25

Chapter 2

BACKGROUND THEORY AND LITERATURE REVIEW

2.1. Introduction

With the growth in population crossing 2 billion, health care is one

of the niche areas where in lots of importance is being paid by both

the government and private sector industries. In a society, public

health plays an important role. There is no human being who is

ideally healthy. Every family will have some ill health, sickness and a

need exists for medication. In every locality a hospital is an essential

establishment for the healthcare of its people. Various schemes are

being offered by Indian government to offer health care at subsidized

rates. Another major challenge is the availability of doctors. It is being

projected that in 2020, the ratio of patients to doctors in India will be

1000:1 and worldwide will be 800:1. This puts lots of pressure and

demand for doctors and medical practioners will only rise [63-64].

Automated drug delivery is one possible solution to overcome the gap

between demand and supply in health care sector. With the growth in

technology and emergence of nanotechnology, biochips (consisting of

biosensors, signal processing and conditioning circuits and controllers

for drug diffusion) provide easy and reliable solutions to mankind in

tackling the health issues. Automated drug delivery is an

interdisciplinary domain that involves biosensors (biology/electronics)

for detection of virus, neural network for disease classification,

26

embedded system for control and drug delivery and electrical system

for motor drive. In this work, an attempt is made to review literatures

in all the above domains and the literature summary is presented at

the end of this section. Automated drug delivery unit has been in

research domain for the past 15 years [42]. However no significant

breakthrough has been attained yet due to the limitations in

development of technologies. Rapid growth in nanotechnology and

availability of biosensors has led to an unprecedented demand in

automated drug delivery units. This work is an attempt to design and

build one such unit for particular disease detection. An attempt is also

made to design and model various biosensors that help in disease

detection.

In this chapter, background theory and review of existing systems

for automated drug delivery is discussed. At the end of this chapter,

literature review summary is presented based on which the problem

statement, definition and methodology adopted to carry out this work

is discussed in detail.

2.2. Traditional Drugs and Diagnosis

Age old practices of disease detection and curing are through

seeking appointment with doctors and being tested as per the

standard procedures. Treatment is provided based on the

recommendations and observations made by the doctor during the

test procedures. The drug is given to the patient as prescribed by the

doctor and the patient is kept under observation. There are very few

27

procedures for drug delivery and these include oral prescription and

injections apart from the treatment procedure which a doctor may

prescribe [65]. When the patient goes to a doctor seeking diagnosis for

a disease, the doctor generally prescribes drugs which are

administered through oral delivery or injections. Problems with Oral

Delivery [66] are:

Often the liver cleans out a large portion of the drug.

The acidic pH of the stomach can often destroy the drug, well

before it is absorbed

Does not work for ionic water soluble drugs

The drugs can also be administered through:

Injections viz. Intravenous, Intramuscular, Subcutaneous

Implantable Devices (ID): Food and Drug Administration has identified

some of the implantable ID chips in humans, for medical purposes

besides security, financial and personal identification or safety

applications.

2.3. Categorization of Implantable Devices

Implantable devices [67] can be categorized as Medical or Non-

medical devices, both either Passive or Active devices. In

Implantable medical devices the most passive implants are structural

devices such as artificial joints, vascular grafts and artificial valves.

On the other hand, active implantable devices require power to replace

or augment an organ’s function or to treat an associated disease.

28

2.3.1. Medical Devices

The "Active Implantable Medical Device" means any active

medical device which is intended to be totally or partially introduced,

surgically or medically, into the human body or by medical

intervention into a natural orifice, and which is intended to remain

after the procedure.

2.3.2. Non-medical Devices

An example of a passive device is the Radio Frequency

Identification (RFID) device. Active devices may use electrical impulses

to interact with the human’s nervous system.

2.3.3. Implantable Devices available in the Market

Current active medical devices available in market are

cardiovascular pacers for patients with conduction disorders or heart

failure, Cochlear and brainstem implants for patients with hearing

disorders.

2.3.4. Implantable Programmable Drug Delivery Pumps

Insulin pump for Diabetes

Neuroleptic /antipsychotic drugs also called "Psychiatric

Implants"

2.3.5. Implantable Neurostimulation Devices

Spinal cord stimulation for chronic pain management

29

Sacral nerve stimulation for control of urinary incontinence

Vagus Nerve Stimulation (VNS) for seizure control in epilepsy

and mood control in severe depression cases. The small

generator and lead are surgically attached to the rib cage, with

the wires traveling under the skin up to the neck and wrapped

around the left Vagus nerve from where the generator sends

electrical signals via the Vagus nerve to the brain.

2.3.6. Medical Devices and Implants

Implantable transponders to locate tumors

Computer-based accurate model of brain

Implantable electronic devices that stimulate nerves to treat

chronic conditions

Rechargeable devices with replacement time of 5-10 years

Adding nanosensors on pumps to deliver drugs

Converging and changing rapidly with computer, nano and bio-

science

Bionic Peepers

Artificial corneas, lungs, glands

2.3.7. RFID Devices

Millions of Radio Frequency Identification Device (RFID) tags

[68] have been sold since the early 1980s. They are used for livestock,

pets, laboratory animals, and endangered-species identification. This

30

technology contains no chemical or battery. The chip never runs down

and has a life expectancy of 20 years.

2.3.8. Implantable Devices under Development

The Micro-Electro Mechanical Systems device (MEMS) [69-72] is

an implantable micro-sensor that can send data to a hand-held

receiver outside the body, alerting doctors to a potential medical crisis,

without using any wire or battery. This GPS monitoring could be used

for several purposes, such as,

• In case of Medical emergencies

Heart attack, Epilepsy, Diabetes

Implantable GPS microchip

2.3.8.1. Biodegradable Implants

Advantages of using biodegradable implants are as follows:

Eliminates additional surgery to remove an implant after it

serves its function

Ideal when the “temporary presence” of the implant is desired

Replaced by regenerated tissue as the implant degrades

Bio Degradation is used for short term applications like

o sutures

o drug delivery

o orthopedic fixation devices (requires exceptionally strong

polymers)

31

o adhesion prevention (requires polymers that can form soft

membranes or films)

o temporary vascular grafts (development stage, blood

compatibility is a problem)

There are four main types of degradable implants:

o the temporary scaffold

o the temporary barrier

o the drug delivery device

o multifunctional devices

The bio - degradable implants provide support until the tissue

heals which is weakened by disease, injury or surgery. It also

supports healing wound, broken bone, damaged blood vessel, sutures,

bone fixation devices and vascular grafts. The rate of degradation is

that the implant should degrade at the same rate of the tissue healing.

Drug-Coated Stents slowly leach medication into blood vessels to keep

them from squeezing shut. They are better than the plain old metal

stents that were used in the past. Big blockages in very small vessels

which are nearly two inches long can be fixed with such stents.

Certain type even dissolves in the body once its job is done.

2.4. Automated Drug Delivery System

Automated drug delivery also called as implantable drug delivery

device or biochip have demonstrated their potential in vast areas of

medical applications as they provide reliable, controllable and

accurate delivery of prescribed drugs without medical intervention.

32

Use of MEMS technology for disease detection, disease monitoring and

drug diffusion applications have been demonstrated for chronic illness

[43]. A typical Automated Drug Delivery Unit is shown in Figure 2.1.

The major components in the system are nanobio sensors, control

unit, interfaces, signal conditioning and processing circuits, expert

system and control unit for monitoring and drug dosing.

Figure 2.1. Block Diagram of Automated Drug Delivery Unit

2.5. Challenges of Drug Dosing

Drug dosing is a technique that is done to cure the diseases through

proper prescription and control of drugs that have been identified to

its corresponding disease based on diagnosis. Drug dosing is a very

critical and challenging step which needs to be correctly monitored

and prescribed by the doctor. Researchers have carried out survey on

identifying new drugs for various diseases and have experimented on

animals and human beings. The fundamental problem is the scaling

of drugs from animal system to human being. For example, scaling up

of drugs tested on mice to human being is still not very much proven.

Preprocessing Signal

conditioning

Decision

making logic

Classification Drug

selection

Drug dosing

and monitoring

Interface Sensors

Control

33

Drug doses have cured mice but many of the drugs tested on human

beings have still not cured humans. Scaling up of drugs from animals

to humans is a challenge all together.

Drug dosing in human beings also depends upon size, area,

weight and volume of the recipient. There exist vast differences

between adults in terms of genetics, metabolism and their physical

structure. Techniques that are adopted for adults cannot be extended

to children. In children, physical structure of human body is

fundamentally different from adults, they are no small adults, their

organs are different and their metabolism is different. Thus the

experiences of prescribing drugs to humans cannot be adapted to

children. Drug dosing is gender diverse, male, female and children

have to be treated differently. Patient’s history is required as body

reactions to drugs are different and thus treatment should be

different. In conventional dosing of drugs, medicine induced into

patient’s body goes everywhere, therefore it is necessary to calculate

volume of drug required and also direct the diffusion of drug, i.e.,

where the drugs can go as in some regions of the body the drug is to

be restricted to enter. These are determined based on calculations and

experiences of doctor. There are various ways of drug dosing with

targeted therapies being the best predictor in terms of perceiving the

dosing. Nanobio systems are used for targeted therapies. FDA has

recommended the use of nanobio systems for targeted therapy and

drug dosing, but they do not have many models, very few exist.

Currently the best way to predict drug dosing is based on experiences.

34

Traditional drug dosing techniques are decided based on area,

weight, volume of patient and then drugs are being recommended. For

some patient body weight or height or volume might be determined

and drug is prescribed, but the patient may have missing organs and

thus the recommendations can be blind. Thus drug dosing is patient

specific and hence a detailed analysis of patient is required before

drug dosing. Control engineering principles are adopted to automate

the system. In control engineering, engineers consider existing data of

patient’s behavior to a given drug and then build models based on

control theory and use them for drug delivery [73]. A typical control

theory model is shown in Figure 2.2. In this model, exchanges

between central compartment and peripherals are controlled by use of

mathematical models that have been developed based on patient’s

history and it is a sophisticated model. The central compartment,

which is the site for drug administration, is generally thought to be

comprised of the intravascular blood volume as well as highly perfuse

organs such as the heart, brain, kidney, and liver. The central

compartment exchanges the drug with the peripheral compartments

comprised of muscle, fat, and other organs and tissues of the body,

which are metabolically inert as far as the drug is concerned.

35

Figure 2.2. Controlled Drug Deliveries

2.5.1. Side Effects in Drug Dosing

During the process of drug dosing, the medicine or drug that is

used for curing a particular kind of disease also reaches other parts in

the body, affecting their molecular structure resulting in side effects

[8]. Drug dosing may also introduce side effects if the diffused drug

exceeds more than the controlled limits and may also cause other

effects to the organs.

In general, the more common side effects caused by drugs include

[74]:

Allergic reactions, such as hives or itching

Flu-like symptoms, including chills, fatigue, fever and muscle

aches and pains

Low blood cell counts, which may lead to bleeding, fatigue and

infection

Nausea

Diarrhea

Skin rashes

Central

Compartment

2

3 n-1

n

36

2.5.2. Genetic responses to Drug Dosing

Drug dosing is based on target therapy and is decided based on

previous history of patients. All humans are not genomically

equivalent and therefore, drug dosing is predicted based on family

tree, and their reactions to drugs. Thus history of family is identified

and then drug dosing is carried out. This is a challenging task as most

of the time family history would not be available. An alternate

technique is to locate particular genes called as SNP- Single

Nucleotide Polymorphisms [75,76]. Predicting drug dosing based on

SNPs has advantages and have been used, but identifying the SNPs is

a major challenge and therefore research is being carried out in this

domain and in another five years this technology may be viable.

Nanobio medical systems are being researched for directed therapies.

In this technique right target cells are identified and drug dosing is

carried out only in this specific region. This minimizes the dose and

also prevents other organs being affected due to the drug and is one of

the best models which have received approval from FDA [77,78,79].

2.6. Survey of Literature on Automated Drug Delivery

Unit

Valcke and Chizek (1997) developed a Closed Loop Drug

Delivery (CLDD) system for use with coronary artery disease [80]. In

this system feedback control algorithms is used to deliver and monitor

37

the drugs to the patient. The patient's response to the delivered drug

is monitored and the drug infusion is adjusted. The system developed

is safer as it constantly monitors the response and when the patient's

response is not within the "safety limits" or becomes hazardous, the

system notifies the operator and stops the drug delivery altogether.

The control unit consists of a Proportional-Differential (PD) controller.

The time response for drug delivery introduces delay due to

implementation complexities. As the experiment was conducted on

animals, delay would not matter but on human being for any real time

it is required to minimize the delay. Since animal experiments are a

lengthy and expensive process, only simulation models are discussed.

Woodruff et al. in 1997 [81] have developed a simulator that can

be used to model closed loop cardiovascular drug delivery unit

considering multiple model factors. It has 1) a nonlinear, pulsatile-

flow cardiovascular model, 2) a physiological regulatory model of the

baroreceptors, 3) a pharmaco kinetics model, and 4) pharmaco

dynamic models of the drugs. Individual building blocks for the

simulator model was developed and individually validated. It was

validated through published data and physician perceptions. Five

animals were tested and realistic simulations were obtained. In this

work, the developed model was validated against real time results and

was published on web.

Yu et al. 1992 [82] developed a controller that can be used to

monitor the cardiac output of a congestive heart failure patient and

administer vasodilation and inotropic agents (Nitroprusside and

38

Dopamine). The controller model used data from 6 "most probable"

patient models to calculate the control algorithms and thus simplified

the computations. Until the advent of a controller such as the one

described in this article, multiple controllers had been used to infuse

drugs such as Sodium Nitroprusside and Dopamine separately. This

new controller greatly simplified the tuning of the overall process by

lumping more than one process together. Most of the models

developed in earlier days were simulation models and used recorded

real time data to make decisions. They are also called as passive

devices as they cannot be controlled in real time.

With the emergence of VLSI technology and MEMS attaining

maturity, a variety of implantable devices have been designed and

demonstrated for chronic diseases, Prescott, et al. 2006 [83]. MEMS

based devices had advantages of active control and were practically

possible to perform the drug delivery process. MEMS based drug

controlled devices provide advantages over passive devices, as the

drug delivery process can be controlled actively after implantation and

even monitored using telemetry, as opposed to passive devices that

depend on the degradation chemistry of the specific device materials

in the intended implantation region.

Very recently MEMS technology based drug delivery devices

were developed to actively control disease (in vivo) in multiple clinical

applications. In order to build such systems, micro-pumps, Nguyen, et

al. 2002 [84], electro-chemical or electrical degradation of membranes

39

for multiple-reservoir drug delivery chips, Santini, et al. 1999 [85];

Grayson, et al. 2003 [86] were used.

The availability of MEMS technology enabled miniaturization of

micro-pumps, storage and delivery of drugs from single and multiple

reservoirs making detection of disease possible with use of sensors.

However, the developed system suffer from major limitations such as

low delivery rates, low reliability due to mechanical moving parts, high

power consumption. Thus drug delivery units based on electrical

mechanisms provide more reliable results. These devices rely on

electro-thermal actuation to rupture a reservoir sealing membrane as

a result of an applied electrical potential, allowing the drugs inside of

reservoirs to freely diffuse into the region of interest, Maloney, et al.

2005 [87]; Grayson, et al. 2005 [88]. Automated drug delivery systems

have been investigated and developed for use in chronic and non-

chronic diseases such as cancer, diabetes and osteoporosis, but still

there is a long way to go in making this device more acceptable and

reliable. From the discussions presented in this section, it is

identified that, drug delivery requires the following:

Sensors for disease detection

Decision making logic

Controller for drug actuation and monitoring

Drug storage and diffuser

Next section comprises of biosensors for disease detection. Various

biosensors and their properties employed for disease detection is

presented in detail. Fabrication methods and material properties for

40

biosensor design are discussed. Based on the discussions presented,

the gaps in literature and scope of this work are presented.

2.7. Biosensors

The ability to detect pathogenic and physiologically relevant

molecules in the body with high sensitivity and specificity offers a

powerful opportunity in early diagnosis and treatment of diseases.

Early detection and diagnosis can be used to greatly reduce the cost of

patient care associated with advanced stages of many diseases. These

costs have been estimated to be $75 billion [21] and $90 billion [22]

for cancer and diabetes, respectively. The diseases are being presently

detected by monitoring the concentration of certain antigens present

in the bloodstream or other bodily fluids and through tissue

examinations. Biosensors are devices that are used to detect diseases

of biological activity in a human body. A biologically sensitive device

with a physical or chemical transducer is used to detect the presence

of specific compounds in a given environment, Vo-Dinh and Cullum,

2000 [89]. Biosensors integrated with other electronic and mechanical

modules are also called as biochips and are used for delivery,

processing, analysis, or detection of biological molecules and species’,

Bashir, 2004 [90]. These devices are used to detect cells,

microorganisms, viruses, proteins, DNA and related nucleic acids, and

small molecules of biochemical’s present in human body. Typically

biosensors comprises of three components: (1) the detector, which

41

identifies the stimulus; (2) the transducer, which converts this

stimulus to a useful output; and (3) the output system, which involves

amplification and display of the output in an appropriate format [3].

Using a biosensor, various samples containing viruses or bacteria are

detected and analyzed as shown Figure 2.3. A sample or stimulus to

be analyzed can be water, food, air, body, blood and fluid that consists

of corresponding virus (called as target) to be detected is preprocessed

and is passed on a sensor consisting of a detector (called as receptor),

the biological activity between the target and receptor constitutes

change in physical property of the transducer or sensor thus

producing equivalent electrical signal.

Figure 2.3. Biosensor for Detection and Analysis of Samples [90]

The electrical signal captured is analyzed for virus detection as shown

in Figure 2.4.

Sample Processing Detection/ Data Analysis/ Separation ID Results

Water

Food

Air

Body

Fluids

42

Figure 2.4. Biosensor function for a Biochip [90]

The biological components present in a sample that is used by

biosensors to detect the presence of a virus or disease can be divided

into five major mechanisms. They are

Antibody/antigen,

Enzymes,

Nucleic acid,

Cells and viruses and

Biomimetic

In antibody/antigen based mechanism, the binding of an antigen

with antibody occurrences under nonspecific interactions are

minimized. The high specificity between an antibody and an antigen

can be utilized in this type of sensor technology [91]. Changes in

refractive index, reflectivity through fluorescent labeling is observed

due to binding and thus the presence of a disease id detected.

Enzyme-based biosensors are composed of enzyme bioreceptors that

Controlled Microenvironment

In a bio-chip

Stimulus

Real time Bio-chemical Communication Electrical or Optical Signals

Cell

43

use their catalytic activity and binding capabilities for specific

detection [92]. The catalytic activity of the enzymes provides these

types of biosensors with the ability to detect much lower limits than

with normal binding techniques. The complementary relationships

between adenosine and thymine and cytosine and guanosine in DNA

form the basis of specificity in nucleic acid-based biosensors. These

sensors are capable of detecting trace amounts of microorganism DNA

by comparing it to a complementary strand of known DNA [91]. For

accurate analysis, polymerase chain reaction (PCR) is often used to

create multiple copies of the sample DNA. DNA is composed of a

phosphate back-bone where each phosphate radical has a negative

charge, a Deoxyribose (D in DNA) sugar and 4 types of bases or

nucleotides. These are adenine (A), thymine (T), cytosine (C), Guanine

(G). The binding property is such that A binds to T and G binds to C

and thus they are called as complimentary base pairs. The base

structure of DNA [93] is shown in Figure 2.5.

DNA based biosensors are very popular in the market and have

shown better performance compared to any other principle of

biomarkers. Polymerase Chain Reaction (PCR) is a technique to

amplify (make multiple copies) DNA molecules. Some of the diseases

that have been diagnosed are based on PCR techniques.

44

5’end BASE

Deoxyribose

3`linkage 5`linkage 3`end

Phosphodiester bond

Figure 2.5. DNA Structure [93]

Microorganisms such as bacteria and fungi can be used as

biosensors to detect specific molecules or the overall ‘‘state’’ of the

surrounding environment [94]. Proteins that are present in cells can

also be used as bioreceptors for the detection of specific analytes [95,

96]. A biomimetic biosensor is an artificial or synthetic sensor that

mimics the function of a natural biosensor. These can include

aptasensors, where aptasensors use aptamers as the biocomponent

[91]. Aptamers are synthetic strands of nucleic acid that can be

designed to recognize amino acids, oligosaccharides, peptides, and

proteins [97].

Further biosensors can also be classified according to the

method of signal transduction. There are four different types of signal

transduction such as optical-detection, electrochemical, mass-

sensitive and thermal detection [98]. Optical detection biosensors are

CH20

CH20

CH20

CH20

H

G

C

T

A

H

H

45

the most diverse class of biosensors because they can be used for

many different types of spectroscopy, such as absorption,

fluorescence, phosphorescence, Raman, SERS, refraction, and

dispersion spectrometry [91]. In addition, these spectroscopic methods

can measure different properties, such as energy, polarization,

amplitude, decay time, and/or phase. Amplitude is the most

commonly measured as it can easily be correlated to the concentration

of the analyte of interest [91]. Electrochemical biosensors measure the

current produced from oxidation and reduction reactions. This

current produced can be correlated to either the concentration of the

electro active species present or its rate of production/consumption

[91]. Biosensors that are based on mass-sensitive measurements

detect small mass changes caused by chemical binding to small

piezoelectric crystals. Initially, a specific electrical signal can be

applied to the crystals to cause them to vibrate at a specific frequency.

This frequency of oscillation depends on the electrical signal frequency

and the mass of the crystal. As such, the binding of an analyte of

interest will increase the mass of the crystal and subsequently change

its frequency of oscillation, which can then be measured electrically

and used to determine the mass of the analyte of interest bound to the

crystal [92]. Thermal biosensors measure the changes in temperature

in the reaction between an enzyme molecule and a suitable analyte

[99]. This change in temperature can be correlated to the amount of

reactants consumed or products formed.

46

In this work, biomarkers based on antigen/antibody and DNA

mechanisms are being used to develop biochip, detection principles

based on electrochemical and mass sensitive techniques are used to

develop sensors for automated drug delivery system. Next section

discusses briefly nanotechnology principles based nanosensors that

have been used as biosensor. Bio sensors that have been developed

based on nanotechnology principles are also called as nanobio

sensors. A detailed discussion on various nanobio sensors are

presented in next section.

2.7.1. Nanobiosensors

The emergence of nanoscale technologies for biology has a great

potential to lead to the development of next generation biosensors with

improved sensitivity and reduced costs [100-103]. Modern biosensors

based on nanoscale techniques have the potential to greatly enhance

methods of detecting foreign and potentially dangerous toxins and

may result in cheaper, faster, and easier-to-use analytical tools.

Furthermore, nanoscale biosensors may be more portable and

scalable for point-of-care sample analysis, controlled drug delivery

and real-time diagnosis. With development in fabrication technology

and due to unique properties of various nano devices such as

nanowires, carbon nanotubes, quantum dots, nanoparticles has

become the focus of intensive research.

47

2.7.2. Nanoparticles

Nanoparticles based on Gold Material (also called as GNPs) have

been used for biosensors due to their special optical and absorption

properties. GNPs are found to have properties such as biocompatible,

less cytotoxic, and resistant to bleaching and can absorb or scatter

light from the visible to near-infrared region. GNP based sensors have

been used in colorimetric biosensor [104-106], cancer imaging

[107,108], drug delivery [109] and cancer therapy [110-112]. Figure

2.6 shows cancer cell detection using GNPs. A GNP is coated with

Raman reporters and conjugated with ScFv antibody and is used in

detecting cancer cells.

GNP ScFv E GFR antibody

Raman reporter DTTC SH-PEG COOH

(a) PEG-SH Shell (b) Cancer cell

Figure 2.6. Cancer cell targeting and spectroscopic detection

using antibody-conjugated SERS nanoparticles (a) Modified gold

nanoparticle with Raman reporter and targeting molecule (b)

Schematic illustration of the nanoparticles targeting the cancer

cells

Gold nanoparticles are used as molecular imaging agents and are

found to be very useful in detecting cancer proteins. An antibody

called as anti-EGFR (epidermal growth factor receptor) conjugated

48

with gold nanoparticles was used to detect EGFRs in cell membranes.

Thus GNPs have found wide use in detecting cancer cells in human

body. Using anti-Prostate Specific Antigen (PSA) antibody combined

with GNPs and gold nanorods was used as one-step homogeneous

immunoassay for cancer biomarker detection. Most of the sensors

developed using GNPs have been used for photothermal cancer

treatment.

2.7.3. Quantum Dots

Quantum dots (QDs) are semiconducting, light-emitting

nanocrystals that have emerged as a powerful molecular imaging

agent and are significantly used as a nanobio sensor. QDs due to their

unique optical properties compared with organic fluorescent labels

have found to have wide applications. Exceptional optical properties of

QDs have made them an exciting field of study for many researchers

in search of molecular imaging tools for better cancer diagnosis. A

multifunctional nanoparticle with biomolecules conjugated to QDs has

been used in cancer targeting and drug delivery [113] as shown in

Figure 2.7.

A10 RNA APTAMER DOX PSMA

Figure 2.7. Cancer Biomarker using Quantum Dots

49

Recent study on QDs has suggested that functionalized

nanoparticles are used to detect alpha-fetoproteins (AFPs) [114] in

immunoassays. QDs have also been integrated into nano-bio-chips

(NBCs) for detecting multiple cancer biomarkers [115].

2.7.4. Carbon Nanotubes

After the discovery of fullerene, carbon nanotubes (CNTs) have

been re-discovered in 1991 by Iijima [116]. CNTs provide high surface-

to-volume ratios, mediate fast electron-transfer and can be

functionalized with almost any desired chemical species. It is also of

great advantage to use carbon nanotubes because label-free detection

of cancer biomarkers is possible. Current developments in CNT-based

cancer detection focus on a more precise and sensitive detection of the

cancer biomarkers. Single-wall carbon nanotubes (SWNT) with [Ru-

(bpy)3]2+- doped silica nanoparticles have been studied as an

electrochemiluminescent immunosensor for PSA detection as shown

in Figure 2.8.

Figure 2.8. SWNT forest with ECL nanoparticles as sandwich

immunoassay for PSA detection

50

CNT Field Effect Transistor (FET) based biosensors are of great

interest [117]. CNTs form the conducting channel in a transistor

configuration and interact with introduced analytes. Analytes interact

with CNTs where 1) charge transfer may occur from analyte molecules

to the CNTs or 2) the analytes can act as scattering potential across

the CNTs. A relationship between the concentration of the analyte and

conductance or potential can be monitored and biosensing can be

done based on this system. A real-time detection of PSA-ACT complex

with CNT-FET has been developed and by providing sufficient space

between each antibody on the CNT, the sensitivity of the system was

maximized as shown in Figure 2.9.

Drain Source

Gate

Linker PSA-ACT Antibody Spacer

Figure 2.9. Set up of CNT-FET with a Linker and a Spacer for the Maximized Sensitivity

This spacing is a crucial element of sensing biomarkers in a buffer

solution since the target is required to be at a distance within the

Si Substrate

AU

PR

SiO2

Analyzer

Pt

AU

PR

51

Debye length to give a large gating effect. Using 1-pyrene butanoic

acid succinimidyl ester (PASE) as a linker and 1-pyrenebutanol (PB) as

a spacer at a specific linker-to-spacer ratio on SWNTs, conductance of

CNT-FET could be largely enhanced [118].

2.7.5. Nanowires

Nanowires have been the object of interest in the recent studies.

When used in biosensing devices, these nanowires have two main

advantages over carbon nanotubes (CNTs). Firstly, the material

properties can be more precisely controlled by manipulating the

conditions during synthesis using well-developed doping techniques.

Secondly, the native oxide layer that forms on the outside of

nanowires allows the use of a broad class of already well developed

fictionalization and blocking chemistries. Application of nanowires in

biosensors shows the most interesting studies [119-127]. In this

section, the biosensor devices made of nanowires as an active sensing

material have been discussed.

Electron transport properties of nanowires are very important for

electrical and electronic applications as well as for understanding the

unique one-dimensional carrier transport mechanism. It has been

noticed that the wire diameter, wire surface condition, crystal

structure and its quality, chemical composition, crystallographic

orientation along the wire axis etc., are important parameters, all of

which influence the electron transport mechanism of nanowires. For

example, the I-V characteristic studies of the Cu nanowires both at

52

room temperature and at 4.2 K showed the linear ohmic behavior [53].

However, by oxidation when the metallic Cu nanowires has been

transformed into semiconducting Cu2O nanowires and placed between

two electrodes, it forms two Schottky diodes in series and in turn

produces a double diode like I–V characteristic curve. It has been

reported that quasi one-dimensional nanowires exhibit both ballistic

and diffusive type electron transport mechanism, which depends upon

the wire length and diameter. It is found that the conductance of

nanowires strongly depends on their crystalline structure. For

example, in the case of perfect crystalline Si nanowires having four

atoms per unit cell, generally three conductance channels are found

[119].Various nanowires have also been applied to biomarker

detection based on silicon nanowires [127], In2O3 nanowires [128],

gold nanowires [129], and conducting polymer nanowires [130].

Silicon nanowires (SiNW) are semiconducting nanowires with

exceptional physical, optical, electronic properties, and excellent

biocompatibility. Since silicon is a well studied material, the surface of

the nanowires can be modified with well-known methods. This

advantage makes itself a promising platform for sensitive detection of

biomarkers [128]. SiNWs modified with peptide nucleic acids (PNAs)

have been used to detect miRNAs that were extracted from Hela cells

as shown in Figure 2.10.

53

SiNW SiNW

miRNA PNA Target miRNA Hybridization

Figure 2.10. SiNW Based Target miRNA Detection via PNA

Voltammetric detection of cancer biomarkers using silica

nanowires (SiO2-NWs) is also developed using nanowires. A lung

cancer biomarker, interleukin-10 (IL-10) and osteopontin (OPN), has

been detected using silica nanowires as templates, through

electrochemical alkaline phosphatase (AP) assay [132] as shown in

Figure 2.11.

Alkaline Phosphotase

Detector antibody OPN/IL-10

Capture antibody

Figure 2.11. Detection of Cancer Biomarker through Sandwich Immunoassay using SiO2-NW

Nanowire sensors have been used as biosensors for detecting

hydrogen and ethanol. Nanowires impregnated with Pt also show high

sensitivity for 1000 ppm of ethanol at or below 150 ºC, with short

recovery and response times. Noble metal nanowires with enzyme

SiO2-NW

54

modified electrodes have been used for detecting H2O2 and glucose.

Gold nanowires with cholesterol oxidase (COx) and cholesterol

esterase (CE) possess excellent properties for the cholesterol detection

and sensors have been developed to detect cholesterol present in

human body. La(OH)3 Nanowires Modified Carbon Paste Electrode

(CPE) sensor was developed to detect Inosine. The system was applied

to inosine determination in both pharmaceutical preparations and

human serum. Gold Nanowires with Glycoconjugate (Antibody) was

developed to detect Bacterial and Octadecanethiol. Two types of

sensors that composed of gold nanowires, one sensor made up of gold

nanowires with glycoconjugate (antibody) were used for bacterial

detection [74]. Another application of gold nanowires as reported by

Liu et al. was for octadecanethiol sensing [75]. Silica Nanowires were

developed to detect cancer Biomarkers Exemplified by Interleukin-10

and Osteopontin. Ramgir et al. reported real-time voltammetric

detection of potential lung cancer biomarkers, namely, interleukin-10

(IL-10) and osteopontin (OPN), using localized silica nanowires as

templates, through an electrochemical alkaline phosphatase (AP)

based assay [76]. Silver Mesowires were developed to detect Amine

Vapor Sensor. Silver mesowires with diameters ranging from 150 to

950 nm and lengths of 100 µm or more were prepared by

electrochemical step edge decoration [77]. The unique properties of

nanowires such as mechanical stability, light weight, enhancement in

the current, reduction in potential are very helpful for the

development of efficient electrochemical sensors and biosensors.

55

Functional nanowire arrays constitute one of the most fascinating

current issues in the nanotechnology field. Given that most of the

applications described above are concerned with pushing the limits of

how few molecules can be detected in a given volume of solution.

Nanowire sensors are inherently useful so long as the entire platforms

are designed in such a way that the entire sample volume can be

interrogated by the sensor. Also, their electrical, thermal, magnetic

and optical properties are excellent. However, the real value of sensors

lies in their detection limit range, sensitivity, etc., and all these are

enhanced by the nanowire modified electrodes in the sensors. These

overall properties provide additional benefits, which enable

development of sensors made of multifunctional, structural materials.

In a few more years nanowire sensors will become prominent among

researchers and in market. They will be small in size and high in

sensitivity.

In this work nanowire based biosensors are used to detect diseases

in human body. In most of the research carried out and reported in

literature it is found that the sensors that are developed have been

first fabricated using any of the advanced technologies and with

appropriate use of materials. Fabrication of biosensors requires

sophisticated lab facilities and is very expensive. Suitable funds

provided by Government or Universities help in setting up state of the

art lab facilities to manufacture bio sensors. It is also found in the

literature that many of the work carried out on biosensor design are

not performed on physical devices but have used mathematical

56

models to validate their design. There are many tools that have been

developed to help researchers to design, model and analyze the

performances of biosensors as per their interest. The tools provided

are developed by various eminent scientists and industry experts and

provide a platform for carrying out research activities in the field of

biosensors. In this work, as the focus of the research is towards

developing a system that can automatically detect the presence of a

particular disease in human body, classify the detected disease and

take necessary action to diffuse the drug, it becomes very difficult to

prove the working model of the entire system due to non availability of

required facilities. However, it is found that in near future, there

would be a huge demand for automated drug delivery and thus the

work carried out in this research would play a vital platform for

scientists, researchers and engineers to adopt the developed model

and fabricate the design for real time applications. Resources available

at nanohub.org [133] are used to develop the proposed automated

drug delivery model. For real time application it is required to develop

a model that resembles a physical model, thus the biosensor model

developed using the nanohub resources are further integrated with

Matlab and Simulink model to develop the automated drug delivery

unit. The expert system that is required to classify and detect the

disease based on the biosensor signals are realized on hardware and

the design is optimized for area, power and speed performances.

Biosensors are extensively used in medical field for disease

detection. In order to perform experimental analysis on the

57

performances and verify functionality of biosensors various tools are

provided in biosensors lab available in nanohub.org. BioSensor Lab

available in nanohub.org is a numerical simulator to predict the

performance metrics for various types of label-free, electronic

biosensors. The BioSensor Lab focuses only on those sensors that can

detect the presence of charged biomolecules near the sensor surface

by electrostatic interaction. The experiments carried out using

biosensor lab provide information on sensor properties and its

electrical characteristics. The development of the automated drug

delivery unit necessitates integration of the biosensor with the expert

system to detect and classify the diseases based on the signals

captured from the biosensors. The required drug, based on the

classification is then diffused into the human body. Thus an

integrated environment is necessary to design, model simulate and

validate the proposed research work. Figure 2.12 shows the integrated

environment developed in this work.

Figure 2.12. Integrated Software Environments for Automated Drug Delivery Unit

BioSensorLab Matlab

Neural

network

based expert

system for

disease

classification

and detection

Simulink

Drug

diffusion

control

unit

Fluid or analyte

solution

Biosensors

58

The biosensors that are modeled in Biosensor Lab are to be

integrated with Matlab and Simulink models. It is necessary to

develop a mathematical model similar to the biosensor available. In

this work, a mathematical model that can model the functionality of

biosensor is developed in Matlab, thus helping in integrating the

expert system with biosensor for construction of automated drug

delivery unit. From the literature review carried out the following are

the summary of literature review:

Mathematical models have been developed for various

sensors, integration of sensors for disease detection have not

been reported

Sensors are not characterized for disease detection,

particularly cancer detection

Integration of sensors with classification and drug diffusion

have not been reported

Biosensors are not available for monitoring or measuring all

possible diseases or activities of human body (limits the

monitoring of patients parameters)

Characterization of available biosensors is a very critical

activity before system deployment

Multiple biosensors are required to measure patients

activities, hence there is a need for an expert system to

measure, monitor and control the drug diffusion system

59

Neural network based approaches for patient monitoring of

multiple parameters and decision making unit is not

reported in the literature.

Intelligent control unit that can control drug diffusion based

on the variations in patients parameter is not reported in the

literature

In this work an attempt is made to bridge the gaps identified in the

literature, the work developed would serve as a platform for many of

the future researchers to pursue research activities in the area of

automated drug delivery unit. Also the model developed can be

realized using VLSI technology for real time applications. This

research work proposes a miniaturized nanotechnology based

automated drug delivery unit based on biosensors for disease

detection. Mathematical models for biosensors, biosensor device

simulations, embedded unit for decision making and drug delivery

unit are modeled and analyzed for its performances and

characteristics. Work developed in this research, is a first attempt

towards automated drug delivery unit for multiple disease

2.8. Problem Statement

“Modeling and Simulation of Biosensor Arrays for Automated

Cancer Detection”

60

2.8.1. Objectives

• To carry out literature review on the following:

– Diseases and diagnosis

– Experimental procedures for diagnosis, Existing

techniques and its limitations

– Bio-sensors and instrumentation circuits for disease

detection

– Diseases detection methods using different sensors

– Embedded control systems, Software and hardware

modeling

– Limitations of existing diagnostic techniques

• To develop the top level block diagram for automated drug

delivery system based on literature review and required

specifications

• To design, model and analyze nano-devices as bio-sensors for

drug delivery

• To develop, implement and analyze software reference model for

drug delivery system

• To analyze the performances of developed software reference

model for cancer detection

• To build the hardware model for automated drug delivery

system and verify its functionality and its performance.

2.8.2. Methodology Adopted to carry out the Proposed Work

This work is interdisciplinary involving domains such as

61

Biomedical Signal Processing, VLSI, nanotechnology and embedded

Systems. The major focus of the work is more towards application

development in the field of medical electronics using VLSI technology.

The methodology adopted in this work is discussed with the flow chart

shown in Figure 2.13. The methodology developed in this work is a

generalized technique that can be adopted to integrate biosensors into

Matlab environment and also develop hardware modules for

automated drug delivery unit.

The mathematical model developed based on the methodology

adopted is used as a Golden Reference Model that can be used in the

Matlab environment to analyze the performances of the biosensor. As

Matlab does not have a biosensor model, this approach acts as a

useful platform to model and analyze the performances and

characteristics of biosensor models. Using the above approach various

nanobiosensors for medical, industrial and defense applications can

be verified for its functionality and can be used in system

development.

62

Figure 2.13. Methodology Adopted for Developing Golden Reference for Biosensor

Literature review on biosensors and development of mathematical

approach for sensor modelling

Setting up of nanosensor lab setup for simulation of biosensors and

verification of functionality of biosensors

Understanding the properties of nano-biosensors based on the

experimental results and characterization of biosensors for disease

detection

Development of mathematical models for biosensors and modelling

the sensors using Matlab

Development of environmental setup to simulate and verify the

functionality of mathematical models developed using Matlab for

biosensors

Validation of developed mathematical models with biosensor models

and fine tuning mismatches

Development of golden software reference model for biosensors and

verifying its functionality for various changes in the input parameters

63

Figure 2.14. Methodologies for Expert System Design for Disease Classification

Use golden reference model developed for biosensor; develop sensor

array network and also the control unit

Setting up of suitable environment to simulate the sensor array in

Matlab to detect diseases

Develop expert system using Neural Networks and train the network

to classify and detect diseases

Interface sensor array model with the expert system and develop test

environment to verify the functionality of the developed model

Development of environmental setup to simulate and verify the

functionality of biosensor interfaced with the expert system

Design of hardware modules for the expert system and synthesize the

same on FPGA and identify its performances

Identify the performances of the developed expert system and optimize

the system for area, timing and power

64

The biosensors developed are further used to build an array of

sensors to detect various diseases; the sensor array is interfaced with

the expert system and is used in system development. Figure 2.14

show the methodology adopted to develop the expert system and drug

diffusion system.

The output of the expert system provides details of the disease

detected and based on the output generated a control unit is designed

to control the drug diffusion unit. A PID based controller is designed

to control the drug diffusion unit. The PID controller is designed and

implemented on FPGA. The design is optimized for area, power and

speed performances. The characteristics of PID controller are

estimated and the controller is interfaced with a motor to drive the

peristaltic pump to diffuse the drug stored. The feedback control unit

monitors the motor and also diffuses the drug accordingly. Detailed

discussion on biosensor modelling, biosensors lab, sensor modelling,

Matlab interface, design of expert system, PID controller and system

design is discussed in succeeding chapters.

2.9 Summary

Nanowire sensors based biosensors have been considered as one of

the appropriate methods for disease detection. Nanowire sensors have

been extensively used for detection of various diseases, they have also

been used to detect virus in the fluid, blood and cells. For system

development using Matlab, currently there are no biosensor models

available. Thus a methodology is developed in this chapter that helps

65

in developing an integrated software environment to model, simulate

and analyze the performances of biosensors and design of automated

drug delivery unit. Next chapter discusses design and development of

biosensors for disease detection.