Fabrication of cacbon nanotube based field effect transistors for DNA sensor application

Fabrication of cacbon nanotube based field effect transistors for DNA sensor application Fabrication of cacbon nanotube based field effect transistors for DNA sensor application Fabrication of cacbon nanotube based field effect transistors for DNA sensor application luận văn tốt nghiệp,luận văn thạc sĩ, luận văn cao học, luận văn đại học, luận án tiến sĩ, đồ án tốt nghiệp luận văn tốt nghiệp,luận văn thạc sĩ, luận văn cao học, luận văn đại học, luận án tiến sĩ, đồ án tốt nghiệp

International Training Institute for Material Science

PHUONG TRUNG DUNG

FABRICATION OF CARBON NANOTUBE BASED FIELD EFFECT

TRANSISTORS FOR DNA SENSOR APPLICATION

MASTER OF SCIENCE THESIS

Supervisors:

DR PHUONG DINH TAM

DR MAI ANH TUAN

Hanoi – 2012

ACKNOWLEDGEMENTS

First of all, I give deepest thank to my two supervisors, Dr Mai Anh Tuan and Dr Phuong Dinh Tam for giving me the opportunity to study on carbon nanotube electronic matter, thank you for your patience, support, faith, and encouragement

My thanks and honors to my lectures at Hanoi University of Science and Technology for mentoring me, and teaching me new things during my two years in master program at ITIMS

I wish to thank to the old and new members of research groups in ITIMS, especially biosensor and gas sensor group, who share valuable experiences and help

me a lot with kindness and friendship

My family, my friends, thank you for the joyful times, unforgettable moments, and thank for being my patient companions during all times

SUMMARY

The incidence of diseases cause by harmful pathogen has increased in recent years that threaten human lives Their appearance causes serious morbidity and immune-compromised to people Therefore, in require of society, many methods have been proposed so far to detect these pathogens

The conventional culture is a standard method which very sensitive and inexpensive and can give both qualitative and quantitative information regarding the analytical present in a sample However, it cost time and require skilled labor The molecular biology-based methods are highly specific and accurate but costly and also time-consuming

A simple, reliable device for analytical application has been developed as alternative Bio-sensing technology started to be most used method after PCR and ELISA due to many advantages including small sized, high sensitivity, short detecting time at low cost

The general objective of this thesis is the development of DNA sensor applying for detection of E Coli bacteria in which SWNTs are used as the transducer elements The sensor will be prepared by deposition of functionalized SWNT over patterned metal electrodes of a silicon wafer which had been pre-fabricated by simple photolithography The molecular recognition mechanism lies

in DNA-DNA hybridization in which single strand DNA probe/SWNT act as the sensing layers

This thesis is organized into three chapters

- Chapter 1 will cover the essential theoretical background; start with brief

introduction on physic of carbon nanotube and some of the detail of the physic for example hybridization Then, the short overview of the single

walled carbon nanotube based transistor (SWNTFET) is introduced Its basic properties such as electrical transport, SWNT-metal interfaces and gate hysteresis also presented

- Chapter 2 will summarize the preparation process of the SWCNFET sensors

and functionalization of carbon nanotube for final device Further, the electrical measurement setup and its used is introduced

- Chapter 3 after completely preparation, SWNTFET is applied to the

detection of E Coli virus sample The results characteristics will be analyze and discuss

- Finally, concludes of the whole thesis and the long-term propose or recommendation for future work will be given

TABLE OF CONTENTS

ACKNOWLEDGEMENTS 2

SUMMARY 3

LIST OF SYMBOL 10

LIST OF FIGURES 12

LIST OF EQUATIONS 16

LIST OF TABLES 17

CHAPTER 1 20

FUNDAMENTAL 20

1.1 Introduction to Carbon nanotubes 20

1.1.1 Electronic structure 20

1.1.2 Band structure 22

1.1.3 Electron mobility 24

1.2 Introduction to deoxyribonucleic acid (DNA) 25

1.2.1 DNA structure 25

1.2.2 Hybridization properties 26

1.3 Introduction to Carbon nanotube field-effect transistor (CNTFET) 27

1.3.1 The origins, MOSFET 27

1.3.2 CNTFET structure 29

1.3.3 CNTFET Working principle 29

1.3.4 Type of CNTFET 33

1.3.5 Application of CNTFET in biosensor 37

1.3.6 Conclusion 42

CHAPTER 2 43

EXPERMENTAL 43

2.1 Reagents 43

2.2 SWNTFET preparation 44

2.2.1 Mask design 44

2.2.2 Sensor fabrication process 45

2.2.3 Nanotube deposition on sensor surface 49

2.3 Equipment and measurement setup 51

2.3.1 Semiconductor parameter analyzer, Keithley 4200 52

2.3.2 Measurement setup 52

CHAPTER 3 54

RESULS AND DISSCUSSION 54

3.1 Device characteristics 54

3.1.1 I d -V ds characteristics of SWNTFET 54

3.1.2 I d -V gs characteristics of SWNTFET 55

3.1.3 Influence of environment exposing to sensor 57

3.2 SWNT – DNA immobilization characteristics 59

3.2.1 Dispersion of SWNTs in DNA solution 59

3.2.2 UV-Vis spectra characterization of SWNTs dispersed in DNA solution 60

3.2.3 Morphology of SWNTs dispersed in DNA solution 62

3.2.4 FTIR spectra of SWNT – DNA films 62

3.3 I d -V gs characteristics of probe DNA immobilization onto sensor 63

3.4 Label – free detection of DNA sequences using SWNTFETs 64

3.4.1 I ds -V ds characteristics 64

3.4.2 Sensitivity and response time of the sensor 66

CONCLUSION 68

REFERENCE 69

ARTICLES 75

HRTEM High Resolution Transmission Electron Microscope

ITIMS International Training Institute for Material Science

MOSFET Metal Oxide Semiconductor Field-Effect Transistor

NTFET Nanotubes Field-Effect Transistor

s-SWNT Semiconducting Single Walled Carbon Nanotube

SWNTFET Single Walled Carbon Nanotube Field Effect Transistor

ΦB, P Potential barrier for holes at the semiconductor/metal contact

q Elementary charge, i.e., magnitude of the charge of an electron

Wnom Nominal device channel width

LIST OF FIGURES

Figure 1.1: HRTEM published images of CNTs by Iijima (a) MWNTs [34], (b) Individual single-shell nanotubes [11] 20 Figure 1.2: Schematic illustrated bonding in SWNT lattice Localized (σ) and delocalized (π, π’) network 21 Figure 1.3: Example of building a nanotube by rolling up a graphene sheet (a) Possible lattice point in monatomic layer, Hamada [25] (b) Chiral vector and θ angle are unique for nanotubes, Saito [30] 22 Figure 1.4: First 2D Brillouin zone of graphene (shaded area) with high symmetric points K, M, and Γ and 1D Brillouin zone of a (6,3) carbon nanotube (parallel lines) Adapted from C Roman [7] 23 Figure 1.5: Calculated SWNT band diagram and density of states (DOS), (a) of a semi-conducting zig-zag (11, 0) (b) of a metallic armchair (6, 6), Code adapted from C Roman [7] 24 Figure 1.6: Schematic illustrated double-helix structure of DNA molecule 26 Figure 1.7: Schematic illustrated a traditional MOSFET structure (a, c) and electrical characteristic (b) Adapted from Ref [47] 28 Figure 1.8: Schematic described a nanotube field-effect transistor (CNTFET) device (a) Atomic force microscopy (AFM) image of a device consisting of a single semiconducting SWNT (b), Scanning electron microscopy (SEM) image of a device consisting of a random array of carbon nanotubes (c), Typical transfer characteristic – dependence of the source-drain conductance on the gate voltage (d) Adapted from Ref [2] 29 Figure 1.9: Schematic of the first back-gated CNTFET (a) and characterization (b) presented by Martel et al 1998 30

Figure 1.10: Schematic of the band alignments between metal and SWNT for metal with a low work function χs < φm < χs + Eg /2, where E g is the band gap of the s- SWNT for (a) V g = 0 V, (b) V g > 0 V, (c) V g < 0V, Adapted from Ref [7] 31 Figure 1.11: Schematic of the band alignments between metal and SWNT for metal with a high work function, Adapted from C Roman [7] 32 Figure 1.12: The structure of back-gate CNTFET (a), output characteristic of CNTFET with Vg = -6 V  5 V and transfer characteristic of back gate CNTFET for Vds= 10  100 V in steps of 10 mV (b, c) (Tans et al., 1998) 33 Figure 1.13: Back gated CNTFET with improved contacts (Martel et al., 2001) 34 Figure 1.14: Top gated CNTFET structure (a) and transfer characteristic of top gate CNTFET with Vds = -0.1 -1.1 V (b) (Wind, 2002 [36] 35 Figure 1.15 Schema illustrated a ambipolar CNTFET structure (a) and electrical characteristic of the potassium doping CNTFET (b) (Martel et al., 2001) 36 Figure 1.16: GAA CNTFET structure (Chen et al., 2008)[10] 37 Figure 1.17: Schematic diagram illustrates transfer curves of CNTFETs with (a) SWNTs-Au, (b) SWNTs-Cr, and (c) dual contact before immobilization and after immobilization and upon hybridization with DNA target sequence [12] 38 Figure 1.18: Schematic illustration of an NTFET coated with a biotinylated polymer layer for specific streptavidin binding (a) The source-drain current dependence on the gate voltage in the absence and presence of streptavidin (b), Adapted from Ref [41] 40 Figure 1.19: (a) CNTFET device for electronic monitoring of the enzymatic degradation of glucose starch after deposition, as well as after hydrolysis with amyloglucosidase (AMG (b) Isd-Vg characteristics from +10 V to – 10 V gate voltages with a bias voltage of 0.6 V (Adapted from Ref [24]) 41 Figure 2.1 Layout of a sensor 44

Figure 2.2: Steps of SWNTFETs fabrication process 45 Figure 2.3 a: Fabrication step 1 – wafer surface cleaning and natural silicon dioxide removal 46 Figure 2.3b: Fabrication steps 2 – Thermally silicon dioxide growth, thickness of SiO2 ~ 100 nm 46 Figure 2.3c: Fabrication steps 3 – Resist coating 47 Figure 2.3d: Fabrication steps 4 – Exposure and developing 47 Figure 2.3e Fabrication steps 5 – Cr and Pt deposition and lift-off, Cr thickness ~

15 nm, Pt thickness ~ 200 nm 48 Figure 2.3 f: Fabrication final steps 6 – Back-gate fabrication, Al thickness ~ 300

nm 48 Figure 2.4: Keithley Semiconductor Characterization System 4200 (including probe station) at HUST (a), Keithley Interactive Testing Environment (KITE) software with integrated CNTFET project for CNTFET device characterization (b) 52 Figure 2.5 Probe station used for electrical characterization of the CNTFET devices (a) Tungsten tips contacting the source, drain and gate electrode of the sensor (b) 53 Figure 3.1: Output characteristics of a SWNTFET, I d - V ds curves measured for Vgs=-5, -3, 0, 3, and 5V DC 54 Figure 3.2: Transfer characteristics of a fabricated SWNTFET, oxide thickness ~

100 nm 55 Figure 3.3: Effects of hysteresis on transfer characteristics of a fabricated SWNTFET exposed to ambient air (25 o C, moist 70%) before (a) and after adsorbed water removing (b) SiO 2 thickness of 100 nm 58

Figure 3.4: Photograph displaying of dispersed SWNTs in DNA solution at different sonication time (a) and different pH values (b) 15 mg SWNTs, 20 mM DNA, sonication power at 125 W, 120 minutes 60 Figure 3.5: UV-Vis spectra of dispersed SWNTs in DNA solution at different sonication times (I) and pH values (II) (Power: 125 W, 20 mM DNA, 15 mg CNTs) 61 Figure 3.6: FE-SEM (a) and TEM (b) images of CNTs dispersed in DNA solution after 2 months (power: 125 W, 20 mM DNA, pH7, 15 mg CNTs, 120 min) 62 Figure 3.7: The FTIR spectra of the SWNTs (a), nature DNA (b), DNA-CNT complex (c), power: 125 W, 20 mM DNA, pH7, 15 mg CNTs, 120 min 63 Figure 3.8: Schematic illustration of drain current dependence of gate voltage after probe immobilization 64 Figure 3.9: I d -V gs characteristics of complementary target DNA detection, exposed concentration of 5 µM 65 Figure 3.10: I d -V gs characteristic of a fabricated SWNTFET immobilized with probe DNA of E Coli bacteria, hybridized in complementary target sequences analyte of different concentration 5 µM, 0.5 µM, and 0.05 µM (a) sensor sensitivity deduced from concentration dependence of drain current curve (b) 66

LIST OF EQUATIONS

Chiral vector = n + m (1.1) 22

Circumference = a C-C (1.2) 22

Diameter = = (1.3) 22

(1.4) 24

LIST OF TABLES

Table 2.1: DNA sequences information 43 Table 2.2: The parameters of the sputtering process 48 Table 2.3: The parameters of the aluminum evaporation process 49

INTRODUCTION

In recent years, the outbreaks of pathogens such as A influenza or diarrhea that attracted the attention of the whole society The source of them can be from food, water, animals, environment, even the human body Thus, for health and safety reasons, investment in medical treatments increases continuously, and scientists and physicians are always at priority to find out the treatments, also to early detection and prevention

For detection of pathogen, several methods have been proposed Basically, conventional methods are always highly selective and sensitive but they also require several days to yield the results Enzyme-linked immuno-sorbent assays (ELISA) can be applied for the direct identification of pathogen in real samples This immuno-based method has been widely used in many fields with high sensitivity However, this method is time-consuming since a pre-enrichment of the sample is often required in order to achieve low limits of detection

Molecular biology-based methods, specifically polymerase chain reaction (PCR) and real-time PCR are nowadays the most common tools used for pathogen detection They are highly sensitive and allow the quantification of sample In addition, microarray platforms of DNA have been developed in order to analyze hundreds of targets simultaneously However, this technique is costly and reagent consuming due to the pre-processing steps of culturing bacterial cells and extracting the DNA before the amplification procedure

Since carbon nanotubes (CNTs) were discovered by Iijima, many publish has been reported about their unique electronic and optical properties and their nano size which make them valuable in the development of bio-sensing platforms Beside, their high capacity for charge transfer between heterogeneous phases makes them suitable as components in electrochemical sensors Finally, the highly

sensitive to changes in chemical environment is truly worthy for molecular recognition processes

Among numerous applications of carbon nanotubes, carbon nanotube based sensing device is rapidly emerging into an independent research field The promising advantage of these devices lies in high selectivity, sensitivities, and rapid results at low cost Moreover, they can operate at room temperature and in ambient conditions In Vietnam, at the beginning of this study, single walled carbon nanotube based sensor has not been successful development and applied Therefore,

it is expected that, this study will be a preliminary for long-term development

CHAPTER 1

FUNDAMENTAL

1.1 Introduction to Carbon nanotubes

Over a decade ago, since the carbon nanotubes were first discovered by the Japanese researcher Sumio Iijima [34], the interest and the enthusiasm for this new material have grown considerably [14] Recent advances in the synthesis and purification of carbon nanotubes have turned them into commercially available materials Subsequently, many experiments have been undertaken to study the physical and electrical properties of carbon nanotubes

(a) MWNTs [34], (b) Individual single-shell nanotubes [11]

1.1.1 Electronic structure

Carbon nanotubes (CNTs) are one-dimensional molecules with shape look like seamless hollow cylinders which made of rolled up sp2 graphitic carbon layer so-called graphene that is only composed of C-C bonds arranged in hexagonal honeycomb lattice structure Many characteristics of the material vary with the

(a) (b)

layout and arrangement of atoms in this structure Therefore, the graphene structure should be first described

There are two types of bonds in CNTs lattice called σ and π The σ bonds formed by the overlap of sp2 hybridization orbitals (the 2s orbital mixes with two 2p orbital to form three equity sp2) at their cone ends which leads to the planar structure (see Fig 2.2 below) Remember, each excited carbon atom have one unoccupied electron left in the non-hybridization orbital 2pz, which is perpendicular

to the σ bonds plane These electrons also form bonds that so-called π bonds with each other by side-by-side orbital overlap However, they are not strong enough as the σ to remain stabilized, identified network Consequently, they form a delocalized network which is responsible for electronic transport

Figure 1.2: Schematic illustrated bonding in SWNT lattice Localized (σ) and delocalized (π, π’) network

In the publication of Hamada et al from NEC Corp [25], only two integer called (n,m) are used to represent for unique order in the lattice They are multiplier coefficient of defined unit vectors in hexagonal honeycomb lattice of graphene By few minor changes, the more clearly describes were given by Saito et al., from MIT [30] not long after which is more common with scientific nowadays That is, the angle between two unit vectors has been replaced to 60o instead of 120o for easy usage And, their name once called and by Hamada et al [25] now called and then were brought together with (n,m) into one total vector so-called “chiral

π network, for conduction

vector” which point the direction to roll up a graphene sheet into nanotube(see Fig 1.3)

The length of the chiral vector corresponds to the circumference and diameter of the nanotube, given in equation(1.2) and (1.3), where aC-C is the carbon-carbon bond length, aC-C =1.44 Å

Figure 1.3: Example of building a nanotube by rolling up a graphene sheet (a) Possible lattice point in monatomic layer, Hamada [25] (b) Chiral vector and θ angle are unique for nanotubes, Saito [30]

1.1.2 Band structure

Similarly, to understand and describe the structural properties of SWNTs, the electronic properties of SWNTs can be considered to be derived from the electronic properties of graphene This procedure is called zone folding approximation [23, 29,

1], which roughly means that the electronic band structure of graphene is cut into

“pieces” of allowed wave vectors These sliced parts will be used to form the SWNT electronic band structure by supplementary boundary conditions included The band structure of graphene can be computed by solving Schrödinger’s equation for all electrons with some simplifications

Figure 1.4: First 2D Brillouin zone of graphene (shaded area) with high symmetric points K, M, and Γ and 1D Brillouin zone of a (6,3) carbon nanotube (parallel lines) Adapted from C Roman [7]

As showed in Fig 1.4, the Brillouin zone of a CNT is represented by many parallel line segments Therefore the energy bands consist of a set of one-dimensional energy dispersion relations in which the allowed energy states are cuts

of the graphene band structure When these cuts pass through a K point of the graphene’s Brillouin zone, the tube is metallic, otherwise is semiconducting If (n,m) is a multiple of 3, the one-dimensional energy bands have a zero energy gap, and an individual CNT is metallic [29] In contrast, if not, the one-dimensional energy bands have an energy gap, then that CNT is semi-conducting Thus, it is clearly to conclude that, armchair CNTs (n, m) are always metallic, and zigzag CNTs (n,0) are only metallic when n is a multiple of 3

Figure 1.5: Calculated SWNT band diagram and density of states (DOS), (a)

of a semi-conducting zig-zag (11, 0) (b) of a metallic armchair (6, 6), Code adapted from C Roman [7]

1.1.3 Electron mobility

The electronic transport in SWNTs is known to be ballistic A ballistic transport occurs in a finite semiconductor when the length of the conductive path (e.g distance between two electrodes) is shorter than the mean free path of electrons, so that electrons move through the material without any scattering process The scattering effects that create a reduction in carrier conductance (can be interpreted as increasing in total resistance of the nanotube) are attributed to electron interaction with impurities, intrinsic phonons, and other electrons For several micrometers SWNTs, the ballistic properties can be maintained, however,

on larger scales, scattering effects should be taken into consideration

Ballistic transport in CNT is possible due to the unique one-dimensional structure in which electrons move along the tube axis without backscattering However, the conductivity of the tubes has an absolute maximum as in all one dimensional transport systems The conductance through the tube is described by the Landauer relation:

Where “q” the elementary charge (in this case, the carrier is electron),

is called quantum conductance, and M is the number of active modes, and

“h” the Planck constant The nanotube intrinsic conductivity of is due to the spin degeneracy [19]

This means the maximal conductivity of 155 µS or minimal resistance of 6.5

kΩ However, in practice that almost never be archived due to parasitic additional contact resistances Some reports about CNTFETs indicate that with appropriate metal for electrodes, like Pd, Pt as well as a right fabrication method can reduce the total resistances to near theoretical limit [2, 3]

1.2 Introduction to deoxyribonucleic acid (DNA)

Since this work aims to develop DNA sensor to realize a fast, simple, and reliable detection, it is meaningful to investigate the specificity and ability of DNA for understanding of processes and mechanisms

1.2.1 DNA structure

Deoxyribonucleic acid (DNA) is a long polymer, which contains the genetic information in living organisms and some viruses The main function of DNA is to store and translate the information It plays an important role in rebuilding cells from generation to generation

A DNA molecule has a double-helix structure as shown in Fig 1.6 The sides

of the ladders are composed of alternating sugars (deoxyribose and phosphates); the rungs of the ladder are composed of nucleotides Nucleotides are basic building units of DNA molecule and different in nitrogen bases There are four kinds of nitrogen bases: Adenine (A), Thymine (T), Guanine (G) and Cytosine (C) Two specific pairs are exclusively formed between bases: A linked T by two hydrogen bonds (non-covalent); G linked C by three hydrogen bonds Each side of the ladder

is defined as a single strand and two long strands entwine like vines, in the shape of

a double helix It is a specific sequence of bases that transmits the genetic codes Hybridization between two strands is a selective recognizing process, thus one strand can be attached on the sensor tip as a probe strand to detect a complementary target strand in the solution

Figure 1.6: Schematic illustrated double-helix structure of DNA molecule

1.2.2 Hybridization properties

Each type of base on one strand bonds with just one type of base on the other strand, as A bonding to T, and C bonding to G This is called complementary base paring Hydrogen bonds between two strands can be broken and rejoined due to the external conditions At high temperature, hydrogen bonds turn to be thermodynamically unfavorable and complementary strands separate Under normal conditions, complementary strands re-anneal and form a “hybrid” molecule This is

a critical feature for DNA molecule to realize information translation from generation to generation Hybridization is a process that two complementary single DNA strands combine into a single molecule The specific interaction between complementary strands provides a feasible method to detect specific genes of DNA

Based on the exclusive combination between complementary DNA strands, label-free DNA detection scheme was proposed The surface of a micro-gap sensor was modified with single strands DNA of known sequence as a probe And, hybridization would happen if the probe was dipped into a complementary strand

DNA solution Since the hybridization between two strands DNAs are exclusive process, only the complementary DNA (target DNA strand) can hybridize with the probe DNA The successful hybridization resulted in the change of surface conductivity

1.3 Introduction to Carbon nanotube field-effect transistor (CNTFET)

The last five decades witnessed the amazing advancement of Silicon (Si) microelectronics Developments in Si microelectronics are achieved by miniaturization scaling the electronic components implemented on the ICs However, miniaturization of ICs are approaching its limit for several reasons including electron tunneling through short channels and thin insulator films, variations in device structure and doping, the effects of crystal misalignments and the increment of interface scatterings on the electron motion since the mean free path (MFP) of electrons become comparable to component dimensions From a circuit point of view, the performance degradation due to miniaturization appears as the increment in the resistance of metallic on-chip interconnects and short channel effects in MOSFETs By these disadvantages, therefore, for the near future, devices that have current-voltage characteristics similar to novel materials and nano-scale operating principles are explored (Avouris et al., 2004) These devices include graphene field effect transistors (FETs), silicon nanowire FETs and carbon nanotube FETs Among these, CNTFETs take more attention due to its superior electrical transport properties, thus, are proposed as the alternatives of the current bulk MOSFETs

1.3.1 The origins, MOSFET

The most famous field effect transistor is the MOSFET Semiconductor field effect transistor) It is a three terminal device based on Si which acts as a switch where the current between two of the terminals is controlled

(Metal-Oxide-by the voltage applied on the third First implemented in 1960s and then has been

widely employed in the integrated circuits and electronic applications such as microprocessors or memory designs Their basic structure is a combination of a capacitor and a transistor The transistor consists of the doped silicon layer in contact with two spatially distinct regions of doped silicon having opposite conductivity which called the source and drain terminals The capacitor is formed

by a dielectric layer of SiO2 placed between a semiconducting layer of p- or n- type doped silicon, and a metallic electrode or highly doped poly-silicon electrode known as “gate” electrode

(a)

(c) (b)

Figure 1.7: Schematic illustrated a traditional MOSFET structure (a, c) and electrical characteristic (b) Adapted from Ref [47]

The working principle of MOSFET is that whenever bias is applied to the gate electrode, there forms field strength across the insulating SiO2 layer and enhance or deplete the density of the mobile charge carriers in the doped silicon (holes in case of p-type or electrons in case of n-type) depending on the polarity of the field Therefore, the changes in the channel resistance of the transistor, consequently, can be detected in the measured conductivity between source and drain electrodes

1.3.2 CNTFET structure

The first carbon nanotube field effect transistors were reported in 1998 [4] The physical structure of CNTFET is very similar to that of MOSFETs This was fabricated by depositing CNT solution onto oxidized Si wafer that had been pre-patterned with gold or platinum electrodes The electrodes acts as source and drain that connected by using the nanotube channel The doped Si substrate served as the gate Fig 1.8 presents a CNTFET structure and transfer characteristic

Figure 1.8: Schematic described a nanotube field-effect transistor (CNTFET) device (a) Atomic force microscopy (AFM) image of a device consisting of a single semiconducting SWNT (b), Scanning electron microscopy (SEM) image of a device consisting of a random array of carbon nanotubes (c), Typical transfer characteristic – dependence of the source-drain conductance on the gate voltage (d) Adapted from Ref [2]

1.3.3 CNTFET Working principle

The operation principle of carbon nanotube field-effect transistors is similar

to the conventional MOSFETs This three terminal device consists of a semiconducting nanotube that acting as conducting channel, the source and drain terminal, the device is turned on/off through the gate terminal Gate voltage (V )

controls the conductive channel between source and drain When the voltage applied to the CNTFET’s gate is varied higher than threshold of device, the channel current is changed due to a semiconducting nanotube switches from the insulating state to the conducting state The device is turned on; otherwise the device is turned off

Figure 1.9: Schematic of the first back-gated CNTFET (a) and

characterization (b) presented by Martel et al 1998

1.3.3.1 Metal-CNT contact

At a metal/nanotube contact, right in the interface, there forms a potential barrier which has the ability to limit carriers come back and forth the junction It can either reduced or increased whenever a bias was applied between metal and semiconductor, then as a result, only the carriers with sufficient energy to overcome its height can pass through This barrier is usually called Schottkey barrier (SB), or Schottkey barrier height (SBH) As a result, the current came through the junction is different between forward (non-negligible current) and reserve polarity (negligible), that is the reason the metal/semiconductor contact so-called Schottkey-diode and the current versus voltage characteristic is always highly non-linear

In general, Schottky barriers (SB) were reported by most experiments [33,

32, 18, and 22] For the ambipolar case there is a voltage drop associated with each contact which forms the Schottky barriers However, nearly Ohmic contacts between a metal and a semiconducting CNT under some conditions are indicated by

(a)

some experiments It can be interpreted as the Schottky barrier to the valence bands

of a semiconducting CNT is equal to zero, and the Schottky barrier to the conduction bands is equal to the energy gap [1, 2] In some case, the contact conductance is even much higher than the channel conductance [44]

Figure 1.10 shows the energy diagrams at the interface of a semi-conducting SWNT and a low work function metal (e.g Al, φAl ≈ 4.2 eV) Φm is the metal work function (χs < φm< χs + Eg/2) , χs is the electron affinity of the SWNT, EF , Ei , Evand Ec is the Fermi energy of metal, the intrinsic Fermi energy, and the energy of the valence and conduction band edges of SWNT, respectively ΦnSB represent for the SB for the electrons and φpSB is the SB for holes Quantum mechanical tunneling through the barrier φB is one component of the electrical transport across a metal/SWNT interface with barriers while thermionic emission (at T > 0 K) over φB

is the other contribution As such, quantum mechanical tunneling depends on the width and thermionic emission on the height of the barrier φB

Figure 1.10: Schematic of the band alignments between metal and SWNT for metal with a low work function χs < φm < χs + E g /2, where E g is the band gap of the s- SWNT for (a) V g = 0 V, (b) V g > 0 V, (c) V g < 0V, Adapted from Ref [7]

For Vg> 0 V the bands in the energy diagram of Fig 1.10 (b) are moved down-wards, which brings the Fermi level closer to the conduction band within the bulk of the SWNT For a low WF metal, a reduced barrier results for electrons

while the barrier for holes is increased Therewith, thermionic emission of electrons is increased and at the same time it is decreased for holes Once Ec lines

up with EF, = = φm - χs and = Eg As Vg is further increased, the Fermi level is moved more closely to the conduction band and becomes thinner Consequently, at the same time, quantum mechanical tunneling of electrons also begins and the electron current through the SWNT increases

For Vg< 0 V, as shown in Fig 1.10 (c), the energy bands are moved upwards, which brings the Fermi level closer to the valence band As a result, becomes smaller and thinner, thus, hole conduction is dominant Once Ev lines up with

Master of Science Thesis Hanoi University of Science and Technology

1.3.4 Type of CNTFET

1.3.4.1 Back gate CNTFET

Utilization of CNTs as the main component of field effect transistors was realized in 1998 by Tan group (Tans et al., 1998) The structure of back gate was showed in Fig.1.12 This transistor has been built with a semiconducting single wall carbon nanotube bridging the platinum source and drain terminal, which are deposited on gate oxide film on a doped silicon wafer The Si wafer is used as a back gate

Figure 1.12: The structure of back-gate CNTFET (a), output characteristic

of CNTFET with Vg = -6 V  5 V and transfer characteristic of back gate

CNTFET for Vds= 10  100 V in steps of 10 mV (b, c) (Tans et al., 1998)

The practicality of this structure, however is not the best since the control of the back gate on the energy bands of the CNT channel is low due to the thickness of the gate oxide which has a value of 100 nm resulting in low conductance (1nS) Moreover, due to semiconducting CNT lays on the metal contacts causes high contact resistance (1MΩ) (Avouris et al., 2004)

(a)

These disadvantages of the first back-gated CNTFET structure directed researchers to develop CNTFET geometries with circuit parameters suitable for use in ICs A better CNTFET structure employs the method that patterns source and drain contacts on the top of CNT channel as shown in Figure 1.13 (Martel et al., 2001) Thermal annealing is also used to improve CNT-metal interface to decrease contact resistance The contact resistance of these type of devices are around 30kΩ, the saturation current is 1μA and the transconductance is around 0.3μS (Avouris et al., 2004)

Figure 1.13: Back gated CNTFET with improved contacts (Martel et al., 2001)

1.3.4.2 Top gate CNTFET

CNTFETs discussed so far use the conductive substrate as a back-gate electrode, usually with gate dielectrics of considerable thickness (~100 nm or more) As a result, high gate voltages are required to turn these devices on Thus, it

is obvious that, gate insulator thickness has to be small in order to achieve useable gate control A top-gated CNTFET structure was shown in Fig 1.14 [36] The CNT

is completely embedded within the gate insulator, offering the full advantage of the gate dielectric A further disadvantage of the open geometry is that exposure of CNTs to air leads to p-type characteristics In contrast, this top gate structure allows the fabrication of both n-type as well as p-type devices An additional advantage is that with only slight modification, it can be made suitable for high-frequency operation, which is not possible with back-gated devices due to the large overlap capacitance between the gate, source, and drain These features make the top gate devices the most technologically relevant CNT transistors fabricated so far, and they allow for a more direct comparison with mainstream silicon-based MOSFETs

The device shows excellent turn on and saturation at gate voltages ~1 V The maximum transconductance is 3.25 μS, which is an extremely high value for a CNTFET device as compared to previously reported CNTFETs The linearly extrapolated threshold voltage is –0.5 V and the inverse sub-threshold slope for top gate operation is ~130 mV/decade

a) Ambipolar CNFET

As in conventional Si MOSFETs, p-type and n-type CNTFETs can be obtained by doping Intrinsically, the CNT channel behaves as ambipolar, in other words permitting both electron and hole transport depending on the polarity

of the gate voltage

(b)

(a)

The capability to produce n-type transistors is important technologically, as it allows the fabrication of CNT-based complementary logic devices and circuits A back gated n-type nanotube transistor can be obtained by doping the nanotube with potassium vapors The details of the procedure are reported by Liu et al [50] the mechanism is that electron transfer from adsorbed potassium atoms to the nanotube can shift the Fermi level of the tube from the valence-band edge to the conduction-band edge, thus reverting the doping from p- to n-type [28]

Figure 1.15 Schema illustrated a ambipolar CNTFET structure (a) and electrical characteristic of the potassium doping CNTFET (b) (Martel et al., 2001)

Derycke et al [46] reported that p- to n-type conversion of the CNTFETs can

be made either by doping the surface of the tube using alkali metals as mentioned earlier by Liu et al [50] or by simply annealing the device in vacuum or in an inert gas The annealing in vacuum removes the adsorbed oxygen and results in the direct modification of the Schottky barrier height at the contacts In contrast, doping changes the barrier thickness and introduces significant shifts of the threshold voltage of the device The conversion from n-type to p-type with the intermediate ambipolar stages is fully reversible

b) Warp gate CNTFET

In order to achieve better gate control, CNTFETs with cylindrical geometric shape, also named as gate all around (GAA) CNTFETs, are proposed The geometry

of a GAA CNTFET is shown in Figure 1.16 (Chen et al., 2008) [10] Obtaining

n-type or p-n-type devices in GAA CNTFET structures is easy since polarity can

be set by doping the semiconducting CNTFET sections between gate insulator and source/drain contacts by n-type and p-type materials, respectively

The concept is totally difference compared to traditional MOSFET vertical gate configuration However, based on the prediction/research estimation being made by the researcher, they strongly believe that the GAA CNTFET is an alternative to Si-MOSFET due to it will provide a better device characteristics compared to 1-D ballistic MOSFET

Figure 1.16: GAA CNTFET structure (Chen et al., 2008)[10]

1.3.5 Application of CNTFET in biosensor

1.3.5.1 DNA sequence hybridization detection

To date, due to high biocompatibilities, the CNTFETs are a promising candidate for DNA sensor Subramanian et al reported a DNA sensor based on CNTFET arrays to detect Escherichia coli O157 [40] In this work, the CNTFET was fabricated by CNTs grown on 500 m thick silicon oxide substrate and contacted by two electrodes labeled source and drain The DNA probe was immobilized onto the CNTFET by covalent attachment The results showed that DNA CNTFET can determine DNA target concentration as low as 1pg.L-1 The carbon nanotubes network field effect transistors (NTNFETs) to detect DNA hybridization published Star et al [43] The DNA probe was immobilized on CNTs

by non-covalent attachment They found that the DNA hybridization with complementary DNA target sequence take place on the device surface that results in reduction of NTNFETs conductance This point was also demonstrated by Hwang