) represents cut (b) of single chip, after photolithography and RIE (oxide) etch; (ii) represents back side view of the chip (p.58)

Fig. 3.20:(i) represents cut (b) of single chip, after 300 µm ICP etch; (ii) represents back side view of the chip (p.58)

Fig. 3.21: (i) represents cut (b) of single chip, after RIE etch; (ii) represents back side view of the chip (p.58)

Fig. 3.22:(i) represents cut (b) of single chip, after gold LPCVD sputtering; (ii) represents back side view of the chip (p.59)

Fig. 3.23:(i) represents cut (b) of single chip, after gold lift-off; (ii) represents back side view of the chip (p.59)

Fig. 3.24:represents cut (b) of single chip, after gold pads protection.(p.60)

Fig. 3.25: (i) represents cut (b) of single chip, after ICP isolation rings etch; (ii) represents back side view of the chip (p.60)

Fig. 3.26: represents cut (a) of single chip, after back side process. (p.60)

Fig. 3.27: (i) represents cut (b) of single chip, after cleaning process; (ii) represents back side view of the chip (p.61)

Fig. 3.28:represents cut (b) of single chip, after front side RIE process. (p.61)

Fig. 3.29:represents cut (b) of single chip, after front side ICP etch. (p.62)

Fig. 3.30: (i) represents cut (b) of single chip, after back side oxide removal by HF; (ii) represents back side view of the chip (p.63)

Fig. 3.31: represents cut (b) of single chip, after front side over etch by ICP. (p.63)

Fig. 3.32:SEM picture showing the stress deformed tether of laminar sensor. (p.64)

Fig. 3.33: represents cut (b) of single chip, after front side oxide removal by HF vapor. (p.64)

Fig. 3.34: SEM picture of the device structure. (p.65)

Fig. 3.35: Combdrive bridges on the device. (p.65)

Fig. 4.1: Probe station. (p.68)

Fig 4.2: Frequency response of turbulent sensor. Operating voltage of ~9V. (p.69)

Fig 4.3: Frequency response of laminarsensor. Operation voltage of ~4V. (p.69)

Fig 4.4: Stiffness calculation of laminarsensor. (p.70)

Fig 4.5: Stiffness calculation of turbulent sensor. (p.71)

Fig 4.6: Force-rebalancing calibration experiment setup. (p.73)

Fig. 4.7: Force-rebalancing calibration experiment setup (zoom-in view). (p.74)

Fig 4.8: Hot wire calibration experiment setup. (p.74)

Fig 4.9: Hot wire calibration experiment setup (zoom-in view). (p.75)

Fig 4.10: Calibration of laminar sensor (Shear-stress vs. Deflection). (p.75)

Fig 4.11:Calibration of laminarsensor (Shear-stress vs. Voltage). (p.76)

Fig 4.12: Calibration of turbulent sensor(Shear-stress vs. Deflection). (p.76)

Fig 4.13: Calibration of turbulentsensor (Shear-stress vs. Voltage). (p.77)

Fig 4.14: Effective applied stress on floatingelement. [96] (p.78)

Fig 4.15: Pressure gradient effect on oursensor (section view sketch). (p.78)

Fig 4.16:Pressure gradient effect on our sensor as it appears thru microscope eyepiece.The angle ofrotation of the floatingelement is expressed with light reflection. (p.79)

Fig 4.17:Pressure gradient effect on our sensor as it appears thrumicroscope eyepiece. The sensordeflected distance is larger than 20 μm. This means thatthe front “Mechanical stopper” dived underthe device layer plane. (p.79)

Fig 4.18:Pressuregradient effect on our sensor as it appears thru microscope eyepiece. The front “Mechanicalstopper” dived under the device layer plane. (p.80)

Fig 4.19:Wiredsensor. (p.81)

Fig 4.20: Sensor diskillustration. (p.81)

Fig 4.21: Sensorcap illustration. (p.82)

Fig 4.22:Sensor disk illustration. (p.84)

Fig 4.23: Experimental plateillustration. (p.85)

Fig 4.24:Dummy diskillustration. (p.85)

Fig 4.25:Experimental setup. (p.86)

Fig. 5.1:Micro ruler feature, new design (L-edit). (p.88)

Fig. 5.2: Teeth overlap and different tether geometry, new design (L-edit). (p.89)

Fig. 5.3:Reduced gap geometry, new design (L-edit). (p.89)

List of tables

Table 1.1: Operation and detection principle of silicon micromachined shear stress sensors (in chronological order of development). (p.18)

Table 2.1: Sensor dimensions. (p.24)

Table 2.2: Summary of all the calculated sensor characteristics. (p.29)

Table 2.3:Summary of the natural frequency for the first mode calculated by ANSYS. (p.30)

Table 2.4: Summary of the natural frequencies of modes 2-5 calculated by ANSYS. (p.31)

Table 2.5: Guide User Interface (GUI) parameters. (p.35)