142
829
THS

 

This is to certify that the
thesis entitled

CMOS INSTRUMENTATION FOR ON-CHIP
BIOFLUORESCENCE AND

BIOELECTROCHEMICAL ASSAYS

presented by

WAQAR AHMED QURESHI

has been accepted towards fulfillment
of the requirements for the

MS degree in ELECTRICAL ENGINEERING

 

 

MK“ I
__/l
Major Professor’s Signature

414214174 8; 20/0

Date

 

MSU is an Affinnative Action/Equal Opportunity Employer

 

LIBRARY
Michigan State
University

 

 

 

-.-c—--._.--.— u

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:/Prolecc&PrelelRClDateDue indd

CMOS INSTRUMENTATION FOR ON-CHIP
BIOFLUOROSCENCE AND BIOELECTROCHEMICAL
ASSAYS

By

Waqar Ahmed Qureshi

A THESIS

Submitted to
Michigan State University
In partial fulfillment to the requirements
for the degree of

MASTER OF SCIENCE
Electrical Engineering

2010

ABSTRACT

CMOS INSTRUMENTATION FOR ON-CHIP
BIOFLUOROSCENCE AND BIOELECTROCHEMICAL ASSAYS

By
Waqar Ahmed Qureshi
Membrane proteins embedded in bilayer lipid membranes (BLM) play a vital role
in biological functions. The molecular mechanisms by which these nanostructured
bio-interfaces efficiently perform many critical tasks are poorly understood. Thus,
there is a pressing need for advanced tools to measure structural and functional
properties of bio-interfaces for applications including proteomics, drug screening
and nanosafety. Transduction techniques including electrochemical, optical and
thermal methods are widely used for bio-interface characterization. However,
biological understanding could be greatly improved by a versatile instrumentation
system capable of performing multiple characterization methods simultaneously.
This dissertation addresses challenges to combining optical and electrochemical
detection schemes on a monolithic chip in order to simultaneously monitor both
structural characteristics and functional activities. Several possible architectures for
electrochemical amperometry and optical measurement were analyzed, and key
challenges in integrating these methods were identified. A multi-mode voltammetry
CMOS instrumentation chip tailored to BLM interfaces was developed. For optical
detection, a CMOS imager array suitable for a combined opto-electrochemical
system was developed with a novel on-chip optical filter utilizing CMOS metal
layers. The results of this research lay a solid foundation for the future

implementation of a fully integrated opto-electrochemical detection system.

To my parents and my (late) sister

iii

or m
Hem

0M1 ‘

Elm
on u
Stil-

Wot

IDOL

lint

tlpl

ACKNOWLEDGMENTS

All glory be to the most praiseworthy who showered His countless blessings
on me throughout the span of my life. With the prolific praise of the Owner of
Honor, I desire to begin a limitless praise, with which He is pleased and I cite His
own words; “All praise to Allah, the Lord of the worlds ". The Quran (39:75).

I am honored to have one of the most distinguished scholars, Professor
AndrewJ. Mason, as my advisor. I would like to thank him for his support
throughout my stay at MSU. I found him an extremely intelligent researcher,
exceptionally intellectual advisor and an excellent teacher. This work could not
have been possible without his ever critical attitude towards my neophyte research
ideas, followed by his tremendous encouragement to heighten my confidence.

I am obliged to my family, especially my parents, for their selfless support
throughout my life, and particularly during the span of this work. I am running short
on words to express my sentiments towards my parents for their gracious deeds and
self-sacrificing exertions to empower my advancement. Accomplishment of this
work is an implicit consequence of my Fathers vigorous attitude to boost my
motivation and enthusiasm, and my Mothers affectionate prayers. I am also
thankful to my other family members including my brothers and sister for their
unbounded yet unconditional love, and having absolute faith in me.

Life in the US could not have been easier in the absence of steadfast friends
and wonderful colleagues and they deserve a special mention here. Stating in

alphabetical v order, I found companions like Ahmed, Awais, Brig. Qaiser, Hassan

iv

._ 5F .ld'

 

Aqeel, Jawad, Khawar, Maaz, Sajjad zaidi, Shahzad, Tariq Abdul-Rehman, and
Zubair who kept my morale high and made my stayla superb experience, here at
MSU.

. Advanced Microsystems and Circuits lab (AMSAC), leaded by Prof.
Mason, is a fantastic team exploring the cutting-edge research. I am proud to be a
part of it. I am grateful to Daniel Rairigh, Xioayi Mu, Xiaowen Liu, Li Haitao, Li
Lin, Yue Huang and Yuning for their help and support. It is my pleasure to know
and work with them.

Lastly I want to extend my gratitude to my advisory committee members,
Professor Shantanu Chakrabartty and Professor Michael Shanblatt, for their
uninterrupted patronization during the course of this work. Their extensive
feedbacks not only opened new horizons of research for me, but also endowed me
with insights into intricate engineering concepts. I am fortunate to have their

persistent contribution and involvement in my work.

Table of Contents

List of Tables ................................................................................... ix
List of Figures .................................................................................... x
Chapter 1
Introduction ..................................................................................... 1
1.1- Overview of Biosensors ............................................................... 1
1.1.1- Introduction ............................................................................ 1
1.1.2- Definition ........................................................................... 1
1.1.3- Classes of Biosensors .............................................................. 2
1.1.4- Advantages ........................................................................... 4
1.1.5- Applications ........................................................................ 5
1.2- Membrane protein-based biosensors ................................................ 6
1.2.1- Motivation .......................................................................... 6
1.2.2- Advantages of Microsystems .................................................... 7
1.2.3- Challenges .......................................................................... 8
1.3- Goals ................................................................................... 11
1.4- Thesis Outline ......................................................................... 13
Chapter 2
Review of On-chip Biosensor Technologies .............................................. 15
2.1- Signaling Mechanisms ............................................................... 15
2.1.1- Luminescence .................................................................... 15
2.1.2- Fluorescence ..................................................................... 16
2.1.3- Electrochemical detection ...................................................... 18
2.1.3.1- Voltammetry ............................................................... 19
2.1.3.2- Impedance spectroscopy ................................................. 20
2.2- Approaches for on-chip fluorescence detection .................................. 21
2.2.1- Fluorescence detection using CCD image sensors .......................... 22
2.2.2- Fluorescence detection using CMOS image sensors ........................ 22
2.2.3- Optical filters ..................................................................... 23
2.2.4- Filterless techniques .............................................................. 26
225- Analysis and Discussion ......................................................... 28
2.3- Approaches for on-chip electrochemical detection .............................. 29
2.3.1- Voltammetry ..................................................................... 29
2.3.2- Impedance spectroscopy ........................................................ 30
2.3.3- Analysis and Discussion ........................................................ 30
2.4- Integration of optical and electrochemical detection ............................ 31

vi

Chapter 3

Integration of Optical and Electrochemical detection circuits ........................... 33
3.1- Motivation ............................................................................. 33
3.2- Background ........................................................................... 33
3.3- Challenges for opto-electrochemical detection ................................... 35

Chapter 4

CMOS microsystem for multimode voltammetry ........................................ 39
4.]- Chip Architecture ..................................................................... 39
4.2- Programmable Waveform Generator .............................................. 40

4.2.1- Control Block ..................................................................... 42
4.2.2- Bidirectional Counter ............................................................ 42
4.2.3- Digital-to-Analog Converter ................................................... 44
4.2.4- Operation of Waveform Generator ............................................ 45
4.3- Single-ended potentiostat ............................................................ 46
4.4- Amperometric readout circuit ...................................................... 48
4.5- Test Setup .............................................................................. 49

Chapter 5

CMOS instrumentation system for optical detection ..................................... 51
5.1- Chip Architecture ..................................................................... 52
5.2- Pixel Architecture .................................................................... 55

5.2.1- Photodiode ........................................................................ 56
5.2.2- Front-end readout electronics .................................................. 57
5.3- On-chip Filter .......................................................................... 61
5.3.1- Introduction ....................................................................... 61
5.3.2- Design methodology ............................................................ 61
5.3.3- Design constraints ............................................................... 64
5.3.3.1- Design and fabrication rules ............................................. 64
5.3.3.2- Filter transmission factor ................................................ 66
5.3.3.3- Process variations ......................................................... 67
5.3.4- Effect of feature size ............................................................ 60
5.4- Physical design ........................................................................ 68
5.5- Summary ............................................................................... 69

Chapter 6 v

Simulation and Experimental Results ...................................................... 71
6.1- Measurement of multi-mode voltammetry CMOS microsystem .............. 71

6.1 .1- Waveform generator ............................................................ 71
6.1.1.1- Test setup .................................................................. 71
6.1.1.2- Test results ................................................................. 72

6.1.2- Potentiostat ........................................................................ 74
6.1.2.1- Test Setup ................................................................... 74
6.1.2.2- Test results ................................................................. 75

6.1.3- Amperometric readout circuit .................................................. 77

6.1.4- Discussion ........................................................................ 78

vii

6.2- Optical detection system ............................................................ 78

6.2.1- Simulation results ................................................................ 80
6.3- On-chip filter .......................................................................... 81
6.3.1- Simulation results ................................................................ 81
6.3.2- Effect of extra metal layers on the selectivity of filter ...................... 84
6.3.3- Process variations ................................................................ 86
6.3.4- Discussion ........................................................................ 88
6.4- Conclusion ............................................................................ 90
Chapter 7
Summary and Future Work .................................................................. 91
7.1- Summary of the contributions ....................................................... 91
7.1.1- Analysis of challenges for opto-electrochemical detection system. . .....91
7.1.2- CMOS microsystem for multimode voltammetry ........................... 92
7.1.3- Optical detection circuit ......................................................... 92
7.2- Future Work ........................................................................... 93
Bibliography ................................................................................... 94

viii

List of Tables

Table I: Calculated diffraction angle for excitation and emission light from
fluorophore and M2 layer width and separation between M2 stripes for CSN
process .......................................................................................... 65

Table H: Diffraction angle for excitation and emission light from fluorophore for
IBM 0.13um CMRFSSF process ........................................................... 68

Table III: Waveform generator characteristics ............................................ 73

Table IV: Types and range of metal layers in CSN process used for MATLAB
simulation ...................................................................................... 83

Table V: Types and range of metal layers used in IBM 0.13um process for
MATLAB
simulation ....................................................................................... 85

Table VI: Different cases for process variations in metal widths for C5N
process .......................................................................................... 87

ix

List of Figures

Figure 1.1: Main components of biosensor ................................................. 2
Figure 1.2: Sensitivity of silicon at different wavelengths ............................... 11

Figure 1.3: Conceptual block diagram of a CMOS opto-electrochemical detection
system .................................................................................. 12

Figure 2.1: Absorption and emission of energy due to change in energy

level .................................................................................... 16
Figure 2.2: Fluorescence occurrence due to change in states ........................... 17
Figure 2.3: Absorption and emission phenomena in fluorescence ..................... 18
Figure 2.4: 3-electrode electrochemical system using cyclic voltammetry ........... 20

Figure 4.1: Block diagram of the autonomous amperometric instrumentation

system .................................................................................. 40
Figure 4.2: Block diagram of the multi-mode waveform generator .................... 43
Figure 4.3: Functional diagram of waveform generator block .......................... 43
Figure 4.4: Schematic of the bidirectional counter ....................................... 44
Figure 4.5: Signal flow from counter to DAC ............................................. 44
Figure 4.6: Block diagram of hybrid digital to DAC used in waveform

generator .............................................................................. 45
Figure 4.7: Input and output waveforms of waveform generator ....................... 46

Figure 4.8: Single-ended potentiostat designed for the instrumentation

system ................................................................................. 47
Figure 4.9: Amperometric readout circuit ................................................. 49
Figure 4.10: Test setup for the instrumentation chip with electrode array ............ 50
Figure 5.1: Block diagram of optical detection system .................................. 52

Figure 5.2: Signal flow path from emitted photons to the output ...................... 52

Figure 5.3: Block diagram of the CMOS optical detection chip ........................ 54
Figure 5.4: Cadence layout of pixel including photodiode and readout circuit

components ........................................................................... 56
Figure 5.5(a): Schematic diagram of the photodiode structure ......................... 57
Figure 5.5(b): Cadence layout of n-well/p-sub 150x200pm2 photodiode. . . . . . ..57
Figure 5.6: Sigma delta (2A) ADC used for readout ..................................... 59
Figure 5.7; Schematic of the folded cascode OTA ....................................... 60
Figure 5.8: Schematic of the voltage comparator ......................................... 60

Figure 5.9: Conceptual block diagram of a CMOS optical interference

filter .................................................................................... 63
Figure 5.10: Cadence layout of the l.5x1.5mm2 optical detection chip ............... 69
Figure 6.1: Die photograph of 3x3mm2 multi-mode voltammetry system. . . . . . . . ....72
Figure 6.2: Experimental test setup for the on-chip waveform generator ............. 72
Figure 6.3: Simulation results for the waveform generator .............................. 73
Figure 6.4: Experimental test setup for on-chip potentiostat ............................ 74

Figure 6.5: CV measurement of 0.1M potassium ferrocyanide (K4[Fe(CN)6]-3H20)
for on-chip electrode ................................................................ 76

Figure 6.6: labview test program interface to test the functionality of amperometric

readout circuit ........................................................................ 77
Figure 6.7: Current readout circuit results for input current from l-9uA. . . . ...78
Figure 6.8: Cadence layout of CMOS optical detection IC .............................. 79
Figure 6.9: Die photograph of CMOS optical detection IC .............................. 79

Figure 6.10: Cadence simulation of a single pixel for 70nA with integrator analog
out and digital output from the comparator ...................................... 80

xi

Figure 6.11: Calculated and simulated time for the discharge of photocurrent using
single pixel sigma-delta ADC ..................................................... 81

Figure 6.12: Physical model used for optical simulation in MATLAB ................ 82

Figure 6.13: Locations of incident light of wavelength on n-well with different
colors showing the light coming from adjacent slits ........................... 83

Figure 6.14: Range of incident light wavelength allowed to pass through the filter
shown by a digital signal ............................................................ 84

Figure 6.15: Locations of incident light of wavelength on n-well with different
colors showing the light coming from adjacent slits ........................... 86

Figure 6.16: Range of incident light wavelength allowed to pass through the filter
shown by a digital signal ............................................................ 86

Images in thesis are presented in color.

xii

Chapter 1
Introduction

1.1 Overview of Biosensors
1.1.] Introduction

Biosensors play an important role in our daily lives through healthcare,
environmental analysis and industrial process controls. Recent advances in sensor
technologies and microelectronics have resulted in a rapid growth in the development
of new biomaterials such as conducting polymers, copolymers and sol gels, and many
improvements in sensing techniques resulting in miniaturized and portable
instruments have been reported. The potential growth in the biosensor industry is
significant and is becoming one of the fastest growing sectors in the world. New
analysis from Frost & Sullivan, World Biosensors Markets, reveals that the biosensor
market is expected to grow from $6.72 billion in 2009 to $14.42 billion in 2016. The
main reason for this fast growth is the breadth and depth at which biosensor
technology has been increasingly applied to both medical and non-medical
applications. A major part of biosensor revenue is due to applications in health care
and clinical diagnostics, especially in “glucose testing”.
1.1.2 Definitions

A biosensor is commonly defined as an analytical device that uses a
biological recognition system to identify molecules and macromolecules. Typically
biosensors are comprised of three components, as shown in Figure.1: (1) the
detector, which is a biological sensing element that gives a response to only the

specified target analyte, (2) the transducer, which converts the detector response

  

I> . T " :f”;}’:,~._, ,

. I ; , . ' -. Output
.>. > it ...4; m“..- Si nal
. .I > g

D. . :>
p D

O I b

 

Figure 1.1 Main components of biosensor.

into a useful (generally optical or electrical) output signal; and (3) the output
system, which involves amplification,
filtering, signal processing, and display of the output in an appropriate format. The
detector typically may be an enzyme, an antibody, a receptor or a cell. The
performance of biosensors is characterized by its speed, sensitivity, accuracy and
portability.
1.1.3 Classes of Biosensors

Biosensors can be classified either by the type of biological sensing element
they utilize or by the method employed for detection and transduction [1]. Common
biological sensing elements and transduction methods are summarized below.
1.1.3.1 Biological sensing elements
(a) Antibody/antigen (Ab/Ag): The non-covalent antibody-antigen interaction is
similar to a “lock” and “key” effect between an enzyme and a substrate. This exact
and specific interaction is used for detection, identification and measurement of
target molecules[2, 3]. This interaction can be detected either through fluorescent
labeling or by observing a refractive index or reflectivity change [4].
(h) Enzymes: Enzyme-based biosensors operate either by producing or

consuming a protein and/or some electroactive species. Based on their binding

capabilities and catalytic activity, enzymes are used for specific detection and
measurement, which can be done directly or in conjunction with an indicator [2, 4].
(c) Nucleic acids: The complementary relationships between adenosine and
thymine and cytosine and guanosine in DNA are used in nucleic acid-based
biosensors to identify small concentrations of DNA in large samples. This
identification is done by comparing a DNA sample with DNA of known organism
also called DNA probe [5].
(d) Cells and viruses: Microorganisms such as bacteria and fungi are used as
biosensors to detect specific molecules or the overall “state” of the surrounding
environment and have been used in the development of biosensors and biochips [2].
Biosensors based on these engineered bacteria are used to detect stress conditions,
toxicity or DNA damaging agents and other specific organic and non-organic
compounds [6]. Also cellular biosensors are used for testing and monitoring the
effectiveness of drugs and therapies [7].
(e) Biomimetic materials based: A biomimetic biosensor is an artificial or
synthetic sensor that is fabricated and designed to mimic the function of a natural
biosensor[2]. Several different technologies including genetic engineering of
molecules, artificial cell membrane fabrication, and molecular imprinting have been
developed to construct these biomimetic receptors [8-12].
1.1.3.2 Methods of Transduction

Based on the method of signal transduction and detection biosensors can be

classified into following categories [2, 13].

(a) Optical Detection: Optical detection biosensors are the most diverse class of
biosensors and can be used for many different types of spectroscopy, such as
absorption, fluorescence, chemiluminescence, RAMAN, SERS, refraction and
dispersion spectrometry [14-18]. These techniques are usually used to measure
different properties, such as energy, amplitude, polarization, phase and detection
time.
(b) Electrochemical Detection: Electrochemical detection provides another possible
means of transduction [19-29] and electrochemical biosensors are designed to
measure the current produced from reduction-oxidation (redox) reactions. The
resulting current produced is directly correlated to the bulk concentration of
electroactive species present or its production/consumption rate [5].
(c) Mass-sensitive: These biosensors are used to detect small changes in the mass
of and interface due to chemical binding of target analytes. The mass changes are
commonly measured indirectly through a change in oscillation frequency of a
resonating crystal [5].
((1) Thermal Detection: Thermal biosensors measure the changes in temperature in
the reaction between an enzyme molecule and a suitable analyte. These changes in
temperature translate to the amount of reactants consumed or products formed [30].
1.1.4 Advantages

Biosensors promise significant advantages over non-biological sensing
technologies. First, biosensors exhibit impressive sensitivity because biomolecules
often possess high affinity toward their targets [31]. For example, antibodies

capture antigens with a dissociation constant at the nanomolar scale, and DNA-

DNA interactions are even stronger than those of antigen-antibody [32]. Secondly,
biological recognition is usually very selective because most biosensors only react
in the presence of the target analyte. Moreover, due to advances in modern
electronics especially CMOS technology, it has become possible to develop
inexpensive, integrated, portable and ready-to-use miniaturized biosensor devices.
Miniaturized biological sensors improve the ability to detect pathogens or perform
genetic analysis in hospitals; more importantly, they are particularly useful for
small clinics and point-of care analysis.

In today’s world of rapid information access, miniaturized biosensors can
serve exceptionally well in emergency situations or for on-site field applications.
Typically, the smaller the device, the faster and more sensitive the response is. In
most cases, biosensors can easily detect analytes in the rnicromolar to picomolar
range [33-37].

Ease-of-use and reduced labor requirements provide additional advantages
of miniaturized biosensors as compared to other traditional techniques such as
atomic absorption spectrometry, inductively coupled atomic electron spectrometry,
and sequential extraction procedure. The results obtained from miniaturized
biosensors are compatible with and comparable to chemical analysis, while being
free of chemical extractions and analytical procedures [38].

1.1.5 Applications

There is an increasing demand for inexpensive and reliable sensors to allow

not only routine monitoring in a central or satellite laboratory but also analysis with

greater patient contact, such as in a hospital ward, emergency room, or operating

room. Biosensors play a major role in our daily lives in a wide number of
applications including healthcare and biomedical diagnostics, environmental
monitoring, industrial process control, and military purposes. In healthcare,
biosensors are used to measure metabolites, ions, blood oxygen, and many other
parameters; to monitor patients during intensive care and surgery; and to investigate
genetic disorders [39, 40]. In environmental monitoring biosensors monitor the
efficacy of physical, chemical or biological pollution and measure toxic chemicals
and analytes resulting from biochemical and chemical transformations [41].
Biosensors are also used in industrial process control to improve product quality,
increase product yield, improve plant performance, reduce raw materials and
optimize energy. Biosensors are used in the creation of bio-fuel alternative energy
sources. They are also critical for detection and warning of chemical or biological
agents in warfare and civilian security.
1.2 Membrane-protein-based biosensors
1.2.1 Motivation

Proteins embedded in bilayer lipid membranes (BLM) play a vital role in
biological functions and are highly useful because of their unique roles in cellular
function and drug discovery [42]. These nanostructured membrane proteins are used
for the transport of substances across membranes, enzymatic activity, intracellular
joining, recognition of other cells and signal transduction. Physical and chemical
properties of membrane proteins need to be better understood because they dictate
the molecular mechanisms by which membranes perform their cellular functions.

To study these mechanisms, new biosensors are needed to characterize the multi-

protein membrane complexes that are regulated by a wide range of interacting
partners. Such protein array biosensors are highly desired for preclinical,
toxicological and clinical studies in order to ascertain candidates for drug screening,
monitor efficacy and toxicity, and identify diagnostic markers. Despite the critical
need to monitor membrane protein activities, the existing tools and techniques are
too time consuming, labor intensive and expensive to permit wide-spread
application and high-throughput operation.

Several labeled and label-free detection techniques are currently used to
measure the multiple activities and characteristics of membrane proteins. Some of
these are measured by optical detection schemes e.g. fluorescence,
bioluminescence. Some are measured by electrochemical detection schemes, e.g.
voltammetry and impedance spectroscopy, which have the advantage of direct
transduction to electrical data, being less prone to interference, and the potential to
miniaturize the entire measurement system [43]. An instrumentation system to
measure the activities and characteristics of membrane proteins must include
multiple detection schemes and ideally incorporate both optical and electrochemical
methods.

1.2.2 Advantages of Microsystem

Sensitivity and portability are among the most demanding requirements for
miniaturized biosensors. The rapid development of microelectronics technologies
has created many opportunities to develop highly integrated, on-chip, electrical
biosensors with many advantages in terms of sensitivity, quantification and

accuracy. Many semiconductor devices and fabrication processes exhibit high

compatibility with biological materials that has enabled the expanding use of
microelectronic and MEMS devices to integrate a large number of biological
recognition elements within a single package to form integrated microsystems. The
high level of integration results in enhancing the functionality and provides
substantial parallelism that enables high throughput. On-chip biosensors facilitate
miniaturized, low power, hand-held and portable instrumentation systems that are
vital for applications such as point-of care diagnostics and clinical applications.
Moreover, they provide faster analysis and lower response times because of the
short distance between the sensor and the target and higher surface to volume ratios.
Because of their small volume, they need lesser amounts of reagents, resulting in
low cost and a compact structure. In addition, miniaturized on-chip biosensors have
very low manufacturing costs when fabricated in mass production.
1.2.3 Challenges

The development of new miniaturized biosensors utilizing membrane
protein interfaces and microsystem structures faces several challenges in design,
integration and fabrication. These challenges must be overcome to fulfill the
requirements of specific applications, especially for high density biosensor arrays.
High density arrays require routing of hundreds of raw signals off the array chip,
introducing bandwidth challenges as well as performance-limiting noise. Similarly,
measurement speed has an impact on the performance of the system because the
microsystem should be able to record all important data [44]. Moreover, the

minimum detectable signal is an important parameter that is affected by

biochemical noise (shot noise of binding/reaction events), interference of
background molecules and readout circuit performance [45].

Different applications favor different interrogation techniques and analysis
methods, so there is a growing need to design diverse and versatile systems that are
capable of performing multiple interrogation techniques to support a range of
analysis suitable for many applications. For example, combining optical and
electrochemical detection schemes in a single CMOS IC would allow simultaneous
optical and electrochemical analysis of on-chip biosensors. This is highly desirable
because it would combine the advantages of both these techniques and result in a
high throughput system with added functionality, improved accuracy and
minimized cost. However, realization of such a microsystem is challenging in terms
of electrical design, integration of techniques, fabrication and packaging. Both
optical and electrochemical detection system needs to be developed and integrated
in an efficient way that permits simultaneous optical and electrochemical detection.
To realize a miniaturized and multimode electrochemical detection system, it is
desirable to realize the following techniques within the readout IC: cyclic
voltammetry (CV), linear sweep voltammetry (LSV), constant potential and normal
pulse voltammetry (NPV). These techniques require different input waveforms, so a
programmable waveform generator is needed to generate these waveforms. Namely,
a triangular, linear ramp, and constant potential generator are needed for the above-
mentioned techniques to increase the throughput of the system. Although industry-
standard CMOS fabrication processes provide low manufacturing costs, a high level

of integration and significant design flexibility, they introduce several important

limitations. For example, another significant challenge in on-chip electrochemical
transducers is the lack of a good electrode material in the standard CMOS process.
In CMOS ICs, only the top metal layer (aluminum) is exposed to the surface, and
aluminum is not as useful with potential electrolytes as gold, platinum, titanium or
Ag/AgCl [25, 46]. Thus, post-CMOS fabrication processes are required to protect
(insulate) the CMOS top-layer metal and create useful electrodes on the surface of
the chip where direct exposure of electrolytes will occur.

For optical detection, CMOS is suitable for visible range photons (320-
780nm), although photons in the UV (100nm-350nm) and IR (700nm-300pm)
range cannot be effectively detected using CMOS photodiodes because of silicon
material characteristics as shown in Figure 1.2 [47]. Proper choice and design of
photodiode with efficient readout is needed to increase its responsivity and quantum
efficiency for a specific application. Also, to obtain the desirable optical
information, filters are required to be incorporated within the optical detection
system. For example, in fluorescence detection, only the emitted light from a
fluorophore is required to be processed while the excitation light should be blocked
using a filter. Typically, post-CMOS processing is carried out to incorporate these
filters in the microsystem [48-50], increasing the cost and also increasing the risk of
effecting other components and hence functionality of the system. Furthermore, for
a combined opto-electrochemical system with on-chip biointerfaces, incorporating
filter using post processing is not feasible as it hinders electrochemical detection by
separating the electrode from the on-chip readout circuitry. Thus, for optical

detection either a filterless approach is required or a new filter design is needed that

10

1.0

 

      
  
     
 

  

0.9 - ------ i ------- *3 ------- J; -------- i -------- i --------
0.8 — ....... 4; ....... J ....... .5 ..... 7
r— ' : ' /
E 0.7 '" """ i """" T """"" a: Rph,m23(,¢,€: ------
: : : t’ :
__. 0.6 _./ ........ I ........................
: : ,t’ : :
.9; 0.5 — ------ e ------- 1, /-----: ------- 4 -------- s --------
> :,/ : : : :
E 0.4 "‘3“ ....... 4 ------- 4 ------- -: -------- : --------
a) I E 5 l 5
c _ .....................
(D 0.3 t 5
CD 0.2 ..... g ..................... ' ........
0.1 ....... ‘ ............................
i L n

 

 

 

0.4 0.5 0.6 0.7 0.8 0.9 1
Wavelength [um]

Figure 1.2. Sensitivity of silicon at different wavelengths [47].

can be incorporated in the combined system without the need for post-CMOS
processing to allow for simultaneous optical and electrochemical detection.
1.4 Goals

The operation of membrane protein sensors requires the presence of a fluidic
environment. Optical detection shows high compatibility with microfluidics and
does not require physical interconnection between detector unit and the microfluidic
device, resulting in minimal disturbance to reaction system; however, at the same
time it suffers from relatively poor detection limits [48]. Electrochemical detection
has the advantage of excellent sensitivity with the realization of on-chip electrodes.
A combined opto-electrochemical detection system would combine the advantages
of both these techniques resulting in an efficient, versatile and high throughput

detection system that can be used for a variety of applications.

11

This research is based on establishing the design of a combined opto-
electrochemical detection system for membrane-based protein biosensor array that
simultaneously implements both optical and electrochemical detection by first
analyzing the challenges to implementing such a system and, secondly, developing
a solution for some of the resulting electronics design challenges. A block diagram
of the CMOS opto-electrochemical detection system concept is shown in Figure
1.3. The CMOS instrumentation chip assembly is mounted on a biomembrane array
module daughterboard that is plugged into a receiver module motherboard for
connection to a standard computer. The receiver module has an open window over
the biomembrane array area permitting access to an optical excitation source, a top-
side reference electrode and top-side fluid delivery/removal. A data acquisition card
(DAQ) within the computer provides programmable control of stimulus,

measurement, and fluidic operations.

   
  
 
  

LIGHT fluid
EXCITA‘NON Ole/”90’

§ v window microf/uidics

 

 

 

CMOS chip

dielectric w/ array site reservoirs
electrode/biomembrane site

Figure 1.3 Conceptual block diagram of a CMOS opto-electrochemical detection
system

12

To meet the requirements of a combined opto-electrochemical detection
system that can be used for a diverse set of biosensors, one goal of this research is
to design a multi-mode voltammetry instrumentation system that includes a
programmable waveform generator to generate different input waveforms i.e.
triangular, linear ramp and constant potential with programmable frequency and
amplitude. These waveforms are needed to perform a variety of voltammetry
applications including cyclic voltammetry, linear sweep voltammetry, constant
potential and pulse voltammetry.

Another goal of this research is the design of a highly efficient optical
detection system that meets the requirements of an opto-electrochemical detection
system. To avoid the necessity of an optical filter using post-CMOS processing, the
final goal of this research is to explore an on-chip filter that relies on a novel
technique using the metal layers inherently present in the standard CMOS process.
Fabrication of miniaturized on-chip electrodes and development of miniaturized
biosensors and bio-interfaces on a CMOS IC for a combined opto-electrochemical
detection system are beyond the scope of this research. These challenges are being
addressed by other fellow students in the Advanced Micro-Systems and Circuits lab
at Michigan State University.

1.4 Thesis Outline

Chapter 2 reviews optical and electrochemical detection methods relevant to
the goals of this thesis research. Chapter 3 outlines the issues and challenges
regarding the design of a combined monolithic opto-electrochemical detection

system along with the background and importance of CMOS opto-electrochemical

l3

detection systems. Chapter 4 discusses the design and VLSI realization of the
proposed electrochemical detection system for different voltammetry applications.
Chapter 5 describes the proposed system for optical detection using fluorescence
and a CMOS interference filter, with an in-depth view of the circuit-level
implementation of the proposed technique. Chapter 6 presents and discusses the
simulation results and experimental results for several VLSI circuits developed for
the proposed fluorescence and electrochemical detection system for membrane
protein assays. Chapter 7 provides the summary of the completed research work and

gives some suggestions for future work in this area.

14

Chapter 2
Review of On-chip Biosensor Technologies

2.1 Signaling Mechanisms

Because membrane proteins play a vital role in biological functions and
drug discovery, an ability to characterize their activities and incorporate them into
sensors is very important. Membrane-protein-based biosensors are used to detect
drugs, neurotransmitters, hormones, toxins, and inhibitors such as amiloride [42].
Characterization of membrane proteins is done by using both labeled and label-free
detection schemes, each with their own advantages. Label-free detection schemes
provide direct detection of analytes without requiring labels, marker compounds
added incorporated within a test solution. It generally uses a transducer to measure
some physical property or characteristics of a biochemical target, i.e. DNA
molecule, protein, virus, or cell. Electrochemical detection is a widely used
continuous label-free detection scheme for characterizing membrane proteins. On
the other hand, labeled detection methods require modifying the target molecules
with labeling compounds. Optical detection methods can use both labeled and label-
free detection schemes for characterizing membrane proteins, e.g.
chemiluminescence and fluorescence. Fluorescence detection is one of the most
widely used labeled optical detection schemes to characterize the membrane
proteins.
2.1.1 Luminescence

Luminescence is defined as a process by which light is produced my means
other than heating or incandescence. This light can be emitted at normal or lower

temperatures and is also named as “cold light”. This phenomenon occurs due to a

15

change in the energy level of an atom, an ion or a molecule. For example, by
absorbing energy from an excitation source, an electron of an atom is driven from
its lower energy level or ground state to a higher energy level, as shown in Figure
2.1(a). This higher energy state, which is unstable, is called an excited state, and
whenever the electron comes back to its lower energy level then a photon of light
may be emitted. The photon energy corresponds to the difference between energy
levels, shown in Figure 2.1(b). Usually the emitted light has less energy as
compared to the energy of excitation light. Based on the source of excitation light,
there are several different types of luminescence: fluorescence, chemiluminescence,
optically simulated luminescence, cathodoluminescence, radio luminescence
phosphorescence, electroluminescence, triboluminescence and therrno

luminescence.

LLLPhoton

Photon
W

     

Ground State Excited State Excited State Ground State
(a) (b)

Figure 2.1. Absorption and emission of energy due to change in energy level.
2.1.2 Fluorescence
According to the Columbia Encyclopedia; fluorescence is defined as the
“luminescence in which light of a visible color is emitted from a fluorophore under
stimulation or excitation by light or other forms of electromagnetic radiation or by

certain other means”. This emitted light or radiation persists as long as the

16

stimulating light continues, and it has a longer wavelength as compared to the
excitation light. This process of changing energy states or levels is shown in Figure
2.2. Some of the important characteristics of fluorescence are as follows.

(a) Stokes Fluorescence: Fluorescence can be defined as the re-emission of longer
wavelength photons by a molecule that has absorbed photons of shorter
wavelengths. In other words, emitted light has less energy than the excitation light

or absorption light because of loss of vibrational energy in the excited state [49].
Thus, the fluorescence occurs at longer wavelengths i.e. hVF< hVA as shown in

Figure 2.3. This change is wavelength is typically referred to as the Stokes Shift.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Vibrational
5 / relaxation
/
4
i A l
V = O A ll 1
3; Absorption Fluorescence
'- ‘ hVF hVA
2
I.”
4
3
2 ’ Ground State
1
V = 0

Wavelength——>
Figure 2.2. Fluorescence occurrence due to change in states.

17

Stokes shift
l-—-l

  
   

absorption emission

Intensity

 

 

Wavelength —>
Figure 2.3. Absorption and emission phenomena in fluorescence.

(b) Lifetime: One of the important characteristics of fluorescence is its lifetime. The
lifetime of the excited state is determined by the average time the molecule spends
in the excited state until it returns back to the ground state or lowest energy level.
Fluorescence is considered to be a very short lifetime phenomena, on the orders of
nanoseconds. This lifetime can change to some extent with changes in the
fluorophore environment.

(0) Quantum Yield: The fluorescence quantum yield, CD, is the ratio between the
number of fluorescence photons emitted and the number of photons absorbed by a

detector.

_ No of photons errutted (2.1)

 

— N o of photons absorbed

2.1.3 Electrochemical detection

Because of its high sensitivity and excellent selectivity, electrochemical
detection is used in many biomedical and environmental applications. Output
depends on the nature of the current flow, i.e. oxidation or reduction current due to

the input voltage applied and properties of the electrode. The electrochemical

18

detector typically contains three electrodes: a working electrode where the chemical
reaction takes place, counter electrode that supplies current, and reference electrode
which acts as zero potential for the working electrode. Because of its highly stable
characteristics, Ag/AgCl is the most widely used reference electrode. The counter
(or auxiliary) electrode is usually made of an inert material such as noble metal or
graphite.

A potentiostat is the basic electronic element in electrochemical detection
schemes. It is used as a biasing circuit and provides an electronic interface to the
detector. It controls the potential between the working and reference electrodes
while a current readout circuit measures the current between the working and
counter electrodes. Different approaches for both single-ended and differential
CMOS potentiostats with a variety of readout circuits have been reported [50-60].
Two of the most widely used electrochemical detection techniques using these
potentiostats are voltametry and impedance spectroscopy.
2.1.3.1 Voltammetry

Voltammetry is a widely used electrochemical technique in which different
types of voltage waveforms are applied at the working electrode and response
current is measured. Based on the input waveform, the different voltammetric
techniques used include cyclic voltametry (CV), linear sweep voltametry (LSV),
constant potential or DC voltametry and pulse voltametry. Out of these techniques,
CV is the most effective and widely used technique to study the formal redox
processes for many applications. In CV, a triangular waveform with a specific rate

and amplitude, referred to as scan rate and scan range, respectively, is applied at the

19

WE (w.r.t the RE) and the response current is measured between WE and CE, as
shown in Figure 2.4. In linear sweep voltametry, a linear ramp is applied at the WE,
and in the case of constant potential or dc voltametry, a constant potential is applied

and response current is measured.

L\
/otential Current
Controlled flow
A A

 

 

 

 

 

 

 

 

/\/\-—-)

Input
519%“ RE WE CE

 

 

 

 

Figure 2.4. 3-electrode electrochemical system using cyclic voltammetry.

2.1.3.2 Impedance Spectroscopy

lrnpedance spectroscopy (IS) is another widely used non-destructive
electrochemical detection technique in which the relationship between the stimulus
potential, response current, and frequency of stimulus is measured to characterize
the impedance. In IS, an AC potential is applied to an electrochemical cell and the
response current containing both real and imaginary components is measured for
each frequency of input stimulus. Sometimes these measurements are made in the

time domain and then fourier transformed into the frequency domain.

20

Electrochemical impedance spectroscopy (EIS) involves measurement and
analysis of materials in which ionic conduction strongly predominates. This
technique can be used to analyze different processes including electronic/ionic
conduction in the electrode and electrolytes, interfacial charging either at the
surface film or the double-layer, charge transfer processes and the mass transfer
effects, if any [61]. As these parameters have different responses at different
frequencies, their features will also show at different frequencies of the impedance
spectrum.

2.2 Approaches for On-chip Fluorescence Detection

Of the reported optical detection methods, fluorescence has emerged as the
most attractive detection technique for biomedical applications due to its excellent
sensitivity and specificity as well as its low cost and compatibility with CMOS
technology. Fluorescence detection can be used for different applications in
biophysics, biochemistry and medical diagnostics [62-66] . The basic components
of an on-chip instrumentation system for fluorescence detection include an
excitation source as a source of light of a specific wavelength, a photo detector to
detect light, and a transducer to convert the optical response signal into voltage so it
can be interfaced with a readout circuit to measure the intensity that correlates to
concentration. In a biosensor with optical detection, photodiodes are typically used
to transform biological information into electrical signals.

Different approaches have been reported for fluorescence detection and
imaging including various types of fluorophores with different configurations of

photodiodes. The fluorophores reported for fluorescence detection are derivatives of

21

Rho

E‘s—J

I“ a)

1311
ha

ha‘

16;

3p:

Rhodarnine, Fluoroscein, Coumarin, Europium and Cyanine [67-71], as best suits
each application. The two primary types of sensors for imaging applications are
charged coupled devices (CCD) and complementary metal oxide semiconductors
(CMOS) imagers. Both CCD and CMOS image sensors can be manufactured in
silicon technology foundries using standard materials and equipment. Commonly,
fluorescence detection utilizes CCD detectors and external signal processing
circuitry. However, because of fabrication differences, CCD detectors cannot be
integrated on the same chip as normal integrated circuits (IC) used for signal
processing. In contrast, CMOS-based optical detectors can detect fluorescent
emission within the same IC as signal processing circuitry [72]. Additional
advantages of using CMOS-based detectors include ease of design, low cost, low
power consumption and small-size detection circuits [73].
2.2.1 Fluorescence detection using C CD image sensors

Because of better image quality and better performance in terms of dynamic
range and noise, a variety of on-chip and off-chip fluorescent detection systems
have been developed using CCD image sensors [74-76]. These detection systems
have been used for different applications including pathogen detection and
myoglobin detection by immobilizing antibodies on the surface of immunosensors.
2.2.2 Fluorescence detection using CMOS image sensors

Fluorescence detection using CMOS image sensors have widely been
reported using different types of photodiodes and readout topologies tailored to
specific applications. The main advantages of CMOS photodiodes over CCDs

include low cost, high speed, and high responsivity. CMOS image sensors also offer

22

superior integration capabilities with lower power consumption at the expense of
image quality.

For fluorescence measurements using CMOS imagers, different photodiodes
are available to sense light intensity which is measured using on-chip
microelectronic circuits. Light intensities as low as ~105 photons/cm2 have been
reported using advanced circuit topologies [77-79]. The different photodiodes that
have been used include p-n, PIN and avalanche photodiodes [80-82]. Similarly
different readout topologies including differential photodiode (PD) [83], pseudo-
differential active pixel sensor(APS), 3T APS and 4T APS [47, 78] have also been
reported to measure low light intensities. Although very large pixel array using
CMOS photodiodes have been reported [84-86], fluorescence applications often
emphasize quantification of light rather than pixel density. Thus, moderate sized
arrays that permit greater area per pixel can be chosen to loosening design
constraints.

2.2.3 Optical filters

Optical filters are commonly used to selectively block excitation frequencies
while transmitting the fluorescent emission signal. Different materials and
processing steps have been reported to incorporate optical filters into fluorescence
detection systems. In [87] a fluorescence detection microsystem has been proposed
by integrating a Cadmium sulfide(CdS) thin-film filter, an (In,Ga)N thin-film blue
LED and a disposable PDMS microfluidic device onto a Si PIN photodetector

substrate. The process includes deposition of a l-3um thin-film CDS filter on a

stacked ITO/Si02 layer using pulsed layer deposition (PLD). SiOz layer with a

23

thickness of 1500A° was deposited by PECVD on silicon photodiode substrate,
while ITO film with a thickness of 2000A° was deposited using sputtering
mechanism. Deposable microchannels were fabricated on a PDMS mold that was
prepared using deep reactive ion etching (DRIE). Although this microsystem was
able to detect fluorescence and also CdS film filtered the emitted signal but it has
many limitations in terms of performance as an appreciable amount of light is
reflected from the backside of microfluidic device. Also for better performance it
needs addition of antireflection coating to the CdS film to suppress the reflection of
emission signal. Moreover larger PIN photodetector area is required to increase the
collection of emitted signal. These factors severely affect the performance and cost
of this detection system as it needs extra processing steps to achieve good
performance. Similarly in [88] a monolithic capillary electrophoresis system with
an integrated on-chip fluorescence detection system has been proposed which
includes a Si photodiode combined with an on-chip multilayer interference filter to
prevent excitation light from hindering the fluorescence detection. This 3pm filter
consists of 20 alternating layers of Si02 and Ti02 at nearly quarter wavelength
thicknesses. Although this technique integrates photodiode with the filter on a
single substrate, it requires extra processing steps that add cost and complexity and
are not available through a standard CMOS foundry. In [89] a fluorescence
detection system has been proposed in which a cross polarization scheme is being

used to separate excitation light form emission spectrum. A bi-layer organic
photodiode (OPD) consisting of duplicate layers of CuPC and C60 with ITO anode

and aluminum cathode increasing the photodiode quantum efficiency. Similarly a

24

monolithically-integrated sensor on a microfluidic platform is presented integrating
a vertical-cavity surface-emitting laser (VCSEL), PIN photodetectors and optical
emission filters on a GaAs substrate and coupled to a glass microfluidic channel
using a discrete micro-lens which is used to both focus the laser beam into the
microfluidic channel and adjust the emitted fluorescence into the photodetector
[90].

The concept of diffraction grating has been used for optical modulation and
filtering by introducing grooves on the top layer to diffract light at different angles
based on their wavelength and angle of incidence. In [91] a monolithic optical
phase shift detector on silicon has been reported that generates a current based on
relative phase difference between two incident beams. It consists of a single
diffraction grating layer etched at the surface of p-n photodiode. The proposed
system need extra elements e. g. a mixing element to project the two incident beams
into one or more common directions where they interfere, and also requires proper
placement of photodetectors at the right place to pick-up the interference products.
In [92] a micromachined optical modulator with electrostatic actuation has been
reported, fabricated by the conventional CMOS process. The optical modulator is
Operated by the interaction of single-layered fixed stationary gratings and moveable
sliding gratings, where the period of gratings is determined by the slide of the
moveable part which allows different diffraction patterns of reflected light. These
diffraction patterns differ according to the wavelength, incident angle, depth and
period of grating. This fabricated optical modulator still needs extra post-processing

to obtain a high-aspect-ratio microstructure.

25

For optical filtering another diffraction grating approach is reported in [93]
where analytical models were designed by using available metal layers in the
standard CMOS process. These metal layers are patterned in the complex structure
of CMOS photodiodes and act as a color filter to select a desirable wavelength. The
width of these metal layers and separation between them are manipulated to act as a
diffraction grating. For selecting a proper wavelength the pattern of metal layers
needs to be optimized. Although this approach is used for color filtering, there is no
discussion about blocking scheme of undesirable light and also about the number of
metal layers i.e. whether single-layered or multilayered diffraction grating is used.
Also there are no actual measurement results from the chip that validate the
analytical model.

2.2.4 Filterless techniques

Although filters enable fluorescence detection and are compatible with
many biomedical applications, standard filters add bulk and complexity that are
incompatible with integration into a microsystem platform. To address this
problem, techniques have been developed to bypass the need for optical filters.
Detection systems based on time-resolved fluorescence spectroscopy have been
proposed which do not need optical filters. In [94] real-time photodetection has
been implemented by extracting the transient fluorescence decay response. Laser
excitation was controlled by a fast clock from the chip to measure the exponential
decay of a fluorophore. Photocurrent response from the photodiode was integrated
repeatedly at different start times, and results of multiple measurements were

averaged to achieve high signal-to-noise performance. Similarly to improve

26

detection sensitivity, multiple measurements at the same start time were taken and
averaged. The output transient current, which is directly proportional to
fluorescence intensity, was generated by numerical differentiation. An 8x4 pixel
array was divided into four banks of eight pixels with a current-mode EA analog-to-
digital converter (ADC) sampling the output from a current-mode sample-and-hold
circuit (SH) within each bank. Signal current from each pixel, which consists of an
n-well/p-sub photodiode and a front end readout circuit, are time-multiplexed onto a
single current-mode SH composed of a differential transconductor with two
feedback storage capacitors. This CMOS sensor array build in a 0.5pm process was
able to measure fluorescence with a sensitivity of 108 photons/cmz, a dynamic range
of 74dB, and sub-nanosecond timing resolution. In [95] a relatively large 64x64
sensor array was built in a 0.18pm process utilizing the same technique with a
differential photodiode and per-column ADC, reporting a sensitivity of 8.8x106
photons/cm2 and sub-picoseconds tinting resolution.

A similar technique was utilized for fluorescence detection by using
different readout topologies and photodiodes. In [96] a two-chip vertically
integrated CMOS microsystem consisting of 8x8 AlInGaN blue micro-LED array
for the excitation pulse, and a 16x4 array of single-avalanche photodiodes (SPAD)
to detect fluorescence emission was proposed for time-resolved fluorescence
analysis. This microsystem was fabricated in a 0.35pm high-voltage CMOS process
and included in-pixel time-gated photon counting circuitry. Detection sensitivity
down to concentrations of 10nM was demonstrated. Although this system provides

good performance, the SPADs require high voltages ~20V that are not suitable for

27

low-power and portable microsystems. The same is true for [82, 97, 98] where
CMOS SPADs were used with different readout topologies for photon counting; all
of these microsystem require high voltage for the SPADs to detect photons.

2.2.5 Analysis and Discussion

Although integration of filter layers using post-CMOS processing has better
performance and blocks the excitation light efficiently, it also adds to the cost of
detection system and needs proper isolation so that it does not affect the other
components of the system. Also, this technique is not suitable for a combined opto-
electrochernical system where no filter layer can be inserted between photodiode
and the fluorophore because of fluidic environment.

Single-layered diffraction gratings compatible with CMOS process are
useful for optical modulation of light. But for multi-color imaging that requires
filtering, multi-layered diffraction grating is required to select the specific
wavelength of light. Moreover selectivity of the filter can be increased to a large
extent by using extra metal layers patterned in an optimized manner that can be
used for 1) extra blocking of undesirable light 2) increasing the selectivity by
blocking the light that is passed thorough the edges of top metal layers.

Time-resolved fluorescence spectroscopy has the advantage of fluorescence
detection without using optical filters; it reduces the cost and does not require post-
CMOS fabrication process. However, this technique also suffers from limitations in
the design. As fluorescent light has a very short life time (nsec-psec), time-resolved
fluorescence spectroscopy need a very high clock rate to separate the excitation

light from emission light. Also this clock needs to be strictly accurate with

28

minimum skew and jitter. On-chip high frequency clock generation leads to extra
design complexity with adverse effects on noise, CMOS area and power
consumption, while off-chip clock generation adds to system cost and bulkiness.
Moreover, because of some overlap of excitation and emission light, there is some
loss of information when the clock is used to tum-on and tum-off the arrival of
input photons to the photodiode. Thus, there is a need for an alternative efficient
fluorescence detection system that is highly sensitive, accurate and portable.

In time-resolved fluorescence, the response time of the photodiode also has
a critical role. With the high speed clock that alternately turns on/off the
photodiode, the photodiode response should be fast enough to generate electron-
hole pairs (EHP) within that time. Larger diode area increases the sensitivity and fill
factor, but at the same time increases diode capacitance and reduces response time
due to large RC time constant. Thus, time-resolved fluorescence may not be a
suitable technique when large diode areas are required.
2.3 Approaches for on-chip Electrochemical detection
2.3.1 Voltammetry

Different on-chip CMOS electrochemical detection systems have been
proposed and used for voltametry applications. In [99] an electrochemical array
microsystem was proposed that uses a high performance CMOS potentiostat and
amperometric readout circuit that supports cyclic voltammetry assay techniques.
Post-CMOS fabrication was done to develop an electrode array on the readout chip.
Similarly, [100] reports the development of a sensor using diamond-coated, micro-

needle electrodes and a readout chip designed in 0.5um CMOS for neurochemical

29

monitoring in the nervous system using fast-scan voltammetry. Moreover,
interdigitated gold electrodes have been fabricated with a CMOS circuit for
electrochemical detection of dopamine, and CV was performed using these
microelectrodes to identify the respective oxidation and reduction potentials for
dopamine in a phosphate buffer solution [101]. Similarly a differential potentiostat
circuit with differential recording electrodes implemented in 0.35um CMOS was
reported for fast scan cyclic voltarrunetry (FSCV) with improvement in the dynamic
range [102].
2.3.2 Impedance Spectroscopy

In [103] a compact, high sensitivity impedance to digital converter (IDC)
was reported with sensors array for impedance spectroscopy. Similarly, in [104] a
fully integrated electrochemical impedance spectroscopy (BIS) biosensor array was
presented, enabling flexible, parallel and electronic bio-molecular detection.
Measurement results of label-free detection for DNA, BSA and protein G in
different frequency ranges was presented. Moreover, in [105] a 4.4mm x 4.4mm
chip consisting of 4 x 4 electrode array with a fully integrated interface circuit was
implemented for stimulation and recording of the activity from electrogenic cells.
The electrode-electrolyte interface was characterized using EIS to model the
interface with an equivalent circuit.
2.3.3 Analysis and Discussion

Although efficient electrochemical detection systems have been reported, a
versatile and multimode electrochemical detection system that can be utilized for a

variety of applications is not currently available. There remains a need for a

30

versatile detection system that is capable of multiple voltammetry techniques. These
techniques require different input waveforms to be applied to the potentiostat, so a
programmable waveform generator capable of controlling the frequency (scan rate)
and amplitude (scan range) is an essential component for an efficient and
multimode voltammetry instrumentation system. Moreover, highly sensitive and
efficient CMOS readout circuit needs to be implemented that are capable of
measuring currents within the fA range.

2.4 Integration of optical and electrochemical detection

Membrane proteins play a vital role in cellular functions and there is a
pressing need to understand the structure and functions of these membrane proteins.
For better understanding and more accurate analysis of membrane proteins different
detection techniques could be combined that can simultaneously measure the
molecular mechanism from different aspects. Optical and electrochemical detection
techniques show great promises to be combined on a monolithic chip resulting in
multiple analysis of membrane proteins at the same time from different aspects with
an improvement in the detection accuracy and analysis.

In order to realize a versatile miniaturized opto-electrochemical
instrumentation system, all of the challenges described above need to be addressed.
This task requires a multidisciplinary approach coving biointerface development,
microfabrication, and CMOS circuit design. The following chapters describe efforts
to address the most critical challenges that can be solved through innovative design
of CMOS circuitry. These efforts form the basis of this MS thesis research and

include 1) a thorough analysis of the challenges for combining optical and

31

electrochemical detection in one system, 2) the design of a multi-mode voltammetry
circuit enabling different voltammetric techniques and 3) the design of an optical
detection circuit with an integrated on-chip optical filter that does not require post-

CMOS processing and also does not hinder electrochemical detection.

32

Chapter 3
Integration of Optical and Electrochemical detection

3.1 Motivation

Combining optical and electrochemical detection in a single chip permits
simultaneous determination of multiple biomembrane processes and is therefore
highly valuable, e.g., to characterize multi-protein membrane complexes that are
regulated by a wide range of interacting partners. This combination results an
improvement in the ability to measure biomembrane phenomena and helps to better
understand these phenomena with significant impact, for example, on field of
materials science, where there is a need to develop functional nanomaterials that
exhibit a desirable safety profile. Multiple analyses can be carried out
simultaneously on a particular membrane protein and interaction between different
molecular phenomena across a membrane-protein or an array of different
membrane proteins can be measured. This results in a versatile, efficient, high
throughput instrumentation system with the advantages of both optical and
electrochemical detection techniques. Also, these two techniques can be compared
to verify the output of each other, increasing the accuracy and reliability of the
overall system.
3.2 Background

Some detection systems have been reported for opto-electrochemical
detection using off-chip imagers for optical detection and off-chip electrodes for
electrochemical detection[106, 107]. However, these systems are bulky, expensive

and less efficient than an on-CMOS microsystem realization. Despite their

33

advantages, CMOS opto-electrochemical detection systems on a monolithic chip
introduce several challenging obstacles that are only beginning to be addressed by
research.

In [108] a label-free optical approach was proposed to quantitatively
measure “localized electrochemistry” on a lOum spot on a single planar electrode
using a single-channel potentiostat by monitoring the modulation of charge at the
electrode interfacial layer (EIL) during the redox process. Although the proposed
optical and electrochemical setup is not a CMOS-based approach, it highlights the
opportunity to develop a CMOS-based instrumentation system for opto-
electrochemical detection system.

An instrumentation chip that combines optical and electrochemical detection
and implemented in 0.35pm CMOS process has been reported [109]. The sensor
uses a combined 128 x 128 optical and 8 x 8 electrochemical pixel array with
readout circuitry. The optical detection circuit uses off-chip color filter resist layer
to separate fluorescence light form excitation light which require post-CMOS
processing resulting in an increase the cost of the system and complexity. Also,
these off-chip filters severely limit the functionality of the proposed system within a
microsystem platform. Moreover, for electrochemical detection the readout circuit
suffers from low sensitivity (nA) which needs to be improved to the pA-fA range
for an effective system. Moreover, the CV profile obtained from the on-chip
electrodes is unbalanced and unstable, which has a deteriorating effect on the
performance of this microsystem. In addition, the proposed system requires an

external signal generator to perform cyclic and differential pulse voltammetry,

34

which limits its use in portable applications. Finally, no performance results have
been shown for differential pulse voltammetry. For an autonomous and efficient
multi-mode voltammetry system, a waveform generator needs to be integrated on
the same chip that can be programmed to generate waveforms for different types of
voltammetry e.g. CV, LSV, constant potential and pulse voltammetry. All of these
limitations should be taken care of for an efficient opto-electrochemical detection
microsystem.
3.3 Challenges for opto-electrochemical detection

The integration of optical and electrochemical detection for simultaneous
operation on a monolithic chip is tremendously beneficial for measuring multiple
sensor interface activities. Such a system would be highly valuable for many
applications, but its implementation is challenging. Apart from the requirements
and challenges discussed in Chapter 1 for electrochemical and optical sensors
individually (high selectivity, sensitivity, accuracy, portability and high
throughput), a combined instrumentation system introduces critical additional
requirements and challenges. These include the design of: l) a CMOS-compatible
sensor interface suitable for both optical and electrochemical detection, 2) a highly
selective optical filter that blocks the excitation light source from the optical
detector without hindering electrochemical detection, 3) a multi-mode
electrochemical detection system with added functionality including a waveform
generator for different voltammetry techniques, 4) a highly sensitive and efficient

readout circuits to measure very low currents, 5) a microsystem scheme that

35

minimizes interference between electrochemical and optical sensors, and 6) proper
system integration and packaging.

To maximize the advantages of combined optical and electrochemical
detection, these measurements are required to be taken at the same time. One of the
main considerations for combining these techniques is their operating frequency.
Electrochemical sensing for bilayer membranes is carried out at very low
frequencies (mHz) while fluorescence detection is carried out at relatively high
frequencies because of the very short lifetimes (decay rate ~nsec-psec) of emitted
light. For fluorescence detection, the photodiode is exposed to light for a very short
time, and larger time may cause saturation of the photodiode. Thus, the photodiode
needs to be turned on/off relatively with a fast pace. In contrast, for electrochemical
detection, sensing electrodes are exposed for a much longer time to the electrolyte.
80 both these sensing mechanisms are required to be carried out independently and
at different rates with no impact on each other.

Optical and electrochemical detection techniques require different stimulus
signals with different response times and different output ranges, and they are
affected by different noise sources. Thus, the readout circuits for both
electrochemical and optical sensors will have different architectures and different
requirements. For example, simple detection of fluorescence places more
significance on the minimum detectable signal (MDS) while quantification of
fluorescent light has more significance on SNR and dynamic range (DR). On the
other hand, for electrochemical detection, DR requirements are less significant

while MDS or sensitivity requirements are high [110]. These requirements and

36

parameters dictate the design of readout circuits for these techniques. Also for
simultaneous measurements, performance of the instrumentation system might be
affected by the interference between both readout circuits. Noise from both readout
circuits need to be isolated and minimized so that it does not affect the performance
of the adjacent circuit.

Optical detection circuits need filters to get the desired information. For
example, for fluorescence detection, filters are incorporated in the design to block
the excitation light while emission light at a different wavelength is allowed to pass
through it. Typically, these filters are built between the photodiode and the
fluorophore using post-CMOS fabrication. However, for a combined on-chip opto-
electrochemical system, these techniques cannot be used because they would isolate
the on-chip electrode from the readout circuit making electrochemical detection
ineffective. Thus, to realize a combined opto-electrochemical system, either a filter-
less approach is required for optical detection or a filter is required that does not
provide any hindrance to electrochemical detection.

For both measurement techniques, the geometry (physical structure) of the
system has critical effect on performance and functionality. The working electrode
for electrochemical detection can be placed on the top of the photodiode, but in that
case it should be transparent so that emitted light can pass through it and get

absorbed by the photodiode. Because of the relatively large size of the working
electrode (~0.1mm2), this topology would relieve some constraints for optical

detection by permitting a photodiode with large area to receive more photons and

increase dynamic range. Another approach may be to place both sensors (electrode

37

and photodiode) side by side with the membrane protein interface on top. This
would not require transparent electrodes because there is no hindrance in the path of
light, but at the same time it would decreases the fill factor of the pixel as the
electrode would cover significant area beside the photodiode. Also, because light
travels in all directions, this approach would result in loss of emitted light due to the
area covered by electrochemical electrode. Moreover response from adjacent pixels
would need to be properly isolated from each other to increase the selectivity and
accuracy of the instrumentation system.

The post-CMOS processing required to develop electrochemical electrodes
on the chip’s surface needs to be highly accurate and selective so that it does not
affect the photodiode array on the CMOS IC. Also proper isolation is required not
only between electrodes for electrochemical detection but also between photodiodes
and electrodes, to minimize the effect of adjacent sensors on the measurement that
may impact on accuracy of the system.

Packaging is also an important issue to be considered for the combined
opto-electrochemical instrumentation system. Optical detection needs to be carried
out with minimum background light to increase the minimum detection limits.
Efficient and accurate fluid delivery is also required for a membrane protein
biosensor system. Both optical and electrochemical readout circuits need to be
protected from the adverse effects of incoming light and fluid interfaces that could

adversely affect the performance of the microsystem.

38

Chapter 4

On-Chip Instrumentation System for Electrochemical Detection

This chapter describes a versatile and multimode CMOS instrumentation
system specifically designed for electrochemical detection using multiple
voltametry techniques including cyclic voltametry (CV), linear sweep voltammetry
(LSV), constant potential voltammetry and normal pulse voltametry (NPV).

4.1 Chip Architecture

The primary targeted application for the on-chip electrochemical detection
system is the characterization of thin-film biomaterials, such as redox enzymes,
membrane proteins and DNA receptors. These applications utilize very low
electrochemical (EC) potential scan rates, in the range of mV/sec, and moderate
scan range, below 2V. The three main blocks of a voltammetry-based
electrochemical instrumentation system are the voltage waveform generator, the
potentiostat and the amperometric readout circuit. For a complete electrochemical
detection system, all of these components need to be integrated monolithically on a
CMOS chip. An autonomous and versatile CMOS chip would contain all of the
necessary instrumentation electronics and a communication interface that permits
user control of measurement operations and reporting of measurement results. The
Chip-scale miniaturization and integration of electrochemical sensors and their

instrumentation electronics has many advantages, most significantly extending the
litnits of detection by improving the signal to noise ratio. The proposed
instrumentation chip was designed in a 3M, 2P 0.5um CMOS process with a 3.3V

supply voltage. As shown in Figure 4.1, it includes all three main components of an

39

electrochemical detection system, namely a digitally programmable signal generator
to produce a variety of stimulus signals for different voltammetry measurements, a
2x2 array (4-channel) of single-ended potentiostat for electrode biasing, and an
amperometric readout circuit to measure response currents resulting from the
stimulus voltage. The chip has the capability to integrate electrodes and sensors
directly on the die surface using post-CMOS processing. On-chip measurements
permit improvements in sensitivity and immunity from environmental interference
by minimizing the length of electrical connections between electrodes and

instrumentation circuit and eliminating external wiring.

Working Electrode
array

I
I
I
I
I
I
I
J

 

   

Triangular
1AA, .9...
EMA-Linear ramp.

Constant
L: potential

. Normal
Multi-mode waveform llllllfl pulse

generator ........

Digital
Input

   

Reference
electrode

Counter
electrode

 
    

'U
9.
(D
3
2".
O
(D
o-c-
m
o-c-

 
       

0
IT!

 

 

 

 

 

 

R7

 

Potentiostat
Readout

 

 

Output

 

Figure 4.1. Block diagram of the autonomous amperometric instrumentation system.

4.2 Programmable Waveform Generator

A waveform generator is an important component of an autonomous
electrochemical instrumentation system that requires different input waveforms for
different voltammetric techniques. To fulfill the requirements of the proposed

electrochemical instrumentation system, a multi-mode waveform generator is

40

desired to produce signals of various shapes for different types of voltammetry,
amplitudes (scan range) and frequencies (scan rates). Programmable analog circuits
could be constructed to generate continuous waveforms, but different analog
circuits would be needed for each desired signal shape (triangle, linear ramp, pulse
etc.) with an undesirable impact on power and chip size. To avoid the adverse
impact of multiple analog waveform generators, a digital solution was explored.

A DAC with a digital controller could implement all shapes, but the output
would be quantized and the complexity of the controller could be significant for the
desired set of signal shapes. This issue was carefully considered and a variety of
designs were explored. A thorough examination of the impacts of stimulus
quantization on electrochemical measurements was also performed. Tests
demonstrated that the steps (abrupt changes) in quantized stimulus voltage signals
generated noise in the output current. However, experiments also demonstrated that
the noise reduced as voltage step size was decreased. Through electrochemical
experiments, it was determined that, for a 2V scan range, a step size of 8mV or
lower produced optimal output results comparable to continuous stimulus signals.
This 8mV step size can be realized by using an 8-bit DAC. Although a higher
number of bits, i.e. 12, 14, could be used, that would also increase the complexity,
power and area of the chip, which is undesirable for electrode arrays on chip. It was
also determined that, with careful design, a very efficient digitally controlled

waveform generator could be realized with low power and chip area. Therefore, a

DAC-based multi-mode waveform generator was developed.

41

The digitally programmable waveform generator is shown in Figure 4.2. It is
composed of a digital control block, a lO-bit comparator, a 10-bit bidirectional
counter, and a 10-bit DAC.
4.2.1 Control Block

The control block was designed to generate non-overlapping clock signals
and control signals to: I) reset the waveform generator, 2) control the counting
direction of the bidirectional counter using a comparator, and 3) select from
multiple waveform modes to enable a specific voltammetry technique. A single pin
was used to serially shift the 22-bit configuration input data to the control block as
shown in Figure 4.3. Based on this input, the MUX and comparator select the
desired signal frequency (Clock), amplitude maximum (S_high), and amplitude
minimum (S_low). This on-chip control block also includes a clock divider that
provides fine tuning of the scan rate; the output clock can be selected from CLK,
CLK/2, CLK/4, or CLK/8, where CLK is the external input master clock.
4.2.2 Bidirectional Counter

For cyclic voltammetry (CV), the input potential is scanned in both positive
and negative slopes to stimulate reduction-oxidation (redox) reactions. A
bidirectional counter is needed to perform this function. Based on the comparator’s
output, a 10-bit counter that counts in either direction is needed to produce the
required waveform. To meet this need, the counter shown in Figure 4.4 was

designed based on a reported counter

42

 

 

  

 

 

 

 

 

 

 

 

 

Digit I I; Bi- — o t t
Inpu Control Comp- . . . u pu
_, arator dgicjtfigpl Latch —> DAC
Clock *I A
I I

 

Figure 4.2. Block diagram of the multi-mode voltage waveform generator.

 

 

 

 

 

 

 

 

* , 22-bit external Data
| 2-bit | 10-bit L‘IIO-bit |¢_—_
U
requency Counter Output
Divider
MUX VComparavtor
Clock S_ llow S _hligh

Figure 4.3. Functional diagram of waveform generator block.

design [111-113] that was modified to fit this application. The binary output of the
counter is latched and is given to a binary-to-thermometer decoder that generates a
thermometric coded digital word. This thermometric code is used by DAC in the
next stage and is used to control the switching of DAC analog signal currents. The 4
MSB binary bits of counter output are decoded into a thermometric code that is
given to the DAC along with its inverted output, as shown in Figure 4.5, to convert

into a discrete analog signal.

43

Master stage V d d grave Stage

‘\
\

Vbias>

1.1L

T
T
l ,
T
T
1
W

............................

 

 

 

mild!”

P HIE] F—TnpuLclk

 

 

 

 

Swap

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

" l I.
lnput_clk 1 IOUTP
rOUTN
S.input fit—— Iii-— MEL 4E,
Shift_en J
INN = J
mp = / 4.7 at
Count_En r 1
Figure 4.4. Schematic of the bidirectional counter.
6-bit LSB —‘— l
4) |
10-bit 4-16 |
Bidirect- Binary 10 , 'To DAC Input
ional Thermometer 16'b'ts:
counter * 4-bit Decoder _ _ l
MSB

Figure 4.5. Signal flow from counter to DAC.

4.2.3 Digital to Analog Converter(DAC)
The basic function of the DAC is to covert the digital output of the counter

into a staircase analog signal that can be applied to the potentiostat. As discussed

above, an 8-bit DAC is sufficient for these waveforms; however, to compensate for
process variations in chip fabrication, a 10-bit segmented DAC has been designed
to generate a staircase analog signal. It uses a combination of an R-2R ladder
network and a current mode thermometer decoder, with the lower six counter bits
setting the R-2R ladder and the upper four bits controlling the thermometer decode,
as shown in Figure 4.6. This hybrid topology enables the DAC to operate with
small currents, improves matching between resistors, and requires less layout area
than a full R-2R ladder network for lO-bit. The DAC output is then buffered and
sent to the potentiostat. The DAC output spans the range of +1 to -1 relative to
analog ground with a step size of 2mV. This provides a stimulus signal up to 2V pk-
pk, which is suitable for most electrochemical measurements.
4.2.4 Operation of Waveform Generator

Waveform generator operates in two phases as shown in Figure 4.7. In the
first phase, a 22-bit data word is shifted serially into the control block to select scan
rate, scan range and type of output signal (waveform). In the second phase, that

output signal is generated based on the input signals setup in the first phase.

DAC

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l t
"p” 4-bitMSBI e-bithel
Rf
L L L I I I I
1 + _V>out
Thermometer R-2R Ladder -
Code Network Vref

Figure 4.6. Block diagram of hybrid digital to analog converter (DAC) used in
waveform generator.

45

Clock

Clock_shifl

RST_signal

Voltammetry_LS
Voltammetty_DC
Voltammetry_NP

S_high

DAC_output

DAC_output

 

 

 

Low/high

 

Low/high

 

 

Low/high

 

“ IL

 

 

 

 

DAC_output

 

 

 

 

 

Wax peak

Min peak
If Voltammetry_LS is high

O

If Voltammetry_DC is high

 

 

l_f_Voltammetry_NP is high

‘I

 

WWII

 

DAC_output

 

 

nflfllllllll ..............

Figure 4.7. Input and output waveforms of waveform generator.

4.3 Single-ended potentiostat

The basic function of an electrochemical potentiostat circuit is to control and

mai ntain the voltage between the working electrode (WE) and the counter electrode

46

 

 

 

 

 

 

VDAco—m— 15kt) Vref2
Vref1'—é\l’(\g— VRE
———’VV\/——T $1 5k!)
ska ska _
0P3 VCE

Figure 4.8. Single-ended potentiostat designed for the instrumentation system.

(CE) under varying current conditions. Figure 4.8 shows the schematic diagram of

the 3-electrode single-ended potentiostat following the approach in [114]. Output

from the waveform generator (VDAC) is level shifted by OP] to match the input
range of 0P2 by summing it with Vrefl.
The output from 0P1, Vouu, is given by

VDAC We 1
Voutl = 2 rf (4.1)

 

The output of 0P2, which is buffered through negative feedback using 0P3, is

connected to the CE. For microscale sensors, it can be assumed that the solution

resistance between the reference electrode (RE) and the CE is negligible so that CE

and RE are at the same potential. The input of 0P3 is connected to RE and no
current flows through the 0P3. The output current resulting from the input stimulus

Voltage at CB can flow only between CE and WE. Vrefz is a DC offset control input

47

to adjust the scan range. A single bias circuit is used for all three op-amps,
decreasing the layout area considerably compared to the potentiostat in [114].
4.4 Amperometric readout circuit

To measure the electrochemical response current at the WE, a capacitive
current readout circuit was developed [115]. This circuit was designed by
colleagues within our research lab and is included here only to complete the
description of the chip. As shown in Figure 4.9, a switched-capacitor (SC) charge

integrator converts the current to a voltage. The voltage is sent to a programmable
gain amplifier (PGA) with a gain of C2/C3, where C2 is a digitally programmable

on-chip capacitor array. The output of the PGA is sampled and held and then fed to

an analog-to-digital converter (ADC) after a low pass filter. The non-overlapping

clock signals for switches phl, phz and ph3 are generated by an on-chip clock

generator block. The output voltage of the current readout circuit is given by:

I WE C 2
V : C 4.2
0‘“ f, . C1 C3 ( )

 

where [WE is the current at the WE and f8 is the frequency of switch phz. Thus, the

input current range can be adjusted to fit different application through
programmable selection of the clock frequency and the PGA gain.

To reduce noise and amplifier offset, correlated double sampling [116] was utilized
in the current readout circuit. All switches were realized with minimum size to
curtail charge injection. To reduce clock feedthrough errors, dummy switches, with

Precise time sequence control, were used.

48

 

 

 

 

 

 

 

 

 

 

 

VWE o— i CZ — Vrefz Ply — :out

V'efz CS-Array +

’ ref2
Vref2 fi—J

Figure 4.9. Amperometric readout circuit.

 

 

 

 

 

4.5 Test Setup

To characterize the functionality of the instrumentation chip, measurements
using an electrochemical solution are required. Digital control signals must be
generated for the chip along with reference voltages for the potentiostat. In the
proposed system, the waveform generator, potentiostat and amperometric readout
circuit can each be tested and their functionality can be verified independently. The

test setup for the instrumentation chip is shown in Figure 4.10. Test electrodes

fabricated on a silicon chip substrate were immersed in the ferro-cyanide (FE3CN2)

solution, with a copper strip used as the CE and a commercial Ag/AgCl electrode
used as the RE. To verify the functionality of chip that is mounted on the test PCB,
different waveforms for CV, LSV and constant potential voltammetry were
generated, and test electrodes were scanned with different scan rates and scan
ranges. A data acquisition card (NT DAQ-6259) was used in conjunction with
LabVIEW to generate the control signals for the chip and to acquire the test data at

different conditions and waveforms. Test results are presented in Chapter 6.

49

 

 

 

Computer

 

 

 

 

 

w
R r__,)

 

 

 

 

 

K%\I

DAQ

 

 

 

PC running labview and

connected to DAC

Test PCB

 

 

    

. 1" -‘ "*
‘-.' ",1 ,,
.1... .\ .

Electrolytic solution with
WE array, RE and CE

Figure 4.10 Test setup for the instrumentation chip with electrode array.

50

 

Chapter 5
CMOS Instrumentation System for Optical Detection

This chapter describes a CMOS optical detection system designed to
measure fluorescent emissions from an on-chip biosensor. The main components of
a fluorescence detection system include an excitation source, an emission filter, a
photodiode, and readout circuit to amplify photodiode currents and generate an
output voltage. Usually light detection in CMOS technology is accomplished using
a reverse-biased PN junction diode that is comprised of a substrate-to-well junction
and a diffusion-in-well junction. Figure 5.1 shows a block diagram of an on-chip
optical detection system that includes a fluorescence excitation source, a sensor
material that emits fluorescence light, a CMOS IC that includes an integrated filter,
and a photo detector with readout circuit to absorb the light energy and further
process it. The fluorescence excitation source may be a laser source or LED source
to shine light of a specific wavelength on the target/sample containing a
fluorophore. The fluorophore emits light at a wavelength that is different from the
excitation source. The light energy is detected by the on-chip photodetector and
readout circuit within the detector pixel, as shown by the signal flow path in Figure

5.2.

51

Light Source

External
Emitted Light Sample Casing

Excitation
Light

 

W. '

 

 

 

 

 

CMOS '0 IC Holder Interference Filter

Figure 5.]. Block diagram of CMOS optical detection system

 

 

 

- U . &
Photocurrent Integration Charge conversion

’ Voitage

 

 

 

 

 

 

 

 

 

 

Generation COnvérsion . I. . ;
_ :1) (C) —V(V) :1)
aphlA
Fluorescence Photodiode WW

source -
Front-end electromcs

Figure 5.2. Signal flow path from emitted photons to the output.

5.1 Chip Architecture

Many engineering tradeoffs are involved in combining optical and
electrochemical detection circuits on a monolithic chip. These tradeoffs affect the
fill factor for photodiode and its dynamic range, interference and noise from
combined circuit, and orientation of both optical and electrochemical sensor.
Design for the optical detection prototype chip was done taking into considerations

the requirements of a combined opto—electrochemical detection circuit on a

52

monolithic chip. For the intended application of biofluorescence assays, i.e. DNA
analysis and characterization of membrane proteins including planar bilayer lipid
membrane (pBLM) and tethered bilayer lipid membrane (tBLM), the goals is
typically not to create an image of the sensor but rather detect and quantify the light
emitted from each sensor cell. Thus, a moderate or small pixel array density can be
used [117].

A block diagram of the proposed CMOS chip is shown in Figure 5.3. The
chip was designed in a 3M, 2P 0.5pm CMOS technology with a 3.3V supply and
includes a 4x3 pixel array, with each pixel including: 1) a photodiode to collect
photons, 2) a readout circuit to condition the input signals, 3) addressing circuitry to
access individual array cells, 4) output buffers to isolate the circuit form loading
effects, and 5) an optical interference filter employing a novel approach that relies
on layers within the standard CMOS process. A key feature of this chip-scale
system compared to other CMOS imager applications is the ability to provide
optical filtering within the CMOS chip, which enables biomaterials to be formed on
electrodes directly on the chip’s surface, as desired for the electrochemical mode of
detection in the desired opto-electrochemical system. In the proposed design, each
individual pixel is based on a sarne-sized photodiode, with the same cell size and
constant pixel pitch. Compared to traditional CMOS imager applications, a large
photodiode detector has been used in order to match the pixel density to the
anticipated electrode density of a future combined opto-electrochemical system.
This permits each pixel in the array to contain both electrochemical detection

Circuitry along with a photodiode. Because the electrochemical pixels are large,

53

~200um diameter, the photodiode can be quite large to maximize the fill factor of
the cell with the added advantage of higher optical sensitivity per pixel.

Because the pixel array is small, optical readout is done per-pixel using a
circuit based on a first order 2A ADC so that each pixel can directly convert photo
charge to a digital output that can be directly read by an off-chip counter. The
advantages of a pixel-level ADC over a chip-level or column-level ADCs include

lower noise and low power

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I
'5 r? ‘ I i” I
o
* I! :5
m
. a I- ~ ‘ I g T _.;f: 43%? .
_'1'-' . 3
I». ~
L ‘ Peripheral Circuitry -—>
, to PC

 

 

Figure 5.3. Block diagram of the CMOS optical detection chip.

dissipation [118]. Also each pixel can be processed and addressed independently
and rapidly without having stringent speed requirements for the circuits within each
pixel.

For the optical interference filter, a novel technique has been introduced that

Utilizes standard CMOS metal layers to realize a diffraction grating. The physical

54

 

orientation of the metal layers has been exploited to block excitation light while
allowing fluorescent emission light to pass through and be absorbed by the
photodiode. This filter is implemented inherently within the CMOS design and does
not need any extra post-CMOS processing steps or filtering materials inserted
between the biosensor films and the CMOS detectors[87-89]. Although filtering is
necessary to ensure the photodiodes respond to fluorescent emissions and are not
saturated by background current due to the excitation source, the use of filters does
adversely impact the responsivity of the system. The responsivity can be improved
to some extent by applying relatively a high reverse bias voltage as implemented in
[119] where reverse bias voltages of 1.2V and 14.2V increase the responsivity by
1.6X and 2.8X respectively. At the same time, it has an advantage of lower diode
capacitance, resulting in a smaller RC delay and increasing the operating speed of
the photodetector.

The final element of the optical detection chip is the peripheral circuitry.
This includes biasing circuitry, intermediate test circuits, and address decoders to
select a particular pixel.
5.2 Pixel Architecture

The amount of current generated in the photodiode in the pixel mainly
depends on the incident light intensity, quantum efficiency (OE) and area of the
photodiode. Output voltage from the pixel depends on the photocurrent generated in
photodiode, the photodiode’s capacitance and the sensitivity of the readout circuit.
Design for the pixel architecture takes into considerations the effect of all of these

Parameters. Each pixel produces a digital output corresponding to the current

55

 

 

generated in the photodiode by incident photons. To match the anticipated structure
of the sensor array on the surface of the chip, each pixel consisting of photodiode
and front-end readout circuit covers an area of 220x250pm2. The active photodiode
area is 150x200um2 and the readout circuit for each pixel covers an area of
70x50um2, as shown in Figure 5.4.
5.2.1 Photodiode

A photodiode is the heart of a CMOS image sensor and is used to absorb
photons. After an analysis of the different types of photodiodes available in CMOS
image sensors including p-n, PIN and avalanche diodes, a standard n-well/p-sub
photodiode shown in Figure 5.5(a) was selected. This n-well/p-sub photodiode has
better performance in terms of fill factor and quantum efficiency, and it also can be
fabricated using the standard CMOS process without requiring any additional

processing steps. This photodiode uses a

Photodiode Sal/lir'gcyh

 

Compa
rator
Integration
\

Figure 5.4. Cadence layout of pixel including photodiode & readout circuit.

56

 

lightly doped large area n-well diffused into a p-substrate to create a pn junction.
The n-well acts as cathode while p-sub acts as anode. The photodiode is reverse
biased to absorb the incident light that carries energy to create electron-hole pairs

(EHP) in the depletion region of the diode, resulting in a current flow. This reverse-

biased diode current ( l D ) is given by
[DzlS—IPH (5.1)
where l S is the saturation current and [pH is the photocurrent due to the incident

light. A layout view of the designed photodiode is shown in Figure 5.5(b).
5.2.2 F rant-end Readout Electronics

The basic function of the readout circuit is to collect the output current from
photodiode, amplify it, and convert it into a voltage so that it can be processed in a
later stage. Different topologies for the front-end readout pixel have been reported
taking into considerations the size of pixel array, dark current and thermal noise

[78-82]. The chosen readout circuit for the pixel is based on first order sigma-delta

      

 

 

953"” .

 

Figure 5.5. (a) Schematic diagram of the photodiode structure, (b) Cadence layout
of n-well/p-sub 150x200pm2 photodiode.

57

lute;

C011

liar

ADC (EA), as shown in Figure 5.6. Main components of this 22A ADC are 1) an
integrator composed of an op- amp and a capacitor in a feedback loop, 2) a
comparator with digital output, 3) a clocked latch and 4) a feedback circuit. This 2A
ADC has the advantage of low sensitivity to transistor noise and large charge
handling capability. Also, noise in the comparator is reduced by the feedback circuit
[118].

Figure 5.7 shows the schematic of a high gain operational transconductance
amplifier (OTA) with single-ended output used in the integrating amplifier. A
single stage amplifier with folded cascode topology was used for a high gain with
low demand on hardware area. Also, this cascode gain stage gives immunity to the
Miller effect at high frequencies. M1 & M2 form an input differential pair; M3 &
M4 form an NMOS current source; M5 & M6 are NMOS cascode transistors to the
input differential pair; M7-M10 form a wide-swing cascoded PMOS current source;
and M11 is a current source for the differential input pair. To achieve low dark
current and high linearity, PMOS transistors are chosen for the differential input
stage. The simulated performance of the proposed design shows a DC gain of 95dB
and unity gain bandwidth of 25MHz with a phase margin of 50°. Simulated power
dissipation of this OTA with a lpF load is 1mW.

The second main component of the EA ADC is the comparator. Propagation
delay is an important parameter in the design of the comparator because it limits the
highest clock rate that can be used in the ADC. The delay in the comparator was
carefully considered in the proposed design. The proposed comparator consists of

three stages; an input preamplifier, a decision circuit and an output buffer [120], as

58

shown in Figure 4.8. This topology has the advantage of relatively small
propagation delay, in the nanosecond range. The preamplifier consists of a
differential amplifier with active loads. Gain of the input stage is set by the
transconductance, while input capacitance is determined by the size of input
transistors. Decision circuit of the comparator is used to discriminate between
signals. This circuit uses positive feedback from cross-gate connection of M7 and
M8 to increase the gain of the decision circuit. Output from the decision circuit is
given to output buffer which is a self biasing differential amplifier. Output buffer
also includes an output inverter as an additional gain stage and to isolate any load
capacitance from the self-biasing differential amplifier. Complete schematic of the
comparator is shown in Figure 5.8. Output of the comparator is passed through a
latch and digital output after passing through output buffer is given to off-chip

counter to measure the width of output pulse.

VddID
lref

O

ff-chip
Counter

 

 

N.-
Photodiode

 

 

 

 

 

 

Figure 5 .6. Sigma Delta (2A) ADC used for readout.

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1’ M9 If]: {a M10
Vpl: 1.1M11
*1 sz I‘
Vin+ M7 Hi ll-I M8
: :EM1 M2 j|__
: VOUI
Vin- " ’
M5 -It VH1 :1-
«I 'k M6
M3 :1: Vnz II: M4

 

 

 

 

I

Figure 5.7. Schematic of the folded cascode OTA.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Input
preamplifier —]—
’11,..M4 :41; 3. 4% MS
I I M12
3 M11::: i II:
. I ..,—--r-__ . .
1 . p . . , s,
V, v- 033:;ng I—I: M15 finals; M18
M19
I: I— LI: 3.1 x
.. ’ M6,. M7 M8 M9 /

.. ‘\-.LI..- ’1 "X.
Ibiasr—Ii M3 § LI? M10 .

A “ _r

I

Figure 5.8. Schematic of the voltage comparator.

 

 

 

60

5.3 On-chip Filter
5.3.] Introduction

The basic function of the on-chip interference filter is to selectively inhibit
transmission of the excitation light and allow the transmission of light emitted from
the fluorophore in bilipid membrane (BLM). Emitted light from fluorophore has a
longer wavelength than the excitation light due to stokes shift and this change in
wavelength is exploited to get the required wavelength of light emitted from
fluorophore. The amount of change in wavelength depends on the type fluorophore
used in the analysis. In many fluorescent systems, the filter is implemented using a
specific material, inserted between the fluorophore and the detector that allows
transmission of only a specific wavelength of light. Traditionally, filters that have
been integrated on CMOS ICs rely on post-CMOS processing to build optical filters
on the top of the chip. However, for a combined opto-electrochemical system post-
CMOS processing is not feasible to integrate the filter as biomembranes are stored
in fluidic environment
5.3.2 Design Methodology

In the proposed optical detection system, a novel technique has been
developed for the on-chip filter by using the metal layers in the standard CMOS
fabrication process to create an interference filter without any post-CMOS
processing. Metal inherently blocks the visible light to a large extent based on its
permittivity and this property of metal has been used in the proposed design of

filter. The placement, width and horizontal gap between the CMOS metal layers

61

have been manipulated to create a multi-slit diffraction grating that passes the
emission light and blocks the excitation light from the light source.

A conceptual block diagram of the on-chip filter that uses CMOS metal
layers as a diffraction grating in C5N process is shown in Figure 5.9. The top metal
layer, M3, is used as a transmission grating with openings between M3 strips acting
as slits with constant widths and constant separations. Similarly, the M2 layer is
used as a reflecting or blocking grating. The third metal layer M1 is used for an
improvement in the selectivity of the filter. Also, M1 serves the purpose of blocking
additional amount of light if passed though M2 layer because of its permittivity.

Each slit in M3 is considered to be a point source generating the same flux
intensity distribution. When incident light is perpendicular to the diffraction grating,
the angle of diffraction for bright regions of light is given by the diffraction grating
equation

d sin 6m = m}. (5.2)

where d is the distance between the centers of adjacent slits, it is the wavelength of

incident light, 6m is the angle of diffraction, and m is an integer denoting diffraction

or spectral order. Equation (5.2) shows that the angle of diffraction increases with
increase in wavelength.

This multi-slit grating can be considered as an array of N identical slits, each
slit having a width x and equally spaced. Calculations were done for the above
mentioned model (figure 5.9) and based on some ideal assumptions with unity
diffraction order i.e. m=l. Incident light was assumed to be perpendicular to the

photodiode. The reason for this assumption was that, in the proposed detection

62

Emission light Excitation light

iiiiiilllllllliililllllluuu

43 = 15W“; :1 = OWE, Blocked Light <—d = 2.4 m —»
9n

l- -
= 0.
1 Optical
7

Y1
length (L)

      
  

   

  

2

  
 

 

 

 

’ \
C = 091"" X3 = 2.4um
Figure 5.9. Conceptual block diagram of a CMOS optical interference filter.

system light source is too close to the photodetector and also incident light at the
edges of slit would be blocked by the slit thickness to some extent.
From equation (2) and the trigonometric relation of the dispersed angle, the

following equation is derived
x

tan 6 = — 5.3
L ( )

where L is the optical path length and x is the distance along the surface of the
photodiode. Equations (2) and (3) relate the dispersed distance, x, to the wavelength
1. of incident light.

Depending on the fluorophore, the Stokes shift generally ranges between
lOnm and 150nm [121]. Calculations were done using the above equations for the

visible range of light (320-780nm) to determine the angle of dispersion. Based on

63

the diffraction angle and vertical distance between M3, M2 and M1 layers,
placement and width of M2 and M1 layer strips was determined that is required to
block a specific wavelength of light in the visible range.

5.3.3 Design Constraints

5.3.3.1 Design and fabrication rules

Filter design based on these calculations is highly dependent on the “design rule
check (DRC)” rules in the layout for a specific CMOS technology. Small slit width
results in a wider diffraction angle which makes it easier to separate the diffracted
wavelengths and block the excitation light, but both the slit width and separation
between the slits are limited by the DRC rules. Similarly slit thickness also has a
significant effect on the transmission of light which is not considered in Equation
5.2 to determine the diffraction angle. Due to this thickness, incident light with

larger diffraction angle is also blocked. Thickness of different metal layers i.e. M1,
M2, M3 and vertical distance between these metal layers i.e. y1, y2 and substrate L

are governed by the specific CMOS process technology and cannot be revealed
because of the proprietary rights. Thus, there is a limitation in the transmission of
emission wavelength range and an effect on selectivity of the filter. Taking into
considerations the DRC rules, optimal wavelength of light calculated for C5N
process was approximately 600nm-780nm, while thickness of metal layers has an
additional effect on the calculated range. Table. I shows an example of the

limitation of selecting wavelength using DRC rules for C5N process. This table

I

shows the calculated diffraction angle “0” with a slit width “xl’ of 0.9pm for

excitation light (600nm), minimum (610nm) and maximum emission light (750nm)

I,

along with M2 stripes width “b” and opening between M2 stripes (slit width) “x2

to allow the emission light. However, due to DRC rules for C5N CMOS process,
minimum width of M2 layer used was 1.2um with a minimum opening Of 0.9um
between M2 stripes which limits the minimum wavelength to be transmitted
through opening between M2 stripes. Results of the effect of these DRC limitations
and metal thickness have been presented in chapter 6 which conclude that an
excitation wavelength less than 730nm to be blocked by M2 stripe width. For
advanced CMOS processes that uses less metal widths because of different DRC
rules, diffracted angle increases for the same wavelength of light and also the
transmitted wavelength range is also increased as presented in Section 5.3.4.

Also proper placement of M1 layer can increase the performance of filter by
blocking a percentage of excitation light that is transmitted due tO the permittivity

of M2 layer and also increasing the selectivity of the required wavelength of light.

TABLE I. CALCULATED DIFFRACTION ANGLE FOR EXCITATION AND EMISSION LIGHT
FROM FLUORPHORE AND M2 LAYER WIDTH AND SEPARATION BETWEEN M2 STRIPS FOR

 

 

 

 

c5N PROCESS
Wavelength Diffraction M2 layer width M2 layer opening
1 angflfi) (b) (x2)
600nm
(Excitation 14.48 °
light) 0.98pm( to block ~0.27um ( to
610nm the 600nm light) transmit light
(minimum 14.72 ° between 610nm-
emitted light) 750nm)
750nm
(maximum 18.21 °
emitted light)

 

 

 

 

 

 

65

5.3.3.2 Filter Transmission Factor

Due to use of metal layers on the top surface of photodiode for filter and
limitation from DRC of CSN CMOS process, amount of transmitted light on the
photodiode is also affected. In the proposed technique incident light (both excitation
and emission) is alternately blocked and allowed by the top metal layer M3. Fill
transmission factor was determined based on some assumptions and a set Of
following equations.

Assuming I as excitation light from the light source and with E as emission
light from the fluorophore distributed in all directions while X is the incident
emission light on top metal layer Of filter. As emission light intensity is always less
than by some factor and also background light has also an effect on emission light,
so emission light from fluorophore is given by

E1 = I ' aF ' “background (5-4)

where a}: is the emission factor from fluorescence and abackground is the

background light factor. As photodetector is placed only under the fluorophore so

half of the emission light is incident on photodiode area an is given by

1
X =—E
l 2 1 (5.5)

where X 1 is the emission light incident on photodiode. As filter is covered by
alternating layers of top metal layer M3 and it was determined from the covered
area that 35% of X1 is transmitted through the top metal layer. Assuming 50% Of

emission light incident on the pixel, calculated transmitted light through the filter is

9% Of emitted light from fluorophore for C5N process. For IBM 0.13pm process

66

which uses less metal width at the tOp, fill transmission factor calculated is 20%
using the same set of equations as used for C5N process.

Similarly, analysis was performed to determine the minimum amount of
excitation light that can be quantified by the readout circuit by using the following
equation

N E
_ P p
1p - R[—t——] (5.6)

where lp is the photonic current that can be detected by readout circuit, R is the

responsivity of photodiode, N pis the number of photons and E p is the energy of

photon.
5.3.3.3 Process Variations

Filter design is also constrained by the process variations of fabrication
process. As a result, metal widths may change that have an impact on the
functionality and performance of filter. An analysis has been done to determine the
effect of process variations of on the functionality of filter. A process variation Of
15% for metal widths was assumed for simulations. Details of this analysis have
been presented in chapter 6.
5.3.4 Effect of CMOS technology feature size

For CSN process only 3 metal layers are allowed but for advanced CMOS
technologies e.g. 0.35pm, 0.18 pm process, additional metal layers can be included
in the layout and also these processes have different DRC rules which can be
manipulated to increase the functionality and selectivity of the filter. Smaller slit

width and separation can increase the diffraction angle, which makes it easier for

67

the filter to block/allow the incident light. To determine the effect of extra metal
layers and advanced CMOS technologies, filter design was also implemented in
IBM 0.13pm CMRF8SF process. Calculations were done for the diffraction angle
and width of metal layers for the visible range Of light. Table. II shows the
diffraction angle for the same wavelengths Of light used in C5N process. For IBM
0.13pm process four metal layers i.e. M1, M2, M3 and MG were used with slit
width for the top layer used was 0.6um and a slit separation of 0.8pm. It is evident
from the Table H that for the same wavelength Of light diffraction angle is wider
and separated which makes it easier for metal layers tO discriminate between
wavelengths. The effect of these extra metal layers on the selectivity Of filter for

IBM 0.13pm process has been demonstrated in Chapter 6.

TABLE II. DIFFRACTION ANGLE FOR EXCITATION AND EMISSION LIGHT FROM
FLUORPHORE FOR IBM 0.13pm CMFRFSSF PROCESS

 

 

 

 

 

 

 

Wavelength Diffraction angle
(3») (9°)
600nm (Excitation light) 25.38 °
610nm (minimum emitted light) 25.83 °
750nm (maximum emitted light) 32.39 °

 

5.4 Physical design

The Optical detection circuit was implemented on a l.5xl.5mm2 chip in a
2P, 3M 0.5um CMOS process, as shown in Figure 5.10. The diffraction grating
filter was implemented using M1, M2 and M3 layers within the active photodiode
area. Due to fabrication process restrictions vertical distance between the layers
cannot be changed. Similarly, the width of metal strips and the horizontal distance

between metal strips is also constrained by the minimum design rules for a specific

68

CMOS technology. TO permit experimental evaluation Of different filter design
parameters, the diffraction grating over each photodiode in the array utilize
different slit widths and slit separations. Layout for the readout and addressing
circuitry was done using only M1 and M2 layers. Readout circuits have been fully
covered by the tOp CMOS metal layer M3 blocking the light tO protect the circuits

from laser light that can change the properties of the circuits.

 

Figure 5.10. Cadence layout Of the l.5xl.5mm2 Optical detection chip.

5.5 Summary

An Optical detection system has been presented for fluorescence detection
with in-pixel readout. Intended application Of this system is characterization Of
biofluorescence assays. Main components of this system are photodiode, on-chip
filter and readout circuit. An n-well/p-sub photodiode is used tO absorb photons.
The in-pixel readout is based on a first order 2A ADC. A novel technique is used

for interference filter by exploiting the orientation Of metal layers in layout that

69

does not require any post-CMOS processing. Design of this optical detection
system was done to meet the demands of on-chip OptO-electrochemical sensor array.

The first prototype chip was designed in 2P, 3M 0.5pm CMOS process.

70

Chapter 6
Simulation and Experimental Results

The autonomous on-chip multimode voltammetry microsystem (Chapter 4)
and Optical detection circuit (Chapter 5) were developed to demonstrate the
feasibility of an Opto-electrochemical detection microsystem and to explore the
relevant analog/mixed signal IC design methodology for such a combined system
on a monolithic chip with different functionalities. Both circuits were fabricated in
AMI 0.5um C5N CMOS technology using a 3.3V power supply. This chapter
presents the simulation and experimental test results of the proposed design and
fabricated chips respectively and an analysis of their functionality and
performances.

6.1 Measurement of Multi-mode Voltammetry CMOS Microsystem

Figure 6.1 shows the die photograph of the 3mm X 3mm prototype
multimode voltammetry CMOS IC. A 2 x 2 array of working electrodes (WE), with
single counter electrode (CE) and reference electrode (RE) are integrated on IC for
on-chip biosensor measurements. Three main components of the microsystem i.e.
waveform generator, 2 x 2 potentiostat array and readout circuit were tested
individually.

6.1.1 Waveform generator
6.1.1.] Test Setup

To verify the functionality of programmable waveform generator, a test

setup comprised of the instrumentation chip, and a PC with a DAQ card Nl-USB

6259 and running a LabVIEW user interface was prepared as shown in Figure 6.2.

71

This LabVIEW software is used tO generate the digital control signals and also to

acquire data from the CMOS chip.

”I“ I‘t-‘I‘Lo- '. ‘:‘ i'fl'f‘i.‘ IJjJI‘I-."3';‘:'c"' . .

Amperometric
Readout Circuit

Pote ntiostat
Array

 

Waveform Generator

 

}o\o;0’-b'U-"O'p.‘r\\o h. 'v-th-I-bvlhst-‘S ~§.\ 90“"
I. O Q ' - .

Figure 6.1. Die photograph Of 3mm x 3mm multi-mode voltammetry system.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Control Si nals
comPUter NI-DAC g Waveform
(:3: 6259 e 0 generator We”
'r ‘_ utput , , su ]
/=E_—"\ waveform . CMOS pp y
unstmmentatronlfl.
[ I

 

 

 

Figure 6.2. Experimental test setup for the on-chip waveform generator.

6.1.1.2 Test results

Figure 6.3 shows simulated output signals from the waveform generator for
each Of the four implemented waveform modes. All Signals were generated under the
same parameters with a scan range Of 2V (max) and scan rate Of 4V/sec using a 2

kHz input master clock. This translates tO amplitude Of 2Vpk-pk and output

72

frequency of le. The characteristics of the waveform generator are summarized in

Table HI.

 

 

 

/

 

 

 

 

 

 

 

 

 

0
0 0.5 1-0 1-5 2-0 00 0.5 1.0 1.5 2.0
Time (sec) Time (see)
(a) (b)
H
2.5 1'17: _
22.0 g -
81.5 8.125 ‘
:3 a 1.0
61.0 a
> >075
0.5
0 0.5 _ _
0 0.5 .1-0 1-5 2-0 0 0.5 1.0 1.5 2.0
Time (390) Time (sec)
(C) (d)

Figure. 6.3. Simulation results for the waveform generator with time on the x-axis
and voltage amplitude on the y-axis. (a) triangular signal, (b) linear sweep, (c)
constant potential, (d) square pulse with increasing amplitude.

TABLE III. WAVEFORM GENERATOR CHARACTERISTICS

 

 

 

 

 

 

 

Area 0.44mm2
DAC Resolution 10 bit
Scan range 0- 2V pk-pk, lO-bits
Scan rate lmV/sec-400V/sec
Step size 2mV
Power consumption 0.15mW

 

 

 

73

6.1.2 Potentiostat
6.1.2.1 Test Setup

TO verify the electrochemical measurement capability Of potentiostat, the
test setup that was prepared, comprised of the instrumentation chip and PC with a
DAQ card (NI-6259) running the LabVIEW software. Initially the potentiostat
circuit was characterized without the electrochemical setup using an external
waveform generator (Tabor Electronics, Waveform Generator-WW5062).

For testing with electrochemical setup a typical electrolyte solution
consisting of 0.1M (molar) potassium ferrocyanide and 1M potassium chloride
buffer was prepared. Test setup for on-chip potentiostat with electrochemical test
setup is shown in Figure 6.4. Measurements were performed at 25°C and a scan rate
Of 100mV-s'l with a liquid junction Ag/AgCl reference electrode. First a
commercial instrument (CH Instruments 700C) was used to take a cyclic
voltammetry (CV) measurement with an on-chip WE. The instrumentation chip was
then connected tO an external electrochemical cell composed Of commercial
components, and a CV measurement was performed using on-chip electronics.

Input stimulus from

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

j 4 C21) 60:53 T Output
E ‘ 1......
L J
Computer running LABVIEW
and connected to DAC
Electrolytic
solution

Figure 6.4 Experimental test setup for on-chip potentiostat.

74

6.1.2.2 Test Results

A series of tests were performed to compare the functionality of the
commercial potentiostat with on-chip potentiostat.

Test results of the potentiostat with electrochemical test setup are shown in
Figure 6.5. Figure 6.5(a) Show the results Of cyclic voltammetry (CV) test using the
commercial potentiostat (CH Instruments 700C) at a temperature of 25°C with a
scan range of 0.4V and scan rate Of 100mV/sec, while Figure 6.5(b) shows the
result of on-chip potentiostat for a CV curve with oxidation-reduction peak
separation of 0.14V. This compares well with measurements of the same
electrochemical cell using commercial instrument, where peak separations of
0.1~0.15V were typically Observed. The x-axis values for Figures 6.5(a) and 6.5(b)

are different because of internal biasing of the CMOS potentiostat.

75

 

OPRRRNN
mmnmmAA

0.3 1
0 :
-0.3 f
-0.6 i
-0.9 1
-1.2 I
-1.5 -
-1.8
-2.1 r . , .

0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0

Current / le-SA

 

 

 

 

Potential / V

(a)

 

Normalized output (V)

 

 

 

 

0 I . i
O -O.5 -1.0 -1.5 -2.0 -2.5
Potential at WE (V)
(b)

Figure. 6.5 CV measurement of 0.1M potassium ferrocyanide (K403H20) for on-
chip electrode (a) using commercial instruments showing a peak separation of
~0. 10V, (b) using on-chip potentiostat showing a peak separation of ~O.14V.

76

6.1.3 Amperometric readout circuit

Amperometric readout circuit [115] was also tested to check the
functionality and performance of the circuit. This amperometric circuit was
designed and tested by fellow student Xiaowen Liu, at AMSaC lab, Michigan State
University. The circuit was tested using DAQ card (NI-6259) and PC running the
LabVIEW software to generate control signals for the circuit and also acquire the
output signal from the chip. Figure 6.6 Shows the LABVIEW test program
interface to test the functionality of amperometric readout circuit. A source meter
was used to sweep the working electrode (WE) current from luA to 9uA. As shown
in Figure 6.7 for a 100 kHz sampling clock, the circuit provides a linear output
voltage over a wide current range; it was found to perform correctly with inputs up
to lOOuA. Thus, the amperometric readout circuit supports about 180dB dynamic

range by changing the frequency of the sampling clock and the gain of the PGA.

‘4' 7 _

i1. ' -. -
gl-"i-i ,
E " _ ,
t .

L‘: ' -

‘1- .

5‘-

 

Fig 6.6 LABVIEW test program interface to test the functionality of amperometric
readout circuit

77

1uA_to_ 90A, 100KHz

 

 

 

 

 

 

9

751.4

C”13 ’*

m.

+1 0.

§I.2 -,

~11 're
3

.91

D

0 123456789

Input current (uA)
Figure 6.7 Current readout circuit results for input current from lpA to 9uA.

6.1.4 Discussion

The simulation results of the on-chip waveform generator and experimental
results from potentiostat and amperometric readout circuit demonstrate that the
circuits developed in this dissertation can be used to characterize membrane
proteins using a platform that uses multiple voltammetric techniques i.e. CV, LSV,
constant potential and normal pulse voltammetry (NPV). Also design of these
circuits is small enough to support a large array of such sensors with (or on top of) a

single microelectronic chip.

6.2 Optical detection system

Figure 6.8 shows the cadence layout of 1.5mm x 1.5mm CMOS optical
detection chip with on-chip filter. Readout circuit in the chip was implemented
using M1 and M2 layers and was later covered by M3 layer in the layout to protect
the circuit from the adverse effects Of laser light that may change the characteristics
of the readout circuit. Die photograph of the fabricated chip with readout circuit

covered with M3 layer is shown in Figure 6.9.

78

     

- I - ”V m
f -' . . g .
I I I I g . I! I I I I _ I
=m3§¥-0€fixmmmn
‘ I ~ I I u 'I"

I g
.. 1
Pixel f , I. l.—

photodiode

 

 

readout
circuit
row 81
column access
bias
output buffers generation

Figure 6.8 Cadence layout Of CMOS Optical detection IC.

 

Figure 6.9 Die photograph of CMOS Optical detection IC.

79

6.2.] Simulation results

The optical detection CMOS chip was simulated for different values of
photocurrent ranging from lOnA to lOOnA. Different test points were added in the
design to test different parameters of the chip. Integrator output and digital output
from sigma delta ADC were verified. Figure 6.10 shows the cadence simulation
results for a 70nA photocurrent with integrator output as well as digital output from
the latch. Width of the digital signal can be measured by using an off-chip digital
counter. A range of discharging time at the output of integrator for a range of
photocurrent from lOnA to lOOnA is shown in Figure 6.11.

W_:r_an_$l_el1t_resoonse at 70nA

 

 

A “If"? - w-” - _ _ _ ___,__.w-. _- ___ __. __ .._n

a 2.4”

8) 2'3; Input reference

g 1.8a

g 1.6~ m
3.5.

E 3.0.

a $2»

3 120‘.“ CLK

6 0.5-

> o. -

 

~ A_ La

_/ fi/ fi/ fi/
/'

Integrator output

L ‘<-— Output latch J J

- -~_ _ gum W _ g -59- W __WW60_ WWWW __ -70. 80
Time(usec)

 

 

 

 

 

 

 

 

 

 

Voltage (V) Voltage (V)
Peewpu b peeww

Figure 6.10. Cadence simulation of a single pixel for 70nA with integrator analog
output and digital output from the comparator.

80

Current vs Discharging time

 

110

 

 

 

 

 

 

 

l l l l l l ' l
100 _, e 1 _______ ‘5 ________ t _______ t _______ J: ________ E ________ —6— calculated time _
: : : : : : —*-- simualted time
90 -----i ----- «z -------- i ------- i ------- i -------- a -------- . ------- ~. -------- imm-
80--------§-~ -~' -------- -------- i -------- ------- -i ------ -
A 70 -------- i ----- -. ------- i ------- i ------- i -------- s -------- i ------- i -------- E ------- .
3 50 ........ i ....... . ....... i ........ s ........ i ....... ....... -
g : : : : : : : : :
‘5 so -------- ------- e ------- i ----- i -------- i -------- ------ -
o i l i l l l l l l
40 -------- i ------- 1 -------- t-----t -------------- : -------- i ------- ‘. -------- : ------- -
30 -------- i -------- E -------- E ------ E -------- 'i ------- i -------- ' ------- -
20 -------- i ------- i -------- i ------- i ------- -:- ------- i -------- i- ------ -
w -------- i ------- i -------- i ------- i ------- i -------- E -------- .~ ----- i ------ time--
i i i i i i 1 L 1
U 0.5 1 15 2 25 3 35 4 45 5
Time(usec)

Figure 6.11. Calculated and simulated time for the discharge of photocurrent using
single pixel sigma-delta ADC.

6.3 On-chip filter
6.3.] Simulation results

To test the functionality of on-chip filter, simulations were done in
MATLAB for different values of excitation and emission wavelength in the visible
range of light that can be detected by the photodiode. Figure 6.12 shows the
physical model used for the simulation that contains three metal layers with the
same parameters and values as dictated by the “design rule check (DRC)” rules for
CSN CMOS process. These simulations were done to determine the minimum
wavelength of incident light allowed to pass through the filter. Table IV shows the
minimum and maximum values used for the three layers used in simulation which
follow DRC rules of CSN process. These simulations were based on the same

diffraction grating equation (2) as discussed in chapter 4.

81

These simulations give a digital response to show the blocking of excitation
light and allowing the emission light. A plot of emitted wavelengths that were
allowed to pass through the diffraction grating from three adjacent slits has been
shown in Figure 6.13 where x-axis represent the normalized location on n-well
where light after passing through the filter enters into n-well to create electron-hole
pair (EHP), while y-axis shows the wavelength of incident light. Different colors
represent different wavelengths and data points when projected on the x-axis show
the location of transmitted light falling on n-well from a series of adjacent slits s], s;
and 53. These different locations for different wavelength of light are due to the
change in diffraction angle for different wavelengths of light. These simulations
reveal that using the above configuration a minimum emission light that can pass
through the designed configuration is 740nm, while the light with a wavelength less

than 740nm is blocked. Empty regions between emission light stripes show that

Um Incident light—l L U
\

._XJ_.
Optical length
(L)

   
    

 

Location of transmitted n-well
wavelength entering into n-well

 

Figure 6.12 Physical model used for optical simulation in MATLAB.

82

light is blocked by either M2 or M1 layer stripes. Figure 6.14 shows the digital
output from the above configuration showing the wavelength range of incident light
passing through it.

In addition, for other advanced CMOS process i.e. 0.18pm, 0.13 pm
processes which use different number of metal layers with different separation and
thickness of metal layers, this topology can be used as a band pass filter. Analysis
for the effect of extra metal layers on the selectivity of filter is shown in next

section.

TABLE IV. TYPES AND RANGE OF METAL LAYERS IN CsN CMOS PROCESS USED FOR

 

 

 

 

 

 

 

 

 

 

 

MATLAB SIMULATION
Metal layer Slit width Slit separation
M3 0.9Lm 1.5um
M2 0.9um (min) 0.9um (min)
1.31m (max) 1.2um (max)
Ml 1.2um (min) 0.9um(min)
1.5um (max) 1.2pm(max)
800 _
790 >- I I I I I I
A 775 t v v v v v ' Transmitted
:5: 765 wavelength
5 755 x I I I I I
8’
2 745 l- I e I e I O
a)
g 735 . . . . . .
725
715 _ Blocked
wavelength
705 -
0 1 2 3 4 5 6 7

location on incident light falling on n-well (um)

Figure 6.13. Locations of incident light of wavelength on n-well with
different colors showing the light coming from different adjacent slits.

83

Transmitted light

 

Blocked light

 

 

640 660 680 700 730 740 760 780 800
Wavelength (nm)

Figure 6.14. Range of incident light wavelength allowed to pass through the filter
Shown by a digital signal.

6.3.2 Effect of extra metal layers on the selectivity of filter

To study the effect of extra metal layers and advanced CMOS fabrication
processes on the selectivity of filter, simulations were done in MATLAB. As
advanced fabrication processes support more metal layers, these metal layers can be
manipulated to be used as band pass filter. Placements of these metal layers select
the required wavelength of light to pass through the filter. These Simulations were
done to determine the range of wavelength of incident light that can be allowed to
pass through the filter. For these simulations, same parameters and values i.e. metal
thickness, optical length and separation between metal layers as dictated by the
“design rule check (DRC)” and fabrication rules for IBM 0.13pm CMRFSSF
process. Table.V shows the range of values for different metal layers used for
simulations. As these values for IBM 0.13pm process are different as compared to
CSN process, so these metal gratings are expected to have different results using the

same design and configuration. Although 8 metal layers are allowed in the IBM

84

0.13pm CMRFBSF process, only 4 metal layers were used i.e. M1, M2, M3, MG to
make the analysis simple. Simulations were done for 3 adjacent slits S], 82 and s3
with the same size and equally spaced. Figure 6.15 Shows different locations on the
n-well (x-axis) fora range of transmitted wavelength through the filter while y-axis
shows the wavelength range used in simulation. These simulations were done for
visible light (320nm-780nm) and it was observed that light with a wavelength range
of 400nm - 530nm is allowed to pass through the filter confirming the use of filter
as band pass filter as shown in Figure 6.16. Similarly addition of extra metal layers
and their placement can be varied to change the allowed wavelength range to pass
through filter and also increase the selectivity by narrowing the wavelength range of

allowed light.

TABLE .V. TYPES AND RANGE OF METAL LAYERS USED IN IBM 0.13pM PROCESS FOR

 

 

 

 

 

 

 

MATLAB SIMULATION
Metal layer Slit width Slit separation
MG 0.6pm (min) 0.4um (min)
1pm (max) 0.8pm (max)
M3 0.6pm(min) 0.4um (min)
1pm(max) 0.8um(max)
M2 0.4um(min) 0.4pm(min)
0.6pm(max) 0.8um(max)
M1 0.4um 1pm

 

 

 

85

 

0.15“ t "is é 2‘5 5 3.15 44
location of incident light on n-well (um)

Blocked
wavelength

Transmitted
wavelength

Blocked
wavelength

Figure 6.15. Locations of incident light of wavelength on n-well with different
colors Showing the light coming from different adjacent slits.

Transmitted light

l

 

Output

Blocked light Blocked light

 

 

AAAAAALL‘A AAAAAAAAAL

300 350 400 450 500 550 600 650
Wavelength (nm)

 

Figure 6.16. Range of incident light wavelength allowed to pass through the filter

Shown by a digital signal.

6.3.3 Process variations

As discussed in chapter 5 process variations also have an impact on the

performance of filter, as design is inherently dependent on the fabrication process

and also limited by the DRC rules. T 0 determine the effect of process variations on

the filter, filter design was simulated in MATLAB with different physical

86

configurations that were implemented in the CMOS chip for CSN process. In the
proposed filter design for CSN process, top M3 transmission layer and M2 blocking
layer have critical role in the functionality and performance of filter. So initially
only M3 and M2 layer were considered for the Simulation with M1 having constant
slit width(x3) of 2.4um. A process variation of 15% was assumed for both M3 and
M2 layers for the analysis. Two different cases were studied for the process
variation in M3 (a) and M2 (b) metal layers. These process variations are
underdevelopment (case 1) and overdevelopment (case 2) of both M3 and M2 layer.
Table VI shows different cases that were analyzed and their effect on the light
wavelength range allowed to pass through the filter. These simulations were done
for a maximum wavelength of 800nm to verify the response of filter for visible light

(320nm-780nm).

TABLE VI. DIFFERENT CASES FOR PROCESS VARIATIONS IN METAL WIDTHS FOR C5N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PROCESS
Cases M3 layer M2 layer Minimum
width width transmitted
wavelength
a - 0.225pm b = 1.5um no light
Case 1 a - 0.225um b +0 .225 pm no light
(under-
development) a - 0.225um b - 0.225 um 775nm
a - 0.225pm b = 1.2 pm 740nm
a - 0.225um b + 0.18 pm no light
a - 0.225um b - 0.18 pm 635nm
a + 0.225um b = 1.5um 800nm
5388912 (over; a + 0.225um b +0 .225 um no light
eve om“ ) a + 0.225um b - 0.225 um 680nm
a + 0.225um b = 1.2 pm 735nm
a + 0.225um b + 0.18 pm no light
a + 0.225pm b - 0.18 pm 630nm

 

87

 

6.3.4 Discussion

The simulated plot in Figure 6.11 Show that a large percentage of visible
light range is blocked using the configuration in Figure 6.10 aS compared to the
calculated values in Chapter 5. Main reasons for this blockage of incident light is
due to the thickness of M3 layer that blocks maximum amount of light due to large
dispersion angle for smaller wavelengths of light. This M3 layer thickness is
comparable with the vertical distance between M3 and M2 layer i.e. y]. As a result
there is a difference in allowing the emission wavelength range between the
calculated and Simulated value, as diffraction grating equation does not include the
slit thickness into account by assuming the slit thickness to be negligible as
compared to slit width and slit separation.

There is a trade off between the slit width for M3 layer and M2 blocking
layer width. Larger slit width allows a large range of wavelength of light to pass
through it without being blocked by M3 layer thickness but at the same time it
decreases the angle of dispersion. This decrease in dispersion angle decreases the
selectivity of the filter as it becomes difficult for M2 layer to only block the
excitation light and some wavelength of emission light is also blocked due to
smaller angle deviation between lights of different wavelength.

Simulation results in Figure 6.13 Showthat addition of metal layers in the
design increases the selectivity of the filter and can be manipulated to use it as band
pass filter. Also the transmitted wavelength for both designs are different i.e. CSN

and IBM 0.13pm process, because of different DRC rules and also because of

88

different thickness of metal layers that has a significant role in order to allow/block
the incident light.

Process variations also have a prominent effect on the performance of filter.
As Shown in Table VI, where different cases are Shown for process variation in both
M3 and M2 layers. It is evident from the table that both M3 and M2 layer variation
have an effect on the transmission wavelength. Another interesting fact that is given
from the above table is that for both underdevelopment of M2 layer has a positive
impact on the transmission wavelength for both underdevelopment and
overdevelopment of M3 layer. So minimum width of M2 layer used in the design
results in better efficiency and performance of filter.

The above analysis also shows the effect of using advanced CMOS process
on the efficiency and performance of the proposed filter. Advanced CMOS
processes use smaller feature size and extra metal layers with smaller process
variations that enhances the performance of filter in terms of flexibility and
selectivity. Simulation plots in Figures 6.14 and 6.15 Show that, for IBM 0.13pm
process, smaller feature Size and more metal layers with less metal thickness allows
for more wavelength transmission due to more freedom of design in Slit width and
Slit separation and also increases the selectivity of filter as compared to the filter
implemented in CSN process. Also filter implemented in IBM 0.13pm process can
act as a band-pass filter as compared to C5N process that only acts as a low pass
filter in visible range of light. So it is concluded that the filter designed in advanced
CMOS processes i.e. 0.35pm, 0.18pm and 0.13pm would exhibit better

performance as compared to the filter designed in 0.5um CMOS process.

89

2.2 Conclusion

The feasibility of the VLSI realization of a programmable waveform
generator and potentiostat with amperometric readout circuit with the capability of
integrating on-chip electrodes is critical for the success of a multi-mode
voltammetry biosensor array microsystem. In this chapter, the silicon realization of
these two critical blocks i.e. waveform generator and potentiostat were
characterized individually and verified with a series of experiments. Similarly
realization of an optical readout circuit with integrated on-chip filter that requires
no post-CMOS processing was demonstrated. Hand calculations were done for the
proof of concept in utilizing the metal layers in layout as interference filter to allow
specific wavelengths of light to pass through it. These calculations were verified
with the simulations in MATLAB. Also it was demonstrated through simulations
the effect of advanced CMOS technology on the performance and selectivity of the

filter.

90

Chapter 7
Summary and Future Work

7.1 Summary of the contributions

Optical and electrochemical detection are considered to be the most widely
used interrogation techniques for miniaturized biosensor arrays. Simultaneous
implementation of these two techniques to determine multiple biomembrane
processes is highly advantageous and their VLSI realization on a single chip would
result in a versatile instrumentation system offering more accurate analysis and high
throughput. But at the same time realization of this integration also face many
challenges in terms of circuit design, fabrication of on-chip electrodes.
This dissertation provides a detailed analysis of the challenges that need to be
resolved to develop an efficient on-chip instrumentation system that combines both
these techniques to characterize bilipid membranes (BLP). Two separate
instrumentation system for electrochemical and optical detection were developed
and fabricated in CSN CMOS process. The results of this research provide a solid
basis for future research in opto-electrochemical detection systems for biosensor
arrays.
7.1.] Analysis of challenges for opto-electrochemical detection system

A thorough analysis of a combined opto-electrochemical detection system to
characterize membrane proteins has been done. Main challenges in the electrical
design to readout both optical and electrochemical information and interference
between these circuits has been discussed. Similarly issues pertaining to the
integration of on-chip electrodes for electrochemical detection, filters for optical

detection have been discussed. This analysis forms a solid basis to design a

91

combined on-chip opto-electrochemical detection system for miniaturized
biosensors arrays.
7. I .2 CMOS microsystem for multimode voltammetry

An on-chip instrumentation system that supports a variety of voltammetric
techniques i.e. cyclic voltammetry, linear sweep voltammetry, constant potential
and normal pulse voltammetry with different scan ranges and scan rates has been
developed. These techniques require different kinds of waveforms i.e. triangular
waveform, linear ramp, constant potential and differential waveform for their
operation. A digitally programmable waveform generator was developed to
generate these waveforms. A 2 x 2 potentiostat array with amperometric readout
circuit was also developed having the capability of integrating electrodes on CMOS
chip.
7.1.3 Optical detection circuit

In this dissertation an optical detection circuit was also developed keeping in
view the requirements of a combined opto-electrochemical detection system. A 4 x
3 pixel array was developed with n-well/p-sub photodiode and in-pixel readout
based on Sigma-delta (2A) ADC. A novel technique was used for the on-chip filter
that does not need any post-CMOS processing and does not provide any hindrance
to electrochemical detection. The placement, width and horizontal gap between the
CMOS metal layers were manipulated to create a multislit diffraction grating that

passes the emission light and blocks the excitation light from the light source.

92

7.2 Future Work

Based on the results of this dissertation, the following suggestions for future
research are made.

Although separate optical and electrochemical detection system have been
developed keeping in view the requirements of combined opto-electrochemical
detection system, VLSI realization of such a combined system on a monolithic chip
is still a challenge. So the next step is the integration of both these techniques on a
monolithic chip.

In terms of circuitry, a potential challenge for a combined opto-
electrochemical system is the interference between the two readout circuits. As both
these techniques are carried out at different frequencies, readout circuits for these
techniques have different issues and need to work independently.

Filter performance for fluorescence detection needs to be improved for high
selectivity and accuracy. Other techniques for filter design can also be considered
for possible options to develop opto-electrochemical detection system.

Other electrochemical detection schemes i.e. impedance spectroscopy can

also be integrated with the optical circuit for a combined instrumentation system.

93

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[3]

[9]

[10]

[11]

Bibliography

K. E. Gonsalves, C. L. Laurencin, C. R. Halberstadt, and L. S. Nair,
Biomedical Nanostructures. NewYork: John Wiley & Sons Inc November
2007.

Vo-Dinh. T and B. Cullum, "Biosensors and Biochips: Advances in
Biological and Medical Diagnostics," Fresenius J Anal Chem, vol. 366,
pp. 540-551, 2000.

E. Gizeli and C. R. Lowe, "Immunosensors," Current Opinion in
Biotechnology, vol. 7, pp. 66-79, 1996.

D. J. Monk and D. R. Walt, "Optical fiber-based biosensors," Analytical
and Bioanalytical Chemistry, vol. 379, pp. 931-945, 2004.

T.Kubick, K. Bogunia-Kubik, and M. Sugisaka, "Nanotechnology on Duty
in Medical Applications," Current Pharmaceutical Biotechnology, vol. 6,
pp. 17-33, 2005.

S.Kohler, S.Belkin, and R.D.Schmid, "Reporter gene bioassays in
environmental analysis," Fresenius J Anal Chem, vol. 366, pp. 769-779,
2000.

A.Reininger-Mack, H.Thielecke, and A.A.Robitzki, "3D-biohybrid
systems: applications in drug screening," Trends in biotechnology vol. 20,
pp. 56-61, 2002.

W.Zhang, G.Canziani, C.Plugariu, R.Wyatt, J .Sodroski, R.Sweet,
P.Kwong, W.Hendrickson, and I.Chaiken, "Conformational changes of
gplZO in epitopes near the CCR5 binding site are induced by CD4 and a
CD4 miniprotein mimetic," Biochemistry, vol. 29, pp. 9405-16, 1999.

K.Yano and I.Karube, "Molecularly imprinted polymers for biosensor
applications," Trends Anal. Chem, vol. 18, pp. 199-204, 1999.

G. J. Cotton, B.Ayres, R.Xu, and T.W.Muir, "Insertion of a synthetic
peptide into a recombinant protein framework: a protein biosensor "
Journal of the American Chemical Society vol. 121, pp. 1100-1101 1999.

R. F. Costello, I. R. Peterson, J. Heptinstall, and D. J. Walton, "Improved

gel-protected bilayers," Biosensors & Bioelectronics, vol. 14, pp. 265-271,
1999.

94

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

J. J. Rarnsden, "Biomimetic protein immobilization using lipid bilayers,"
Biosensors & Bioelectronics, vol. 13, pp. 593-598, 1998.

J. M. Kohler, A. Csaki, J. Reichert, R. Moller, W. Straube, and W.
Fritzsche, "Selective labeling of oligonucleotide monolayers by metallic

nanobeads for fast optical readout of DNA-chips," Sensors and Actuators
B Chem, vol. 76, pp. 166-172, 2001.

S.S.Pathak and H.F.Savelkoul, "Biosensors in immunology: the story so
far," Immunology Today, vol. 18, pp. 464-467, 1997.

L. D. Stefano, P. Arcari, A. Lamberti, C. Sanges, L. Rotiroti, I. Rea, and I.
Rendina, "DNA Optical Detection Based on Porous Silicon
Technologyzfrom Biosensors to Biochips," Sensors vol. 7, pp. 214-221,
2007.

T. Vo-Dinh, "SERS chemical sensors and biosensors: new tools for
environmental and biological analysis," Sensors and Actuators B Chem,

vol. 29, pp. 183-189, 1995.

J. Dostalek and W. Knoll, "Biosensors based on surface plasmon-
enhanced fluorescence spectroscopy," Biointerphases, vol. 3, pp. 12-22,
2008.

V. Sanz, S. d. Marcos, and J. Galban, "A reagentless optical biosensor
based on the intrinsic absorption properties of peroxidase," Biosensors &
Bioelectronics, vol. 22, pp. 956-964, 2007.

J. Zhang, Y. Huang, N. Trombly, C. Yang, and A. Mason,
"Electrochemical Array Microsystem with Integrated Potentiostat,"
presented at IEEE Int.Conf. on Sensors, Irving CA, 2005.

J. Wang, X. J. Zhang, and M. Prakash, "Glucose microsensors based on
carbon paste enzyme electrodes modified with cupric hexacyanoferrate,"
Analytica Chimica Acta, vol. 395, pp. 11-16, 1999.

J. Kulys, "The carbon paste electrode encrusted with a microreactor as
glucose biosensor," Biosensors & Bioelectronics, vol. 14, pp. 473-479,
1999.

V. G. Andreou and D. P. Nikolelis, "Electrochemical transduction of
interactions of aflatoxin M1 with bilayer lipid membranes (BLMS) for

the construction of one-Shot sensors " Sensors and Actuators B Chem,
vol. 41. Pp. 213-216, 1997.

95

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

C. Yang, Y. Huang, B. L. Hassler, R. M. Worden, and A. J. Mason,
"Amperometric Electrochemical Microsystem for a Miniaturized Protein
Biosensor Array," presented at IEEE Biomedical Circ. Systems, 2009.

M. Breten, T. Lehmann, and B. Braun, "Integrating data converters for
picoampere currents from electrochemical transducers," presented at IEEE
International Symposium on Circuits and Systems, Geneva, 2000.

A. Hassibi and T. H. Lee, "A Programmable Electrochemical Biosensor

Array in 0.18pm Standard CMOS," presented at International Symp.
Solid-State Circuits, 2005.

P. Kim, N. Kohli, B. Hassler, N. Dotson, A. Mason, R. M. Worden, and R.
Ofoli, "An Electrochemical Interface for Integrated Biosensors," presented
at IEEE Int. Conf. on Sensors, Toronto, Canada, 2003.

C. Yang, Y. Huang, B. L. Hassler, R. M. Worden, and A. J. Mason,
"Amperometric Electrochemical Microsystem for a Miniaturized Protein
Biosensor Array," IEEE Trans. Biomedical Circ. Systems, vol. 3, pp. 160-
168, 2009. ‘

X. Yang, C. Shi, R. E. Holmlin, S. O. Kelley, M. G. Hill, D. M. Crothers,
and J. K. Barton, "Electrochemical Microsensors for DNA Mutation
Detection," Digest Solid-State Sensor, Actuator and Microsystems
Workshop, pp. 118-121, 2002.

E. Bakker, and M. Telting-Diaz, "Electrochemical Sensors," Anal. Chem,
vol. 74, pp. 2781-2800, 2002.

Y.-H. Zheng, T.-C. Hua, and F. Xu, "A Thermal Biosensor Based on
Enzyme Reaction," presented at [BEE-Engineering in Medicine and
Biology(EMBS), Shanghai,China, 2005.

J. E. Kohn and K. W. Plaxco, "Engineering a Signal transduction
mechanism for protein-based biosensors," presented at Proc of National
Academy of Sciences(PN AS), 2005.

S. Song, Hui Xu, and C. Fan, "Potential diagnostic applications of
biosensors: current and future directions," International Journal of
Nanomedicine, vol. 1, pp. 433-440, 2006.

G. S. Bang, S. Cho, and B. G. Kim, "A novel electrochemical detection

method for aptamer biosensors," Biosensors & Bioelectronics, vol. 21,
pp. 863-870, 2005.

96

[34]

[35]

[36]

[37]

[38]

{39]

[40]

[41]

[42]

[43]

[44]

K. Ikebukuro, C. Kiyohara, and K. Sode, "Novel electrochemical sensor
system for protein using the aptamers in sandwich manner," Biosensors &
Bioelectronics, vol. 20, pp. 2168-2172, 2005.

A. E. Radi, J. L. A. Sanchez, E. Baldrich, and C. K. O'Sullivan, "Reusable
impedimetric aptasensor," Analytical Chemistry(ACS), vol. 77, pp. 6320-
6323, 2005.

K. Maehashi, T. Katsura, K. Kerman, Y. Takamura, K. Matsumoto, and E.
Tarniya, "Label-free protein biosensor based on aptamer-modified
carbon nanotube field-effect transistors," Analytical Chemistry(ACS),
vol. 79, pp. 782-787, 2007.

J. A. Hansen, J. Wang, A. N. Kawde, Y. Xiang, K. V. Gothelf, and G.
Collins, "Quantum-dot/aptamer-based ultrasensitive multi-analyte
electrochemical biosensor," Journal of the American Chemical
Society(JACS), vol. 128, pp. 2228-2229, 2006.

R. Turpeinen, M. Virta, and M. Haggblom, "Analysis of Arsenic
Bioavailability in Contaminated Soils," Environmental Toxilogy and
Chemistry( ET &C ), vol. 22, pp. 1-6, 2003.

G. Marrazza, I. Chianella, and M. Mascini, "Disposable DNA
electrochemical sensor for hybridization detection " Biosensors &
Bioelectronics, vol. 14, pp. 43-51, 1999.

J. Wang, X. Cai, G. Rivas, H. Shiraishi, and N. Dortha, "Nucleic acid
immobilization, recognition and detection at chronoamperometric DNA
chips," Biosensors & Bioelectronics, vol. 12, pp. 587-599, 1997.

M. P. Marco and D.Barcelo, "Environmental applications of analytical
biosensros," Meas. Sci. Tech, vol. 7, pp. 1547-1562, 1996.

P. Yin, C. J. Burns, P. D. J. Osman, and B. A. Cornell, "A tethered bilayer
sensor containing alamethicin channels and its detection of amiloride

based inhibitors," Biosensors & Bioelectronics, vol. 18, pp. 389-397,
2003.

M. Trojanowicz, "Miniaturized Biochemical Sensing Devices Based on
Planar Bilayer Lipid Membranes," Fresenius J Anal Chem, vol. 371, pp.
246-260, 2001.

C. Yang, D. Rairigh, and A. Mason, "On-Chip Electrochemical Impedance

Spectroscopy for Biosensor Arrays," presented at IEEE Int. Conf. on
Sensors, Daejeon, Korea, 2006.

97

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

B. Jang and A. Hassibi, "Biosensor systems in standard CMOS processes:
Fact or fiction?," presented at IEEE Symp on Industrial Electronics(ISIE),
2008.

M. Schienle, C. Paulus, A. Frey, F. Hofmann, B. Holzapfl, P. Schindler-
Bauer, and R. Thewes, "A Fully Electronic DNA Sensor with 128
Positions and In-Pixel AID Conversion," J. Solid-State Circuits, vol. 39,
pp. 2438-2445, 2004.

J. Ohta, Smart CMOS Image Sensors and Applications: Barnes & Noble,
Sep 2007.

C. Yi, Q. Zhang, C.-W. Li, Y. Yang, J. Zhao, and M. Yang, "Optical and
electrochemical detection techniques for cell-based microfluidic systems,"
Analytical and Bioanalytical Chemistry vol. 384, pp. 1259-1268, 2006.

P. Abbyad, W. Childs, X. Shi, and S. G. Boxer, "Dynamic Stokes shift in
green fluorescent protein variants," National Academy of Sciences(PNAS),
vol. 104. pp. 20189-20194, 2007.

A. Bandyopadhyay, G. Mulliken, G. Cauwenberghs, N. Thakor, "VLSI
Potentiostat for Distributed Electrocherrrical Neural Recording," IEEE Int.
Sym. on Circ. and Systems, vol. 2, pp. 740-743, 2002.

A. Frey, M. Jenkner, M. Schienle, C. Paulus, B. Holzapfl, P. Schindler-
Bauer, F. Hofmann, D. Kuhlmeier, J. Krause, J. Albers, W. Gumbrecht, D.
Schmitt-Landsiedel, and R. Thewes, "Design of an integrated potentiostat
circuit for CMOS bio sensor chips," presented at IEEE International
SymposiumCircuits and Systems, Bangkok, Thailand, 2003.

R. Genov, M. Stanacevic, M. Naware, G. Cauwenberghs, and N. Thakor,
"VLSI Multi-Channel Track-and-Hold Potentiostat," presented at
Microtechnologies for the New Millennium 2003, Proc. SPIE, 2003.

R. G. Kakerow, H. Kappert, E. Spiegel, and Y. Manoli, "Low-Power
Single-Chip CMOS Potentiostat," Int. Conf. Solid-State Sensors and
Actuators (Transducers), vol. 1, pp. 142-145, 1995.

E. Katz, 1. Willner, "Probing Biomolecular Interactions at Conductive and
Semiconductive Surfaces by Impedance Spectroscopy: Routes to
Irnpedimetric Irnmunosensors, DNA-Sensors, and Enzyme Biosensors,"
Electroanalysis, vol. 15, pp. 913-947, 2003.

98

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

S. M. Martin, F. H. Gebara, T. D. Strong, and R. B. Brown, "A Low-
Voltage, Chemical Sensor Interface for Systems-On-Chip: The Fully-
Differential Potentiostat," presented at IEEE Int. Symp. Circuits and
Systems, Vancouver, Canada, 2004.

R. Matos, L. Angnes, and C. L. do Lago, ”Development of a four-channel
potentiostat for amperometric detection in flow injection analysis,"
Instrumentation Science & Technology, vol. 26, pp. 451-459, 1998.

H. S. Narula and J. G. Harris, "Integrated VLSI Potentiostat for Cyclic
Voltarnetry in Electrolytic Reactions," presented at IEEE Computer
Society Symp. on VLSI, 2004.

M. Naware, A. Rege, R. Genov, M. Stanacevic, G. Cauwenberghs, and N.
Thakor, "Integrated Multi-Electrode Fluidic Nitric-Oxide Sensor and
VLSI Potentiostat Array," presented at IEEE Int. Symp. Circuits and
Systems, Vancouver, Canada, 2004.

R. J. Reay, S.P. Kounaves, and G.T.A. Kovacs, "An Integrated CMOS
Potentiostat for Miniaturized Electroanalytical Instrumentation," Proc.
Inter. Solid-State Circuits Confi, pp. 162 - 163, 1994.

R. F. B. Turner, D. J. Harrison, and H. P. Baltes, "A CMOS Potentiostat
for Amperometric Chemical Sensors," IEEE J.Solid-State Circuits, vol.
22, PP. 473-478, 1987.

Ratnakumar. B. V, Smart. M. C, and S. S, "Electrochemical Impedance
Spectroscopy and Its Applications to Lithium Ion Cells," presented at
Annual Battery Conf. Appl. Adv, CA, USA, 2002.

R. Sanders and H. C. Gererritsen, "Fluorescence lifetime imaging of free
calcium in single cells," Bioimaging, pp. 1331-138, 1994.

D. Elson, S. Webb, J. Siegel, K. Suhling, D. Davis, J. Lever, D. Phillips,
A. Wallace, and P. French, "Biomedical applications of fluorescence
lifetime imaging," Opt. Photon. News, pp. 27-32, 2002.

K. Dowling, M. J. Dayel, S. Hyde, P. French, M. J. Lever, and J. D. Hares,
"High resolution time-domain fluorescence lifetime imaging for

biomedical applications," Journal of Modern Optics, vol. 2, pp. 199-209,
1999.

M. Elangovan, R. N. Day, and A. Periasamy, "Nanosecond fluorescence
resonance energy transfer-fluorescence lifetime imaging microscopy to
localize the protein interactions in a Single living cell," Journal of
Microscopy, vol. 205, pp. 3-14, 2002.

99

[66]

[67]

[68]

[69]

[70]

[71]

[72]

[73]

[74]

[75]

J. R. Lakowicz, "Emerging Biomedical Applications of Time-Resolved
Fluorescence Spectroscopy " in Topics in Fluorescence Spectroscopy, vol.
4: Springer US, 2006, pp. 1-19.

C. Dongre, R. Dekker, H. J. W. M. Hoekstra, D. Nolli, R. M-Vazquez, R.
Osellame, G. Cerullo, G. A. J. Besselink, and M. Pollnau, "Fluorescence
sensing with femtosecond laser written waveguides in a capillary
electrophoresis chip for monitoring molecular separation," presented at
European Conference On Integrated Optics (ECIO), Eindhoven, The
Netherlands, 2008.

L. Brand, C. Eggeling, C. Zander, K. H. Drexhage, and C. A. M. Seidel,
"Single-Molecule Identification of Coumarin-120 by Time-Resolved
Fluorescence Detection: Comparison of One- and Two-Photon Excitation
in Solution," The Journal of Physical Chemistry, vol. 101, pp. 4313-4321,
1997.

J. Malicka, I. Gryczynski, B. P. Maliwal, J. Fang, and J. R. Lakowicz,
"Fluorescence Spectral Properties of Cyanine Dye Labeled DNA Near
Metallic Silver Particles," Analytical Biochemistry, vol. 317, pp. 136-146,
2003.

C. Shi, H. Gu, and C. Ma, "An aptamer-based fluorescent biosensor for
potassium ion detection using a pyrene-labeled molecular beacon "
Analytical Biochemistry, vol. 400, pp. 99-102, 2010.

H. Harrna, T. Soukka, and T. Lovgren, "Europium Nanoparticles and
Time-resolved Fluorescence for Ultrasensitive Detection of Prostate-
specific Antigen " Clinical Chemistry, vol. 47, pp. 561-568, 2001.

Soares. A and Perry. R, "A Two-Color CMOS Optical Detector Circuit,"
presented at Proceedings of IEEE Southeast Conf, 2007.

Ardeshirpour.Y, Deem.M.J, and Shirani.S, "2-D CMONS Based Image
Sensor System for Fluorescent Detection," presented at Proceedings of
Canadian Conference on Electrical and Computer Engineer, 2004.

Darain. F, Yager. P, Gan. K. L, and S. C. Tjin, "On-chip detection of
myoglobin based on fluorescence," Biosensors & Bioelectronics, vol. 24,
pp. 1744-1750, 2009.

J. L. Fu, Q. Fang, T. Zhang, X. H. Jin, and Z. L. Fang, "Laser-Induced
Fluorescence Detection System for Microfluidic Chips Based on an
Orthogonal Optical Arrangement," Analytical Chemistry(ACS), vol. 78,
pp. 3827-3834, 2006.

100

 

[76]

[77]

[78]

[79]

[80]

[81]

[82]

[83]

[84]

Zourob.M, Mohr.S, Brown.B.J.T, Fielden.P.R, McDonnell.M.B, and
Goddard.N.J, "Bacteria detection using disposable optical leaky

waveguide sensors " Biosensors & Bioelectronics, vol. 21, pp. 293-302,
2002.

M. Beiderrnan, T. Tam, A. Fish, G. A. Jullien, and O. Yadid-Pecht, "A
Low-Light CMOS Contact Irnager With an Emission Filter for Biosensing
Applications," presented at IEEE Biomedical Circ. Systems, 2008.

T. Huang, S. Sorgenfrei, K. L. Shepard, P. Gong, and R. Levicky, "A
CMOS Array Sensor for Sub-800-ps Time-Resolved Fluorescence

Detection," presented at IEEE Custom Intergrated Circuits Conference
2007 .

B. Jang, P. Cao, A. Chevalier, A. Ellington, and A. Hassibi, "A CMOS
Fluorescent-Based Biosensor Microarray," presented at IEEE International
Solid-State Circuits Conference 2009.

M. L. Simpson, G. S. Sayler, G. Patterson, D. E. Nivens, E. K. Bolton, J.
M. Rochelle, J. C. Amott, B. N. Applegate, S. Ripp, and M. A. Guillom,
"An integrated CMOS rnicroluminometer for low-level luminescence
sensing in the biolurrrinescent bioreporter intergated circuit," Sensors and

Actuators, pp. 134-140, 2001.

R. A. Yotter and D. M. Wilson, "A Review of Photodetectors for Sensing
Light-Emitting Reporters in Biological Systems," IEEE SENSORS
JOURNAL, vol. 3, pp. 60, 2003.

C. Niclass, M. Sergio, and E. Charbon, "A single photon Avalanche Diode
Array fabricated in Deep-submicron CMOS Technology," presented at
Proc of Design, Automation and Test in Europe, 2006.

T.-C. D. Huang, S. Sorgenfrei, P. Gong, R. Levicky, and K. L. Shepard,
"A 0.18- m CMOS Array Sensor for Integrated Time-Resolved
Fluorescence Detection," IEEE Journal of Solid State Circuits (JSSC), vol.
44, pp. 1644-1654, 2009.

S. Yoshihara, Y. Nitta, M. Kikuchi, K. Koseki, Y. Ito, Y. Inada, S.
Kuramochi, H. Wakabayashi, M. Okano, H. Kuriyama, J. Inutsuka, A.
Tajima, T. Nakajima, Y. Kudoh, F. Koga, Y. Kasagi, S. Watanabe, and T.
Nomoto, "A 1/1.8-inch 6.4 MPixel 60 frames/s CMOS Image Sensor With
Seamless Mode Change," IEEE J. Solid State Circuits, vol. 41, pp. 2998-
3006, 2006.

101

[85]

[86]

[87]

[88]

[89]

[90]

[91]

[92]

[93]

[94]

[95]

M. Mori, M. Katsuno, S. Kasuga, T. Murata, and T. Yamaguchi, "A 1/4in
2M Pixel CMOS Image Sensor with 1.75 Transistor/Pixel," presented at
IEEE International Solid-State Circuits Conference, 2004.

D. Yang, A. E. Gamal, B. Fowler, and H. Tian, "A 640 x 512 CMOS
Image Sensor with Ultrawide Dynamic Range Floating-Point Pixel-Level
ADC," IEEE J. Solid State Circuits, vol. 34, pp. 1821-1834, 1999.

Chediak. J. A, L. 2., and S. J, "Heterogeneous integration of CdS filters
with GaN LEDS for fluorescence detection microsystems," Sensor and
Actuators A, 2003.

Webster. J. R, Burns. M. A, Burke. D. T, and M. C. H, "Electrocphoresis
system with integrtaed on-chip fluorescence detection," presented at IEEE
MEMS Miyazaki, Japan, 2000.

Shuai. Y, Banerjee. A, Klotzkin. D, and P. I, "On-chip Fluorescence
Detection using Organic Thin Film Devices for a Disposable Lab-on-a-
chip," presented at [BBB Sensors, 2008.

Thrush.E, Levi.O, Cook.L.J, Deich.J, Kurtz.A, Smith.S.J, Moerner.W.E,
and H. Jr.J.S, "Monolithically integrated semiconductor fluorescence

sensor for microfluidic applications," Sensors and Actuators B Chem, vol.
105. PP. 393-399, 2005.

P. Argue], J. Valentin, S. Fourment, F. Lzoes-Dupuy, G. Sarrabayrouse, S.
Bonnefont, Y. Jourlin, S. Reynaud, N. Destouches, A.V. Tishchenko, and
J. Jay, "A monolithic optical phase-shift detector on Silicon," IEEE
Sensors, vol. 5, pp. 1305-1309, 2005.

Dai. C-L, Chen. H—L, Yu. L-B, Lin. C-H, and Chang. P-Z, "Design and
fabrication of CMOS optical modulator " Sensors and Actuators A, vol.
95, pp. 69-74, 2001.

P. B. Catrysse, W. Suh, 8. Fan, and M. Peeters, "One-mode model for
patterned metal layers inside integrated color pixels," Optics letters, vol.
29, pp. 974-976, 2004.

G. Patounakis, K. L. Shepard, and R. Levicky, "Active CMOS Array
Sensor for Time-Resolved Fluorescence Detection," IEEE J. Solid State
Circuits, vol. 41, pp. 2521-2530, 2006.

Ta—chien, D. Huang, S. Sorgenfrei, P. Gong, R. Levicky, and K. L.
Shepard, "A 0.18-um CMOS Array Sensor for Integrated Time-Resolved
Fluorescence Detection," IEEE J. Solid State Circuits, vol. 44, pp. 1644-
1654, 2009.

102

[96]

[97]

[98]

[99]

[100]

[101]

[102]

[103]

[104]

[105]

B. R. Rae, K. R. Muir, D. Renshaw, R. K. Henderson, J. Girkin, Z. Gong,
J. McKendry, E. Gu, and M. D. Dawson, "A Vertically Integrated CMOS
Micro-System for Time- Resolved Fluorescence Analysis," presented at
IEEE Biomedical Circ. Systems, Beijing, 2009.

J. A. Richardson, L. A. Grant, and R. K. Henderson, "Low Dark Count
Single-Photon Avalanche Diode Structure Compatible With Standard
Nanometer Scale CMOS Technology," IEEE Photonics Technology
Letters, vol. 21, pp. 1020-1022, 2009.

D. Stoppa, D. Mosconi, L. Pancheri, and L. Gonzo, "Single-Photon
Avalanche Diode CMOS Sensor for Time-Resolved Fluorescence
Measurements," IEEE Sensors journal, vol. 9, pp. 1084-1090, 2009.

Mason.A, H. Yue, Y. Chao, and Z. J, "Amperometric Readout and
Electrode Array Chip for Bioelectrochemical Sensors," presented at IEEE.
Int Symposium on Circ. and Systems (ISCAS), 2007.

Hintsche.R, AlbersJ, Bemt.H, and Eder.A, "Multiplexing of
Microelectrode Arrays in Voltammetric Measurements," Electroanalysis,
vol. 12. pp. 660-665, 2000.

Chan.F-L, C. W-Y, K. L-M, L. C-H, W. S-W, Y. Y-S, and L. M-S-C, "An
Electrochemical dopamine sensor with a CMOS detection circuit,"
Journal of Micromechanics and Microengineering, vol. 18, 2008.

M. Nazari, H. and R. Genov, "A fully differential CMOS potentiostat,"
IEEE Int. Sym. on Circ. and Systems, pp. 2177-2180, 2009.

C. Yang, S. R. Jadhav, R. M. Worden, and A. J. Mason, "Compact Low
Power Irnpedance-to-Digital Converter for Sensor Array Microsystems,"
IEEE J. Solid State Circuits, vol. 44, pp. 2844-2855, 2009.

A. Manickam, A. Chevalier, M. McDermott, A. Ellington, and A. Hassibi,
"A CMOS Electrochemical Impedance Spectroscopy Biosensor Array for
label-free Biomolecular detection " presented at IEEE Int. Solid State
Circuits Conference(ISSCC), 2010.

F. Heer, W. Franks, A. Blau, S. Taschin, C. Ziegler, A. Hierlemann, and

H. Baltes, "CMOS microelectrode array for the monitoring of electrogenic
cells," Biosensors & Bioelectronics, vol. 20, pp. 358-366, 2004.

103

[106]

[107]

[108]

[109]

[110]

[111]

[112]

[113]

[114]

[115]

A. Chovin, P. Garrigue, P. Vinatier, and N. Sojic, "Development of an
Ordered Array of Optoelectrochemical Individually Readable Sensors
with Submicrometer Dimensions: Application to Remote
Electrocherniluminescence Imaging," Anal. Chem, vol. 76, pp. 357-364,
2004.

L.-C. Chen and K.-C. Ho, "Multimode Optoelectrochemical detection of
cysteine based on an electrochrorrric Prussian blue electrode," Sens.
Actuators B, Chem, vol. 130, pp. 418-424, 2008.

G. Singh, D. Moore, and R. F. Saraf, "Localized Electrochemistry on a
lOum Spot on a Monolithic Large Electrode: An Avenue for
Electrochemical Microarray Analysis," Anal. Chem, vol. 81, pp. 6055-
6060, 2009.

T. Tokuda, K. Tanaka, M. Matsuo, K. Kagawa, M. Nunoshita, and J. Ohta,
"Optical and electrochemical dual-image CMOS sensor for on-chip

biomolecular sensing applications," Sensors and Actuators pp. 315-322,
2007.

A. Agah, A. Hassibi, J. D. Plummer, and P. B. Griffin, "Design
requirements for integrated biosensor arrays," presented at Proc of the
SPIE, 2005.

C. Yang, "Impedance Extraction Microsystem for nanostrucured
electrochemical sensor arrays," in Electrical and Computer Engineering,
vol. PhD. East Lansing: Michigan State University, 2008, pp. 155.

G.E.Hack, "Binary up/down counter," IBM technical Disclosure Bulletin,
vol. 24, pp. 4399-4400, 1982.

X. D. Lu and P. C. Treleaven, "A special-purpose VLSI chip:A dynamic
pipeline up-down counter," J. Microprocess. Microprogram, vol. 10, pp.
1-10, 1982.

C. Yang, Y. Huang, B. L. Hassler, R. M. Worden, and A. J. Mason,
"Amperometric Electrochemical Microsystem for a Miniaturized Protein
Biosensor Array," Proc of IEEE Biomedical Circ. Systems, vol. 3, pp.

160-168, 2009.

X. Liu, "Amperometric readout circuit," in Electrical and Computer
Engineering, vol. M.S. East Lansing: Michigan State University, 2009.

104

[116]

[117]

[118]

[119]

[120]

[121]

C. C. Enz and G. C. Temes, "Circuit techniques for reducing the effects of
op-amp imperfections: autozeroing, correlated double sampling, and
chopper stabilization," Proceedings of the IEEE, vol. 84, pp. 1584-1614,
1996.

H. Eltoukhy, K. Salama, and A. E. Gamal, "A 0.18pm CMOS
Biolurrrinescence Detection Lab-on-Chip," IEEE J. Solid State Circuits,
vol. 41, 2006.

B. Fowler, A. E. Gamal, and D. Yang, "Techniques for Pixel Level
Analog to Digital Conversion," Proc of SPIE conference on Infrared
Readout Electronics, vol. 3360, 1998.

S.-H. Huang and W.-Z. Chen, "A lO-Gbps CMOS Single Chip Optical
Reciever with 2-D Meshed Spatially-Modulated Light Detector," Proc of
IEEE Custom Intergrated Circuits Conference (CICC) vol. 6, pp. 129-132,
2009.

L. Baker, and Boyce, CMOS Circuit Design, Layout, and Simulation:
IEEE Press, 1998.

M. Dandin, P. Abshire, and E. Smela, "Optical filtering technologies for
integrated fluorescence sensors," Lab on a Chip vol. 7, pp. 955-977, 2007.

105

 

"71111117119111illiliiilllllilil“