Research and synthesis of sno2 nanowirestransition metal dichalcogenides heterostructure towards gas sensing applications

TMDs gas sensors are particularly promising due to their ability to operate even at room temperature; therefore, a SnO2/TMD heterostructure gas sensor is believed to have an enhanced gas

HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

MASTER THESIS

Research and synthesis of SnO 2

nanowires/ transition metal

dichalcogenides heterostructure towards gas sensing applications

TRUONG TIEN HOANG DUONG

duong.tth202767M@sis.hust.edu.vn

Materials Science

Institute: International Training Institute for Materials Science

Hanoi, 05/2022

HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

MASTER THESIS

Research and synthesis of SnO 2

nanowires/ transition metal

dichalcogenides heterostructure

towards gas sensing applications

TRUONG TIEN HOANG DUONG

duong.tth202767M@sis.hust.edu.vn

Materials Science

Institute: International Training Institute for Materials Science

Hanoi, 05/2022

Supervisor’s signature

SOCIALIST REPUBLIC OF VIETNAM

Independence – Freedom - Happiness

CONFIRMATION OF MASTER’S THESIS ADJUSTMENT

Full name of author: Truong Tien Hoang Duong

Thesis topic: Research and synthesis of SnO 2 nanowires/ transition metal dichalcogenides heterostructure towards gas sensing applications

Major: Materials Science

Student ID: 20202767M

The author, the supervisor, and the Committee confirmed that the author has adjusted and implemented the thesis according to the report of the Committee on May 19th, 2022 with the following contents:

i

Acknowledgements First of all, I would like to express my deepest gratitude to my supervisor,

Dr Chu Manh Hung, for his thoughtful feedback and conscientious guidance throughout my master’s program His expertise and critical thinking in science is what I want to achieve in my future career I also extend my gratitude to Prof Nguyen Duc Hoa and Assoc Prof Nguyen Van Duy for their valuable suggestions and explanations regarding my research topic

I would also like to take this chance to express my special thanks to my very first advisor, Dr Le Minh Hai, whose perspective and knowledge have inspired and guided my research interests since my undergraduate days

In appreciation of my peers, I would like to thank my lab mates and classmates for their friendly and helpful collaboration, and especially Ho Huu Hau for accompanying me throughout experiments and discussions I am grateful to my senior Nguyen Ngoc Yen for her constant emotional support and mentoring that helped me overcome difficult times during my study To my best friend Tuc Anh, thank you for always being by my side

Last but not least, I would like to express my sincere gratitude to my family for their constant, unconditioned love and care I also acknowledge the financial and motivational support from Vingroup Innovation Foundation during the course

of my master’s program

oC) to exhibit the highest sensing performance, thus increasing power consumption In recent years, TMDs have become a rising star in many electronic applications and devices, because of their unique structural, chemical and electrical characteristics TMDs gas sensors are particularly promising due to their ability to operate even at room temperature; therefore, a SnO2/TMD heterostructure gas sensor is believed to have an enhanced gas sensing performance with a relatively low working temperature This thesis focuses on the fabrication of SnO2 NW/WS2

heterostructure gas sensors via physical approaches followed by a facile drop casting method The as-obtained heterostructure gas sensors showed enhanced gas response and selectivity towards NO2, and the working temperature is lowered compared to that of the SnO2 NW gas sensor This result contributes to the development of a new generation of gas sensors with high performance and low energy consumption

STUDENT

Truong Tien Hoang Duong

iii

TABLE OF CONTENTS

INTRODUCTION 1

CHAPTER 1 LITERATURE REVIEW 4

1.1 SnO2 nanowires 4

1.1.1 Crystal structure of SnO2 4

1.1.2 Electrical properties of SnO2 4

1.1.3 On-chip growth of SnO2 nanowires by thermal evaporation 5

1.1.4 SnO2 NW gas sensors 5

1.2 Two-Dimensional Transition Metal Dichalcogenides (WS2) 7

1.2.1 Crystal structure of WS2 8

1.2.2 Electrical properties of WS2 8

1.2.3 Fabrication methods 9

1.2.4 WS2-based gas sensors 12

1.3 SMO/TMD heterostructure gas sensors 13

CHAPTER 2 EXPERIMENTAL DETAILS 16

2.1 Synthesis of on-chip SnO2 NWs by evaporation method 16

2.1.1 Fabrication of interdigitated Pt chips 16

2.1.2 Fabrication of on-chip SnO2 NWs by the evaporation method 16 2.2 Synthesis of TMD nanomaterials 17

2.2.1 Fabrication of WS2 NSs 17

2.3 Synthesis of SnO2 nanowire/ TMD nanomaterial heterojunction gas sensors 18 2.3.1 Fabrication of WS2@SnO2 gas sensor 18

2.4 Characterization methods 18

2.4.1 X-Ray Diffraction 19

2.4.2 Raman spectroscopy 19

2.4.3 Ultraviolet-Visible Spectroscopy (UV-Vis) 20

2.4.4 Scanning Electron Microscope and Energy-Dispersive X-ray Spectroscopy 20

2.4.5 Transmission Electron Microscopy and Selected-Area Electron Diffraction 20 2.5 Gas sensing measurement 21

CHAPTER 3 RESULTS AND DISCUSSION 23

3.1 SnO2 NW gas sensor 23

iv

3.1.1 Structure and morphology 23

3.1.2 Gas sensing performance 24

3.2 WS2 NS gas sensor 28

3.2.1 Influence of solvent in the LPE of WS2 29

3.2.2 Structure and morphology 29

3.2.1 Gas sensing performance 30

3.3 SnO2@WS2 gas sensor 36

3.3.1 Structure and morphology 36

3.3.2 Gas sensing performance 39

CONCLUSIONS AND FUTURE WORK 45

LIST OF PUBLICATION 46

REFERENCES 47

v

LIST OF ABBREVIATIONS TMDs Transition Metal Dichalcogenides

FESEM Field Emission Scanning Electron Microscopy

HRTEM High Resolution Transmission Electron Microscopy SAED Selected-Area Electron Diffraction

EDX Energy-dispersive X-ray spectroscopy

JCPDS Joint Committee on Powder Diffraction Standards sccm Standard cubic centimeters per minute

ppb Parts per billion

vi

LIST OF FIGURES

Fig 1.1 Unit cell of rutile SnO2 [18] 4

Fig 1.2 Different structures of TMDs: a) 1H, b) 1T, c) 1T’, d) 1T”, e) 2H, and f) 3R [60] 8

Fig 1.3 Layer-dependent band structures of WS2 ultrathin films [65] 9

Fig 1.4 (a) Step-by-step process of TMD hydrothermal synthesis and (b) sequential phenomena appear in hydrothermal reactions [69] 10

Fig 1.5 Liquid-phase exfoliation routes for 2D TMDs Adapted from Ref [85] 11

Fig 1.6 (a-b) TEM and HRTEM of WS2-WO3 hybrid Gas responsive curve to 5 ppm at 25 °C of (c) WS2 and (d) WS2-WO3 gas sensor (e) Responsive curve to different NO2 concentrations of WS2-WO3 gas sensor at 25 °C, (f) Energy band structure of WS2-WO3 heterojunction [110] 14

Fig 2.1 Interdigitated Pt electrodes 16

Fig 2.2 Schematic diagram of the on-chip fabrication of SnO2 NWs 17

Fig 2.3 Schematic illustration of the LPE of WS2 NSs 18

Fig 2.4 Fabrication of WS2@SnO2 gas sensors by drop casting 18

Fig 2.5 (a) Gas sensing measurement setup at ITIMS and (b) Inside the test chamber 21

Fig 3.1 XRD spectrum of SnO2 NWs grown on interdigitated Pt electrode 23

Fig 3.2 (a-c) SEM images of on-chip SnO2 NWs, with inset showing the corresponding interdigitated electrodes (d) EDX spectrum of SnO2 NWs 24

Fig 3.3 (a) Transient resistive curves to different NO2 concentrations of SnO2 NW gas sensor and (b) the corresponding gas response at different temperatures, (c) response and recovery times of the SnO2 NW gas sensor at 200 °C 25

Fig 3.4 I-V characteristic of SnO2 NW gas sensor 26

Fig 3.5 (a) Linear fit of gas response as a function of NO2 concentration and (b) Arrhenius polynomial fit of 40 base resistance data points of SnO2 NW gas sensor at 200 °C 27

Fig 3.6 Selectivity of the SnO2 NW gas sensor at 200 °C 27

Fig 3.7 NO2 sensing mechanism of SnO2 NW gas sensor 28

Fig 3.8 UV-Vis absorption spectra of exfoliated WS2 using different solvent ratios 29

Fig 3.9 SEM images of (a) bulk WS2 and (b) exfoliated WS2 NSs 30

Fig 3.10 TEM images of (a) exfoliated WS2 NS, with (b) higher magnification of the yellow-dashed region (c) HRTEM image of the red-dashed region (d) SAED pattern of WS2 NSs 30

Fig 3.11 (a) XRD spectrum and (b) Raman scattering spectroscopy of bulk and exfoliated WS2 NSs 30

vii

Fig 3.12 (a) I–V characteristics of WS2 NS gas sensor (b) Arrhenius fit of the Ln(I(A)) versus 1/T of the exfoliated WS2 nanosheets 30Fig 3.13 (a) Transient resistances of WS2 NS gas sensor at various NO2

concentrations and temperatures, (b) gas response at different temperatures and (c) response time and recovery time at RT of WS2 NS gas sensor as a function of NO2

concentration 31Fig 3.14 Resistive response of bulk WS2 gas sensor to 5 ppm NO2 at RT 32Fig 3.15 (a) Linear fit of gas response and (b) Fifth polynomial fit of 15 data points of base resistance of WS2 NS gas sensor at RT 33Fig 3.16 (a) Resistance curves to 5 ppm NO2 and (b) corresponding changes in resistance and response of WS2 NS gas sensor under different room humidity at

RT 33Fig 3.17 Selectivity of the WS2 NS gas sensor among different gases 34Fig 3.18 Stability of WS2 gas sensor to 5 ppm NO2 at RT at (a) fresh condition and (b) after 60 days 34Fig 3.19 NO2 sensing mechanism of WS2 NS gas sensor 36Fig 3.20 (a) XRD pattern of WS2@SnO2 and (b) Raman spectra of SnO2 NW and

WS2@SnO2 37Fig 3.21 (a-d) SEM images of WS2@SnO2 gas sensor 38Fig 3.22 EDX pattern and atomic percentage of WS2@SnO2 Inset is the SEM of the examined area 38Fig 3.23 (a) TEM image and (b) corresponding SAED pattern of WS2@SnO2 39Fig 3.24 (a) A response curve to 5 ppm NO2 at 50 °C of four equivalent SnO2

NW gas sensors, (b) comparison of the four gas sensors after drop casting of WS2

NSs (the SnO2 NW gas sensor was denoted SW0) 39Fig 3.25 (a-c) Dynamic response curve of SW1, SW2, and SW3 to different NO2

concentrations from 50 to 150 °C and (d-f) the corresponding gas responses as a function of NO2 concentration, respectively 40Fig 3.26 (a) Stability and (b) Response and recovery time of sample SW2 at optimal working temperature (100 °C) 41Fig 3.27 (a) Comparison of gas response to 5 ppm NO2 as a function of temperature of SnO2, WS2, and WS2@SnO2 gas sensors, and (b) comparison of their sensitivity to NO2 at optimal working temperature 41Fig 3.28 (a) Comparison of selectivity between SnO2, WS2, and WS2@SnO2 gas sensors as percentages of gas response to 5 ppm NO2 at corresponding optimal temperatures 42Fig 3.29 NO2 sensing mechanism of WS2@SnO2 gas sensor 44

viii

LIST OF TABLES Table 1.1 A summary of different SnO2 NW gas sensors and their variations The asterisk-marked publications are from our research group 7Table 1.2 Advantages and disadvantages of different fabrication methods for TMDs Adapted from Ref [95] 12Table 3.1 Comparison of gas response towards NO2 of various WS2-based gas sensors with different morphologies and synthesis methods 35Table 3.2 Comparison of NO2 response and working temperature among various hybrid SMO/TMD and other functionalized sensors 43

1

INTRODUCTION Motivation

The action of breathing is indispensable to living creatures Besides nitrogen and oxygen that takes up 99% of air, the remaining minor proportion is mostly argon (0.93 %), CO (0.038 %), and trace amounts of other gas molecules Despite having such low concentrations in air, many gas molecules such as NO2, SO2, CO,

H2S, and NH3 can be extremely dangerous to humans as well as other living organs For example, NO2 is a common air pollutant from combustion processes

in fossil fuel vehicles or internal gas stoves An exposure to only 5 part-per-million (ppm) of NO2 can severely harm the respiratory system, and in an extended duration may result in diseases such as lung cancer, asthma, or even death [1,2] Over the last century, the level of technology and engineering of humankind has been significantly developed in every industrial sector As a result, more and more toxic compounds have been released into the environment, posing a detrimental threat to the well-being of society Exposure to an excessive concentration of these toxic gases can directly damage human tissues and organs, creating serious or even lethal illnesses In 2019, air pollution became the most threatening problem to global health, according to World Health Organization (WHO) [3] Moreover, on a larger scale, an uncontrollable amount of toxic gases can be extremely harmful to the environment as well as to the integrity of the whole ecosystem Examples can be seen in the mass death of birds caused by toxic gases from wildfires in Colorado [4], or in acid rains caused by NO2 and SO2 gases [5] Therefore, the high risks and damage from air pollution have brought about the demand for modern gas detecting and sensing devices as a way to effectively monitor the environmental status

There are many types of gas sensors, such as electrochemical, magnetic, optical, calorimetric, or mass sensitive Among them, electrical gas sensors (or resistive/conductometric) are one of the most widely studied gas sensors for their simple structure and ease of fabrication The transducer part of resistive gas sensors records and analyzes the change in resistance or other electrical properties

of sensing materials when exposed to analytical gases Meanwhile, the sensing layer often involves the use of nanomaterials whose sizes are comparable to those

of gas molecules Semiconducting nanomaterials are among the most researched material groups for gas sensing devices because their electrical properties can be easily manipulated and measured

Semiconducting metal oxides (SMOs) represent the largest proportion of resistive gas-sensing materials due to their ease of fabrication, high sensitivity, and excellent stability More than 20 % of the gas sensor market was attributed to SMO gas sensors [6] Specifically, SnO2 is considered the most popular and attractive SMO material for gas sensor The first ever commercial gas sensor was invented

by Naoyoshi Taguchi in 1968, using a semiconducting SnO2 film as the sensing material [7,8] SnO2-based gas sensors have excellent gas sensing characteristics

energy-So far, there have been several studies that reported the enhancement in gas sensing characteristics of SMO/TMD heterojunction [14–16] However, up to now, there have not been any particular studies concerning the heterostructure of TMDs with SnO2 nanowires and the effect that it may create to the gas sensing performance of the said hybrid material structure Therefore, we decided to choose the research topic for this thesis with the title: “Research and fabrication of SnO2

nanowires/ transition metal dichalcogenides heterostructure towards gas sensing applications”

Research Objectives

The main objectives of this thesis are:

(i) To successfully fabricate SnO2 nanowires and WS2 nanosheets and

prepare WS2@SnO2 gas sensors by a simple drop casting method (ii) To investigate the gas sensing performances of the WS2@SnO2

heterostructure gas sensors and compare them to those of each individual component gas sensor

Research method

To achieve the proposed research objectives, the research reported in this thesis was conducted following several methods, namely:

(i) The SnO2 NWs were fabricated on-chip via a thermal evaporation

technique The WS2 NSs were fabricated using a liquid phase exfoliation process

(ii) The WS2@SnO2 heterostructure gas sensor was prepared by drop

casting Different relative quantities of each component were conducted

in order to investigate the effect and find out an optimal parameter

3

(iii) The crystal structures and morphologies of the obtained samples were

studied by X-ray diffraction, Raman Scattering, Scanning Electron Microscope, Transmission Electron Microscope, and Energy-Dispersive Spectroscopy

(iv) The gas sensing performances were investigated using the gas-sensing

measurement system at the Research and Development of Nanosensors Laboratory (ITIMS, HUST)

Scientific contribution of the thesis

Conventional SnO2 gas sensors possess good gas sensing characteristics, but still exhibit several drawbacks, especially the high working temperature Meanwhile, the recent TMD-based gas sensors are capable of sensing different gases at room temperature, despite having lower gas response and recoverability

By investigating the influence of the SnO2 NW/TMD heterostructure in gas sensing performance, we believe that the research conducted in this thesis will contribute to the general understanding of the structure concerned, as well as help

to produce an energy-saving, highly sensitive gas sensor generation in the future Thesis outline

The theoretical and experimental research related to the proposed research was reported in this thesis in a 4-chapter organization Specifically:

i Chapter 1 (Literature review) gives general knowledge about semiconducting gas sensors as well as structural and electrical properties of SnO2 and WS2 nanomaterials The literature about each material and their heterostructure in gas sensing applications was also reviewed

ii Chapter 2 (Experimental details) includes details about the experimental procedures conducted during the study The fundamental mechanisms of characterization and gas-sensing measurement methods related to this study were also provided

iii In Chapter 3 (Results and discussion), the structural and morphological results of the obtained gas sensors were presented and discussed in detail Next, the gas sensing performances of the concerned materials were analyzed and compared thoroughly, and a gas sensing mechanism corresponding to each material was proposed Finally, we concluded the main findings presented in the thesis, and gave a few recommendations to further improve the current study

4

CHAPTER 1 LITERATURE REVIEW

In this chapter, we will present fundamental aspects of semiconducting gas sensors, SnO2 nanowires gas sensors, and WS2 gas sensors The crystal structure and electrical properties of each material shall be discussed in detail Moreover, different fabrication methods corresponding to each material will be compared, with the emphasis on the thermal evaporation of SnO2 NWs and liquid-phase exfoliation of WS2 NSs Finally, the characteristic of WS2@SnO2 heterostructure

as well as their gas sensing applications published in the literature will be fully reviewed

1.1 SnO2 nanowires

1.1.1 Crystal structure of SnO2

SnO2 can exist in various polymorphs, such as the rutile-type (P42/mnm), pyrite-type (Pa3), CaCl2-type (Pnnm), α-PbO2-type (Pbcn), fluorite-type (Fm3m), and ZrO2-type orthorhombic phase I (Pbca) or cotunnite-type orthorhombic phase

II (Pnam) [17] Nevertheless, the most stable and commonly available structure of SnO2 is the rutile phase

The tetragonal rutile structure of tin oxide is specialized by a P42/mnm space group and the lattice constants a = b = 4.7374 Å and c = 3.1864 Å [18], as illustrated in Fig 1.1 A unit cell of stoichiometric SnO2 consists of six atoms, with two tin ions in the +4 oxidation state and four oxygen ions in the typical -2 oxidation state The cations are located at the corners (0,0,0) and the center (½, ½,

½) of the unit cell, while the anions are located at the positions (x, x, 0) and (½+x,

½-x, ½), with x = 0.307 Each tin atom is coordinated by six oxygen atoms, and each oxygen atom is coordinated by three tin atoms

Fig 1.1 Unit cell of rutile SnO2 [18]

1.1.2 Electrical properties of SnO2

Non-stoichiometric SnO2 exhibits n-type semiconducting behavior, with a wide direct bandgap of 3.6 eV [19] This electrical property is resulted from the existence of oxygen vacancies that creates donor states in the band gap, specifically 30-34 meV and 140-150 meV below the conduction band as denoted

ED1,2 in the band diagram [20] Therefore, the band gap of SnO2 as well as its electrical properties may vary widely between different fabrication methods and morphologies due to the difference in oxygen vacancies [21]

5

1.1.3 On-chip growth of SnO2 nanowires by thermal evaporation

One of the reasons why SnO2 nanomaterials have been extensively studied is

their diversity of fabrication methods and geometries 1D SnO2 nanofibers and

nanotubes can be prepared via electrospinning [22,23], 2D SnO2 thin films can be

synthesized by sputtering technique [24,25] or chemical vapor deposition [26],

while various SnO2 nanostructures prepared by different hydrothermal processes

were also widely reported [27–30] However, it is well known that the fabrication

of SnO2 nanowires fabricated by direct thermal evaporation technique (also known

as on-chip SnO2 nanowire growth) is one of the easiest and most straightforward

methods Sears was the first to report and explain the growth of fine mercury

whiskers using the thermal evaporation technique in 1955 [31] He observed that

the growth of mercury whiskers was anisotropic, meaning that the length of the

whiskers increased without any significant widening in diameter when the growing

time increased

In general, on-chip fabrication of SnO2 NWs as well as any other SMO NWs

can be classified into two mechanisms:

(a) Vapor-Liquid-Solid (VLS) growth mechanism

The so-called VLS growth mechanism of Si NWs was first proposed by Ellis

and Wagner in 1964 [32] In a VLS growth process, the vapor can be quickly

adsorbed on a liquid metallic catalyst, which can be gold (Au), nickel (Ni), or iron

(Fe), and after the nucleation sites are created, the unidirectional growth of 1D

nanowires will occur as a result of supersaturation

Many VLS processes have been used to fabricate various 1D nanomaterials

with precisely designed positions according to the structure of the deposited metal

catalyst film The growth of SnO2 nanostructures is achieved by the presence of a

tin or tin oxide powder source, with Argon as the carrier gas and Au as the most

popular catalyst To control the size of the nanowires, the size of the liquid catalyst

as well as the quality of the substrate must be carefully considered The length of

the wires as well as the density are controlled by the growth rate, which can be

affected by powder source, source-substrate distance, O2 flow, growth

temperature, growth time, etc (b) Vapor-Solid (VS) growth mechanism

In many cases, the growth of 1D nanostructures can take place directly on

solid surface instead of a liquid nucleation site as in a VLS process This is called

Vapor-Solid (VS) growth mechanism Compared to VLS growth, the VS growth

requires higher energy, i.e higher temperature, for the vapor material to adsorb

directly on the solid surface However, they do not require an additional deposition

of catalysts, thus simplifying the fabrication process

1.1.4 SnO2 NW gas sensors

SnO2 was the first ever material that was applied in commercial gas sensor

[33], and the number of publications related to SnO2-based gas sensors has been

still increasing over the last decade [9] In terms of searching expressions, the

SnO2 NW gas sensors can be designed in either single-wire type [37–39] or multijunction network type [40–42] Although single NW gas sensors can take advantage of their excellent mono crystallinity, low dimension, and straightforward sensing mechanism to exhibit unique gas sensing characteristics [37], networked NW gas sensors are likely to be more sensitive and reliable [43]

In many years, numerous SnO2 gas sensors have been prepared and reported

in literature In general, SnO2-based gas sensors are low cost, easy to prepare, exhibit excellent sensitivity and stability over various gases and volatile organic compounds (VOCs) A comprehensive review about on-chip SnO2 NW gas sensors as well as other SMO NW gas sensors was given by Hung et al [44] Table 1.1 shows the working temperature and gas response of different SnO2 NW-based gas sensors It is evident that most SnO2 NW-based gas sensors require a high operating temperature (more than 150 °C) to exhibit optimal gas sensing performances Some publications have suggested different methods such as using

UV radiation or complementing another material to lower their working temperature or enhance gas response [45][46] Meanwhile, researchers also seek for new sensing materials that can intrinsically interact with toxic gases at low temperatures as well as their combination with conventional SMO gas sensors in order to produce new types of gas sensors with high performance and low power consumption, which will be discussed thoroughly in the next sections

7

1.2 Two-Dimensional Transition Metal Dichalcogenides (WS2)

TMD materials are an interesting material group, among other 2D materials such as graphene, MXenes , etc [55–57] Their names are derived directly from their atomic formula MX2, where M is a transition metal such as molybdenum (Mo), tungsten (W), and X is a chalcogen element (S, Se, Te) When the transition metal is in the group 10 such as Pt or Pd, they are referred to as Noble Metal Dichalcogenides (NMDs), although the number of studies on NMDs so far is still quite low due to their difficulties in fabrication [58]

TMDs have a 2D nature due to their layered crystal structure Transition metal atoms and chalcogen atoms exist in graphene-like honeycomb atomic planes separately, and two chalcogen atom planes together with a transition metal atom plane in between form a TMD layer Within each plane, atoms bond with each other by a strong covalent bonding [59] with no dangling bonds, whereas different atom planes are held together by a weak van der Waal bonding Therefore, they can exist in stable ultrathin 2D structures of 2-3 layers or even single layer Depending on the relative position of atoms in each plane, different crystal structure of TMDs can be generated, as shown in Fig 1.2 Commonly, the 2H

Table 1.1 A summary of different SnO2 NW gas sensors and their variations The asterisk-marked publications are from our research group Materials and

structure

Working Temp.(°C)

Target gas

8

TMDs, in which different layers of TMDs are placed exactly on top of each other, exhibit semiconducting behaviors, whereas the 1T or 1T’ phases exhibit metallic nature [60]

Fig 1.2 Different structures of TMDs: a) 1H, b) 1T, c) 1T’, d) 1T”, e) 2H,

and f) 3R [60]

1.2.1 Crystal structure of WS2

The most stable form of WS2 is the 2H phase, which belongs to the hexagonal space group P63/mmc In the case only a single layer of WS2 presents, it is called 1H phase When the coordination of W changes from trigonal prismatic to octahedral, it forms a metal-stable 1T WS2 phase The interlayer distance of WS2

is around 6.5 Å

1.2.2 Electrical properties of WS2

The most common and stable 2H WS2 phase exhibits a p-type semiconducting behavior, with band gap varied according to the number of layers Bulk WS2 has an indirect band gap of approximately 0.88 eV [61], and with decreasing number of layers, the band gap of WS2 as well as other behaviors increase accordingly [62,63] Interestingly, when thinning down to monolayer, also known as the 1H phase, it has a direct band gap of about 1.95 eV [64] The indirect-to-direct band structure transition of few-layer to monolayer WS2

calculated with or without spin-orbit coupling are described in Fig 1.3

9

Fig 1.3 Layer-dependent band structures of WS2 ultrathin films [65] 1.2.3 Fabrication methods

So far, 2D TMDs have been fabricated using several techniques, depending

on the desired structure and applications They can be divided into “bottom-up approaches” – where nanomaterials are built up from atoms or molecules, and

“top-down approaches” – where nanomaterials are fabricated from the breaking down of bulk materials

(a) Bottom-up approaches

Chemical vapor deposition (CVD) is an effective bottom-up technique to synthesize large-area TMD ultrathin films with high degree of freedom and high uniformity desired for fundamental condensed matter physics study and electronic devices The main growing mechanism of CVD involves the reactions of gaseous

or vaporized materials at elevated temperature to crystalize and form solid products on a substrate There are two main routes for synthesizing TMDs by CVD: The first one involves the reaction of sulfur or selenium vapors with a pre-deposited metal or metal oxide thin film on a substrate (also known as direct sulfurization or selenization) In the case of selenization, H2 may be required to facilitate the reduction of Se due to its low reactivity The second one also involves the vaporization of chalcogen precursors at the upstream zone of the quartz tube, but the metal source is in the form of powder precursors and is placed at the downstream zone Zhou et al reported a salt-assisted CVD technique to produce

47 different TMD materials and heterostructures, in which the salts (NaCl, KI) were used to lower the melting temperatures of metal oxide precursors [66] Nevertheless, 2D thin films synthesized by CVD still suffers from low scalability and low reproducibility, thus limiting its practical application potentials

Traditionally, 2D TMD nanostructures can also be fabricated by the hydrothermal method, which are particularly suitable for electrode materials in supercapacitors and batteries [67–69] The hydrothermal synthesis of TMDs

Fig 1.4 (a) Step-by-step process of TMD hydrothermal synthesis and (b) sequential phenomena appear in hydrothermal reactions [69]

Hydrothermal synthesis is widely applied in many studies because of its low cost, large scale, and their diverse nanostructures available The most common morphology of hydrothermally-prepared TMDs is nanoflowers [70–73], which consist of 2D TMD nanosheets aggregating together The hydrothermal method is especially useful to construct different heterostructures or modify the surface of materials, but its susceptibility to experimental conditions and dangerous risks still remain

(b) Top-down approaches

2D nanomaterials, in terms of fabrication method, have an intriguing advantage over other types of low-dimensional (LP) materials such as 0D, 1D, or 3D: That is, 2D nanomaterials have their own layered forms, i.e the bulk counterparts, whereas the other LP materials do not own Therefore, 2D materials can be fabricated via a top-down approach besides the conventional, well-known bottom-up synthesis of nanomaterials This ease of fabrication does not hold true for other types of LP materials [74]

TMDs can be exfoliated either physically or chemically In most cases, exfoliation techniques can generate large scale, free-standing and near-theory quality 2D TMD nanosheets, in contrast to CVD-synthesized TMD thin films, whose properties might be largely altered by the morphology and strains on the substrate [61] The simplest exfoliation technique – mechanical exfoliation – involves the use of scotch tape to exfoliate TMD ultrathin films directly from the bulk material It was well-known that mechanically exfoliated TMD films exhibit

11

much higher quality than that of CVD-prepared TMD films [75] However, the low yield and low uniformity of the mechanical exfoliation process significantly hinders their application in practical uses [76] Therefore, other exfoliation techniques have been developed to produce 2D TMDs with higher yield and more uniform size [77]

In addition to the mechanical exfoliation method, the exfoliation of 2D TMD nanosheets mainly takes place in an aqueous medium as it can be applied in many positions simultaneously instead of layer by layer There are two main routes of liquid-phase exfoliation (LPE) routes as illustrated in Fig 1.5 In a chemical approach, TMD nanosheets can be exfoliated from bulk materials by the intercalation of lithium ions (the so-called Li-intercalated exfoliation method) When the ions intercalate into the layered structure of TMD bulk materials, the interlayer van der Waals bonding is weakened by the gaseous H2 gaseous bubbles generated from the hydration reaction of alkaline compounds with water, leading

to a large expansion of the crystal structure Other chemical exfoliation strategies for TMDs have also been developed, following the same mechanism [78,79] Several chemically exfoliated TMD nanosheets have been reported with significantly enhanced properties [80–82] Nevertheless, the intercalation of ions are time consuming and may cause structural deformation such as winkles as well

as may create doping effects that irreversibly alter the pristine conditions of obtained TMD nanosheets [83,84]

as-Fig 1.5 Liquid-phase exfoliation routes for 2D TMDs Adapted from Ref

[85]

While intercalation exfoliation methods rely on the structural changes within adjacent layers, a more direct and physical approach involves the use of external sonication, which is called sonication-assisted exfoliation, but often referred to simply as LPE In a typical LPE process, TMD bulk material is mixed in a solution, and a sonication probe is immersed in the mixture The exfoliation mechanisms are attributed to the collapse of high-intense vacuum bubbles followed by micro jets and stress waves created during the sonication [86] Interestingly, the matching

of surface tension components between the 2D material and that of the solvent is supposed to play a critical role It was reported by Shen et al that a minimal

12

distinction between the two will prevent the 2D materials from reaggregating [87]

On the other hand, Jawaid et al stated that a close value of surface tension does not fully explain the high yield of different solutions in LPE processes, but rather

a chemical mechanism in each solvent itself as well [88]

Many organic solvents, such as N-methyl-2-pyrrolidone (NMP), dimethylformamide (DMF), or dimethyl sulfoxide (DMSO), have been proven to

N,N-be suitable for the exfoliation of 2D nanomaterials [89] However, their high boiling temperatures (typically > 200 °C), high cost, and high toxicity to the environment have hindered their practical applications [90] Therefore, more and more LPE approaches to a more environmentally friendly fabrication have been reported [91–93] Zhou et al proposed a strategy where an environmentally friendly mixed solution of water and ethanol with different compositions can be efficiently used for LPE of any 2D nanomaterials [94]

In summary, the advantages and disadvantages between different fabrication methods for TMDs are given in Table 1.2

1.2.4 WS2-based gas sensors

Besides a wide range of applications such as energy storage [96,97], hydrogen evolution reaction [98], biosensors [99,100] and optoelectronics [101,102], TMDs have also been extensively studied for gas sensing applications TMDs are particularly promising for low-temperature gas sensing devices, due to their unique physical and electrical properties With the ability to detect and measure trace concentrations of different toxic gases without an external heat source, they are excellent sensing materials for low-power and even flexible gas sensors, which show great potential in the Internet-of-Things (IoT) era [11,103]

Table 1.2 Advantages and disadvantages of different fabrication methods

for TMDs Adapted from Ref [95]

exfoliation

Solution-processable, no phase transition

Wide thickness distribution, low yield, difficulty in removing the organic solvent Intercalation-

based liquid

exfoliation

Solution processable, larger lateral size, high yield

intercalation-Longer time, more steps

CVD growth High crystallinity, larger

lateral size

High temperature

Wet-chemical

synthesis

Solution-processable Synthesis condition sensitive,

small lateral size

13

Their gas sensing characteristics come directly from the charge transfer process at surface defects of chalcogen atoms Along with a high surface-to-volume ratio, they are able to sense various toxic gases even at room temperature

However, like other TMD-based gas sensors, WS2-based gas sensors still suffer from low gas response and slow or incomplete recovery, which pose a practical challenge to compete with conventional SMO-based gas sensors [104] Moreover, they may easily lose their original sensing performance over time due

to their high tendency to be oxidized even at room temperature [105] along with their twisting tendency that decreases the number of active sites [106]

1.3 SMO/TMD heterostructure gas sensors

Since the unique characteristics of TMD nanomaterials were discovered and appeared promising to researchers, several studies have focused on the combination of TMDs with traditional SMO nanomaterials in order to look for potentially new interesting synergistic effects from their heterostructures Tekalgne et al reported an enhanced photocatalytic activity of SnO2@WS2

heterostructure, which is promising for solar water splitting application [107] Other publications also reported SnO2/WS2 hybrid humidity sensors with enhanced performances [108,109]

The influence of SMO/TMD heterojunctions in enhancing gas sensing performances are also very attractive In a study by Han et al., WS2 nanoflakes fabricated by LPE were decorated with WO3 by chemical oxidation (Fig 1.6(a-b)) [110] Compared to the low response, long response time and virtually no recovery of WS2 nanoflake gas sensor towards 5 ppm NO2 at 25 °C (Fig 1.6(c)), the WS2-WO3 nanohybrid gas sensor showed significantly enhanced gas response, response time and recoverability (Fig 1.6(d)), reaching 16.7 towards 5 ppm NO2

(Fig 1.6(e)) The enhance in gas sensing performance was attributed to the extra adsorption sites as well as the potential barrier (built-in electrical field) at the interface between WS2 and WO3, as illustrated in Fig 1.6(f)

14

Fig 1.6 (a-b) TEM and HRTEM of WS2-WO3 hybrid Gas responsive curve

to 5 ppm at 25 °C of (c) WS2 and (d) WS2-WO3 gas sensor (e) Responsive curve to different NO2 concentrations of WS2-WO3 gas sensor at 25 °C, (f) Energy band structure of WS2-WO3 heterojunction [110]

Several studies have also been reported regarding the decrease in working temperature of SMO/TMD heterojunction gas sensors Jha et al fabricated a

WS2/WO3 hybrid gas sensor by a grinding and sonication process The heterojunction between WS2 and WO3 was shown to have a major role in decreasing working temperature compared to that of WO3 (from 350 °C to 250 °C)

as well as enhanced the sensor’s gas response and selectivity towards NH3 [111] The integration of MoS2 and SnO2 has also been reported in several papers, with improved gas response, selectivity, and lowered operating temperature [14,112] Another study reported on the implementation of a SnO2 nanofiber(NF)-based gas sensor with MoS2 NSs also showed an improved optimal operating temperature of

150 °C compared to that of pristine SnO2 NF counterpart (300 °C), along with improved gas response towards SO2 [113] It is clearly that the operating

of a self-heated WS2-SnO2 core-shell structure [114] The gas response to 10 ppm

CO of the bulk WS2 gas sensor was highest (1.23) at 120 °C, but increased to 8 after the coating of ALD-delivered SnO2 shells The self-heating effect also lowered the sensor’s working temperature (equivalent to an operating temperature

of about 52 °C) In another study by the same research group, they decorated a bare WS2 gas sensors with SnO2 and Au nanoparticles by a simple drop coating step The SnO2-WS2 gas sensor showed an enhanced gas response of 1.952 towards

50 ppm CO at room temperature compared to only about 1.07 of the bare WS2 gas sensor [115] Nevertheless, the influence of SnO2/WS2 heterojunctions is still quite unclear due to the appearance of Au as a major catalyst, and the comparison in working temperature was also not reported

So far, there have been no specific studies investigating the gas sensing characteristics of heterostructures between WS2 and SnO2 NWs In the following chapters, a comprehensive study of gas sensing performance of each material, namely SnO2 NWs and WS2 NSs, as well as of their heterostructure will be presented in detail The synthesis methods of each sensing material are facile and scalable, and the investigation of their heterostructure gas sensor will provide a significant contribution to the development of SMO/TMD hybrids as a new type

of high-performance and energy-efficient gas sensor

16

CHAPTER 2 EXPERIMENTAL DETAILS

In this chapter, the fabrication process of on-chip SnO2 NWs, WS2 NSs, and SnO2 NW/ WS2 NS gas sensors will be presented in detail In addition, the characterization methods and the set-up of gas sensing measurement system will also be provided

2.1 Synthesis of on-chip SnO2 NWs by evaporation method

2.1.1 Fabrication of interdigitated Pt chips

The interdigitated Pt electrodes were prepared on a thermally oxidized silicon wafer using a DC sputtering system Each pair of Pt electrodes consists of 37 digits

of 780 µm in length and 20 µm in width The gap between two adjacent digits is

100 °C to evaporate the solvents, ready for further use

2.1.2 Fabrication of on-chip SnO2 NWs by the evaporation method The fabrication of SnO2 NWs was carried out using a horizontal quartz-tube furnace (Lindberg/Blue M, model TF55030A, USA) at ITIMS, HUST through a thermal evaporation process The growth of SnO2 NWs occurred directly on pretreated Pt chips using a pure Sn powder source (Alfa Aesar, 99.8%) and a flow

of ultrapure O2 (99.99 %) The trace content of O2 in the growth process was controlled by a Mass Flow Controller (MFC, Aalborg, model GFC17S-VALD2-A0200, USA) The schematic diagram of the on-chip fabrication of SnO2 NWs is depicted in (Fig 2.2)

The synthesis of SnO2 NWs was described in detail elsewhere Firstly, 0.1 g

Sn powder was put into an aluminum boat, with the interdigitated Pt electrodes being placed face-up on top The boat was loaded into the center of a horizontal quartz-tube furnace The tube’s ends were sealed with silicon O rings and the tube was subsequently evacuated It was thoroughly purged for 2 hours with alternate

17

on-off switching of 5-minute flow of 300 sccm Ar in order to remove all the remaining dust particles and gas molecules Then, the furnace temperature was increased from room temperature to 730 °C in 20 minutes It should be noted that during the whole process, the Ar gas flow should not be introduced in order to ensure a sufficient level of vacuum that is required for the large-scale growth of SnO2 NWs When the temperature was about to reach the targeted value, a 0.5 sccm flow of O2 was initiated, and the growth process was maintained at 730 °C for 5 minutes, after which the whole system was allowed to cool down naturally

Fig 2.2 Schematic diagram of the on-chip fabrication of SnO2 NWs 2.2 Synthesis of TMD nanomaterials

2.2.1 Fabrication of WS2 NSs

WS2 NSs were fabricated via a liquid-phase exfoliation (LPE) method using

an environment-friendly solvent Schematic illustration of the LPE process is given in Fig 2.3 Specifically, 90 mg bulk WS2 powder (NAYATE, 99.99%) was mixed into a 30 ml solvent of 2:1 ethanol/DI water The mixture was ultrasonicated

at 420 W for 8 hours in total using an ultrasonic probe sonicator During the exfoliation process, a continuous iced water-cooling system and a 5 s ON/OFF regime were applied in order to prevent the system from overheating After the process was complete, the sample was left undisturbed overnight before being centrifuged at 2000 rpm for 15 minutes Next, two-thirds of the supernatant were carefully collected using a micropipette, and the sediment at the bottom of the centrifuge tubes, which were unexfoliated bulk powder, was discarded Another centrifugation at 6000 rpm was applied to the obtained supernatant for 30 minutes, and after carefully removing the solvent, the WS2 NSs were eventually obtained, ready for further use

18

Fig 2.3 Schematic illustration of the LPE of WS2 NSs

2.3 Synthesis of SnO2 nanowire/ TMD nanomaterial heterojunction gas sensors

2.3.1 Fabrication of WS2@SnO2 gas sensor

The preparation of WS2@SnO2 gas sensor was conducted using a facile casting method, as illustrated in Fig 2.4 In prior to the drop-casting, the same amount of WS2 sediments collected from the exfoliation process were dispersed in ethanol at different volumes, namely 10 µl, 5 µl and 1 µl, and three SnO2 NW gas sensors were prepared Note that the three gas sensors was examined to have relatively the same base resistance as well as a similar gas sensing performance compared to the reference sample Using a micropipette, a drop of about 0.4 µl in volume was dropped onto the three samples for all three concentrations of WS2

drop-NSs The samples were denoted SW1, SW2, and SW3 from the most diluted to the least diluted WS2 solution The samples were baked at 120 °C shortly before being heat treated in vacuum at 300 °C (ramping speed 2 °C/min) in 3 hours to stabilize the contact between WS2 NSs and SnO2 NWs During the heat treatment, a flow

of 50 sccm Ar was introduced to blow out any unwanted vapors

Fig 2.4 Fabrication of WS2@SnO2 gas sensors by drop casting

2.4 Characterization methods

19

The crystal structure of the samples were investigated by X-Ray Diffraction (XRD, D8 Advance, Bruker, Germany) using the Cu-Kα radiation Raman spectroscopy was conducted using an InVia confocal microscope UV-Vis absorbance spectra were recorded using a V-650 UV/VIS spectrophotometer (Jasco, USA) The morphology of the sample was examined by Field-Effect Scanning Electron Microscopy (FESEM, JEOL JSM7600F), and the elemental composition was evaluated by Energy-Dispersive X-ray Spectroscopy (EDX) attached to the FESEM system The microstructure of the samples was investigated using a High-Resolution Transmission Electron Microscope (HRTEM, Tecnai G2

20 S-TWIN/FEI)

2.4.1 X-Ray Diffraction

XRD is the most efficient and widely used method for characterizing the crystal structure of materials It is based on the phenomenon of wave interference, specifically of X-rays When a single-wavelength X-ray beam interacts with the sample under a continuously changing angle, the X-ray beam will be diffracted by the crystal planes of the crystalline solid, and a spectrum of diffraction intensity will be recorded The diffracted waves are called in-phase when the path difference between them equals an integer number of the X-ray wavelength, i.e when the following relationship is satisfied:

nλ = 2dsinθ (1)

where λ is the wavelength of the X-ray source, and d is the spacing between parallel atomic planes Equation 1 is called the law of diffraction, also known as the Bragg’s Law The constructive interference as a result of in-phase waves will create a peak in the diffraction spectrum, and a collection of multiple diffraction peaks will be compared with reference spectra given by the Joint Committee on Powder Diffraction Standards (JCPDS) in order to determine the crystal structure

as well as the phase of the sample

2.4.2 Raman spectroscopy

Raman spectroscopy is a very powerful nondestructive method for characterizing the structural and electrical properties of nanomaterials It concerns about the Raman scattering, i.e the inelastic scattering of a high-intensity laser beam out of the sample The amount of inelastic scattering is very small compare

to that of elastic scatting, or Rayleigh scatting, in which the wavelength of the scattered light is identical to that of the incident light, thus bringing no information about the structure of materials Instead, the Raman scattering phenomenon includes scattered lights with wavelengths different from those of the incident light

as a result of the interaction with the chemical bonds inside the material Thus, Raman spectroscopy can provide useful information about the vibrational and rotational properties of materials Hence, Raman spectroscopy has been an essential tool for investigating the structural characteristics of 2D materials, such

as graphene or TMDs [116]

20

2.4.3 Ultraviolet-Visible Spectroscopy (UV-Vis)

The Ultraviolet-Visible spectroscopy uses a range of continuous electromagnetic wavelengths in ultraviolet–visible regions to characterize absorption behaviors of matters, usually carried out in a solution The UV-Vis spectrum gives various valuable measurements, such as absorbance, transmittance,

or reflectance Moreover, the wavelengths of the absorption peaks can refer to different types of bonds in the molecule And the absorbance obtained from UV-Vis spectroscopy can be used to examine the relative absorption between different samples using the same setup (i.e the same incident light’s intensity – Io, and the same path length – L Ultimately, the parameters can be used to calculate the concentration (c) of the solution using the Beer-Lambert law:

c = A

ϵL (2) where ε is the molar absorptivity, a constant belonging to a particular species

2.4.4 Scanning Electron Microscope and Energy-Dispersive X-ray Spectroscopy

Scanning Electron Microscope (SEM) is a widely used technique to examine the morphology and topography of nanomaterials In SEM instruments, an electron beam is focused and accelerated by multiple electromagnetic coils and scanned over the surface of the sample The reflected secondary electrons will be collected

to give an image of the sample’s morphology Other types of signals are also collected and analyzed, such as backscattered electrons that give a contrast image

of different elemental compositions in the sample The relative elemental compositions are calculated by analyzing the intensity of the characteristic X-ray, which is emitted after an excited electron migrates to an inner shell This technique

is called Energy-Dispersive X-ray Spectroscopy (EDX or EDXS), and EDX is often integrated into SEM instruments

2.4.5 Transmission Electron Microscopy and Selected-Area Electron Diffraction

Transmission Electron Microscopy (TEM) uses a high energy electron beam

to transmit through an ultrathin specimen and give many interesting information

at the atomic scale of materials Whereas SEM uses knocked-off electrons to visualize the morphology and surface composition of specimen, TEM uses transmitted electrons to give information about the inner structure of specimen The high-resolution TEM (HRTEM) is a powerful imaging mode that can show atomic structure of samples

Selected-Area Electron Diffraction (SAED) is a powerful crystallography technique that is typically integrated with TEM systems Based on SAED patterns, one can identify crystal structures, crystal orientation, and other crystal characteristics of specimen SAED patterns can appear in the form of patterned bright spots (single crystal), diffraction rings (polycrystal), or diffused rings (amorphous)

21

2.5 Gas sensing measurement

The investigation of gas sensing properties was carried out using a built gas sensing system at the International Training Institute for Materials Science (ITIMS), HUST in dynamic mode The setup of the gas sensing measurement system was shown in Fig 2.5(a) It consists of a Mass Flow Controller (MFC) system, which is used to precisely mix and control the concentration of analytical gases, a Keithley system model 2450 for electrical measurements, a test chamber, and a Labview control software The interior of the test chamber was shown in Fig 2.5(b)

In dynamic mode, gases are continuously flown through the test chamber, while the resistance of the sensor is continuously recorded When the analytical gas is injected into the chamber, it is mixed with dry air with a controlled concentration so that the flow is quantitively unchanged (i.e same stable gaseous pressure) Accordingly, the concentrations of analytical gases are calculated by the formula:

22

this thesis, the flow rate is maintained at 400 sccm, and the analytical gas is mixed with dry air using a Mass Flow Controller (MFC) system Sensors are placed on a hot plate capable of reaching 450 °C, and their electrical properties are continuously measured using a Keithley system no 2450

in this thesis for reference

3.1.1 Structure and morphology

The crystal structure of SnO2 NWs was investigated by X-ray diffraction In the XRD spectrum of the bare SnO2 NWs sample (Fig 3.1), there appeared several diffraction peaks at 2θ = 26.62 °, 33.94 °, 51.78 °, and 54.85 ° that correspond to the (110), (101), (211), and (220) planes of rutile SnO2 phase (JCPDS card #88-0287), respectively In addition, two intense peaks at 39.79 ° and 44.79 ° that belong to the diffraction peaks of Pt, according to the JCPDS card #87-0636, and the diffraction peak of the (400) plane of the Si substrate (JCPDS #75-0589) were also present This result confirmed the successful on-chip growth of SnO2 on Pt electrodes

Fig 3.1 XRD spectrum of SnO2 NWs grown on interdigitated Pt

electrode

To examine the morphology and density of as-grown SnO2 NWs, SEM images of the SnO2 NWs at several magnifications were taken Fig 3.2(a) shows the gap between two adjacent electrode digits, with the pair of interdigitated electrodes given inset As can be seen, the SnO2 NWs covered the entire surface

of the electrodes, and numerous SnO2 NWs grew horizontally in the gap, creating