淡江大學機械與機電工程學系博士班

博士論文
指導教授：楊 龍 杰 博士

整合式金氧半微機電流速計及其在拍翼之
應用
INTEGRATED CMOS MEMS FLOW
SENSOR AND ITS APPLICATION TO
FLAPPING WINGS

研究生：衛瑞紗 撰
中華民國 111 年 6 月

I
INTEGRATED CMOS MEMS FLOW SENSOR AND ITS
APPLICATION TO FLAPPING WINGS
A THESIS SUBMITTED IN PARTIAL SATISFACTION OF
THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
MECHANICAL AND ELECTRO-MECHANICAL ENGINEERING DEPARTMENT
TAMKANG UNIVERSITY

Waikhom Reshmi
June 2022

II
 ACKNOWLEDGEMENT

It’s a genuine pleasure to express my deep sense of thanks and gratitude to my

mentor and advisor Prof. Yang Lung-Jieh and his wife Mrs. Yang. Professor’s
dedication, dynamism and a keen interest in his overwhelming attitude to help his
student had been solely and mainly responsible to successfully complete my work on
time. I admire his nature to push his students to the greater potential right from
providing a scholarship to the time of submitting the thesis. His scholarly advice and
scientific approach have helped me to a very great extent to accomplish this task. Also,
I would also like to express my heartfelt gratitude to Mrs. Yang, who had been kind and
dedicated to the welfare of her students.
I would also like to extend my gratitude to Prof. Shih Horng-Yuan for sharing
his enormous knowledge on several topics on how to design an Instrumentation
Amplifier, I would like to thank Prof. Shih H-Y, Prof. Yang W-B for granting us the
permission to use their department’s laboratory for measuring the Gain of the
Instrumentation Amplifier. Heartfelt thanks to Prof. Chao and Prof. Kang for their
advices in seminar classes. Apart from my amazing Professors, I’d like to thank all the
staff at the department office, for being extremely considerate and helpful with both
academic and non-academic matters.
The CMOS MEMS foundry service provided by Taiwan Semiconductor
Research Institute (TSRI) / United Microelectronics Corporation (UMC) are highly
appreciated.
Thanks to the Chemistry Engineering Department of Tamkang University for
providing SEM (Scanning Electron Microscope) service.
I would like to thank my lab mates Vivek Jabaraj Joseph, Chandrashekhar
Tasupalli, Abhishek Chauhan, Ken, Neethish Kumar Unnam and Paritala
Veeranjaneyulu for being there for me in carrying out the experiment. Thanks to
Chandrashekhar Tasupalli for helping in Wind Tunnel Test. Also, Thanks to Prof. Dai
Ching-Liang, Prof. Shih Horng-Yuan, Prof. Yang Wei-Bin and Prof. Shih Wen-Pin to
serve as an oral test committee for the final defense of the Doctor’s thesis.
Last, but not least, I would like to thank my Grandparents, Parents, and Siblings
for inspiring me every day. I dedicate this work to all my family members, teachers and
friends.

III
論文名稱：整合式金氧半微機電流速計及其在拍翼之應用 頁數:121
校 系 所 組 別 ： 淡 江 大 學 機 械 與 機 電 工 程 學 系 博 士 班
畢 業 時 間 及 提 要 別 ： 110 學 年 度 第 2 學 期 博士學位論文提要
研究生：衛瑞紗 指導教授：楊龍杰 博士

論文提要內容：

本研究工作的主要目的是設計和研究 CMOS MEMS 流量傳感器，然後安裝

在撲翼上以測量其靈敏度。開發了自熱式流量傳感器，並通過風洞試驗進行
了驗證。 CMOS 傳感器設計採用 U18MEMS 工藝實現，是 UMC 0.18µm 1P6M
工藝與 MEMS 後製工藝的結合。圖層設計是使用 CADENCE 軟件完成的。用
於設計電阻器的材料是多晶矽。作者提出使用 STI（淺溝槽隔離）氧化物來保
護多晶矽免受 MEMS 開放區域中的 XeF2 攻擊。儘管將多晶矽圖案放置在
MEMS 區域最初違反了設計規則，但作者還是獲得了 TSRI 的許可。考慮了適
當的 TCR（電阻溫度係數）值，該值對應於實現對溫度變化和不同氣流速度
的高靈敏度。
1. 第一批 自熱式流量傳感器採用熱線式設計，運行過程中不使用任何加器。
我們只採用串聯 2 個 RTD 組成的橋式電路。一個 RTD 置於基板上，另一
個 RTD 置於熱隔絕良好的 MEMS 區域；故接受風吹時，兩個 RTD 形成
溫差與其阻值差異。自熱或設式之輸出電壓不再是噪聲，而是 RTD 半橋
的偏壓，是器件運行所必需的。
2. 在第二批 IC 製造中，加入儀表放大器設計，於自熱式流量傳感器之旁，
作為同一晶片的電壓訊號放大之用。
3. 先將傳感器保持在風洞內進行校正測試，然後再將其安裝在翼展 70cm 的
撲翼上進行升力與推力的量測。
單獨的傳感器設計在 0-10 m/s 的速度範圍內具有 138 µV/V/(m/s)/mW 的歸
一化靈敏度。儀表放大器的電壓增益測量為 30dB。在接續進行放大器和傳感
器集成測量後，觀察到 20dB 的電壓增益，因此自熱式流量傳感器的整體靈敏
度, 顯著提高到 1371 µV/V/ (m/s)/mW 於流速範圍 0-5 m/s。

IV
 除了風洞校正測試，集成傳感器芯片通過安裝在撲翼上進一步測試不同
的撲動頻率。撲翼的測力儀數據也在 MATLAB 軟件中收集和處理。輸出電壓
相對於升力與淨推力的平均靈敏度分別為 90 µV/V/gf 與 159 µV/V/gf。最後，
結合測試結果對 CMOS MEMS 流量傳感器的優勢進行了論證和討論 : 本
CMOS MEMS 流速計晶片結合於撲翼所測得之電壓波形，較類似於淨推力之
時變波形，表現出剪應力感測器的特徵。

關鍵詞：CMOS，MEMS，UMC 0.18 µm，流量傳感器，儀表放大器，

撲翼，靈敏度

＊依本校個人資料管理規範，本表單各項個人資料僅作為業務處理 使用，並於
保存期限屆滿後，逕行銷毀。
表單編號：ATRX-Q03-001-FM030-03

V
Title of Thesis: Integrated CMOS MEMS Flow Sensor and Total pages:121
Its Application to Flapping Wings
Key word： CMOS; MEMS; UMC 0.18 µm; flow sensor; instrumentation
amplifier; flapping wing; sensitivity
Name of Institute: Doctoral program, Department of Mechanical and Electro-
Mechanical Engineering, Tamkang University
Graduate date： June, 2022 Degree conferred: Doctor

Name of student： Waikhom Reshmi Advisor: Dr. Lung-Jieh Yang

衛瑞紗 楊龍杰 博士

Abstract：
The main objective of this research work is to design and study
complementary metal-oxide semiconductor microelectromechanical systems
(CMOS MEMS) flow sensor, which is later, mounted on flapping wing to measure
its sensitivity. The self-heating type flow sensor was developed and verified by wind
tunnel testing. The CMOS sensor is implemented by using U18MEMS process,
which is a combination of UMC 0.18µm 1P6M process along with MEMS post
process in Taiwan Semiconductor Research Institute (TSRI). The layout design is
done by CADENCE software. The material used for designing the resistive thermal
detectors (RTDs) is polysilicon. The author proposed the use of shallow trench
isolation (STI) oxide to protect polysilicon from the XeF2 attack in the MEMS open
region. We received the permission from TSRI even though placing the polysilicon
patterns in MEMS region is originally a violation to the design rule. Appropriate
temperature coefficient of resistance (TCR) value was considered which corresponds
to achieve a high sensitivity to the temperature changes and that of different air flow
speeds.
1. The self-heating flow sensor design, the 1st batch of CMOS MEMS tape out
herein, was done with a hot-wire type where it did not use any heater during
the operation. The author only adopt a bridge circuit composing of 2 RTDs
connected in series. One RTD is on silicon substrate and the other is on the
MEMS region with good thermal isolation. Therefore, apparent temperature
difference and resistance change shows from these 2 RTDs. The self-heating

VI
 output voltage is no more a noise but the bias to the RTD half bridge and is
necessary for the device operation.
2. In the 2nd batch of CMOS MEMS tape out, Instrumentation Amplifier (IA)
design was added into the self-heating flow sensor as the on-chip circuitry
to magnify the voltage output.
3. Flow speed calibration was conducted by mounting the sensor inside the
wind tunnel, later it is also mounted on a flapping wing of 70 cm wingspan
to correlate the lift and net thrust forces.

The sensor design alone has a normalized sensitivity of 138 µV/V/(m/s)/mW

within the speed range of 0-10 m/s. The gain of the IA alone is measured to be 30dB.
After being integrated with the IA and the flow sensor, a gain of 20dB was observed.
Hence, the overall sensitivity of the self-heating flow sensor was measured and
substantially improved to 1371 µV/V/ (m/s)/mW for the flow speed range of 0-5
m/s. Apart from wind tunnel test, the integrated sensor chip is further tested for
different flapping frequency by mounting it on a flapping wing. Standard force gauge
data of the flapping wing were also collected simultaneously and processed in the
MATLAB software. The sensitivity of the output voltage with respect to the lift and
the net thrust are measured as 90 and 159 µV/V/gf respectively. Finally, the
advantages of the CMOS MEMS flow sensors are justified and discussed. The
testing results show that the CMOS MEMS flow sensor signals on the flapping wing
is more similar to the net thrust signal, rather than the lift signal, from the standard
force gauge, and behave with a characteristic of shear stress sensor.
Keywords: CMOS, MEMS, UMC 0.18 µm, flow sensor, instrumentation
amplifier, flapping wings, sensitivity

According to “TKU Personal Information Management Policy Declaration “, the

personal information collected on this form is limited to this application only. This
form will be destroyed directly over the deadline of reservations.
表單編號：ATRX-Q03-001-FM031-02

VII
Contents

ACKNOWLEDGEMENT ..................................................................................... III

ABSTRACT ........................................................................................................... VI

CONTENTS ........................................................................................................ VIII

LIST OF FIGURES ................................................................................................ XI

LIST OF TABLES .............................................................................................. XIV

CHAPTER 1: INTRODUCTION ............................................................................ 1

1.1 Brief history of flapping flight and aerodynamic force measurement .. 2

1.1.1 Inertial effect on the time averaged lift of flapping wings .............. 4

1.1.2 Aerodynamic forces generated by corrugated flapping wing ......... 4

1.1.3 Check valve design on wing............................................................ 5

1.1.4 Soap film visualization on flapping wing ....................................... 6

1.2 Relationship of flow speed and FWMAV’s lift and net thrust ................ 9

CHAPTER 2: CMOS MEMS FLOW SENSOR DEVELOPMENT ..................... 12

2.1 Literature survey on CMOS MEMS technology ................................... 12

2.2 Classification and evolution of CMOS MEMS technology in Taiwan .. 16

2.2.1 UMC ................................................................................................ 17

2.2.2 TSMC .............................................................................................. 19

2.3 Educational CMOS foundry service provided by TSRI ......................... 20

2.4 Design flow of CMOS MEMS sensors .................................................. 22

CHAPTER 3: SELF-HEATING FLOW SENSOR ............................................... 25

3.1 Transfer block of flow sensor operation ................................................ 25

3.2 Design of self-heating/hot-wire flow sensor .......................................... 26

3.2.1 Working principle............................................................................ 26

3.2.2 One-dimensional thermo-resistance analysis .................................. 27

3.2.3 Fluid structure simulation ................................................................ 32

VIII
 3.3 Sensor implementation and testing ........................................................ 37

3.3.1 Layout and tape out ....................................................................... 37

3.3.2 Scanning electron microscope (SEM) inspection ......................... 38

3.3.3 Wind tunnel test............................................................................. 41

CHAPTER 4: ON-CHIP AMPLIFIER FOR FLOW SENSOR ........................... 46

4.1 Instrumentation amplifier pre and post design simulation ..................... 47

4.2 Layout and tape out of instrumentation amplifier.................................. 52

4.3 Instrumentation amplifier gain test ........................................................ 53

4.4 Advantages summary ............................................................................. 55

4.5 Other technical issues of CMOS MEMS process .................................. 56

CHAPTER 5: FLOW SPEED MEASUREMENT OF FLAPPING WING .......... 59

5.1 Sensor mounted on flapping wing using package of TSRI/UMC ......... 60

5.2 Sensitivity Test....................................................................................... 61

5.2.1 Wind tunnel experiment on the integrated flow sensor ................. 61

5.2.2 Application of flow sensor to a flapping wing .............................. 65

5.2.3 Comment on conversion of sensor output voltage to flow speed .. 71

5.3 Packaging issues of MEMS flow sensors .............................................. 71

5.4 Experimental findings ............................................................................ 72

5.4.1 Flow sensor output speed vs. lift or net thrust from force gauge .. 72

5.4.2 Summary on flapping wing experiment ........................................ 81

CHAPTER 6: CONCLUSION ............................................................................... 84

6.1 Summary about the flow sensor ............................................................. 84

6.2 Summary about the instrumentation amplifier....................................... 85

6.3 Summary about the flapping wing with flow sensor experiment .......... 86

6.4 Future Work ........................................................................................... 86

REFERENCES ....................................................................................................... 89

Appendix A: Calorimetric flow sensor ................................................................ 103

IX
 A1 Working principle .................................................................................... 103

A2 Layout and tape out ................................................................................. 104

A3 Testing ..................................................................................................... 105

Appendix B: More graphs of integrated chip test in wind tunnel ........................ 110

Appendix C: MATLAB code ............................................................................... 116

C.1 Cut-off frequency..................................................................................... 116

C.2 FFT........................................................................................................... 117

Appendix D: Sensor on wind turbine blade ......................................................... 118

Publications .......................................................................................................... 120

Journal papers..................................................................................................120
Conference.......................................................................................................121
Patents..............................................................................................................121

X
List of Figures

Figure 1.1 Check-valve on wing: (a) close; (b) open; (c) & (d) lift waveforms ......... 5
Figure 1.2 Image processing of the soap film visualization ....................................... 7
Figure 1.3 Frame processed using image processing.................................................. 7
Figure 1.4 Sensiron flow sensor ............................................................................... 10
Figure 2.1 Various post processes of CMOS MEMS devices .................................. 19
Figure 2.2 Fabrication process of the CMOS MEMS foundry ................................. 22
Figure 2.3 CMOS microsensor fabrication process .................................................. 23
Figure 3.1 Transfer block for the flow sensor operation .......................................... 25
Figure 3.2 Cross sections for the current CMOS process of TSRI/UMC ................. 26
Figure 3.3 Self-heating flow sensor design............................................................... 27
Figure 3.4 Cross sectional view of polysilicon on RTD substrate and cavity .......... 30
Figure 3.5 Plotting the TCR graph ............................................................................ 31
Figure 3.6 Theoretical calculation of the flow sensor ............................................... 32
Figure 3.7 Boundary conditions for the 2D CFD simulation.................................... 32
Figure 3.8 Velocity field around the flow sensor (far field view) ............................ 33
Figure 3.9 Velocity field around the flow sensor (Near field view) ......................... 34
Figure 3.10 Velocity field near the PCB surface and flow sensor .............................. 34
Figure 3.11 Line plots normal to surface of the resistor R1 and R2 ........................... 35
Figure 3.12 Output voltage profile for the 2D CFD simulation ................................ 37
Figure 3.13 Layout diagram of CMOS flow sensor .................................................. 37
Figure 3.14 (a) SB 18 connection; (b) confocal optical microscope image ............... 38
Figure 3.15 Scanning electron microscope (SEM) machine ..................................... 39
Figure 3.16 (a) Holes on MEMS open region; (b) depth of the cavity ...................... 40
Figure 3.17 Cross sectional view of chip using confocal optical microscope ........... 41
Figure 3.18 Wind tunnel experimental set up and testing ......................................... 42
Figure 3.19 (a) 1D theoretical data; (b) 2D CFD simulation; (c) experimental result 43
Figure 4.1 (a) IA circuit diagram; (b) CADENCE schematic diagram .................... 47
Figure 4.2 Pre layout simulation AC response (magnitude) ..................................... 49
Figure 4.3 Pre layout simulation AC response (phase margin) ................................ 49
Figure 4.4 Pre layout simulation common mode gain .............................................. 50
Figure 4.5 Post layout Simulation AC Response (magnitude) ................................. 50

XI
Figure 4.6 Post layout simulation AC response (phase margin) ............................... 51
Figure 4.7 Post layout simulation common mode gain............................................. 51
Figure 4.8 (a) Single stage Op-Amp layout; (b) IA layout ....................................... 52
Figure 4.9 (a) Integrated chip layout; (b) tape out SB 40 package of TSRI ............. 53
Figure 4.10 (a) Pin out circuit diagram IA; (b) set up to test the IA ........................... 54
Figure 4.11 Frequency gain measured oscilloscope .................................................. 55
Figure 5.1 Flow sensor mounted on the wing surface ............................................. 60
Figure 5.2 (a) several IC chips; (b) SB 18 Package dimensions ............................... 61
Figure 5.3 Integrated flow sensor and IA test ........................................................... 63
Figure 5.4 Sensitivity graph for self-heating sensor with IA .................................... 64
Figure 5.5 PCB board to use the minimal required area ........................................... 65
Figure 5.6 IC chip being mounted on the PCB board ............................................... 65
Figure 5.7 Flow sensor with configuration (a) “A1” and “A2”; (b) “B1” and “B2” 66
Figure 5.8 Averaged output voltage vs. wind speed for (a) “A1”; (b) “A2”; ........... 67
Figure 5.9 Averaged output voltage vs. wind speed for (a) “B1”; (b) “B2” ............. 68
Figure 5.10 Amplified signal output for config. A1 and A2 with diff. frequencies ... 69
Figure 5.11 Amplified signal output for config. B1 and B2 with diff. frequencies.... 70
Figure 5.12 Lift and thrust measurement using force gauge ...................................... 73
Figure 5.13 Lift measured w.r.t different flapping frequency .................................... 74
Figure 5.14 Lift raw data of FWMAV processed in the MATLAB ........................... 75
Figure 5.15 (a)-(f) Lift measured in force gauge (a1)-(f1) calibrated data of sensor . 76
Figure 5.16 Net thrust measured with respect to different flapping frequency .......... 77
Figure 5.17 Net thrust raw data of FWMAV processed in the MATLAB..................78
Figure 5.18 (a)-(f) force gauge net thrust data (a1)-(f1) calibrated data of sensor ..... 79
Figure 5.19 Graph for various output voltages subjected to (a) lift and (b) net thrust 83
Figure 6.1 Wind farm with partially damaged turbine blades ................................... 87
Figure 6.2 Flexible PCB board design sensor chips are mounted on it ..................... 88
Figure A1 Calorimetric CMOS MEMS Flow sensor .............................................. 104
Figure A2 (a) Layout diagram; (b) Cross sectional view; (c) Packaged chip .......... 105
Figure A3 Windspeed vs. Vout................................................................................ 107
Figure A4 (a) Test set up; (b) graph for different heating voltages ......................... 108
Figure A5 Graph of integrated IA and calorimetric sensor ..................................... 109
Figure B1 Graph of integrated self-heating sensor and IA for 0-6m/s .................... 110

XII
Figure B2 Graph of integrated self-heating flow sensor and IA ............................... 111
Figure B3 Graph of integrated calorimetric flow sensor and IA .............................. 112
Figure B4 Graph of integrated calorimetric flow sensor and IA .............................. 113
Figure B5 Lift measurement in force gauge and calibrated data of flow sensor ...... 114
Figure B6 Thrust measured in force gauge and calibrated data of flow sensor..........115

XIII
List of Tables

Table 1.1 Averaged force comparison ......................................................................... 8

Table 1.2 Survey of existing MEMS flow sensor ...................................................... 11
Table 3.1 Specification of self-heating flow sensor design ....................................... 31
Table 3.2 Actual free stream velocity observed after the 2D CFD simulation .......... 36
Table 3.3 Output voltage w.r.t. inlet velocity after the 2D CFD simulation ............. 36
Table 3.4 Resistance values ....................................................................................... 44
Table 3.5 Output reading measured in wind tunnel test ............................................ 44
Table 5.1 Output voltage reading for wind speed 0-6m/s.......................................... 63
Table 5.2 Sensitivity and power consumption summary ........................................... 64
Table 5.3 Lift data measured with the MAV using force gauge ................................ 73
Table 5.4 Net thrust measured with the MAV by using force gauge ........................ 77
Table 5.5 Lift measured in force gauge and output voltage measured in sensor ....... 80
Table 5.6 Net thrust measured in force gauge and Vout measured in sensor ............ 81
Table A1 Vout for different heating power supply .................................................. 106
Table A2 Sensitivity comparison for different heating power supply ..................... 106
Table A3 Output voltage of integrated calorimetric sensor and IA ......................... 109
Table B1 Output voltage of integrated self-heating sensor and IA ......................... 110
Table B2 Output voltage of integrated self-heating flow sensor and IA ................. 111
Table B3 Output voltage of integrated calorimetric flow sensor and IA ................. 112
Table B4 Output voltage of integrated calorimetric sensor and IA ......................... 112

XIV
 CHAPTER 1: INTRODUCTION

As the current engineering systems like the offshore wind turbines and aircrafts
getting much more sophisticated and expensive than before, the on-site measurement
technology could benefit these systems with intelligent control and provide bountiful
information good to the operation and maintenance. Even as early as 1990s, Dr. C. M.
Ho of UCLA and Dr. Y.C Tai of Caltech have ever developed on-site measurement by
microelectromechanical systems (MEMS) to study the low Reynolds-number
microchannel flow [1-3] and the aerodynamic characteristics of micro air vehicles
(MAVs) [4-7]. They used the concept of discrete sensor/actuator array to investigate the
fundamental mechanics and to greatly modify the aerodynamic performance of MAVs
with the tailless configuration.
Our research team has ever done the research project sponsored by Ministry of
Science and Technology (MOST), Taiwan research in 2004 about the on-site lift
measurement using PVDF sensor film on the flapping wings. This idea is somewhat
different from the UCLA/Caltech’s discrete sensor/actuator array idea. The PVDF is
continuously layout on the flapping wing and can summarize the total lift force as a real-
time and global output [8]. The more recent papers regarding the on-site sensors using
MEMS technology and their applications were widely reviewed [9-23].
The MEMS sensors used in the previous on-site measurement are mostly
pressure sensors, tactile force sensors and flow sensors. The flow sensors which got
involved thermo-fluidic interaction with the environment under on-site monitoring have
more difficult packaging issues and were selected to be investigated in this research
work. Application on the MAV will be conducted to show the practical value of the on-
site flow sensor technology.
Compared to other MEMS sensors, the flow sensors are seldom discussed in
Taiwan. CMOS MEMS sensor technology will be used to speed up the development
progress of many research works. The finished flow sensor may have some commercial
applications as the following [24].
a. Air conditioner – Air filter, clogged filter detection, leak detection, flow rate
detection.

b. Burning control – Household fuel cell, control for gas mixing.

1
 c. Medical equipment – Oxygen concentration device respirator, breathing quantity
monitor, control for gas supply, measurement of gas consumption.

d. Long-term monitoring the wind speed of the wind turbine [25-30];

e. Others – Chemical analysis equipment, etc.

Secondly, the on-site sensing technology also needs more research work by
considering the sensor technology, sensor signal coordination and packaging techniques.
Flow sensors are extensively used for flow measurements in additional applications to
aerospace, automotive, biomedical, environmental, hydrodynamics and the chemical
and process industries. They can be classified as thermal and non-thermal. Non-thermal
flow sensors can be grouped, as per their transduction scheme, into differential pressure-
based, lift force-based, cantilever deflection-based and resonating frequency-based. A
comprehensive review of these non-thermal flow sensors is presented by Wang et al.
[31]. Thermal flow sensors, on the other hand, use temperature as the main measurement
parameter [29, 32-37].

1.1 Brief history of flapping flight and aerodynamic force

measurement
Caltech released the first IC-integrated flexible shear stress sensor skin in 2003,
which can instantly respond to aerodynamics drag. The measurement signals are mixed
for adaptive flow control and flight control of fixed-wing MAV [4, 38, 39]. MEMS
multi-sensor chip for gas flow sensing was also developed by Xu et al [40].
The above-mentioned sensor-actuated array chip is applied to a 2 µm CMOS
MEMS foundry service. Recently, Hong Kong University of Science and Technology
(HKUST) and National Chao-Tung University (NCTU) have co-developed a
calorimetric flow sensor using a CMOS 0.18 µm 2-ploy-4-metal (2P4M) MEMS foundry
service provided by the Chip Implementation Center (CIC) of Taiwan to completely
manufacture the overall chip [41, 42].
CMOS MEMS flow sensors is an utmost important need in emerging
applications such as determining the fluid flow rate in wind turbines, flapping wing
micro air vehicles (FWMAV’s), various industries, etc. [24, 43]. Ever since the MEMS
technologies were established in the 1980s, many efforts have been made by researchers
to develop MEMS flow sensors [3, 44]. The advantages of these kind of sensors are that

2
it has low power consumption, high accuracy, and small sized, good sensitivity up to
some extent of flow range. The first ever experimental data of the pressure distribution
of gaseous flows in microchannels using MEMS technology were reported by Liu et al.
[3] Several groups have studied the temperature change or distribution inside
microchannels [45, 46] . Out of different flow sensing methods, thermal flow sensors
have simple design and easy to realize structures. Thus, it offers a practical solution for
integrated sensing in microfluidic applications. The basic elements in a thermal flow
sensor are heating and sensing elements. Heat transferred from the heating elements to
the working fluid results in resistance variation in sensing element of thermal flow
sensor. Better sensitivity can be achieved if all the heat generated by the heating element
can be transferred to the fluid and detected by the sensing element [36]. Apparently, due
to packaging difficulties and flow induced signal fluctuations, most thermal flow sensors
were built on the wall of the flow channel and sometimes it suffered inevitable heat loss
[33, 47]. Heat loss phenomena should be taken good care otherwise there will be more
power consumption which will again lead to decreased in sensitivity [39]. MEMS multi-
sensor chip for gas flow sensing were also developed by Xu et al [38]. Although, some
processes cannot be fabricated in the simple clean room of Tamkang University, we can
seek assistance for Taiwan CMOS foundry service.
Motivation for flow sensor design came from flow sensing technique which is
conventionally necessary to industrial control of fluid transportation. In the last 3 decade,
the similar mass flow control (MFC) to flow sensing is important to the semiconductor
processing [35, 48]. The flow sensing control of the respiration machines is urgently
needed during Covid-19 pandemic period of course and attracts people’s attention
recently [49, 50].
The application scenario of the flow sensor in this thesis is different from the
general engineering issues and is the on-site in-situ speed measure regarding the flapping
wing flow fields. This idea comes from the in-situ MAV flight control and flow
monitoring of UCLA/Caltech MEMS group in 1990s [1, 2, 4, 5]. As the size and the
total mass of flapping wings herein are very limited, the flow sensor cannot be bulky but
must be miniaturized by MEMS technology. Therefore, the flow feature and patterns
will not be altered by the flow sensors and the related electronic package [51-54] .
MEMS devices including flow sensors should be developed and manufactured their
prototypes in professional micromachining facilities, which cannot be afforded by the

3
author’s research team. The author then resorts to the free foundry service of CMOS
MEMS provided by TSRI in Taiwan [55]. The MEMS flow sensors developed by
commercial CMOS process lines are still scarcely heard [56]. Prior to the flow sensor
design, some studies related to flapping wing, aerodynamics and its fluid flow
phenomena which the author got involved are as follows:

1.1.1 Inertial effect on the time averaged lift of flapping wings

The time-averaged inertial force which relates to the velocity difference between
the final and initial states of a flapping wing motion were studied [10, 57]. For the
periodic flapping motion with identical final and initial velocities, there are no inertial
force contribution to the time-averaged lift. Therefore, the wake capture mechanism
proposed by Dickinson justifies more convincing than Sunada’s added mass or Sun’s
rapid acceleration at the stroke onset of hovering [58-60]. The vanishing inertial force to
the time-averaged lift is also beneficial to the concise signal processing of lift data from
the wind tunnel test.

1.1.2 Aerodynamic forces generated by corrugated flapping wing

Another research work, which deals with the fabrication, aerodynamic
measurement, and performance evaluation of the corrugated wing patterns on a
FWMAV, was also studied. The corrugated wing pattern is generally seen in insects.
Insect wings with their corrugated topological features give themselves high load-
bearing capacity during flapping and hovering. It is believed that appropriate corrugated
structures on insect wings enhance aerodynamic performance. A new fabrication process
using polydimethylsiloxane (PDMS)-molding and parylene coating was proposed and
implemented for a 24 cm-span FWMAV, which is composed of corrugated wing
inspired by the dragonfly wing. The fabrication includes the master wing template made
by three-dimensional (3D) printing, the PDMS molding process for wing template
duplication, and the final parylene coating to obtain corrugated wings. The parylene-C
thickness was selected as 40 μm. For comparing the aerodynamic performance, two
types of flapping wings were implemented on the FWMAVs: first with a pair of parylene
corrugated wings and the other with polyethylene terephthalate (PET) flat membrane
(without corrugation) wings. The lift signal was then measured by the load cell in a wind
tunnel. Regarding the aerodynamic performance of flapping wings, a systematic

4
performance estimation of finding the cruising conditions and the cruising lift for
FWMAVs was investigated through the massive wind tunnel data. The parylene flapping
wings with corrugation attained lift coefficients of 7.8-8.0, which the PET flat membrane
wing cannot achieve so far [61-65].

1.1.3 Check valve design on wing

A FWMAV demands high lift and thrust generation for a desired payload. In
view of this work focuses on a novel way of enhancing the lift characteristics through
integrating check-valves in the flapping wing membrane [66-70]. Figure 1.1 shows the
illustration of check-valve closing and opening on the wing surface: (a) check-valve
closing during downstroke; (b) check-valve opening during upstroke; (c) expected lift
wave form without check-valves; (d) expected lift wave form with check-valves.

(a) (c)

(b) (d)

Figure 1.1 Check-valve on wing: (a) close; (b) open; (c) & (d) lift waveforms

Modal analysis and static analysis are performed to determine the natural
frequency and deformation of the check-valve. Based on the inference, the check-valve
opens and closes during the upstroke flapping and downstroke flapping, respectively.
Wind tunnel experiments were conducted by considering the two cases of wing design,
i.e., with and without a check-valve for various driving voltages, wind speeds and
inclined angles. A 20 cm-wingspan PET membrane wing with two check-valves,

5
composed of central disc-cap with radius of 7.43 mm, supported by three S-beams,
actuated by Evans mechanism to have 90° stroke angle, is considered for the 10 gf (gram
force) FWMAV study [8, 44, 57, 62-64, 67, 71-78]. The aerodynamic performances,
such as lift and net thrust for these two cases are evaluated. The experimental result
demonstrates that an average lift of 17 gf is generated for the case where check-valves
are attached on the wing membrane to operate at 3.7 V input voltage, 30° inclined angle
and 1.5 m/s wind speed. It is inferred that sufficient aerodynamic benefit with 68% of
higher lift is attained for the wing membrane incorporated with check-valves. Thus,
flapping wing with check valves, its lift during the upstroke can be turned to positive.

1.1.4 Soap film visualization on flapping wing

FWMAVs animate the small-space dexterous flight, hovering, and energy-saving

characteristics of birds and insects, and are believed to have enlightenment for the
development of bionic flight in the future. When designing FWMAVs, detailed unsteady
aerodynamic information is required. Besides the computational fluid mechanics (CFD)
technology study, the flow visualization is also needed to assist this research. This work
innovatively used soap film visualization with high-speed photography to record two
kinds of the 2D flow fields laterally and longitudinally, respectively, generated by a
flapping wing of 10 cm span [45, 66, 79-92]. Different from the qualitative comparison
of soap film imaging with the conventional smoke tracing method, the subsequent
processing of the soap film images was demonstrated. This work explains how to
quantify the soap film imaging into lift and thrust forces, and the corresponding results
are compared with the wind tunnel force measurement data preliminarily. Figure 1.2
shows the flow chart for image processing of the soap film visualization and Figure 1.3
represents the frame processed using image processing: (a) select a rectangular domain
(A = y*  z*) from a colorful image; (b) convert into a thickness matrix 559  393 shown
by grey scale (unit: nm). Table 1.1 is the averaged force comparison between the wind
tunnel force gauge and the soap film experiment in this work.

6
 Figure 1.2 Image processing of the soap film visualization

(a) (b)

Figure 1.3 Frame processed using image processing

7
 Table 1.1 Averaged force comparison
Force Wind Tunnel Soap film
Type Test Experiment
−0.8 gf (7.58 Hz, 0.15
Lift 1.7–1.8 gf (33 Hz, 1.2 m/s)
m/s)
Thrust 2.4 gf (33 Hz, 1.2 m/s) 3.6 gf (20.8 Hz, 1.16 m/s)

Some findings from this topic are discussed and summarized as below:
1. This work preliminarily presents a soap film visualization technique to capture
the unsteady flow images around a 10 cm-span FWMAV in a dynamic manner.
Several flapping cycles have been recorded before the collapse of the soap film.
Two experimental setups with the soap film plane perpendicular to starboard and
streamwise directions have been demonstrated. The patterns of inverted Karman
vortex and wing tip vortex were both successfully observed.
2. A theoretical formulation about the relationship between the thickness matrix and
the color matrix of the soap film is performed for evaluating the jet speed and the
downwash speed. It is useful for calculating the thrust and lift of a FWMAV. The
soap film visualization in this work can emergingly output quantitative results.
3. After filtering the ripple noise deduced by the free-surface wave propagation
which can be apparently observed during flapping motion, the resultant thrust
and lift waveforms come from the soap film images and the thickness fields are
similar to the wind tunnel data preliminarily.
4. The lift signal of the soap film experiment for the 10 cm-span FWMAV is similar
to the classical waveform. Both of them have a trend agreed with the twin-peak
phenomenon which is a manifestation of the “delayed stall” and “wake capture”
lift mechanisms of wing flapping motion proposed by Dickinson [60]. However,
the soap film time-averaged lift value (−0.8 gf) still deviates from the measured
result (1.75 gf) of wind tunnel testing very much. The main reason is that the
traveling speeds for the two cases (1.2 m/s vs. 0.15 m/s) are distinctly different.
5. On the thrust signal of the soap film experiment, it has a similar trend to the
classical signal and verifies that there are two positive thrust actions per flapping
cycle in the real manner. In other words, either in downstroke or upstroke periods,
the flapping wing always generates positive thrust to accelerate the FWMAV.
Meanwhile, FWMAV generates positive lift only during the downstroke in

8
 general [10]. However, it is observed that for a single cycle, the time-averaged
thrust of soap film (3.6 gf) is more than the wind tunnel force-gauge data (2.4
gf). It may be due to the inaccurate lift estimation from soap film experiment and
the thrust calculation needs assistance from the lift data.
6. In addition, the FWMAV traveling speed is limited under 1.5 m/s now and not
exactly appropriate to the case of forward flight. (Further, it is also hard to find
out the wind tunnel testing counterpart for result comparison.) However, this
method may have potential applications to investigate the flow field of hovering
FWMAVs in the future.
7. Globally, this soap film visualization is a low-cost method without dangerous
concerns to study the unsteady flows of FWMAVs, and it can be as an option of
flow visualization techniques besides the smoke tracing and PIV (particle image
velocimetry)[85, 93, 94].

1.2 Relationship of flow speed and FWMAV’s lift and net

thrust
Relationship between the flow speed measurement and the lift and net thrust
generation of flapping wings will be discussed in this section. There are two ways to
investigate the fluid mechanics of flapping wing. One is to use a wind tunnel to measure
the global lift and thrust externally. The other is to use PIV [85, 93, 94] or flow
visualization [90, 95] to show the detailed flow field internally. The force gauge output
of the former wind tunnel testing can be referred to the ornithopter book written by Yang
& Balasubramanian [10]. The author also participated in the research of soap film flow
visualization about a flapping wing [95]. Unfortunately, the time-consuming original
processing of PIV and the image processing of soap-film experiment stuck the real-time
access of the above aerodynamic information into the MAV flight control and the on-
board avionics integration [62, 64, 71]. Mounting the MEMS flow sensors directly on
the flapping wing for detecting the downwash speed by the 1st hand is obviously much
better than the previous non-MEMS sensing technologies. To measure the real-time, on-
site lift of a flapping wing in terms of (downwash) flow speed, the model formulation is
necessary. In Ref [8], Yang et al. have ever derived the PVDF voltage output linearly
proportional to the flapping lift. In Ref. [89, 95], the author’s team have also derive the
soap-film thickness time variation related with the downwash flow speed of the flapping

9
wing. In this thesis, the author plans to directly measure the local speed by CMOS
MEMS flow sensors on wing surface, to denote the real-time, on-site aerodynamic force
of a flapping wing.
Section 1.1.2, 1.1.3 used force gauge in the wind tunnel to do the measurement.
The output of the force gauge denotes the overall lift and thrust of a flapping wing. The
local flow change on the wing surface needs the help of high-speed photography. Section
1.1.4 is about the soap film visualization, which is similar to PIV and can analyze the
downwash and horizontal jet locally. However, the 2D soap film feature, low successful
rate of the experiment and inaccuracy limits this kind of application method. Based on
the flow field measurement experience of the soap film visualization technique and other
previous work related to sensors and actuators wing technique development by
UCLA/Caltech in 1990s [2-7], the author hopes to develop the CMOS MEMS flow
sensor on the flapping wing.
The simplest way to detect the on-site local speed of a flapping wing is to mount
commercial MEMS flow sensors of Figure 1.4 on the flapping wing surface. However,
due to the very bulky size of the outside package of the commercial MEMS flow sensors
compared to relatively small size of flapping wing, the above direct integration is not
feasible. Another reason to give up using commercial MEMS flow sensors is also due to
the sensor package. Any approach to change the flow sensor package to match the need
for the flapping wing application will easily attenuate the output sensitivity seriously.
Therefore, it is necessary to develop and implement MEMS flow sensor by ourselves in
the following chapters.

Figure 1.4 Sensiron flow sensor [70]

Some of the existing MEMS flow sensors and their sensitivities using different
materials are shown in Table 1.2.

10
 Table 1.2 Survey of existing MEMS flow sensor
Sensing Configuration Fluid Detection Sensitivity Ref.
element type range
3C-SiC thin film Hot film Air 0-9 m/s 0.091 1/(m/s) [96]
heater
Pt Hot wire Air 1.5-11 CV mode 0.1433 [97]
m/s × 10-4 A(m/s)-1/2
CC mode 7.98 ×
10-3 V(m/s)-1/2
Al/Si bond wire Hot wire Air .01-17.5 LF mode: [98]
m/s 1.61V/(m/s)
HF mode:
0.58V/(m/s)
Au/Cr detector Hot film Gas .01-5 m/s 6.8 × 10-3 1/(m/s) [99]
Al/Poly Hot film Gas, N2 0-0.4 m/s 6 × 10-3 [100]
thermocouples V/(m/s)W
Poly heater
Ni resistors Hot film Diff.gas 0-5 m/s 2.6 × 10-2 [101]
mixture V/(m/s)
B-doped poly Hot film DI 0-8.25 × 0.87 V/(m/s) [102]
heater water 10-3 m/s
P- doped Calorimetric Air 0-50 m/s Silicon substrate: [103]
polysilicon 3.04 × 10-2
thermopile V/(m/s)
W/Ti heater Soldered: 5.1 ×
10-4 V/(m/s)
Adhesive
bonded: 2.01 ×
10-2 V/(m/s)
Polysilicon Cantilever Air 0-6 m/s 7.4 × 10-6 1/(m/s) [104]
strain gauges
Polysilicon Cantilever Gas 0-70 7.4 × 10-6 V [105]
strain gauges l/min l/(m/s)2
ZnO transducers Piezoelectric Water 30-300 2.08(uL/h)/mV [106]
polyimide pressure ul/h
sensor
PDMS micro Piezoelectric Water 10-6m/s × 0.3V/(m/s) [107]
pillars hair cell sensor 10-2 m/s
PVDF nanofiber
tips

11
 CHAPTER 2: CMOS MEMS FLOW SENSOR
DEVELOPMENT

In this chapter, CMOS MEMS flow sensor design development will be explained
in details.

2.1 Literature survey on CMOS MEMS technology

When humans entered the 21st century from the 20th century, cross-sector
integration of science and technology is constantly emerging; in conjunction with the
trend of miniaturization, multi-functionality, and precision. It is the development of
various cutting-edge technologies and industries (including nanotechnology, materials,
chemicals, electronics, computers, information, photovoltaics, machinery, medicine,
biotechnology, health care, environmental protection, and energy, etc.), and will have a
huge impact on human life, even economic and wealth distribution. The first to propose
the concept of microelectromechanical systems (MEMS) or micro system technology
(MST) was the physicist, R. Feynman [108]. In 1959, at the annual meeting of the
American Physical Society, Feynman delivered a speech entitled "There is plenty of
room at the bottom" to attract people's attention to small-scale science and technology.
Shortly after Feynman’s second speech “Infinitesimal machinery” in 1983, this
emerging technology measured in units of “micrometer” (μm).
The invention of the transistor at Bell Telephone Laboratories in 1947 sparked a
fast-growing microelectronic technology [109, 110]. Piezoresistive silicon strain gauges
were introduced in the late 1950s by Kulite Semiconductor, Bell Lab’s first licensee of
patents on semiconductor piezoresistance reported in 1954 [111, 112]. It was discovered
that the piezoresistive effect in Ge and Si had the potential to produce Ge and Si strain
gauges with a gauge factor (i.e., instrument sensitivity) 10 to 20 times greater than those
based on metal films. As a result, Si strain gauges began to be developed commercially
in 1958. The first high-volume pressure sensor was marketed by National Semiconductor
in 1974. This sensor included a temperature controller for constant-temperature
operation.
In 1982 silicon was called as a mechanical material in K. Petersen’s famous
review article [113]. During 1987-1988, a turning point was reached in micromachining

12
when, for the first time, techniques for integrated fabrication of mechanisms (i.e. rigid
bodies connected by joints for transmitting, controlling, or constraining relative
movement) on Si were demonstrated. The resonant gate transistor (RGT) [114] was
dissimilar with conventional transistors in that it was not fixed to the gate oxide. As an
alternative, it was movable and cantilevered with respect to the substrate used. In 1967,
the RGT was the earliest demonstration of micro electrostatic actuators. It was also the
first demonstration of surface micromachining techniques.
In 1971, Intel publicly introduced the world's first single chip microprocessor,
the Intel 4004. The 4004 powered the Busicom calculator and was Intel's first
microprocessor [115]. The story of the invention of the charge-coupled device (CCD) is
easier to unravel L.F.J. Sangster [116] of Philips Research Laboratory and W.S. Boyle
and G.E. Smith [117] of Bell Laboratories at Murray Hill independently came up with
the CCD structure. In United States, J.B Angell, Stanford University [118-120], K. D.
Wise from University of Michigan has also made astonishing remark with their
work[118, 120-124]. Though there were many initiatives in the 1960s, it is almost certain
that Stanford University’s Integrated Circuits Laboratory, under the guidance of Prof. J.
Angell, can be credited with being the renowned research group to be active in the field
of silicon sensors and micromachining. Another enthusiastic group had also been
working on silicon sensors for biomedical applications under the guidance of sensor
pioneer Prof. W. H. Ko of Case Western Reserve University [125, 126]. The first of the
gate-controlled diode designed by Zemel et al. [127] at the University of Pennsylvania
can also be noted. In Sweden, I. Lundstrom et al. invented Pd-gate led many interesting
gas sensitive devices University of Linkoping in 1975 [128], G. Stemme also made
contributions in this field who worked at Royal Institute of Technology, Sweden [129].
The first to point out the necessity of applying CAD to sensors was S. D Senturia from
MIT, UK in 1976 [130]. The first silicon flow sensor based on the temperature sensitivity
of a diffused Wheatstone bridge in silicon was designed by van Putten and Middelhoek
of Delf University of Technology in 1974 [131]. P. Bergveld of University of Twente
[132] and R Puers of K. U Leuven are also remarkable mentions from Netherland [133].
Institute of Integrated Micro and Nano Systems, School of Engineering and Electronics,
Scottish Microelectronics Centre, University of Edinburgh, Scotland; T. Matsuo of
Tokohu University, Japan [122]; M. Bao of Fudan university, China [125]; W. Gopel
from University of Tubigen, [134-136] and E. Obermeier T.U Berlin, Germany 1987

13
[137, 138] are some of the renowned Professors from the distinguished Universities from
several countries whose era should be remembered. However, the most brilliant
technology milestone of MEMS is that Muller, Fan and Tai invented the world's first
micrometer-sized electrostatic motor (micromotor) in 1989 at the University of
California, Berkeley, using the IC process technology before the official terminology of
MEMS [139, 140]. R. S. Muller of UC Berkley started the IEEE Electron Devices
Society of IEEE/ASME Journal of Microelectromechanical Systems in 1990.
Famous professor of MEMS H. Guckel of UW Madison, USA; R Howe of UC
Berkley, USA; M. Esashi of Tohoku University, Japan; H. Fujita of University of Tokyo,
Japan; O. Tabata of Kyoto University, Japan; C. M. Ho of UCLA, USA; C. J. Kim of
UCLA, USA; V. M. Bright of U Colorado Border, USA and G. K. Fedder of UM
Annaber, USA are very active in conducting MEMS research projects and hosting
important international conferences like IEEE MEMS, Transducers, MicroTAS, etc. In
1994, Cornell University introduced a bulk micromachining process called as single
crystal reactive etching and metallization (SCREAM) [141] Special mention is also
made of Petersen’s review paper [113], a truly seminal work that gave the field a much-
needed impetus. From 2005 to 2021 with the advancement in MEMS fabrication,
manufacturing technologies and processes, the various applications of MEMS structures
are explored and developed, some of which are airbag accelerometers, intelligent tires,
vehicle security systems, inertial brake lights, headlight leveling, rollover detection; or
inkjet printer heads, projection screen and televisions, earthquake detection and gas
shutoff, and instrumentation, voltage controlled oscillators (VCOs); or surveillance,
arming systems, embedded sensors, data storage, aircraft control, tanks control; or blood
pressure sensor and so on.
Thus, we can observe from last several decades, there has been rapid growth of
MEMS fabrication technology, which enables manufacturing of micro devices of
various sensors and actuators. The term MEMS originally meant miniaturized
electromechanical actuators, but now ap-plies to a broad family of micromachined
sensors, actuators, and systems with coupled electrical, mechanical, radiant, thermal,
magnetic, and chemical effects. The technologies are based on the sequence of the
fabrication of complementary metal-oxide-semiconductor (CMOS) circuitry and MEMS
elements, while silicon-on-insulator (SOI) and CMOS MEMS are discussed separately
[112, 142-144]. Starting from 0.8 µm double-polysilicon-double-metal (DPDM) CMOS

14
process, tremendous efforts have been made to continuously improve process yield and
reliability, while minimal feature sizes and fabrication cost continue to decrease [112,
145-147].
The concept of producing MEMS devices in a standard CMOS process was first
described by H. Baltes [148, 149] . He conducted the feasibility study followed by the
launch of several national agencies that had the ability to provide the CMOS foundry.
The motivation for its research and development (R&D) is to directly use the off-the-
shelf IC foundry equipment, integrate the MEMS sensors and actuators with sensing
signal processing on the same silicon chip (on-chip circuitry). Designers of MEMS
devices only need to design the layout and after the completion of the foundry performed
a post-processing to obtain the desired MEMS devices. The benefits of the MEMS
sensors, actuators and the signal processing circuits can be linked on-chip together to
reduce the influence of external noise to a minimum and improve component
performance. After verifying the stability and reliability of the CMOS MEMS devices
by using CMOS foundry, it is quick to enter the production market and create business
opportunities. Semiconductor roadmaps show the current state and, more important,
outline the future performance of CMOS technology with ever-increasing integration
density and decreasing feature size [150-154]. With the breakthrough of technologies, it
has enabled to integrate MEMS structures with ICs on a single CMOS substrate, so-
called monolithic CMOS MEMS integration.
CMOS compatible MEMS technologies includes bulk micromachining (wet or
dry, isotropic or anisotropic, from wafer front or back), surface micromachining
(sacrificial dielectric and/or metal or polysilicon), micromachining before/after
completion of IC process (pre-CMOS/post-CMOS), use of IC materials for
microstructures with electro-thermo or opto-mechanical functions, thin film deposition,
chemical-mechanical polishing (CMP), and wafer bonding. It is indeed important to
mention about the trending materials where a number of researchers has started
exploring more about Silicon carbide materials and nanostructures. Study of this
materials requires a good understanding of physical, biological and chemical properties,
fundamental studies, as well as applied, theoretical and/ or computational studies related
to several new applications are highly motivated. Bulk silicon carbide is a wide-bandgap
semiconductor with different crystalline forms owing to excellent electronic
characteristics, it has mainly been used for high-temperature, high-frequency, and high-

15
power electronic devices. It is also recognized as one of the best biocompatible materials,
especially in cardiovascular and blood-contacting implants and other biomedical devices
[155-162]. However, if the above cannot be exactly fabricated in the simple clean room
of researchers’ laboratory, commercial CMOS foundry service could be found as help.

2.2 Classification and evolution of CMOS MEMS technology

in Taiwan
The research on CMOS MEMS devices in Taiwan were started in National
Taiwan University (NTU). It began with Chang, Lu and Liu to ask help of CIC’s CMOS
foundry to guide students to attend the IC design courses held in CIC for 1994 summer,
and to start the layout design of capacitive micro accelerometers, piezoresistive micro
pressure sensors and sensing signal reading circuits [142, 163]. The first difficulty faced
during the development of CMOS MEMS devices is that the architecture of Cadence
software is entirely for the needs of IC design and with its fixed sequence of "schematic,"
"H-Spice simulation", "circuit layout", and "layout verification". Due to the complexity
of ICs, the area per unit square centimeter at that time may have tens of millions of
transistors connected together. Therefore, the layout designers must design IC on one
hand, and use the software to repeatedly check, simulate and verify to minimize the
possible bugs to bring the ultimate solution. For this reason, the IC design software
focuses on the integration of transistor components and processes with known
characteristics and completes the huge IC system. The designers do not receive much
stress of the micro-structured layers of the wafer which are subjected to, and they will
not be told how to do the conversion of physical signals into electrical signals for
sensing/actuating, either [164]. In other words, if they would like to use commercial IC
design software to design or even manufacture CMOS MEMS devices, they must first
clarify the fundamental differences between transistor IC and CMOS MEMS devices. In
addition, they need to communicate this CMOS MEMS concept to CIC leaders who
originally work on the standardization and protocol of IC foundry. Otherwise, the IC
designer will be amazed about the “strange” layout design methodology of CMOS
MEMS devices.
For example, the CMOS MEMS designers still use the Cadence software to draw
layouts, but they need not go through the complete verification procedure of the standard
IC because MEMS structures are not ICs layout. The exact verification of the complete

16
ICs should include the followings. (1) design rule check (DRC): Check whether the
layout patterns are too dense, and to ensure the foundry's production correctness and
there will be no short circuit problem; (2) electrical rule check (ERC): Check electrical
characteristics; (3) layout versus schematic (LVS): Compares and map the layout with
the logic circuit of schematic; (4) layout parameter extraction (LPE): Captures the
electrical parameters of the layout.
For the MEMS device of the non-signal processing circuit, the DRC verification
is necessary to ensure that the microstructure can be successfully made, but the other
three verifications (2)(4) to MEMS sensor are not required (; for example, a variable
capacitor membrane subject to a pressure loading cannot be given or simulated for its
capacitance change.) Furthermore, for an IC designer, it seems a waste of chip area of
400 × 400 µm2 to make a single sensor capacitor because making an 8086
microprocessor may use the similar chip area. Even though the dimensions of the MEMS
device are well simulated and determined in advance by the finite element analysis
(FEA), which has not yet been integrated into the circuit design software so far. The
MEMS engineers also need to conduct cross-domain communication because CMOS
MEMS are different from the traditional circuit design training. At the end of 1995, by
the help of CIC and TSMC, Dai, Lu and Chang of the NTU group successfully
developed the first CMOS MEMS mechanical sensors including a piezoresistive
pressure sensor and a capacitive accelerometer with a switched capacitor circuit for
capacitance measurement. The first two mechanical sensors are with the chip areas of 2
mm × 3 mm and 3 mm × 3 mm, respectively; the tiny switched capacitor circuit was
directly placed among the mechanical sensors [142, 163].

2.2.1 UMC

United Microelectronics Corporation (UMC) was founded as Taiwan’s first IC

fabrication company in 1980 as a spin-off from the government-sponsored Industrial
Technology Research Institute (ITRI) in Hsinchu, Taiwan. UMC being significant
supplier to the automotive industry, it has four 300 mm fab: one in Tainan, Taiwan; one
in Singapore; one in Xiamen, China; one in Japan. UMC also has several 200 mm fab in
Hsinchu and Suzhou, China. In 1995 the foundry service just started. In 1996, UMC 0.35
µm CMOS process was released successfully. In 1997 and 1999, the minimum line width
or the critical dimension (CD) was upgraded to 0.35 µm and 0.18 µm, respectively. At

17
that time, UMC focused on improving their CMOS IC process capability only. The
CMOS MEMS researchers need to carry out their die-level post process of CMOS
MEMS personally, e.g., by performing their own plasma-enhanced chemical vapor
deposition (PECVD) process in which the deposition of thin films of various materials
takes place at a lower temperature (<350C) than that of standard CMOS process. The
various post processes in Figure 2.1 after CMOS standard process may be also followed
by die-level photolithography, wet etching (backside), wet etching (front side), wet
etching followed by electroplating nickel, or wet etching (fronts side) followed by
gelatin coating [165, 166] or perylene coating [167], etc. Figure 2.1 (a) shows the CMOS
standard process which composed of four kinds of different thin films including silicon
nitride (Si3N4), silicon dioxide (SiO2), aluminum, and polysilicon on the silicon
substrate; (b) is for fabricating a membrane, the backside of the silicon substrate were
lapped and polished to reduce its thickness from 680 µm to 200 µm. This procedure
decreases much time for backside V-groove etching and preserves the aluminum pads
from being attacked by the exposed anisotropic etching solution. Figure 2.1 (c) is the
acid etch metals (surface micromachining); (d) is the metallic suspension bridges can be
created from the front side and the sacrificial material is changed to oxide accordingly.
Herein the plasma etching is done after the passivation nitride is removed and the top
metal layer is exposed. Therefore, the top metal layer is used as the etch-resistant
material during the subsequent dry etching that create the laminated microstructure.
Without wet etching, surface tension sticking does not occur and the laminated-
suspension microstructures are achieved. The anisotropic oxide plasma etching through
to the silicon substrate by CF4/O2 RIE. For Figure 2.1 (e), acid is used to etch metals and
TMAH used to etch the Si substrate; (f) is electroplating Ni process of step (c); In Figure
2.1 (g), parylene or gelatin is coated in the process (e); lastly, (h) is the anisotropic and
isotropic etching with STI protection.

18
 (a) (e)

(b) (f)

(c) (g)

(d) (h)

Figure 2.1 Various post processes of CMOS MEMS devices

2.2.2 TSMC

Taiwan Semiconductor Manufacturing Company, Ltd. (TSMC) is another

Taiwanese multinational semiconductor contract manufacturing and design company
with its headquarters and main operations located in the Hsinchu Scientific Park in
Taiwan. On 300 mm wafers, currently, TSMC has silicon lithography on node sizes
ranging from 0.13 µm to 5 nm and ranked the number one of the IC foundry providers

19
in the world. The capability of IC foundry service in Taiwan proves its top leading place
in the globe. The reason is neither due to the strong IC design houses nor due to the
reliable equipment supply of the semiconductor process. However, the success comes
from the high utility rate of the process equipment and high product yield rate. For
TSMC especially, they additionally created the advantage of flexibility in process
sequence and process conditions, and set up a very high entry barrier, which other IC
fab companies cannot easily surpass.
Regarding the IC designer training for students in Taiwan, Ministry of Science
and Technology (MOST; or National Science Council) founded the CIC in Hsinchu
Scientific Park. CIC provided low-price or free access of Cadence software to the
academics and called for IC designs from all universities in Taiwan regularly. After
review and defense, CIC granted the real chip implementation for the selected IC designs
and merges the selected IC layouts into a multi-project-wafer (MPW). Finally, CIC sent
the photomasks to TSMC or UMC for fabricating a batch of wafers with these selected
IC designs. After the IC foundry, the IC designers could receive 20-40 pieces of CMOS
IC chips for testing and verifying their IC functionality. The student designer haves their
obligations to submit their test reports to CIC within months after they received the IC
chips. Through this kind of education training using real IC foundry service with several
times of call-for-designs every year, CIC trained many IC student designers for IC design
houses as well as developing or exploring some advanced IC circuitries at the same time.

2.3 Educational CMOS foundry service provided by TSRI

The Taiwan Semiconductor Research Institute (TSRI) under the National
Applied Research Laboratories (NARLabs) is now a consolidation of the CIC and Nano
Device Laboratories (NDL) since 2019 with the hope of creating an integrated
semiconductor research environment in which excellent academic research and
innovative technological development can be conducted to sustain the competitiveness
of the semiconductor industry in Taiwan. It is also to keep pace with international
technology trends as well as the 3-nm node and the rapid development of new
applications in, e.g., artificial intelligence, quantum computers, next-gen magnetic
random-access memory, high-speed computers, and 5G network.
At the heart of sensor products is the CMOS technology using the latest TSRI/
UMC 0.18µm single-poly-six-metal (1P6M) process, which enables us to combine the

20
sensor component with amplifier circuitry on a tiny CMOS silicon chip. The CMOS
MEMS technology provides error-free gas flow metering that remains stable over long
period and generates a fast and high-precision sensor signal recently [34, 37, 41, 55, 168,
169].
Step by step fabrication process of the latest CMOS MEMS foundry process is
shown in the following Figures 2.2 (a) depositing the polysilicon above the substrate;
(b) the CMOS layer with additional pattern of contact layer; (c) the standard CMOS layer
with amorphous SiO2, metal M1-M6, metal 7 hard mask and passivation layer; (d)
depositing the thick photoresist layer to perform the patterning; (e) anisotropic silicon
oxide etching; (f) XeF2 isotropic silicon etching.
Figure 2.2(f) shows UMC 0.18 m CMOS process proceeded with isotropic XeF2
undercut etching for the MEMS open area; light blue color in the cross-sectional area
shows the use of shallow-trench-isolation (STI) to protect polysilicon during etching
process. It is because that the gate oxide thickness is only 4 nm and hard to protect the
polysilicon sensor from the XeF2 attack. Therefore, the authors proposed the using of
STI oxide to protect polysilicon from the XeF2 attack. The authors received the
permission from TSRI even though placing the polysilicon patterns above STI in MEMS
region is originally a violation to the design rule.
Fabrication technologies during CMOS and CMOS MEMS processes include
thermal conversion [33, 37, 53, 170-172], chemical vapor deposition (CVD) [112, 155,
173], epitaxy [137, 174-178], physical vapor deposition (PVD) [179-183], atomic layer
deposition (ALD) [184, 185], spin-on films/dielectrics [186-188], bulk micromachining
[189], surface micromachining [150, 152, 190, 191], photolithography module [192],
dry etching and wet etching. Most of the materials used for structural, sacrificial and
passivation layers include silicon and its oxide, nitride, silicon-germanium, carbide
derivatives and other notable semiconductor and dielectric materials [193-195].

21
 (a) (d)

(b) (e)

(c) (f)

Figure 2.2 Fabrication process of the CMOS MEMS foundry

2.4 Design flow of CMOS MEMS sensors

The process design flow for fabrication of MEMS sensor and CMOS integrated
circuit (IC) is shown in the following Figure 2.3. Analysis of analog circuit behavior is
done by H-Spice, Spectre and ADS software, analysis of simulated sensor behavior is
done by COMSOL or ANSYS which is combined with related sensing principles. Both
the circuit layout and sensor layout use the same CMOS material layers. However, in the
layout verification part, the sensor must pass DRC at most (under some special
conditions can even relax certain DRC rules.)

22
 Figure 2.3 CMOS microsensor fabrication process

This flow chart above is prepared by the author as an easy-to-understand tool to

help any researcher/person to identify and implement the design idea/processes.
Additionally, the detailed CMOS standard U18 1P6M CMOS MEMS fabrication

23
process is also explained in Figure 2.2. In this flowchart, we can see the process flow of
a CMOS sensor or IC designer acknowledging the design principle by identifying the
specification initially. Later, the simulation of the design can be done in H-spice, spectre,
ADS etc. To understand the user interface and experience, we can also run the FEA
COMSOL multyphysics. The layout of the finalised circuit or sensor layout can be made
in CADENCE virtuoso. DRC, LVS, ERC check is necessary before the tape out inorder
to make sure of the proper design functionality. Moreover, another important step that is
needed at final stage is post processing. The final product is then package and testing are
done accordingly.

24
 CHAPTER 3: SELF-HEATING FLOW SENSOR

This chapter will explain about the self-heating flow sensor design in details.

3.1 Transfer block of flow sensor operation

In the following, the authors took the flow sensor design as a main topic to verify
the CMOS MEMS process. The work builds on a fabrication process reported at
Transducers-2021 [55] revealing that CMOS MEMS technology can generates minimal
space utilization micromachined flow sensors. The flow sensor was designed with a self-
heating half-bridge RTD or a hot wire which did not use any heater during the operation.
This concise device size can be in the range of 100 µm and ideally fabricated by CMOS
MEMS process. The amplifier will also be added later to the sensor design. Numerous
micromachined thermal flow sensors have been surveyed for different materials and
applications including battery-free, wireless devices and calorimetric sensors [37, 40-42,
51-53, 168, 169]. The high-quality CMOS foundry services provided by TSMC and
UMC foundry were highly welcome over the past two decades. The CMOS MEMS
foundry is therefore one of the best candidates so far to implement the new MEMS sensor
design and fabrication. Another major advantage for this kind of flow sensor is its low
power consumption [30, 51].
The primary level of strategy implementation is shown below in Figure 3.1. The
design depiction of the sensor is first calculated theoretically by setting up the desired
values which is the conception and construction of the framework that underpins the
major operation plan and its subsequent execution. It decreases the resistance value when
the velocity is gradually increased, creating an unbalance voltage output.

Figure 3.1 Transfer block for the flow sensor operation

25
3.2 Design of self-heating/hot-wire flow sensor
Initially, the design of the flow sensor was done with a self-heating or hot-wire
type where it did not use any additional heater during the operation. Self-heating of RTD
bridge circuit is originally a noise problem and a shortcoming to the thermal flow
sensors. Figure 3.2 shows the chip cross sections for the current CMOS process provided
by TSRI and UMC: (a) typical UMC 0.18 m CMOS MEMS process with isotropic
XeF2 undercut etching for the MEMS open area; (b) alternative way in MEMS region
using shallow-trench-isolation (STI; light blue) to protect polysilicon (dark blue) during
XeF2 etching. We conventionally need to eliminate this noise by ways of reducing
excitation (sensing) current or using pulse measurements [196]. We only adopt a bridge
circuit composing of 2 RTDs R1 and R2 in series as shown in Figure 3.3. The self-heating
voltage Vout is no more a noise but the bias to the RTD half bridge and is necessary for
the device operation. Resistor R1 is on the substrate and R2 is on the MEMS cavity or
the free-standing plate region, which provides a good thermal isolation [55]. T1 is the
local temperature on the substrate, and T 2 on the MEMS cavity with air gap underneath,
accordingly.

(a) (b)

Figure 3.2 Cross sections for the current CMOS process of TSRI/UMC

3.2.1 Working principle

Based on the principle of heat transfer by one-dimensional steady state thermal

resistance analysis of heat conduction between polysilicon on substrate and heat
convection on polysilicon resistor on MEMS cavity. As temperature changes, its
resistance value changes. This work manifests a new concept of MEMS based flow
sensor, which epitomizes the principle of convective heat transfer from an electrically
heated resistive sensing element. Of many flows sensing methods, thermal flow sensors
are simple and easy to implement structures. As temperature changes, its resistance value

26
changes. When the velocity increases, resistance value decreases, thereby creating an
unbalance voltage output. Since the relationship between resistance and temperature is
consistent in this type of sensors, measuring the resistance value across the sensing
element at any given temperature allows the output to calculate with high degree of
accuracy. The CMOS layer are designed by CADENCE software. The material used for
designing the resistors is polysilicon. Appropriate temperature coefficient of resistance
(TCR) value is calculated which corresponds to achieve a high sensitivity to the
temperature changes and that of different flow properties. The CMOS sensor design is
made using U18MEMS process, which is combination of UMC 0.18µm 1P6M process
along with MEMS post process. Our intelligible yet coherent design is unique from the
other existing flow sensors.

Figure 3.3 Self-heating flow sensor design.

3.2.2 One-dimensional thermo-resistance analysis

By the one-dimensional (1D), steady state thermal resistance analysis of heat

conduction about the silicon substrate and the MEMS cavity area, the simplified models
are assumed in Figures 3.3 and 3.4, respectively. The heating or bias power q from the
RTDs diffuses to ambient by ways of the convective cooling on the top and the
conduction on the bottom (“//” means two thermal resistances in parallel). Using Fourier
law of heat conduction [197], the temperature differences T1,2= (T1,2-T) relates the
heat transfer rate q are shown in Eq. (1) and Eq. (2).
𝑇1 −𝑇∞ 𝑉2
𝐿𝑆𝑖 1 𝐿 = 𝑞 = 4𝑅0 (1)
( )//( + 𝑀𝑂 ) 0
𝑘𝑆𝑖𝐴 ℎ∞ 𝐴 𝑘𝑀𝑂 𝐴

27
 The thermal resistance in Eq. (1) can be explained schematically by Figure 3.4(a) and
(c).

𝑇2 −𝑇∞ 𝑉2
𝐿𝑆𝑖 𝐿𝑆𝑇𝐼 𝐿 1 𝐿 = 𝑞 = 4𝑅0 (2)
( + + 𝑐𝑎 )//( + 𝑀𝑂 ) 0
𝑘𝑆𝑖𝐴 𝑘𝑆𝑖𝑂2 𝐴 𝑘𝑓𝐴 ℎ∞ 𝐴 𝑘𝑀𝑂 𝐴

The thermal resistance in Eq. (2) can be explained schematically by Figure 3.4
(b) and (d).

In addition, the heat transfer coefficient ℎ∞ relates to the flow speed 𝑈∞ by the
laminar boundary layer theory Eq. (3) or Eq. (3a) subject to a flat plate under medium
value of Reynolds number flow [198]. Substitute Eq. 3(a) into Eqs. (1) and (2) can
corelate the temperature difference ∆T1,2 and the flow speed 𝑈∞ .

ℎ∞ 𝑥 𝑈∞ 𝑥
𝑁𝑢 = = 0.664(𝑅𝑒 1/2 )(𝑃𝑟 1/3) = 0.664𝑃𝑟 1/3√ (3)
𝑘𝑓 

𝑈∞
ℎ∞ = 0.664𝑘𝑓 𝑃𝑟 1/3 √ (3a)
𝑥

where,
𝐿
Rth,si = 𝑘 𝑆𝑖𝐴 = thermal resistance of silicon
𝑆𝑖

𝐿
Rth,MO = 𝑘 𝑀𝑂𝐴 = thermal resistance of MOS metals
𝑀𝑂

𝐿𝑆𝑇𝐼
Rth,STI = 𝑘 = thermal resistance of STI oxide
𝑆𝑖𝑂2 𝐴

𝐿𝑐𝑎
Rth,air = = thermal resistance of air gap
𝑘𝑓 𝐴

A: Area of chip (300m × 250m)

∆𝑇: Temperature difference.
h: Convective heat transfer coefficient.
Lsi: Thickness of silicon (400 m)
LMO: Thickness of metal (9.76 m)
LSTI: Thickness of STI (0.4 m)
Lca: Thickness of air cavity (45 m)
Kf: Thermal conductivity of air (0.025 W/m K).
Ksi: Thermal conductivity of silicon (148 W/m K)
KMO: Thermal conductivity of metal (237 W/m K)
KsiO2: Thermal conductivity of silicon dioxide (1.3 W/m K)
Nu: Nusselt number

28
Pr: Prandtl number (0.73 for air).
q: Heat transfer rate.
Re: Reynolds number
R0: Heater resistance in ohm.
𝑇1,2 : Surface temperature of positions 1 and 2.
𝑇∞ : Ambient temperature (25C)
𝑈∞ : Freestream flow speed (1-10 m/s)
Vo: Power supply (1.8 V)
: Kinematic viscosity (viscosity/density) of the air at 25C (15.5210-6 m2/s).
x: Sensor bridge length occupied by the sensor on the chip. (167.5 m)

We assume R0 as the design resistance value (1kΩ) of one RTD. As the flow
speed increases across the sensor, both RTDs have temperature drop, but the cooling
effect of R1 on the silicon substrate is better than R2 on the MEMS cavity with air gap
thermal isolation. So, the temperature difference T2=T2-T of R2 is greater than
T1=T1-T of R1. The values of R1 and R2 after application of flow speed can be shown
below.

𝑅2 = 𝑅0 [1 + (𝑇𝐶𝑅)(𝛥𝑇2 )] (4a)

𝑅1 = 𝑅0 [1 + (𝑇𝐶𝑅)(𝛥𝑇1 )] (4b)

Where T1 and T2 are depicted by Eqs. (1), (2) and 3(a).

The output voltage of the half-bridge is therefore approximated as Eq. (5)

following the temperature difference between two RTDs by the voltage divider formula
and the binominal theorem as below.

𝑉𝑜𝑢𝑡 /𝑉0 (0.5)[1 + (𝑇𝐶𝑅)(0.5)(𝛥𝑇2 − 𝛥𝑇1)] (5)

Combining Eqs. (1), (2), (3a) and (5), we can correlate the output voltage Vout
and flow speed U. By this theoretical formulation, the output voltage change shows
the measurable feasibility by the first estimation as 0.125-0.5 mV with 1V DC bias per
1m/s flow speed change if the TCR values are 500-2000 ppm/K.

Based on the UMC 0.18 m standard 1P6M CMOS foundry with MEMS post
process shown in Figure 3.2, the flow sensor was implemented with cavities underneath
the freestanding plates of multiple metals M1-M6 and ILD. Figure 3.4 (a) shows the

29
cross sectional view of polysilicon on substrate ; (b) Cross sectional view of Polysilicon
on cavity; (c) 1D thermal resistance(Rth) model of the RTD R1 with T1 on Si substrate;
(d) 1D thermal resistance model of the RTD R2 with T2 on the floating MEMS area
with air gap [197]. Appropriate TCR value is also calculated using the technology file
provided by TSRI as shown in Figure 3.5. Table 3.1 denotes the specifications of the
flow sensor design consideration. The author moreover predicts the theoretical output
voltage in Figure 3.6.

(a)

(b)

(c)

(d)
Figure 3.4 Cross sectional view of polysilicon on RTD substrate and cavity

30
 Figure 3.5 Plotting the TCR graph

The specifications for making the self-heating flow sensor design are shown in
the Table 3.1.
Table 3.1 Specification of self-heating flow sensor design

Specification Value

Length of R1,2 354 µm

Width of R1,2 3 µm

Free stream velocity 0-10 m/s

Supply voltage 1.8 V and 3.3 V

We used 1D thermal resistance analysis, which is different from a 3D analysis,

and why we are considering the main difference here is that 3D analysis shows the
interaction of individual components with their surroundings, whereas 1D analysis
shows the entire design of a system and the interactions of the different components of
this larger system. This makes a 1D aims simulation excellent for optimizing the design
of an entire system; while a 3D analysis is also relatable as it is used for determining the
optimum, design characteristics of individual components like flow pattern around a

31
solid object or heat transfer with internal cooling air, etc. 1D analysis aim to increase
system efficiency, by helping researcher understand the interaction of the different
components within the system. The theoretical calculation of the flow sensor for it’s
estimated performance is shown in Figure 3.6.

Figure 3.6 Theoretical calculation of the flow sensor

3.2.3 Fluid structure simulation

2D CFD simulations were carried out in ANSYS fluent for analyzing near and far
field flow physics around the thermal flow sensor. It helps to determine the velocity
distribution on the surface of resistor R1 and R2, and will be further used for calculating
average convective heat transfer coefficients. SIMPLE method with second order
upwind scheme has been implemented for this model, a residual error value of 10−6 is
selected for achieving stable and converged numerical solutions. Figure 3.7 represents
the boundary conditions for the 2D CFD simulation of thermal flow sensor, installed in
a wind tunnel facility mentioned in section 3.3.

Figure 3.7 Boundary conditions for the 2D CFD simulation

32
 A fluid domain of size 100 cm × 30 cm similar to wind tunnel dimension is
designed within which the complete package consisting of flow sensor wired on a
breadboard has been kept at an inclination of 60° with respect to the incoming flow
stream. Inlet velocity varies from 1 m/s to 15 m/s and the pressure at the outlet is assigned
a value of zero. CFD simulations of the flow sensor shows the variations in the flow
velocity across the geometry, velocity increases as the streamlines gets converged when
encountering edges of printed circuit board (PCB) and breadboard. It is followed by the
formation of wake and vortices at the corners of PCB surface as shown in Figure 3.8
path lines representing velocity field around the flow sensor (far field view). Figure 3.9
is the path lines representing velocity field around the flow sensor (near field view), and
Figure 3.10 is the path lines representing velocity field near the PCB surface and flow
sensor respectively.

Figure 3.8 Velocity field around the flow sensor (far field view)

33
 Figure 3.9 Velocity field around the flow sensor (Near field view)

Figure 3.10 Velocity field near the PCB surface and flow sensor

A boundary layer is formed on the surface of PCB and hence on the surface of
resistors R1 & R2. Velocity field decreases significantly while going from far field to
near field of the flow sensors due to the formation of boundary layer. Though the
geometry is inclined on the flow direction, the flow stream appears to be incoming in a

34
parallel direction towards the resistors. To calculate the velocity distribution on the
surface of resitor R1 and R2, a line plot which is kept normal to the surface of resistors
was employed to know the flow velocity variation within the boundary layer. line plot
describing velocity variation along the line normal to PCB surface near the leading edge
of resistors helped to determine the average magnitude of freestream velocity. Free
stream velocity can also be obtained by studying the velocity profile at any point on the
surface of the resistors. In this study velocity profile is obtained at the trailing edge of
the resistors where the length of the boundary layer thickness is maximum.
To study about the velocity boundary layer on the surface of resistors, parallel
lines normal to the surface of resistor were drawn having an equal finite gap in between
two successive lines. Velocity profile ahead of the leading edge was also observed. Line
plots normal to surface of resistor R1 and R2 is shown in Figure 3.11. Thickness of the
boundary layer increases as the distance from the leading-edge increases, it represents
the velocity profile on each line. An iterative method is developed to calculate the free
stream velocity on the surface of the resistors. The method is based on assuming a
boundary layer thickness at the trailing edge of the resistor R1 and R2 surface initially
then performing CFD simulations to get the assumed free steam velocity. Boundary layer
thickness is recalculated using boundary layer thickness relations derived using
momentum integral equations. If the recalculated boundary layer thickness matches with
the initially assumed boundary layer thickness value, the free stream velocity is
considered for further evaluation if not the assumptions are corrected and the whole
process is repeated.

Figure 3.11 Line plots normal to surface of the resistor R1 and R2

35
 Free stream velocities calculated for the surface of resistor for different inlet
velocity conditions is given in the Table 3.2. In addition, Table 3.3 shows output voltage
w.r.t inlet velocity after CFD simulation. Average Nusselt number has been calculated
using the actual free stream velocity obtained from the section. Calculation of output
voltage requires average convective heat transfer coefficient which can be deduced from
the Nusselt functions. Output voltage was calculated for each of the inlet velocity shown
in Table 3.3. Output voltage of the CFD simulations is plotted in Figure 3.12.

Table 3.2 Actual free stream velocity observed after CFD simulation
Wind tunnel inlet Velocity (m/s) Velocity near the sensor chip (m/s)

1 0.1388

5 1.1204

10 1.5695

15 2.0610

Table 3.3 Output voltage w.r.t. inlet velocity after the 2D CFD simulation
Wind tunnel inlet Velocity (m/s) Output volatge after CFD simulation (V)

1 0.9217

5 0.9166

10 0.9156

15 0.9147

36
 Figure 3.12 Output voltage profile for the 2D CFD simulation

3.3 Sensor implementation and testing

3.3.1 Layout and tape out

The layout diagram of the self-heating flow sensor design is shown below in
Figure 3.13. As the available space we can use is more than sufficient, we have added a
dummy sensor and heater across the empty region to utilize the total 1.5 mm × 1.5 mm
which is permitted for free educational foundry service. Figure 3.14(a) denotes the SB
18 pin package provided by TSRI. Figure 3.14(b) represents the image taken after the
tape out, i.e, by using confocal microscope.

Figure 3.13 Layout diagram of CMOS flow sensor

37
 (a)

(b)

Figure 3.14 (a) SB 18 connection; (b) confocal optical microscope image

3.3.2 Scanning electron microscope (SEM) inspection

A scanning electron microscope (SEM) is a type of electron microscope that

produces images of a sample by scanning the surface with a focused beam of electrons.

38
The high-quality images were captured in Tamkang University Chemistry Engineering
Department. Figure 3.15 shows the SEM electronic machine that was used to collect the
images.
Figure 3.16(a) shows the 8 m  8 m etch-holes which measured spacing is
averaged as 9.09  0.28 m on CMOS layer. It also shows how smooth the sidewall or
the trench wall of 8 m  8 m etch-holes. The etch-hole is with the shape of a nozzle
and has somewhat grass residue accordingly. The whole MEMS open area is actually a
freestanding plate membrane mesh.

Figure 3.15 Scanning electron microscope (SEM) machine

Figure 3.16(b) shows the SEM photo of the so-called MEMS open area or the
cavity after the isotropic XeF2 undercut etching provided by TSRI/ UMC. (All over the
MEMS open area, thousands of 8 m  8 m square etch-holes were designed for XeF2
to go through and etch the cavity underneath this MEMS open area.) The etched depth
by XeF2 is measured as 44.94 m, almost 4 times of the CMOS thickness 11 m and is
large enough to ensure the fully freestanding of the MEMS structures without any spikes
or pillars underneath. This etched depth also allows the lateral separation distance
between two square etch-holes can be as far as 20 m so that the sensor connection lines

39
can zigzag among the etch-holes on the levitated MEMS open area. In Figure 3.17 (a)
and (b), cross sectional view of the chip was captured using confocal optical microscope.
Of course, the images captured in SEM were better than the confocal optical microscope.

(a)

(b)
Figure 3.16 (a) Holes on MEMS open region; (b) depth of the cavity

40
 (a)

(b)
Figure 3.17 Cross sectional view of chip using confocal optical microscope

3.3.3 Wind tunnel test

Thermal flow sensors design was made using CADENCE software. The
designed, fabricated CMOS MEMS flow sensor after packaging is validated by keeping

41
the sensor packaged chip in a low-speed wind tunnel and calibration is done for the wind
speed ranging from 0-15 m/s. Figure 3.18 represents (a) sensor on big PCB board; (b)
reduce weight and size PCB board with minimal wire connections; (c) sensor inside the
wind tunnel with wires connection; (d) Sensor board on an inclined stage in the wind
tunnel. The flow sensor setup in Figure 3.18 is apparently output signal changed with
wind velocity under the inclined angle more than 60. The wind tunnel test section of 30
cm × 30 cm × 100 cm.

(a) (c)

(b) (d)
Figure 3.18 Wind tunnel experimental set up and testing

With application of the bias voltage of 1.8 V for the half bridge RTDs (1 K
each), the overall self-heating power is 1.62 mW from the RTDs themselves. The nearby
center heater dissipates only 0.324 mW and almost has no thermo-fluidic effect on the
flow sensor temperature change. Figure 3.19 (a) and Figure 3.19 (c) presents that is the
theoretical calculation of 1D model along with the experimental two output voltages vs.
wind velocity of the left sensor and right sensor, respectively. Both of the two self-
heating flow sensors are with the same normalized sensitivity of 138 V/V/(m/s)/mW if
counting only one RTD dissipating power of 0.81 mW. Figure 3.19 (b) is the 2D CFD
simulation result shown in Figure 3.12 compared to the claimed resolution towards
micrometer per second and the normalized sensitivity of 155 V/V/(m/s)/mW of Lee’s
work in MEMS’20, the design herein without any signal filtering and amplification
achieved 89% of Ref. [37, 53].

42
 (a)

(b)

(c)
Figure 3.19 (a) 1D theoretical data; (b) 2D CFD simulation; (c) experimental result

The design resistance value and measured resistance value for RTD’s of Figure
3.3 is shown in the following Table 3.4. Table 3.5 is the output reading measured in wind
tunnel test.

43
 Table 3.4 Resistance values
Resistor Design resistance value Measured resistance value
R1 1000 ohm 933±20 ohm
R2 1000 ohm 930±19 ohm

Table 3.5 Output reading measured in wind tunnel test

Wind Speed Output Voltage Left Sensor Output Voltage Right Sensor
(m/s) in Fig. 3.14 (a) (V) in Fig. 3.14 (a) (V)
0 0.944 0.925
5 0.943 0.924
10 0.942 0.923
15 0.941 0.922

According to theoretical values of Figure 3.19(a), the output voltage of the

fabricated flow sensor by UMC 0.18 m MEMS process was verified by a wind tunnel
facility in Figure 3.18(a-d). Preliminarily, the theoretical sensitivity of 113
µV/V/(m/s)/mW is near to the measured sensitivity 138 µV/V/(m/s)/mW under the flow
speed of 15 m/s; the linearity of the measured output voltage is obviously better than the
theoretical one. The reason why the sensitivity of the theoretical output voltage was
underestimated is that the actual flow speed in table 3.2 was also lower than the
freestream speeds 𝑈∞ . The amount of flow speed retarding from the freestream inside
the boundary layer of the flow sensor surface is furthermore investigated by FEA like
ANSYS in Sec 3.2.3 and the corresponding graph is plotted in Figure 3.19 (b).

Restated, the reasons why the measured data (linear output) is different from the
nonlinear output of theoretical prediction is that:
1. The theoretical prediction of 1D model has a singular point at zero flow speed.
The theoretical model is more proper for the medium speed range (>5m/s.)
2. The measured output of the zero-flow speed could be also ignored as well. Only
the data at 5, 10, 15 m/s were reserved.

44
3. For the flow speed range U of 5-15 m/s, we did the straight-line fitting for both
the above two figures. It found that the sensitivity for the theoretical 1D model
is 113 µV/V/(m/s)/mW and for the measure data is 138 µV/V/(m/s)/mW.

4. Now the sensitivity value of the theoretical 1D model was underestimated than
the measured data. The excuse may be that the actual flow speed above the sensor
chip in Table 3.2 was lower than the free-stream speed U due to the package
issue.
5. As we know in the 1D model, the smaller speed predicted by the 2D simulation
will give larger voltage output and sensitivity values as Table 3.3 and Figure
19(b) and a larger sensitivity than 113 uV/V/(m/s)/mW by the 2D simulation can
be obtained and closer to the measured sensitivity of 138 uV/V/(m/s)/mW.

Another concept of thermo resistive calorimetric MEMS-based flow sensor

design was made using CADENCE software in collaboration with HKUST Professor
Y.K Lee. The material used for designing the resistors and heater is polysilicon. This
CMOS sensor design is realized by using U18MEMS process, which is a combination
of UMC 0.18µm 1P6M process along with MEMS post process. The sensor comprises
of two RTDs connected in series as bridge circuit. Measuring the resistance value
across the sensing element at given temperature or constant heating power allows the
output to reversely predict the flow speed with acceptable accuracy. In this case the
sensitivity is measured to be 160 µV/V/(m/s)/mW. The explanation of the calorimetric
design is briefed in Appendix A.

45
 CHAPTER 4: ON-CHIP AMPLIFIER FOR FLOW
SENSOR

The instrumentation amplifier (In-Amp or IA) shown in Figure 4.1(a) can

amplify the output signal of the above-mentioned CMOS MEMS flow sensors. MEMS
flow sensors are often based on temperature detection. For most materials, the electrical
resistivity changes with temperature. IA is a type of differential amplifier that has been
outfitted with input buffer amplifiers, which eliminate the need for input impedance
matching and thus making the amplifier particularly suitable for use in measurement and
test equipment. Additional characteristics include very low DC offset, low drift, low
noise, very high open-loop gain, very high common-mode rejection ratio (CMRR), and
very high input impedances. IAs are used where great accuracy and stability of the circuit
for short-term and long-term are both required [199-201]. Using materials with
appropriate TCR permits us to achieve a high sensitivity to temperature changes and thus
to flow speed. Integration on the same chip will contribute to more precise output,
sources of error will be minimized, and fault prone solder points will be eliminated [202].
The IA layout consists of three operational amplifiers circuits in Figure 4.1(a),
sometimes has the difficulty in equalizing two R1~R3 values in the layout of Figure 4.1(b)
for real CMOS fabrication. It therefore gives a testing result of 30dB gain only under the
bandwidth of 10 kHz; the DC level also deviate from the design value of V0/2. With the
supply voltage as V0=1.8V and reference voltage as 1V, the total current consumption is
1.335mA.
The half-bridge output voltage of Eq. (5) has a DC level of 0.5V0. Although it
does not affect the sensitivity, the DC level should be reset to zero. Using the IA of
Figure 4.1, the DC level is easy to eliminate by assigning V1=sensor output voltage and
V2= 0.5V0. The pre and post layout simulation were also done during the design of IA.
By the resistor values designed in Figure 4.1 (b), the voltage gain of the IA is
about 50,000. Actually, the initial voltage gain was designed as 10,000 as below:
Final sensitivity = (voltage gain) (sensor sensitivity)
= (10000) (100 µV/V/(m/s)/mW
= 1 V/ V/ (m/s) /mW
A safety factor of 5 (G = 10000→ 50,000) is prepared in advance to deal with
the degradation or uncertainty happened in the IA implementation.

46
 vo∪t 𝑅3 𝑅1
G =v = (1 + 2 𝑅 )
2 −v1 𝑅2 𝑔𝑎𝑖𝑛

G = 50,000

(a)

(b)

Figure 4.1 (a) IA circuit diagram; (b) CADENCE schematic diagram

4.1 Instrumentation amplifier pre and post design

simulation
The main difference between pre-layout and post-layout is that pre-layout
simulations take place before completing the layout design, while post-layout

47
simulations use the completed layout as their basis. In pre-layout and post-layout, we
translate physical parameters into circuit elements and other mathematical models for
simulation. However, for pre-layout simulation, we must build up a circuit schematic to
include all elements of the simulation. Post-layout simulation involves extraction of
physical information from the routed layout dseign. We add models for the ICs,
connectors, and other connected components to run simulations. Moreover, both types
of simulations have the same type of simulation results.
For the amplifier circuit, it is observed to have gain about 68.8 dB. Voltage gain
≈ 3000 below 20 Hz of the flapping frequency. Even though this voltage gain is much
smaller than the initial design value of 50,000 (the reason is under investigation), the
simulated gain still works well to amplify the sensor output to a level of easy detection
about the low-speed air flow. In Figure 4.2, it can be also found that the magnitude decay
rate is about -20 dB/decade, that is the feature of a 1st order dynamic system with
experimental decay behavior. From Figure 4.3, the phase lag of IA at 20 Hz (flapping
frequency of FWMAV) is less than -13.5° and not serious. However, if we need a larger
sampling rate to pick up the unsteady lift/thrust signals, e.g., 200 Hz, the phase lag would
get worse to -67.5°. AC response for the magnitude, phase margin and common mode
gain of the amplifier circuit for pre layout simulation and is shown below as Figure 4.2,
Figure 4.3 and Figure 4.4 respectively. The simulation for analog design environment
was done in HSPICE/SPECTRE simulator. The available schematic editor is
Composer/ADS. The layout design is made in Virtuoso is verified in Calibre.

48
Figure 4.2 Pre layout simulation AC response (magnitude)

Figure 4.3 Pre layout simulation ac response (phase margin)

49
 Figure 4.4 Pre layout simulation common mode gain
For the amplifier circuit, the post layout simulation of AC Response for
magnitude, phase margin and common mode gain of the amplifier circuit are shown in
Figure 4.5, Figure 4.6 and Figure 4.7 respectively. Common mode voltage gain results
from the same signal being given to both the inputs of an op-amp. If both signals flow
in the same direction, it creates common mode interference, or noise.

Figure 4.5 Post layout Simulation AC Response (magnitude)

50
Figure 4.6 Post layout simulation AC response (phase margin)

Figure 4.7 Post layout simulation common mode gain

51
4.2 Layout and tape out of instrumentation amplifier
The layout diagram of the IA is made using U18 MEMS process in CADENCE
software. Figure 4.8 (a) is the layout design of single stage operational amplifier; Op-
Amp Figure 4.8 (b) is the layout design of Instrumentation amplifier using three
operational amplifiers. Again, due to the difficulty in equalizing two R1~R3 values of
Figure 4.1(a) in the layout of Figure 4.8 for real CMOS fabrication, the DC level cannot
be exactly eliminated. This deficiency can be observed from the output voltage of the
amplified output voltage of the flow sensor will be discussed more. Figure 4.9 (a) shows
layout design of integrated IA, self-heating and calorimetric flow sensor; Figure 4.9 (b)
shows self-heating, calorimetric flow sensor and IA integrated together in a single tape
out chip with tape out SB 40 pin package of TSRI.

(a)

(b)
Figure 4.8 (a) single stage Op-Amp layout; (b) IA layout

52
 (a)

(b)
Figure 4.9 (a) integrated chip layout; (b) tape out SB 40 package of TSRI

4.3 Instrumentation amplifier gain test

The fabricated IA were tested in Tamkang University Electrical Engineering
Laboratory under the supervision of Professor H.-Y. Shih. Figure 4.10 (a) and (b) shows
the pin out diagram to check the gain of the amplifier and set up to test the IA gain
performance using power supply, multimeter, waveform generator and oscilloscope.

53
(a)

(b)
Figure 4.10 (a) pin out circuit diagram IA; (b) set up to test the IA
Frequency gain measured in cathode ray oscilloscope (CRO) is observed to have
30 dB gain only under the bandwidth of 10 kHz, the DC level also deviate from the
design value of V0/2. With the supply voltage as V0=1.8V and reference voltage as 1V,
the total current consumption is 1.335mA. IA gain is proportional to the ratio between
resistors. In Figure 4.1(b) we have 500kΩ and 1kΩ resistor. The resistance with larger

54
value occupies larger area and leads to huge process variation in semiconductor
processing. If it changes its value after the fabrication, therefore it contributess to
variation in measured gain value. Another reason for decrease in gain could be due to
1/f noise. Figure 4.11 represents the graphical plot of the frequency gain. The IA design
alone occupies the area of 250 × 350 m2 whereas the total area occupied by the IA and
sensor integration is 600 × 550 m2.

Figure 4.11 Frequency gain measured oscilloscope

The voltage gain of 30dB in Figure 4.11 is very different from the simulated gain
of 68.8 dB in Figure 4.2. The magnitude decay rate in Figure 4.11 is only -5.5 dB/decade,
also much smaller than the simulated value of -20 dB/decade in Figure 4.2. Beside the
possible reasons of resistance mismatch and the 1/f noise, more investigations on the
voltage gain variation are needed in the future.

4.4 Advantages summary

• Smaller device size
Technology has pushed to the point that we can build a device so small that it
can hardly see with our human eye. The typical size of MEMS devices is usually
measured in micrometers and nanometers. Using similar fabrication techniques as
building ICs, people are now able to build sensors and actuators on the same microscopic
level with the processor chip. Measured in micrometers, almost all MEMS sensors and

55
actuators can be batch produced together on the same chip with circuitry. They indeed
compose a system on a chip as small as possible according to the current CMOS
technology [181].
The flow sensor is made with the latest U18MEMS CMOS MEMS process which
was discussed in Chapter 3. The dimensions of the flow sensor were miniaturized to save
the chip area and achieve acceptable sensitivity and linearity. The sensor size herein is
300m × 250m in Figure 3.3 and self-heating part which is smaller than many prior
arts. Within the flow speed range of 0-15 m/s in a wind tunnel, a normalized sensitivity
of this small CMOS MEMS sensor was obtained as 138 V/V/(m/s)/mW which is with
the same order of magnitude to the best value of 160 V/V/(m/s)/mW of the previous
works [37, 40, 52, 53, 168, 169, 203-205].

• Less power consumption

With enhanced sensitivity and less power consumption with only one RTD self-
dissipating power of 0.81 mW, we can expect: In a digital viewpoint, the small voltage
swing results in small noise margin. In analog domain, the strength of signal level is
weakened due to the low supply voltage. The total power consumed by the flow sensor
with IA is 3.24 mW. Other design such as calorimetric type flow sensors consumes more
power due to the additional center heater for adjusting the output performance [206].

• Enhancing the sensing sensitivity by on-chip amplifiers

Adding on-chip amplification circuits to the RTD flow sensors may results in a
different behavior of the sensor output sensitivity compared to the one without amplifier
circuitry. Addressing the above challenge requires collaboration of MEMS and analog
circuit design engineers. The case study of the self-heating RTD flow sensor alone gives
a normalized output sensitivity of 138 µV/V/(m/s)/mW under the flow speed of 15 m/s.
With the measured gain of IA as 30dB calibrated at 10kHz, and by integrating it into the
self-heating RTD sensor, the overall sensitivity is expecetd to be improved and shown
in the next chapter.

4.5 Other technical issues of CMOS MEMS process

In the surface micromachining, reducing the residual stress of the microstructure
is an issue involving the warping of the microstructure after its releasing and suspension

56
by removing the sacrificial layer, or even causing the adhesion or stiction to the silicon
substrate. The phenomenon of the device failure is that the suspended micro structures
after CMOS fabrication and post-processing have some unavoidable residual stress
where the CMOS process is not able to adjust the process parameters as the polysilicon
structure layer is grown in the LPCVD of surface micromachining. The residual stress
problem can be minimized by avoiding the cantilever structure (because the free end will
arbitrarily turn up or bend down) or I-beam configuration [207]. Instead, use a thin
bridge or membrane design to resist the residual stress by the rigidity of the structure.
The residual stress state of each structural layer of CMOS materials, including
information such as tensile stress or compressive stress, magnitude etc. cannot be
obtained from the technology file of the foundry service. Therefore, a concept of “test
key” may be necessary to develop the mechanical parameter monitoring technology of
the semiconductor process by an onsite sensing way [144, 164, 208]. The microstructure
of the in-situ gauge or test key showing mechanical behavior such as residual
stress/strain and mechanical properties such as Young's modulus and Poisson's ratio.
Similarly, the resistivity or TCR values of polysilicon and metal layers in CMOS
layers cannot be identified very accurately from the technology file of the foundry
service, either. This is the reason why the theoretical sensitivity of 1D thermal resistance
model in Figure 3.19(a) deviates 18% away from the measured data in Figure 3.19(c) so
obviously. For perfecting the CMOS MEMS sensor design in the future, allocating some
small area with test keys for in-situ measuring the resistivity and TCR values are
mandatory.
However, from the various testing process, our research team found some major
drawbacks as below:
1. We need to tilt up the sensor to solve flow speed retarding inside the boundary
layer of the sensor package.
2. Fine particles get accumulated on the surface of the mesh structure of the MEMS
region which in return decreases the sensitivity or even malfunctions the sensor.
3. Electro static discharge (ESD) problem reduces the lifetime of the sensors.
ESD is the most prominent single event, as about 35% of chip failures can be
attributed to such an incident. Modern and future technologies call for complex and
customized ESD protection solutions that have small footprint, high robustness, low
loading effect, high latch-up immunity, and low cost. ESD design windows are shrinking

57
and ESD design costs are increasing dramatically with decreasing technology node
(thinner gates, shallower junctions). The operational voltage, oxide breakdown voltage,
determines ESD design window thermal failure effects. The amount of charge depends
on several factors, one of which is the humidity. Three requirements for ESD occurrence:
source of charge (i.e. human body), source of storage (i.e. metal knob), and conducting
paths between the two i.e skin. The severity of ESD (magnitude of ESD induced current)
depends on the amount of charge (measured in voltage), amount of storage (measured in
capacitance), and the resistance of the conducting path. Two types of ESD protection
scheme are there namely off chip and on chip. On chip ESD is a must have due to the
handling and assembly at the component level, and the off chip is optional and is often
added to enhance the ESD robustness at the system level. On chip ESD protection
objectives:
1. Shunt ESD current away from core circuit.
2. Clamp I/O & power pins voltages to a safe level
All ESD devices must operate within a window defined by the operating voltage
(lower bound) and breakdown voltage (upper bound). ESD devices require high
robustness and transparency. Frequently used ESD protection devices in
CMOS/BiCMOS technologies are (a) SCR: Robust but difficult to scale for different
ESD requirements, (b) ggNMOS/PMOS: simple structure, but gate oxide can be
damaged by high voltage, (c) diodes: Most widely used I/O protections. Simple
structure, but high leakage current, very small trigger voltage, no snapback, and large
aera. The above mentioned several techniques are surveyed from (ESD) protection for
low-voltage consumer electronics presented by Prof. Juin J. Liou from University of
Central Florida.
For time being, our research team have used ESD band to protect the CMOS chips
from further damage during the actual experimental test in the wind tunnel and the
FWMAV.

58
CHAPTER 5: FLOW SPEED MEASUREMENT OF
FLAPPING WING

A flapping wing micro air vehicle (FWMAV) demands high lift and thrust
generation to achieve good flight maneuverability. Researches regarding the FWMAV
improvement need the on-site lift information and measurement. In this case,
microelectromechanical systems (MEMS) flow sensors on a flapping wing may be
helpful in detecting the real-time downwash and lift. The conventional MEMS flow
sensor design is composed of a floating plate with resistive temperature detectors (RTDs)
heated by another heater aside, both made by polysilicon. By using 0.18 m-linewidth
complementary-metal-oxide-semiconductor (CMOS) IC foundry with MEMS post
processing provided by UMC/ TSRI Taiwan, the sensor chips were implemented. We
adopt the self-heating of the RTD half-bridge with the area of 300×250 m2 for sensor
functionality and omit the heater aside. This new sensor design generates a normalized
output sensitivity of 138 V/V/(m/s)/mW within the speed range of 0-15 m/s in a low-
speed wind tunnel. In addition, instrumentation amplifier is added beside the flow sensor
to form a monolithic integrated circuit. The gain of the amplifier is measured to be 30dB.
After being integrated with this IA, the overall sensitivity of the flow sensor was hope
to be substantially improved. With this new design, a test is being conducted to measure
the flow speed of a FWMAV. A flapping mechanism was used for the FWMAV to do
the flapping with wing rotation. In this chapter, the author apply the flow sensor on wing
membrane to measure the lift and net thrust during downstroke and upstroke of a MAV.
Figure 5.1 shows the idea about flow measurement during downstroke flapping and
upstroke flapping with sensors mounted on the wing surface.

59
 Figure 5.1 Flow sensor mounted on the wing surface

5.1 Sensor mounted on flapping wing using package of

TSRI/UMC
Researches regarding the FWMAV improvement need the on-site lift
information and measurement. In this case, MEMS flow sensors on a flapping wing can
be helpful in detecting the real-time downwash and lift. Furthermore, this type of
CMOS-MEMS flow sensor will be promising as a measuring device for various ranges
of wind speeds and different kind of applications [24, 209]. It became utmost important
for this type of flow sensors to emerge its design in applications for on-site lift
measurement in MAVs, wind turbines, biomedical purpose, and so on. Lift data can be
observed from the downstroke and upstroke motions of a full flapping cycle. There are
several ways for lift improvement, for instance, by increasing the wing foil thickness and
the camber [77], by adding wing corrugation [64], by enlarging the stroke angle [67],
using wing rotation mechanisms [60], using morphing wings [78, 210, 211], by changing
the wing materials [72, 76, 212], or by adjusting the wing stiffness along the chord-wise/
span-wise direction [213]. Figure 5.2 (a) shows several packaged IC chips to measure
the flow speed of a FWMAV. Figure 5.2 (b) is the SB 18 Package dimensions.

60
 (a)

(b)
Figure 5.2 (a) several IC chips; (b) SB 18 Package dimensions

5.2 Sensitivity Test

5.2.1 Wind tunnel experiment on the integrated flow sensor

As mentioned previously, we have ever obtained the sensitivity of the self-heating

flow sensor as 0.138 mV/V/(m/s) by testing in the wind tunnel facility available in our
laboratory. Herein the author continued to do the wind tunnel flow testing and sensor
calibration with one result exemplified as the voltage output vs. flow speed. The offset voltage

61
V0ffset is 1.178 V and the sensitivity (negative slope) S is 1.388 mV/V/(m/s). These values will
be depicted in the following.The linearized formula for the amplified flow sensor to measure
air speed u is

Vout=V0ffset - S·u (6)

For all the packed flow sensors from TSRI, we cannot guarantee all of them to have
the same offset voltage V0ffset and the same sensitivity S. The reasons may come from the
resistance variation during the CMOS MEMS fabrication and packaging process. We also
found that our packaged flow sensors are very weak to resist the degrading issue attacked by
air moisture, particles and electrostatic discharging damage (ESD) subject to the direct contact
of our sensor with the ambient air.

Finally, the ideal monolithic MEMS device, although it can be directly

implemented in a CMOS foundry service, has some problems that need to be solved.
The biggest difficulty lies in the integration of the circuit functions because the IC design
software cannot directly simulate the output signal changes caused by the MEMS sensor.
For example, the sensing capacitor or RTD is only a value-fixed capacitance or
resistance in Cadence. H-Spice may predict the relationship between the pre-drive bias
voltage and the output voltage but cannot capture the model in which the sensing
capacitance or resistance changes. The so-called "on chip" in the circuit function still
lacks the real integration with MEMS sensors. It means that only the MEMS sensing
element and the signal processing circuit are fabricated on the same silicon chip
physically (; for the CMOS MEMS chip, the process and the material are already
compatible.) However, there is no circuit model available for the MEMS devices.
Therefore, the overall system optimization of MEMS and IC devices cannot be done yet
in the Cadence software so far. In the existing operation of the Cadence software, the
signal processing circuit first completes all the verification procedures, and then it is put
together with the layout of the MEMS sensor that is verified by the DRC only. Finally,
it was sent directly to the foundry production. After the foundry service, the basic
functions of ICs are firstly tested, and then the “hybrid” connection is performed nearest
to the wire bonding pads to realize the physical sensor test totally. For the successful
integration of MEMS and circuitry function by CMOS MEMS in the future, there are
several difficulties needed to be overcome [42, 143, 154, 214-221]. The experimental
testing for the final integrated flow sensor chip is done as shown in the Figure 5.3. The

62
authors have plotted the graph, which is achieved after the integration test of the
amplified flow sensor output as shown in Figure 5.4.

Figure 5.3 Integrated flow sensor and IA test

The calculated sensitivity value for self-heating flow sensor according to the
Table 5.1 and Figure 5.4 graph is 1371µV/V/(m/s)/mW considering power dissipation
as 0.8 mW. Table 5.1 shows the output voltage reading of integrated chip for self-heating
flow sensor and IA for wind speed 0-6 m/s.

Table 5.1 Output voltage reading for wind speed 0-6m/s

Wind speed (m/s) Vout (V)
0 1.180
1 1.173
2 1.173
3 1.175
4 1.169
5 1.170
6 1.169

63
 Figure 5.4 Sensitivity graph for self-heating sensor with IA

Sensitivity is improved to 1371 µV/V/(m/s)/mW for flow speed ranging from 0-

5 m/s in Figure 4.6 which is plotted with reference to Table 5. The overall voltage gain
is 20 dB smaller than the calibrated value of 30dB. In other words, CMOS flow sensor
with on-chip IA was observed to have 10x better sensitivity. Table 4.1 shows the
sensitivity and power consumption of the flow sensor without any readout circuit and
with amplifier circuit.

Table 5.2 Sensitivity and power consumption summary

Flow sensor Sensitivity(µV/V/(m/s)/mW) Power consumption(mW)

Without amplifier 138 1.62

With amplifer 1371 4.86

64
5.2.2 Application of flow sensor to flapping wing

Figure 5.5 shows cutting of PCB board to use the minimal required area. After
inserting the chip on the PCB board, the total weight along with the wires is 9.02 g. Figure 5.6
shows the IC chip being mounted on the PCB board with wire connections to put on MAV.

Figure 5.5 PCB board to use the minimal required area

Figure 5.6 IC chip being mounted on the PCB board

65
 To analyze the experimental data by the application of self-heating flow sensor
to flapping wing of 70 cm-span, is shown in Figure 5.7(a-b). The flapping wing was
independently controlled by a pair of servo motors with the flapping frequency below 2
Hz. Two of the CMOS MEMS integrated flow sensors developed in this work were
respectively time-averaged with the wind speed as Figure 5.8 (a-b) and were mounted
on the ribs of the flapping wings marked with A1 and A2 shown in Figure 5.7 (a).
Similarly, another sensor arrangement of Figure 5.7 (b) assigns two of the CMOS
MEMS flow sensors B1 and B2 developed in this project and time averaged with the
wind speed as Figure 5.9 (a-b) which were mounted on the rib and the leading edge on
left flapping wing. The linearized Eq. (6) for different flow sensors were depicted inside
Figures 5.8 (a-b) and 5.9 (a-b). (For easy understanding, the sensor locations are
identified as configuration A1, A2 and B1, B2 in Figure 5.7.).

(a)

(b)

Figure 5.7 Flow sensor with configuration (a) “A1” and “A2”; (b) “B1” and “B2”

66
(a)

(b)

Figure 5.8 Averaged output voltage vs. wind speed for (a) “A1”; (b) “A2”

67
 (a)

(b)

Figure 5.9 Averaged output voltage vs. wind speed for (a) “B1”; (b) “B2”

After the conversion by the linearized Eq. (6), we obtained the local flow speed
history in Figure 5.10 from output voltage data of the flow sensors mounted as Figures
5.7 (a) with configuration “A1” and with configuration “A2” ; we also obtained the local
flow history shown in Figure 5.11 corresponding to output voltage data of the flow
sensors mounted as displayed in Figure 5.7 (b) with configuration “B1” and with
configuration “B2”. Thus, Figure 5.10 shows the amplified flow sensor signal output
from the flapping wing motion (stroke angle of 174°) of Figure 5.7 (a) with configuration

68
“A1” and with configuration “A2”. The flapping frequency: (a)1.23 Hz; (b)1.33 Hz;
(c)1.4 Hz; (d)1.46 Hz; (e)1.5 Hz; (f)1.6 Hz. Similarly, Figure 5.11 shows the amplified
flow sensor signal output from the flapping wing motion (stroke angle of 174°) of Figure
5.7(b) with configuration “B1” and with configuration “B2”. The corresponding flapping
frequency are: (a) 1.23 Hz; (b) 1.33 Hz; (c) 1.4 Hz; (d) 1.46 Hz; (e) 1.5 Hz; (f) 1.6 Hz.

(a) (d)

(b) (e)

(c) (f)

Figure 5.10 Amplified signal output for config. A1 and A2 with diff. frequencies

69
(a) (d)

(b) (e)

(c) (f)

Figure 5.11 Amplified signal output for config. B1 and B2 with diff. frequencies

The two sensors of Figure 5.7 (a) have the symmetrical mounting configuration,
and their output local speeds should be similar to each other. The cases with frequencies
(a) 1.23 Hz and (d) 1.46 Hz reveal this almost identical performance. However, case (e)
1.5Hz shows the worst output, on the other hand. Due to the very limited capability of
DAQ (20 channel signals per second) so far, we only captured 15 points in approximately
two flapping cycles of data. It is not easy to pick up detailed waveforms of the flow
sensor signals and even harder to discuss the periodic behavior accordingly. Data in

70
Figures 5.10 and 5.11 will be checked again with a standard force gauge in the wid tunnel
in the following section.

5.2.3 Comment on conversion from flow sensor output voltage to flow speed
The flow sensor is assumed linear for certain narrow speed range and hoped to
have a constant value of sensitivity during the measurement; the sensitivity value is
important to obtain the flow speed correspondingly. However, in the matter of fact, it is
hard to linearly fit the calibration curve of the flow sensor in its full range for easy
conversion to speed value because the flow sensor output is intrinsically nonlinear. The
sensitivity was found to be largely influenced by the CMOS MEMS sensor chip package
of TSRI and the testing environment. In the future we hope to collect more data points
using keysight DAQ 970A with 20 channels by a better frame rate (improved from 80
channels per second using DAQM901A to 450 channels per second using DAQM900A)
for further analyzing the flow speed feature of the flapping wing and even more of the
wind turbine flow field with higher degree of accuracy.

5.3 Packaging issues of MEMS flow sensors

So far, the flow sensor chips after performing several experiments, their
performance starts to degrade or even being invalid. This problem may be due to the
accumulation and choking of many thermo-conductive micro-scale particles inside the
MEMS etch holes. Therefore, that the thermal isolation of the MEMS cavity vanishes
and the RTD change is no more. Comprehensive research studies also found that flow
sensor’s proper packaging will be necessary to achieve accurate flow measurement to
prolong the lifetime of the sensor devices and hence, result will be accomplished with
even better sensitivity in the future.

•Dummy sensor test

The dummy flow sensor, which was kept on the upper half of the MEMS region
without thermo-isolated MEMS cavity in Figure 3.13, is also tested by the wind blowing
in the wind tunnel. No signal changing according to wind velocity was found because of
the same cooling condition or convective behavior of the two RTDs of the self-heating
half-bridge, i.e., T2 = T1.

71
•Orientation test
The author also tested the orientation effect of the self-heating flow sensor. In
other words, we changed the airflow direction in from “west→east” to “north→south”.
The output sensitivity degradation or variation is found about 20-40%. The output
performance variation denotes the intrinsic influence by the packaging issue of flow
sensors [52, 54, 149, 169, 202, 222].

•Aging observation
So far, all the CMOS MEMS flow sensor chips are without protection epoxy or
functional layer on the chip and pad surfaces, and the bonded wire oxidation and
moisture penetration issues cannot be avoided. Keeping in account the phenomenon of
aging, the sensor was tested time after time. Hence the good sensitivity output of the
sensor only lasts for 6 months. The aging effects are normally due to accumulation of
particles inside the measuring chamber, conditions such as high humidity or extreme
temperatures. MEMS device is subjected to vibration, high or low temperature, thermal
and moisture cycling, etc. Comprehensive research studies also found that flow sensor’s
proper packaging [52, 54, 149, 169, 202, 222] will be necessary to achieve accurate flow
measurement to prolong the lifetime of the sensor devices and hence, result will be
accomplished with better sensitivity.

5.4 Experimental findings

5.4.1 Flow sensor output speed vs. lift or net thrust from force gauge

Table 5.3 and Figure 5.13 shows the lift force vs. flapping speed according to
different throttle % of power supply. In the following Figure 5.12, it shows the
experimental test set up for lift and net thrust measurement using force gauge (Bertec,
OH USA) on the base [10, 62, 63, 64, 71, 73] without any free stream velocity but only
flapping speed.

72
 Figure 5.12 Lift and thrust measurement using force gauge

Table 5.3 Lift data measured with the MAV using force gauge
Throttle % Flapping speed (m/s) Lift (gram force)
10 0.678 3.0
20 0.715 22.5
30 0.793 24.1
40 0.806 32.2
50 0.825 32.4
60 0.880 32.3

73
 Figure 5.13 Lift measured w.r.t different flapping frequency

The lift and thrust data were measured with the same mechanism earlier used for
flow sensor test by mounting on to the wing, but this time without counting the flow
sensor output on the FWMAV, we just noted the output data of the force gauge. Table
5.3 is the lift data measured with the flapping wing MAV by using force gauge. The lift
data is plotted with respect to different windspeed as shown in Figure 5.13. In addition,
Figure 5.14 shows the lift raw data of FWMAV processed in the MATLAB for flapping
frequency: (a) 1.23 Hz (throttle 10%); (b) 1.33 Hz (throttle 20%); (c) 1.4 Hz (throttle
30%); (d) 1.46 Hz (throttle 40%); (e) 1.5 Hz (throttle 50%); (f) 1.6 Hz (throttle 60%)
respectively. The calibrated MATLAB processed data for the lift measurement in force
gauge (a-f). Calibrated data from the flow sensor on centre of both side of wing of
FWMAV with configuration A1 and A2 for 1.6 s time period is shown in Figure 5.15
(a1~f1). Comments on the lift force comparison is Figure 5.15 are as below:
1. The force gauge data in Figure 5.15 (a), (e) and (f) are saturate values which can
be justified by the lift values of 32 gf in Table 5.3.
2. The flow sensor data in Figure 5.15 (c1) is not consistent for the symmetric
sensor location at A1 and A2 in Figure 5.7 (a).
3 Finally, only the data of Figure 5.15 (a) and (a1) and Figure 5.15 (b) and (b1)
could be compared to each other. However, the periodic lift signals of about 2
flapping cycles in fore gauge do not also shown in the flow sensor case.

74
 Therefore, data in Figure 5.15 was concluded without apparently good
relationship between the flapping lift and the on-site flow sensor output voltage.

(a) (d)

(b) (e)

(c) (f)

Figure 5.14 Lift raw data of FWMAV processed in the MATLAB

75
(a) (a1)

(b) (b1)

(c) (c1)

(d) (d1)

(e) (e1)

(f) (f1)
Figure 5.15 (a)-(f) Lift measured in force gauge and (a1)-(f1) calibrated data of sensor

76
 Table 5.4 shows the thrust data measured with the MAV by using force gauge.
The thrust data is plotted with respect to different windspeed as shown in Figure 5.16.
Figure 5.17 shows the thrust raw data of FWMAV processed in the MATLAB for
flapping frequency: (a) 1.23 Hz; (b) 1.33 Hz; (c) 1.4 Hz; (d) 1.46 Hz; (e) 1.5 Hz; (f) 1.6
Hz respectively. This data is processed in MATLAB software using an LPF (low pass
filter) with a noise cut-off frequency of 5 Hz.

Table 5.4 Net thrust measured with the MAV by using force gauge
Throttle % Flapping wind speed(m/s) Net thrust (gram force)
10 0.678 -7.0
20 0.715 -16.1
30 0.793 -5.3
40 0.806 -20.0
50 0.825 -22.8
60 0.880 -20.6

Figure 5.16 Thrust measured with respect to different flapping frequency

Figure 5.18 shows MATLAB processed data for the thrust measurement in force
gauge (a)-(f) and calibrated data from the flow sensor on centre of FWMAV for (a1)-
(f1) with configuration A1 and A2 for 1.6 s time.

77
 A low-pass filter is a filter that allows signals below a cutoff frequency (known
as the passband) and attenuates signals above the cutoff frequency (known as the
stopband). Low-pass filters, especially moving average filters or savitzky golay filters
are often used to clean up signals, remove noise, create a smoothing effect, and perform
data averaging, and design decimators and interpolators. Low-pass filters produce slow
changes in output values to make it easier to see trends and boost the overall signal to
noise ratio with minimal signal degradation. We can use MATLAB to design finite
impulse response (FIR) based and infinite impulse response (IIR) based filters, two
common low pass filter methods. Which we have standard filters with change in the
range of smoothness of the curve fitting with the original signal data. The original
flapping raw data with respective to the time frame noise reduction sample method. after
importing the raw data into the MATLAB programming platform, we need to make sure
the fast Fourier transform (FFT) function passes the required noise filtering condition
from the imported data. Therefore, from both proper cutoff frequency and the error
percentage we can be able to do the trial-and-error method to identify the perfect noise
reduction from the raw data.

(a) (d)

(b) (e)

(c) (f)
Figure 5.17 Net thrust raw data of FWMAV processed in the MATLAB

78
(a) (a1)

(b) (b1)

(c) (c1)

(d) (d1)

(e) (e1)

(f) (f1)
Figure 5.18 (a)-(f) force gauge net thrust data and (a1)-(f1) calibrated data of sensor

79
 Table 5.5 shows the FWMAV’s raw data of lift measured in force gauge and
output voltage measured in sensor whereas Table 5.6 depicts the FWMAV’s raw data of
thrust measured in force gauge and output voltage measured in sensor. Figure 5.19 (a) is
the comparison graph of various output voltages subjected to lift force; Figure 5.19 (b)
is the various output voltages subjected to thrust force.
We have omitted the configuration B1 and B2 type data in the main test as the
sensitivity value are not attractive. However, the graph for the same is shown in the
Appendix B.

Table 5.5 Lift measured in force gauge and output voltage measured in sensor
Lift 1st pt 2.158 6.474 3.333 -0.209 4.404
Vout A1 1st pt 1.126 1.126 1.126 1.126 1.126
Vout A2 1st pt 1.126 1.126 1.126 1.126 1.126
Lift 2nd pt 20.269 22.039 25.292 23.798 19.643
Vout A1 2nd pt 1.125 1.125 1.125 1.125 1.125
Vout A2 2nd pt 1.125 1.125 1.125 1.125 1.125
Lift 3rd pt 26.602 27.871 23.933 17.798 18.417
Vout A1 3rd pt 1.124 1.124 1.124 1.124 1.124
Vout A2 3rd pt 1.123 1.123 1.123 1.123 1.123
Lift 4th pt 32.246 36.171 32.255 30.496 33.279
Vout A1 4th pt 1.123 1.122 1.121 1.122 1.120
Vout A2 4th pt 1.122 1.122 1.122 1.121 1.121
Lift 5th pt 32.804 33.267 30.222 30.605 35.887
Vout A1 5th pt 1.122 1.121 1.122 1.121 1.122
Vout A2 5th pt 1.122 1.122 1.122 1.122 1.122

80
 Table 5.6 Net thrust measured in force gauge and output voltage measured in sensor
Net Thrust 1st pt -7.769 -6.357 -7.706 -8.023 -6.926
Vout A1 1st pt 1.126 1.126 1.126 1.126 1.126
Vout A2 1st pt 1.126 1.126 1.126 1.126 1.126
Net Thrust 2nd pt -17.291 -15.836 -15.460 -14.846 -17.356
Vout A1 2nd pt 1.125 1.125 1.125 1.125 1.125
Vout A2 2nd pt 1.125 1.125 1.125 1.125 1.125
Net Thrust 3rd pt -5.052 -3.905 -5.723 -5.848 -7.596
Vout A1 3rd pt 1.124 1.124 1.124 1.124 1.124
Vout A2 3rd pt 1.123 1.123 1.123 1.123 1.123
Net Thrust 4th pt -21.636 -19.660 -20.509 -20.343 -20.375
Vout A1 4th pt 1.123 1.122 1.121 1.122 1.120
Vout A2 4th pt 1.122 1.122 1.122 1.121 1.121
Net Thrust 5th pt -20.963 -24.2 -24.047 -24.131 -22.009
Vout A1 5th pt 1.122 1.121 1.122 1.121 1.122
Vout A2 5th pt 1.122 1.122 1.122 1.122 1.122

Comments on the net thrust force comparison in Figure 5.18 are as below:
1. The force gauge data in Figure 5.18 (d), (e), (f) are saturate values which can be
justified by the net thrust values of 20-22 gf in Table 5.4.
2. The flow sensor data in Figure 5.18 (c1) is not consistent for the symmetric
location at A1 and A2 in Figure 5.7 (a).
3. Finally, only the data of Figure 5.18 (a) and (a1) and Figure 5.18 (b) and (b1)
could be compared to each other. Even though the periodic behavior of the flow
sensor signals in Figure 5.18 (b1) and (c1) are not apparent enough, the signal up
and down behavior between the force gauge and the flow sensor have some
positive correlation.
Therefore, data in Figure 5.18 was concluded reasonably to describe the time
varying signal of the net thrust for a flapping wing.

5.4.2 Summary on flapping wing experiment

1. Standard force gauge data of the flapping wing were collected at the same time
with the CMOS MEMS flow sensor output.

81
2. The sensitivity of the output voltage with respect to the lift is measured and
linearized as 90 µV/V/gf in Figure 5.19 (a). The net thrust is also measured and
linearized as 159 µV/V/gf in Figure 5.19(b). From the sensitivity values of output
voltage, mounting the flow sensor on the flapping wing is better assigned to
detect net thrust rather than lift.
3. The test results in Figure 5.15 (lift data) and in Figure 5.18 (net thrust data) show
that the CMOS MEMS flow sensor signals on the flapping wing is more similar
to the net thrust signal, rather than the lift signal, from the standard force gauge.
4. In conclusion, the CMOS MEMS flow sensor herein is not only sensitive enough
to detect the aerodynamic forces of a flapping wing, but also behaves with a
characteristic of shear stress sensor for detecting the net thrust forces. This
argument matches with the application of MEMS shear stress sensor to detect the
drag force of a fixed wing MAV in Caltech/UCLA work during 1994 to 2005 [2,
4-7, 38, 39].

82
(a)

(b)

Figure 5.19 Graph for various output voltages subjected to (a) lift and (b) net thrust

83
 CHAPTER 6: CONCLUSION

The main objective of this research work is to design and study CMOS MEMS
flow sensor and test the sensitivity in the wind tunnel and later mounted on to flapping
wing to measure its flow speed. Self-heating type flow sensor were developed using 2
RTD’s connected in series. The resistors value is 1 kΩ and the supply voltage is 1.8V.
The CMOS sensor design is realized by using U18MEMS process, which is a
combination of UMC 0.18 µm 1P6M process along with MEMS post process. The layer
design is done using CADENCE software. The material used for designing the resistors
are polysilicon. The author used of STI oxide to protect polysilicon from the XeF2 attack
in the MEMS open region. The authors received the permission from TSRI even though
placing the polysilicon patterns in MEMS region is originally a violation to the design
rule. Appropriate TCR value were considered which corresponds to achieve a high
sensitivity to the temperature changes and that of different air flow properties. In order
to enhanced the sensitivity of the sensor, we have also designed an IA. Using SEM
machine, the etched depth by XeF2 is measured as 44.94 m, almost 4 times of the
CMOS thickness 11 m and is large enough to ensure the fully freestanding of the
MEMS structures without any spikes or pillars underneath and the 8 m  8 m etch-
holes measured spacing has averaged value of 9.09  0.28 m on CMOS layer.

6.1 Summary about the flow sensor

The Self-heating flow sensor design was done with a hot-wire type where it
didn’t use any heater during the operation. We only adopt a bridge circuit composing of
2 RTDs that is R1 and R2 connected in series, R1 is kept directly above the substrate and
R2 is above the MEMS cavity. Based on the principle of heat transfer by one dimensional
steady state thermal resistance analysis of heat conduction between polysilicon on
substrate and heat convection on polysilicon resistor on MEMS cavity, the self-heating
output voltage is no more a noise but the bias to the RTD half bridge and is necessary
for the device operation. As temperature changes, its resistance value changes. When the
velocity increases, resistance value decreases, thereby creating an unbalance voltage
output.

84
1. Preliminarily, the theoretical sensitivity of 113 µV/V/(m/s)/mW is near to the
measured sensitivity. The sensor design alone has a normalized sensitivity of 138
µV/V/(m/s)/mW within the speed range of 0-10 m/s.
2. One RTD dissipates 0.81 mW of power.
3. The sensor size is as small as 300 × 250 m2.

6.2 Summary about the instrumentation amplifier

In the next batch of IC fabrication, Instrumentation Amplifier design was made
and we adopt self-heating flow sensor output signal to the amplifier input. The outcomes
are:
1. The gain of the IA is measured to be 30dB under the bandwidth of 10 kHz. The
benefits of adding amplifier are to improved SNR and lower voltage noise
density.
2. After being integrated with an on-chip amplifier and the sensor, experimental test
was conducted by keeping the sensor inside the wind tunnel, later it is also tested
by mounting on to the wing of the FWMAV. Gain of 20dB was observed with
the integrated-on chip test, the overall sensitivity of the self-heating flow sensor
was measured and substantially improved to 1371 µV/V/ (m/s)/mW for the flow
speed range of 0-5 m/s.
3. Apart from wind tunnel test, the integrated sensor chip is further tested for
different flapping frequency by mounting on to the wing. Force gauge data of the
FWMAV were also collected and processed in the MATLAB software. The
sensitivity of the lift sensing is 90 µV/V/gf and the sensitivity of net thrust sensing
is 159 µV/V/gf. With the supply voltage as V0=1.8V and reference voltage as 1V,
the total current consumption of the IA is 1.335mA.
4. The IA design alone occupies the area of 250 × 350 m2 whereas the total area
occupied by the IA and sensor integration is 550 × 600 m2.

Finally, the advantages of the CMOS MEMS monolithic integration flow sensors
and amplifier circuit are justified and discussed with the testing results.

85
6.3 Summary about the flapping wing with flow sensor
experiment
Some of the important observations from the integrated test of flapping wing and
flow sensor are as follows:
1. Standard force gauge data of the flapping wing were collected at the same time
with the CMOS MEMS flow sensor output.
2. The sensitivity of the output voltage with respect to the lift is measured as 90
µV/V/gf in Figure 5.19(a), and to the net thrust is measured as 159 µV/V/gf in
Figure 5.19(b).
3. The test results in Figure 5.15 (lift data) and in Figure 5.18 (net thrust data) show
that the CMOS MEMS flow sensor signals on the flapping wing is more similar
to the net thrust signal, rather than the lift signal, from the standard force gauge.
4. In conclusion, the CMOS MEMS flow sensor herein not only sensitive enough
to detect the aerodynamic forces of a flapping wing, it also behaves with a
characteristic of shear stress sensor.

6.4 Future Work

1. Application on wind turbine blade
Wind turbine blades flutter in extreme weather and fatigue loading conditions
because of random wind speeds. Usually, the wind turbines are placed in remote areas
with very high altitude that makes difficult to be reached and do maintenance process.
Time to time the structural health monitoring has to be done for wind turbine for its
better performance. Factors mainly affect the wind turbines are heavy rainfall,
temperature, humidity, erosion etc. the damaged blade directly affects the efficiency of
power generation. The wind turbine has complexity in maintenance and expensive too.
The manufacturing of wind turbine blade is a challenge. It needs much care and patience.
In addition, it needs a big capital of investment for fabrication [223-227].
In Figure 6.1 (a), we can see the view of a wind farm in Taiwan, in Figure 6.1 (b)
we can see a partially broken wind turbine blade, if proper initiative is not taken at the
right time, more damage will occur as shown in Figure 6.1 (c) due to lack of surveillance
which will lead to extra expenses than the expected. More details of wind turbine is
explained in Appendix D.

86
 (a) (b)

(c)
Figure 6.1 Wind farm with partially damaged turbine blades

2. Flexible and light PCB board design for application on smaller MAV
Flexible printed circuit boards were designed using EasyEDA, which is a web-
based EDA tool suite that enables hardware engineers to design and discuss schematics,
simulations and printed circuit boards. Due to ESD problem during the wire bonding, it
could not work. In future, after solving the ESD problem, we can test our sensor in
smaller bird with extremely light packaging. Hence, performance can then be measured
accordingly. Figure 6.2 (a) shows the Flexible PCB board design dimensions and Figure
6.2 (b) shows the sensor chips mounted on the flexible PCB board after the wire bonding.

87
 (a)

(b)
Figure 6.2 Flexible PCB board design sensor chips are mounted on it

3. Combination pressure sensor and flow sensor on flapping wing can be tested to
measure the lift and net thrust force simultaneously

The sensitivity of the lift sensing is 90 µV/V/gf and the sensitivity of net thrust
sensing is 159 µV/V/gf. This sensor can be better assigned as net thrust force sensing
device. Flow sensor is like a shear stress sensor, good for detecting thrust or drag. In
future, combination of pressure sensor and flow sensor on flapping wing can be tested
to measure the lift and net thrust force simultaneously.

88
REFERENCES

[1] M. Gad-el-Hak, "The fluid mechanics of microdevices—the Freeman scholar

lecture," J. Fluids Eng., vol. 121, no.1, pp.5-33, 1999.
[2] F. Jiang, Y.-C. Tai, C.-M. Ho, R. Karan, and M. Garstenauer, "Theoretical and
experimental studies of micromachined hot-wire anemometers," in Proceedings
of 1994 IEEE International Electron Devices Meeting, pp. 139-142.
[3] J. Liu, Y.-C. Tai, and C.-M. H.-C. Pong, "MEMS for pressure distribution studies
of gaseous flows in microchannels," in Proceedings IEEE Micro Electro
Mechanical Systems. 1995, 1995: IEEE, p. 209.
[4] N. Pornsin-Sirirak, M. Liger, Y.-C. Tai, S. Ho, and C.-M. Ho, "Flexible parylene-
valved skin for adaptive flow control," in Technical Digest. MEMS 2002,
Fifteenth IEEE International Conference on Micro Electro Mechanical Systems,
2002, pp. 101-104.
[5] F. Jiang, G.-B. Lee, Y.-C. Tai, and C.-M. Ho, "A flexible micromachine-based
shear-stress sensor array and its application to separation-point detection,"
Sensors and Actuators A: Physical, vol. 79, no. 3, pp. 194-203, 2000.
[6] G.-B. Lee et al., "Robust vortex control of a delta wing by distributed
microelectromechanical-systems actuators," Journal of Aircraft, vol. 37, no. 4,
pp. 697-706, 2000.
[7] G.-B. Lee, A. Huang, C.-M. Ho, F. Jiang, C. Grosjean, and Y.-C. Tai, "Sensing
and control of aerodynamic separation by MEMS," Journal of Mechanics, vol.
16, no. 1, pp. 45-52, 2000.
[8] L.-J. Yang, C.-K. Hsu, J.-Y. Ho, and C.-K. Feng, "Flapping wings with PVDF
sensors to modify the aerodynamic forces of a micro aerial vehicle," Sensors and
Actuators A: Physical, vol. 139, no. 1-2, pp. 95-103, 2007.
[9] Y. Javed, M. Mansoor, and I. A. Shah, "A review of principles of MEMS
pressure sensing with its aerospace applications," Sensor Review, 2019.
[10] L.-J. Yang and B. Esakki, Flapping Wing Vehicles: Numerical and Experimental
Approach. CRC Press, 2021.
[11] C.-M. Ho and Y.-C. Tai, "MEMS and its applications for flow control," 1996.
[12] N.-T. Nguyen, X. Huang, and T. K. Chuan, "MEMS-micropumps: a review," J.
Fluids Eng., vol. 124, no. 2, pp. 384-392, 2002.
[13] M. Tanaka, "An industrial and applied review of new MEMS devices features,"
Microelectronic Engineering, vol. 84, no. 5-8, pp. 1341-1344, 2007.
[14] S.-H. Liao, W.-J. Chen, and M. S.-C. Lu, "A CMOS MEMS capacitive flow
sensor for respiratory monitoring," IEEE Sensors Journal, vol. 13, no. 5, pp.
1401-1402, 2013.
[15] H.-Y. Shih, C.-T. Hsin, C.-W. Yang, H.-L. Chen, and J.-Y. Kuo, "A CMOS
MEMS Pressure Sensor for Blood Pulse and Pressure Measurement
Applications," in 2021 International Symposium on Intelligent Signal Processing
and Communication Systems (ISPACS), 2021: IEEE, pp. 1-2.
[16] B. T. Chia et al., "Temperature sensor array using flexible substrate," in 2007
IEEE 20th International Conference on Micro Electro Mechanical Systems
(MEMS), 2007: IEEE, pp. 589-592.
[17] C.-Y. Lee, S.-J. Lee, Y.-C. Hu, W.-P. Shih, W.-Y. Fan, and C.-W. Chuang, "Real
time monitoring of temperature of a micro proton exchange membrane fuel cell,"
Sensors, vol. 9, no. 3, pp. 1423-1432, 2009.

89
[18] C.-Y. Lee, S.-J. Lee, W.-P. Shih, C.-H. Lin, C.-C. Chang, and P.-Z. Chang,
"Fabrication of micro temperature sensor and heater in a stainless steel-based
micro reformer," in 2009 4th IEEE International Conference on Nano/Micro
Engineered and Molecular Systems, 2009, pp. 172-175.
[19] W.-P. Shih et al., "Flexible temperature sensor array based on a graphite-
polydimethylsiloxane composite," Sensors, vol. 10, no. 4, pp. 3597-3610, 2010.
[20] A. Y.-T. Wang et al., "Thermal manipulation utilizing micro-cantilever probe in
scanning electron microscopy," in Progress in Optomechatronic Technologies:
Springer, 2014, pp. 169-180.
[21] K.-H. Cheng, L.-J. Tzou, W.-B. Yang, and S.-S. Sheu, "A CMOS low power
voltage controlled oscillator with split-path controller," in ICECS 2001. 8th IEEE
International Conference on Electronics, Circuits and Systems (Cat. No.
01EX483), 2001, vol. 1, pp. 421-424.
[22] W.-B. Yang, Y.-Y. Lin, and Y.-L. Lo, "Analysis and design considerations of
static CMOS logics under process, voltage and temperature variation in 90nm
CMOS process," in 2014 International Conference on Information Science,
Electronics and Electrical Engineering, 2014, vol. 3: IEEE, pp. 1653-1656.
[23] W.-B. Yang, Y.-Y. Lin, C.-H. Wang, K.-N. Chang, C.-H. Chen, and Y.-L. Lo,
"Analysis and design considerations of static CMOS logics under process,
voltage and temperature variation in UMC 0.18 µ m CMOS process," in 2015
International Symposium on Intelligent Signal Processing and Communication
Systems (ISPACS), 2015: IEEE, pp. 57-61.
[24] F. Ejeian et al., "Design and applications of MEMS flow sensors: A review,"
Sensors and Actuators A: Physical, vol. 295, pp. 483-502, 2019.
[25] C. Loraux and E. Brühwiler, "The use of long term monitoring data for the
extension of the service duration of existing wind turbine support structures," in
Journal of Physics: Conference Series, 2016, vol. 753, no. 7: IOP Publishing, p.
072023.
[26] M. Benedetti, V. Fontanari, and D. Zonta, "Structural health monitoring of wind
towers: remote damage detection using strain sensors," Smart Materials and
Structures, vol. 20, no. 5, p. 055009, 2011.
[27] W. Yang, P. J. Tavner, C. J. Crabtree, Y. Feng, and Y. Qiu, "Wind turbine
condition monitoring: technical and commercial challenges," Wind Energy, vol.
17, no. 5, pp. 673-693, 2014.
[28] A. Stoica et al., "WindBots," 2016.
[29] T. Leu, J. Yu, J. Miau, and S. Chen, "MEMS flexible thermal flow sensors for
measurement of unsteady flow above a pitching wind turbine blade,"
Experimental Thermal and Fluid Science, vol. 77, pp. 167-178, 2016.
[30] M. M. Pedersen, T. J. Larsen, H. A. Madsen, and G. C. Larsen, "Using wind
speed from a blade-mounted flow sensor for power and load assessment on
modern wind turbines," Wind Energy Science, vol. 2, no. 2, pp. 547-567, 2017.
[31] Y.-H. Wang et al., "MEMS-based gas flow sensors," Microfluidics and
Nanofluidics, vol. 6, no. 3, pp. 333-346, 2009.
[32] J. T. Kuo, L. Yu, and E. Meng, "Micromachined thermal flow sensors—A
review," Micromachines, vol. 3, no. 3, pp. 550-573, 2012.
[33] S. Dalola et al., "MEMS thermal flow sensor with smart electronic interface
circuit," IEEE Sensors Journal, vol. 12, no. 12, pp. 3318-3328, 2012.
[34] N. Djuzhev et al., "Measurement system for wide-range flow evaluation and
thermal characterization of MEMS-based thermoresistive flow-rate sensors,"
Sensors and Actuators A: Physical, vol. 330, p. 112832, 2021.

90
[35] L. Hinkle and C. F. Mariano, "Toward understanding the fundamental
mechanisms and properties of the thermal mass flow controller," Journal of
Vacuum Science & Technology A: Vacuum, Surfaces, and Films, vol. 9, no. 3,
pp. 2043-2047, 1991.
[36] C.-H. Wu, D. Kang, P.-H. Chen, and Y.-C. Tai, "MEMS thermal flow sensors,"
Sensors and Actuators A: Physical, vol. 241, pp. 135-144, 2016.
[37] W. Xu, X. Wang, X. Zhao, and Y.-K. Lee, "Two-dimensional CMOS MEMS
thermal flow sensor with high sensitivity and improved accuracy," Journal of
Microelectromechanical Systems, vol. 29, no. 2, pp. 248-254, 2020.
[38] Y. Xu, C.-W. Chiu, F. Jiang, Q. Lin, and Y.-C. Tai, "A MEMS multi-sensor chip
for gas flow sensing," Sensors and Actuators A: Physical, vol. 121, no. 1, pp.
253-261, 2005.
[39] Y. Xu, F. Jiang, S. Newbern, A. Huang, C.-M. Ho, and Y.-C. Tai, "Flexible
shear-stress sensor skin and its application to unmanned aerial vehicles," Sensors
and Actuators A: Physical, vol. 105, no. 3, pp. 321-329, 2003.
[40] W. Xu, K. Song, S. Ma, B. Gao, Y. Chiu, and Y.-K. Lee, "Theoretical and
experimental investigations of thermoresistive micro calorimetric flow sensors
fabricated by CMOS MEMS technology," Journal of Microelectromechanical
Systems, vol. 25, no. 5, pp. 954-962, 2016.
[41] W. Xu, L.-J. Yang, and Y.-K. Lee, "CMOS compatible MEMS air velocity
sensor with improved sensitivity and linearity for human thermal comfort sensing
applications," IEEE Sensors Journal, vol. 21, no. 21, pp. 23872-23879, 2021.
[42] M. Ahmed, W. Xu, S. Mohamad, M. Duan, Y.-K. Lee, and A. Bermak,
"Integrated CMOS-MEMS flow sensor with high sensitivity and large flow
range," IEEE Sensors Journal, vol. 17, no. 8, pp. 2318-2319, 2017.
[43] B. Abbasnejad, W. Thorby, A. Razmjou, D. Jin, M. Asadnia, and M. E. Warkiani,
"MEMS piezoresistive flow sensors for sleep apnea therapy," Sensors and
Actuators A: Physical, vol. 279, pp. 577-585, 2018.
[44] N. Nguyen, "Micromachined flow sensors—a review," Flow measurement and
Instrumentation, vol. 8, no. 1, pp. 7-16, 1997.
[45] L. Zhang et al., "Measurements and modeling of two-phase flow in
microchannels with nearly constant heat flux boundary conditions," Journal of
Microelectromechanical Systems, vol. 11, no. 1, pp. 12-19, 2002.
[46] L. Jiang, M. Wong, and Y. Zohar, "Phase change in microchannel heat sink under
forced convection boiling," in Proceedings IEEE Thirteenth Annual
International Conference on Micro Electro Mechanical Systems (Cat. No.
00CH36308), 2000: IEEE, pp. 397-402.
[47] Z. Tan, M. Shikida, M. Hirota, K. Sato, T. Iwasaki, and Y. Iriye, "Experimental
and theoretical study of an on-wall in-tube flexible thermal sensor," Journal of
Micromechanics and Microengineering, vol. 17, no. 4, p. 679, 2007.
[48] J. Xie, J. Shih, and Y.-C. Tai, "Integrated surface-micromachined mass flow
controller," in The Sixteenth IEEE International Conference on Micro Electro
Mechanical Systems, 2003. MEMS-03 Kyoto, pp. 20-23.
[49] J. Alsalaet et al., "Laminar flowmeter for mechanical ventilator: manufacturing
challenge of covid-19 pandemic," Flow Measurement and Instrumentation, vol.
82, p. 102058, 2021.
[50] Z. Fang, A. I. Li, H. Wang, R. Zhang, X. Mai, and T. Pan, "Covid19 AmbuBox:
a fast-deployable low-cost ventilator for COVID-19 emergent care," SLAS
TECHNOLOGY: Translating Life Sciences Innovation, vol. 25, no. 6, pp. 573-
584, 2020.

91
[51] M. Ahmed, W. Xu, S. Mohamad, F. Boussaid, Y.-K. Lee, and A. Bermak, "Fully
integrated bidirectional CMOS-MEMS flow sensor with low power pulse
operation," IEEE Sensors Journal, vol. 19, no. 9, pp. 3415-3424, 2019.
[52] W. Xu et al., "A wafer-level encapsulated CMOS MEMS thermoresistive
calorimetric flow sensor with integrated packaging design," in 2017 IEEE 30th
International Conference on Micro Electro Mechanical Systems (MEMS), 2017:
IEEE, pp. 989-992.
[53] W. Xu, X. Wang, R. Wang, J. Xu, and Y.-K. Lee, "CMOS MEMS thermal flow
sensor with enhanced sensitivity for heating, ventilation, and air conditioning
application," IEEE Transactions on Industrial Electronics, vol. 68, no. 5, pp.
4468-4476, 2020.
[54] J.-B. Sun, M. Qin, and Q.-A. Huang, "Flip-chip packaging for a two-dimensional
thermal flow sensor using a copper pillar bump technology," IEEE sensors
journal, vol. 7, no. 7, pp. 990-995, 2007.
[55] R. Waikhom, L.-J. Yang, H.-Y. Shih, and C.-R. Kuo, "Self-heating CMOS flow
sensor," in 2021 21st International Conference on Solid-State Sensors, Actuators
and Microsystems (Transducers), 2021: IEEE, pp. 1279-1282.
[56] R.-H. Ma, P.-C. Chou, Y.-H. Wang, T.-H. Hsueh, L.-M. Fu, and C.-Y. Lee, "A
microcantilever-based gas flow sensor for flow rate and direction detection,"
Microsystem Technologies, vol. 15, no. 8, pp. 1201-1205, 2009.
[57] L.-J. Yang, B. Esakki, and R. Waikhom, "Inertial effect on the time-averaged lift
of flapping wings," Journal of Applied Science and Engineering, vol. 23, no. 2,
pp. 357-359, 2020.
[58] M. Sun and J. Tang, "Unsteady aerodynamic force generation by a model fruit
fly wing in flapping motion," Journal of Experimental Biology, vol. 205, no. 1,
pp. 55-70, 2002.
[59] S. Sunada and C. Ellington, "Approximate added-mass method for estimating
induced power for flapping flight," AIAA Journal, vol. 38, no. 8, pp. 1313-1321,
2000.
[60] M. H. Dickinson, F.-O. Lehmann, and S. P. Sane, "Wing rotation and the
aerodynamic basis of insect flight," Science, vol. 284, no. 5422, pp. 1954-1960,
1999.
[61] L.-J. Yang, "The micro-air-vehicle golden snitch and its figure-of-8 flapping,"
Journal of Applied Science and Engineering, vol. 15, no. 3, pp. 197-212, 2012.
[62] L.-J. Yang, B. Esakki, U. Chandrasekhar, K.-C. Hung, and C.-M. Cheng,
"Practical flapping mechanisms for 20 cm-span micro air vehicles," International
Journal of Micro Air Vehicles, vol. 7, no. 2, pp. 181-202, 2015.
[63] L.-J. Yang, A.-F. Kao, and C.-K. Hsu, "Wing stiffness on light flapping micro
aerial vehicles," Journal of Aircraft, vol. 49, no. 2, pp. 423-431, 2012.
[64] L.-J. Yang, N. Kapri, R. Waikhom, and N. K. Unnam, "Fabrication, aerodynamic
measurement and performance evaluation of corrugated flapping wings," J.
Aeronaut. Astronaut. Aviat, vol. 53, pp. 83-94, 2021.
[65] D. J. Pines and F. Bohorquez, "Challenges facing future micro-air-vehicle
development," Journal of Aircraft, vol. 43, no. 2, pp. 290-305, 2006.
[66] M. Karásek, F. T. Muijres, C. De Wagter, B. D. Remes, and G. C. De Croon, "A
tailless aerial robotic flapper reveals that flies use torque coupling in rapid
banked turns," Science, vol. 361, no. 6407, pp. 1089-1094, 2018.
[67] M. Keennon, K. Klingebiel, H. Won, and A. Andriukov, "Tailless flapping wing
propulsion and control development for the nano hummingbird micro air

92
 vehicle," in American Helicopter Society Future Vertical Lift Aircraft Design
Conference, 2012, pp. 1-24.
[68] H. V. Phan and H. C. Park, "Mechanisms of collision recovery in flying beetles
and flapping-wing robots," Science, vol. 370, no. 6521, pp. 1214-1219, 2020.
[69] F.-Y. Hsiao, L.-J. Yang, S.-H. Lin, C.-L. Chen, and J.-F. Shen, "Autopilots for
ultra lightweight robotic birds: automatic altitude control and system integration
of a sub-10 g weight flapping-wing micro air vehicle," IEEE Control Systems
Magazine, vol. 32, no. 5, pp. 35-48, 2012.
[70] A. Muhammad, Q. V. Nguyen, H. C. Park, D. Y. Hwang, D. Byun, and N. S.
Goo, "Improvement of artificial foldable wing models by mimicking the
unfolding/folding mechanism of a beetle hind wing," Journal of Bionic
Engineering, vol. 7, no. 2, pp. 134-141, 2010.
[71] L.-J. Yang, C.-K. Hsu, H.-C. Han, and J.-M. Miao, "Light flapping micro aerial
vehicle using electrical-discharge wire-cutting technique," Journal of Aircraft,
vol. 46, no. 6, pp. 1866-1874, 2009.
[72] L.-J. Yang and S. Marimuthu, "Acoustic comparison of PET and latex wings for
flapping micro-air-vehicles," in 10th IEEE International Conference on
Nano/Micro Engineered and Molecular Systems, 2015: IEEE, pp. 172-174.
[73] L.-J. Yang et al., "Check-valve design in enhancing aerodynamic performance
of flapping wings," Applied Sciences, vol. 11, no. 8, p. 3416, 2021.
[74] W. He, T. Meng, X. He, and C. Sun, "Iterative learning control for a flapping
wing micro aerial vehicle under distributed disturbances," IEEE Transactions on
Cybernetics, vol. 49, no. 4, pp. 1524-1535, 2018.
[75] M. Bronz, J. M. Moschetta, P. Brisset, and M. Gorraz, "Towards a long
endurance MAV," International Journal of Micro Air Vehicles, vol. 1, no. 4, pp.
241-254, 2009.
[76] Q. V. Nguyen, Q. T. Truong, H. C. Park, N. S. Goo, and D. Byun, "Measurement
of force produced by an insect-mimicking flapping-wing system," Journal of
Bionic Engineering, vol. 7, pp. S94-S102, 2010.
[77] M. Ashraf, J. Young, and J. Lai, "Reynolds number, thickness and camber effects
on flapping airfoil propulsion," Journal of Fluids and structures, vol. 27, no. 2,
pp. 145-160, 2011.
[78] N. Ismail, A. Zulkifli, M. Abdullah, M. H. Basri, and N. S. Abdullah,
"Optimization of aerodynamic efficiency for twist morphing MAV wing,"
Chinese Journal of Aeronautics, vol. 27, no. 3, pp. 475-487, 2014.
[79] J. Shang, S. A. Combes, B. Finio, and R. J. Wood, "Artificial insect wings of
diverse morphology for flapping-wing micro air vehicles," Bioinspiration &
Biomimetics, vol. 4, no. 3, p. 036002, 2009.
[80] Q. Truong, Q. Nguyen, V. Truong, H. Park, D. Byun, and N. Goo, "A modified
blade element theory for estimation of forces generated by a beetle-mimicking
flapping wing system," Bioinspiration & Biomimetics, vol. 6, no. 3, p. 036008,
2011.
[81] G. C. De Croon, M. Groen, C. De Wagter, B. Remes, R. Ruijsink, and B. W. van
Oudheusden, "Design, aerodynamics and autonomy of the DelFly,"
Bioinspiration & Biomimetics, vol. 7, no. 2, p. 025003, 2012.
[82] T. Liu, B. T. Campbell, S. P. Burns, and J. P. Sullivan, "Temperature-and
pressure-sensitive luminescent paints in aerodynamics," Applied Mechanics
Reviews, vol. 50, pp. 227-246, 1997.

93
[83] R. H. Engler, C. Klein, and O. Trinks, "Pressure sensitive paint systems for
pressure distribution measurements in wind tunnels and turbomachines,"
Measurement Science and Technology, vol. 11, no. 7, p. 1077, 2000.
[84] Y. Egami et al., "Quantitative visualization of the leading-edge vortices on a
delta wing by using pressure-sensitive paint," Journal of Visualization, vol. 4,
no. 2, pp. 139-150, 2001.
[85] M. Ramasamy and J. G. Leishman, "Phase-locked particle image velocimetry
measurements of a flapping wing," Journal of Aircraft, vol. 43, no. 6, pp. 1867-
1875, 2006.
[86] T. Flint, M. Jermy, T. New, and W. Ho, "Computational study of a pitching bio-
inspired corrugated airfoil," International Journal of Heat and Fluid Flow, vol.
65, pp. 328-341, 2017.
[87] Y. Couder, J. Chomaz, and M. Rabaud, "On the hydrodynamics of soap films,"
Physica D: Nonlinear Phenomena, vol. 37, no. 1-3, pp. 384-405, 1989.
[88] M. Rutgers, X. Wu, and W. Daniel, "Conducting fluid dynamics experiments
with vertically falling soap films," Review of Scientific Instruments, vol. 72, no.
7, pp. 3025-3037, 2001.
[89] M. Rutgers, X. l. Wu, R. Bhagavatula, A. Petersen, and W. Goldburg, "Two‐
dimensional velocity profiles and laminar boundary layers in flowing soap
films," Physics of Fluids, vol. 8, no. 11, pp. 2847-2854, 1996.
[90] M. Gharib and M. Beizaie, "Visualization of two-dimensional flows by a liquid
(soap) film tunnel," Journal of Visualization, vol. 2, no. 2, pp. 119-126, 1999.
[91] J. Zhang, S. Childress, A. Libchaber, and M. Shelley, "Flexible filaments in a
flowing soap film as a model for one-dimensional flags in a two-dimensional
wind," Nature, vol. 408, no. 6814, pp. 835-839, 2000.
[92] M. R. Haque, M. Fayed, A.-L. Gunter, O. Smadi, L. Kadem, and H. D. Ng,
"Numerical simulation and flow visualization using soap film of the self-
organized vortex structure in the wake of an array of cylinders," Journal of
Visualization, vol. 14, no. 4, pp. 311-314, 2011.
[93] L. Hopkins, J. Kelly, A. Wexler, and A. Prasad, "Particle image velocimetry
measurements in complex geometries," Experiments in Fluids, vol. 29, no. 1, pp.
91-95, 2000.
[94] F. Söderberg and B. Johansson, "Particle image velocimetry flow measurements
and heat-release analysis in a cross-flow cylinder head," SAE Technical Paper,
0148-7191, 2002.
[95] L.-J. Yang, C. Tasupalli, R. Waikhom, and N. Panchal, "Soap Film Visualization
of a 10 cm-Span Flapping Wing," Fluids, vol. 6, no. 10, p. 361, 2021.
[96] V. Balakrishnan, T. Dinh, H.-P. Phan, D. V. Dao, and N.-T. Nguyen, "Highly
sensitive 3C-SiC on glass based thermal flow sensor realized using MEMS
technology," Sensors and Actuators A: Physical, vol. 279, pp. 293-305, 2018.
[97] S.-T. Hung, S.-C. Wong, and W. Fang, "The development and application of
microthermal sensors with a mesh-membrane supporting structure," Sensors and
Actuators A: Physical, vol. 84, no. 1-2, pp. 70-75, 2000.
[98] M. Sadeghi, R. Peterson, and K. Najafi, "Air flow sensing using micro-wire-
bonded hair-like hot-wire anemometry," Journal of Micromechanics and
Microengineering, vol. 23, no. 8, p. 085017, 2013.
[99] E. Yoon and K. D. Wise, "An integrated mass flow sensor with on-chip CMOS
interface circuitry," IEEE Transactions on Electron Devices, vol. 39, no. 6, pp.
1376-1386, 1992.

94
[100] G. Kaltsas and A. Nassiopoulou, "Novel CMOS compatible monolithic silicon
gas flow sensor with porous silicon thermal isolation," Sensors and Actuators A:
Physical, vol. 76, no. 1-3, pp. 133-138, 1999.
[101] L. Qiu, S. Hein, E. Obermeier, and A. Schubert, "Micro gas-flow sensor with
integrated heat sink and flow guide," Sensors and Actuators A: Physical, vol. 54,
no. 1-3, pp. 547-551, 1996.
[102] S. Wu, Q. Lin, Y. Yuen, and Y.-C. Tai, "MEMS flow sensors for nano-fluidic
applications," Sensors and Actuators A: Physical, vol. 89, no. 1-2, pp. 152-158,
2001.
[103] H. Sturm and W. Lang, "Membrane-based thermal flow sensors on flexible
substrates," Sensors and Actuators A: Physical, vol. 195, pp. 113-122, 2013.
[104] N. Svedin, E. Stemme, and G. Stemme, "A new bidirectional gas-flow sensor
based on lift force," in Proceedings of International Solid State Sensors and
Actuators Conference (Transducers' 97), 1997, vol. 1: IEEE, pp. 145-148.
[105] N. Svedin, E. Kalvesten, E. Stemme, and G. Stemme, "A new silicon gas-flow
sensor based on lift force," Journal of Microelectromechanical Systems, vol. 7,
no. 3, pp. 303-308, 1998.
[106] A. Kuoni, R. Holzherr, M. Boillat, and N. F. de Rooij, "Polyimide membrane
with ZnO piezoelectric thin film pressure transducers as a differential pressure
liquid flow sensor," Journal of Micromechanics and Microengineering, vol. 13,
no. 4, p. S103, 2003.
[107] M. Asadnia et al., "From biological cilia to artificial flow sensors: Biomimetic
soft polymer nanosensors with high sensing performance," Scientific Reports,
vol. 6, p. 32955, 2016.
[108] R. P. Feynman, "There’s plenty of room at the bottom," Engineering and Science,
vol. 23, no. 5, 1959.
[109] W. Shockley, "The theory of p‐n junctions in semiconductors and p‐n junction
transistors," Bell System Technical Journal, vol. 28, no. 3, pp. 435-489, 1949.
[110] J. Bardeen and W. H. Brattain, "Physical principles involved in transistor action,"
Physical Review, vol. 75, no. 8, p. 1208, 1949.
[111] C. S. Smith, "Piezoresistance effect in germanium and silicon," Physical Review,
vol. 94, no. 1, p. 42, 1954.
[112] R. Ghodssi and P. Lin, MEMS materials and processes handbook. Springer
Science & Business Media, 2011.
[113] K. E. Petersen, "Silicon as a mechanical material," Proceedings of the IEEE, vol.
70, no. 5, pp. 420-457, 1982.
[114] H. C. Nathanson, W. E. Newell, R. A. Wickstrom, and J. R. Davis, "The resonant
gate transistor," IEEE Transactions on Electron Devices, vol. 14, no. 3, pp. 117-
133, 1967.
[115] F. Faggin, M. E. Hoff, S. Mazor, and M. Shima, "The history of the 4004," IEEE
Micro, vol. 16, no. 6, pp. 10-20, 1996.
[116] F. L. J. Sangster and K. Teer, "Bucket-brigade electronics: new possibilities for
delay, time-axis conversion, and scanning," IEEE Journal of Solid-State Circuits,
vol. 4, no. 3, pp. 131-136, 1969.
[117] W. S. Boyle and G. E. Smith, "Charge coupled semiconductor devices," Bell
System Technical Journal, vol. 49, no. 4, pp. 587-593, 1970.
[118] S. Samaun, K. Wise, E. Nielsen, and J. Angell, "An IC piezoresistive pressure
sensor for biomedical instrumentation," in 1971 IEEE International Solid-State
Circuits Conference. Digest of Technical Papers, 1971, vol. 14: IEEE, pp. 104-
105.

95
[119] S. C. Terry, J. H. Jerman, and J. B. Angell, "A gas chromatographic air analyzer
fabricated on a silicon wafer," IEEE Transactions on Electron Devices, vol. 26,
no. 12, pp. 1880-1886, 1979.
[120] K. D. Wise, J. B. Angell, and A. Starr, "An integrated-circuit approach to
extracellular microelectrodes," IEEE Transactions on Biomedical Engineering,
no. 3, pp. 238-247, 1970.
[121] J. Borky and K. D. Wise, "Integrated signal conditioning for silicon pressure
sensors," IEEE Transactions on Electron Devices, vol. 26, no. 12, pp. 1906-1910,
1979.
[122] T. Matsuo and K. D. Wise, "An integrated field-effect electrode for biopotential
recording," IEEE Transactions on Biomedical Engineering, no. 6, pp. 485-487,
1974.
[123] K. Najafi, K. Wise, and T. Mochizuki, "A high-yield IC-compatible multichannel
recording array," IEEE Transactions on Electron Devices, vol. 32, no. 7, pp.
1206-1211, 1985.
[124] E. Yoon and K. D. Wise, "A multi-element monolithic mass flowmeter with on-
chip CMOS readout electronics," in IEEE 4th Technical Digest on Solid-State
Sensor and Actuator Workshop, 1990: IEEE, pp. 161-164.
[125] W. H. Ko, M.-H. Bao, and Y.-D. Hong, "A high-sensitivity integrated-circuit
capacitive pressure transducer," IEEE transactions on electron devices, vol. 29,
no. 1, pp. 48-56, 1982.
[126] W. H. Ko, J. Hynecek, and S. F. Boettcher, "Development of a miniature pressure
transducer for biomedical applications," IEEE Transactions on Electron Devices,
vol. 26, no. 12, pp. 1896-1905, 1979.
[127] C.-C. Wen, T. Chen, and J. Zemel, "Gate-controlled diodes for ionic
concentration measurement," IEEE Transactions on Electron Devices, vol. 26,
no. 12, pp. 1945-1951, 1979.
[128] K. Lundström, M. Shivaraman, and C. Svensson, "A hydrogen‐sensitive Pd‐gate
MOS transistor," Journal of Applied Physics, vol. 46, no. 9, pp. 3876-3881, 1975.
[129] A. C. Fischer et al., "Integrating MEMS and ICs," Microsystems &
Nanoengineering, vol. 1, no. 1, pp. 1-16, 2015.
[130] M. G. Allen, S. D.Senturia, "Microfabricated structures for the measurement of
mechanical properties and adhesion of thin films," In L. H. Lee (eds) Adhesives,
Sealants, and Coatings for Space and Harsh Environments, Polymer Science and
Technology, vol. 37, Springer, Boston, MA., 1998.
[131] A. Van Putten and S. Middelhoek, "Integrated silicon anemometer," Electronics
Letters, vol. 10, no. 21, pp. 425-426, 1974.
[132] P. Bergveld, "Development of an ion-sensitive solid-state device for
neurophysiological measurements," IEEE Transactions on Biomedical
Engineering, no. 1, pp. 70-71, 1970.
[133] S. Reyntjens and R. Puers, "A review of focused ion beam applications in
microsystem technology," Journal of Micromechanics and Microengineering,
vol. 11, no. 4, p. 287, 2001.
[134] H. Baltes, W. Göpel, and J. Hesse, Sensors Update, vol 8." Sensor Review, vol.
22, no. 3, pp. 265-265.
[135] W. Göpel et al., "Surface defects of TiO2 (110): a combined XPS, XAES and
ELS study," Surface Science, vol. 139, no. 2-3, pp. 333-346, 1984.
[136] W. Göpel, G. Rocker, and R. Feierabend, "Intrinsic defects of TiO 2 (110):
Interaction with chemisorbed O2, H2, CO, and CO2," Physical Review B, vol. 28,
no. 6, p. 3427, 1983.

96
[137] G. Krötz et al., "Heteroepitaxial growth of 3C-SiC on SOI for sensor
applications," Materials Science and Engineering: B, vol. 61, pp. 516-521, 1999.
[138] E. Obermeier and P. Kopystynski, "Polysilicon as a material for microsensor
applications," Sensors and Actuators A: Physical, vol. 30, no. 1-2, pp. 149-155,
1992.
[139] L.-S. Fan, Y.-C. Tai, and R. S. Muller, "IC-processed electrostatic micromotors,"
Sensors and Actuators, vol. 20, no. 1-2, pp. 41-47, 1989.
[140] R. Muller and J. Conragan, "Transducer action in a metal‐insulator‐piezoelectric‐
semiconductor triode," Applied Physics Letters, vol. 6, no. 5, pp. 83-85, 1965.
[141] K. A. Shaw, Z. L. Zhang, and N. C. MacDonald, "SCREAM I: a single mask,
single-crystal silicon, reactive ion etching process for microelectromechanical
structures," Sensors and Actuators A: Physical, vol. 40, no. 1, pp. 63-70, 1994.
[142] C.-L. Dai, S.-S. Lu, and P.-Z. Chang, "Design and processing of integrated micro
accel-erometers using standard CMOS process," Journal of the Chinese Institute
of Engineers, vol. 20, no. 1, pp. 47-55, 1997.
[143] H. Qu, "CMOS MEMS fabrication technologies and devices," Micromachines,
vol. 7, no. 1, p. 14, 2016.
[144] C.-L. Dai and P.-Z. Chang, "In-situ micro strain gauges for measuring residual
strain of three CMOS thin films using only one maskless post-processing step,"
Journal of the Chinese Institute of Engineers, vol. 20, no. 5, pp. 539-548, 1997.
[145] Z.-Y. Tsai, P.-J. Shih, Y.-C. Tsai, and C.-L. Dai, "Manufacturing and testing of
radio frequency MEMS switches using the complementary metal oxide
semiconductor process," Sensors, vol. 21, no. 4, p. 1396, 2021.
[146] H. Baltes, O. Brand, A. Hierlemann, D. Lange, and C. Hagleitner, "CMOS
MEMS-present and future," in Technical Digest. MEMS 2002 IEEE
International Conference. Fifteenth IEEE International Conference on Micro
Electro Mechanical Systems, 2002: IEEE, pp. 459-466.
[147] H.-J. Wang, T.-Y. Cheng, C.-C. Huang, C.-Y. Su, C.-L. Dai, and Y.-C. Tsai,
"High sensitivity and flexible fabric strain sensor based on electrochemical
graphene," Japanese Journal of Applied Physics, vol. 60, no. SC, p. SCCD04,
2021.
[148] H. Baltes, "CMOS as sensor technology," Sensors and Actuators A: Physical,
vol. 37, pp. 51-56, 1993.
[149] H. Baltes and O. Brand, "CMOS-based microsensors and packaging," Sensors
and Actuators A: Physical, vol. 92, no. 1-3, pp. 1-9, 2001.
[150] W. P. Eaton and J. H. Smith, "Micromachined pressure sensors: review and
recent developments," Smart Materials and Structures, vol. 6, no. 5, p. 530,
1997.
[151] H. Dudaicevs, M. Kandler, Y. Manoli, W. Mokwa, and E. Spiegel, "Surface
micromachined pressure sensors with integrated CMOS read-out electronics,"
Sensors and Actuators A: Physical, vol. 43, no. 1-3, pp. 157-163, 1994.
[152] S. Middelhoek, "Celebration of the tenth transducers conference: the past,
present and future of transducer research and development," Sensors and
Actuators A: Physical, vol. 82, no. 1-3, pp. 2-23, 2000.
[153] C. Chang et al., "A wideband electrostatic microwave switch fabricated by
surface micromachining," Journal of the Chinese Institute of Engineers, vol. 23,
no. 6, pp. 781-787, 2000.
[154] W. Fang et al., "CMOS MEMS: A key technology towards the “More than
Moore” era," in 2013 Transducers & Eurosensors XXVII: The 17th International

97
 Conference on Solid-State Sensors, Actuators and Microsystems, 2013: IEEE,
pp. 2513-2518.
[155] M. A. Fraga, R. S. Pessoa, M. Massi, and H. S. Maciel, "Applications of SiC-
based thin films in electronic and MEMS devices," Physics and Technology of
Silicon Carbide Devices, vol. 1, pp. 313-336, 2012.
[156] M. Fraga and R. Pessoa, "Progresses in synthesis and application of sic films:
from CVD to ALD and from MEMS to NEMS," Micromachines, vol. 11, no. 9,
p. 799, 2020.
[157] Y. Zhai et al., "Design, fabrication and test of a bulk SiC MEMS accelerometer,"
Microelectronic Engineering, p. 111793, 2022.
[158] N. Wright and A. Horsfall, "SiC sensors: a review," Journal of Physics D:
Applied Physics, vol. 40, no. 20, p. 6345, 2007.
[159] J. Fan and P. K. Chu, Silicon carbide nanostructures: fabrication, structure, and
properties. Springer, 2014.
[160] S. Nakata et al., "Strain engineering of core–shell silicon carbide nanowires for
mechanical and piezoresistive characterizations," Nanotechnology, vol. 30, no.
26, p. 265702, 2019.
[161] A. Sciuto et al., "4H-SiC pn Junction-Based Near IR Photon Source," IEEE
Sensors Journal, vol. 21, no. 2, pp. 1504-1509, 2020.
[162] S. E. Saddow, Silicon carbide biotechnology: a biocompatible semiconductor for
advanced biomedical devices and applications. Elsevier, 2012.
[163] L.-J. Yang, C.-C. Lai, C.-L. Dai, and P.-Z. Chang, "A piezoresistive micro
pressure sensor fabricated by commercial DPDM CMOS process," Journal of
Applied Science and Engineering, vol. 8, no. 1, pp. 67-73, 2005.
[164] C.-L. Dai, "In situ electrostatic microactuators for measuring the Young's
modulus of CMOS thin films," Journal of Micromechanics and
Microengineering, vol. 13, no. 5, p. 563, 2003.
[165] H.-H. Wang, C.-W. Hsu, W.-H. Liao, L.-J. Yang, and C.-L. Dai, "Micro pressure
sensors of 50 μm size fabricated by a standard CMOS foundry and a novel post
process," in 19th IEEE International Conference on Micro Electro Mechanical
Systems, 2006: IEEE, pp. 578-581.
[166] L.-J. Yang, W.-Z. Lin, T.-J. Yao, and Y.-C. Tai, "Photo-patternable gelatin as
protection layers in low-temperature surface micromachinings," Sensors and
Actuators A: physical, vol. 103, no. 1-2, pp. 284-290, 2003.
[167] L.-J. Yang and T.-Y. Lin, "A PDMS-based thermo-pneumatic micropump with
Parylene inner walls," Microelectronic Engineering, vol. 88, no. 8, pp. 1894-
1897, 2011.
[168] W. Xu, X. Wang, Y. Chiu, and Y.-K. Lee, "High sensitivity and wide dynamic
range thermoresistive micro calorimetric flow sensor with CMOS MEMS
technology," IEEE Sensors Journal, vol. 20, no. 8, pp. 4104-4111, 2019.
[169] W. Xu, L. Pan, B. Gao, Y. Chiu, K. Xu, and Y.-K. Lee, "Systematic study of
packaging designs on the performance of CMOS thermoresistive micro
calorimetric flow sensors," Journal of Micromechanics and Microengineering,
vol. 27, no. 8, p. 085001, 2017.
[170] P. Michalik, D. Fernández, M. Wietstruck, M. Kaynak, and J. Madrenas,
"Experiments on MEMS integration in 0.25 μm CMOS Process," Sensors, vol.
18, no. 7, p. 2111, 2018.
[171] N. Maluf and K. Williams, An introduction to microelectromechanical systems
engineering. Artech House, 2004.

98
[172] Y. Song, S. Dhar, L. C. Feldman, G. Chung, and J. Williams, "Modified deal
grove model for the thermal oxidation of silicon carbide," Journal of Applied
Physics, vol. 95, no. 9, pp. 4953-4957, 2004.
[173] L. J. Hornbeck, "Deformable-mirror spatial light modulators," in Spatial Light
Modulators and Applications III, 1990, vol. 1150: International Society for
Optics and Photonics, pp. 86-103.
[174] P. Gennissen, M. Bartek, P. French, and P. Sarro, "Bipolar-compatible epitaxial
poly for smart sensors: stress minimization and applications," Sensors and
Actuators A: Physical, vol. 62, no. 1-3, pp. 636-645, 1997.
[175] S. Greek, F. Ericson, S. Johansson, M. Fürtsch, and A. Rump, "Mechanical
characterization of thick polysilicon films: Young's modulus and fracture
strength evaluated with microstructures," Journal of Micromechanics and
Microengineering, vol. 9, no. 3, p. 245, 1999.
[176] O. De Sagazan, M. Denoual, P. Guil, D. Gaudin, and O. Bonnaud, "Micro-
electro-mechanical systems fast fabrication by selective thick polysilicon growth
in epitaxial reactor," Microsystem Technologies, vol. 12, no. 10, pp. 953-958,
2006.
[177] D. J. Young, J. Du, C. A. Zorman, and W. H. Ko, "High-temperature single-
crystal 3C-SiC capacitive pressure sensor," IEEE Sensors Journal, vol. 4, no. 4,
pp. 464-470, 2004.
[178] N. Moronuki, M. Kojima, and A. Kakuta, "Single-crystal SiC thin-film produced
by epitaxial growth and its application to micro-mechanical devices," Thin Solid
Films, vol. 516, no. 16, pp. 5344-5348, 2008.
[179] T. Abe and M. L. Reed, "Low strain sputtered polysilicon for micromechanical
structures," in Proceedings of Ninth International Workshop on Micro
Electromechanical Systems, 1996: IEEE, pp. 258-262.
[180] K. A. Honer and G. T. Kovacs, "Integration of sputtered silicon microstructures
with pre-fabricated CMOS circuitry," Sensors and Actuators A: Physical, vol.
91, no. 3, pp. 386-397, 2001.
[181] N. Ledermann, J. Baborowski, P. Muralt, N. Xantopoulos, and J.-M. Tellenbach,
"Sputtered silicon carbide thin films as protective coating for MEMS
applications," Surface and Coatings technology, vol. 125, no. 1-3, pp. 246-250,
2000.
[182] S. Inoue, T. Namazu, H. Tawa, M. Niibe, and K. Koterazawa, "Stress control of
a-SiC films deposited by dual source dc magnetron sputtering," Vacuum, vol. 80,
no. 7, pp. 744-747, 2006.
[183] V. Bhatt and S. Chandra, "Silicon dioxide films by RF sputtering for
microelectronic and MEMS applications," Journal of Micromechanics and
Microengineering, vol. 17, no. 5, p. 1066, 2007.
[184] N. D. Hoivik, J. W. Elam, R. J. Linderman, V. M. Bright, S. M. George, and Y.
Lee, "Atomic layer deposited protective coatings for micro-electromechanical
systems," Sensors and Actuators A: Physical, vol. 103, no. 1-2, pp. 100-108,
2003.
[185] W. S. Yang and S. W. Kang, "Comparative study on chemical stability of
dielectric oxide films under HF wet and vapor etching for radiofrequency
microelectromechanical system application," Thin Solid Films, vol. 500, no. 1-2,
pp. 231-236, 2006.
[186] M. M. Rahman, R. Tero, and T. Urisu, "Shrinking of spin-on-glass films induced
by synchrotron radiation and its application to three-dimensional

99
 microfabrications," Japanese Journal of Applied Physics, vol. 43, no. 6S, p.
3941, 2004.
[187] A. A. Yasseen, J. D. Cawley, and M. Mehregany, "Thick glass film technology
for polysilicon surface micromachining," Journal of Microelectromechanical
Systems, vol. 8, no. 2, pp. 172-179, 1999.
[188] R. H. Liu, M. J. Vasile, and D. J. Beebe, "The fabrication of nonplanar spin-on
glass microstructures," Journal of Microelectromechanical Systems, vol. 8, no.
2, pp. 146-151, 1999.
[189] L.-J. Yang and S.-W. Kang, "The SOI-like method of reducing the die size of
bulk-micromachined sensors," Sensors and Materials, vol. 14, no. 1, pp. 023-
034, 2002.
[190] W. Yun, R. T. Howe, and P. R. Gray, "Surface micromachined, digitally force-
balanced accelerometer with integrated CMOS detection circuitry," in Technical
Digest IEEE Solid-State Sensor and Actuator Workshop, 1992: IEEE, pp. 126-
131.
[191] C. L. Dai and P. Z. Chang, "A CMOS surface micromachined pressure sensor,"
Journal of the Chinese Institute of Engineers, vol. 22, no. 3, pp. 375-380, 1999.
[192] C. Aydogan et al., "Fabrication of optical nanodevices through field-emission
scanning probe lithography and cryogenic etching," in Novel Patterning
Technologies 2018, 2018, vol. 10584: International Society for Optics and
Photonics, p. 105841G.
[193] A. Franke et al., "Post-CMOS integration of germanium microstructures," in
Technical Digest. IEEE International MEMS 99 Conference. Twelfth IEEE
International Conference on Micro Electro Mechanical Systems, 1999: IEEE,
pp. 630-637.
[194] S.-H. Tseng, M. S. Lu, P.-C. Wu, Y.-C. Teng, H.-H. Tsai, and Y.-Z. Juang,
"Implementation of a monolithic capacitive accelerometer in a wafer-level 0.18
µm CMOS MEMS process," Journal of Micromechanics and Microengineering,
vol. 22, no. 5, p. 055010, 2012.
[195] A. E. Franke, J. M. Heck, T.-J. King, and R. T. Howe, "Polycrystalline silicon-
germanium films for integrated microsystems," Journal of
Microelectromechanical Systems, vol. 12, no. 2, pp. 160-171, 2003.
[196] G. T. A. Kovacs, Micromachined Transducers Sourcebook, Mc Graw-Hill 1998.
[197] F. P. Incropera et al., Principles of Heat and Mass Transfer, 8th ed., John Wiley
& Sons Singapore Pte. Ltd., 2017.
[198] H. Schlichting and K. Gersten, Boundary-layer theory. Springer Science &
Business Media, 2003.
[199] H.-Y. Shih, S.-K. Lin, and P.-S. Liao, "An 80times analog-implemented time-
difference amplifier for delay-line-based coarse-fine time-to-digital converters in
0.18µm CMOS," IEEE Transactions on Very Large Scale Integration (VLSI)
Systems, vol. 23, no. 8, pp. 1528-1533, 2014.
[200] H.-Y. Shih, C.-N. Kuo, W.-H. Chen, T.-Y. Yang, and K.-C. Juang, "A 250 MHz
14 dB-NF 73 dB-gain 82 dB-DR analog baseband chain with digital-assisted DC-
offset calibration for ultra-wideband," IEEE Journal of Solid-State Circuits, vol.
45, no. 2, pp. 338-350, 2010.
[201] W.-B. Yang, H.-Y. Shih, Y.-Y. Lin, M.-H. Hong, C.-H. Wang, and Y.-L. Lo, "A
1.8-V 4.36-ppm/° C-TC bandgap reference with temperature variation
calibration," in 2013 International SoC Design Conference (ISOCC), 2013:
IEEE, pp. 103-106.

100
[202] L. Bowman and J. D. Meindl, "The packaging of implantable integrated sensors,"
IEEE Transactions on Biomedical Engineering, no. 2, pp. 248-255, 1986.
[203] H. Baltes et al., "IC MEMs microtransducers," in International Electron Devices
Meeting. Technical Digest, 1996: IEEE, pp. 521-524.
[204] W. Xu, X. Wang, X. Zhao, Z. Ke, and Y.-K. Lee, "An integrated CMOS MEMS
gas flow sensor with detection limit towards micrometer per second," in 2020
IEEE 33rd International Conference on Micro Electro Mechanical Systems
(MEMS), 2020: IEEE, pp. 200-203.
[205] W. Xu, X. Wang, Z. Ke, and Y.-K. Lee, "Bidirectional CMOS-MEMS airflow
sensor with sub-mW power consumption and high sensitivity," IEEE
Transactions on Industrial Electronics, vol. 69, no. 3, pp. 3183-3192, 2021.
[206] T. Glatzl et al., "Hot-film and calorimetric thermal air flow sensors realized with
printed board technology," Journal of Sensors and Sensor Systems, vol. 5, no. 2,
pp. 283-291, 2016.
[207] L.-J. Yang, T.-W. Huang, and P.-Z. Chang, "CMOS microelectromechanical
bandpass filters," Sensors and Actuators A: Physical, vol. 90, no. 1-2, pp. 148-
152, 2001.
[208] P.-Z. Chang and L.-J. Yang, "A method using V-grooves to monitor the thickness
of silicon membrane with m resolution," Journal of Micromechanics and
Microengineering, vol. 8, no. 3, p. 182, 1998.
[209] S. Silvestri and E. Schena, "Micromachined flow sensors in biomedical
applications," Micromachines, vol. 3, no. 2, pp. 225-243, 2012.
[210] E. Chang, L. Y. Matloff, A. K. Stowers, and D. Lentink, "Soft biohybrid
morphing wings with feathers underactuated by wrist and finger motion,"
Science Robotics, vol. 5, no. 38, p. eaay1246, 2020.
[211] L. Y. Matloff et al., "How flight feathers stick together to form a continuous
morphing wing," Science, vol. 367, no. 6475, pp. 293-297, 2020.
[212] G. Wei et al., "Self-powered hybrid flexible nanogenerator and its application in
bionic micro aerial vehicles," Nano Energy, vol. 54, pp. 10-16, 2018.
[213] S. Combes and T. Daniel, "Flexural stiffness in insect wings I. Scaling and the
influence of wing venation," Journal of Experimental Biology, vol. 206, no. 17,
pp. 2979-2987, 2003.
[214] L.-J. Yang, H.-H. Wang, P.-C. Yang, Y.-C. Chung, and T.-S. Sheu. "New
packaging method using PDMS for piezoresistive pressure sensors." Sensors and
Materials 19, no. 7, 391-402, 2007.
[215] Y.-S. Liu and K.-A. Wen, "Monolithic low noise and low zero-g offset
CMOS/MEMS accelerometer readout scheme," Micromachines, vol. 9, no. 12,
p. 637, 2018.
[216] J. Smith, S. Montague, J. Sniegowski, J. Murray, and P. McWhorter, "Embedded
micromechanical devices for the monolithic integration of MEMS with CMOS,"
in Proceedings of International Electron Devices Meeting, 1995: IEEE, pp. 609-
612.
[217] Y.-C. Lin, P.-H. Hong, S.-K. Yeh, C.-C. Chang, and W. Fang, "Monolithic
integration of pressure/humidity/temperature sensors for CMOS-MEMS
environmental sensing hub with structure designs for performances
enhancement," in 2020 IEEE 33rd International Conference on Micro Electro
Mechanical Systems (MEMS), 2020: IEEE, pp. 54-57.
[218] A. Morris, "Monolithic integration of RF-MEMS within CMOS," in 2015
International Symposium on VLSI Technology, Systems and Applications, 2015:
IEEE, pp. 1-2.

101
[219] C.-M. Sun, C. Wang, M.-H. Tsai, H.-S. Hsieh, and W. Fang, "Monolithic
integration of capacitive sensors using a double-side CMOS MEMS post
process," Journal of Micromechanics and Microengineering, vol. 19, no. 1, p.
015023, 2008.
[220] H. Baltes, O. Brand, G. K. Fedder, C. Hierold, J. G. Korvink, and O. Tabata,
CMOS-MEMS. John Wiley & Sons, 2013.
[221] H. Baltes, O. Brand, G. K. Fedder, C. Hierold, J. G. Korvink, and O. Tabata,
"CMOS-MEMS," Advanced Micro and Nanosystems, vol. 2, p. 608, 2005.
[222] L.-J. Yang, H.-H. Wang, P.-C. Yang, Y.-C. Chung, and T.-S. Sheu, "New
packaging method using PDMS for piezoresistive pressure sensors," Sensors and
Materials, vol. 19, no. 7, pp. 391-402, 2007.
[223] A. M. Kaynia, "Seismic considerations in design of offshore wind turbines," Soil
Dynamics and Earthquake Engineering, vol. 124, pp. 399-407, 2019.
[224] R. Wlezien et al., "The aircraft morphing program," in 39th
AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials
Conference and Exhibit, 1998, p. 1927.
[225] S. Morris and J. Foss, "Transient thermal response of a hot-wire anemometer,"
Measurement Science and Technology, vol. 14, no. 3, p. 251, 2003.
[226] M. M. Pedersen, T. J. Larsen, H. A. Madsen, and G. C. Larsen, "Using wind
speed from a blade-mounted flow sensor for power and load assessment on
modern wind turbines," Wind Energy Science, vol. 2, pp. 547-567, 2017.
[227] M. M. Pedersen, T. J. Larsen, H. A. Madsen, and S. J. Andersen, "Free flow wind
speed from a blade-mounted flow sensor," Wind Energy Science, vol. 3, pp. 121-
138, 2018.

102
Appendix A: Calorimetric flow sensor

During the course of the research work, the author had the privilege to work with
another concept of thermo resistive calorimetric MEMS-based flow sensor design in
collaboration with HKUST Professor Y.K Lee research team.
This work manifests another concept of thermo resistive calorimetric MEMS-
based flow sensor. The sensor comprises of two RTDs connected in series as bridge
circuit. Measuring the resistance value across the sensing element at given temperature
or constant heating power allows the output to reversely predict the flow speed with
acceptable accuracy. The layer design is again done by using CADENCE software. The
material used for designing the resistors and heater is polysilicon. This CMOS sensor
design is realized by using U18MEMS process, which is a combination of UMC 0.18µm
1P6M process along with MEMS post process. The dimensions of the sensor were
optimized to achieve improved sensitivity [34, 40, 52, 168].

A1 Working principle
The author adopted a bridge circuit where two RTDs are connected in series. In
our previous work [55], the design of the sensor was done with a self-heating hot-wire
type where it didn’t use any heater during the operation. In this new design, an additional
heater is added beside the bridge sensing resistors of calorimetric type so that sensitivity
improvement can be seen. The material used for the design is polysilicon. The sensing
operation follows the principle of heat convection as shown in Figure A1. When air
velocity passes across the resistors, it creates a differential temperature which in turn
gives a voltage signal output. The CMOS flow sensor drawn in CADENCE software.

103
 Figure A1 Calorimetric CMOS MEMS Flow sensor

A2 Layout and tape out

Layout design is made with 0.18 m standard CMOS foundry with MEMS post
etching in UMC shown in Figure 2.1. with metal layers M1-M6. However, the process
flow for MEMS device with polysilicon sensing elements which will be damage during
the isotropic undercut etching in the final step of the post process and the gate oxide
thickness is probably 4 nm and hard to protect the polysilicon gate/sensor from the final
undercut etching. Figure A2 represents (a) Layout of the CMOS flow sensor drawn in
CADENCE; (b) Cross sectional view of the flow sensor; (c) Packaged chip received
after tape out. Therefore, throughout the sensor design, shallow-trench-isolation (STI)
oxide is used to protect polysilicon during the final undercut etching, shown in Figure
A2 (b).

104
 (a)

(b)

(c)
Figure A2 (a) Layout diagram; (b) Cross sectional view; (c) Packaged chip

A3 Testing
This work presents a solitary design of thermal flow sensor which is fabricated
using UMC 0.18 µm 1P6M CMOS MEMS technology. The dimensions of the sensors
were optimized to achieve better sensitivity. Moreover, to improve the thermal-isolation
functionality of the sensor, plenty of 88 µm2 holes were created using the MEMS post-
etching. Packaging is done using the SB28 package provided TSRI. As shown in Figure
A3 (a), the wind tunnel test was done with the application of 1.5 V for bridge sensing
resistor and different voltages for heater power supply. Figure A3 (b) shows the output
voltage vs. wind speed graph. Various ranges were analyzed among which the sensor
achieved a maximum sensitivity of 160 µV/(m/s)/mW at 2.5 V heating power. On the
other hand, the heater power supply to 1.8 V and 3.3 V couldn’t achieve very good

105
sensitivity. These results can be supported by chart which is plotted corresponding to the
observation made in Table A1. Table A2 shows the different sensitivities resulting due
to different heating voltages. The improved sensor layout has increased sensitivity
compared to the previous work (138µV/(m/s)/mW) [55]. Since the resistivity of the
polysilicon material cannot be adjusted to an optimized value, the best sensitivity hasn’t
been achieved yet. Figure A4 (a) shows the wind tunnel test set up; A4 (b) shows the
sensitivity comparison graph for different heating voltages.

Table A1 Vout for different heating power supply

Wind Speed Vout(Rh=1.8V) Vout(Rh=2.5V) Vout(Rh=3.3V)
(m/s)
1 0.736 0.758 0.784
2 0.736 0.758 0.783
3 0.736 0.757 0.783
4 0.736 0.756 0.783
5 0.736 0.756 0.783
6 0.736 0.754 0.782
7 0.736 0.752 0.782
8 0.736 0.751 0.782
9 0.735 0.749 0.779
10 0.735 0.748 0.779

Table A2 Sensitivity comparison for different heating power supply

Sensor power Measured Normalized
Heater power
supply (V) Speed (m/s) Sensitivity sensitivity
supply (V)
(mV/(m/s)) (µV/(m/s)/mW)
1.5 1.8 0-10 0.1 30
1.5 2.5 0-10 1.0 160
1.5 3.3 0-10 0.5 46
1.8 No heater 0-10 0.2 138

106
(a)

(b)

(c)
Figure A3 Windspeed vs. Vout

107
 (a)

Comparison
0.79
Output Voltage(V)

0.78

0.77

0.76 Rh = 1.8 V

0.75 Rh = 2.5 V
Rh = 3.3 V
0.74

0.73
0 1 2 3 4 5 6 7 8 9 10 11 12

Wind Speed(m/s)
(b)
Figure A4 (a) Test set up; (b) graph for different heating voltages

Wind tunnels are large tubes with air moving side in a controlled manner. Wind
tunnel is used to measure the aerodynamic forces produced and experienced by the
experiment object inside it. The sensor was successfully tested in a wind tunnel for
various ranges of wind speed. This sensor is used for measuring air velocity ranging
from 0-10 m/s with three different heating voltages applied to the center heater of 1K
ohm. The sensor achieved a maximum sensitivity at 2.5 V resulting in 1 mV/(m/s) giving
a normalized output sensitivity of 160 µV/(m/s)/mW.

108
 Table A3 Output voltage of integrated calorimetric sensor and IA
Wind speed(m/s) Vout(V)
0 0.951
1 0.929
2 0.931
3 0.926
4 0.924
5 0.922

Figure A5 Graph of integrated IA and calorimetric sensor

The calculated sensitivity value for calorimetric flow sensor according to the
Table A3 and Figure A5 graph is 146.4µV/V/(m/s)/mW considering power dissipation
as 22 mW.

109
Appendix B: More graphs of integrated chip test in
wind tunnel

Table B1 shows the output voltage of integrated self-heating sensor and IA and
it’s respective graph is plotted in Figure B1. Similarly, another case where Table B2
represents the output voltage of integrated self-heating flow sensor and IA and the graph
is plotted in Figure B2.

Table B1 Output voltage of integrated self-heating sensor and IA

Wind speed(m/s) Vout(V)
0 1.176
1 1.174
2 1.174
3 1.173
4 1.175
5 1.173
6 1.173

Figure B1 Graph of integrated self-heating sensor and IA for 0-6m/s

110
 Table B2 Output voltage of integrated self-heating flow sensor and IA
Wind speed(m/s) Vout(V)
0 1.165
1 1.164
2 1.166
3 1.163
4 1.167
5 1.163
6 1.162

Figure B2 Graph of integrated self-heating flow sensor and IA

Table B3 is the output voltage of integrated calorimetric flow sensor and

instrumentation amplifier and Figure B3 is its corresponding graphical representation.
Similarly, Table B4 is another output voltage of integrated calorimetric sensor and IA
and its graph is plotted in Figure B4.

111
Table B3 Output voltage of integrated calorimetric flow sensor and IA
Wind speed(m/s) Vout(V)
0 0.998
1 0.969
2 0.972
3 0.971
4 0.974
5 0.973
6 0.973

Figure B3 Graph of integrated calorimetric flow sensor and IA

Table B4 Output voltage of integrated calorimetric sensor and IA

Wind speed(m/s) Vout(V)
0 1.072
1 1.14
2 1.15
3 1.133
4 1.001
5 1.148
6 1.143

112
 Figure B4 Graph of integrated calorimetric flow sensor and IA

Figure B5 MATLAB processed data for the lift measurement in force gauge (a-f) and
calibrated data from the flow sensor on leading edge and mean chord length of FWMAV
(a1-f1) for configuration B1, B2 for 1.6s time period with reference to Figure 5.7.
Figure B6 MATLAB processed data for the thrust measurement in force gauge (a-f)
and calibrated data from the flow sensor on leading edge and mean chord length of
FWMAV for (a1-f1) for configuration B1, B2 for 1.6s time period with reference to
Figure 5.7.

113
(a) (a1)

(b) (b1)

(c) (c1)

(d) (d1)

(e) (e1)

(f) (f1)
Figure B5 Lift measurement in force gauge and calibrated data of flow sensor

114
(a) (a1)

(b) (b1)

(c) (c1)

(d) (d1)

(e) (e1)

(f) (f1)
Figure B6 Thrust measured in force gauge and calibrated data of flow sensor

115
Appendix C: MATLAB code
This code was used to reduce the noise in the MATLAB processed data.

C.1 Cut-off frequency

clc;clear
xlsFile_force ='file name.xlsx';
force = xlsread(xlsFile_force);
fz= force (:,4);
N = size(fz,1);
fs = 1000;
freqStep = fs/N;
% f = 10*freqStep;
time = (0:N-1)/fs;
cutOffFreq=5; % Cutoff frequency
filterOrder=8; % Order of filter
[b, a]=butter(filterOrder, cutOffFreq/(fs/2), 'low');
fz_cut_8= filter(b,a,fz);
% Plot spectral magnitude
% subplot(4,2,1);
plot(time, fz_cut_8, '.-');grid on
title('filter 8');
xlabel('Time (seconds)'); ylabel('Lift force (gw)');
axis([0 inf -200 200])
%axis square
Plot single-sided amplitude spectrum.
subplot(4,2,1);
filter_f_z_8 = fft(fz_cut_8,NFFT)/N;
fr = fs/2*linspace(0,1,NFFT/2+1);
plot(fr,2*abs(filter_f_z_8(1:NFFT/2+1)), '.-b'); grid on
axis([0 300 0 inf])
title('Amplitude Spectrum')
xlabel('Frequency (Hz)')
ylabel('|Fz(f)|')

116
C.2 FFT
xlsFile_force ='file name.xlsx';
force = xlsread(xlsFile_force);
fz= force (:,4);
N = size(fz,1);
fs = 1000;
freqStep = fs/N;
f = 10*freqStep;
time = (0:N-1)/fs;
subplot(4,2,1);
plot(time, fz, '.-');
title('Fz');
xlabel('Time (seconds)'); ylabel('Lift force (gw)');
axis([0 inf -50 50])
axis square
Plot spectral magnitude
NFFT = 2^nextpow2(N);
Y = fft(fz,NFFT)/N;
fr = fs/2*linspace(0,1,NFFT/2+1);
Plot single-sided amplitude spectrum.
subplot(4,2,3);
plot(fr,2*abs(Y(1:NFFT/2+1)), '.-b'); grid on
axis([0 300 0 inf])
title('Amplitude Spectrum')
xlabel('Frequency (Hz)')
ylabel('|Fz(f)|')
axis square
cutOffFreq=4;
filterOrder=2;
[b, a]=butter(filterOrder, cutOffFreq/(fs/2), 'low');
fz_cut_2= filter(b,a,fz);
subplot(4,2,2);
plot(time, fz_cut_2, '.-');grid on
title('filter 2');
xlabel('Time (seconds)'); ylabel('Lift force (gw)');
axis([0 inf -50 50])

117
Appendix D: Sensor on wind turbine blade

Applications on offshore wind turbine will be conducted to show the practical

value of the on-site flow sensor technology. Take the example of 8MW offshore wind
turbine, the diameter is up to 160 m. The longer turbine blade will cause the smaller
natural frequency and the lower mechanical stiffness according to the scaling law [223].
Therefore, the turbine blade is no longer regarded as a rigid body, and the corresponding
problem of fluid-structure interaction (FSI) will be serious (even though the upstream
velocity is regarded as a constant value.)
Moreover, the next-generation wind turbine may be so large as to include several
wind blowing clusters with non-uniform flow field. In other words, the boundary
condition of “uniform free stream velocity” is no more valid, and we need to investigate
the non-uniform wind loading on the big turbine blade surface. The very costly (1 billion
TWD per station) wind turbines of course deserve us to optimize its operation in a state-
of–art manner. And the distributed flow sensing technology is necessary to the above
research investigation [224].
The conventional anemometers [225] wind speed meters with configuration of
hemispherical cup, vane type, hot-wire, LDA (Laser-Doppler anemometer), ultrasonic
type, and pitot-tube are all lumped single sensors to measure the air speed of a single
point. These sensors need not to mention their spatial resolution any more. They are not
proper to measure the detailed velocity field around the big wind turbine with the future
application scenario as below.
1. Long-term monitoring the wind speed of the wind turbine [25];

2. Real-time measuring the abnormal deformation or RPM value of the wind

turbine [26];

3. Precisely predicting the fracture location on the turbine blade from the on-site
flow sensor array signal [27];

4. With adding the self-energy harvester, the flow sensor array on the turbine
blade surface don’t need the external power supply [28].

5. Flexible packaging way is necessary as the turbine blade having large

deformation subject to huge wind [33].

6. A robust instrument and measurement system is required to obtain accurate and

reliable wind speed recordings [38, 39].

118
 It has been a while ago Taiwan wind industry association was officially founded
on April 2012 for facilitating the development of wind power industry in Taiwan. Its
purpose is to combine personnel from research institutions, industries and academia
together to promote wind power industry development strategy, to provide information
and technical exchanges, to actively promote industrial vertical and horizontal
integration and thus to enhance wind power industry. The Global Wind Energy Council
(GWEC) is promoting Wind Force 12, a global industry blueprint that aims to enable
1,250 GW of wind power installed worldwide by 2020 supplying 12 % of the world's
electricity. This year's Wind Energy Taiwan will try to offer the best platform to connect
key players in various fields, including raw material suppliers, component/subsystem
suppliers, balance of plant (BoP), offshore and onshore wind turbine systems, and
developers and operators. But, one major risky factor involve in this project is that the
maintenance of wind turbine is very cost effective, lack of maintenance will lead to
damage of turbine blades.
Wind turbines may be as high as a skyscraper, but they owe their ability to
operate efficiently and safely to the dozens of tiny, inexpensive sensors that can monitor
their health. We know that the aerodynamics analysis phenomena of wind turbines are
very much similar with Flapping wing micro air vehicles, we can firstly test the flow
sensor in smaller prototype, later, we can examine the same in wind turbine
too. Furthermore, it is concluded that a robust instrument and measurement system is
required to obtain accurate and reliable wind speed recordings [226, 227]. So, our flow
sensor may be helpful to detect the defected area in the wind turbine blades and try to
reduce further damages. Finally, by designing the best one we can show the ability to
compete with the rest of the world.

119
Publications
Journal papers
1. Lung-Jieh Yang, Reshmi Waikhom, Horng-Yuan Shih, and Yi-Kuen Lee,
"Foundry Service of CMOS MEMS processes and the Case Study of the Flow
Sensor," Processes, vol.10, no. 7, 1280, 2022.
2. Lung-Jieh Yang, Chandrashekhar Tasupalli, Reshmi Waikhom, and Nikhil
Panchal, "Soap film visualization of a 10 cm-span flapping
wing," Fluids, vol. 6, no. 10 (2021): 361.
3. Lung-Jieh Yang, Reshmi Waikhom, Wei-Chen Wang, Vivek Jabaraj Joseph,
Balasubramanian Esakki, Neethish Kumar Unnam, Xiu-Han Li, and Chi
Yuan Lee, "Check-valve design in enhancing aerodynamic performance of
flapping wings," Applied Sciences, vol. 11, no. 8 (2021): 3416.
4. Lung-Jieh Yang, Niroj Kapri, Reshmi Waikhom, and Neethish Kumar
Unnam, "Fabrication, aerodynamic measurement and performance
evaluation of corrugated flapping wings," J. Aeronaut. Astronaut. Aviat, vol.
53 (2021): 83-94.
5. Lung-Jieh Yang, Balasubramanian Esakki, and Reshmi Waikhom, "Inertial
effect on the time-averaged lift of flapping wings," Journal of Applied
Science and Engineering, vol. 23, no. 2 (2020): 357-359.
6. Saravana, Kompala, Balasubramanian Esakki, Lung-Jieh Yang, Wei-Chen
Wang, Reshmi Waikhom, and Chih-Yu Jen. "Fabrication of flapping wing
mechanism using fused deposition modeling and measurement of
aerodynamic forces." Journal of Aeronautics, Astronautics and
Aviation, vol. 51, no. 1 (2019): 131-140.

120
Conference papers
1. Reshmi Waikhom, Chandrashekhar Tasupalli, Lung-Jieh Yang, and Horng-
Yuan Shih, “Preliminary integration test of a CMOS flow sensor with on-
chip instrumentation amplifier,” Symposium on Nano Devices and Circuit
Technology (SNDCT-2022), TSRI, Hsinchu, Taiwan (R.O.C), May 19-20,
2022, pp. 63.
2. 楊龍杰、強度、衛瑞紗” 以皂膜顯像術分析拍翼升力與推力”，台灣機電
工程國際學會全國學術研討會，2022 年 4 月 22-23 日，彰師大，ISME
2022, ES10-327。
3. Lung-Jieh Yang, Reshmi Waikhom, and Horng-Yuan Shih, "Developing
CMOS MEMS flow sensors on a flapping wing surface," The 17th
International Conference on Intelligent Unmanned Systems (ICIUS-2021),
Ho Chi Minh City, Vietnam, Aug. 25-27, 2021, paper no. 29.
4. Waikhom, Reshmi, Lung-Jieh Yang, Horng-Yuan Shih, and Cai-Rong Kuo.
"Self-heating CMOS flow sensor." In 2021 21st International Conference on
Solid-State Sensors, Actuators and Microsystems (Transducers) 20-25, June
2021, pp. 1279-1282.
5. 衛瑞紗, 施鴻源, 楊龍杰, “安裝於拍翼表面之金氧半微機電流速計開發”,
台灣機電工程國際學會全國學術研討會, 2021 年 5 月 28-29 日, 中原大學,
ISME 2021, paper no. 103.
6. Reshmi Waikhom, Lung-Jieh Yang, Wei-Chen Wang, Saravana Kompala,
“Enhancing the Aerodynamic performance of flapping wings using check-
valves,” 44th National Conference on Theoretical and Applied Mechanics,
Yilan, Taiwan, Nov. 26-27, 2020.

Patents
1. 楊龍杰, 費約瑟, 衛瑞紗, 微流道矽膠口罩, Taiwan Patent I739591, Sep. 11,
2021 (Approved).
2. 楊龍杰 , 衛瑞紗 , 互補金氧半微機電感測器晶片的製造方法及其裝置 ,
Taiwan Patent Pending 109146849, Dec. 30, 2020 (Application).

121