Development of Transition Metal Compound Electrodes … · II Development of Transition Metal Compound Electrodes for Supercapacitors Haoran Wu Master of Applied Science Materials

Development of Transition Metal Compound Electrodes for

Supercapacitors

by

Haoran Wu

A thesis submitted in conformity with the requirements for the degree of Master of Applied Science

Materials Science and Engineering University of Toronto

© Copyright by Haoran Wu 2014

II

Development of Transition Metal Compound Electrodes for

Supercapacitors

Haoran Wu

Master of Applied Science

Materials Science and Engineering

University of Toronto

2014

Abstract:

Many transition metal compounds are found to possess promising pseudocapacitive

properties. Research on these transition metal compounds, especially Mo, V and W

based materials, is still limited. This project aimed to investigate cost-effective methods

utilizing inexpensive transition metal compounds to produce EC electrodes with high

energy density and power capability. Mo(O,N)x was developed as a pseudocapacitive

electrode material by electrodeposition of Mo oxide followed by low-temperature thermal

nitridation in a N2 environment. The partial nitridation significantly improved the

electrochemical properties, including capacitive behavior, power capability and cycle life.

The device performance can be improved with an asymmetric configuration using an

electrodeposited WO3 electrode. An electrolessly deposited V mixed oxides electrode

also showed ideal capacitive behaviors and good cycle life. By tuning the fabrication

methods, the electrodes (i.e. Mo(O,N)x, WO3 and VOx.yH2O) were developed with low

cost processes.

III

Acknowledgements:

Foremost, I would give my sincere thanks to my supervisor, Professor Keryn Lian, for

her kind support, guidance, encouragement and patience throughout this project. I have

not only learned professional skills and knowledge but also grown up, gained

confidence and refined my personality.

I would like to thank the members in the Flexible Energy and Electronic Laboratory

(FEEL) for their generous support, especially the helpful discussions and brilliant

advises from Han Gao, Sanaz Ketabi and Matthew Genovese.

I am grateful to Sal Boccia and Dan Grozea for general support in the lab supply; Neil

Coombs and Ilya Gourevich for the help in performing ESEM; George Kretschmann for

helping me conduct XRD and Rana Sodhi for XPS analyses.

I very much appreciate the Ontario Graduate Scholarship (OGS), the NSERC Graduate

Scholarship (CGS) and the Hatch Graduate Scholarship for Sustainable Energy

Research for the financial support.

Last but not least, I owe my deep gratitude to my parents Minyan Zhou and Jianming

Wu for their patience, support and guidance. Meanwhile, I would like to thank my

girlfriend Qi Zhang for her understanding, encouragement and love.

IV

Table of Contents:

Abstract: ....................................................................................................................... II

Acknowledgements: ....................................................................................................III

Table of Contents: ...................................................................................................... IV

List of Tables: ............................................................................................................. VI

List of Figures: ........................................................................................................... VII

List of Appendices: ...................................................................................................... X

Abbreviations: ............................................................................................................. XI

Chapter 1: Introduction to Electrochemical Capacitors ................................................ 1

Chapter 2: Literature Review ....................................................................................... 6

2.1 Electric Double Layer Capacitors (EDLC) .......................................................... 6

2.2 Pseudocapacitors .............................................................................................. 7

2.3 Transition Metal Compounds as Electrode Materials ......................................... 9

Chapter 3: Objectives .................................................................................................26

3.1 Objectives for Mo Oxide-nitride .........................................................................27

3.2 Objectives for W Oxide .....................................................................................27

3.3 Objectives for V Oxide ......................................................................................27

Chapter 4: Experimental .............................................................................................29

4.1 Fabrication Method of Mo Oxide-nitride Electrode ............................................30

4.2 Fabrication Method of W Oxide Electrode ........................................................31

4.3 Fabrication Method of V Oxide Electrode .........................................................31

4.4 Characterizations of the Electrodes ..................................................................31

Chapter 5: Results and Discussion ............................................................................36

5.1 Molybdenum Oxide-nitride ................................................................................36

5.2 Tungsten Oxide ................................................................................................51

V

5.3 Vanadium oxide ................................................................................................58

Chapter 6: Conclusions ..............................................................................................68

Chapter 7: Future Work ..............................................................................................70

References .................................................................................................................71

Appendices: ................................................................................................................76

VI

List of Tables:

Table 1-1: Comparison of the properties of electrostatic capacitor, ECs and batteries,

adapted from [3] .............................................................................................................. 2

Table 2-1: Comparison between EDLCs and pseudocapacitors, adapted from [3] ......... 8

Table 2-2: Products from the reactions of molybdenum oxides with ammonia, adapted

from [56] .........................................................................................................................18

Table 4-1: Summary of electrode fabrication methods ...................................................29

Table 4-2: List of chemicals used in the fabrication process ..........................................29

Table 5-1: Capacitance of the nitrided Mo oxide electrodes through different

electrodeposition conditions ...........................................................................................37

Table 5-2: Compositions of electrode surfaces treated in different gases ......................40

Table 6-1: Summary of Mo(O,N)x, WO3 and VOx.yH2O electrodes ................................69

Table 6-2: Summary of ECs based on the fabricated electrodes ...................................69

VII

List of Figures:

Fig. 1-1 Ragone chart of specific power vs. specific energy for energy storage devices,

dash line represents the time constant of the devices, adapted from [1]. ....................... 2

Fig. 1-2: Schematic configurations of a typical EC and cross-section of the main

components, adapted online: www.ustudy.in/node/2303. ............................................... 3

Fig. 1-3: Categories of ECs by charge storage mechanisms and corresponding

electrode materials. ......................................................................................................... 4

Fig. 2-1: Schematic diagram of EDLC and its charge storage mechanism, adapted from

[6] .................................................................................................................................... 7

Fig. 2-2: Schematic of charging mechanism in a pseudocapacitor ................................. 8

Fig. 2-3: Cyclic voltammograms of sol-gel derived RuO2.xH2O at a scan rate of 2mV/s in

0.5 M H2SO4, annealed in various temperatures, adapted from [22]. ............................10

Fig. 2-4: Cyclic voltammograms of electrodeposited MnO2 at scan rates of 10mV/s,

50mV/s, 100mV/s and 200mV/s in 0.5M Na2SO4, adapted from [28] ............................11

Fig. 2-5: E-pH diagram of a Mo-H2O system, adapted from [36] ...................................13

Fig. 2-6: Cyclic voltammograms of potentiodynamically deposited MoxO at various scan

rates, adapted from [39] .................................................................................................14

Fig. 2-7: Passive and transpassive region for Mo oxides from pH 1.5 to pH 8, adapted

from [40]. ........................................................................................................................15

Fig. 2-8: Cyclic voltammograms of CVD deposited MoxN (γ-Mo2N and δ-MoN) at 10

mV/s in 0.5 M H2SO4, dot line (fresh), solid line (after 2 weeks) and dash line (after 3

weeks), adapted from [21]. ............................................................................................17

Fig. 2-9: E-pH diagram of a W-H2O system, adapted from [63] .....................................19

Fig. 2-10: Cyclic voltammograms of template method prepared WO3 at 5 mV/s in 2 M

H2SO4, adapted from [60], b- stands for non-porous bulk and m- for meso-porous. ......20

Fig. 2-11: Cyclic voltammograms of W2N at 50 mV/s in 1 M KOH, adapted from [75]. ..21

Fig. 2-12: E-pH diagram of a V-H2O system, adapted from [86] ...................................22

Fig. 2-13: Cyclic voltammograms of VOx at (1) 25 (2) 100 (3) 300 (4) 500 mV/s in 3 M

LiCl, adapted from [82] ...................................................................................................24

VIII

Fig. 2-14: Cyclic voltammograms of VN in 0.1M of various electrolytes, adapted from [50]

.......................................................................................................................................25

Fig. 4-1: Example temperature profile when heat up to 400oC .......................................30

Fig. 4-2: Schematic diagrams of the testing set-up (a) for single electrode (b) beaker cell

EC (c) filter paper EC .....................................................................................................32

Fig. 4-3: Cyclic voltammograms of reactions (a) diffusion control with slow kinetics (b)

thin-layer condition with fast kinetics, adapted from [99]. ...............................................33

Fig. 4-4: Example CV profile of an ideal EDLC. .............................................................34

Fig. 4-5: Example cyclic voltammograms of a Ru dioxide electrode, adapted from [19].

.......................................................................................................................................34

Fig. 4-6: Example charging/discharging profile for an ideal EC. ....................................35

Fig. 5-1-1: Cyclic voltammograms in 0.5 M H2SO4 at 100 mV/s: (a) Mo oxide vs. Ti

substrate between -0.55 V and 0 V (b) Mo oxide, Mo oxide heat treated in N2, ammonia

and forming gases between -0.15 V and 0.45 V. ...........................................................39

Fig. 5-1-2: High resolution XPS spectra for Mo 3d: (a) Mo oxide, (b) Mo oxide heat

treated in N2, (c) forming gas, (d) ammonia. ..................................................................41

Fig. 5-1-3: Cyclic voltammograms in 0.5 M H2SO4 at 100 mV/s of Mo oxide-nitride heat

treated at various temperatures in N2. ...........................................................................42

Fig. 5-1-4: ESEM micrographs of Mo electrodes: (a) as-deposit Mo oxide, (b) nitrided

Mo oxide, (c) side-view of as-deposit Mo oxide, (d) side-view of nitrided Mo oxide. ......43

Fig. 5-1-5: XRD patterns of (a) as-deposit Mo oxide (b) Mo oxide treated in air and (c)

Mo oxide-nitride treated in N2, with identical temperature profiles at a peak temperature

of 400 oC. .......................................................................................................................44

Fig. 5-1-6: Cyclic voltammograms in 0.5 M H2SO4: (a) Mo oxide at 1st, 2500th, and

5000th cycle at 100 mV/s; (b) Mo oxide-nitride at 1st, 2500th, and 5000th cycle at 100

mV/s. ..............................................................................................................................45

Fig. 5-1-7: galvanic charge-discharge curves of Mo oxide-nitride electrode at 1 mA/cm2

for 5000 cycle in 0.5 M H2SO4. ......................................................................................46

Fig. 5-1-8: cyclic voltammograms of a complete beaker EC in 0.5 M H2SO4 at various

scan rates made from (a) two identical Mo oxide-nitride electrodes (b) two identical

carbon electrodes (b) carbon as positive electrode and Mo oxide-nitride as negative

electrode. .......................................................................................................................48

IX

Fig. 5-1-9: cyclic voltammograms of (a) Ti substrate and Mo oxide-nitride in 0.5 M

Na2SO4 at 100 mV/s; Mo oxide-nitride at 1st, 5000th, and 10000th cycle at 100 mV/s. ...49

Fig. 5-2-1: ESEM micrographs of W oxide electrodes (a) low magnification, (b) high

magnification ..................................................................................................................51

Fig. 5-2-2: High resolution XPS spectra for W 4f on the surface of W oxide electrode. .52

Fig. 5-2-3: Cyclic voltammograms of as-deposit WO3 vs. Ti substrate in 0.5 M H2SO4 at

100 mV/s. .......................................................................................................................53

Fig. 5-2-4: Cyclic voltammograms of as-deposit WO3 at 1st, 5000th and 15000th cycles in

0.5 M H2SO4 at 100 mV/s. .............................................................................................54

Fig. 5-2-5: Cyclic voltammograms of as-deposit WO3 in 0.5 M H2SO4 at 100 mV/s and 1

V/s. .................................................................................................................................55

Fig. 5-2-6: Cyclic voltammograms of beaker ECs with various configurations in 0.5 M

H2SO4 at 100 mV/s. .......................................................................................................56

Fig. 5-2-7: Cyclic voltammograms at 100mV/s of as-deposit WO3 in (a) 0.5 M Na2SO4

and 0.5 M H2SO4 (b) 0.5 M Na2SO4 for 5000 cycles. .....................................................57

Fig. 5-3-1: Homogeneities of the electroless deposited vanadium electrodes in various

deposition concentration and time. ................................................................................59

Fig. 5-3-2: (a) Capacitance vs. deposition time in 0.25 M, 0.5 M, and 0.75 M VOSO4

baths for vanadium oxide electrodes, measured at 5 mV/s; (b) CVs of 0.25M-5days,

0.5M-5days, 0.75M-5days, and Ti electrodes in 1 M LiCl at 5 mV/s. .............................60

Fig. 5-3-3: Surface morphologies of the electrodes in Fig. 5-3-2(b); (a)-(c) top view and

(d)-(f) side view of 0.25M-5days, 0.5M-5days, and 0.75M-5days electrodes. ................61

Fig. 5-3-4: XRD patterns and cyclic voltammograms for 0.5M-5days vanadium

electrodes (a) as-deposit (b) heat treated in argon at 120oC (c) heat treated in argon at

350oC (d) heat treated in air at 350oC. ...........................................................................63

Fig. 5-3-5: High resolution XPS spectra for V2p and O1s of 0.5M-5days electrode.......64

Fig. 5-3-6: Cyclic voltammograms of as-deposit V oxides 1 M LiCl at 10 mV/s. ............65

Fig. 5-3-7: Cyclic voltammograms of 0.5M-5days vanadium oxide electrode in 1M LiCl

(a) at 1st, 1000th, 2000th and 3000th cycle 5 mV/s (b) at 10 mV/s and 100 mV/s. ...........66

Fig. 5-3-8: Cyclic voltammogram of an EC cell made from two identical 0.5M-5days

vanadium oxide electrodes in 1 M LiCl at 5 mV/s. .........................................................67

X

List of Appendices:

Appendix A-1: List s of existing electrode materials for ECs

Appendix B-1: Effects of Electrodeposition Parameters for Mo oxide

Appendix B-2: XPS Analysis on Mo(O,N)x Electrode on Mo3p Orbitals and N1s Orbitals

Appendix B-3: SIMS Analysis on Mo(O,N)x Electrodes

Appendix B-4: Reproducibility of Mo(O,N)x Electrodes

Appendix C-1: Effects of Hydrogen Peroxide in Electrodeposition of WO3

Appendix C-2: XRD of WO3 Electrodes

Appendix C-3: Reproducibility of WO3 Electrodes

Appendix D-1: Reproducibility of V Mixed Oxides Electrodes

XI

Abbreviations:

ALD: Atomic Layer Deposition

CE: Counter Electrode

CNT: Carbon Nanotubes

CV: Cyclic Voltammogram

CVD: Chemical Vapor Deposition

ECs: Electrochemical Capacitors

EDLC: Electric Double Layer Capacitors

EDS: Energy Dispersive X-ray Spectroscopy

ESEM: Environmental Scanning Electron Microscopy

ESR: Equivalent Series Resistance

OLC: Onion Like Carbon

PANI: Polyaniline

PPy: Polypyrrole

PVA: Poly Vinyl Alcohol

P3MT: poly(3-methylthiophene)

RE: Reference Electrode

SIMS: Secondary Ion Mass Spectroscopy

WE: Working Electrode

XPS: X-ray photoelectron Spectroscopy

XRD: X-ray Diffraction

1

Chapter 1: Introduction to Electrochemical Capacitors

The pressing demand of sustainable energy is one of the most vital challenges imposed

on the current generation. While many technologies have been developed to harvest

sustainable energy, storing the energy and delivering it whenever and wherever needed

is another more important concern. This is because some sustainable energy sources

are highly site-specific, such as hydro- and geothermal energy; the others such as wind

and solar energy are intermittent and do not coincide with the demand. Therefore, the

development of off-grid electrical energy storage devices such as batteries and

supercapacitors are very important for a society to rely on sustainable energy.

Energy storage devices can be categorized into electrostatic capacitors, electrochemical

capacitors (ECs) and batteries. The main differences of these devices can be

summarized in a Ragone chart shown in Fig. 1-1[1] in terms of specific power vs.

specific energy. Electrostatic capacitors have the highest specific power that delivers a

large amount of energy in a short period of time, but it is extremely limited by the

specific energy (less than 0.1 Wh/kg). Batteries such as lithium-ion batteries have large

specific energy. However, they suffer from a slow power delivery or uptake (less than

100 W/kg). Compared to electrostatic capacitors and batteries, ECs have balanced

specific power and energy. An EC can be fully charged within seconds and deliver

substantial amount of energy, which bridges the gap between electrostatic capacitors

and batteries. In addition, ECs enable long cycle life and shelf life [2], making them

complementary to batteries.

Energy storage devices have been found in many applications. Table 1-1[3] lists a

detailed comparison of the properties among electrostatic capacitors, ECs and batteries.

ECs are used in various applications, such as hybrid electric vehicles, electronic devices,

uninterrupted power systems, power regulations [4-6]. A recent example is the

applications of ECs in the emergency doors on an Airbus A380 airplane[1]. However,

the applications of ECs are still constrained by relatively high cost and low specific

energy [1, 7, 8], which are the current direction of improvement.

2

Fig. 1-1 Ragone chart of specific power vs. specific energy for energy storage devices,

dash line represents the time constant of the devices, adapted from [1].

Table 1-1: Comparison of the properties of electrostatic capacitor, ECs and batteries,

adapted from [3]

Electrostatic

Capacitors ECs Batteries

Discharge time 10-3 – 10-6 s 0.3-30 s 0.3-3 h

Charge time 10-3 – 10-6 s 0.3-30 s 1-5 h

Specific energy

(Wh/kg) <0.1 1-10 10-100

Specific power

(W/kg) >10000 1000-10000 50-200

Efficiency 1 0.85-0.98 0.7-0.85

Cycle life >500000 >100000 500-2000

3

Similar to batteries, the configuration of an EC (shown in Fig. 1-2) includes two

electrodes, known as anode and cathode that allow physical or chemical energy storage

on the surface; a layer of electrolyte that permits ion exchanges; current collectors to

conduct electricity for external circuit, and a layer of separator (optional) to prevent

electric short. Among these components, electrodes and electrolytes are the most

important parts that enable the functions of an EC. The development of the electrodes

attracts more attention because it directly determines the electrochemical behaviors

such as the capacitance, conductivity and cost of an EC. Therefore, this work mainly

focuses on the development of the electrode materials.

Fig. 1-2: Schematic configurations of a typical EC and cross-section of the main

components, adapted online: www.ustudy.in/node/2303.

The capacity of an EC is conventionally denoted as capacitance (C). The capacitance of

the entire cell Ccell can be expressed by Canode and Ccathode in equation (1)[9]:

=

+

(Equ-1)

If Canode = Ccathode in a symmetric cell, the Ccell equals to ½ Canode or ½ Ccathode. If Canode ≠

Ccathode in an asymmetric cell, the Ccell is dominated by the electrode with smaller

capacitance. The maximum energy and power can be delivered by an EC are

expressed in equation 2 and 3 [10, 11]:

4

(Equ-2)

Pmax

(Equ-3)

Where E is the energy, V is the maximum voltage of the cell and R is the equivalent

series resistance (ESR). Equation (2) is derived by integrating the voltage change in the

discharge process while equation (3) is a conventional notation. It is suggested that the

energy and power are directly related to the voltage window of the cell (V), the charges

that can be stored on the electrode surface (C) and the ESR (R) of the electrode and

electrolyte.

ECs can be categorized into electric double layer capacitors (EDLC) and

pseudocapacitors by the charge storage mechanisms as illustrated in Fig. 1-3. EDLC

employs double layer that physically separates the charges while the pseudocapacitors

enable surface redox reactions that store the energy electrochemically. The charge

storage mechanisms are directly governed by the electrode materials. Carbon based

electrodes such as activated carbon, carbon aerogels, carbon nanotubes (CNT) and

grapheme are electrode materials for EDLC due to large surface area for double layer

formation, while metal compounds and conductive polymers are pseudocapacitive

materials attributing to their reactive surface. Detailed list of existing electrode materials

was summarized in Appendix A-1.

Fig. 1-3: Categories of ECs by charge storage mechanisms and corresponding

electrode materials.

5

In general, an EC electrode should have the following performance criteria: high

capacitance, fast kinetics, high conductivity, long cycle life, large voltage window and

low cost.

The outline of the thesis is given as follows:

Chapter 1: This chapter gives a brief introduction to electrochemical capacitors,

including their structures, materials and performance criteria.

Chapter 2: This chapter offers a comparison between of electric double layer capacitors

and pseudocapacitors. Particularly, transition metal compounds as pseudocapacitors

electrodes are listed; as well as their advantages, disadvantages and direction of

improvement.

Chapter 3: This chapter states the objectives of this research, including the detailed

approaches for each transition metal compound investigated.

Chapter 4: Experimental details are given in this chapter, including fabrication methods

for the three investigated transition metal compounds. The characterizations methods

are also specified in this chapter.

Chapter 5: This chapter discusses the results obtained. Material characterizations give

the full understanding of the chemistry, structure and morphology of the synthesized

materials; while electrochemical characterizations evaluate their performance in EC

applications.

Chapter 6: Conclusions based on chapter 5 are given in this chapter, followed by

recommendation of the additional work.

6

Chapter 2: Literature Review

2.1 Electric Double Layer Capacitors (EDLC)

The EDLC uses a physical surface charge accumulation to store electric energy in a

Helmholtz layer (i.e. double layer) as shown in Fig. 2-1[6]. Carbon materials with large

surface area, such as activated carbon, are usually used as electrode material, allowing

large amount of charge accumulation. This is because the capacitance is proportional to

the surface area of an electrode according to equation (4):

C=

(Equ-4)

Where C is the capacitance, A is the electrode surface area, ε is the permittivity and d is

the separation distance between the two electrodes. During discharging process,

electrons travel from the negative electrode to the positive electrode through an external

circuit, while the cations move towards positive electrode and anions move the opposite

direction in the electrolyte. During charging, the reversed processes occur. The cations

and anions accumulate on respective electrode surface without any charge exchange

between electrode/electrolyte interfaces. The charging()/discharging() can be

expressed by equations (5)[12, 13]:

Es1 + Es2 + A- +C+ ↔ Es1+//A- + Es2

-//C+ (Equ-5)

Where Es1 and Es2 represent two electrode surfaces, A- and C+ are anions and cations,

respectively. When an EC is fully charged, the potential distributes symmetrically across

two electrode/electrolyte interfaces within 2-10 Angstroms as illustrated in Fig. 2-1.

EDLC using inexpensive material, such as activated carbon, is the current technology

and has already been commercialized. However, the theoretical threshold of double

layer capacitance is limited at 20-50 uF/cm2. And the activated carbon based electrodes

are also limited by their low conductivity. Although researchers have developed next

generation EDLC electrodes based on more advanced carbon materials such as carbon

nanotubes (CNT), graphene and onion like carbon (OLC) that enable higher conductivity

7

[1, 9, 14], problems still exist. For example, same advanced carbon materials have

smaller surface area and higher cost compared to that of activated carbon. Therefore,

pseudocapacitors attract significant attentions due to their large capacitance and high

conductivity.

Fig. 2-1: Schematic diagram of EDLC and its charge storage mechanism, adapted from

[6]

2.2 Pseudocapacitors

Pseudocapacitors are analogue to high power batteries since they enable fast and

reversible electrochemical redox reactions on the electrode surface, which allow charge

transfer across the electrode electrolyte interfaces as depicted in Fig. 2-2. Due to the

redox reaction on the electrode surface or sometimes within near-surface bulk, the

pseudocapacitors can deliver 10-100 times higher capacitance than the EDLC[15].

Though redox reactions occur on the electrode, the electrochemical behaviors of

pseudocapacitors are similar to that of EDLCs by showing a rectangular cyclic

voltammograms (CV) (see section 4.4.1) and a linear voltage increase/decay in a

constant current charging/discharging process (see section 4.4.2)[3].

8

Fig. 2-2: Schematic of charging mechanism in a pseudocapacitor

Table 2-1 lists the comparisons between EDLC and pseudocapacitors [3]. It is clear that

pseudocapacitors have significant advantage in capacitance compared with EDLCs.

However, due to the kinetics of the redox reactions, the specific power of

pseudocapacitors may be limited. Therefore, the current trend is to develop

pseudocapacitors with both high capacitance and high power (fast kinetics and high

conductivity).

Table 2-1: Comparison between EDLCs and pseudocapacitors, adapted from [3]

EDLCs Pseudocapacitors

1. Physical charge separation Faradaic charge exchange

2. 20-50 uF/cm2 200-2000 uF/cm2

3. Highly reversible

Charging/discharging

Reversible but may be limited by the

discharging rate

4. Has restricted voltage range Has restricted voltage range

6. Exhibit mirror-image cyclic

voltammogram

exhibit mirror-image cyclic

voltammogram

Various materials have demonstrated pseudocapacitive behaviors. They can be

categorized into conductive polymers and transition metal compounds. Conductive

9

polymers, such as poly(3-methylthiophene) (P3MT)[16], polypyrrole (PPy)[17] and

polyaniline (PANI) [18], were developed as pseudocapacitive electrode material and

showed substantial capacitance. Moreover, conductive polymers are relatively

inexpensive and are easy to be manufactured. The charge storage mechanism is

Faradaic that typically involves doping and de-doping processes, in which the ions in the

electrolyte intercalate in and out of polymer matrix [19]. However, this leads to cycling

problems like swelling and shrinkage, as the ions have to frequently travel across the

polymer matrix that degrades the structures. Moreover, conductive polymers are

relatively poor electron conductors [1, 3, 5].

Transition metal compounds, on the other hand, are considered the best candidates as

pseudocapacitive electrode materials. They can possess high electric conductivity, high

capacitance and good chemical stability, making them very attractive for

pseudocapacitor applications [3]. Many transition metal compounds have already been

investigated such as ruthenium dioxide (RuO2) and manganese dioxide (MnO2). Others

such as molybdenum, tungsten and vanadium based compounds are currently under

investigation and have shown some promising pseudocapacitive behaviors. The

following section reviews the development of the transition metal compounds as

electrode materials for pseudocapacitors.

2.3 Transition Metal Compounds as Electrode Materials

2.3.1 Ruthenium Dioxide (RuO2)

Among the transition metal compounds, ruthenium dioxide (RuO2), especially in its

hydrated form (RuO2 .xH2O), is known as the benchmark due to the high specific

capacitance, high conductivity, long cycle life and high charging/discharging efficiency

[20, 21]. A typical example of sol-gel derived hydrous ruthenium dioxide is shown in Fig.

2-3 [22]. The CV cycles were highly reversible that showed mirror-image profiles. The

electrode showed the highest capacitance of 720F/g at relatively low annealing

temperature of 150oC, suggesting that the ruthenium dioxide in a hydrous form offers

the best electrochemical performance [23]. This was verified by Hu et al. as they

10

reported a capacitance of RuO2.xH2O as high as 1340F/g at a higher scan rate of

25mV/s by a similar method [23].

Fig. 2-3: Cyclic voltammograms of sol-gel derived RuO2.xH2O at a scan rate of 2mV/s in

0.5 M H2SO4, annealed in various temperatures, adapted from [22].

The reaction mechanism of a RuO2.xH2O in acidic electrolyte has been well studied,

and is expressed by the following reaction (1a) reported by C. Hu [23]:

RuOa(OH)b + δH+ + δe- RuOa-δ(OH)b+δ (Rx-1a)

The rectangular CV actually contains multiple overlapping redox peaks and each peak

represents a change in the oxidation state of Ru in the oxide. At least three redox

couples occurred: Ru(VI)/Ru(IV), Ru(IV)/Ru(III) and Ru(III)/Ru(II). Take Ru(IV)/Ru(III) for

example, the reaction can be re-written as (1b):

RuO(OH)2 + H+ + e- Ru(OH)3 (Rx-1b)

11

These reactions are highly reversible, contributing to the ideal capacitive behaviors.

RuO2.xH2O is considered as the gold standard material for pseudocapacitor. However,

it cannot be commercialized due to the high cost. Therefore, researchers are

investigating other transition metal compounds as alternatives to hydrous ruthenium

dioxide.

2.3.2 Manganese Dioxide (MnO2)

Manganese dioxide (MnO2) is the most investigated transition metal compound as low-

cost alternative to RuO2.xH2O[1]. In general, MnO2 is inexpensive, environmentally

friendly, and exhibits high theoretical capacitance ranging from 1100 to 1300F/g [24-27].

Fig 2-4 shows typical CV profiles of an electrodeposited MnO2 at various scan rates [28].

The electrode showed a rectangular-shape CV at a lower scan rate of 10mV/s,

suggesting MnO2 a suitable pseudocapacitive electrode material. However, the CV

deteriorated quickly as the scan rate increased, indicating a low conductivity.

Fig. 2-4: Cyclic voltammograms of electrodeposited MnO2 at scan rates of 10mV/s,

50mV/s, 100mV/s and 200mV/s in 0.5M Na2SO4, adapted from [28]

/cm

2

12

It was also found that the performance of MnO2 based electrodes strongly depended on

the synthesis methods. The methods, such as electrodeposition [28], hydrothermal [29],

micro-emulsion [30] and template-synthesis [31], directly controlled the crystal

structures (i.e. α-, β-, γ-type), morphologies, electrode thickness and pore structures,

which had significant effects on the electrochemical performance [5]. As suggested in

Fig. 2-4, the main challenge to commercialize the MnO2 is the low ionic and electronic

conductivity. In additions, problems such as dissolutions and low surface area have also

impeded the use of MnO2 as pseudocapacitive electrode material. As a result, the

current trend is to develop nano-structure MnO2 [32, 33] to increase the specific surface

area and to enhance the ion diffusions; or to grow MnO2 onto carbon materials such as

CNT and graphite [34, 35] as a composite material to increase the conductivity of the

electrode. However, these approaches may lead to problems such as poor bonding

between MnO2 and the substrate. Alternative transition metal compounds are needed to

be developed.

2.3.3 Molybdenum Oxides (MoO2/MoO3) and Nitride (Mo2N/MoN)

Molybdenum has multiple oxidations states ranging from –II to +VI, contributing to a rich

electrochemistry that is suitable for electrochemical applications. The E-pH diagram of

Mo-water system is shown in Fig. 2-5 [36]. The stable Mo oxides are MoO2 and MoO3 in

neutral and acidic environment.

13

Fig. 2-5: E-pH diagram of a Mo-H2O system, adapted from [36]

The pseudocapacitive behavior of Mo has been investigated in its oxide and nitride

forms. The electrochemical properties of the Mo oxides have been studied extensively.

A typical CV of potentiodynamically deposited Mo mixed oxides (Mo4+ and Mo6+) in

acidic electrolyte is shown in fig. 2-6. A large redox peak was observed at around -0.5 V

vs. Ag/AgCl, which was identified by Wang et al. [37] as partly overlapping double

reductions that can be expressed by the reactions (2 and 3):

MoO3 + xH+ + xe- MoO3-x(OH)x (0 < x < 1.6) (Rx-2)

MoO3 + 2yH+ + 2ye- MoO3-2y + yH2O (0 < y < 1) (Rx-3)

In fact, redox reactions were observed for Mo oxides in both acidic and neutral

electrolytes [38] due to a molybdenum bronze formation, expressed in reaction (4):

MoO3 + xA+ + xe- AxMoO3 (0 < x< 2) (Rx-4)

Where A+ can be cations such as H+, Li+, Na+ and K+. Liu et al. suggest that the redox

reactions are controlled by the diffusion process of the electrolyte cations within the Mo

oxides [38].

14

Fig. 2-6: Cyclic voltammograms of potentiodynamically deposited MoxO at various scan

rates, adapted from [39]

Mo oxides exhibit some charge storage at around -0.5 V vs. Ag/AgCl, but the

asymmetric anode and cathode current density and lack of mirror-image CV profile lead

to a poor capacitive behavior. Moreover, the large separation between the cathodic and

anodic peaks suggests a low electrical conductivity and poor reversibility. Studies also

demonstrated a limited stability window as shown in Fig. 2-7 [40]. For example, the

voltage window of Mo oxides in acidic electrolyte (pH=1.5) is limited from -0.5 V to 0.1 V

vs. Ag/AgCl. A potential lower than -0.5 V results in hydrogen evolution and oxide

reduction, while a potential higher than +0.1 V beaks down the passive layer. Therefore,

it can be concluded that Mo oxides are not, albeit they were described in some literature

as, an ideal pseudocapacitive electrode material for EC applications.

15

Fig. 2-7: Passive and transpassive region for Mo oxides from pH 1.5 to pH 8, adapted

from [40].

In general, Mo oxides and their thin films can be produced by methods such as vacuum

evaporation [41], sputtering[42], plasma spray [43], chemical vapor deposition

(CVD)[44], sol-gel technique [45] and electroplating [37, 46-49]. Among these

techniques, vacuum evaporation, sputtering, plasma spray and CVD require large-scale

vacuum system that increase the fabrication cost. Sol-gel route involves multiple

complicated precursors and require post heat treatment as well as long aging time.

Electrodeposition is the most cost-effective and the simplest way to produce Mo oxides.

The E-pH diagram in Fig. 2-5 well explains the feasibility of electrodeposition of MoO2

from MoO42- anions. For example, electrodeposition can occur by lowering the potential

below 0 V in a bath pH=2. The deposited MoO2 could be oxidized into MoO3 in air. Many

groups have investigated Mo oxide thin films, produced by electroplating processes, and

examined their electrochemical properties in acidic and neutral electrolytes [37, 46-49].

The films showed some charge storage capability and CVs similar to the one shown in

Fig. 2-6. However, as concluded previously, Mo oxides are limited as pseudocapacitive

material due to their low electrical conductivity and the poor reversibility.

16

Mo nitride, on the other hand, is a known pseudocapacitive material which exhibits

metal-like conductivity and reversible pseudocapacitive behavior similar to hydrous

ruthenium dioxide [21, 50-53]. In general, γ-Mo2N and δ-MoN were investigated in

pseudocapacitors applications. The active redox reaction for δ-MoN was proposed by

Liu et al. [21] as shown in reaction (5):

MoN + H+ + e- ↔ Mo2+ + NH2- (Rx-5)

The reaction is active in acidic electrolytes as the proton is the key factor that triggers

the redox reactions. Although it was not shown in the literature, the redox reaction of γ-

Mo2N is expected to be similar to reaction 5, which can be inferred in [21] as different γ-

Mo2N/δ-MoN ratios resulted in similar CVs. The formations of γ-Mo2N and δ-MoN

strongly depend on the synthesis conditions. For example, Liu et al. [21] conducted

CVD in ammonia at various temperature. The γ-Mo2N was formed at low temperature

400oC, which was attributed to insufficient decomposition of NH3 gas. In contrast, the N-

rich phase δ-MoN was formed at higher temperature 700oC due to sufficient supply of N.

In general, a mixture of γ-Mo2N and δ-MoN, namely MoxN, was produced in the study.

The γ-Mo2N/ δ-MoN ratio was determined by the amount of N supplied. CV of the Mo

nitride electrode prepared by the study is shown in Fig. 2-8 [21]. The profile exhibited

ideal rectangular charging/discharging curves and high reversibility with a good

capacitance of 250 F/g, suggesting Mo nitride to be a very promising electrode material

in pseudocapacitors applications.

17

Fig. 2-8: Cyclic voltammograms of CVD deposited MoxN (γ-Mo2N and δ-MoN) at 10

mV/s in 0.5 M H2SO4, dot line (fresh), solid line (after 2 weeks) and dash line (after 3

weeks), adapted from [21].

Similar to other transition metal compounds, the electrochemical behavior strongly

depends on the synthesis methods. In general, these methods can be categorized into

direct and indirect methods. Direct methods include chemical vapor deposition (CVD)[21]

and sputtering [54]. Indirect methods convert Mo oxides into Mo nitrides, examples are

multi-step pyrolysis [55], and powder chemical/thermal synthesis [50]. For instance,

Jaggers et al. [56] used a series of Mo oxides precursors to react with NH3 to produce

Mo nitrides. The reactants and products are listed in table 2-2. The majority of the

product was face center cubic γ-Mo2N with nitrogen atoms randomly occupying half of

the octahedral sites. Two possible routes were proposed in the study, which are

expressed in reactions (6 and 7):

MoO3 MoOxN1-x Mo2N (Rx-6)

MoO3 MoO2 Mo2N (Rx-7)

Where the MoO3 could be first converted to oxynitride (MoOxN1-x) or reduced to MoO2

before the final Mo2N phase is formed.

18

Table 2-2: Products from the reactions of molybdenum oxides with ammonia, adapted

from [56]

Reactants in NH3 environment Products

MoO3 γ-Mo2N

H0.04MoO3 γ-Mo2N

H0.13MoO3 γ-Mo2N

H0.31MoO3 γ-Mo2N

H0.80MoO3 γ-Mo2N and δ-MoN

MoO2 γ-Mo2N and δ-MoN

Among the mentioned approaches, powder synthesis produced a high surface area and,

thus, high capacitance. However, it requires attaching the active materials to the current

collectors with binder and filler (mostly carbon), which often result in reduced

conductivity. Multi-step pyrolysis involved 50% tantalum oxide and several heat-assisted

deposition steps prior to 700 oC nitridation [55]. Chemical vapor deposition and

sputtering could generate intimate contact between the active material and the current

collectors. However, the fabrication of large sized electrodes might be limited to the size

of ultrahigh vacuum systems. Therefore, it is necessary to develop a method that can

cost-effectively produce pseudocapacitive Mo nitride electrode with electrochemical

behaviors similar to the one shown in Fig. 2-8.

2.3.4 Tungsten Oxide (WO3) and Nitrides (W2N)

Tungsten is another transition metal with various oxidation states range from -II to +VI,

out of which the most stable one is +VI. Metallic tungsten and its oxides are widely used

in the industry such as in cutting tools and catalysts [57]. Tungsten trioxide (WO3) has

been investigated as electrochromic material [58, 59] due to the rich oxidation states

and induced color effects. It also attracts considerable attention as electrode material in

pseudocapacitor application [60-62].

The E-pH diagram [63] of tungsten-water system is shown in Fig. 2-9. It is clear that

WO3, W2O5 and WO2 are stable at a potential higher than 0 V in an acidic environment.

19

Fig. 2-9: E-pH diagram of a W-H2O system, adapted from [63]

The electrochemical process of WO3 can be expressed according to the following

reaction (8) to form tungsten bronze [64]:

WO3 + xH+ + xe- HxWO3 (Rx-8)

It was also confirmed by Changshin et al. [65] that the electrochemical process is

activated in an acidic environment as the proton plays an important role in redox

reactions. And the kinetics of the tungsten bronze formation are determined by the

diffusion of the protons through amorphous or crystalline WO3 materials [64].

Various methods such as laser vaporization [66], CVD [67], multi-step pyrolysis [60] and

electrodeposition [68-71] have been used to produce tungsten oxides particles and thin

films. Laser vaporization employed laser that directly vaporize tungsten metal with

controlled condensation of corresponding vapor phase so that the particle size and

aggregation could be well controlled. However, it required high energy laser beam and

vacuum chamber. And the electrode production was limited to a relatively small

laboratory scale. CVD could produce WO3 film with extremely high purity and the film

20

composition and characteristics could be easily tailored by changing deposition

parameters. However, the drawbacks of CVD include the requirements of vacuum

chamber, complicated gas feed and larger amount of energy to evaporate metallic

tungsten electrode. Multi-step pyrolysis is a relatively simple method to synthesize WO3.

For example, Yoon et al. [60] reported the electrochemical properties of the bulky (b-)

and meso-porous (m-) tungsten oxide produced by a template method. The

performance is shown in figure 2-10.

Fig. 2-10: Cyclic voltammograms of template method prepared WO3 at 5 mV/s in 2 M

H2SO4, adapted from [60], b- stands for non-porous bulk and m- for meso-porous.

The pseudocapacitance was mostly at the negative voltage with a total capacitance

around 200F/g. The charge storage decreased dramatically as the voltage increased.

However, the technique required silica templates and high calcination temperature.

Electrodeposition could be the most common method to fabricate WO3 due to low cost

and easy processing [68-71]. In electrodeposition, the solution bath is usually made by

dissolution of tungsten metal into hydrogen peroxide. Alternatively, tungstic acid

(H2WO4) was used as electrolyte. The electrodeposition normally occurs in low pH

around 1-2 and low potential. For example, in a plating bath solution where pH=1, the

electrodeposition occurs below 0 V according to Fig. 2-9.

21

Tungsten nitrides have also been investigated as electrode material as the nitrides often

have higher electric conductivity compare to oxides. W2N can be synthesized though

some one-step methods such as atomic layer deposition (ALD)[72, 73] and CVD[74].

Again these one-step deposition methods require complicated precursors and large

vacuum systems, making them not feasible for massive industrial productions. In

contrast, simple conversion from commercial W oxides through heat treatment in N-

containing environment is more practical and promising when taking cost into

consideration. For example, Ko et al. reported the electrochemical behaviors of W2N

synthesized by heat treatment of WO3 in an ammonia environment, as shown in Fig. 2-

11 [75].

Fig. 2-11: Cyclic voltammograms of W2N at 50 mV/s in 1 M KOH, adapted from [75].

The CV was indeed more reversible and rectangular compared to that of WO3. However,

the pseudocapacitive reactions occurred in an alkaline electrolyte (e.g. 1 M KOH),

suggesting a completely different charge storage mechanism. The scan rate applied in

Fig. 2-11 was 50 mV/s, which was much higher compare to that of Fig, 2-10, suggesting

a higher conductivity of W2N. Nonetheless, this work mainly focuses on the

development of WO3 electrode for proton conducting electrolyte using an

electrodeposition process. The exploration of tungsten nitride is subjected to future

investigation.

22

2.3.5 Vanadium Oxides (V2O5, VO2 and H2V3O8) and Nitride (VN)

Vanadium oxides, especially V2O5, VO2, and H2V3O8, have been investigated as

electrodes for pseudocapacitors due to their low cost, abundance and their potential

pseudocapacitive characteristics [76-85] . The E-pH diagram of V species is shown in

Fig. 2-12[86]. Except the V2O5, which is stable in acidic environment; the oxides with

lower V oxidation states are stable in neutral and alkaline environments, including mix

oxides H2V3O8 and VO2. Researchers have found that the V oxides are capable of

intercalating Li ions. For instance, the crystalline V2O5 enables the insertion of small

guest species such as Li+ into its perovskite-like cavities and form vanadium bronze

according to the following reaction (9) [87]:

V2O5 + xLi+ + xe- LixV2O5 (Rx-9)

Fig. 2-12: E-pH diagram of a V-H2O system, adapted from [86]

In addition to crystalline V2O5, amorphous V2O5 such as V2O5.H2O xerogels [88] and

aerogels[89] also have Li-intercalation properties. Recent studies further discovered

similar intercalations characteristics on VO2 [85], H2V3O8 [78] and VOx.H2O [79].

However, the electrochemical performances of the V oxides based electrodes are

23

limited by the low electrical conductivity and low diffusion coefficient of Li+. Many

projects were developed to address this problem. For example, Wang et al. [90] blended

Ag nanowire with V2O5 to form composite Ag0.08V2O5 to increase the electrical

conductivity. West et al. reported [88] that adding more water content expanded the

distance between oxide layers that enhanced the capability of Li ion diffusion. Hibino et

al. [91] concluded that, in general, the V oxides should be structurally-modified towards

a more open structure, allowing easy diffusion of Li ions. Overall, the electrochemical

performance of the V oxide electrodes depends strongly on their structures, chemistries,

and morphologies [76, 78, 80], which are directly governed by the methods for synthesis.

Vanadium oxides have been produced via sol-gel formation[77], hydrothermal

synthesis[78], atomic layer deposition [79] and electrodeposition [80-84]. The electrodes

synthesized by these methods showed a large variation in electrochemical behavior. For

example, V2O5 aerogel prepared from a sol-gel method [77] showed a particularly high

capacitance of 1300 F/g. The synthesized V2O5 aerogel was an ideal bulk material for

energy storage due to large surface area, but it was less suitable for high rate

applications as a very low scan rate of 0.1 mV/s must be used in the study. H2V3O8

synthesized by a hydrothermal process showed good specific capacitance of 236

mF/cm2 or 121 F/g but with a non-ideal CV [78]. This might be due to the formation of

well-crystalized material that resulted in a battery-like electrode. Mixed oxide VOx (VO2,

V2O3, and V2O5) produced by atomic layer deposition [79] exhibited a high capacitance

of 600 F/g and high power capability at 1 V/s. However, atomic layer deposition involved

high cost, which is not viable for scaling up. Electrodeposition [80-84] is an attractive

low-cost method of producing vanadium electrodes. Since electrodeposition occurs in

aqueous plating bath, the produced oxides were usually hydrated and ideal for ion

transportation. For example, Fig. 2-13 shows the CVs of a mixed vanadium oxides

electrode prepared by electrodeposition on a graphite substrate [82]. The profile

exhibited an ideal pseudocapacitive behavior from -0.2 V to 0.8 V in a neutral electrolyte

(3 M LiCl). As the scan rate increased from 25 mV/s to 500 mV/s, the CV retained its

rectangular shape, indicating a high conductivity. Although electrodeposition is a low-

cost method to produce V oxides electrode with ideal electrochemical behaviors in EC

applications, alternative methods such as electroless deposition to further simplify the

24

production process are also possible and would be more appealing. However, vanadium

oxides electrode produced by such methods (i.e. electroless deposition) for

pseudocapacitors applications has never been reported in the literature.

Fig. 2-13: Cyclic voltammograms of VOx at (1) 25 (2) 100 (3) 300 (4) 500 mV/s in 3 M

LiCl, adapted from [82]

Vanadium nitride (VN) has also been studied as electrode material for EC. Since the VN

electrodes are eventually used in alkaline electrolyte in ambient condition. Partial

oxidation of the nitride is inevitable. Choi et al. [92] proposed the electrode process in

1M KOH, shown by the following reaction (10):

VNxOy + OH- VNxOy//OH- + VNxOy-OH (Rx-10)

Where VNxOy//OH- is the double layer and VNxOy-OH represents the successive

oxidation of the OH- group on the double layer interface. Therefore, both double layer

and redox reaction contribute to the capacitance, and the latter is the predominant factor.

There are several approaches to produce VN, including some one-step methods such

as mechanochemical synthesis [93], magnetron sputtering [94] and chemical reaction of

VCl4 and NH3 [92]. A common two-step method is also available: to synthesize V2O5

first, followed by conversion to VN via temperature-programmed heat treatment in

25

reducing gas such as ammonia [95-98]. Fig. 2-14 shows the CVs of VN synthesized by

converting V2O5 into VN via a temperature-programmed heat treatment. The highest

capacitance (234 F/g) was obtained in a KOH alkaline electrolyte, inferring the OH-

cation activates the pseudocapacitive reactions.

Fig. 2-14: Cyclic voltammograms of VN in 0.1M of various electrolytes, adapted from [50]

26

Chapter 3: Objectives

In chapter 2, the development of pseudocapacitive material can increase both

capacitance and power capability for an EC. Hydrous ruthenium dioxide is the

benchmark but limited by the high cost. Alternative transition metal compounds,

especially Mo oxide-nitride, W oxide and V oxide have been investigated and have

shown some promising electrochemical properties. However, the behavior strongly

depends on the fabrication method. So far low-cost methods to produce high

performance pseudocapacitive electrodes have not yet been achieved. Thus further

investigations are needed and improvements can be made. Ideally, the developed

electrodes should have the following properties:

1. Ideal capacitive behavior (i.e. rectangular CV and linear charging/discharging

profile, see section 4.4.1 and 4.4.2)

2. Reasonable area-specific capacitance

3. Fast redox kinetics

4. High conductivity

5. Long cycle life

6. Large voltage window

7. Stable in various electrolytes

8. Inexpensive material and process

An electrode with a single chemistry can hardly meet all criteria. However, through

careful tailoring of the electrode chemistry, adjusting the surface structure and using

advanced cell techniques such as asymmetrical configurations, optimized

electrode/device performance can be achieved. In this work, three compounds (i.e. Mo

oxide-nitride, W oxide and V oxide) were explored. To meet the low cost and easy-to-

process objectives, techniques such as electrodeposition, low temperature heat

treatment and electroless deposition were employed. The materials were tuned to have

the optimum pseudocapacitive behaviors.

27

3.1 Objectives for Mo Oxide-nitride

It was discussed that the environmentally benign N2 can be used to nitride Mo oxide in a

previous study. However, the chemistry, crystal structures and optimum process

parameters were unclear. It is also necessary to compare N2 with other common

nitridation gases such as forming gas and ammonia. In this part of the study, the

following approaches to further investigate the material were undertaken:

1. Perform a comparative study of heat treatment in different gas environments.

2. Understand the detailed chemistry, surface morphology and crystal structure of

the electrode synthesized by the best method for synthesis.

3. Apply the electrodes in EC cells to test their performance.

3.2 Objectives for W Oxide

The electrodeposition techniques for tungsten oxide have been well studied in the

literature. The goal of this study is to use a technique from the literature to fabricate a W

oxide electrode and to compare with Mo oxide-nitride electrodes. To identify any

similarities and complementary properties, asymmetrical devices were also studied to

improve performance. The following approaches were proposed:

1. Utilize an electrodeposition technique to synthesized tungsten oxide electrode.

2. Perform material and electrochemical characterizations on the electrodes to

understand their properties.

3. Explore the performance of the asymmetric device using both a W oxide

electrode and a Mo oxide-nitride electrode.

3.3 Objectives for V Oxide

The objective of this part of the study is to develop an inexpensive and easy-to-process

method to produce vanadium oxide electrodes suitable for EC applications. The

following approaches were proposed:

28

1. Develop a low cost method to synthesize V oxide electrodes and to optimize the

process conditions.

2. Perform material and electrochemical characterizations on the electrodes to

study the properties.

3. Apply the electrodes in EC cells to test their performance.

29

Chapter 4: Experimental

This work mainly focuses on electrodeposition, electroless deposition and low

temperature heat treatment. Table 4-1 is a brief summary of the methods utilized in this

study to produce the electrodes. Table 4-2 lists the chemicals used in the processes.

Table 4-1: Summary of electrode fabrication methods

Mo oxide-nitride W oxide V oxide

Electrodeposition: Electrodeposition: Electroless deposition:

0.4 M Na2MoO4, 0.12 M Na2SO4

and 5 mM H2SO4

0.06 M Na2WO4, 0.1M

H2SO4 and 0.09 M H2O2

0.25 – 0.75 M VOSO4 and

0.3 ml 1 M NaOH

-0.85V to 0 V at 0.01 V/s 37 min -0.85V to 0 V at 0.01 V/s 37

min 3 - 7 days

Heat treatment of

electrodeposited Mo oxide:

10oC/min to target temperature

for 3 hours and cooled naturally

In N2, forming and NH3

Table 4-2: List of chemicals used in the fabrication process

Mo oxide-nitride W oxide V oxide

Substrate Ti foil 0.13mm thick ( McMaster-Carr)

Main deposition

chemical Na2MoO4 (Alfa Aesar) Na2WO4 (Alfa Aesar) VOSO4 (Alfa Aesar)

Additional

deposition

chemicals

H2SO4 (Caledon)

Na2SO4 (Alfa Aesar)

H2SO4 (Caledon)

H2O2 (Life) NaOH (EMD)

Heat treatment

gas

N2 grade 5 (Linde)

NH3 grade 5 (Linde)

Forming gas 10%H2 90%

N2 (Linde)

30

4.1 Fabrication Method of Mo Oxide-nitride Electrode

Mo based electrodes were synthesized by electrodeposition of Mo oxide onto Ti

substrates followed by thermal nitridation. A thin titanium foil was ultrasonically cleaned

for 5 min each sequentially in acetone, methanol, and DI water. The electroplating bath,

comprised of 0.4 M Na2MoO4 and 0.12 M Na2SO4, was adjusted to a pH of 2 by the

addition of H2SO4. The Mo oxide film was deposited onto a titanium substrate via

potential cycling between -0.85 V and 0 V (vs. Ag/AgCl) at a rate of 0.01 V/s. 13 cycles

of electrodeposition were used, which translates to about 37 min electroplating.

For nitridation, the as-deposit Mo oxide electrodes were heated in a tube furnace (MTI

GSL-1100) in 3 different gas atmospheres including nitrogen, ammonia and forming gas

(10% H2 and 90% N2) at 400 oC (Fig. 4-1). Once the most promising gas was identified,

heat treatment temperature was varied between 200 to 700 oC to obtain an optimum

process condition from both performance and energy standpoints. This optimized

condition was used to process all the samples reported in the remaining study of Mo.

During nitridation, the furnace temperature was ramped up at a rate of 10 oC/min and

was held at the target temperature for 3 hours before it was allowed to cool down

naturally to room temperature. The gas flow rate was approximately 0.1 L/min during the

entire heat treatment process. For comparison, an as-deposit Mo oxide and a heat-

treated pure titanium foil were also used as electrodes.

Fig. 4-1: Example temperature profile when heat up to 400oC

31

4.2 Fabrication Method of W Oxide Electrode

W oxide electrodes were also synthesized via electrodeposition. The electroplating bath

was comprised of 0.06 M Na2WO4, 0.1M H2SO4 and 0.09 M H2O2, which needed to be

fresh-made as the concentration of H2O2 varied significantly with time. Similar to the

deposition of Mo oxide, a thin titanium foil was ultrasonically cleaned for 5 min each

sequentially in acetone, methanol, and DI water. Cyclic deposition was used with cycling

from -0.85V to 0V, at a scan rate of 10 mV/s for 13 cycles. A large piece of titanium foil

was used as the counter electrode and Ag/AgCl as the reference electrode.

4.3 Fabrication Method of V Oxide Electrode

The substrate for vanadium oxide deposition was a thin titanium foil, ultrasonically

cleaned for 5 minutes each in acetone, methanol, and DI water. Each Ti substrate was

immersed in VOSO4 solutions in a selected concentration. 0.3 ml 1 M NaOH was added

into each solution to initiate the precipitation process. The Ti substrate was kept in this

solution under ambient conditions until the substrate was covered by a greenish film.

The effects of VOSO4 concentration and deposition time on the capacitance of the

electrode were studied. Deposition baths were prepared with VOSO4 concentrations of

0.25 M, 0.5 M, and 0.75 M. Three Ti substrates were immersed into each bath. A coated

sample was removed from the solution and measured at the same time on day 3, 5, and

7, respectively. The experiment was repeated a few times to ensure the reproducibility.

The electrodes in the following are referred to by their experimental conditions, e.g.,

“0.5M-5days” for the electrode plated in 0.5 M VOSO4 for 5 days.

4.4 Characterizations of the Electrodes

The surface morphology of the electrodes was examined by an FEI Quanta FEG 250

environmental scanning electron microscope (ESEM). X-ray diffraction (XRD) analyses

were conducted using a Philips XRD system with a monochromatized Cu-Kα anode.

32

The surface chemistry of the electrodes was analyzed with X-ray photoelectron

spectroscopy (XPS) using a thermo scientific K-Alpha instrument.

The electrochemical properties of the electrodes were characterized in various

electrolytes by cyclic voltammetry (CV) (see section 4.4.1) utilizing a CHI760D

potentiostat and an Ivium CompactStat in a 3-electrode beaker cell shown in Fig. 4-2a.

Charging and discharging tests were also employed for cycle life performance using the

Ivium CompactStat. Ag/AgCl was used as the reference electrode and a large Ti foil

was used as the counter electrode. In addition, symmetric and asymmetric two-

electrode EC cells were assembled. As for Mo and W based ECs, two electrodes were

immersed into 0.5 M H2SO4 and kept approximately 0.7 cm apart (beaker EC) as shown

in Fig. 4-2b. For the V based EC, the two electrodes were separated by a 0.17 mm thick

filter paper soaked in electrolyte (filter paper EC) as illustrated in Fig. 4-2c. CV tests

were conducted at multiple scan rates from 5 mV/s to 2 V/s.

Fig. 4-2: Schematic diagrams of the testing set-up (a) for single electrode (b) beaker cell

EC (c) filter paper EC

*WE: working electrode, CE: counter electrode, RE: reference electrode

4.4.1 Cyclic Voltammetry

Cyclic voltammetry (CV) is a frequently used dynamic method to assess the

electrochemical behaviors of a single electrode or a complete EC. It can be used to

evaluate general capacitive behavior, capacitance, kinetics of the electrode, power

33

performance and cycle life. A typical 3-electrode cell includes a working electrode (WE),

a counter (CE) and a reference electrode (RE); where the WE is the testing electrode,

CE allows the current to pass through and RE serves as a potential reference. The

potential scan is programed at an initial voltage where no electrolysis should occur, then

the scan sweeps at a fixed scan rate in units of V/s to the switching potential and

reverses direction back to the initial point at the same scan rate. The current is recorded

during the scan and plotted against the voltage, known as a cyclic voltammogram (CV)

as illustrated in Fig. 4-3. For example, Fig. 4-3a shows reversible oxidation/reduction

reactions. Two current peaks indicate the voltage where the main reactions occur. If the

reactions are controlled by diffusion or the kinetics are not fast enough, a peak

separation can be observed in Fig. 4-3a. Fig. 4-3b shows the reversible

oxidation/reduction reactions on very thin layers where fast kinetics apply. Therefore, no

peak separation can be observed.

Fig. 4-3: Cyclic voltammograms of reactions (a) diffusion control with slow kinetics (b)

thin-layer condition with fast kinetics, adapted from [99].

An ideal EDLC has a rectangular CV profile as shown in Fig. 4-4. The capacitance can

be calculated by equation (6):

(Equ-6)

Where C is the capacitance, I is the current and v is the scan rate. In most case, the

obtained capacitance values are divided by the area or the weight of the electrode to

34

obtain area-specific capacitance (F/cm2) or the weight specific capacitance (F/g),

respectively.

Fig. 4-4: Example CV profile of an ideal EDLC.

In a pseudocapacitor, the CVs are more complicated. For example, Fig. 4-5 depicted

the CV of a Ru dioxide electrode. Each current peak represents a change of oxidation of

Ru from (II) to (III) and to (IV). The current peaks overlap to give a CV shape that is

similar to that of an EDLC. In fact, the CVs of a pseudocapacitor can never be perfect,

the capacitance must be calculated by equation (7):

(Equ-7)

Where Q is the total charge stored in the electrode and V is the voltage window

scanned.

Fig. 4-5: Example cyclic voltammograms of a Ru dioxide electrode, adapted from [19].

35

The conductivity and kinetics of the electrode can be obtained by evaluation of CVs at

high scan rates. Maintaining an ideal rectangular shape for the CV profile at high scan

rates indicates a high conductivity and fast kinetics. The cycle life of the electrodes can

be analyzed through CVs after repeated cycles. A good cycle life can be seen if

thousands of cycles of CV overlap. Overall, CV is a very powerful technique to evaluate

electrochemical performance of an electrode, which is the main tool utilized in this

project.

4.4.2 Charging and discharging

Charging and discharging is an electrochemical method complementary to CV. It

charges and discharges the electrode at a constant current density and the

corresponding voltage is recorded. An ideal capacitor obeys the following equation 8:

(Equ-8)

Where I is the current density, V is the voltage and t is the time. If the Capacitance and

the current density are constant, the voltage must increase or decrease linearly as

illustrate in Fig. 4-6. Since the charging/discharging is more analogous to real working

conditions than CV, it is a complementary tool to CV, especially for evaluating cycle life.

For example, a good cycle life can be seen if both t and dV/dt for each

charging/discharging cycle are constant after long repeated cycling.

Fig. 4-6: Example charging/discharging profile for an ideal EC.

36

Chapter 5: Results and Discussion

In this chapter, 3 transition metal compounds: Mo oxide-nitride, W oxide and V oxide are

discussed. They are relatively new as electrode materials in EC applications. Recent

studies have shown some promising electrochemical behaviors, yet improvements can

be made through developing better methods for synthesis. Moreover, these materials

have properties in common that could be leveraged into an EC for synergistic or

complementary performance.

5.1 Molybdenum Oxide-nitride

In previous work [100, 101], a Mo oxide electrode nitrided by N2 was reported to have

close-to-ideal pseudocapacitive behaviors and was used to form solid EC cells through

a highly conductive proton-conducting polymer electrolyte. However, the focus of that

work was to prove the concept that the electrode can be synthesized with N2 at low

temperature (400oC) and demonstrate the capability of the solid electrolyte on the

electrodes. In this section, a comparative study was performed to develop a deep

understanding of the effects of nitridation on Mo oxide electrode and their

pseudocapacitive behaviors in both acidic and neutral electrolytes.

5.1.1 Effects of Electrodeposition

Before any nitridation, Mo oxide was deposited on the Ti substrate by electrodeposition.

A cyclic deposition was used. It was found that the electrodeposition was strongly

affected by parameters such as the voltage window, deposition speed and total

deposition time. Therefore, the effects of these parameters were studied to identify the

optimized electrodeposition method for the Mo oxides, while the as-deposit Mo oxide

electrodes were nitrided in N2 following the methods from the previous work [100, 101].

Capacitance of the electrodes was measured and used as output to evaluate the

deposition parameters (table 5-1):

37

Table 5-1: Capacitance of the nitrided Mo oxide electrodes through different

electrodeposition conditions

Experiment No. Voltage Window Deposition Rate Average Capacitance

1

-0.75 V to 0 V

0.3 V/s 3.2 mF/cm2

2 0.05 V/s 5.5 mF/cm2

3 0.01 V/s 6.3 mF/cm2

4

-0.85 V to 0V

0.3 V/s 7.1 mF/cm2

5 0.05 V/s 8.0 mF/cm2

6 0.01 V/s 8.2 mF/cm2

*capacitance measured through CV at a scan rate of 0.1 V/s in 0.5 M H2SO4, the total

deposition time was fixed at 37 min

The highest capacitance was from electrodes deposited in a voltage window from -0.85

V to 0 V and a scan rate of 0.01 V/s. A larger voltage window allowed for more charge

transfer, and resulted in more deposition of Mo oxide onto the Ti substrate. However,

further extension of the voltage window led to hydrogen evolution that reduced the

current efficiency. Therefore, the voltage window from -0.85 V to 0 V was selected.

Results also showed that a slower deposition rate led to a higher capacitance, probably

due to sufficient time for ions to replenish the Nernst diffusion layer. A potential window

from -0.85 V to 0 V and a deposition rate of 0.01 V/s were finally selected to produce the

as-deposit Mo oxide electrodes for the following study. Patil et al. [102] concluded that

the as-deposit Mo oxide is usually hydrous and the deposition can be expressed by the

following reaction (11):

MoO42- + 2e- + 4 H+ + n H2O MoO2-n(OH)2n + 2 H2O (Rx-11)

Besides the voltage window and deposition rate, the deposition time was also

investigated and a total of 37 min was selected. Nonetheless, compared with nitridation,

electrodeposition only produces an intermediate product and is not the focus of this

investigation. Details of the electrodeposition analyses are given in Appendix B-1.

38

5.1.2 Comparative Study of the Effects of Gas for Nitridation

The effects of nitridation were investigated by comparing among 3 nitridation gases.

Two characterization tools were applied: CV was used to obtain the electrochemical

properties, while the XPS was used to analyze the surface chemistry. These

characterizations were applied to determine the best nitridation condition and to

investigate the reasons behind.

Before studying any nitrided electrodes, baseline CVs of an as-deposit Mo oxide

electrode and a Ti substrate were first established in Fig. 5-1-1a. The CV of the Ti

substrate showed a much lower current density than that of the Mo oxide, suggesting

the electrochemical behavior of Ti was limited by double layer charging/discharging. The

CV profile of the Mo oxide matched reports in the literature [37, 39, 46-49], showing a

strong electrochemical redox activity from -0.55 V to 0 V. At more positive potential, the

Mo oxides were not stable. Although some Faradaic oxidation/reduction occurred on the

electrodeposited Mo oxide and some charge could be stored, the irreversible

anodic/cathodic peaks and the lack of “mirror-imaging” profile made it a non-ideal

pseudocapacitive electrode material compared to RuO2 [22] and Mo2N [52].

The CVs of Mo oxide electrodes treated in different nitridation gas, including nitrogen,

ammonia and forming gas (10% H2 and 90% N2), are shown in Fig. 5-1-1b. The CVs of

these electrodes were obtained in a voltage window between -0.15 V and +0.45V, to

match the window reported for Mo nitrides in H2SO4 by Liu et al. [21], Deng et al. [55]

and Li et al. [52]. Comparing Fig. 5-1-1a to 5-1-1b, the considerable difference in

electrochemical behavior between Mo oxide and nitrided Mo oxide was a strong

evidence of change in their surface composition and chemistry. Examining Fig. 5-1-1b,

the CV of ammonia treated Mo oxide showed a tilted profile while the electrode treated

with forming gas had relatively small capacitance. The electrode treated in N2 exhibited

the most symmetrical and rectangular CV profile and the highest area-specific

capacitance among the three electrodes, showing the best pseudocapacitive behavior.

39

Fig. 5-1-1: Cyclic voltammograms in 0.5 M H2SO4 at 100 mV/s: (a) Mo oxide vs. Ti

substrate between -0.55 V and 0 V (b) Mo oxide, Mo oxide heat treated in N2, ammonia

and forming gases between -0.15 V and 0.45 V.

The surface compositional analyses of Mo oxides before and after nitridation were

performed using XPS to reveal the chemistry that contributed to various electrochemical

behaviors. Since the N 1s peak overlaps with Mo 3p peaks, it was difficult to

deconvolute the nitrogen signal; Mo 3d spectra were analyzed. Fig. 5-1-2 shows the Mo

3d spectra for as-deposit Mo oxide (2a), Mo oxide treated in nitrogen (2b), in forming

gas (2c) and in ammonia (2d). The change in intensity and the shape of the Mo 3d

spectra suggested a change in the chemistry of Mo cations after heat-treatment in those

gas atmospheres. Due to the spin-orbit effects, the Mo 3d splits into Mo 3d 5/2 and Mo

3d 3/2. A detailed analysis of Mo oxides in Fig. 5-1-2a revealed the Mo 3d 5/2 peaks at

232.0 eV and 230.7 eV, which corresponded mainly to Mo6+ with a small amount of Mo4+

[103]. This suggested that the electrodeposition process produced a mixture of MoO3

and MoO2. For all nitride samples in Fig. 5-1-2 b-d, a third Mo 3d 5/2 peak appeared at

229.3 eV corresponding to the formation of Mo cations at a lower oxidation state,

referred to as Moδ+ (0<δ<4) [103-105]. The Moδ+ peak can be attributed to the formation

of a Mo2N-type compound as reported by Kim et al. [103], McKay et al. [106], Shi et al.

[107, 108] and Becue et al. [104]. The overall surface compositions of the as-deposit Mo

oxide and all of the nitrided Mo oxides are listed in Table 5-2.

40

Table 5-2: Compositions of electrode surfaces treated in different gases

Species Mo

6+(MoO3)

(at.%)

Mo4+

(MoO2)

(at.%)

Moδ+

(Mo2N-

type) (at.%)

Mo oxide 79.3 20.7 0

Mo oxide heat treated in N2 62.9 17.7 19.4

Mo oxide heat treated in

forming gas 58.0 17.4 24.6

Mo oxide heat treated in NH3 43.3 8.6 48.1

In Mo oxide, the percentage ratio of Mo6+ to Mo4+ was 79.3% to 20.7%. After various

heat treatments in nitrogen containing media, the surface Mo6+ and Mo4+ species

decreased and replaced by Moδ+. This indicated that Mo oxide was partially reduced

and was likely converted to nitride after the heat treatment in all three gas atmospheres.

The surfaces of these electrodes were covered with a mixture of Mo species,

corresponding to different oxidation states in Mo oxides and nitride. Among the three

nitrided samples, the Mo oxide after ammonia had the highest Moδ+ content followed by

that in forming gas and then pure N2. This trend can be attributed to the reduction power

of both ammonia and forming gas. However, what was unexpected was the effect of N2,

which also reduced part of surface oxides, albeit only less than 20 at.%. Nevertheless,

the higher Moδ+ from ammonia and forming gas did not lead to a better electrochemical

performance compared to that in N2 as illustrated in Fig. 5-1-1b. Moreover, since N2 is a

cheaper, safer and more environmentally benign gas compared to forming gas and

ammonia; and heat treatment in N2 led to the best electrochemical performance, it was

selected for further investigation.

41

Fig. 5-1-2: High resolution XPS spectra for Mo 3d: (a) Mo oxide, (b) Mo oxide heat

treated in N2, (c) forming gas, (d) ammonia.

The effect of the heat treatment temperature on the pseudocapacitance of nitrided Mo

oxide was further studied by comparing the CV profiles of Mo oxide after being heat

treated at 200, 400, 600, and 700 oC in N2 gas (Fig. 5-1-3). Mo oxide treated at 400 oC

demonstrated the highest capacitance; additional analyses showed a performance

plateau between 400 and 450 oC. Thus, a heat treatment temperature of 400 oC was

considered optimal as it yielded the most energy efficient process and products.

Compared to conventional heat treatment which was typically carried out in NH3 at 700

oC or higher [52, 55, 106, 109, 110], this approach used a far more benign gas and at a

much lower temperature. The significant improvement in electrochemical behavior

demonstrated the effectiveness of this low-temperature, N2-based heat treatment

process.

42

Fig. 5-1-3: Cyclic voltammograms in 0.5 M H2SO4 at 100 mV/s of Mo oxide-nitride heat

treated at various temperatures in N2.

5.1.3 Characterizations on Optimized Electrode in Acidic Media

An effective low temperature ( 400oC ) heat treatment in N2 was demonstrated to

produce nitrided Mo oxide electrodes for ECs. To further study the nitrided electrodes,

morphology and structure characterizations such as ESEM and XRD, electrochemical

analyses such as cycle life and charging/discharging were performed.

5.1.3.1 Material Characterizations

Surface properties such as porosity, thickness and coating/substrate interface

determine the capacitance and cycle life. It is important to investigate the electrode

surface through imaging methods. The surface morphologies of the as-deposit Mo oxide

and the heat-treated Mo oxide at 400 oC in N2 were examined by ESEM. The as-deposit

Mo oxide electrode (Fig. 5-1-4a) showed a smoothly coated film with dry “mud-like”

cracks [102, 111], indicating that the Mo oxide was uniformly deposited onto the Ti

substrate and that the oxide was hydrated. The cracks were the result of

drying/dehydration of the Mo oxide films after electrodeposition [102, 112], which was

also observed in other electrodeposited metal oxide films such as RuO2.xH2O [111, 113].

43

Fig. 5-1-4: ESEM micrographs of Mo electrodes: (a) as-deposit Mo oxide, (b) nitrided

Mo oxide, (c) side-view of as-deposit Mo oxide, (d) side-view of nitrided Mo oxide.

The cross-section of the Mo oxide was examined after bending the electrode in liquid

nitrogen to mechanically induce film delamination. The average thickness of the Mo

oxide layer was approximately 300 nm (Fig. 5-1-4c). After heat treatment at 400 oC in N2,

the Mo oxide film appeared partially dehydrated (Fig. 5-1-4b) as suggested by Patil et al

[102] with a ca. 20% reduction in thickness (Fig. 5-1-4d). The film showed some heat

induced sintering after heat treatment, which might also promote a strong bonding

between the film and the substrate. The crystal structure of the as-deposit and heat

treated Mo electrodes in N2 was further characterized by XRD.

The XRD patterns of the as-deposit Mo oxide exhibited an amorphous structure

probably due to its high level of hydration (Fig. 5-1-5a). Thus, an as-deposit Mo oxide

sample was subjected to heat treatment in air using the same temperature and time

(400 C, 3 hr) to serve as reference for that heat treated in N2. The XRD patterns of both

air treated Mo oxide and N2 treated Mo oxide are shown.

44

Fig. 5-1-5: XRD patterns of (a) as-deposit Mo oxide (b) Mo oxide treated in air and (c)

Mo oxide-nitride treated in N2, with identical temperature profiles at a peak temperature

of 400 oC.

Although the deposited films were just a few hundred nanometers thick so that the

peaks of the substrate titanium were dominating, the characteristics of the deposited

films were clearly visible. The air treated Mo oxide (Fig. 5-1-5b) exhibited a structure

very similar to MoO3 (PDF #01-076-1003), indicating a high oxidation state of Mo

species after oxidation in air. In contrast, the Mo oxide after heat treatment in N2 (Fig. 5-

1-5c) showed a fingerprint that can be broke down into patterns of MoO3 (PDF #01-076-

45

1003), MoO2 (PDF # 01-073-1249) and γ-Mo2N [114-116]. The characteristic peaks of γ-

Mo2N were not as strong as it had only less than 20 at.% after the heat treatment in N2.

Based on the XPS and XRD analyses, the chemical composition of the N2 heat treated

Mo oxide was defined as Mo oxide-nitride or Mo(O,N)x, where x < 3. Other material

characterizations such as additional XPS study on Mo 3p orbital and Secondary Ion

Mass Spectroscopy (SIMS) were also undertaken to study the optimized Mo oxide-

nitride electrode, and the details are given in Appendix B-2 and B-3, respectively.

Results showed a trace amount of N on the Mo oxide-nitride electrode surface, again

demonstrating the effectiveness of using N2 to produce Mo oxide-nitride electrodes.

5.1.3.2 Electrochemical Characterizations

An EC is required to be capable of enduring thousands of charging/ discharging cycles

without significant degradation of its electrochemical behaviors. In this work, cycle life

was used to test its stability. The cycle lives of both Mo oxide and Mo oxide-nitride

electrodes were tested for up to 5000 CV cycles. The 1st, 2500th, and 5000th cycles are

shown in Fig. 5-1-6.

Fig. 5-1-6: Cyclic voltammograms in 0.5 M H2SO4: (a) Mo oxide at 1st, 2500th, and

5000th cycle at 100 mV/s; (b) Mo oxide-nitride at 1st, 2500th, and 5000th cycle at 100

mV/s.

For the Mo oxide electrode (Fig. 5-1-6a), the capacitance was reduced by 50% after

5000 cycles, indicating relatively poor cycle life and showing that the Mo oxide electrode

might not be suitable for the long-term cycling as expected for ECs. In contrast, the

46

Mo(O,N)x electrode exhibited a much more stable cycle life. As shown in Fig. 5-1-6b, the

CVs of Mo(O,N)x for the 1st, 2500th, and 5000th cycles were almost overlapping. The

capacitance of Mo oxide-nitride remained constant beyond 5000 cycles, even showing a

slight (6%) increase, probably due to further activation of Mo oxide-nitride in the

electrolyte. The heat treatment process has not only induced the desired

pseudocapacitive properties but also promoted higher stability of Mo oxide-nitride over

Mo oxide.

The cycle life test demonstrated by CV was based on charging/discharging at a

constant scan rate. However, in a real application, the current, rather than the scan rate,

is a fixed input. Therefore, in order to demonstrate the pseudocapacitive properties of

the Mo(O,N)x electrode under constant current charging/discharging analogous to real

conditions, galvanic charge-discharge was conducted at 1 mA/cm2 for up to 5000 cycles

as shown in Fig. 5-1-7. The charge-discharge curve showed capacitive behavior by

exhibiting close-to-linear charging/discharging curves. The same amount of charging

and discharging time within a cycle showed the high efficiency. In addition, the Mo(O,N)x

electrodes exhibited excellent reproducibility shown in Appendix B-4.

Fig. 5-1-7: galvanic charge-discharge curves of Mo oxide-nitride electrode at 1 mA/cm2

for 5000 cycle in 0.5 M H2SO4.

47

A small loss of capacitance was observed after 5000 cycles probably due to the slight

dissolution of the bulk oxides during cycling. However, the Mo(O,N)x electrode is still

considered chemically stable and can be expected to have a long cycle life in EC

applications. Moreover, cycle life performance can be further improved by using

advanced solid polymer electrolytes (41).

5.1.3.3 Cell Performance Based on Optimized Electrodes

In order to test the fabricated Mo(O,N)x electrodes in EC applications, they were

assembled in a two-electrode configuration to mimic a symmetric EC cell. CVs at scan

rates of 0.5, 1, and 2 V/s are shown in Fig. 5-1-8a, The current increased almost linearly

from 0.5 V/s to 2 V/s, which demonstrated the high rate and power capability of the

device in 0.5 M H2SO4. The CV at a rate of 2 V/s still exhibited nearly ideal rectangular

capacitive behavior, indicating that the surface redox reaction had fast kinetics.

However, the voltage window for the symmetric cell was limited to 0.7 V, which can be

improved by leveraging an asymmetric configuration as suggested by Li et al [82] and

Deng et al. [117]. An asymmetric device uses two different electrodes, where one acts

as anode and the other acts as cathode. Both electrodes are required to be stable in the

same electrolyte and use the same type of ion in their charge storage mechanisms.

Since each electrode has its own redox potential, an asymmetric EC enables a larger

cell voltage compared to its symmetric counterparts.

An asymmetrical device was assembled using a carbon electrode (graphite with Poly-

vinyl-alcohol (PVA) binder) as the positive electrode and the Mo oxide-nitride electrode

as the negative electrode. The electrochemical behaviors are shown in Fig. 5-1-8c. A

symmetric cell using two identical carbon electrodes were also assembled as a

reference shown in Fig. 5-1-8b. Compared to either 0.7 V for the Mo oxide-nitride

symmetric EC or 1.5 V for the carbon symmetric EC, the asymmetric EC exhibited the

widest voltage window of 2 V. It also showed the highest capacitance, based on the

capacitance measured at 0.5 V/s (i.e. 3.4 mF/cm2, 4.6 mF/cm2 and 7.2 mF/cm2 for Mo

oxide-nitride symmetric, carbon symmetric, Mo-carbon asymmetric cells). The

asymmetric EC showed an energy density of 2.8 times over carbon symmetric and 17

times over Mo oxide-nitride symmetric ECs based on equation (2). Since both

48

electrodes can take full advantage of their capable potential windows, the asymmetric

device optimizes the energy storage.

Fig. 5-1-8: cyclic voltammograms of a complete beaker EC in 0.5 M H2SO4 at various

scan rates made from (a) two identical Mo oxide-nitride electrodes (b) two identical

carbon electrodes (b) carbon as positive electrode and Mo oxide-nitride as negative

electrode.

49

5.1.4 Further Exploration of Mo Oxide-nitride Electrode in Neutral Electrolyte

It was known that protons activate the pseudocapacitive reactions on a Mo oxide-nitride

electrode. However, it is also interesting to explore the capability of the Mo oxide-nitride

electrode in neutral electrolytes for paring with other electrodes that are only stable in

neutral electrolytes to offer more asymmetrical cell configurations. Therefore, the Mo

oxide-nitride electrode was tested in 0.5 M Na2SO4. The CVs are shown in Fig. 5-1-9a.

Compared to the CV of the Ti substrate, the CV of Mo oxide-nitride in 0.5 M Na2SO4

exhibited a much higher capacitance, although smaller than that of the Mo oxide-nitride

electrode in 0.5 M H2SO4. Although the capacitance in 0.5 M Na2SO4 was lower, the

voltage window can be extended to ca. 1.1 V. This was probably due to a lack of

protons in the 0.5 M Na2SO4 that limited the hydrogen evolution reactions at low

potential. Overall, the Mo oxide-nitride in 0.5 M Na2SO4 had an energy density that was

4.7 times higher than that in 0.5 M H2SO4 based on equation (2).

Fig. 5-1-9: cyclic voltammograms of (a) Ti substrate and Mo oxide-nitride in 0.5 M

Na2SO4 at 100 mV/s; Mo oxide-nitride at 1st, 5000th, and 10000th cycle at 100 mV/s.

The cycle life performance of the Mo oxide-nitride electrode is shown in Fig. 5-1-9b.

Similar to the stability in acidic electrolyte (i.e. 0.5 M H2SO4), the capacitance of the

electrode remained constant after 10,000 cycles, suggesting an excellent stability and

long cycle life in a neutral electrolyte (i.e. 0.5 M Na2SO4). Therefore, it is clear that the

synthesized Mo oxide-nitride electrode can be used in neutral environment, making it

50

possible for more asymmetrical configurations with electrodes that only available in

neutral electrolyte.

5.1.5 Summary

A thin film Mo oxide-nitride pseudocapacitive electrode was synthesized by

electrodeposition of Mo oxide on Ti followed by low-temperature (400 oC) thermal

nitridation. Three nitridation gases, N2, forming gas and ammonia, were employed and

compared. Electrochemical analyses showed the N2-treated film had the best

pseudocapacitive behavior, outperforming the Mo oxide nitrided in forming gas and

ammonia. Surface analyses of these films showed about 20% to 50% conversion of Mo

oxide to nitrides, while the film nitrided in N2 had about 20% conversion. Cycle life and

stability of the resultant N2-treated Mo oxide-nitride were also much improved over Mo

oxide. A symmetric cell using the Mo oxide-nitride electrodes was demonstrated and

showed good performance at high scan rate. But the voltage window was limited at 0.7

V, which was improved by leveraging an asymmetric configuration using a carbon

electrode as the positive electrode. Further exploration showed the Mo oxide-nitride

electrode was stable in neutral electrolyte.

51

5.2 Tungsten Oxide

As discussed in section 5.1.3.3, Mo oxide-nitride electrode has a limited voltage window.

Thus a carbon electrode was added, forming an asymmetric EC, to address this

problem. However, the capacitance of the carbon electrode is limited by the double

layer capacitance. To further increase the capacitance, it would be interesting to have

another pseudocapacitive electrode to pair with Mo oxide-nitride electrode. WO3 is a

promising candidate due to its known pseudocapacitive behaviors. Moreover,

electrodeposition methods to deposit WO3 on metallic substrates were well studied [68-

71]. The focus of this section is not on the deposition parameters as the procedure from

the literature was followed (see Appendix C-1). Rather, the material and electrochemical

characterizations of the deposited tungsten oxide were performed to identify the

suitability for a tungsten oxide asymmetric pseudocapacitor.

5.2.1 Material Characterizations of as-deposit Tungsten Oxide

The as-deposit tungsten oxide electrode was first examined by ESEM. Images were

taken at different magnifications, as shown in Fig. 5-2-1 a-b. Similar to that the of as-

deposit Mo oxide electrode, the “dry mud” like surface morphology suggested

successful coating. The rough surface may increase the specific surface area for redox

reactions, resulting in increased capacitance. The coating thickness was estimated to

be about 100nm.

Fig. 5-2-1: ESEM micrographs of W oxide electrodes (a) low magnification, (b) high

magnification

52

The surface composition of the as-deposit tungsten electrode was examined by XPS.

The XPS spectra are shown in Fig. 5-2-2. A detailed analysis showed that the two major

peaks were at binding energies of 37.12 eV and 34.94 eV, corresponding to W6+ 4f5

and W6+ 4f7 [118], respectively. It clearly showed that the as-deposit tungsten electrode

was WO3. There were also two minor peaks co-existing in the XPS spectra, which were

identified as W5+ species. Further analysis showed that the atomic ratio between W6+

and W5+ was 92.5%: 7.5%. Therefore, WO3 was the dominant compound on the

electrode surface.

Fig. 5-2-2: High resolution XPS spectra for W 4f on the surface of W oxide electrode.

In addition, the XRD pattern indicated an amorphous surface structure, probably due to

hydration (see Appendix C-2). Based on the XPS and XRD results, the

electrodeposition process can be expressed in the following reactions (12 and 13a,b):

2WO42- + 2e- + 6H+

W2O5 + 3H2O (Rx-12)

W2O5 – 2e- + (2n+1)H2O 2 WO3 .nH2O + 2H+ (Rx-13a)

H2O2 + 2e- + 2H+ 2H2O (Rx-13b)

Since hydrogen peroxide was used during the electrodeposition, the homogeneity of the

coating was affected, causing variation of the capacitance. (see Appendix C-3)

53

5.2.2 Electrochemical Characterizations of as-deposit WO3 Electrode

The CVs of the as-deposit WO3 electrode in 0.5 M H2SO4 are shown in Fig. 5-2-3. For

comparison, the CV of Ti substrate was superimposed, which showed a very low current

density attributed to double layer capacitance. Since various voltage windows of WO3

electrode were reported in the literature [61, 62, 65], an incremental CV scan was

employed to identify the best voltage window. As shown in the figure, no

pseudocapacitive reactions were observed beyond 0.4 V. However, as the voltage

decreased, the pseudocapacitive current increased significantly until hydrogen evolution

at around -0.4V. This suggested the best voltage window for as-deposit WO3 electrode

was -0.4 to 0.4 V vs. Ag/AgCl. An average capacitance was measured to be 12.3

mF/cm2 in 0.5 M H2SO4.

Fig. 5-2-3: Cyclic voltammograms of as-deposit WO3 vs. Ti substrate in 0.5 M H2SO4 at

100 mV/s.

The cycle life of the electrode was examined by charging/discharging up to 15,000

cycles, as shown in Fig. 5-2-4. From 1st to 5,000th and to 15,000th cycles, the CVs

overlapped, implying a high stability of the as-deposit WO3 electrode in acidic electrolyte.

The mirror-imaged CV indicated highly reversible proton-associated pseudocapacitive

54

reactions in 0.5 M H2SO4. The cycling test demonstrated the WO3 electrodes

synthesized by this electrodeposition method were well suitable in EC application for

long term cycling.

Fig. 5-2-4: Cyclic voltammograms of as-deposit WO3 at 1st, 5000th and 15000th cycles in

0.5 M H2SO4 at 100 mV/s.

In order to test the electrode performance at high discharging rate for high power

applications, the scan rate was increased 10 times from 0.1 V/s to 1 V/s. The resultant

CV is showin in Fig. 5-2-5. Compare to that at 0.1 V/s, the CV at 1 V/s maintained a

reversible and rectangular shape, indicating fast kinetics of reactions of WO3.

Nonetheless, there was a small capacitance drop from 12.3 mF/cm2 to around 11.3

mF/cm2, which is acceptable as the ionic movement is limited at high scan rates.

55

Fig. 5-2-5: Cyclic voltammograms of as-deposit WO3 in 0.5 M H2SO4 at 100 mV/s and 1

V/s.

5.2.3 Symmetric and Asymmetric EC Devices

A symmetric EC using WO3 electrodes was assembled and the CV is shown in Fig. 5-2-

6. The symmetric WO3 EC exhibited a very narrow voltage window around 0.3V and a

low capacitance due to the limitations of WO3. In the symmetric EC using two Mo(O,N)x

electrodes (shown in Fig. 5-2-6 also in Fig. 5-1-8 a), both capacitance and voltage

window were wider than that of WO3 symmetric EC. Yet the performance was not

satisfactory. When using WO3 as the negative electrode and Mo oxide-nitride as the

positive electrode, the voltage window was further extended to around 0.8V with the

highest capacitance (see Fig. 5-2-6). The asymmetrical configuration optimized the

performance of each electrode. However, there was a significant overlap in the voltage

window between WO3 and Mo(O,N)x electrodes, so that the improvement was not

significant.

56

Fig. 5-2-6: Cyclic voltammograms of beaker ECs with various configurations in 0.5 M

H2SO4 at 100 mV/s.

5.2.4 Further exploration of WO3 electrodes

To further explore the capability of WO3 electrode in different environment, the WO3

electrodes were tested in neutral electrolyte (i.e 0.5M Na2SO4). The CVs are shown in

Fig. 5-2-7. Fig. 5-2-7a overlays the CVs of a WO3 electrode tested in 0.5 M H2SO4 and

Na2SO4, in which the CV in neutral electrolyte shifted to a more negative voltage. No

pseudocapacitive reaction was observed when the voltage exceeded 0.2 V, and the

hydrogen evolution reaction was not observed until around -0.7 V, indicating a negative

shift of voltage window in a neutral environment. This is expected due to the

thermodynamic and overpotential in neutral media.

The cycle life behavior was poor in 0.5 M Na2SO4 as depicted in Fig. 5-2-7b. The

capacitance significantly decreased after 5,000 cycles, suggesting severe electrode

dissolution. It is expected due to the WO3 is unstable in neutral environment based on

the E-pH diagram shown in Fig. 2-9. The results showed the as-deposit WO3 was not

suitable in a neutral electrolyte such as 0.5 M Na2SO4.

57

Fig. 5-2-7: Cyclic voltammograms at 100mV/s of as-deposit WO3 in (a) 0.5 M Na2SO4

and 0.5 M H2SO4 (b) 0.5 M Na2SO4 for 5000 cycles.

5.2.5 Summary

A thin film WO3 pseudocapacitive electrode was synthesized by electrodeposition.

Electrochemical analyses showed the as-deposit WO3 electrode had symmetric and

reversible CV profile, which is suitable for EC application. Surface analyses of these

films showed the surface composed of mainly WO3. Electrochemical characterizations

also showed a long cycle life and high rate capability in acidic electrolyte. An

asymmetric device was demonstrated using Mo(O,N)x as positive electrode and WO3 as

negative electrode, and results showed improvements in voltage window and

capacitance of the asymmetric EC over that of the symmetric cells. Further exploration

showed the produced WO3 electrode is not stable in neutral electrolyte.

58

5.3 Vanadium oxide

While Mo(O,N)x and WO3 have exhibited some promising characteristics as

pseudocapacitive electrodes, their active voltage windows are relatively narrow and the

redox reactions are not active or stable in neutral electrolytes. V oxides seem to be able

to complement these shortcomings. Since electrodeposition is a low cost method and

was used in previous study, initially it was planned to use electrodeposition to deposit V

oxide. While developing the electrodeposition parameters, a novel electroless

deposition was discovered and became the focus of the study. Therefore, an electroless

method to produce vanadium based electrodes for pseudocapacitors is presented in this

section.

5.3.1 Optimization of the Electroless Deposition

Since the electroless deposition of V oxide has never been investigated, deposition

parameters were tuned and optimized to yield the best pseudocapacitive behaviors for

ECs. The resultant electrodes were characterized using CV, ESEM and XRD to

optimize the method for synthesis.

5.3.1.1 Effects of Deposition Conditions: Concentration and Deposition Time

The appearance and the capacitance of the electrolessly deposited vanadium oxides

were found closely related to the concentration of VOSO4 and the deposition time. Thus,

they were used in process optimization, in which 3 concentrations (i.e. 0.25, 0.5 and

0.75 M) and 3 deposition times (i.e. 3, 5 and 7 days) were employed. After 3 days of

electroless deposition, a greenish coating appeared on the Ti substrates. From day 3 to

day 5, the coating became visibly thicker, especially in the 0.25 M and 0.5 M plating

solutions. However, inhomogeneities and island-like structures were observed after

further soaking; in the 0.75 M VOSO4 solution the coating thickness even seemed to

decrease (Fig. 5-3-1).

59

Fig. 5-3-1: Homogeneities of the electroless deposited vanadium electrodes in various

deposition concentration and time.

In addition to their visual differences, the electrochemical behaviors of the electrodes

should also show variations. The area-specific capacitances of the samples (measured

by CV) as a function of deposition time are shown in Fig. 5-3-2a. A significant increase

in specific capacitance was observed from day 3 to day 5, with the capacitance

approaching a plateau after day 5. Based on the coating coverage and the capacitance,

the 5 days of electroless deposition appeared to be the best condition, and thus the

following discussions are based on the 5-day samples.

The CVs of the vanadium oxide electrodes (after 5-day deposition in 0.25 M, 0.5 M, and

0.75 M VOSO4 plating solutions) in 1 M LiCl electrolyte are shown in Fig. 5-3-2b. The

CV of a bare Ti substrate was overlaid as a baseline. The Ti substrate had a small

capacitance of 0.03 mF/cm2. In contrast, the vanadium oxide electrodes exhibited much

higher area-specific capacitance of 60.3 mF/cm2, 67 mF/cm2, and 41.9 mF/cm2 for the

0.25M-5days, 0.5M-5days, and 0.75M-5days electrodes, respectively. Although the

0.25M-5days electrode showed a high capacitance, its CV profile was distorted and far

from ideal when compared to the 0.5M-5days electrode which exhibited an ideal

60

rectangular CV in the voltage window from 0 V to 0.6 V. On the other hand, the 0.75M-

5days electrode also showed good CV profile, but its capacitance was low.

Fig. 5-3-2: (a) Capacitance vs. deposition time in 0.25 M, 0.5 M, and 0.75 M VOSO4

baths for vanadium oxide electrodes, measured at 5 mV/s; (b) CVs of 0.25M-5days,

0.5M-5days, 0.75M-5days, and Ti electrodes in 1 M LiCl at 5 mV/s.

To better understand the factors that contribute to the discrepancy in capacitance, the

surface morphology of the electrode surfaces was investigated by ESEM. Fig. 5-3-3

shows the surface morphologies of the vanadium oxide electrodes produced by 5-day

deposition in the three concentrations. All electrodes exhibited a highly porous surface

of mixed meso- and macro-pores, similar as reported by Hu et al. using

electrodeposition [81, 82, 84]. Figs. 5-3-3 a-c show the ESEM top view of the 0.25M-

5days, 0.5M-5days, and 0.75M-5days electrodes. The 0.5M-5days electrode had the

finest microstructure and porosity in Fig. 5-3-3b, which corresponded to a high area-

specific capacitance. The side views of these electrodes are provided in Figs. 5-3-3 d-f:

All electrodes showed a dense layer interfacing between the Ti substrate and the

porous vanadium oxide layers. The 0.5M-5days electrode had the highest coverage and

the most intimate contact between the porous layer and the compact layer. In contrast,

the 0.25M-5days electrode in Fig 5-3-3d showed a much looser bonding between the

coating and the Ti substrate, and Fig 5-3-3f revealed low oxide coverage on the 0.75M-

5days electrode. These observations supported the results in Fig. 5-3-2 and suggested

that a 5-day deposition in 0.5 M VOSO4 (i.e. 0.5M-5days) yielded the best morphology

61

to achieve high capacitance. The reproducibility was good using the 0.5M-5days

deposition method, and the details are given in appendix D-1.

Fig. 5-3-3: Surface morphologies of the electrodes in Fig. 5-3-2(b); (a)-(c) top view and

(d)-(f) side view of 0.25M-5days, 0.5M-5days, and 0.75M-5days electrodes.

5.3.2 Characterizations on Optimized Electrodes

An optimized electroless deposition of V oxide electrode in 0.5 M VOSO4 for 5 days

(0.5M-5days) was established in section 5.3.1. In this section, detailed material

characterizations were performed on the electrode synthesized by the optimized method.

5.3.2.1 Effects of Heat Treatment

The as-deposit vanadium oxide electrode (0.5M-5days) demonstrated very promising

electrochemical performance for pseudocapacitors application. Additional heat

treatment was performed to see if any further improvement could be achieved. The

0.5M-5days electrodes were heat treated in argon or air at various temperatures for 5

hours. The XRD pattern and corresponding CVs are shown in Fig. 5-3-4. In Fig. 5-3-4a,

62

the XRD pattern showed the structure of the as-deposit vanadium oxide electrode

appeared to be amorphous, likely due to the high level of hydration. The electrode was

then heat treated in an argon environment at a temperature of 120oC, slightly above the

boiling point of water. The XRD pattern and the CV are shown in Fig. 5-3-4b, the heated

electrode still exhibited an amorphous structure with a rectangular CV that was similar

to that of the as-deposit V oxides electrode, suggesting no change on the surface

chemistry and crystal structure. Subsequently the electrode was heated up to around

350oC in argon. Interestingly, the XRD pattern (shown in Fig. 5-3-4c) changed

dramatically to a pattern which matched no existing vanadium oxides in the database,

likely due to new phase of mixed V oxides formation. In addition, the CV of the electrode

treated 350oC in argon was not identical to any known electrode in the literature.

Nonetheless, the CV deteriorated from rectangular shape to a non-symmetrical one,

suggesting the electrode was not ideal in EC application. Lastly, the electrode was

heated in air up to 350oC. As shown in Fig. 5-3-4d, the XRD pattern matched the known

structure of V2O5 (PDF # 01-076-1803), clearly indicated that the amorphous vanadium

oxide had transformed into orthorhombic V2O5 when heated up to 350oC in air.

Moreover, the CV in Fig. 5-3-4d was very similar to that reported in the literature for

V2O5 [119], confirming the surface was crystalized V2O5.

In Fig. 5-3-4, with the increase of crystallinity, the electrodes became more “battery-like”.

It can be seen in the figures that the crystalline V2O5 and the one treated in argon at

350oC had large charge storage, represented by the significant current peaks. However,

the current peaks revealed that the heat treated electrodes were showing slow redox

kinetics and were battery-like rather than capacitor-like. In fact, crystalline V2O5 were

investigated as Li-ion battery electrode materials [120]. The CV profile in Fig. 5-3-4d

was very similar to that reported in the literature[119]. Overall, it was suggested that

amorphous and nanocrystalline V2O5 , rather than crystalline V2O5, exhibit better

pseudocapacitive behavior [76], which is the case in this work.

63

Fig. 5-3-4: XRD patterns and cyclic voltammograms for 0.5M-5days vanadium

electrodes (a) as-deposit (b) heat treated in argon at 120oC (c) heat treated in argon at

350oC (d) heat treated in air at 350oC.

5.3.2.2 Surface Characterizations

The surface composition of the electrodes was obtained by XPS analyses. Since the

binding energies of V2p and O1s are very close, the high resolution spectra of V and O

for the 0.5M-5days electrode are shown together in Fig. 5-3-5.

64

Fig. 5-3-5: High resolution XPS spectra for V2p and O1s of 0.5M-5days electrode.

The V2p3/2 spectrum showed two oxidation states with binding energies at 515.9 eV for

V4+ and 517.4 eV for V5+ [118, 121, 122]. The O1s spectrum was deconvoluted into 3

peaks at 531.7 eV, 530.4 eV and 529.8 eV, which corresponded to H-O-H, V-OH, and

V-O-V, respectively [84, 118]. Based on the V2p and O1s spectra, the electrode

material was likely mixed hydrous V2O5 and VO2 similar to the materials obtained from

electrodeposition by Hu et al., who suggested the composition to be VOx.yH2O [81, 82,

84]. Based on the experimental observations and XPS analysis, it was suggested that

the VO2+ ions reacted with OH- to precipitate VO(OH)2 according to a possible reaction

(14) [123]:

2NaOH + VOSO4 VO(OH)2 + Na2SO4 (Rx-14)

The precipitations initially dispersed in the solution and then condensed on the Ti

substrate. Due to the electroless reactions were exposed to air, vanadium was partially

oxidized from IV to V as observed in XPS spectra. Hu reported a V5+ content of 89 mol%,

which is comparable to the V5+ concentration of 83 mol% observed in this work.

65

5.3.2.3 Electrochemical Characterizations

The voltage window of the electrode has to be determined to avoid damage such as

overcharging to the electrodes. The voltage window of V oxides electrode was

determined by incremental scan shown in Fig. 5-3-6. An as-deposit V oxides electrode

was incrementally scanned up to 0.8 V in 1 M LiCl. When the voltage exceeded 0.6 V,

a permanent damaged was made to the electrode, leading to a change of CV shape

from a rectangular to an ellipse, especially when the voltage was further increased to

0.8 V. High voltage (i.e. 0.7 V and 0.8 V) could permanently change the internal

structure of the as-deposit V oxides, which led to the deterioration of the electrochemical

behaviors. Therefore, 0 to 0.6 V was selected as the voltage window for the V oxides

electrode in this work.

Fig. 5-3-6: Cyclic voltammograms of as-deposit V oxides 1 M LiCl at 10 mV/s.

In order to test its cycle life, the 0.5M-5days electrode was subjected to 3000 CV cycles

in 1 M LiCl solution, as shown in Fig. 5-3-7a. Vanadium oxide was reported to have poor

structural stability during charging-discharging due to material pulverization and

dissolution in the electrolyte [78, 124, 125]. Instead of the severe degradation reported,

the electrolessly deposited electrode showed only around 10% decrease in capacitance

66

over 3000 cycles. Moreover, the CV profile maintained its rectangular shape, suggesting

that the electrolessly deposited vanadium oxide is structurally stable.

The power performance of the electrode was tested through CV at higher scan rates

(e.g. 10 mV/s and 100 mV/s) as shown in Fig. 5-3-7b. At 10 mV/s, the CV maintained a

rectangular charging and discharging profile. The CV, however, deteriorated as the scan

rate further increased to 100 mV/s. The elliptical CV profile suggested the kinetics of the

electrode was not fast enough to be capable at fast charging/discharging. Compare to

Mo oxide-nitride and tungsten oxide electrode, the vanadium mix oxides electrode has

higher resistance and lower rate capability. Nonetheless, this could be improved by

additives to increase the electrode conductivity and the bonding between electrode

material and the substrate.

Fig. 5-3-7: Cyclic voltammograms of 0.5M-5days vanadium oxide electrode in 1M LiCl

(a) at 1st, 1000th, 2000th and 3000th cycle 5 mV/s (b) at 10 mV/s and 100 mV/s.

To further examine its suitability for EC applications, a two-electrode cell was

assembled using two 0.5M-5days electrodes separated by a thin filter paper soaked

with 1 M LiCl. The CV of this device is shown in Fig. 5-3-8 and displayed a rectangular

profile. The cell voltage in Fig. 5-3-8 was limited to 0.6 V, which could be improved

further by leveraging an asymmetric configuration as suggested by Deng et al. [117].

67

Fig. 5-3-8: Cyclic voltammogram of an EC cell made from two identical 0.5M-5days

vanadium oxide electrodes in 1 M LiCl at 5 mV/s.

5.3.3 Summary

A simple and inexpensive electroless deposition method was leveraged to produce

vanadium oxide electrodes for EC applications. An optimized processing condition was

identified. Material characterizations revealed a meso- and macro-porous structure on

the electrode surface. Surface and structural analyses showed that the composition

involved both VO2 and V2O5 and that the mixed vanadium oxide was in the amorphous

state. Heat treatments on the as-deposit electrode further indicated the

pseudocapacitive behaviors on amorphous and hydrous vanadium oxides. Although the

power capability was limited by the intrinsic resistance of vanadium oxides, rectangular

CVs and good cycle life of the electrodes suggest that vanadium oxides produced by

the described electroless deposition method are promising as high performance EC

electrode materials.

68

Chapter 6: Conclusions

The objective of this work is to explore low cost pseudocapacitive transition metal

compounds and develop facile fabrication methods for EC application. The materials

should possess capacitive behaviors, large capacitance, fast kinetics, large voltage

window and long cycle life. Three compounds (Mo(O,N)x, WO3 and VOx.yH2O )were

successfully developed using simple, environmental benign and inexpensive methods.

The electrodes were characterized and summarized in table 6-1. It is concluded that

each electrode has its own advantages and disadvantages.

1. Mo(O,N)x electrode has strong bonding to the substrate, which leads to fast

kinetics and excellent cycle life. It is applicable in both acidic and neutral

electrolytes. However, the capacitance is relatively low compare to that of WO3

and VOx.yH2O electrodes. In addition, the CV profile is not perfectly ideal.

Nonetheless, it is the first time that nitridation of Mo oxides using N2 is reported,

offering a low-cost and promising route of producing Mo oxide-nitride electrode.

2. WO3 electrode fabricated by electrodeposition is very stable in acidic electrolyte,

as well as capable in high rate charging/discharging. The drawbacks are the non-

rectangular CV, relatively small capacitance and narrow voltage window.

However, by using an asymmetric configuration, WO3 can serve as a

complementary electrode to Mo(O,N)x that improves the device performance.

3. VOx.yH2O electrode was synthesized by a low cost electroless route for the first

time in this project. Its highly porous surface results in large capacitance.

Moreover, the CV profile is ideally rectangular and reversible. The disadvantages

include the low conductivity, and limited electrolytes (i.e. neutral electrolyte).

The three transition metal compounds all showed, in some way, promising

electrochemical behaviors as electrode materials for EC application. However, they all

have limited voltage window. Through the development of asymmetric cell, the voltage

window can be expanded. This was demonstrated in this work (see table 6-2).

69

Table 6-1: Summary of Mo(O,N)x, WO3 and VOx.yH2O electrodes

Name Mo oxide-nitride W oxide V mixed oxides

Chemistry MoO2 + MoO3 +

Mo2N WO3 + WO2 V2O5 + VO2

Method Electrodeposition +

heat treatment in N2

Electrodeposition Electroless

deposition

Crystal structure crystalline amorphous amorphous

Capability in various

electrolytes Acidic/neutral Acidic Neutral

Capacitance Medium Medium High

Kinetics or rate Fast Fast Slow

Voltage window (V) 0.6 0.8 0.6

Cycle life Excellent Excellent Good

Table 6-2: Summary of ECs based on the fabricated electrodes

Configuration Voltage Window (V) Electrolyte

Symmetric cell

Mo (ON)x 0.7 0.5 M H2SO4

W oxide 0.4 0.5 M H2SO4

V mixed oxide 0.6 1 M LiCl

Asymmetric cell (-)Mo(O,N)x // (+) carbon 2.0 0.5 M H2SO4

(+)Mo(O,N)x // (-) W oxide 0.8 0.5 M H2SO4

70

Chapter 7: Future Work

The three types of electrodes showed promising electrochemical behaviors for EC

applications. However, each of them has limitations. Therefore, several approaches are

proposed to improve these electrodes and further explore their potentials.

1. The CV profile of the Mo(O,N)x electrode is governed by the ratio between MoO2 ,

MoO3 and Mo2N. The current ratio leads to a close-to-ideal capacitive behavior

that could be further improved by tuning the heat treatment process using more

precise heat treatment time and temperature.

2. The capacitance of the Mo(O,N)x electrode is relatively low due to the limited

surface roughness. The electrodeposition can be tailored to increased surface

roughness by changing the deposition parameters and substrate pre-treatment.

3. WO3 needs more analyses and could be co-deposited with Mo(O,N)x. It would be

interesting to try sequential or layer by layer deposition to integrate WO3 and

Mo(O,N)x for synergistic effects.

4. WN is a known pseudocapacitive electrode material. It would be interesting to

use stronger nitridation gases such as ammonia to convert as-deposited WO3

into WN and extend to neutral and alkaline electrolytes.

5. The electroless deposition of V oxides discovered in this study opens a new route

of such synthesizing V oxides based electrodes for EC applications. Due to the

precipitation process in the deposition, the electroless deposition can be

potentially applied to any substrates, including carbon substrates and conductive

polymers.

6. The VOx.yH2O electrode showed a low power capability due to the high electrical

resistance and weak bonding to the substrates. This could be addressed by

adding additives such as metal dopants to increase the conductivity, or bonding

enhancers such as dopamine to strengthen the coating/substrate interface.

7. The performance of the ECs assembled by developed electrodes can be further

improved using more advanced solid polymer electrolytes developed in our lab.

71

References

[1] P. Simon, Y. Gogotsi, Nature Materials, 7 (2008) 845-854. [2] A. Burke, Journal of Power Sources, 91 (2000) 37-50. [3] Y. Zhang, H. Feng, X. Wu, L. Wang, A. Zhang, T. Xia, H. Dong, X. Li, L. Zhang, International Journal of Hydrogen Energy, 34 (2009) 4889-4899. [4] J.R. Miller, A.F. Burke, Electrochemical Society Interface, 17 (2008) 53-57. [5] G. Wang, L. Zhang, J. Zhang, Chemical Society Reviews, 41 (2012) 797-828. [6] R. Kötz, M. Carlen, Electrochimica Acta, 45 (2000) 2483-2498. [7] D. Cericola, R. Kötz, Electrochimica Acta, 72 (2012) 1-17. [8] P.J. Hall, E.J. Bain, Energy Policy, 36 (2008) 4352-4355. [9] A.G. Pandolfo, A.F. Hollenkamp, Journal of Power Sources, 157 (2006) 11-27. [10] L.L. Zhang, X.S. Zhao, Chemical Society Reviews, 38 (2009) 2520-2531. [11] H. Pan, J. Li, Y. Feng, Nanoscale Research Letters, 5 (2010) 654-668. [12] J.P. Zheng, J. Huang, T.R. Jow, Journal of The Electrochemical Society, 144 (1997) 2026-2031. [13] Y.M. Vol'fkovich, T.M. Serdyuk, Russian Journal of Electrochemistry, 38 (2002) 935-959. [14] P. Simon, Y. Gogotsi, Accounts of Chemical Research, 46 (2012) 1094-1103. [15] B.E. Conway, V. Birss, J. Wojtowicz, Journal of Power Sources, 66 (1997) 1-14. [16] A. Balducci, W.A. Henderson, M. Mastragostino, S. Passerini, P. Simon, F. Soavi, Electrochimica Acta, 50 2233-2237. [17] B. Muthulakshmi, D. Kalpana, S. Pitchumani, N.G. Renganathan, Journal of Power Sources, 158 1533-1537. [18] K.R. Prasad, N. Miura, Electrochemical and Solid-State Letters, 7 (2004) A425-A428. [19] B.E. Conway, Electrochemical Supercapacitors: Scientific Fundamentals and technological Applications, 1 ed., Springer, New York, 1999. [20] I.-H. Kim, K.-B. Kim, Journal of The Electrochemical Society, 153 (2006) A383-A389. [21] T.C. Liu, W.G. Pell, B.E. Conway, S.L. Roberson, Journal of The Electrochemical Society, 145 (1998) 1882-1888. [22] J.P. Zheng, P.J. Cygan, T.R. Jow, Journal of The Electrochemical Society, 142 (1995) 2699-2703. [23] C.C. Hu, Y.H. Huang, Journal of The Electrochemical Society, 146 (1999) 2465-2471. [24] S. Devaraj, N. Munichandraiah, Electrochemical and Solid-State Letters, 8 (2005) A373-A377. [25] S.C. Pang, M.A. Anderson, T.W. Chapman, Journal of The Electrochemical Society, 147 (2000) 444-450. [26] M. Toupin, T. Brousse, D. Bélanger, Chemistry of Materials, 16 (2004) 3184-3190. [27] J.-K. Chang, M.-T. Lee, W.-T. Tsai, Journal of Power Sources, 166 (2007) 590-594. [28] T. Xue, C.-L. Xu, D.-D. Zhao, X.-H. Li, H.-L. Li, Journal of Power Sources, 164 (2007) 953-958. [29] Y. Yang, C. Huang, Journal of Solid State Electrochemistry, 14 (2010) 1293-1301. [30] S. Devaraj, N. Munichandraiah, Journal of The Electrochemical Society, 154 (2007) A80-A88. [31] M. Lai, D.J. Riley, Journal of Colloid and Interface Science, 323 (2008) 203-212.

72

[32] H.-Q. Wang, G.-f. Yang, Q.-Y. Li, X.-X. Zhong, F.-P. Wang, Z.-S. Li, Y.-h. Li, New Journal of Chemistry, 35 (2011) 469-475. [33] H. Xia, J. Feng, H. Wang, M.O. Lai, L. Lu, Journal of Power Sources, 195 (2010) 4410-4413. [34] S.W. Lee, J. Kim, S. Chen, P.T. Hammond, Y. Shao-Horn, ACS Nano, 4 (2010) 3889-3896. [35] Z. Fan, J. Yan, T. Wei, L. Zhi, G. Ning, T. Li, F. Wei, Advanced Functional Materials, 21 (2011) 2366-2375. [36] M. Pourbaix, Atlas of electrochemical equilibria in aqueous solutions, Pergamon Press, Oxford, 1966. [37] W. Baoxing, D. Shaojun, Journal of Electroanalytical Chemistry, 379 (1994) 207-214. [38] S. Liu, Q. Zhang, E. Wang, S. Dong, Electrochemistry Communications, 1 (1999) 365-369. [39] H. Farsi, F. Gobal, H. Raissi, S. Moghiminia, Journal of Solid State Electrochemistry, 14 (2010) 643-650. [40] V.S. Saji, C.-W. Lee, ChemSusChem, 5 (2012) 1146-1161. [41] Y. Saito, C. Kaito, T. Naiki, Journal of Crystal Growth, 79 (1986) 436-442. [42] D. Mutschall, K. Holzner, E. Obermeier, Sensors and Actuators B: Chemical, 36 (1996) 320-324. [43] M. Golozar, K. Chien, K. Lian, T. Coyle, Journal of Thermal Spray Technology, 22 (2013) 710-722. [44] T. Ivanova, A. Szekeres, M. Gartner, D. Gogova, K.A. Gesheva, Electrochimica Acta, 46 (2001) 2215-2219. [45] W. Dong, B. Dunn, Journal of Non-Crystalline Solids, 225 (1998) 135-140. *46+ B. Więcek, U. Twardoch, Journal of Physics and Chemistry of Solids, 65 (2004) 263-268. [47] D. Guay, G. Tourillon, G. Laperriere, D. Belanger, The Journal of Physical Chemistry, 96 (1992) 7718-7724. [48] H. Farsi, F. Gobal, H. Raissi, S. Moghiminia, Journal of Solid State Electrochemistry, 14 (2010) 681-686. [49] A. Guerfi, R.W. Paynter, L.H. Dao, Journal of The Electrochemical Society, 142 (1995) 3457-3464. [50] P. Pande, P.G. Rasmussen, L.T. Thompson, Journal of Power Sources, 207 (2012) 212-215. [51] C. Chen, D. Zhao, D. Xu, X. Wang, Materials Chemistry and Physics, 95 (2006) 84-88. [52] X.-L. Li, Y. Xing, H. Wang, H.-L. Wang, W.-D. Wang, X.-Y. Chen, Transactions of Nonferrous Metals Society of China, 19 (2009) 620-625. [53] S.L. Roberson, D. Finello, R.F. Davis, Journal of Applied Electrochemistry, 29 (1999) 75-80. [54] V.P. Anitha, S. Major, D. Chandrashekharam, M. Bhatnagar, Surface and Coatings Technology, 79 (1996) 50-54. [55] C.Z. Deng, R.A.J. Pynenburg, K.C. Tsai, Journal of The Electrochemical Society, 145 (1998) L61-L63. [56] C.H. Jaggers, J.N. Michaels, A.M. Stacy, Chemistry of Materials, 2 (1990) 150-157. [57] G.R. Bamwenda, H. Arakawa, Applied Catalysis A: General, 210 (2001) 181-191. [58] P.K. Shen, K.Y. Chen, A.C.C. Tseung, Journal of The Electrochemical Society, 141 (1994) 1758-1762. [59] A. Azens, A. Hjelm, D. Le Bellac, C.G. Granqvist, J. Barczynska, E. Pentjuss, J. Gabrusenoks, J.M. Wills, Solid State Ionics, 86–88, Part 2 (1996) 943-948. [60] S. Yoon, E. Kang, J.K. Kim, C.W. Lee, J. Lee, Chemical Communications, 47 (2011) 1021-1023.

73

[61] D. Wang, J. Li, X. Cao, G. Pang, S. Feng, Chemical Communications, 46 (2010) 7718-7720. [62] S.H. Park, Y.H. Kim, T.G. Lee, H.K. Shon, H.M. Park, J.Y. Song, Materials Research Bulletin, 47 (2012) 3612-3618. [63] M. Anik, K. Osseo-Asare, Journal of The Electrochemical Society, 149 (2002) B224-B233. [64] J.P. Randin, R. Viennet, Journal of The Electrochemical Society, 129 (1982) 2349-2354. [65] C. Jo, I. Hwang, J. Lee, C.W. Lee, S. Yoon, The Journal of Physical Chemistry C, 115 (2011) 11880-11886. [66] S. Li, M. Samy El-Shall, Nanostructured Materials, 12 (1999) 215-219. [67] X.P. Wang, B.Q. Yang, H.X. Zhang, P.X. Feng, Nanoscale Research Letters, 2 (2007) 405-409. [68] L. Kondrachova, B.P. Hahn, G. Vijayaraghavan, R.D. Williams, K.J. Stevenson, Langmuir, 22 (2006) 10490-10498. [69] S.H. Baeck, T. Jaramillo, G.D. Stucky, E.W. McFarland, Nano Letters, 2 (2002) 831-834. [70] B.-X. Zou, Y. Liang, X.-X. Liu, D. Diamond, K.-T. Lau, Journal of Power Sources, 196 (2011) 4842-4848. [71] M. Deepa, A.K. Srivastava, S.N. Sharma, Govind, S.M. Shivaprasad, Applied Surface Science, 254 (2008) 2342-2352. [72] J.W. Klaus, S.J. Ferro, S.M. George, Journal of The Electrochemical Society, 147 (2000) 1175-1181. [73] J.S. Becker, S. Suh, S. Wang, R.G. Gordon, Chemistry of Materials, 15 (2003) 2969-2976. [74] T. Nakajima, K. Watanabe, N. Watanabe, Journal of The Electrochemical Society, 134 (1987) 3175-3178. [75] A.R. Ko, S.-B. Han, Y.-W. Lee, K.-W. Park, Physical Chemistry Chemical Physics, 13 (2011) 12705-12707. [76] V. Augustyn, P. Simon, B. Dunn, Energy & Environmental Science, 7 (2014) 1597-1614. [77] W. Dong, D.R. Rolison, B. Dunna, Electrochemical and Solid-State Letters, 3 (2000) 457-459. [78] G. Wang, X. Lu, Y. Ling, T. Zhai, H. Wang, Y. Tong, Y. Li, ACS Nano, 6 (2012) 10296-10302. [79] S. Boukhalfa, K. Evanoff, G. Yushin, Energy & Environmental Science, 5 (2012) 6872-6879. [80] C.-H. Lai, C.-K. Lin, S.-W. Lee, H.-Y. Li, J.-K. Chang, M.-J. Deng, Journal of Alloys and Compounds, 536, Supplement 1 (2012) S428-S431. [81] C.-M. Huang, C.-C. Hu, K.-H. Chang, J.-M. Li, Y.-F. Li, Journal of The Electrochemical Society, 156 (2009) A667-A671. [82] J.-M. Li, K.-H. Chang, C.-C. Hu, Electrochemistry Communications, 12 (2010) 1800-1803. [83] I.-H. Kim, J.-H. Kim, B.-W. Cho, K.-B. Kim, Journal of The Electrochemical Society, 153 (2006) A1451-A1458. [84] C.-C. Hu, C.-M. Huang, K.-H. Chang, Journal of Power Sources, 185 (2008) 1594-1597. [85] G. Armstrong, J. Canales, A.R. Armstrong, P.G. Bruce, Journal of Power Sources, 178 (2008) 723-728. [86] K. Post, R.G. Robins, Electrochimica Acta, 21 (1976) 401-405. [87] Y. Wang, K. Takahashi, K.H. Lee, G.Z. Cao, Advanced Functional Materials, 16 (2006) 1133-1144. [88] K. West, B. Zachau-Christiansen, T. Jacobsen, S. Skaarup, Electrochimica Acta, 38 (1993) 1215-1220. [89] K. Salloux, F. Chaput, H.P. Wong, B. Dunn, M.W. Breiter, Journal of The Electrochemical Society, 142 (1995) L191-L192.

74

[90] Y. Wang, K. Lee, H. Shang, B. Wiley, Y. Xia, G. Cao, physica status solidi (a), 202 (2005) R79-R81. [91] M. Hibino, M. Ugaji, A. Kishimoto, T. Kudo, Solid State Ionics, 79 (1995) 239-244. [92] D. Choi, G.E. Blomgren, P.N. Kumta, Advanced Materials, 18 (2006) 1178-1182. [93] M.A. Roldan, V. López-Flores, M.D. Alcala, A. Ortega, C. Real, Journal of the European Ceramic Society, 30 (2010) 2099-2107. [94] J.C. Caicedo, G. Zambrano, W. Aperador, L. Escobar-Alarcon, E. Camps, Applied Surface Science, 258 (2011) 312-320. [95] X. Zhou, H. Chen, D. Shu, C. He, J. Nan, Journal of Physics and Chemistry of Solids, 70 (2009) 495-500. [96] R. Kapoor, S.T. Oyama, Journal of Solid State Chemistry, 99 (1992) 303-312. [97] C.M. Ghimbeu, E. Raymundo-Pinero, P. Fioux, F. Beguin, C. Vix-Guterl, Journal of Materials Chemistry, 21 (2011) 13268-13275. [98] C.M. Ghimbeu, F. Sima, R.V. Ostaci, G. Socol, I.N. Mihailescu, C. Vix-Guterl, Surface and Coatings Technology, 211 (2012) 158-162. [99] J.F. Rusling, S.L. Suib, Advanced Materials, 6 (1994) 922-930. [100] Y.-J. Ting, K. Lian, N. Kherani, ECS Transactions, 35 (2011) 133-139. [101] H. Gao, Y.-J. Ting, N.P. Kherani, K. Lian, Journal of Power Sources, 222 (2013) 301-304. [102] R.S. Patil, M.D. Uplane, P.S. Patil, Applied Surface Science, 252 (2006) 8050-8056. [103] G.-T. Kim, T.-K. Park, H. Chung, Y.-T. Kim, M.-H. Kwon, J.-G. Choi, Applied Surface Science, 152 (1999) 35-43. [104] T. Bécue, J.-M. Manoli, C. Potvin, G. Djéga-Mariadassou, M. Delamar, The Journal of Physical Chemistry B, 101 (1997) 6429-6435. [105] J.G. Choi, D. Choi, L.T. Thompson, Applied Surface Science, 108 (1997) 103-111. [106] D. McKay, J.J. Hargreaves, R. Howe, Catalysis Letters, 112 (2006) 109-113. [107] C. Shi, A.M. Zhu, X.F. Yang, C.T. Au, Applied Catalysis A: General, 293 (2005) 83-90. [108] C. Shi, A.M. Zhu, X.F. Yang, C.T. Au, Applied Catalysis A: General, 276 (2004) 223-230. [109] K.-H. Lee, Y.-W. Lee, A.R. Ko, G. Cao, K.-W. Park, Journal of the American Ceramic Society, 96 (2013) 37-39. [110] J.B. Claridge, A.P.E. York, A.J. Brungs, M.L.H. Green, Chemistry of Materials, 12 (1999) 132-142. [111] B.E. Conway, Electrochemical Supercapacitors: Scientific Fundamentals and technological Applications, in: Electrochemical Supercapacitors: Scientific Fundamentals and technological Applications, Springer, New York, 1999, pp. 277. [112] D. Golodnitsky, V. Yufit, M. Nathan, I. Shechtman, T. Ripenbein, E. Strauss, S. Menkin, E. Peled, Journal of Power Sources, 153 (2006) 281-287. [113] C.-C. Hu, M.-J. Liu, K.-H. Chang, Journal of Power Sources, 163 (2007) 1126-1131. [114] Z.B.Z. Wei, P. Grange, B. Delmon, Applied Surface Science, 135 (1998) 107-114. [115] Y. Li, Y. Fan, Y. Chen, Catalysis Letters, 82 (2002) 111-116. [116] J.D. Wu, C.Z. Wu, Z.M. Song, F.M. Li, Thin Solid Films, 311 (1997) 62-66. [117] L. Deng, G. Zhu, J. Wang, L. Kang, Z.-H. Liu, Z. Yang, Z. Wang, Journal of Power Sources, 196 (2011) 10782-10787. [118] C. D. Wagner, L. E. Davis, J. F. Moulder and G. E.Muilenberg, Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics Division, Perkin-Elmer Corp., Minnesota, 1979.

75

[119] K. Takahashi, S.J. Limmer, Y. Wang, G. Cao, The Journal of Physical Chemistry B, 108 (2004) 9795-9800. [120] A. Pan, J.-G. Zhang, Z. Nie, G. Cao, B.W. Arey, G. Li, S.-q. Liang, J. Liu, Journal of Materials Chemistry, 20 (2010) 9193-9199. [121] K.-H. Chang, C.-C. Hu, Acta Materialia, 55 (2007) 6192-6197. [122] J. Mendialdua, R. Casanova, Y. Barbaux, Journal of Electron Spectroscopy and Related Phenomena, 71 (1995) 249-261. [123] R. Minch, M. Es-Souni, CrystEngComm, 15 (2013) 6645-6647. [124] Y. Liu, M. Clark, Q. Zhang, D. Yu, D. Liu, J. Liu, G. Cao, Advanced Energy Materials, 1 (2011) 194-202. [125] D. Liu, Y. Liu, S.L. Candelaria, G. Cao, J. Liu, Y.-H. Jeong, Journal of Vacuum Science and Technology A, 30 (2012) A1231-A1236. [126] L.L. Zhang, R. Zhou, X.S. Zhao, Journal of Materials Chemistry, 20 (2010) 5983-5992.

76

Appendices:

Appendix A-1: List s of existing electrode materials for ECs

The electrodes materials were categorized into carbons, conductive polymers, metal

oxides and ruthenium dioxides and their capacitances were compared in Fig. A-1-1.

Carbons are electrode materials for EDLC, while conductive polymers, metal oxides and

ruthenium dioxides are pseudocapacitive electrode materials.

Fig. A-1-1 Capacitances of various electrode materials for ECs, adapted from [126]

77

Appendix B-1: Effects of Electrodeposition Parameters for Mo Oxide

Coating a thin layer of Mo oxide onto the Ti substrate is the first step to fabricate a Mo

based electrode in which a cyclic deposition method was applied. Deposition voltage

window, rate and time are important parameters. Fig. B-1-1 shows the effects of

changing the deposition voltage window and rate in the electrodeposition process.

Fig. B-1-1: Capacitance of Mo(O,N)x electrodes measured at 100 mV/s in 0.5M H2SO4

showing effects of deposition voltage and rate in electrodeposition.

When fixing the deposition time at 37min, it was suggested that the deposition voltage

window from -0.85V to 0V was better than -0.75V to 0V. It can be attributed to the extra

amount of charge transfer, represented by the addition area shown in the CV (see Fig.

B-1-2a). When fixing both deposition time and deposition voltage window, it can be seen

that slower deposition rate resulted in higher capacitance due to overall more charges

were accumulated (see Fig. B-1-2b).

78

Fig. B-1-2: (a)Cyclic voltammograms of electrodeposition profile at 0.01 V/s in Mo

plating bath, deposition window from 0 V to -0.75 V vs. 0 V to -0.85 V.(b) accumulated

charges through the electrodeposition process at 0.01, 0.05 and 0.3 V/s from 0V to -

0.85V for 37 min.

The voltage window from -0.85 V to 0 V and the scan rate of 0.01 V/s were selected.

The deposition time was then set as a variable. Fig. B-1-3 suggests the current

efficiency decreased linearly with time, resulting in a similar trend in the measured

capacitance of the Mo(O,N)x electrodes.

Fig. B-1-3: Charge transfer per cycle through a 0 V to -0.85 V electrodeposition at 0.01

V/s for various deposition time.

79

The capacitance of the Mo(O,N)x electrode (produced at electrodeposition from 0 V to -

0.85 V at 0.01 V/s then treated in N2 for 3 h) can be predicted in the following equation

(9):

C (uF/cm2) = -0.237 t2 (min) + 117.5 t + 4177 (Equ-9)

Where C is the capacitance in uF/cm2, t is the deposition time in min.

80

Appendix B-2: XPS Analysis on Mo(O,N)x Electrode on Mo3p Orbitals and N1s

Orbitals

In addition, the Mo 3p spectra were analyzed to supplement the results from the Mo 3d

spectra for Mo(O,N)x electrode treated in N2 (Fig. B-2-1). The detailed compositions of

Mo 3p and N 1s for both Mo oxide and Mo oxide-nitride are listed in Table B-2.

Fig. B-2-1: High resolution XPS spectra for Mo 3p (a) Mo oxide (b) Mo oxide-nitride

treated in N2.

Table B-2: Compositional analysis on Mo oxide and Mo oxide-nitride electrodes

Sample Species Peak

Identification

Binding Energy

(eV) At%

Mo oxide

Mo6+(MoO3) 3p 1/2 415.1

79.3 3p 3/2 397.7

Mo4+(MoO2) 3p 1/2 413.6

20.7 3p 3/2 396.2

Nitrided Mo oxide

Mo6+ 3p 1/2 415.1

55.0 3p 3/2 397.7

Mo4+ 3p 1/2 413.6

15.5 3p 3/2 396.2

Moδ+(Mo2N-

type

compound)

3p 1/2 412.6 17.0

3p 3/2 395.2

N 1s 397.5 12.5

81

Appendix B-3: SIMS Analysis on Mo(O,N)x Electrodes

The surfaces of the Mo oxide and Mo(O,N)x electrodes were examined by Secondary

Ion Mass Spectroscopy (SIMS) as complement to XPS and XRD analyses. The depth

profiles are shown in Fig. B-3-1. It was observed the nitrogen content increased at the

very outer layer of the electrode surface, which suggested the effectiveness of the

nitridation process in N2.

Fig. B-3-1: SIMS depth profile of (a) Mo oxide electrode (b) Mo(O,N)x electrode (c)

nitrogen species of both Mo oxide and Mo(O,N)x electrode (d) high resolution of nitrogen

species of both Mo oxide and Mo(O,N)x electrode.

82

Appendix B-4: Reproducibility of Mo(O,N)x Electrodes

Reproducibility should be taken into consideration when finalizing a production method.

Five electrodes deposited by the same method (i.e. electrodeposition from 0 V to -0.85

V at 0.01 V/s for 37 min then nitrided in N2 for 400oC) were tested and their CVs are

shown in Fig. B-4-1. It can be seen that the 5 CVs almost overlaps. The capacitances

calculated from the CVs were 7.4, 7.84, 8.93, 8.46 and 7.48 mF/cm2, which resulted in

an average value of 8.1 mF/cm2. The method was considered fairly stable as the

produced electrodes had similar CV profile and capacitance.

Fig. B-4-1: Cyclic Voltammograms of five Mo(O,N)x electrodes produced by the

optimized method.

83

Appendix C-1: Effects of Hydrogen Peroxide in Electrodeposition of WO3

Since contradiction exists in the literature, the effects of using H2O2 in the

electrodeposition were tested by adding 0.09M of H2O2. The electroplating bath without

H2O2 was used a reference. The coated electrodes are shown in Fig. C-1-1. The

electrode deposited in plating bath with hydrogen peroxide showed a “rainbow” color

distribution due to oxygen bubbles on the substrate. The CVs of the 3 electrodes are

shown in Fig. C-1-2. Although the plating bath with H2O2 produced a non-uniform

coating, it yielded the best electrochemical performance and the highest capacitance.

Fig. C-1-1: Appearance of Ti substrate, coated W oxide electrode without H2O2 and

coated W oxide electrode with 0.09M H2O2.

Fig. C-1-2: Cyclic voltammograms of Ti substrate, coated W oxide electrode without

H2O2 and coated W oxide electrode with 0.09M H2O2 at 10 mV/s in 0.5 M H2SO4.

84

Appendix C-2: XRD of WO3 Electrodes

The XRD pattern of as-deposit WO3 electrode was obtained and shown in Fig. C-2-1.

The XRD pattern of WO3 electrode almost overlapped with that of Ti substrate. From 16

to 38 (2θ) degree, where characteristic peaks of WO3 should locate, no tungsten peaks

were observed. Hence the as-deposit WO3 had an amorphous structure.

Fig. C-2-1: XRD patterns of as-deposit WO3 electrode overlapping with that of Ti

substrate.

85

Appendix C-3: Reproducibility of WO3 Electrodes

Eight electrodes deposited by the same method (i.e. electrodeposition from 0 V to -0.85

V at 0.01 V/s for 37 min) were tested and their CVs are shown in Fig. C-3-1. It can be

seen that the 8 CVs have large discrepancy in capacitance. The capacitances

calculated from the CVs were 4.8, 13.6, 6.6, 5.6, 10.2, 16.7, 4.1 and 4.3 mF/cm2, which

resulted in an average value of 8.2 mF/cm2. The considerable difference in capacitance

was due to the oxygen bubbles on the electrode surface during the electrodeposition.

However, the all CV shapes were similar, indicating these electrodes had the same

chemistry.

Fig. C-3-1: Cyclic Voltammograms of eight WO3 electrodes produced by the optimized

method.

86

Appendix D-1: Reproducibility of V Mixed Oxides Electrodes

Reproducibility of the synthesized V mixed oxides electrode was evaluated using the

0.5M-5days deposition method. 3 electrodes deposited by the optimized process were

tested and their CVs are shown in Fig. D-1-1. It can be seen that the 3 CVs almost

overlapped. The capacitances calculated from the CVs were 76.2, 72.3 and 67.0

mF/cm2, which resulted in an average value of 71.8 mF/cm2. The method was

considered fairly stable as the produced electrodes had similar CV profile and

capacitance.

Fig. D-1-1 Cyclic Voltammograms of 3 vanadium oxides electrodes produced by the

optimized method.