CMOS Integrated Lab-on-a-chip System for Personalized Biomedical Diagnosis
Description
A thorough examination of lab-on-a-chip circuit-level operations to improve system performance
A rapidly aging population demands rapid, cost-effective, flexible, personalized diagnostics. Existing systems tend to fall short in one or more capacities, making the development of alternatives a priority. CMOS Integrated Lab-on-a-Chip System for Personalized Biomedical Diagnosis provides insight toward the solution, with a comprehensive, multidisciplinary reference to the next wave of personalized medicine technology.
A standard complementary metal oxide semiconductor (CMOS) fabrication technology allows mass-production of large-array, miniaturized CMOS-integrated sensors from multi-modal domains with smart on-chip processing capability. This book provides an in-depth examination of the design and mechanics considerations that make this technology a promising platform for microfluidics, micro-electro-mechanical systems, electronics, and electromagnetics.
From CMOS fundamentals to end-user applications, all aspects of CMOS sensors are covered, with frequent diagrams and illustrations that clarify complex structures and processes. Detailed yet concise, and designed to help students and engineers develop smaller, cheaper, smarter lab-on-a-chip systems, this invaluable reference:
- Provides clarity and insight on the design of lab-on-a-chip personalized biomedical sensors and systems
- Features concise analyses of the integration of microfluidics and micro-electro-mechanical systems
- Highlights the use of compressive sensing, super-resolution, and machine learning through the use of smart SoC processing
- Discusses recent advances in complementary metal oxide semiconductor-integrated lab-on-a-chip systems
- Includes guidance on DNA sequencing and cell counting applications using dual-mode chemical/optical and energy harvesting sensors
The conventional reliance on the microscope, flow cytometry, and DNA sequencing leaves diagnosticians tied to bulky, expensive equipment with a central problem of scale. Lab-on-a-chip technology eliminates these constraints while improving accuracy and flexibility, ushering in a new era of medicine. This book is an essential reference for students, researchers, and engineers working in diagnostic circuitry and microsystems.
More details
Other editions
Additional editions
Persons
Hao Yu, Southern University of Science and Technology, China, is an assistant professor and area director of the VIRTUS/VALENS Centre of Excellence.
Mei Yan, Consultant, developed lab-on-a-chip biomedical sensor circuits and systems for personalized biomedical diagnosis as a Research Fellow at Nanyang Technological University in Singapore.
Xiwei Huang, Hangzhou Dianzi University, China, is an assistant professor at the School of Electronics and Information.
Content
Preface x
1 Introduction 1
1.1 Personalized Biomedical Diagnosis 1
1.1.1 Personalized Diagnosis 1
1.1.2 Conventional Biomedical Diagnostic Instruments 3
1.1.2.1 Optical Microscope 3
1.1.2.2 Flow Cytometer 4
1.1.2.3 DNA Sequencer 5
1.2 CMOS Sensor-based Lab-on-a-Chip for System Miniaturization 7
1.2.1 CMOS Sensor-based Lab-on-a-Chip 7
1.2.2 CMOS Sensor 8
1.2.2.1 CMOS Process Fundamentals 8
1.2.2.2 CMOS Sensor Technology 10
1.2.2.3 Multimodal CMOS Sensor 13
1.2.3 Microfluidics 14
1.2.3.1 Microfluidic Fundamentals 14
1.2.3.2 Microfluidics Fabrication 16
1.3 Objectives and Organization of this Book 20
1.3.1 Objectives 20
1.3.2 Organization 20
References 21
2 CMOS Sensor Design 25
2.1 Top Architecture 25
2.2 Noise Overview 25
2.2.1 Thermal Noise 26
2.2.2 Flicker Noise 27
2.2.3 Shot Noise 28
2.2.4 MOSFET Noise Model 29
2.3 Pixel Readout Circuit 29
2.3.1 Source Follower 30
2.3.2 Sub-threshold Gm Integrator 33
2.3.3 Ctia 35
2.4 Column Amplifier 38
2.5 Column ADC 39
2.5.1 Single-Slope ADC 39
2.5.2 Sigma-Delta ADC 43
2.6 Correlated Sampling 49
2.6.1 Correlated Double Sampling 49
2.6.2 Correlated Multiple Sampling 51
2.7 Timing Control 52
2.7.1 Row Timing Control 52
2.7.2 Column Timing Control 55
2.8 LVDS Interface 57
References 59
3 CMOS Impedance Sensor 60
3.1 Introduction 60
3.2 CMOS Impedance Pixel 61
3.3 Readout Circuit 63
3.4 A 96 × 96 Electronic Impedance Sensing System 65
3.4.1 Top Architecture 65
3.4.2 System Implementation 67
3.4.2.1 System Setup 67
3.4.2.2 Sample Preparation 68
3.4.3 Results 68
3.4.3.1 Data Fitting for Single Cell Impedance Measurement 69
3.4.3.2 Cell and Electrode Impedance Analysis 71
3.4.3.3 EIS for Single-Cell Impedance Enumeration 71
References 74
4 CMOS Terahertz Sensor 76
4.1 Introduction 76
4.2 CMOS THz Pixel 76
4.2.1 Differential TL-SRR Resonator Design 76
4.2.1.1 Stacked SRR Layout 76
4.2.1.2 Comparison with Single-ended TL-SRR Resonator 80
4.2.1.3 Comparison with Standing-Wave Resonator 82
4.2.2 Differential TL-CSRR Resonator Design 83
4.3 Readout Circuit 84
4.3.1 Super-regenerative Amplification 84
4.3.1.1 Equivalent Circuit of SRA 84
4.3.1.2 Frequency Response of SRA 86
4.3.1.3 Sensitivity of SRA 86
4.3.2 Super-regenerative Receivers 87
4.3.2.1 Quench-controlled Oscillation 87
4.3.2.2 SRX Design by TL-CSRR 89
4.3.2.3 SRX Design by TL-SRR 91
4.4 A 135 GHz Imager 94
4.4.1 135 GHz DTL-SRR-based Receiver 94
4.4.2 System Implementation 95
4.4.3 Results 95
4.5 Plasmonic Sensor for Circulating Tumor Cell Detection 98
4.5.1 Introduction of CTC Detection 98
4.5.2 SRR-based Oscillator for CTC Detection 99
4.5.3 Sensitivity of SRR-based Oscillator 101
References 103
5 CMOS Ultrasound Sensor 106
5.1 Introduction 106
5.2 CMUT Pixel 107
5.3 Readout Circuit 109
5.4 A 320 × 320 CMUT-based Ultrasound Imaging System 110
5.4.1 Top Architecture 110
5.4.2 System Implementation 111
5.4.2.1 Process Selection 111
5.4.2.2 High Voltage Pulser 112
5.4.2.3 Low-Noise Preamplifier and High Voltage Switch 115
5.4.3 Results 116
5.4.3.1 Simulation Results 116
5.4.3.2 Two-channel AFE IC Measurement Results 117
5.4.3.3 Acoustic Transmission Testing with AFE IC and CMUT 121
5.4.3.4 Acoustic Pulse-echo Testing with AFE IC and CMUT 122
References 124
6 CMOS 3-D-Integrated MEMS Sensor 126
6.1 Introduction 126
6.2 MEMS Sensor 127
6.3 Readout Circuit 127
6.4 A 3-D TSV-less Accelerometer 129
6.4.1 CMOS-on-MEMS Stacking 129
6.4.2 Bonding Reliability 132
6.4.2.1 Al-Au Thermo-compression Shear Strength 132
6.4.2.2 Al-Au Thermo-compression Hermeticity 134
6.4.3 Results 135
6.4.3.1 Standalone Validation of the Readout Circuit 135
6.4.3.2 Functionality Testing of CMOS-on-MEMS Chip 136
6.4.3.3 Reliability Testing of CMOS-on-MEMS Chip 138
References 141
7 CMOS Image Sensor 142
7.1 Introduction 142
7.2 CMOS Image Pixel 145
7.2.1 Structure 145
7.2.1.1 FSI 4 T Pixel 145
7.2.1.2 Back Side Illumination Pixel 147
7.2.1.3 Stack Pixel 148
7.2.2 Noise and Model 150
7.2.2.1 Photon Shot Noise 151
7.2.2.2 Reset Noise 152
7.2.2.3 Thermal Noise 152
7.2.2.4 Flicker Noise 154
7.2.2.5 Fixed Pattern Noise 154
7.3 Readout Circuit 155
7.3.1 Global Serial Readout 156
7.3.2 Correlated Double Sampling 156
7.4 A 3.2 Mega CMOS Image Sensor 158
7.4.1 4-way Shared Pixel Unit 158
7.4.2 Top Architecture 159
7.4.3 System Implementation 162
7.4.4 Results 164
7.4.4.1 System Characterization 164
7.4.4.2 Digital CDS for FPN Reduction 164
7.4.4.3 Blood Cell Imaging Experiments 165
References 167
8 CMOS Dual-mode pH-Image Sensor 169
8.1 Introduction 169
8.2 CMOS Dual-mode pH-Image Pixel 170
8.3 Readout Circuit 172
8.3.1 CDS for Optical Sensing 174
8.3.2 CDS for Chemical Sensing 174
8.4 A 64 × 64 Dual-mode pH-Image Sensor 175
8.4.1 Top Architecture 175
8.4.2 System Implementation 177
8.4.3 Results 177
References 184
9 CMOS Dual-mode Energy-harvesting-image Sensor 186
9.1 Introduction 186
9.2 CMOS EHI Pixel 187
9.3 Readout Circuit 191
9.4 A 96 × 96 EHI Sensing System 195
9.4.1 Top Architecture 195
9.4.2 System Implementation 197
9.4.3 Results 203
References 211
10 DNA Sequencing 213
10.1 Introduction 213
10.2 CMOS ISFET-based Sequencing 213
10.2.1 Overview 213
10.2.2 ISFET-based Sequencing Procedure 215
10.3 CMOS THz-based Genotyping 220
10.3.1 Overview 220
10.3.2 THz-based Genotyping Procedure 220
10.4 Beyond CMOS Nanopore Sequencing 221
10.4.1 Overview 221
10.4.2 Nanopore-based Sequencing Procedure 223
10.5 Summary 227
References 230
11 Cell Counting 231
11.1 Introduction 231
11.2 Optofluidic Imaging System 231
11.2.1 Contact Imaging 231
11.2.2 Optofluidic Imaging System Model 232
11.2.2.1 Resolution Model 232
11.2.2.2 Dynamic Range Model 233
11.2.2.3 Implication to SR Processing 234
11.3 Super-resolution Image Processing 234
11.3.1 Multi-frame SR Processing 235
11.3.2 Single-frame SR Processing 236
11.4 Machine-learning-based Single-frame Super-resolution 237
11.4.1 Elmsr 238
11.4.2 Cnnsr 242
11.5 Microfluidic Cytometer for Cell Counting 245
11.5.1 Microfluidic Cytometer System 245
11.5.1.1 System Overview 245
11.5.1.2 Microfluidic Channel Fabrication 246
11.5.1.3 Microbead and Cell Sample Preparation 246
11.5.1.4 Microfluidic Cytometer Design 247
11.5.1.5 Cell Detection 248
11.5.1.6 Cell Recognition 249
11.5.1.7 Cell Counting 250
11.5.2 Results 250
11.5.2.1 Counting Performance Characterization 250
11.5.2.2 Off-Line SR Training 251
11.5.2.3 On-line SR Testing 253
11.5.2.4 On-line Cell Recognition and Counting 254
References 255
12 Conclusion 258
12.1 Summaries 258
12.2 Future Works 260
Index 262
1
Introduction
1.1 Personalized Biomedical Diagnosis
1.1.1 Personalized Diagnosis
The world's older population continues to grow at an unprecedented rate. The proportion of people aged 60 years and over grows faster than any other age group, as shown in Figure 1.1 [1]. According to the expectation of the World Health Organization [2], between 2000 and 2050, the world's population of over 60 years of age will double from about 11% to 22%, and the absolute number of such people will also increase from 605 million to 2 billion. Among the aging countries, the most dramatic changes are now taking place in low and middle income countries with limited biomedical infrastructure, incomplete healthcare systems, and shortage of funds and resources. The current aging society also comes with special healthcare challenges due to limited hospital resources, doctors and related facilities. A portable and low-cost biomedical diagnosis instrument is thereby in high demand to meet the needs of the growing aging population in the form of personalized biomedical diagnosis.
Figure 1.1 Global aging trends for the percentage of the total population at 60 years of age or over in 2012 and 2050 [1].
Over the past several decades, biomedical diagnostic techniques such as the microscope [3], ultrasound [4-6], flow cytometry [7-9], and genetic sequencing [10-12], have improved the accurate monitoring of existing diseases, and also the understanding of the underlying causes of those diseases. However, to obtain highly sensitive measurements, current diagnosis instrument systems are usually bulky and expensive, their complicated operation requiring professional personnel. As such, they are usually only available in established hospitals or clinics, and hence are not flexible for multiple functionality diagnoses for on-site personalized diagnosis. These problems pose significant challenges for the personalized healthcare of aging populations, especially in low-income developing countries. As such, portable and affordable biomedical instruments that can be miniaturized are required to provide a point-of-care (POC) diagnosis [13-15].
A POC biomedical instrument is meant to perform the diagnosis at the site of patient care by a clinician or by the patient without the need for clinical laboratory facilities. The tests are rapid, portable, non-invasive, and easy-to-use, with timely testing results, which allow rapid clinical decision-making and also mitigate treatment delay. The development of these POC diagnostic and monitoring instruments thereby allow individuals, especially these older people, to monitor their own health. It leads to a paradigm shift from conventional curative medicine, to predictive and personalized diagnostics [16]. Moreover, because factors such as medication adherence, genetics, age, nutrition, health, and environmental exposure can vary, and also the extent of biomedical treatment and drug response of each individual, people are becoming more interested in exploring the biology of the disease and its treatment at his or her own individual level. For example, people can use the information about his or her own genes, proteins, metabolites, etc. at the molecular level, and the leverage with the existing environment to prevent, diagnose, and treat the disease at the individual level. Therefore, such personalized biomedical diagnostics, with the existing supporting instrument platform, has emerged as a significant need for the coming aging world.
1.1.2 Conventional Biomedical Diagnostic Instruments
In the following section, three of the most widely-used traditional biomedical diagnosis instruments, namely the high-resolution optical microscope, flow cytometer, and DNA sequencer, are discussed as the starting point for comparison.
1.1.2.1 Optical Microscope
A microscope is an optical instrument that produces a magnified image of the biomedical object under inspection, compared with what the naked human eye can observe, using visible light with lenses. The optical microscope was invented more than 400 years ago by two Dutch spectacle makers, Hans and Zaccharias Janssen, then improved by Galileo and Antonie van Leeuwenhoek, and is the leading high-resolution visualization tool and the gold standard for biomedical imaging at the cellular level [17].
To achieve micrometer or sub-micrometer resolution, almost all microscopes require precise and expensive optical lenses, as well as a large distance between the object lens and eyepiece lens for the light to travel and reshape, as shown in Figure 1.2. The object to be observed is illuminated by a light source. As light passes through the object, the objective lens (i.e. the lens closest to the object) produces the corresponding magnified object image in the primary image angle. The eyepiece (i.e. the lens that people look into) acts as a magnifier that produces an enlarged image by the objective lens. The overall magnification of the microscope system is the multiplication of both the object and the eyepiece. The principle of magnification is based on the thin lens approximation as follows:
(1.1)where Li and Lo are the image distance and object distance, F is the focal length of the objective lens, M is the magnification factor of the objective lens, and H1 and H2 are the sizes of object and image respectively. Therefore, a significant space Li is usually required to produce a large microscopic magnification, which is the main difficulty for the minimization of the optical microscope.
Figure 1.2 A typical microscope and its optical path with objective lens and eye piece. To reach high-resolution imaging capability, bulky, expensive, and sophisticated lenses are required.
Compared with the earliest compound microscope, the current design has evolved to incorporate multiple lenses, filters, polarizers, beam-splitters, sensors, illumination sources, and a host of other components, aimed at improving resolution and sample contrast. However, this basic microscope design has undergone very few fundamental changes over the centuries, so it bulky, expensive, and complicated, hence not suitable for the desired POC diagnosis.
1.1.2.2 Flow Cytometer
Flow cytometer is another widely-used biomedical instrument for applications in, for example, blood cell counting and sorting. Based on basic working principles, there are two types of flow cytometry methods, optical-based and electrochemical-based, as shown in Figure 1.3. With the help of sheath fluid and fluid dynamics effects, the blood cells are injected and passed through the measuring tunnel one at a time. For the optical-based cytometry shown in Figure 1.3(a), each cell along the path interacts with the laser beam respectively and the light intensity of the scattering is measured by the forward and orthogonal optical detectors. The measurement results depend on the size of the cell and its internal complexity, which can be used for cell differentiation and counting. Whereas, for the electrochemical based cytometry shown in Figure 1.3(b), pairs of electrodes are placed on both sides of a narrow orifice and a low-frequency current is applied. When blood cells are driven through, the impedance values are measured, which vary with the cell size and composition.
Figure 1.3 Diagrams of: (a) Optical-based; and (b) Electrochemical-based flow cytometry.
As an example, the FACSCount, as shown in Figure 1.4, is one commercial optical-based flow cytometer system that can provide absolute and percentage counting results of various types of cells, such as red blood cells (RBCs), leukocytes, and CD4 T-lymphocytes. Clinicians rely on this system to diagnose the stage progression of HIV/AIDS, guide the treatment decision for HIV-infected persons, and evaluate the effectiveness of Antiretroviral Therapy (ART) [18]. However, it is only available in laboratory settings due to the limitations of bulky desktop size, prohibitive equipment cost ($27,000), high maintenance and reagent costs ($5-$20), low throughput (30-50 samples/day), and the need for an experienced operator, etc. [19]. Thus, it cannot meet the needs of personalized diagnosis.
Figure 1.4 BD FACSCountT cell cytometer system.
1.1.2.3 DNA Sequencer
DNA sequencing, which detects the nucleotide order in DNA strands, enables the study of metagenomics and genetic disorders for diseases in aging people on an individual basis. Therefore, it plays an important role in personalized diagnostics. The first widely-applied DNA sequencing technique was the Sanger sequencing in the 1970s. This technique employs DNA polymerase to synthesize double-stranded DNA (dsDNA) from a primed single-stranded DNA (ssDNA) template. Four standard deoxyribonucleoside triphosphates (dNTPs), adenine (A), cytosine (C), guanine (G), and thymine (T), are used to extend the DNA, whereas four radioactively labeled di-dNTP (ddNTP) elements (ddATP, ddGTP, ddCTP, and ddTTP) are used to cease DNA extension. That is, once a ddNTP is attached to the DNA template, polymerase synthesis of this strand is invalid and no more dNTPs (or ddNTPs) can be added.
During sequencing, four chambers are employed for DNA synthesis. Each...
