Min, Guanbo (2022) Optimisation of Triboelectric Nanogenerator

performance in vertical contact-separation mode. PhD thesis.

[Link]

Copyright and moral rights for this work are retained by the author
A copy can be downloaded for personal non-commercial research or study,
without prior permission or charge
This work cannot be reproduced or quoted extensively from without first
obtaining permission in writing from the author
The content must not be changed in any way or sold commercially in any
format or medium without the formal permission of the author
When referring to this work, full bibliographic details including the author,
title, awarding institution and date of the thesis must be given

Enlighten: Theses
[Link]
research-enlighten@[Link]
 Optimisation of Triboelectric
Nanogenerator Performance in Vertical
Contact-Separation Mode

Guanbo Min

A thesis submitted to the James Watt School of Engineering at the University of

Glasgow in partial fulfilment of the requirements for the Degree of Doctor of

Philosophy

February, 2022
 Abstract
Triboelectric nanogenerator (TENG) is one of the most promising energy harvesters – a technology

that uses repeated or reciprocating contact of suitably chosen materials to generate charge via the

triboelectric effect (TE) and utilizes this as usable voltage and current. TENGs are attractive as they

can continuously generate charge over a wide range of operating conditions and have several

valuable advantages such as light weight, simple structure, low cost and high efficiency. Therefore,

TENGs have been explored in a wide range of applications, including self-powered wearable

electronics, powering electronics and even for harvesting ocean wave/wind energy. One of the major

limitations of TENGs is their low power output (usually <500 W/m2). This thesis focuses of a few

specific approaches to optimising TENG output performance. This thesis begins by presenting a

solution to this challenge by optimizing a low permittivity substrate beneath the tribo-contact layer.

The open circuit voltage is found to increase by a factor of 1.3 in moving from PET to the lower

permittivity PTFE. TENG performance is also believed to depend on contact force, but the origin

of the dependence had not previously been explored. Herein, we show that this behaviour results

from a contact force dependent real contact area Ar as governed by surface roughness. The open

circuit voltage Voc, short circuit current Isc and Ar for a TENG were found to increase with contact

force/pressure. Critically, Voc and Isc saturate at the same contact pressure as Ar suggesting that

electrical output follows the same evolution as Ar. Assuming that tribo-charges can only transfer

across the interface at areas of real contact, it follows that an increasing Ar with contact pressure

should produce a corresponding increase in the electrical output. These results underline the

importance of accounting for real contact area in TENG design, as well as the distinction between

real and nominal contact area in tribo-charge density definition. High-performance ferroelectric-

assisted TENGs (Fe-TENGs) are developed using electrospun fibrous surfaces based on P(VDF-

TrFE) with dispersed BaTiO3 (BTO) nanofillers in either cubic (CBTO) or tetragonal (TBTO) form

in this thesis. TENGs with three types of tribo-negative surface were investigated and output

increased progressively. Critically, P(VDF-TrFE)/TBTO produced higher output than P(VDF-

TrFE)/CBTO even though permittivity is nearly identical. Thus, it is shown that BTO fillers boost

output, not just by increasing permittivity, but also by enhancing the crystallinity and amount of the

β-phase (as TBTO produced a more crystalline β-phase present in greater amounts)

1
Table of contents

Abstract .......................................................................................................................... 1
Table of contents ........................................................................................................... 2
List of figures ................................................................................................................ 5
List of tables ................................................................................................................ 10
List of publications and awards .................................................................................. 11
Glossary of abbreviations............................................................................................ 14
Chapter I. ..................................................................................................................... 15
Introduction ................................................................................................................. 15
1.1 Motivation ......................................................................................................................... 15
1.2 Research objectives .......................................................................................................... 17
1.3 Key contributions of this thesis ...................................................................................... 17
1.4 Structure of the thesis ...................................................................................................... 19

Chapter II. ................................................................................................................... 21

Literature Review ........................................................................................................ 21
2.1 New energy resources ...................................................................................................... 21
2.1.1. Renewable energy................................................................................................................ 21
2.1.2. Micro/nano energy .............................................................................................................. 24

2.2 Triboelectric nanogenerator ..................................................................................... 28

2.2.1 Triboelectrification ............................................................................................................... 28
2.2.2 Working principle of triboelectric nanogenerators........................................................... 29
2.2.3 The modes of triboelectric nanogenerators ....................................................................... 31

2.3 Triboelectrification of materials..................................................................................... 37

2.4 Research progress on triboelectric nanogenerators .................................................... 46
2.5 Summary ........................................................................................................................... 61

Chapter III................................................................................................................... 63
Modelling, simulation and optimization for contact and separation mode
triboelectric nanogenerators ....................................................................................... 63
3.1 Introduction ...................................................................................................................... 63
3.2 Model of the working principle of contact and separation mode TENGs ............... 65
3.3 Modelling of surface tribo-charge transfer .................................................................. 72
3.3.1 Metal-Dielectric Mode ......................................................................................................... 73
3.3.2 Dielectric-Dielectric Mode ................................................................................................... 75

3.4 Optimizing CS-TENG performance via a low permittivity substrate ..................... 77

2
 3.4.1 Simulation and fabrication .................................................................................................. 78
3.4.2 Results and discussion.......................................................................................................... 82
3.4.3 Application in a self-powered system by integration with a stretchable photodetector. 85

3.5 Conclusions ....................................................................................................................... 87

Chapter IV. .................................................................................................................. 89

Origin of the contact force-dependent response of triboelectric nanogenerators .... 89
4.1 Introduction ...................................................................................................................... 89
4.2 Fabrication and experiments .......................................................................................... 90
4.2.1 TENG fabrication and surface characterization ............................................................... 90
4.2.2 TENG Electrical measurements ......................................................................................... 91
4.2.3 TENG Contact area measurements .................................................................................... 93

4.3 Results: electrical output and contact area .................................................................. 94

4.4 Implications and applications of pressure and area dependence ........................... 101
4.5 Conclusions .................................................................................................................... 104

Chapter V................................................................................................................... 107

Ferroelectric-Assisted High-Performance Triboelectric Nanogenerators based on
Electrospun P(VDF-TrFE) Composite Nanofibers with Barium Titanate
Nanofillers ................................................................................................................. 107
5.1 introduction .................................................................................................................... 107
5.2 Experiments ................................................................................................................... 108
5.2.1 Preparation of Materials ................................................................................................... 108
5.2.2 Electrospinning ................................................................................................................... 110
5.2.3 Characterization and testing ............................................................................................. 112
5.2.4 Graphene supercapacitor .................................................................................................. 112

5.3. Result and Discussion .................................................................................................. 113

5.3.1 TENG performance ........................................................................................................... 113
5.3.3 Ferroelectric-assisted TENG powered high performance self-charging system .......... 128

5.4 Conclusions .................................................................................................................... 132

Chapter VI. ................................................................................................................ 134

A wide range self-powered flexible pressure sensor based on the triboelectric
nanogenerator ........................................................................................................... 134
6.1 Introduction ................................................................................................................... 134
6.2 Characterization of self-powered pressure sensor ................................................... 135
6.3 Applications .................................................................................................................... 137
6.3.1 Monitoring of human body movements ........................................................................... 137
6.3.2 TENG as self-powered impact force sensor for falling objects ...................................... 140

6.4 Conclusion ...................................................................................................................... 143

3
Chapter VII. .............................................................................................................. 144
Conclusion and future perspectives ......................................................................... 144
7.1 Conclusions .................................................................................................................... 144
7.2 Future perspectives ....................................................................................................... 147

Appendix A: Surface Roughness Metrology ............................................................ 149

Appendix B: MATLAB code ..................................................................................... 150
References: ................................................................................................................ 151

4
List of figures
Fig. 2.1: Renewable energy harvesting including solar energy, wind power, hydropower, biomass energy

and geothermal energy. ............................................................................................................................ 23

Fig. 2.2: Working principles of micro/nano energy harvesting including pyroelectric nanogenerator,

piezoelectric nanogenerator and triboelectric nanogenerator [44, 55, 56]. ............................................. 27

Fig. 2.3: Working principle of contact-mode triboelectric nanogenerators: (a) open circuit voltage (b))

short circuit current.................................................................................................................................. 31

Fig. 2.4: Working principle of the four TENG modes of operation and their equivalent circuits [80, 90,

93, 102, 103]. ........................................................................................................................................... 33

Fig. 2.5: Studies on the effect of contact area in CS-TENGs [160，171，172] .................................... 47

Fig. 2.6: A variety of applications of TENG based self-powered sensors, including: (a) self-powered

velocity and trajectory tracking sensors (VTTS), (b) self-powered velocity dual-mode sensor, (c) self-

powered acceleration sensor, (d) a wide range self-powered pressure sensor, (e) self-powered water wave

sensor and (f) health monitor [47, 208-216]. ........................................................................................... 57

Fig. 2.7: Types of TENG-supercapacitor-based self-charging systems (a) sandwich shape self-charging

power unit (SCPU), (b) output comparison of one TENG based SCPU and two TENG based SCPUs (c)

self-charging screen printed micro-supercapacitor power unit and (d) biodegradable SCPU [219-221].

................................................................................................................................................................. 59

Fig. 3.1: Schematic of the CS-TENG used for numerical study by (a) 2D (b) 3D version [42]. ............ 67

Fig. 3.2: Output performance with separation distance (a) open circuit voltage (b) output charge density

and (c) short circuit current density. ........................................................................................................ 68

Fig. 3.3: The simulation results of (a) open circuit voltage (b) short circuit current density and (c) output

charge density. The effects for the output performances with thickness of PDMS (d) open circuit voltage,

(e) output charge density and (f) short circuit current density. ................................................................ 69

Fig. 3.4: (a) Schematic and device structure information of a metal-dielectric mode CS-TENG and (b)

the energy band diagram for surface tribo-charge transfer at before contact and after contact status. ... 74

Fig. 3.5: (a) Schematic and device structure information for a dielectric-dielectric mode CS-TENG and

(b) the energy band diagram for surface tribo-charge transfer at before contact and after contact status.

................................................................................................................................................................. 76

5
Fig. 3.6. Fabrication steps for the ‘top tribo-layer’ [42]. ......................................................................... 81

Fig. 3.7. Schematic of equivalent circuit for open circuit voltage measurements [49]. .......................... 81

Fig. 3.8: Simulation results: (a) Open circuit voltage vs separation distance with 5cm×5cm areas and

200μm thickness substrate, (b) Open circuit voltage vs separation distance with 2cm×2cm areas and

100μm thickness substrate, (c) Short circuit charge den sity vs time, and (d) Short circuit current density

vs time with substrate materials having a range of permittivity values (at 1mm separation distance for c

and d), (e)–(f) Open circuit voltage, short circuit current density and short circuit charge density versus

relative permittivity for different substrate materials (at 1mm separation) [42]. .................................... 84

Fig. 3.9: Comparison of open circuit voltage signals for the devices with PTFE and PET substrates [42].

................................................................................................................................................................. 85

Fig. 3.10: Bending stability test for TENG devices: (a) Optical image while convex bending, (b) Optical

image while concave bending [50]. ......................................................................................................... 86

Fig. 3.11: (a) Equivalent circuit for TENG-PD based self-powered system, (b) Integration of stretchable

PD with TENG: Output voltage of TENG at different frequency across PD irradiated with and without

UV light and their respective responsivity at genera ted voltage. Photographic demo showing the

connection of LED in series with self-powered photodetector for real-time UV detection [50]. (Courtesy:

the photodetector was designed and fabricated by Dr. Yogeenth Kumaresan)........................................ 87

Fig. 4.1. Schematic of the contact separation mode TENG having copper (Cu) in contact with

polyethylene terephthalate (PET) [152]. ................................................................................................. 91

Fig. 4.2. TENG surface topography (a) Cu foil and (b) PET film. Note: the z-axis is in micrometers for

Cu and nanometer for PET due to the roughness difference [152]. ........................................................ 92

Fig. 4.3. TENG self-aligning test setup: (a) Photo of test rig in mechanical test machine and (b) Sectional

view of the test rig. The numbered parts are (1) TENG device (Cu and PET), (2) Glass plates, (3) Upper

rig platform, (4) Lower rig platform, (5) S pherical bearing ball and (6) Locking screws [152]............ 93

Fig. 4.4: Schematic of equivalent circuit for (a) open circuit voltage measurements, and (b) short circuit

current measurements [152]. ................................................................................................................... 94

Fig. 4.5: Real contact area measurement using pressure sensitive film: (a) Pressure sensitive film

sandwiched between tribo-layers, (b) Compression of layers under contact force, (c) Pressure sensitive

film operation and (d) Sample tested film showing areas of contact in red/pink [152]. ......................... 95

Fig. 4.6. (a) Sample of TENG open circuit voltage signals for increasing contact force and (b) Peak open

6
circuit voltage and peak short circuit current versus nominal contact pressure (and force) [152]. ......... 96

Fig. 4.7: (a) Real contact area as detected by pressure sensitive film with increasing contact pressure

showing as-scanned film with pink indicating solid contact (left) and binarized image with white

indicating solid contact (right) and (b) Contact area ratio (Ar/An) versus nominal contact pressure (and

force). Note: the nominal contact patch size in (a) is 2.5 cm × 2.5 cm [152]. ........................................ 97

Fig. 4.8: (a) Open circuit voltage and contact area ratio (Ar/An) versus nominal contact pressure and (b)

Short circuit current and contact area ratio (Ar/An) versus nominal contact pressure. The blue hatched

regions represent the pressure range in previous literature (as depicted in Table 4.2) [152]. ............... 101

Fig. 4.9: (a) LED array performance versus nominal contact pressure and force (and contact area ratio)

and (b) representative energy harvesting circuit. The LED array represents the word “BEST” – the LEDs

become progressively brighter as the contact pressure (and contact area) is increased [152]. ............. 102

Fig. 4.10: Overall view of the relationship between CS-TENG output and contact pressure [152]. .... 107

Fig. 5.1: Fabrication procedures for (a) P(VDF-TrFE), (b) P(VDF-TrFE)/CBTO and (c) P(VDF-

TrFE)/TBTO NF mats by electrospinning [49]. .................................................................................... 110

Figure 0.2Fig. 5.2: TENG performance versus wt% of TBTO fillers in the tribo-negative P(VDF-

TrFE)/TBTO layer showing (a) open circuit voltage and (b) short circuit current density [49]. .......... 111

Fig. 5.3: indicates the purely piezoelectric output when a piezoelectric nanogenerators is fabricated for

the PVDF-TrFE/TBTO case [49]. ......................................................................................................... 112

Fig. 5.4: (a) Schematic of equivalent piezoelectric nanogenerator (i.e. piezo only) by PVDF-TrFE/TBTO

NFs, (b) output voltage signal with varying applied contact forces and (b) output voltage with varying

operation frequencies [49]. .................................................................................................................... 113

Figure 0.1Fig. 5.5: Characterization of device performance. (a-c) illustrate Fe-TENG device structure

having negative tribo-surface as: (a) P(VDF-TrFE), (b) P(VDF-TrFE)/CBTO and (c) P(VDF-

TrFE)/TBTO (with PET at the positive tribo-layer in each case). (a-c) also illustrate the role of

ferroelectric polarization in affecting the triboelectric surface potential and charge density by indicating

how the bias of the bound charges and the level of tribocharges increases as we move to P(VDF-

TrFE)/TBTO (Note: XC is degree of crystallinity & σT is tribocharge density). Electrical output is

characterized in (d-i) for each of the three device types as follows: (d) open circuit voltage, (e) short

circuit current density, (f) output power density and (g-i) output current and voltage versus resistive load

for (g) pristine P(VDF-TrFE) & PET, (h) P(VDF-TrFE)/CBTO & PET and (i) P(VDF-TrFE)/TBTO &

7
PET. Contact pressure was fixed at 12.8 kPa. (j) transferred charge for three devices. (k) the stability of

P(VDF-TrFE)/TBTO & PET device [49]. ............................................................................................. 117

Fig. 5.6: Comparison of output voltage for the device with (a) PVDF-TrFE and (b) PVDF-TrFE/TBTO

by non-poling, poling and electrospinning processes [49]. ................................................................... 119

Figure 0.3Fig. 5.7: Materials characterization for P(VDF-TrFE) and P(VDF-TrFE)/BTO: (a) X-ray

diffraction (XRD) results of P(VDF-TrFE) (red line), P(VDF-TrFE)/CBTO (black line) and P(VDF-

TrFE)/TBTO (blue line) with crystalline β-phase (110/200) and XRD patterns of the cubic and tetragonal

BTO nanoparticles (001/100); (b) as-measured dielectric properties of the nano-fibrous mats for: P(VDF-

TrFE), P(VDF-TrFE)/CBTO, and P(VDF-TrFE)/ TBTO...................................................................... 121

Figure 0.4Fig. 5.8: COMSOL simulated potential distributions for PVDF-TrFE (left), PVDF-

TrFE/CBTO (middle) and PVDF-TrFE/TBTO (right) devices. Surface potential is clearly highest for the

TBTO case [49]. .................................................................................................................................... 124

Fig. 5.9: (a) Overlapped electron cloud (OEC) model of electron transfer between interface pairs for (a)

P(VDF-TrFE) & PET, (b) P(VDF-TrFE)/CBTO & PET and (c) P(VDF-TrFE)/TBTO & PET [49].... 126

Fig. 5.10: Open circuit voltage versus nominal contact pressure [49]. ................................................. 128

Fig. 5.11: Surface topography and contact analysis: surface height contour plots of (a) P(VDF-TrFE), (b)

P(VDF-TrFE)/CBTO and (c) P(VDF-TrFE)/TBTO. SEM images of nanofiber surfaces for (d) P(VDF-

TrFE), (e) P(VDF-TrFE)/CBTO and (f) P(VDF-TrFE)/TBTO (inset images at higher magnification).

BEM prediction of real contact area (in black) for: (g) P(VDF-TrFE) & PET, (h) P(VDF-TrFE)/CBTO

& PET and (i) P(VDF-TrFE)/TBTO & PET. Note how the predicted contact area follows the peaks of

the fibers in the fiber orientation [49]. ................................................................................................... 129

Fig. 5.12. (a) Schematic of graphene electrode based flexible supercapacitor (SC), (b) energy harvesting

circuit for self-charging system (i.e. integration with the Fe-TENGs), (c) charging and discharging

voltage of SC at different contact forces (freque ncy of 8Hz), (d) the maximum value of voltage charge

into the SC versus nominal contact pressure, (e) charging and discharging voltage of SC by different

frequencies (contact pressure 12.8 kPa), (f) the maximum value of voltage charge into the SC versus

vibration frequencies, (g) representative energy harvesting circuit of Fe-TENGs for lighting LEDs and

(h) photograph of lighted LED array representing the word ‘‘BEST’’ [49]. ......................................... 132

Fig. 6.1: Characterization of TENG based self-powered pressure sensor and its potential applications: (a)

working principles of a CS-TENG, (b) a sample of TENG Voc signal under the contact pressure of

8
1176kPa (735N). (c-e) Voltage output of TENG based self-powered pressure sensor for: (c) low, (d)

medium-high pressure and (e) Super high-pressure ranges [27]. .......................................................... 139

Fig. 6.2: Voltage signal of the pressure sensing TENG when pressed using different numbers of fingers.

............................................................................................................................................................... 140

Fig. 6.3: Gait information from a man of 80kg weight collected by the pressure sensing TENG (a)

walking, (b) jogging, (c) running and (d) jumping. ............................................................................... 141

Fig. 6.4: Gait information from a woman of 45kg weight collected by the pressure sensing TENG (a)

walking, (b) jogging, (c) running and (d) jumping. ............................................................................... 142

Fig. 6.5: (a) Schematic of free-falling object onto TENG impact sensor from different heights of 5cm,

10.5cm, 13cm, 14cm; (b) output voltage peaks of TENG impact sensor showing the consecutive

rebounding of the object from different heights shown in (a); (c) represents the results of output voltage

against heights; (d) depicts impact velocity versus fall height and (e) shows the results for impact force

against heights. ...................................................................................................................................... 145

9
List of tables
Table. 2.1: Comparisons of output performances for energy harvesters [45, 52]. .................................. 26

Table. 2.2: Triboelectric series for common materials following a tendency to lose electrons (positive) to

gain of electrons (negative) [12].............................................................................................................. 40

Table 2.3: Literature comparison of TENG factors and parameters effecting electrical output. ............. 47

Table 2.4: Typical PVDF based TENG performances compared with the present work (final row) ...... 56

Table 3.1. Summary table of the key theoretical formulas describing CS-TENG operation. ................. 74

Table 3.2. Parameters for ‘top tribo-layer’. ............................................................................................. 81

Table 3.3. Substrate relative material permittivity εr, ‘top tribo-layer’ relative equivalent permittivities

ε(r,eq) and ‘top tribo-layer’ absolute equivalent permittivities εeq. ........................................................ 81

Table 3.4: Comparison of fabricated TENG devices and test parameters. .............................................. 82

Table 4.1: Areal surface topography parameters of TENG tribo-contact surfaces (see Appendix A for

information on parameters used). ............................................................................................................ 94

Table 4.2: Typical nominal contact pressures (force/area) used in TENG studies compared with the

present work (final row). ....................................................................................................................... 101

Table 4.3: Typical contact pressures available in possible high load applications for TENGs. ............ 105

Table 5.1: Areal root mean square surface roughness of the tribo-contact surfaces and predicted contact

area percentage (BEM) of the Fe-TENGs tribo-contact pair. ................................................................ 132

Table 5.2: Typical charging performance and charging period for TENG based self-charging systems in

the literature as compared with the present work (final row). ............................................................... 136

No table of figures entries found.

10
List of publications and awards
Journal papers:
1. G. Min, L. Manjakkal, D. Mulvihill and R. Dahiya, Triboelectric Nanogenerator with
Enhanced Performance via an Optimized Low Permittivity Substrate, IEEE Sensors
Journal, (20) 6856 – 6862,2020. (DOI: 10.1109/JSEN.2019.2938605)
2. G. Min, Y. Xu, P. Cochran, N. Gadegaard, D.M. Mulvihill, R. Dahiya, Origin of the
Contact Force- Dependent Response of Triboelectric Nanogenerators, Nano Energy,
(83) 105829, 2021. (DOI: 10.1016/[Link].2021.105829)
3. G. Min, A. Pullanchiyodan, A. S. Dahiya, E.S. Hosseini, Y. Xu, D.M. Mulvihill and R.
Dahiya,Ferroelectric-Assisted High-Performance Triboelectric Nanogenerators based
on Electrospun P(VDF- TrFE) Composite Nanofibers with Barium Titanate Nanofillers,
Nano Energy, (90) 106600, 2021. (DOI: doi:10.1016/[Link].2021.106600).
4. Y. Xu, G. Min, N. Gadegaard, D. Mulvihill and R. Dahiya, A unified contact force-
dependent modelfor triboelectric nanogenerators accounting for surface roughness,
Nano energy, (76) 105067, 2020.(DOI: 10.1016/[Link].2020.105067)
5. Y. Kumaresan, G. Min, A. S. Dahiya, A. Ejaz, D. Shakthivel and R. Dahiya,
Kirigami and Mogul patterned Ultra-Stretchable High-Performance ZnO
Nanowires based Photodetector, Advanced Materials Technologies, 2100804,
2021 (DOI: 10.1002/admt.202100804).
6. Guanbo Min, Gaurav Kandelwal, Abhishek Singh Dahiya, Radu Chirila, Daniel M.
Mulvihill and Ravinder Dahiya, “A wide range self-powered flexible pressure sensor
based on triboelectric nanogenerator.” (Under review, IEEE Sensors Journal)

Conference paper:
1. G. Min, L. Manjakkal, D. M. Mulvihill and R. Dahiya, “Enhanced triboelectric
nanogenerator performance via an optimised low permittivity, low thickness substrate,”
IEEE Sensors conference, N. Delhi, India, Oct 2018.
2. G. Min, A. Dahiya, D. M. Mulvihill and R. Dahiya, “A wide range self-powered flexible
pressure sensor based on triboelectric nanogenerator” the IEEE Int. Conf on Flexible and
Printable Sensors andSystems (FLEPS 2021), Manchester, UK, June 2021.

Awards:
1. Finalist (Top 4) at British Society for Strain Measurement (BSSM), 2021 Young Stress
Analyst Competition.
2. Top 5% paper award at IEEE Sensors 2018 Conference (N. Delhi, India).

11
Acknowledgements
I would like to express my utmost gratitude and thanks to my supervisors Dr. Daniel
Mulvihill and Prof. Ravinder Dahiya for their continuous support through my PhD. I
appreciate the trust and freedom they gave me to explore my own research interests,
without which this research would never have been possible. Further, I am grateful for
the excellent service from the electronics and mechanical workshop and the James Watt
Nanofabrication Centre (JWNC) staff.
I want to extend my sincere thanks to all the members of Bendable Electronics and
Sensing Technologies (BEST) group and Materials and Manufacturing Research Group
(MMRG). I particularly thank Dr. Libu Manjakkal, Prof. Yang Xu, Dr. Abhishek
Dahiya, Dr Abhilash Pullanchiyodan, Dr. Gaurav Khandelwal, Dr Ensieh Seyed
Hosseini and Mr. Radu Chirila who were closely associated in shaping some key
outcomes of the research presented in this thesis and the associated publications. Also,
I would like to thank Dr Yogeenth Kumaresan, Dr Saoirse Dervin, Dr. Bhavani
Yalagala and Mr. Adamos Christou who supported me during the work.
Finally, I would like to express my deepest gratitude to my parents (Mr. Qinggang Min
and Ms. Ling Li) for their unconditional sponsorship, love and support. Especially,
thanks for your encouragement, when I was down, alone and sad. Thank you for your
love and for your complete faith in me; this success would not be possible without you.

12
Author’s declaration
I declare that, except where explicit reference is made to the contribution of others, this
thesis is the result of my own work and has not been submitted for any other degree at
the University of Glasgow or any other institution.

Guanbo Min

13
Glossary of abbreviations
TENG Triboelectric nanogenerator
CS-TENG Contact and separation mode triboelectric nanogenerator
LS-TENG Lateral sliding mode triboelectric nanogenerator
SE-TENG Single electrode contact mode triboelectric nanogenerator
CF-TENG Contact freestanding tribo-layer mode triboelectric nanogenerator
T-TENG Textile triboelectric nanogenerator
Voc Open circuit voltage
ISC Short circuit current
Qsc Short circuit output charge
Jsc Short circuit current density
Ar Real contact area
!! Tribo-charges density
!" Free charge density
C Capacitance
x(t) Separation distance
d Thickness
"# Electric field
# Permittivity
#$ Permittivity of air
#% Relative permittivity
An Norminal contact area
TE Triboelectric effect
FE Ferroelectric effect
PTFE Polytetrafluoroethylene
PDMS Polydimethylsiloxane
PVDF Polyvinylidene fluoride
P(VDF- Poly(vinylidene fluoride-trifluoroethylene)
TrFE)
TBTO Tetragonal Barium titanate
PET Polyethylene terephthalate

PMMA Poly(methyl methacrylate)

PI Kaptan

14
Chapter I.

Introduction
1.1 Motivation
Since energy is the basis for human beings to accomplish their daily activities (e.g.,
powering electronics and commercial activity), its generation and impact on the
environment has become the pre-eminent issue around the world. However, as
consumption of non-renewable energy (i.e., petroleum, coal, etc.) is increasing
exponentially, a world energy crisis is approaching progressively and restricting
economic development [1]. Reportedly, it is roughly estimated that the reserves of coal
and uranium will be exhausted by the mid-21st century and within 170 years [1-3]. The
reserves of petroleum and natural gas are approximately 1.3k hundred million tons and
140 kMcm3 worldwide, respectively [4, 5]. Petroleum is expected to be exploited
exhaustively until 2055 with an exploitation speed of 30 hundred million tons per year,
and natural gas also can only be used for a further 50 years, with a consumption speed
of 2.5 kMcm3/year. Because of the high consumption and the limited reserves of the
raw material, it is a tremendous challenge for humanity to acquire innovational energy
sources for the future. Indeed, the development of such an energy source can enhance
the development of a country, placing it in a better international position with
accompanying influence. Developing a renewable energy source requires significant
pioneering work. Note that an innovational energy source, which has the characteristic
of low pollution and high energy storage, means energy without the traditional
dependency on fossil fuels. This development of renewable energy can respond to the
challenge of resource exhaustion and environmental pollution [6].
Additionally, exploiting and employing microdevices and micro-electromechanical
systems (MEMS) is one of the methods to settle the challenge of resource exhaustion,
because of their low power consumption [7-10]. The external power suppliers (e.g.,
lithium batteries and supercapacitors) are necessary for these microdevices and MEMS
to provide continuous electrical energy supply. Simultaneously, these external power
suppliers increase the limitation of the energy storage and environmental pollution [11].

15
In order to figure out the energy requirements, nanogenerators were suggested to
convert ambient energy into electricity. Because of the separate positive and negative
charges, the differential potential can drive induced charges as the oriented current to
power the applications [12-15]. Most of the nanogenerators depend on properties such
as pyroelectricity, piezoelectricity and triboelectricity [13, 16-19]. Therein,
triboelectrification and electrostatics are part of our daily lives – think for example of
the charging of cloths when rubbed over hair. Generally, triboelectrification is often
regarded as a negative effect, for example, factory dust explosions, separation between
medicine and powders, etc. Additionally, triboelectrification has often been ignored by
researchers, as triboelectricity was difficult to harvest and utilise. The advent of
triboelectric nanogenerators (TENGs) changed the stereotype of the negative effect of
triboelectrification, due to its high energy conversion efficiency for converting ambient
mechanical energy to electricity (the maximum conversion efficiency can be 85%) [12].

Triboelectric energy harvesting depends on converting ambient random mechanical

energy to electricity, and it can reduce the amount of energy waste. Depending on this
definition, TENGs are fabricated by two types of flexible materials with large
differences in the triboelectric properties such as the interface pair [12, 13]. Due to the
rubbing or contact and separation of the contacting tribo-material surfaces, tribo-
charges are generated on the interface surfaces. Because of electrostatic induction and
polarization, applications can be powered by free charges which are induced on the two
sides of the electrodes. Using TENGs, the mechanical energy from daily life can be
transformed efficiently, including vibration, human motion, ocean wave, wind power,
etc [20-23]. Moreover, TENGs can also be applied as self-powered sensors and into
self-charging systems, to increase their coefficient of utilization [24, 25]. Especially,
this can be a response to the challenge of a 24-hour power supply. These results
demonstrate that this technology can be valuable in wearable applications, where higher
power output, more efficient charging and flexibility are paramount.

16
1.2 Research objectives
The main objective of this work was to understand, design and characterize high
performance flexible TENGs and their applications in self-powered/charging systems.
The detailed major objectives are:

1. To understand the principle of contact and separation mode triboelectric

nanogenerators (CS-TENGs) by MATLAB simulation to summarize the effect of
parameters for optimisation of device performance.
2. To characterize the electrical output of TENGs and the effect of real contact area
(Ar) at the tribo-contact interface.
3. To understand the origin of the contact force-dependent response of CS-TENGs.
4. To investigate the properties of TENG contact materials that can improve surface
tribo-charge generation.
5. To understand the principle of optimizing the performance of the tribo-material by
electrospinning and fabricating a high-performance tribo-layer by electrospinning.
6. To design a CS-TENG as a self-powered sensor and integrate a high-performance

TENG with a supercapacitor to develop as a wearable self-charging system.

1.3 Key contributions of this thesis

The key contributions of this thesis can be summarized as follows:

• A simulation study was presented to investigate the effects of varying substrate

relative permittivity upon TENG outputs such as open circuit voltage, short circuit
current density and transferred charge density. In addition to further in-depth
simulation, two fabricated devices with polyethylene (PET) and
polytetrafluoroethylene (PTFE) substrates also were tested to verify the simulation
results.

• The work is the first-time both real contact area and electrical output have been

measured simultaneously between two nominally flat surfaces during the operation
of a TENG device.

17
• The origin of the contact force-dependent response of TENGs is shown to result

from the contact force dependence of the real contact area (Ar), as governed by the
unavoidable presence of surface roughness. The open circuit voltage Voc, short
circuit current Isc and Ar for a TENG, having two nominally flat tribo-contact
surfaces, were found to increase with contact force/pressure. The Ar was notably
small at low pressures (0.25% at 16 kPa) that are typically experienced in wearable
applications. However, it increased 328-fold to as much as 82% when it saturated
beyond about 1.12 MPa pressure - achievable for impact with ocean waves.
Critically, Voc and Isc saturate at the same contact pressure as Ar suggesting that
electrical output follows the evolution of the Ar as governed by the surface
topography.

• This thesis also presents high-performance ferroelectric-assisted TENGs using

electrospun fibrous surfaces based on P(VDF-TrFE) with dispersed BaTiO3 (BTO)

nanofillers in either cubic (CBTO) or tetragonal (TBTO) form. The results shown
were that BTO fillers boost output, not just by increasing permittivity, but also by
enhancing the crystallinity and amount of the β-phase (as TBTO produced a more
crystalline β-phase present in greater amounts).

• The ferroelectric-assisted TENG was integrated with a flexible graphene electrode-

based supercapacitor to produce a self-charging system capable of charging to 1.25

V in just 5 min.

• A CS-TENG was designed as a wide-range, self-powered flexible pressure sensor.

The fabricated TENG can detect a wide range of applied pressures from 3.2 to 1176
kPa. The sensor sensitivities in three pressure regimes, namely low (1-10 kPa),
medium-high (10-500 kPa) and ultra-high (>500 kPa) were extracted as 3.16, 0.023
and 0.031 V/kPa, respectively.

• A wide range pressure sensor has been demonstrated showing how the sensor can

be used to detect finger tapping, collect user gait information and detect impact
forces.

18
1.4 Structure of the thesis
Following the introduction, this thesis is organised as follows:

• Chapter II introduces an overview of triboelectric nanogenerators (TENGs) with

emphasis on the recent progress on contact and separation mode triboelectric

nanogenerators (CS-TENGs). In particular, the power generation mechanisms,
working principles of the four TENG modes and different tribo-materials that have
been used along with their properties and their performances. In addition, the key
parameters, which can optimize TENG performance have been highlighted.
Simultaneously, the thesis determines a rank order of the sensitivity of CS-TENG
output to the different parameters. Finally, because of the results of the comparisons,
three aspects have been afforded particular focus: increasing contact area between
the interface pair, enhancing the crystallinity of the contact materials and TENG
applications in self-powered sensors and self-charging systems.
Ÿ Chapter III presents the formulation and model validation through electrical
characterization of CS-TENGs. Firstly, the progress of CS-TENG models based on
device structure is introduced such as parallel-plate capacitor mode, distance-
dependent electric field model and Load-dependent model in the first part of this
chapter. After that, a numerical study on tribo-charge generation behaviors has been
analyzed by the energy bands between the interface pair materials. Finally, a study
on optimizing CS-TENG performance via a low permittivity substrate by both the
simulations and experiments is carried out.

• Chapter IV describes experiments that uncover the origin of the contact force-

dependent response of TENGs. Their performance is believed to depend on the

contact force, but its origin had not yet been established. Herein, we show that the
contact force-dependent electrical response is due to the contact force-dependent
real contact area that results from the inevitable presence of surface roughness on
the tribo-contact surfaces.

• Chapter V develops high-performance ferroelectric-assisted TENGs using

electrospun fibrous surfaces based on P(VDF-TrFE) with dispersed BaTiO3 (BTO)

19
 nanofillers in either cubic (CBTO) or tetragonal (TBTO) form. TENGs with three
types of tribo-negative surface (pristine P(VDF-TrFE), P(VDF-TrFE)/CBTO and
P(VDF-TrFE)/TBTO) in contact with PET were investigated. Essentially, tribo-
charge transfer is boosted due to increased surface potential owing to enhanced
ferroelectric polarization induced during electrospinning. P(VDF-TrFE)/TBTO
produced higher output than P(VDF- TrFE)/CBTO even though permittivity is
nearly identical. Thus, it is shown that BTO fillers boost output, not just by
increasing permittivity, but also by enhancing the crystallinity and amount of the
β-phase (as TBTO produced a more crystalline β-phase present in greater amounts).

• Chapter VI presents a triboelectric nanogenerator (TENG) based wide-range, self-

powered flexible pressure sensor. The fabricated TENG can detect a wide range of
applied pressures from 3.2 to 1176 kPa. The sensor sensitivities in three pressure
regimes, namely low (1-10 kPa), medium-high (10-500 kPa) and ultra-high (>500
kPa) were extracted as 3.16, 0.023 and 0.031 V/kPa, respectively. Stable and
repeatable TENG responses were achieved for all three pressure regimes and this
is due to how the real contact area at the TENG interface varies reliably with
contact pressure. To the best of our knowledge, this is the first time a TENG as a
self-powered pressure sensor has been characterized over such a wide pressure
range (that includes higher pressures). These results highlight the potential for
TENGs as pressure sensors in a wide range of applications such as: detecting
pressure in wearables, self-powered e-Skin for the sole of humanoid robots to help
them walk and stand, and for detecting impact forces with water waves.

• Chapter VII summarizes the contribution of this research on tribo-materials

design, CS-TENG fabrication and characterization of high-performance CS-TENG

based self-powered sensor/ system and self-charging system. It also provides
suggestions for future work.

20
Chapter II.

Literature Review
2.1 New energy resources
2.1.1. Renewable energy
The utilization of energy was one of the major drivers of social economic development
and increasing standard of living. Wind power, water wave power, firewood and animal
power were the major applied power sources for a classical civilization [26-29]. After
the industrial revolution, fossil fuels gradually became the dominant energy source
around the world [30]. With the consumption of fossil fuels, renewable energy sources
are being explored and developed promptly. However, non-fossil and renewable energy
sources still only cover 20% of the total energy utilization [6]. With energy
requirements growing and the non-renewable traditional energy sources exhausting, an
energy crisis is approaching, even threatening human life. It is estimated that only one
person out of four can survive intro the indefinite future (on average) with this speed of
energy consumption. Therefore, renewable energy sources have been suggested as a
response to these energy challenges, because of their high efficiency, cleanliness,
diversification and globalization.
Additionally, the traditional energy source (e.g., fossil fuel) is falling out of fashion due
to its many harmful effects. Compared with the traditional energy sources, renewable
energy is clean, it can reduce the greenhouse effect and it reduces the problem of
resource shortage and dust pollution [3]. There are nine current renewable energy
sources that satisfy these conditions, including solar energy, geothermal energy, wind
power, hydrogen energy, biomass energy, nuclear energy, nuclear fusion energy, tidal
energy and natural gas hydrate (NGH) [16, 26, 31-37]. The advantages of renewable
energy technologies are: (1) they are an underdeveloped resource with high potential,
(2) there are abundant reserves widely spread geographically, (3) they are
environmentally friendly and (4), the wind power, geothermal energy and solar energy
can be a bottomless supply. In short, renewable energy has more advantages than the
traditional energy sources and application spaces. Currently, solar energy, geothermal

21
energy, tidal energy, wind power and biomass energy are being applied extensively and
have a highly developing prospect. They are explained further here (A to E):

A．Solar energy is one of the most well known renewable energy sources globally,
owing to its advantages of being clean, environmentally friendly and sustainable
etc. It depends on converting solar radiation energy to electricity. Notably, the
average solar radiation energy per hour can supply the energy to use for a year for
all humans sufficiently, which is approximately 430 quintillion joules [16].
However, solar cells have the disadvantage of energy interruption due to the
environmental limit [38]. For example, solar energy is only outstandingly adequate
in the sunny daytime and summer season, irrespective of the geographical location.

B． Wind power generation is the second largest source among the renewable energy
sources. The mechanical energy from wind can be collected by turbines to drive
generators which then transfer electricity into the national grid. Reportedly, more
than 5% of global electricity was from wind power production in 2018 with 591
GW of global capacity [26]. Nonetheless, wind power generation is not suitable for
every property due to its high noise, threatening presence to wildlife, unpredictable
wind intensity and location limitations. Hydropower is the largest renewable
energy source for power generation.

C． Similar to the wind power generation, hydropower generates electricity by the flow
of water to drive a turbine. Because of the barrier placed in the water (the dam),
the potential energy of the river can be controlled and also allow electrical power
to be stored for use when demand reaches a peak [32]. Therefore, compared with
the other renewable energy sources, hydropower is more reliable due to its stability
and controllability [39]. Tidal energy is a variation of hydropower. However, its
stability is lower than hydropower energy.

D．Biomass energy is a sort of energy that is shaped by solar energy in the form of

22
 chemical energy (i.e., renewable carbon source); it is because of photosynthesis
[33]. Simultaneously, biomass includes a wide range of varieties such as plant,
microorganism and the waste produced by the herbivores. There are 1,700 million
tons of organisms produced by photosynthesis per year [33]. Simultaneously, the
energy in these organisms is approximately 15 times the total energy consumption
in the world. However, its coefficient of utilization is only 3%, although biomass
is abundant [40]. Additionally, the cost of biomass power generation is
uncompetitive, and its waste is harming the environment.

E． Power generation by geothermal energy depends on the temperature difference

between the earth’s surface heat and the heat below the earth’s surface [31]. A
geothermal heat pump is used to convert this heat difference to the electricity. A
special application is ocean thermal energy, which depends on the temperature
difference between the ocean surface and deep below the surface [41]. Comparing
with fossil fuels, the heat (or electricity) can be extracted without burning.
Simultaneously, geothermal energy can be used uninterruptedly, compared with
solar energy and wind power. However, the release of H2S is the main disadvantage
of geothermal energy power generation.

Renewable energy is an important route of response to the energy crisis in the world.
In 2020, there are 34% world electricity from renewable sources [42]. However,
renewable energy sources still have weaknesses in their applications such as the energy
interruption from the environmental limit, location limit, low conversion rate and
technical difficulty. In addition, recent renewable energy is only replacing the original
mega energy by a green route. It still requires a direct and repaid energy route for
powering the micro/nanoelectronics/devices such as micro/nano-sensors, robotics,
health monitoring and tribotronic transistors [43]. It is to be noted that these devices
only need relatively low power (around 0.1–100 mW/m2). From the analysis in the
previous section, it can be seen that the method for power generation of renewable
energy involves:

23
1. Collecting the heat from the energy source and converting it into electricity.

2. Transferring the renewable energy sources as the mechanical energy. After that, the
mechanical energy can be converted to electricity (e.g., turbine).

However, their similarity is that a reasonable power terminal is adopted. Therefore,

after determining the collected energy category, designing a power terminal (e.g.,
nanogenerators, solar cells, etc) is the key to micro/nano energy harvesting (see Table
2.1).

Fig. 2.1: Renewable energy harvesting including solar energy, wind power, hydropower, biomass energy
and geothermal energy.

2.1.2. Micro/nano energy

The potential of renewable energy as the energy source for the large-scale power grid
has been discussed in Section 2.1.1. Their characteristics include high energy reserves,
high cost of power generation, large power terminal equipment and high output power.
Moreover, nano-scale energy (i.e., milliwatt) harvesting is also a promising method to
provide power for small electronics and nano/micro-systems, as it is sustainable and
maintains free and self-powered operation [12, 14]. The application of renewable

24
energy in the aspect of nano energy is also developing increasingly with the progression
of nanotechnology, including energy harvesting, energy conversion, energy storage and
applications [14, 24, 44]. Nano energy technology also improves the utilization factor
of renewable energy, as the random and waste energy in human life can also be utilized
such as human motion, gentle breezes and rolling tires. Integrating these varieties of
nano energies can also accumulate as high energy because nano energies also have high
energy reserves.
Therefore, a number of nanogenerator concepts have been designed depending on how
their materials properties might facilitate electricity generation. For instance, solar cells
depend on solar radiation. Pyroelectric nanogenerators depend on the temperature
sensing of materials [45, 46]. Piezoelectric nanogenerator and triboelectric
nanogenerator, on the other hand, can convert mechanical energy [12, 47].
Nanogenerators are innovative self-powered devices that can directly transfer ambient
energy to electricity. Nanogenerators can convert many variations of ambient energy to
electricity, including wind, water flow and human motion [12]. Additionally, this
promising technology has a wide range of applications. Firstly, nanogenerator-based
sensors and sensor arrays can detect location, acceleration, pressure, temperature and
humidity precisely [24]. For example, if we integrate nanogenerators into the slopes of
Parkinson’s disease, patients’ gait information (e.g., normal walking, rapid walking or
tumble) can be analysed effectively [48]. Because of its self-powered and highly
sensitive characteristic, its prospect of application obtains positive responses.
Simultaneously, it is also a beneficial power supplier for low power electronics such as
self-powered systems. Nanogenerators have been demonstrated to power pH sensors,
LED arrays and photodetector-based systems [49-51]. Finally, due to its high output
power (perhaps up to 500W/m2), energy storage systems also can be charged by
nanogenerators, such as supercapacitor and lithium batteries [50, 52]. Energy
harvesting through various mechanisms such as triboelectric, piezoelectric and
pyroelectric has been widely explored. These devices are discussed in this section.

25
Table. 2.1: Comparisons of output performances for energy harvesters [45, 52].

Energy Voltage Current Power density Efficiency Application

harvesters (V) ()A) (mW/cm2) (%)
PyENG 0.01-190 0.0006-11 0-0.014 0.01-14.8 Geothermal energy
and Solar energy
PENG 0.01-200 0.01-10 0.001–30 0.01-25 Wind power and
hydropower
TENG 12-1500 10-2000 0.01-100 10-85 Wind power and
hydropower
Pyroelectric nanogenerators (PyENGs) utilize pyroelectric material to convert ambient
thermal energy to electricity. Regularly, thermoelectric energy harvesting relies on the
Seebeck effect, where the heated electrons flow toward the cooler side of the device
[46]. However, the environmental temperature distribution is spatially consistent rather
than with a gradient. Thus, it is not able to use Seebeck effect to harvest thermal energy
in the environment, as the temperature is changing with time. Owing to this behaviour,
the PyENG is the most suitable selection for thermal energy harvesting. Because of the
anisotropy of the solid material, the pyroelectric material will be self-polarized by the
change of the temperature. A wearable PyENG, which was integrated on a N95 face
mask, can generate a power density of 0.68$W/cm2 [45]. Additionally, there are also
many potential applications for PyENGs such as self-powered temperature sensor,
temperature imaging and medical monitoring and diagnosing.
The piezoelectric nanogenerator (PENG) is the first-generation mechanical energy
harvester; it depends on the piezoelectric property of the material between two
electrodes [47]. Generally, PENGs were fabricated as a sandwich structure. Since
piezoelectric materials exhibit the phenomenon of self-polarization by external applied
pressure, the free charges can be induced on the electrodes by the internal electric field
(i.e., from the piezoelectric material self-polarization). After that, the free charges can
be induced on two sides of the electrodes. Recently, there are many well-known
piezoelectric (PE) materials that have been investigated for nanogenerators, including
zinc oxide (ZnO), zirconate titanate (PZT), barium titanate (BaTiO3) and
polyvinylidene fluoride (PVDF), even the ferroelectric material as Poly(vinylidene
fluoride-co-trifluoroethylene) (P[VDF-TrFE]) [53]. These materials with piezoelectric

26
property are so because of their specific atomic structure. For example, ZnO is a type
of metallic oxide with a piezoelectric property, as the zinc atom and oxide oxygen atoms
present non-symmetrical configuration in the ZnO molecule [47]. In addition, *-phase
PVDF is piezoelectric due to the differences in electronegativity between the fluorine
atom and the hydrogen atom [19]. The piezoelectric material is placed between two
electrodes in a PENG device. After a mechanical pressure is applied to the piezoelectric
material, the electric charges within piezoelectric crystal go ‘out of balance’ (i.e., self-
polarization) [47]. Owing to the material self-polarization, an internal electric field is
formed to induce the free charges on two sides of the electrodes. Because
piezoelectricity is the process of using crystals to convert mechanical energy to
electricity effectively, this principle is applied in a wide range of applications, including
human motion monitoring, voice recording, self-powered active sensors for health care
monitoring and electronic skins for discriminating static/dynamic pressure stimuli.
However, its low output power density (~30mW/cm2) [54] is the largest limit for
extending its applications, as depicted in Table 2.1.
The triboelectric nanogenerator (or TENG) is the second-generation mechanical energy
harvester. Recently, a high performance record for CS-TENGs was 100 W/m2 output
power density, which is over 3-fold higher than the average PENG output. Additionally,
the conversion efficiency of a TENG can be higher than 85%, which is approximately
5 times that of a solar cell [12, 16]. For instance, a spherical TENG with 7 cm diameter
has been fabricated to harvest water wave energy [55]. A single spherical TENG can
generate a 10mW instantaneous output power at 10Kohm load. It is estimated that, if
these types of TENG devices are installed onto a 1km2 area on the sea surface (i.e.,
combining them as a TENG-based energy network), they could produce 27.6GWh of
average output energy. This result envisages a blue energy dream based on TENGs, in
which many TENG-based energy networks combined by cables can provide energy for
a coastal town. This idea also demonstrates the possibility of compound nano/micro
scale energy to be converted as mega/giga scale energy. Overall, these reasons also
demonstrate that TENG technology can be more valuable in the applications of
wearable, water wave energy and sensing technology, where high power, high transfer
27
efficiency and flexibility are paramount.

Pyroelectric nanogenerator Piezoelectric nanogenerator

Micro/nano energy
harvesters

Triboelectric nanogenerator

Fig. 2.2: Working principles of micro/nano energy harvesting including pyroelectric nanogenerator,
piezoelectric nanogenerator and triboelectric nanogenerator [44, 55, 56].

2.2 Triboelectric nanogenerator

The key parameters that govern TENG performance are understanding in this section
2.2-2.3 and then explore ways to boost output in section 2.4.
2.2.1 Triboelectrification
Triboelectrification phenomenon are often considered as a negative effect in human
life. For example, electrostatics usually influence the wireless network during the flight
of a plane at high altitudes, interrupting the communication between the flight and
ground service. In pharmaceutical preparation, tribo-charges attract dust and decrease
medical purity. For electronics, the tribo-charges will breakdown the electronic element
and IC to boost the device burn-in. Additionally, the electrostatic effect easily attracts
dust on electronic surfaces such as TV and phone screens to reduce their resolution. In
factories, tribo-charges result in fires with inflammable and explosive materials.
Therefore, people usually avoid triboelectrification [56]; however, the triboelectric
nanogenerator is one of the applications that has a positive effect via high efficiency
power generation [12].
In the contact-separation triboelectric nanogenerator, it is the coupling of contact

28
electrification and electrostatic induction that defines the operating principle [12].
Therein, the generated tribo-charges are the power source that drives the TENGs to
induce the free charges on the electrodes [12, 57-59]. The generation of tribo-charges
depends on the contact electrification phenomenon between the contacting material
interfaces [58]. Because of that, the transfer of opposite and equal tribo-charges occurs
on each of the contact surfaces. The phenomenon of contact electctrification and
triboelectrification is widely experienced in daily life. For instance, after rubbing a
plastic rod on a woollen sweater, it can be observed to attract paper scraps and hairs.
Also, when people wear out their sweaters, they can sometimes hear the sound of a
charge or even see an emerging electric spark. Tribo-charge generation by contact
electctrification or triboelectrification depends on the differences in the tribo-polarity
between two tribo-materials [12]. Generally, when metal and dielectric or dielectric and
dielectric are in contact, an output electrical signal can be detected [60]. During the
friction between two metals, tribo-charges can also be generated; however, the charges
would be lost promptly due to the metal’s conductivity [60]. The tribo-charge
generation is widely understood as the charge transfer between the contacting atom and
molecule (or molecule and molecule) [12, 56, 60]. Note that the chain structure of a
material surface is different to the chain structure of material bulk, in that there are
dangling chains on a material surface [61, 62]. Furthermore, a temporary chemical bond
is formed during contact, because of the differences in electron affinity [63]. In order
to balance the differences in energy band (i.e. chemical potential) of the contacting atom
and molecule (or molecule and molecule), the electrons can transfer from one material
surface to the other. After the interface pair is separated, opposite and equal tribo-
charges will be accumulated on the two sides of the interface.

2.2.2 Working principle of triboelectric nanogenerators

The reason why the triboelectrification phenomenon could be employed as a
nanogenerator is because triboelectrification contains a vast energy reserve, and it is
occurring in real time. In 1929, Van de Graaff’s generator based on the
triboelectrification principle was born, which can transfer electrostatic charges to the

29
metal ball surface through corona discharging [64, 65]. Simultaneously, two electrodes
could be discharging, once the charge density reached the threshold value [65].
However, in 2012, the first TENG was reported by Zhonglin Wang’s group in Georgia
Tech [66]. When two tribo-materials contact each other by an external force, they
generate opposite and equal tribo-charges on two contact surfaces due to their
difference capability in gaining and losing electrons to produce the device internal
voltage [12, 43, 57]. After connecting the applications with a TENG by the external
wires, the induced free charges can be used for driving the electrons across the
applications. Because the internal voltage is changing with the varying separation
between the contact surfaces, the external voltage is also changing to balance the
internal voltage. Because of the external voltage, the induced free charges flow to form
the current in the circuit [67].
Herein, this thesis concentrates on the vertical contact-separation mode triboelectric
nanogenerator (CS-TENG) as an example to explain its working principles at open
circuit and short circuit conditions, respectively. The working steps applying to the open
circuit voltage and short circuit current condition for a CS-TENG are shown
schematically in Fig. 2.3. Here, we explain the working principle of a contact-
separation TENG (CS-TENG) at open circuit and short circuit conditions step by step:
The open circuit voltage and short circuit current are the significant parameters for
presenting the performance of CS-TENGs [12, 57-59]. There is no free charge on the
two electrodes for the open circuit configuration in Fig. 2.3(a) [12]. The open circuit
voltage is created by the difference in potential between the electrodes as separation
occurs. Simultaneously, these built-in electric fields cause the polarization of the
dielectric material and produce the induced charges on the electrodes [58]. In the short
circuit mode, after the tribo-charges have been generated on the contact surfaces, the
tribo-layers are being polarized to form the bound charges inside of the tribo-layer
materials. Because of the bound charges, there will be induced free charges on two sides
of the electrodes. Neutralizing the negative charge on the bottom electrode, the positive
charges on the top electrode will be conducted to the bottom electrode to form the short
circuit current, as shown in Fig. 2.3(b).
30
Working principles of a CS-TENG is similar to the short circuit condition, when it
connected to applications (e.g., electronics). However, due to connecting the
application to an external circuit, there is a differential potential between two sides of
applications. When the separation is increasing to the amplitude, the internal voltage of
CS-TENG is also increasing, and the top induced free charge is pumped from top to
bottom electrode to balance the internal different potential. Additionally, the output
current from the CS-TENG could be flowing from positive polar to the negative polar
of the application. However, when the separation is reducing from amplitude to zero, it
is equivalent to adding an opposite voltage at two electrodes. For this reason, the
electrons on the bottom electrode are driven back from bottom to the top electrode.
Thus, owing to this periodic vibration by the external applied force, two interfaces are
contacting and separating continuously, and the device is generating alternating current
signal.
2.2.3 The modes of triboelectric nanogenerators
The working principle of the TENG has been introduced with reference to the CS-
TENG, however, there are another three modes of TENG that have been invented to
harvest different modes of vibration effectively, including sliding mode, freestanding
mode and single electrode mode, as shown in Fig. 2.4 [68]. The theoretical module of
TENGs depends on the parallel capacitor module and the concept of displacement
current (i.e. Maxwell’s law) [12, 69, 70]. Note that the current from the TENG
corresponds to the varying of the internal electric field and the tribo-layer polarization
[68]. However, the output voltage for different modes of TENGs related to the device
total capacitance and total surface tribo-charges. Note, device total capacitance can be
calculated by TENGs’ equivalent circuit. We now focus in more detail on the individual
modes.

31
Fig. 2.3: Working principle of contact-mode triboelectric nanogenerators: (a) open circuit voltage (b))
short circuit current.

i. Contact and separation mode triboelectric nanogenerator (CS-TENG)

Contact and separation triboelectric nanogenerators (CS-TENG) couple contact
electrification and electrostatic induction [12, 57]. Fig 2.3 has already explained the
working principles of a CS-TENG with an interface pair of PDMS and PET. After the
tribo-charges are generated on the interfaces, the electric fields inside the TENG device
are shown schematically in Fig. 2.4(a). The total electric field is equal to the sum of the
electric fields inside of the tribo-layers and the air gap electric field [57]. If we assume
that the transferred free charges quantity is Q in Fig 2.4(a), the output voltage could be
calculated by the parallel capacitor module as [57]:
) , , -% .(()
,&'( = − *+ /+ " + + $ + 1(3)5 + +!
(2.1)
! #" #$

where, S is the nominal contact area, d1 and d2 are the thickness of tribo-layers, #%1
and #%2 are the relative permittivity of tribo-layer materials, #$ is the permittivity of
air, x(t) is the separation distance, v(t) is the speed of vibration and !! is tribo-charge
density.

32
Since Q is equal to zero at the open circuit condition, we have
*-% -% .(()
,&3 = 4 = +!
(2.2)
&'(

Depending on the equivalent circuit of a CS-TENG in Fig. 2.4(a), the total capacitance
of the device is deduced from:
1 1 1 1
4)*)+,
=4 +4 +4 (2.3)
" +-# $

Due to the ,&'( = 0 at short circuit condition,

*-% *-% .(()
653 = 4 =, (2.4)
+-# 4)*)+, ! 6.(()

,)./ *- , 7(()
753 = ,(
= (, %6.(())
!
$ (2.5)
!

ii. Lateral sliding mode triboelectric nanogenerator

The working principle of a lateral sliding mode triboelectric nanogenerator (LS-
TENG) has been shown schematically in Fig. 2.4(b), and it couples the
triboelectrification (rubbing interfaces) with electrostatic induction. Comparing with
the CS-TENGs, the interface pair presents relative displacement to replace the
separation step to generate tribo-charges on the material sliding surfaces [71]. The
working principle of a LS-TENG is still based on the parallel capacitor, as shown in
Fig. 2.4(b).
However, it depends on altering the swept area of tribo-layers to form the output signals
by the external applied force [72]. Increasing the swept area, both surface tribo-charges
(Q) and horizontal electric fields of two tribo-layers are rising to form the output voltage
between two side of electrodes, as shown in Fig. 2.4(b) [71, 72]. Therefore, the output
voltage could be calculated by the parallel capacitor module as [71]:
) , , - .(() , ,
,&'( = − 8(9:.(())+ /+ " + + $ 5 + (9:.(())+
%
/+ " + + $ 5 (2.6)
! #" #$ ! #" #$

where, w and l are the width and length of tribo-layer, x(t) is the horizontal relative
displacement, d1 and d2 are the thickness of tribo-layers, #%1 and #%2 are the relative
permittivity of tribo-layer materials, #$ is the permittivity of air, v(t) is the speed of
rubbing and !! is tribo-charges density. The total capacitance of the device can be
calculated from:

33
 1 1 1 +! 8(9:.(())
4)*)+,
=4 +4 = 0" 0$ (2.7)
" $ 6
1#" 1#$

If it is at short circuit condition (,&'( = 0),

6*4 = !! 91(3) (2.8)
-% .(() ," , )23
,&3 = (9:.(())+! +#"
/ + + $5 = 4 (2.9)
#$ )*)+,

,)./ (2.10)
753 = ,(
= !! 9:(3)

(a) (c)

C1
Cair
C2

(b) (d)

C1
C2

Fig. 2.4: Working principle of the four TENG modes of operation and their equivalent circuits [80, 90,

93, 102, 103].

iii. Single electrode contact mode triboelectric nanogenerator

Both CS-TENGs and LS-TENGs all require employing two electrodes to construct a
closed circuit to conduct charges in the TENG-based circuit. However, the single
electrode triboelectric nanogenerator (SE-TENG) overcomes the limits of applications
and increases the diversity of energy harvesting (e.g. [73], [74], [75] and etc.). The SE-
TENG could especially be applied as a human skin based TENG, due to its flexible and
practical power generation model. Therein, it has high potential for use in tactile screen
sensors or generators [76-78]. Fig. 2.4(c) shows the working principles and equivalent
circuit of a SE-TENG schematically. The tribo-layer based on a layer of dielectric was
coated by a layer of conductor as the electrode, and it requires to be connected with a
reference electrode. After the independent tribo-layer makes contact with the bottom
34
 dielectric surface, tribo-charges are generated on the dielectric and independent tribo-
layer material surfaces. Due to the electrical potential based on the tribo-charges, the
electrons flow from the electrode to the reference electrode [79]. When the independent
tribo-layer makes contacts with the bottom interface again, the voltage between Node
#1 and #2 decreases as the x(t) reduces in Fig. 2.4(c). The relationship between the
voltage between Node #1 and #2 (V1), Node #1 and #3 (V2) and Node #2 and #3 (V3) is
[68, 80]:
,2 = ,1 + ,; (2.11)
Both ,2 and ,1 are decreasing, because the x(t) and x(t) + g are reducing. Thus, the
charges are drawn from Electrode 2 back to the Electrode 1. When the device is in the
open circuit condition, the total charges on Node #1, Node #2 and Node #3 are -;!! ,
;!! and 0, respectively. Therefore, depending on the charge conservation, it can be
seen
<2 ,2 + <1 ,1 = −;!! (2.12)

<; ,; − <1 ,1 = ;!! (2.13)

Its open circuit voltage ,; in Fig. 2.4(c) can be calculated by combining Eq. (2.11-
2.13) as:
*-% 4$
,<4 = ,; = 4 (2.14)
" 4$ 64$ 44 64" 44

Because 653 = ,&3 <2,; , the short circuit output charge could be obtained as:
*-% 4$ (2.15)
653 = 4
" 64$

Note that <2,; is the capacitance between Node #2 and Node #3, in which it is equal
to:
4 4
<2,; = <; + 4 "64$ (2.16)
" $

iv. Contact freestanding tribo-layer mode triboelectric nanogenerator

Contact freestanding tribo-layer mode triboelectric nanogenerator (CF-TENG) is
composed of two fixed metal electrodes and one freestanding dielectric layer between
them, as shown in Fig. 2.4(d) [68, 81]. Both sides of the surface of the freestanding
dielectric layer could contact intermittently with the two sides of the electrode
interfaces depending on the external applied forces. When the dielectric layer makes
35
contact with the top metal surface, the top electrode generates tribo-charges; the charges
conducted to the bottom electrode [81]. When the dielectric layer makes contact with
the bottom metal layer, the distance (g - x(t)) between top metal layer and dielectric
layer decreases until they are in contact. The bottom metal surface and dielectric layer
bottom surface transfer tribo-charges and connect to the top electrode [68, 81].
Repeatedly, according to the middle dielectric layer motion between two electrodes, the
alternating current is formed in the external circuit. Fig. 2.4(d) shows the working
principle and equivalent circuit of a CF-TENG schematically. Comparing with the CS-
TENG, the distance between two electrodes is fixed as =. However, due to the varying
of the distance (= − 1(3) and 1(3)) between the independent dielectric layer, the
voltage between two electrodes could be calculated as (assuming the transferred charge
on the electrode is Q) [81]:
0
> " 6?@) 2-% .(()
1"
,&'( = +! *
+ +!
(2.17)
2-% .(()
,&3 = +!
(2.18)

The CF-TENG is equivalent to three capacitors in series; thus, the total capacitance of
the CF-TENG should be:
1
<(&(A9 = " " " (2.19)
6 6
3" 3$ 34

The charge quantity at short circuit stage should be:

2-% *.(()
653 = 0" (2.20)
6?
1!

We have just reviewed the electrical theories of the four TENG modes, which all depend
on the parallel plate capacitor model. However, in reality the plates are not infinite and
the effect of their finite in-plane dimensions must be accounted for [59]. Therefore, a
distance-dependent electric field model (DDEF model) was proposed by Dharmasena
et al, 2017 [59]. In the DDEF model, the variation of the electric field is stimulated
reasonably at different sites in the three-dimensional space firstly, and subsequently, the
total electric field operating in a fixed dimension TENG device is calculated rather than
regarding the triboelectric surfaces as an infinite plate. Integrating the charged micro-
segments along the charged surface could account for the electric field across the tribo-
36
charges charged surface [59, 67]. Therefore, this triboelectric electric field across the
contact surfaces should be corresponding to the separation distance rather than
distributing uniformly in space. According to Dharmasena et al (2017) [59], the electric
field in the vertical position of a surface not only depends on the vertical distance from
the point in the space to the planar surface, but also relates to the dimension of the
contact surface.
5
-% -%
"# = >?@3>A B 6
C= D(E) (2.21)
B+ $7C 7 $ 5 $ B+
D> @ 6> @ 61
6 6 6

where, L and W are the length and width of contact surface. z is the distance from point
to the contact surface, # is the permittivity of medium and !! is the tribo-charge
density.
Combining Eq. (2.21) with the TENG device structure, the potential of two electrodes
can be calculated to investigate the open circuit voltage, short circuit current and output
power even for non-planar devices. Therefore, the parallel capacitor model and the
DDEF model are two fundamental models for prediction of TENG output. These
models can also be used to inform TENG system design.

2.3 Triboelectrification of materials

In Section 2.1.3, we have reviewed theoretical models applicable to the four
triboelectric nanogenerators (TENG) modes. According to these, TENG output
performance depends on a number of factors, including the surface tribo-charge density
[12], separation distance [59], device dimensions [82], permittivity of the tribo-contact
materials [83] and difference in electron affinities of the interface pair [84]. Among
these, the surface tribo-charge density is the key parameter driving TENG performance,
because it is the energy source for a triboelectric device [12]. We have already discussed
how surface tribo-charge generation depends on the contact materials. This is very
much related to the differences in electron affinities (i.e. work function) of the interface
pair [58]. When contact occurs at the interface, the material with high electron affinity
(tribo-negative material) gains electrons from the material with low electron affinity
(tribo-positive material) [57]. After that, the generated negative and positive tribo-

37
charges are fixed on the tribo-negative and tribo-positive material surfaces, respectively.
Thus, it is critical for TENG design to select a pair of contact materials with large
differences in tribo-properties. For the basic materials selection in triboelectric device
design, the triboelectric series is the best reference [12, 85], as shown in Table. 2.2. In
general, there are two positions in the series: positive and negative. Many of the
common polymers are tribo-negative, including Teflon materials (e.g. PTFE and FEP),
PDMS, PI (Kapton) and PVDF [86]. Metals such as gold, aluminum and copper are in
the middle region of the triboelectric series [12]. However, nylon, silk and human skin
are tribo-positive [20]. Generally, researchers adopt a tribo-negative material thin film
(such as Teflon materials, PDMS, PI (Kapton) and PVDF) in contact with a tribo-
positive material (such as metal, nylon or silk, etc.) to fabricate flexible TENGs. We
now review some of the material possibilities in turn.

38
Table. 2.2: Triboelectric series for common materials following a tendency to lose electrons (positive) to
gain of electrons (negative) [12].

(Continue)

39
i. Polytetrafluoroethylene (PTFE) based TENGs
PTFE is a type of synthetic polymer where all the hydrogen atoms in polyethene (PE)
are replaced by fluorine atoms [87]. PTFE is one of the most commonly used tribo-
negative materials in TENGs [88-92]. Simultaneously, because of properties of acid
resistance, alkali resistance and resistance to organic solutions, the application range
for PTFE based TENGs is extended [93]. For example, a Teflon material based single
electrode TENG (SE-TENG) was used to detect the chemical sensing of solutions in
Jiang et al. [94]. The PTFE based SE-TENG was tested on 10 types of ionic standard
solutions, and it showed different sensing output ( ,&3 , 753 and 653 ) with each.
Moreover, due to the super hydrophobic characteristic of PTFE thin films, one can even
harvest water wave or water droplet energy using TENGs. Therein, Yang et al (2018)
[117] used a PTFE based tribo-layer to make contact with DI water, tap water and 0.6M
NaCl solution in a water tank. For the TENG with a pristine PTFE surface, the output
power density could be 5mW/m2. Additionally, the output power density could be
increased to 9.62W/m2, when the hydrophobic property of the PTFE surface was
increased. A high-performance droplet electricity generator with metal/PTFE/ITO
structure has also been designed using the super hydrophobic property of PTFE and the
triboelectricity between the droplet and PTFE [118]. Note that after the droplet
impinged on the PTFE surface, the tribo-charges could be generated by the droplets
spreading, retracting and sliding. Therein, the maximum open circuit voltage and short
circuit current could were approximately 143.5V and 270$F. Therefore, using PTFE to
harvest water energy has potential and requires further investigation.

ii. Polydimethylsiloxane (PDMS) based TENG

Polydimethylsiloxane (PDMS) is a variation of high molecular organic silicon
compound, which is short for organosilicon [95]. Normally, sluggishness, non-toxicty
and uninflammability characterize PDMS [96]. Because of the properties of high
viscosity, workability and ease of sealing with self and other materials, it always been
fabricated as thin films with high surface functionalization [97-99]. For instance,
exploiting the patterned silicon wafer, it is easy for PDMS to define the nanopatterns
40
on its surface. Because of the high surface-to-volume ratio, the contact area of the
interface pair could be increased effectively, especially defining the mating contacting
surface patterns [100]. A stretchable TENG was fabricated with PDMS micro-pyramids
and Al coated PDMS micro-pyramids patterns to increase the effective contact area of
two tribo-layers [93]. Other important properties of PDMS are its transparent and soft
property which could also be applied in the TENG design (especially, using as a tactile
screen sensor). Because of its high output sensitivity and self-power specialties, PDMS
based triboelectric tactile screen sensors have the potential to replace the traditional
electronic tactile screen sensor to reduce the pressure on battery life [101]. Human skin
and PDMS could be combined as a SE-TENG (i.e. the maximum output power density
could be 10mW/cm2) and employed as a tactile screen sensor [102]. Additionally, this
SE-TENG could also be used as a pulse sensor which could perform 2V output voltage
under pulse beating. PDMS is widely employed to design flexible and stretchable
TENGs.

iii. Polyvinylidene fluoride (PVDF) based TENG

Polyvinylidene fluoride (PVDF) is also a material used, for example, in protective films
for glasses, building exteriors and billboards [86]. Simultaneously, as a consequence of
its capability in filtrating impurities (e.g. sediment and germs), PVDF thin films can
also be used for pretreatment of sewage purification and sea water desalination [103].
Furthermore, the separation films, which are inside of many energy storage devices (e.g.
supercapacitor, lithium battery and etc.), are adopted PVDF thin films [104, 105]. In
the case of its application in energy harvesting, PVDF, which is composed of the
repeating –(CH2-CF2)– unit, is becoming one of the more popular polymer materials in
piezoelectric and triboelectric technologies [50, 106]. Because of the all-trans (tttt)
chain conformation for the * phase PVDF and its copolymer, it can generate a
piezoelectric response or piezoelectric energy harvesting by external applied forces
[107]. Therefore, PVDF and its copolymers have drawn tremendous attention as
sensing materials in wearable and flexible sensors. According to the triboelectric series
in Table. 2.2, PVDF is a tribo-negative material [85]. For example, a TENG device in
41
Mi et al, 2018 [108] using PVDF and CNF/PEI as the tribo-negative and positive
materials generated 105V and 4.5$F/cm2 open circuit voltage (Voc) and short circuit
current (Isc), respectively. Interestingly, a pristine * phase PVDF thin film (without
any modifications) contacting Al foil can generate 0.8mW/m2 power density [109].
However, after performing the forward poling process on the PVDF, the output power
could be increased over three times. The output power was decreased 1.3 times by the
reversed poling process on the pristine PVDF thin film. This is because dipole
alignment can be increased by the poling process (from the external electric field) to
increase the material surface potential [109]. Additionally, reversed polarized PVDF
decreased the device performance, because the position of the ferroelectric dipole
moments mismatch with the triboelectric electric field [110]. Owing to this principle,
the piezo/triboelectric nanogenerator has also been designed by inserting one PVDF
layer under the negative tribo-layer to self-align the dipoles inside tribo-layer [111]. As
a results of these, electrospun PVDF and its copolymer have been suggested as the
tribo-negative layer, because of the high surface to volume ratio and surface potential
[112]. Therein, PVDF is also an appropriate electrospinning material, because of its
suitable solution viscosity [113]. Additionally, PVDF films also have other advantages,
including high thermostability, high abrasive resistance (e.g., mechanical strength and
toughness) and high chemical stability (e.g., resisting many chemicals and solvents)
[113].
iv. Polyethylene terephthalate (PET) based TENG
PET is a semi-aromatic polymer that is synthesized by ethylene glycol and terephthalic
acid as well as being one of the most commonly manufactured thermoplastics [114]. Its
melting point and glass transition temperature is 250 to 255 ℃ and 67 to 81 ℃ ,
respectively [115]. Because of its characteristics of excellent abrasion resistance, high
flexibility, creep resistance, high dimensional stability, weak corona resistance and fair
oxygen barrier property, it is applied in industrial applications extensively.
Simultaneously, PET is also a common contact or substrate material selection for use
in flexible TENGs [116-120]. However, the electron affinity of PET is weaker as
compared to the abovementioned polymers [12, 20]. Since it is in the middle region of
42
the triboelectric series, as shown in Table. 2.2, PET could be used as both a tribo-
positive and negative material in the TENG designs [118, 119]. For example, a SE-
TENG composed by a nanowire-structured PET thin film and Cu foil could generate
1000V, 4.3$A and 15W/m2 Voc, Isc and power density by 115L/min air flow [118, 119,
121]. Fan et al (2012) [122] fabricated a high transparent TENG by a regular and
uniform PDMS patterned array based film and ITO coated PET as tribo-negative and
positive materials, respectively. It could generate 18V and 0.13$A/cm2 output voltage
and current density. Because the TENG exhibits a high transmittance of 75%, its
potential application in touchscreen, high-definition LCD and self-powered electronic
displays is obvious.

v. Metals in TENG design

According to the triboelectric series (Table. 2.2), metals are in the tribo-positive region:
examples include aluminum (Al), copper (Cu), nickel (Ni) and gold (Au) [12].
Therefore, aluminum and copper are often employed as the tribo-positive material in
TENG construction. These metals are also often used as electrodes. These can be
prepared by commercial foils or by metallization etc. Furthermore, the foils could also
be used as the substrate to fabricate the tribo-layer [123]. For instance, a solution of
polymers (e.g. PDMS, PVDF, PMMA and PI) can be spin-coated on the foil.
Additionally, the Al foil has been extensively used as the substrate to collect electrospun
nanofibers, because of the conductive property [124]. After forming the nanofibers
based thin film, the Al foil can then be used as the electrode of the tribo-layer [50].

In the previous sections, we have discussed fundamental principles, including the

principle of power generation, modes of TENG operation, equivalent circuits and
important triboelectric materials. The basic common parameters, which influence the
output performance of TENGs, include surface tribo-charge density, permittivity,
separation distance, sliding distance, thickness of tribo-layers, device dimensions,
operation frequency and contact area. These parameters have been demonstrated as
effective methods to improve the performance of TENGs, as shown in Table 2.3. A
43
detailed parametric study would be required to determine a rank order of the sensitivity
of TENG output to the different parameters. Comparing with the dimensional and
operational parameters in a TENG device (see Table 2.3), it could observe gains of 5.8
and 4.9 for VOC and ISC, respectively as real contact area (Ar) was increased from 0.25
to 82% (present work). It is less than the gains of 50 and 8.5 reported in [59] for VOC
and ISC, respectively, as separation distance was varied from 0.1 to 1 mm. However, for
example, the VOC gain of 5.8 is comparable with VOC gains of 2.6 to 8 reported with
changes in tribo-layer thickness (0.02 to 2 mm) in [125, 126], while the ISC gain of 4.9
is of a similar order to ISC gains of 2.4 to 6 reported with change in frequency (0.5 to 5
Hz) in [59, 127]. Considering optimization via the interface materials, the tribo-charge
density, which is key to power generation, (Table. 2.2) depends on the strength of the
triboelectric effect generated by a particular interface pair [12]. Additionally, the major
parameter used to quantify the surface tribo-charge generation performance for the
contact materials is the material’s surface potential [109, 128, 129]. Comparing with
the other dimensional and operational parameters in TENGs, gains of 4 and 4.5 for VOC
and ISC, respectively, when P(VDF-TrFE)’s surface potential was increased from -0.65
to -5.15V can be observed. It is approximate to the gains from Ar reported in the present
thesis. Thus, increasing the materials surface potential is another significantly effective
parameter to enhance CS-TENG performance. For example, after increasing the
crystallinity of P(VDF-TrFE) in [133], its surface potential is enhanced from -0.87 to -
2.37V and VOC and ISC of the device were optimized 1.7 and 1.8 times, respectively.
Consequently, enhancement of the triboelectric performance of an interface surface pair
via optimization of real contact area, surface potential and permittivity are promising
routes to boosting TENG output.

44
Table 2.3: Literature comparison of TENG factors and parameters effecting electrical output.
Voc Gain Isc Gain
Frequency (f) Hz
0.5 to 5 (PET & PDMS [59]) - 6
1 to 4 (Al & FEB [127]) 1 2.4
1 to 5 (Cu & FEP [130]) 1 3.6
0.3 to 3 (Nylon & PTFE [131]) 1 -

Separation distance (d) mm

0.1 to 1 (PET & PDMS [59]) 50 8.5
1 to 20 (Al & PTFE [125]) 9 2.5

Contact Material Permittivity (ε - relative)

7.5 to 20 (PA6 & PVDF-BTO [126]) 1.2 1.4

Thickness of Tribo-Layer (t) µm

200 to 1000 (Nylon & PTFE [131])
t for PTFE 2.8 -
32 to 220 (PDMS & PTFE [132])
t for PDMS 7 2.2
300 to 2000 (Al & PTFE [125])
t for PTFE 8 -
20 to 100 (PA6 & PVDF-BTO [126])
t for PVDF-BTO 2.6 1.6

Real Contact Area Ratio (Ar) %

0.25 to 82 (Cu & PET) [133] 5.8 4.9

Surface potential
Al & (-0.65 to -5.15V) P(VDF-TrFE) [128] 4 4.5
Al & (-0.87 to -2.37V) P(VDF-TrFE) [109] 1.7 1.8
Al & (-240 to -540mV) PVDF/AgNW [129] 1.7 1.7

Crystallinity (FWHM)
PET & (0.95 TO 0.81) P(VDF-TrFE)/BTO [50] 1.4 3.4
Al & (0.79 to 0.7) P(VDF-TrFE) [109] 1.7 1.8

45
2.4 Research progress on triboelectric nanogenerators
i. Real contact area
Surface roughness is known to affect aspects such as the frictional response of surfaces
[134], the failure behavior of materials [135] and the contact stiffness of interfaces [136].
The real contact area between two rough surfaces is generally much lower than the
idealized nominal contact area between two perfectly flat surfaces – this is an
established result in the field of Tribology. What has been observed in the TENG
literature is that CS-TENG electrical output is generally contact force dependent and
generally increases as the force pressing the tribo-contact surfaces together increases
[137-143]. Some researchers have reported results with electrical output (e.g., Voc and
Isc) varying linearly with force [139, 141], while in others, a later-stage saturation occurs
at higher forces [137, 138, 141, 142]. However, a link between contact area (mostly
studied by tribologists) and electrical output in TENGs has never really been proposed
and certainly has not been studied experimentally.
One might presume that charge transfer might increase with contact area since contact
within the interatomic repulsive regime is required for electron transfer [63, 144]. In
order to demonstrate this mechanism of surface tribo-charge transfer, material surface
potential has been measured by Kelvin probe force microscopy (KPFM) between the
centre of the scanned area and the edge after a KPFM scan [144]. Based on a parallel
capacitor assumption, the tribo-charge induced dielectric surface potential change can
be associated with the surface tribo-charge density as [145]:
-% (0-8,8/)#-/
∆, = +! +0-8,8/)#-/
(2.22)

)% E#
!! = (2.23)
E9

QT is the quantity of surface tribo-charges. Therefore, variation of surface tribo-charge

generation could be predicted by investigating the material surface potential after
combining Eq. 23 [133]. Li et al (2016) [144] extended the experiments of Zhou et al
(2013) [145] by changing the phase shift of the KPFM probe vibration (from attractive
to repulsive force regime). Depending on their experimental results, the surface
potential change could be detected when the tip was vibrating at repulsive regime.

46
Inversely, there was no surface potential identified in the attractive regime. Therefore,
these results demonstrated the surface tribo-charge generation is limited to areas of
solid contact (i.e. real contact areas).
The influence of contact pressure on TENG output has been investigated by a TENG
device consisting of a pyramid patterned PDMS surface (negative tribo-layer) in contact
with a flat Ag surface (positive tribo-layer) in Fig. 2.5 (a) [137]. Typical increase in the
open circuit voltage, with load, up to the point of saturation (Fig. 2.5 (b)) was shown
together with microscopic images of the PDMS pyramids in contact with a glass slide
which qualitatively indicated a corresponding growth in contact area (i.e. size of the
square contact patches). From the simulation and experimental results, it can be seen
that only point contacts at the peaks of the pyramids occur in the low contact pressure
range (e.g. from 0.1 to 1 kPa). Because the PDMS based pyramids were deforming with
increasing contact pressure, the contact area between glass slide and PDMS layer are
approaching to full-contact status by the time the high contact pressure is reached.
According to its electrical output (i.e. Voc), it increases close to linear between 10 to
100 kPa and starts to saturate after 100kPa (Fig. 2.5(b)). The relationship between Voc
versus contact pressure has been predicted numerically in Fig. 2.5(c). Note, the
transition point from the qualitative partial contact (PC) to the qualitiative full contact
(FC) condition matches with the transition point to Voc saturation. However these results
were somewhat qualitative – contact area was not measured and no relationship
between the TENG outputs and quantitiative measurement of contact area is evident in
the contribution of Seol et al. [137].

47
 (a) (b) (c)

(d) (e) (f)

(g) (h) (i)

PET

PDMS

Fig. 2.5: Studies on the effect of contact area in CS-TENGs [160，171，172].

Yang et al (2018) [146] modeled the effect of different patterned surface types such as
pyramids etc. (in contact with a flat surface) on the evolution of both contact area and
electrical output (and included the effect of adhesion) – Fig. 2.5(d-f). The deformation
of the pyramid and cone patterns are increased by rising the contact pressure, however,
there is an extremely low deformation of the cube and cylinder patterns by the
increasing contact pressure in Fig. 2.5(d). In addition, the variation of the contact areas
(patterned surface in contact with the same flat surface) were calculated at different
applied contact forces in Fig. 2.5(e). Especially, the effective contact area ratio
increased at the low applied forces range (≤400 N) and saturated after applied forces
≥400N for the pyramid and cone patterned surfaces. Oppositely, the deformations of
the cube and cylinder patterns are roughly constant, and their contact area ratios are
roughly constant in Fig. 2.5(e) as might be expected from the geometry. Note, the
contact area ratio of the cone patterned film is lower than for pyramid patterned film.
The electrical output results for Voc of the CS-TENGs with pyramid and cone patterns

48
predicted the same changes with contact force as the effective contact area, which
increase and later saturate as the surface pattern is flattened in Fig. 2.5(f). However, the
Voc for the CS-TENGs with the cube and cylinder patterns remain roughly constant –
again in line with how the contact area varies for these geometries. Because the
parameters of Voc in Eq. (2.2) (e.g. the separation distance, device dimension and air
permittivity) are constant, the variation of the Voc is related to the change of the effective
contact area ratio. However, this work was numerical and only considered structured
surfaces with certain geometries.
A recent load dependent model (LDM) by the present authors in Xu et al. (2020) [147]
has shown how accounting for a contact force-dependent real contact area (arising from
random surface roughness) produces the commonly observed force-dependent TENG
electrical performance, depending on the DDEF mode of CS-TENG in Fig. 2.5(g-i).
Persson’s contact theory is used to determine the real contact area of the interface pair
and is therefore applicable to contact from first touch up to nearly complete contact (i.e.
nearly 100% contact) as long as deformation remains elastic. Fig. 2.5(h) presents the
contact area fraction as increasing linearly with increasing contact pressure from 0 to
about 0.4MPa and saturating gradually after a contact pressure of over 0.4MPa. The
DDEF model considered the generated tribo-charges distributed on a constant,
unchanging nominal area, but the load-dependent model [147] essentially adapts the
DDEF to a situation where the contact area evolves with contact pressure. The
simulation study is carried out to study the influence of surface roughness (and real
contact area) on the output voltage and short-circuit current of TENGs under different
pressures. Comparing the experimental and simulation results, the simulated results
from load-dependent model is more representative of the experimental results than the
simulated results from DDEF model in Fig. 2.5(i). The important aspect being that the
predicted voltage is now contact force dependent. These previous papers contributed to
qualitative understanding enhanced by theoretical and numerical predictions. However,
no quantitative experimental contact area measurements were plotted in any of these
studies [137, 143, 146] and the experimental study by Seol et al (2015) [137] was based
on qualitative observations of the deformation of idealized highly engineered (pyramid-
49
on-flat) type contacts. Therefore, the role of contact area still remains to be
demonstrated experimentally via measurement and calculation of real contact. Crucially,
this kind of experimental study needs to be applied to the general class of non-
engineered nominally flat rough surfaces (i.e. all non-patterned surfaces).
Chapter IV of this thesis resolves this issue by accurately measuring both TENG real
contact area and electrical output simultaneously over a wide range of contact pressures
(from first touch to heavily loaded) for two nominally flat rough surface tribo-contact
pairs. All other parameters are fixed so that just the effect of the contact force and area
can be elucidated. Experiments are carried out where electrical output (such as Voc &
Isc) were measured together with Ar under increasing contact pressure (or force) with
identical test conditions for a TENG in contact separation mode. Fabrication of
nano/micropatterns on the contact surface has been considered extensively as an
approach to enhance the real contact area of the interface pair such as pyramids, line
patterns, nano-helixes and nanofibers [148, 149] – although many of these approaches
do not succeed in enhancing contact area very effectively (primarily because of the gaps
between the features). This aim of increasing contact area leads partly to the next topic:
that of a TENG with electrospun PVDF-TrFE composite nanofibers. Here, there is also
a second major optimizing factor – permanent material polarization made possible as
PVDF-TrFE is ferroelectric. Hence, the next section reviews the literature related to the
work in Chapter V on a ferroelectric-assisted high performance triboelectric
nanogenerator based on electrospun PVDF-TrFE composite nanofibers with BaTiO3
nanofillers.

ii. Electrospun ferroelectric assisted triboelectric nanogenerator

Many applications place very particular demands on TENG materials such as: high
flexibility, ability to maximize electrical output and robustness in sustaining high
mechanical stress/strain [150-154]. In terms of maximizing electrical output, a tribo-
contact material pair that maximizes change transfer is required. Four key factors that
help optimize this are maximizing: (real) contact area [133, 147], difference in electron
affinities of the interface pair [12], permittivity of the tribocontact materials [99, 155,

50
156] and difference in residual surface potential between the interface surfaces [128,
157]. With these demands in mind, materials that have received particular attention are
the ferroelectric (FE) polymers polyvinylidene (PVDF) and copolymer poly(vinylidene
fluoride-co-trifluoroethylene) (P(VDF-TrFE)) as these are flexible, but also allow for
permanent polarization of the tribo-contact surfaces and are piezoelectric.
In Section 2.2, this manuscript has already discussed PVDF (with its repeating –
(CH2-CF2)– unit molecular structure). PVDF is one of the advantageous tribo-negative
materials, because it is a fluoride. Note, fluorine has the highest electron affinity among
the atoms to increase material triboelectric property [158]. Its co-polymer P(VDF-TrFE)
combines the trifluoroethylamine structure into the PVDF molecular structure (i.e. –
(CH2-CF2) m–(CFH-CF2) n–). In addition, the polar *-phase is crystallized from the
melt form of P(VDF-TrFE) readily comparing with PVDF [109]. According to Nalwa
(1995) [159], crystallinity of P(VDF-TrFE) up to 90% can be achieved by pretreatment.
Simultaneously, the high crystallinity of PVDF-TrFE can increase its negative tribo-
property for the application of PVDF-TrFE in TENGs as well. It has been demonstrated
to promote higher crystallinity and better dipole alignment similarly to how high dipole
moment solvents have been shown to increase crystallinity and dipole alignment for
P(VDF-TrFE) [126]. As expected, a high triboelectric output performance corresponds
to the TENG device with highest crystallinity P(VDF-TrFE) thin film. The higher
crystallinity results in a higher dipole alignment of P(VDF-TrFE), and the presence of
aligned dipoles in the β-phase of ferroelectric (FE) polymers results in the enhancement
of the spontaneous polarization (i.e. higher surface potential) and hence the surface
charge density [160]. Thus, enhancing the crystallinity of the FE polymer is an effective
method to optimize the ferroelectric assisted TENG performance. Comparing with
PVDF, P(VDF-TrFE) has been selected as the tribo-negative polymer in Chapter V,
because PVDF-TrFE could be fabricated at higher crystallinity than PVDF under the
same conditions.
Additionally, recent work has shown significant improvements (1 to 5 times) in
TENG electrical output when polarized PVDF or P(VDF-TrFE) are compared to their
non-polarized equivalent [66, 160, 161]. Especially, Lee et al (2016) [162] analyzed the
51
influence form the polarization of P(VDF-TrFE) thin film for the TENG devices,
including forward, reverse and non-polarized situations. The experiments not only
explained the effect from surface potential for a TENG device, but of most interest is
the tribo-property changed by the direction of the polarization. From their results, the
highest occupied energy level of electrons of original P(VDF-TrFE) was shifted to the
lower or higher level by forward or reverse-polarizing its thin film. After that,
differences of work function between the contacting materials are enhanced and this
boosts the tribo-charge transfer between the interface pair surfaces. The poling process
is the general method to increase the β-phase formation and crystallinity of P(VDF-
TrFE) [163]. However, to promote the β-phase formation and crystallinity,
electrospinning also could be used as it combines uniaxial stretching and electric field
poling in a single step [133, 147, 164]. The jet initiation could not happen for the
viscous PVDF-TrFE solution until applying the high voltage during the electrospinning
fabrication, because the electrostatic force is inadequate to overcome the surface
tension of the PVDF-TrFE solution droplet and shoot out the droplet from the nozzle
[165-167]. It is clear that the high voltage can increase the stretching force, because it
is observed that the diameter of the nanofiber decreases with increasing voltage [165].
Therefore, the PVDF-TrFE molecular chains dipole is oriented as the stretching of a
polymer jet under the high applied voltages during the electrospinning fabrication to
shift it from α to β-phase [166, 167]. For example, experimental results have shown that
the PVDF β-phase content could be increased from 67.8% to 72.4% by increasing the
applied voltage from 12kV to 25kV [168]. Simultaneously, the degree of crystallinity
of the PVDF is enhanced from 48% to 56.3% corresponding to a growth of β-phase
content from 70% to 76%, after applied voltage increases from 14kV to 20kV [169].
Hence, the high applied voltage can increase both the β-phase content and degree
crystallinity of the PVDF-TrFE. Additionally, it has been reported that the real contact
area could be increased by optimisation of the nanofiber surface shape [153]. Film
surfaces include roughness and microscale defects and the nanofibers (from
electrospinning) can potentially match the contours of these during contact. Due to the
high surface-to-volume ratio of the interface pair, the contact area of the nanofiber
52
based TENG device could potentially be improved, compared with their spin-coated
thin film-based equivalent. Therefore, electrospinning seems an attractive method to
fabricate a high performance P(VDF-TrFE) nano-fibourous TENG. This has not been
done before and is the subject of Chapter V.
Third, different material optimization routes have been followed to enhance the
formation of the β-phase of FE polymers. For instance, it has been shown that addition
of filler materials such as metal nanostructures, and ceramic nanoparticles into the FE
polymer matrix can enhance the formation of a highly polar crystalline β-phase, because
of the electrostatic interactions between the nanostructures and the highly aligned
dipoles of the FE polymer chains [126, 128, 129, 170, 171]. The introduction of
nanoparticles has another very important advantage i.e. high permittivity inorganic
fillers can increase the dielectric constant of the tribo-contact materials, which serves
to increase the TENG capacitance and hence the ability to store charge [126, 128, 155,
163]. As an example, by dispersing particles of tetragonal barium titanate (TBTO with
formula BaTiO3) in a PVDF matrix, an increase in open circuit voltage and short circuit
current density of spin coated TENGs by 1.5 to 2.4 and 1.8 to 2.3 times, respectively
has been reported [126, 128] (Table 2.4). Similarly, voltage and current increases of 1.6
and 1.5 times were observed with a similar dispersion of TBTO particles in a P(VDF-
TrFE) matrix [128]. However, a significant observation from Seung et al [163] reports
that the output performance of the P(VDF-TrFE)/TBTO thin film based TENG device
could not be improved much without the poling process. This is because the more
random dipole alignment decreases the intensity of the material intrinsic electric field
and even the material surface potential. The electrospinning fabrication approach holds
the potential to resolve this problem, because a high voltage can be applied during the
fabrication. Considering that TENGs produce high voltages but low currents, the
current density is a vital parameter for performance. If tribo-charge density increases
with real contact area [133, 147], significant gains over the state of the art are potentially
possible if above materials are used in fibrous form. In this regard, electrospinning can
provide better outcomes than the methods such as spin coating used in previous works.
Further, electrospinning could enable even more highly charged P(VDF-TrFE)/TBTO
53
 (nanofiber (NF) based) surfaces capable of further boosting current density. The
premise is that electrospinning the P(VDF- TrFE)/TBTO surface will lead to higher
dipole alignment and crystallinity in the P(VDF-TrFE)/TBTO matrix (due to higher
voltage and extra alignment due to fiber stretching.

Table 2.4: Typical PVDF based TENG performances compared with the present work (final row)
Contact materials Fabrication Category Pressure Freq. Voc Jsc Voc/Pressure Jsc/Pressure Sample size Ref.
! 2
Method (kPa) (Hz) (V) (µ"/$% ) (V/kPa) (µ"/&) cm

PVDF & Al Electrospinning Tribo - 2 94.3 0.28 - - 25 [172]

PVDF-HFP & Silicon Electrospinning Tribo+piezo 20 1 17 0.06 0.85 0.03 4 [173]
Rubber
PVDF-AgNWs & Nylon Electrospinning Tribo+Piezo 12.5 3 240 3.0 19.2 2.46 4 [174]
PVDF & CNF/PEI aerogel Electrospinning Tribo 30 10 105 4.5 3.5 1.5 2 [108]
PVDF & Silk fibroin Electrospinning Tribo+piezo 32.1 4 500 1.5 15.6 0.47 8 [175]

PVDF/MXene & PA6 Electrospinning Tribo+piezo 25 1 50 2.5 2 1 4 [176]

125 4 300 15 2.4 1.2

PVDF/Graphene & PA6 Electrospinning Tribo+piezo 25 5 1373 14.1 54.9 5.64 4 [177]
225 5 1720 21.3 7.6 0.95 4

PVDF/MXene & Nylon 11 Electrospinning Tribo+piezo 23 4 75 3 3.3 1.3 3 [178]

PDMS/BTO & Al Spin-coating Tribo+piezo - 2 375 0.3 - - 16 [179]
PVDF & PVA Spin-coating Tribo 15 5 230 1.5 15.3 1 4 [180]
P(VDF-TrFE) poled & Al Spin-coating Tribo+Ferro 380 5 10 4.0 0.03 0.11 1 [181]
P(VDF-TrFE) & PET Spin-coating Tribo+piezo - 0.5 35 0.6 - - 1 [182]

P(VDF-TrFE)/BTO poled & Spin-coating Tribo+ferro 98 - 45.7 1.25 0.47 0.13 1 [128]
Al

BTO
PVDF-BTO & PA6 Spin-coating Tribo+piezo 250 5 450 3.2 1.8 0.13 4 [126]
P(VDF-TrFE)/BTO & PET Electrospinning Tribo+ferro 12.8 4 315 6.7 24.6 5.23 6.25 [50]

Therefore, in this work, we develop high-performance hybrid-TENGs using

electrospun fibrous tribo-negative surfaces based on P(VDF-TrFE) composites with
dispersed BaTiO3 (BTO) nanofillers in Chapter V. This is the first time that a nano
fibrous P(VDF-TrFE)-based TENG has been developed. We also explore the use of
cubic BTO (CBTO) as a means of probing the differences introduced by the
piezoelectric TBTO as compared to the non-piezoelectric CBTO. Accordingly, three
types of β-phase crystalline P(VDF-TrFE) NF-based films are fabricated (via
electrospinning) as active negative TE layers: pristine P(VDF-TrFE), P(VDF-

54
TrFE)/CBTO and P(VDF-TrFE)/TBTO. PET sheet material is used as the positive
tribo-layer in all three cases to complete the Fe-TENGs. Table 2.4 compares outputs
from the studies discussed above and also includes the present work which exhibits the
highest current density even though the contact pressure is the lowest – suggesting a
significant boosting of current density via the approach in Chapter V.
iii. Self-powered/charging electronics/system based on triboelectric
nanogenerator
TENGs have emerged recently as effective self-powered sensors: examples include
physical sensors (e.g., dynamic pressure sensor, speed sensor, acceleration sensor, fluid
sensor and position sensor) and biomedical or healthcare monitoring sensors [48, 183-
191]. Firstly, a TENG is a displacement dependent energy device whether it be sliding
or vertical separation distance. Because TENG output increases with displacement of
the tribo-surfaces, there is a sufficient reason to use TENGs to detect motion. Fig.2.6(a)
presents a work on multi-dimensional self-powered velocity and trajectory tracking
sensors (VTTS) which are composed of a sliding mode TENG array and a LED array
[184]. The moved location of an object can be regarded as it moves in a plane
rectangular coordinate system on a 2D plane, as characterized by corresponding
coordinates, (x, y). Since TENGs can charge the LED array, the location of the object
can be read out from the LED array with a resolution of 250$m. This multi-dimensional
VTTS can respond to the limitation of the test motion from single-dimensional motion
sensor advantageously. Many modern industrial applications also require measuring
speed effectively, including for transportation, airport baggage carousels and for robotic
automation etc. Simultaneously, the motion or vibration speed (i.e. frequency) of
TENGs is one of the important parameters governing their output [12]. For instance,
Jing et al (2014) [185] reported a self-powered velocity dual-mode sensor to detect both
rectilinear and rotary motion, depending on the variation of operational speed - see
Fig.2.6(b).
Self-powered acceleration sensors based on TENGs have the potential to be applied to
earthquake monitoring, vibration monitoring such as global position systems,
biomedical devices, vehicle safety, intelligent electronic products and mechanical
55
equipment vibration monitoring and troubleshooting. A high performance self-powered
acceleration sensor based on a TENG has been fabricated by a PVDF nanofibers-based
film and a liquid Hg ball for vibration monitoring with a high sensitivity of 0.26 V∙s/m2
by Zhang et al (2017) [190], as shown in Fig. 2.6(c). We say in Section 2.2.2 that
electrical output has mostly been observed to be pressure dependent. Hence, a pressure
sensor is the one of the basic applications of the CS-TENG. However, due to its pulse
AC output signal, the CS-TENG is almost designed as dynamic pressure sensor,
especially when applied as a wide range pressure sensor. Because the electrical output
(e.g. open circuit voltage) of a CS-TENG is always saturated at high contact pressure,
as shown in Fig. 2.5(b), there is a sufficient reason to use it as a wide range self-powered
pressure sensor. For example, the electrical output (e.g., output voltage) shows linear
variation in different contact pressures ranges (e.g., 0 to 5kPa, 5 to 60kPa and 60 to
600kPa), as shown in Fig.2.6(d) [192]. Another advanced TENG sensing application is
to detect the characteristics of fluids, including gas and liquid flow [186]. Based on
converting the mechanical energy from the fluid, the fluid characteristics (e.g., pressure
& flow speed) can be recorded by the TENG (e.g. rotating TENG) effectively [191].
Additionally, wave energy and sensing also could be detected by a single electrode
TENG due to the hydrophobic feature of the triboelectric material surface [188]. A self-
powered water wave sensor was designed by Xu et al (2019) [189] with an accuracy of
10 mm, as shown in Fig 2.6(e). Self-powered sensors based on TENGs could also be
applied in the field of smart medicine and biomedical engineering [193]. The basic
parameters in these applications often depend on pressure and frequency such as heart
and pulse monitoring [187]. Additionally, a textile TENG (T-TENG) was integrated
with a sock to detect the gait information of Parkinson's disease patients [48]. This
design also depends on the walking frequency of the patients and these are different at
normal walking, loss of stride and freezing of gait. Collection of this data could allow
medics to monitor patients in real-time, as depicted in Fig 2.6(f). Moreover, in the field
of biomedical engineering, a PTFE based single electrode TENG was integrated into a
medical infusion tube, and the triboelectric output signal can then represent the flow
velocity of liquid and gas produced from the injector [183]. This can demonstrate the
56
monitoring of the transfusion process for a patient. In this section, the self-powered
sensors based on TENGs have been reviewed from the prospective of physical sensors
and biomedical or healthcare monitoring.
Due to their extensive design in orientation of object, patient monitoring, vibration
monitoring, wide range pressure and speed sensing, the potential of the TENG-based
self-powered sensor applications have been highlighted. In theory, expanded Maxwell
equation demonstrated the reason why TENG-based self-powered sensors can actively
detect the static and dynamic processes arising from mechanical agitation by TENGs
output electric signals [194]. Note that the contact surfaces is deforming via the shear
FG
stresses between the interface pair. The second term (i.e., F(
) of the displacement

current density (LH ) fuction is the current owing to the movement of the changed media
FG
as driven by an external mechanical agitation/force. In Wang’s calculation [195] for F(
,

the time variation of the surface charge density and the movement speed of the medium
were import into the final results. The reults also indicates the variation of the medium
volume and surface are also influencing the TENG outputs. It also illuminates the
potential of CS-TENGs to be a self-powered pressure sensore, because the shear
stresses deformed the surface shape (i.e., the real contact area of the interface pair). This
thesis builds on findings about the effect of contact pressure in Chapter IV to develop
a wide rage pressure sensor for various novel sensing applications in Section VI.

57
 (a) (b)

(c)
(d)

(e)
(f)

Fig. 2.6: A variety of applications of TENG based self-powered sensors, including: (a) self-powered
velocity and trajectory tracking sensors (VTTS), (b) self-powered velocity dual-mode sensor, (c) self-
powered acceleration sensor, (d) a wide range self-powered pressure sensor, (e) self-powered water wave
sensor and (f) health monitor [47, 208-216].

Having explored TENGs as sensors, another important application for TENGs is as a

power supplier to electronics or energy storage devices. Therein, charging energy
storage should be one of the major applications for next generation TENG-based self-
charging systems, because these self-charging systems can facilitate 24-hour powering
by nanogenerators such as TENGs. Owing to the high voltage and low current
characteristics of TENGs, one of the challenges for this TENGs-based self-charging
system is reducing the charging time of the energy storage devices [196, 197].
Supercapacitors are the general selection in self-charging systems, due to their high-
power density (i.e. lower the charging time), as shown in Fig. 2.7. Simultaneously, the
TENG outputs are AC power and must be converted to DC power. Hence, a rectifier
circuit is the most common method to integrate the TENG and supercapacitor [197]. In

58
order to improve the charging performance, the straightforward method is to increase
the output power of the TENG. Two TENGs were integrated at the top and bottom side
of one supercapacitor to enhance the input power, as shown in Fig. 2.7(a-b) [197].
Comparing with the system integrating one TENG and a supercapacitor, the stored
voltage of the supercapacitor with two TENGs was increased 2 times with an 8s
charging time in Fig. 2.7(b). The second challenge for the self-charging system is the
integration between the TENGs and supercapacitors. Lou et al (2015) [196] inserted 4
microampere supercapacitors onto the TENG substrate in series as a single device.
Since they are connected in series, the input currents are the same for each
supercapacitor. Thus, although the required current for charging the supercapacitor
array is not increased compared with a single supercapacitor, the output voltage from
the supercapacitor array is enhanced four times, as shown in Fig. 2.7(c) [196]. This
optimization not only decreased the input current for the supercapacitor array and
improved the output performance of the supercapacitor array, but TENGs and
supercapacitors could be integrated as a single device. By this method, the low input
current from TENGs can be answered commendably. Finally, adopting the same
materials for both TENG tribo-layer and the package of supercapacitor is another
challenge, because this could reduce the complexity of the final integrated system, as
shown in Fig. 2.7(c-d). Thus, it can decrease the number of the system layers and boost
the range of device applications - a biodegradable self-charging system was fabricated
by Shi et al (2019) [198]. Both supercapacitor package and TENG substrate used
cellulose paper, and the interface pair of the TENG were nitrocellulose and
PPy/cellulose paper. This design could combine the TENG substrate and supercapacitor
package as one layer and extend the application range of this self-charging system due
to its biodegradable property. However, the recent designs use an external circuit (or
PCB) to combine TENGs and supercapacitors. Therefore, this is one of the key
challenges for future work.
In this work, we successfully integrate the Fe-TENGs (P(VDF-TrFE)/TBTO) with a
graphene electrode-based flexible supercapacitor to realize a self-charging system that
can charge to 1.25 V in just 5 mins (see the application section of Chapter V).
59
 (a) (c)

(b)

(d)

Fig. 2.7: Types of TENG-supercapacitor-based self-charging systems (a) sandwich shape self-charging
power unit (SCPU), (b) output comparison of one TENG based SCPU and two TENG based SCPUs (c)
self-charging screen printed micro-supercapacitor power unit and (d) biodegradable SCPU [219-221].

In the previous sections, the methods of optimizing TENG performance have been
discussed, including enlarging (real) contact area, difference in electron affinities of the
interface pair, permittivity of the tribocontact materials and difference in residual
surface potential between the interface surfaces. Additionally, as we have just seen,
TENGs have also been developed as self-powered sensors (e.g. acceleration sensor,
dynamic pressure sensor, velocity sensor, trajectory tracking sensor and etc.) due to its
high electric output and sensitivity. Examples are the TENG based self-powered
pressure sensors in the sole of humanoid robots or in the shoes of a person with walking
problems. Another reason why we need to optimize the performance of TENGs is the
current gap between their power output (∼0.1- 100 mW/cm2) [54] and the power
needed to develop self-powered systems for charging sensors and electronics (8.44-
107mW/m2) [199]. Generally, TENGs can generate electrical power from the ambient
energy and use it to charge the sensors, sensor arrays and electronics. Additionally,

60
TENG generated power can also be channeled into energy storage devices such as
supercapacitors and lithium batteries to realize 24-hour power supply to key
applications. For example, the self-powered charging system with a ferroelectric
assisted TENG and graphene electrode supercapacitor is sufficient for wearable
applications as often low power sensors (nanowatt to microwatt power range [200-203])
need to be powered. Thus, the two major applications of TENGs are self-power
electronics (system) and self-power/charging system.

2.5 Summary
This Chapter has presented an overview of TENGs and their applications. It has also
explored key areas that are likely to be critical in leading to significant optimization of
output. TENG output is still low when compared to many device power requirements
and the need for optimization is clear. The chapter discussed the principle of operation
of the four TENGs modes, TENG materials and triboelectric energy generation. Key
areas with promise for significant optimization were identified. The identified areas are:
real contact area, difference in electron affinities of the interface pair, permittivity of
the tribocontact materials and difference in residual surface potential between the
interface surfaces.
It was noted that reducing the permittivity of the tribo-contact materials would be likely
to increase electric field strength and boost output. The problem is that low permittivity
and optimized tribo-electric surface properties (tribo-positivity or tribo-negativity) do
not necessarily coincide. An approach to finding an optimized solution to this problem
is identified in Chapter III. The next route towards optimization emerged from results
in the literature indicating that TENG output increases with contact pressure. There are
some qualitative observations suggesting that a contact force-dependent real contact
area might be the origin of this behavior. However, concurrent measurements of TENG
electrical output and real contact area have never been made. Nor has this phenomenon
been studied for standard engineering surfaces possessing random surface roughness.
Hence, Chapter IV explores the fundamental origins of this behavior and shows how
output can be optimized via contact area. The next impetus for optimization emerged

61
from the body of literature suggesting how permanent polarization of ferrelectrric tribo-
contact materials can be used to boost output. P(VDF-TrFE) with BaTiO3 (BTO)
nanofillers was identified as a suitable high performance ferroelectric material.
Although PVDF/BTO and P(VDF-TrFE)/BTO have been used previously, this has
been to produce a film based spin-coated TENG surface. What has not been explored
is how much output can be boosted if high voltage electrospinning is used to create a
nano-fiborous P(VDF-TrFE)/BTO TENG surface. Furthermore, the enhancing effects
of the BTO nanoparticles has not been adequately explained. Previous studies have
identified the benefit of their high permittivity, but have not explored their effect on the
amount and degree of crystallinity of the β-phase in P(VDF-TrFE). These aspects are
addressed in Chapter V which develops high-performance ferroelectric-assisted
TENGs using electrospun fibrous surfaces based on P(VDF-TrFE) with dispersed BTO
nanofillers in either cubic (CBTO) or tetragonal (TBTO) form.
The need to properly apply these high-performance TENGs in applications was also
touched upon in the literature review. This means integrating with suitable energy
storage and also designing novel TENG based sensors. An overview of the advances in
TENG-based self-powered/charging electronics/system applications and its future
challenges have been presented. In the present work, the advances in Chapters IV and
V lead to the rapid charging TENG based super-capacitor system developed at the end
of Chapter V and the wide-range pressure sensors developed and adapted to key sensing
applications in Chapter VI. In conclusion, despite the plethora of challenges in
optimizing TENG performance, high performance TENGs hold great promise for
advances in many areas where renewable energy, self-powered sensors or self-charging
system are needed. The review hints at a number of approaches that might be taken to
further boosting output and these are explored in the subsequent research chapters.

62
Chapter III.

Modelling, simulation and optimization for contact and separation

mode triboelectric nanogenerators

3.1 Introduction
The triboelectric nanogenerator (TENG) has emerged as an exciting new technology
for localized electricity production. It simply utilizes the mechanical rubbing of certain
surface pairs to produce electricity and has applications in harvesting energy from
several sources [12]. The first significant paper on the topic of contact and separation
triboelectric nanogenerator (CS-TENG) was published as recently as 2012 by Fan et al
[66]. The power generation of CS-TENGs depends on the principle of contact
electrification and electrostatic induction. The tribo-charge density, which is key to
power generation, as well as, driving charge flow in TENGs, (Fig. 2.3) depends on the
strength of the triboelectric effect generated by a particular interface pair [12, 57-59].
Based on the distributions of intrinsic electric field distributions during the power
generation, a recent numerical model of CS-TENG operation has been developed by
Niu et al [57]. The energy source (i.e. tribo-charges) is uniformly distributed over
infinite plates in Niu et al; however, this model does not account for the fact that real
TENGs consist of finitely sized plates and it leads to the artificial result of voltage
indefinitely increasing with separation distance between the plates (Fig. 2.5(g) in [59]).
The distance-dependent electric field (DDEF) model addresses this issue [59] and
accounts for finitely sized plates – this gives a voltage which saturates beyond a certain
separation distance. Xu et al. [147] improved the DDEF model by considering the
theoretical influence from contact area due to the surface roughness on the contact
surfaces and established the so-called load-dependent model (LDM). These
fundamental numerical models were important steps in the development of our
understanding of how TENGs work. They also allow simulation of the effect of various
parameters on TENG performance (Section 3.2). This thesis studies these theoretical
models initially to understand the working principles of the CS-TENG.
However, these numerical models say nothing about the generation of tribo-charge

63
density. Rather, they depend on an experimental value being input for tribocharge
density. The surface tribo-charge density is one of the most important parameters,
because it serves as the energy source for TENGs [12]. Therefore, the energy band
approach was recommended to understand the surface charge transfer and the effected
parameters of the contact materials [58, 60]. The surface tribo-charge transfer depends
on balancing the differences in the contact materials band gap [60]. For the metals, they
are characterized by their Fermi level (Ef), in which all the states are occupied below it.
The states are empty above the Fermi level. For the dielectric, the molecular structure
characteristics of its surface and bulk are different [60, 63]. The dielectric surface can
be characterized by its conduction band (CB) and valence band (VB); however, the
presence of surface/defect states in the bandgap should be assumed, owing to breaking
the symmetry at the surface [63]. Some of the surface states in the bandgap (Eg) with
energy below Ef could be filled up by electrons transferred from the metal into the
dielectric surface during contact between the metal and dielectric. This is because the
edge level of the valence band of the dielectric is lower than the Fermi level of the metal
[58, 60]. It should be noted that this results in a dielectric surface with negative charges.
The tribo-charge generation procedures and the calculation of surface tribo-charge
density are discussed using the energy band apparoch in Section 3.3. The transfer of
surface tribo-charges can simply be explained by the energy band theory. Additionally,
the parameters, which affect surface tribo-charge generation, will also be highlighted
from the tribo-charge density calculation. This modelling will establish the foundation
for selecting the triboelectric interface pair and designing the triboelectric composites
in the future TENG research.
Finally, this chapter presents a study on optimizing a low permittivity substrate below
the tribo-contact surface so that gains from both strong triboelectric effect and low
permittivity can be harnessed in a single device architecture. A simulation-based study
is presented to investigate the effects of varying substrate relative permittivity upon
TENG outputs such as open circuit voltage, short circuit charge density and short circuit
current density. In addition to further in-depth simulation, two TENG devices fabricated
with polyethylene terephthalate (PET) and polytetrafluoroethylene (PTFE) substrates
64
are tested to verify the simulation results. In order to assess the influence of substrate
permittivity only, the tribo-contact materials are fixed as PET and PDMS films in both
the simulations and experiments in Section 3.4. Therefore, the tribo-charge density of
the devices can be considered constant. Other factors affecting the output parameters
where also fixed to enable comparison including separation distance, frequency and
applied force.
This chapter primarily discusses the following points:
Ÿ Introduction of the working principles of a CS-TENG by MATLAB simulation and
summary of the impacted parameters from the device architecture.
Ÿ Analysis of the tribo-charge generation behavior by the energy band theory
between the interface pair materials.
Ÿ A study on optimizing CS-TENG performance via a low permittivity substrate by
both simulations and experiments.

3.2 Model of the working principle of contact and separation mode TENGs
The tribo-charges are the power source that drives TENGs to produce the free charges
on the electrodes. In CS-TENGs, coupling of contact electrification and electrostatic
induction define the operating principle. The transfer of opposite and equal tribo-
charges occurs on each contact surface, owing to the difference in the triboelectric
property of the contacting materials. The working principle applying to the open circuit
voltage and short circuit current condition for a CS-TENG is shown schematically in
Fig. 2.3 in Chapter II. The output performance of a TENG is actually dependent upon
many factors, including surface roughness, thicknesses, separation distance and contact
material properties. The open circuit voltage is created by the difference in potential
between the electrodes as separation occurs. Simultaneously, these built-in electric
fields cause the polarization of the dielectric material and produce the induced charges
on the electrodes. In a short circuit mode, the charge is transferred from one electrode
to the other. Therein, the separation distance between the interface pairs is one of most
important parameters. Different open circuit voltages will be developed with different
separations. The building of potential on the electrodes is also related to the intensity

65
of the internal electric field of TENGs. This depends on both the tribo-charge density
σT and the permittivity ε as [59]:
-%
"= (3.1)
2+

where, E is the electric field perpendicular to the charged surface, !! is tribo-charge

density and # is the permittivity of the medium. However, this electric field model is
appropriate to a conventional capacitor rather than a CS-TENG, because the
conventional parallel capacitor model assumes infinitely large charged planes. Hence,
this model cannot explain the polarization of the dielectrics or the induction of free
charges on the electrodes of a TENG completely. Therefore, the distance-dependent
electric field (DDFE) model has been proposed to complete the explanation of the
electric field variations of oppositely charged finite plates in relation to their contact-
separation movement [59].
-
"# = B+% D(E) (3.2)

5
D(1) = >?@3>A B 6
C (3.3)
$7C 7 $ 5 $
D> @ 6> @ 61
6 6 6

where, L and W are the length and width of contact surface and z is the distance from
point to the contact surface. Because the equivalent and opposite tribo-charges are
located on two sides of the contact surfaces, the total overlapped electric field intensity
on one side electrode can be presented as:
-
"(&(A9 = B+% ND(1) − D(1 + E)O (3.4)
!

where, x is the distance from material surface to the electrode surface (i.e., the thickness
of the tribo-layer). The electric potential can be described by the relationship between
the distance and the total overlapped across the electrode as:
- I I (3.5.1)
P1 = B+% /∫. D(1)R1 − ∫. 6# D(1) R15
"" "

:-% I I
P2 = /∫. D(1)R1 − ∫. 6# D(1) R15 (3.5.2)
B+$ $ $

They can be simplified as:

- . 6#
P1 = B+% /∫. " D(1)R15 (3.5.3)
""

66
 :-% . 6#
P2 = B+$
/∫. $ D(1)R1 5 (3.5.4)
$

where, φ1,2 are the potential of two electrodes, x1,2 are the thickness of the tribo-layers
and ε1,2 are the permittivities of the dielectric materials. Because there is not induced
free charges on the electrodes, the electric potential of the electrodes can neglect the
influences from the induced free charges. The open circuit voltage (Voc) can be
calculated as:
Voc= P1 − P2 (3.6)
Additionally, if the device is considered working at short circuit condition, the induced
free charges on the electrodes will affect the potential of the electrodes. Thus, the
potential of the electrodes could be described as:
:-: . 6.$ 6# - . 6#
P1 = B+"
(∫$ " D(1)R1) + B+% (∫. " D(1)R1) (3.7.1)
" "

P2 =
:-:
(∫$ "
. 6.$ 6# -
D(1)R1) + B+% (∫. $
. 6#
D(1)R1) (3.7.2)
B+$ $ $

where, !" is the induced free charge density on the electrodes

Because of the short circuit condition, P1 − P1 =0
Therefore, the induced free charge density can be calculated as:
" ;" <7 " ; <7
-% ( ∫ L(.),. 6 ∫; $ L(.),. )
1" ;" 1$ $
!" = " " ; <; <7 (3.8)
( 6 ) ∫! " $ L(.),.
1" 1$

The short circuit current (Isc) can be evaluated as:

,-:
Isc = F ,(
, (3.9)

where A is the device area. The separation distance (z) is holding on a sinusoidal
function for the device operation, as:
;B
E = TUVA /2XD3 + 2
5+T (3.10)

From the Eq. (3.5.1-3.6) and (3.8-3.9), the key parameters influencing CS-TENG
output performance are tribo-charge density, separation distance, thickness of the tribo-
layer and contact area. The tribo-charge density is a constant which is related to the
contact materials. A dielectric-dielectric CS-TENG, which is using PDMS and PET as
the interface pair, was selected for the numerical simulation in Fig.3.1. PDMS and PET
films with 20$Y and 200$Y thickness, and another PET sheet with 100$Y was

67
adopted as the substrate under PDMS film. Moreover, the relative permittivity of PDMS
and PET was 2.72 and 3.3, respectively. Note, the relative permittivity of the top tribo-
layer can be calculated as 3.24 by the parallel capacitor model [43]. The tribo-charge
density used was 40.7$</Y2 from [82]. Fig. 3.2(a) plots the result of open circuit
voltage vs separation distance. The open circuit voltage is increasing lineraly when the
separation distance is under 18mm. Additionally, the growth rate of open circuit voltage
is decreasing after the separation distance is over 18mm. Therefore, the separation
distance is one of the important parameters for improving the open circuit voltage of
CS-TENGs. However, the free charges can flow between two electrodes to counteract
the difference in potential of two electrodes at short circuit current stage, as shown in
Fig. 2.3 in Chapter II. Additionally, Fig.3.2(b-c) shows that the free charge and short
circuit current density increase with rising separation distances. The output charge
density can increase from 33.5 to 38.6 μC/m2, when the separation distance increases
from 1 to 10mm. In addition, the short circuit current density is improved 3.4 times,
after the separation distances rise to 10mm. Thus, based on the simulation results and
formula analysis, the increasing separation distance of CS-TENGs can influence the
outputs as indicated.

(a) (b) ++++++++++++

x1
--------------
f(z)
x2 ++++++++++++
--------------

Fig. 3.1: Schematic of the CS-TENG used for numerical study by (a) 2D (b) 3D version [42].

(a) (b) (c)

Output charge density (!%/$! )

Output current density (!"/$! )

Open circuit voltage (V)

Fig. 3.2: Output performance with separation distance (a) open circuit voltage (b) output charge density
and (c) short circuit current density.

68
The effect from the thickness of the tribo-layers for the device output performances was
also investigated by the MATLAB simulation. The thickness of PET sheet is fixed at
100$Y as well as the thickness of PDMS film are defined as 120, 80, 40 and 20$Y.
Fig. 3.3(a-c) shows the simulation outputs (open circuit voltage vs separation distance,
short circuit charge density vs time and short circuit current density vs time) of the
device with different thickness of PDMS films. Simultaneously, Fig. 3.3 then illustrates
these outputs against the thickness of PDMS. The open circuit voltage is increased with
reducing thickness of PDMS, as shown in both Fig. 3(a) and (d). Herein, the open circuit
voltage is raised from 1368V to 1371V under a fixed 1mm separation distance, after
decrease the PDMS thickness from 120 to 20$Y. Moreover, the short circuit current
density and short circuit charge density decline with the growing thickness of PDMS
from Fig 3.3(b) and (e). Also when the thickness of PDMS film is enhanced from 20$Y
to 120$Y, the short circuit current density decreases from 174.6 to 155.4 $F/Y2 (i.e.,
12.3% reduce). This enhancement with 20$Y thickness PDMS film can be 1.12 times
the output of short circuit current density from the device with 120$Y PDMS as the
negative triboelectric interface. A similar trend is observed with short circuit charge
density in which it increases from 32.28 to 33.63$</Y2 over the range of thickness of
PDMS and 1mm separation distance, as depicted in Figs. 3.3(c) and (f). Therefore,
depending on the investigations of the outputs from the devices with different thickness
of the contact material, a low thickness tribo-layer can increase the output performancs
of CS-TENGs.
(a) (b) (c)

(d) (e) (f)

Fig. 3.3: The simulation results of (a) open circuit voltage (b) short circuit current density and (c) output

69
charge density. The effects for the output performances with thickness of PDMS (d) open circuit voltage,
(e) output charge density and (f) short circuit current density.

Because of the inevitable presence of surface roughness, the contact area between the
interface pair cannot achieve 100% contact (Chapter V) and, in many cases will be very
much less. Simultaneously it is reasonable to assume that the tribo-charges can only be
generated by real contact of the materials (i.e. atomic interaction). The contact area (S)
in Eq. (2.1) in section II is inapplicable to the CS-TENG theoretical model, because
tribo-charges can only be generated in the real contact area. Increased applied force
induces growth of the real contact area between the contacting surfaces so that the total
generated tribo-charge is likely to be increased. Depending on this hypothesis, a model
of CS-TENG principle (load-dependent model) has been proposed by Yang et al (2020)
[147]. The model essentially applies the concept of a load-dependent real contact area
from the tribology literature on rough surface contact. The core of the argument in this
model is that the tribo-charge density is redefined as:
)% (3.11)
!! = E#

where, the 6! is the quantity of generated tribo-charges on the contact surfaces, but
F% is the real contact area between the interface pair. Secondly, the surface roughness
was characterized by Persson’s contact theory to determine the load-dependent contact
area (Persson’s contact theory gives a reasonable estimation of real contact area up to
complete contact provided the deformation remains linearly elastic.). An error function
of the normal force expresses the real contact area (Ar) in Persson’s contact theory as
[147]:
E# M 2
E9
= erf ]E ∗ ^|PQ | $ 6|PQ |$
_ (3.12)
9N " $

Where ℎ1 (1, a) and ℎ2 (1, a) are the heights of two contact surfaces roughness and
" ∗ is the composite modulus which can be calculated by:
1 1:7" $ 1:7$ $
N∗
= N"
+ N$ (3.13)
where E1,2 and :1,2 are the Young’s modulus and the Poisson’s ratio of two tribo-
contact layers, respectively. According to Eq. (3.12), the Ar is varying with contact force.
The generated tribo-charges are distributing in the locations of real contact, and the

70
intensity of the formed electric field also depends on the Ar. The distance-dependent
electric field expression in Eq. (3.3) can be improved as:
-%
"# = D(E)
B (3.14)
Therein [147],

B # E 2
D(E) = 2 B− >
+ 1C /E # 5 (3.15)
C # 6(#)$ 9
?

After integrating the "# into Eq. (3.5.1 to 3.6) and (3.7.3 to 3.9), the characteristics of
the CS-TENG depending on the contact force (or pressure) can be investigated,
including open circuit voltage, short circuit current and transfer charge density in Table
3.1.
As a conclusion to the above-discussed formulas, it should be noted that the DDEF
model has been established to improve the conventional CS-TENG, as tribo-charges
are distributing on finitely sized charged planes. The DDEF model is more accurate in
explaining the principle of a CS-TENG depending on Maxwell’s equation, because this
model shows the polarization of the dielectrics or the induction of free charges on the
electrodes of a TENG. The key underpinning equations describing the operation of the
CS-TENG are summarized below in Table 3.1.

71
 Table 3.1. Summary table of the key theoretical formulas describing CS-TENG operation.

Electric field VOC Qsc Ref.

#" *+" +" -(/) *+" *+" -(/) Niu’s

!! = ,() = , = #2) = , =3
$%# %! #$% 1& #$% ,'('#) & 4-(/)

# model
− $" + ((*)
!$%& =
%# [57]
#"
!' =
$%# %'
* -* ./ * - ./
(" Voc= @! − @' +" ( ∫
,* -*
7(-)3-4 ∫- + 7(-)3-)
,+ + DDFE
!5 = .(/) (2) = * * -* .-+ ./
-% (, 4, ) ∫&
* +
7(-)3-
-* 45
(" model
@! = AB .(C)DC E
.(/) -%! -*

6 −(" -+ 45 [59]
7 @' = AB .(C)DCE
= 012*03 ⎛ ⎞ -%' -+
2/ 9 / ' 6 '
⎝ 7 4 ;7< + ;7< + 1 ⎠
(" ,() #2) = ,() L":;< Load-
!5 = .(/)
-
5(/) L":;< dependent
("
- / G& ' = J2%# 9! B .(/)D/
.(/) = ⎛− + 1⎞ H I - # G8 -
2 G8 = model
9G& + (/)' %# 9! ;∫-
5(/)4-*
.(/)D/ + ∫-
5(/)4-+
.(/)D/<
⎝ - ⎠ 5(/)4-* * +

+ %! 9! AB .(/)D/ 5(/)4-* 4-+ -* [147]

5(/) + %! 9! AB .(/)D/ − B .(/)D/E
5(/)4-+ #
-*
− B .(/)D/E 5(/)4-+ 4-* -+
# + %' 9! AB .(/)D/ − B .(/)D/E
5(/)4-* #
5(/)4-+
+ %' 9! AB .(/)D/
5(/)

-+
− B .(/)D/EK
#

3.3 Modelling of surface tribo-charge transfer

In the previous sections, the parameters which influence the CS-TENG performance
have been discussed separately except the generation of surface tribo-charge density.
The tribo-charge generation is related to the contact electrification via tunneling effect
between the contact materials [58]. Contact electrification depends on the charge
transfer from the contacting particles (e.g. atoms or molecules) on the contact material’s
surface [20, 63]. Because of the dangling chains on the material surface (as the surface
defects), a temporary bond between the contact particles will form [61, 62]. The

72
negative charges can be transferred onto the tribo-negative material surface, due to the
high electron affinity. Additionally, the holes (positive charges) are left on the tribo-
positive material surface. The large differences in the tribo-properties of the interface
pair exist so that they can generate more tribo-charges on the contact material’s surface.
The basic materials for the interface pair are usually selected as a pair of materials with
large difference in location in the triboelectric series (see Section 2.3). However,
although the triboelectric series is a respectable reference for the contact material
selection, it cannot explain the mechanism of surface tribo-charge transfer or the
influential parameters involved in charge generation. Thus, according to the mechanism
of the tribo-charge generation, the energy band for illuminating the charge transfer has
been applied to explain these questions explicitly. In terms of contacting materials, there
are two key types for TENGs: metal-dielectric and dielectric-dielectric. Next, tribo-
charge generation in each of these two modes is explained.

3.3.1 Metal-Dielectric Mode

The energy band diagrams for the interface materials representing charge transfer are
shown in Fig. 3.4 [58]. Owing to the differences in work function and electron affinity
between the metal and dielectric, electrons will be transferred from the surface of the
metal to the dielectric [59, 139, 204]. As the work function of the dielectric is generally
larger than for a metal, the electrons from the top metal will be injected into the surface
of the dielectric during the procedure of contact and separation until the surface states
of the dielectric are as high as the Fermi level of the metal [58]. Moreover, there is a
built-up electric field in the TENG. Also, the source of the electric field is from the
tribo-charges when the interface pairs are separated. Due to electrostatic induction, the
top and bottom electrodes are induced with the free charges (the charge density on the
top and bottom electrode are !1 and !2 respectively, as shown in Fig. 3.4(a).) [58].
Owing to the free charges on the top (!1 ) and bottom (!2 ) electrode that were induced
by the tribo-charges (!! ) on the surface of the dielectric, the relationship between tribo-
charge density and induced charge density is [58]:
!! + !1 + !2 = 0 (3.15)

73
 (3.16)
Besides, because of losing electrons and built-in electric field for the metallic surface,
the metallic vacuum energy level reduceds by ∆"733 , which compares with the
dielectric surface in Fig. 3.4. Thus, the ∆"733 can be calculated as:
-" .S
∆"733 = +!
(3.17)

where x is a critical tunneling distance, g is the elementary charge and #$ is the

permittivity of air. Moreover, the surface states of the dielectric should be the same as
the metallic Fermi energy level after contact, owing to the equilibrium of the system.
Hence,
N 6∆N.
!! = −g ∫N ! h5 (") R" (3.18)
!

where, h5 (") is the surface density of states, ∆"5 is the range of filled surface.
Therefore, the average of surface density of states can be described as:
@ <∆@.
∫@ ! U. (N)
iiiiiiii
h 5 (") =
!
(3.19)
∆N.

Thus, simplifying the ∆"5 gives:

:-%
∆"5 = VVVVVVVV
(3.20)
U. (N)S

From Fig. 3.4, the function between the energy of metallic and dielectric surface can be
defined in Eq. (3.21), depending on the energy conservation as:
P1 + ∆"733 =P2 − ∆"5 (3.21)
Simultaneously, the differential potential between two electrodes is:
-" -$
,= +!
1−+ R (3.22)
" +!

where the #1 is the permittivity of the dielectric. Therefore, the tribo-charge density
can be calculated as:
B CB 0
W6( " $ )×(16 )
!! = 0
8
"
1" ;
0
(3.23)
6( )(16 )
1" 1! D. (@)8$ 1" ;

And at the open circuit condition as:

B CB 0
( " $ )×(16 )
!! =
8 1" ; (3.24)
0 " 0
6( )(16 )
1" 1! D. (@)8$ 1" ;

Consequently, the tribo-charge density is related to the difference in work functions

between two contact materials and the surface density of states in Eq, (3.24). However,

74
the tribo-charge transfer between the dielectric-dielectric materials is different from the
metal-dielectric mode, because dielectric-dielectric mode depends on the different VB
and CB structures of two contact dielectric materials [63]. The energy band analysis for
the tribo-charge transfer of the dielectric-dielectric mode is depicted in Section 3.3.2.

(a) + + + + + + + + + + + + + + + !"
x
- - - - - - - - - - - !-!- - - - - - - - -
d
+ + + + + + + + + + + + + + + !#

(b) " #$$

"#$$ ∆"#$$ =
&! '(
)'

!!
!!
!"
!"
"%&
"%&
−&)
∆"( =
+( (

Metal Dielectric Metal Dielectric

(+) x (-) (+) (-)
x

Before contact After contact

Fig. 3.4: (a) Schematic and device structure information of a metal-dielectric mode CS-TENG and (b)
the energy band diagram for surface tribo-charge transfer at before contact and after contact status.

3.3.2 Dielectric-Dielectric Mode

Comparing the contact electrification of metal-dielectric and dielectric-dielectric
materials, the free charges on the top electrode are induced by the tribo-charges on two
sides of dielectric material surfaces rather than the single dielectric material surface, as
shown in Fig.3.5. Simultaneously, after the equal and opposite tribo-charges on two
dielectric surfaces appear, the generated tribo-charges are fixed on the material surfaces.
Also, the transferred charge density should be satisfied by|!1 | = |!2 | = |!| [139].
Where !1 and !2 are the induced charge density on the top and bottom electrodes
respectively in Fig. 3.5 and the ! is the transferred charge density. In order to balance
the differences in VB and CB structures of two contact dielectric materials, the charges
will transfer from one dielectric to another dielectric surface during the contact [60].
Due to the charge transfer, the variations of the Fermi level (compared with the original
Fermi level) are presented as ∆"51 and ∆"52 . ∆"51 and ∆"52 are also related to the

75
surface density of states [60] as.
-%
∆"51 = VVVVVVVVVV
U." (N)S (3.25.2)
:-%
∆"52 = VVVVVVVVVV
U.$ (N)S (3.25.1)
where iiiiiiiii
h51 (") and iiiiiiiii
h52 (") are the surface density of states for Dielectric 1 and 2.
Furthermore, owing to the opposite tribo-charges on each dielectric surface, the
differential potential between two interface surfaces can be determined as:
-" 6-$
,?AY = g 1 = g∆"?AY 1 (3.26)
+!

where x is a critical tunneling distance, g is the elementary charge and #$ is the

permittivity of air. Also, depending on the energy band diagram, the function can be
presented by [60]
% - % - (3.27)
P1 + ∆"?AY 1 + VVVVVVVVVV
U (N)S
= P2 − /− VVVVVVVVVV
U (N)S
5
." .$

Thus, based on Eqs. (3.25 - 3.27), the tribo-charge density can be calculated by [50]
2-S $ . VVVVVVVVVV
U (N)×U
."
VVVVVVVVVV
.$ (N)
!! = (g(P2 − P1 )-( +!
))× /U
VVVVVVVVVV
(N)6U VVVVVVVVVV
(N)
5 (3.28)
." .$
S (Z$ :Z" )
!! = $8$ ;!
"
GGGGGGGGGGG
"
6GGGGGGGGGGG6 (3.29)
D ." (@) D.$ (@) 1!

Consequently, the equations of both metal-dielectric and dielectric-dielectric mode

tribo-charge density have been calculated by analysing their energy band diagrams for
surface tribo-charge transfer. From Eqs. (3.24) and (3.29), the key parameters
governing tribocharge density are clear for both metal-dielectric and dielectric-
dielectric modes. These are the differences in work function of the contacting materials,
the surface states of the contacting materials and the permittivity of the materials. The
difference in work function means the edge of the tribo-negative material Fermi level
is lower than the tribo-positive material Fermi level. Thus, there are empty surface
states that can be filled in the electrons from the tribo-positive materials. The number
of the empty surface states is as high as the large differences in the work function of the
interface pair. Contact materials surface states and permittivity are the key parameters
to improve the tribo-charges density in Eq. (3.24) and (3.29). This result is helpful in
approaching tribo-materials design for high performance TENG devices. For example,
designing composites can increase the contact surface permittivity. Additionally, the
76
poling process on the contact material surface can attempted, because the external
voltage bias can shift the dielectric surface states. It should be noted that the material
surface potential is the parameter to evaluate the shifting of the dielectric surface states
[58, 63]. This aspect is explored in detail in Chapter V.

(a) - - - - - - - - - - - - - - - - - - - - - "#
!!
++++ +++++++++++
x
---------------------
!"
+ + + + + + + + + + + + + + + "$

(b) " #$$

∆'%&&

!!
!" %# %$
"%! ''#
!!
""# # −!!
""$ #
"%" ''$

Dielectric Dielectric Dielectric Dielectric

(+) x (-) (+) x (-)

Before contact After contact

Fig. 3.5: (a) Schematic and device structure information for a dielectric-dielectric mode CS-TENG and
(b) the energy band diagram for surface tribo-charge transfer at before contact and after contact status.

3.4 Optimizing CS-TENG performance via a low permittivity substrate

According to the numerical study in Section 3.2, the output performance of a CS-TENG
has been simulated by MATLAB. Firstly, the simulation-based study is used to
investigate the effects of varying substrate relative permittivity upon TENG outputs
such as open circuit voltage, short circuit charge density and short circuit current density.
In addition to further in-depth simulation, two TENG devices were fabricated and tested
with polyethylene terephthalate (PET) and polytetrafluoroethylene (PTFE) substrates
to verify the simulation results. In order to assess the influence of substrate permittivity
only, the tribo-contact materials are fixed as PET and PDMS films (see Fig. 3.1) in both
the simulations and experiments. Therefore, the tribo-charge density of the devices can
be considered constant. Other factors affecting the output parameters where also fixed
to enable comparison including separation distance, frequency and applied force.

77
Because of the same fabrication for the PDMS thin film, the contact area is also fixed
by using the same applied force value.

3.4.1 Simulation and fabrication

A．Simulation

Using the distance-dependent electric field model of CS-TENGs in Section 3.2 [26, 38],
the open circuit voltage and short circuit current was simulated (using MATLAB) for
different substrate relative permittivity. The simulated TENG consists of two contacting
layers which can be thought of as a parallel plate capacitor (Fig. 3.1). The ‘top tribo-
layer’ is comprised of the tribo-contact material, the substrate and the electrode while
the ‘bottom tribo-layer’ comprises just the partner tribo-contact material and its
electrode. The interface pair (i.e. the tribo-contact material pair) was nominally selected
as PDMS and PET (as in Fig. 3.1) to allow direct comparison with results from [59].

Accordingly, a tribo-charge density of 40.7 μC/m2 was used as in [59]. First, TENG

operation was simulated for a selection of substrate materials and an optimum material
was chosen based on optimum permittivity. The total equivalent permittivity of the top
tribo-layer can be calculated by the parallel plate capacitor model. The capacitance of
the substrate layer (in the top tribo-layer) is given by C = εA /d [205], where C is
capacitance, ε is permittivity, d is thickness of medium and A is the in-plane area. If the
capacitance of the top layer is supposed as two dielectrics in one capacitor, the
equivalent capacitance of the two materials comprising the top tribo-layer can be
presented as:
1 1 1
=4 +4 (3.30)
48H " $

where, C1 and C2 are the capacitances of PDMS and the substrate, respectively.
Combining Eq. (3.30) and the formula for capacitance above, the total relative
permittivity of the ‘top tribo-layer’ can be calculated by [205] as:
+!,#$%& +!,&'()*!+*, (-#$%& .-)'()*!+*, )
!',)* = -#$%& +!,&'()*!+*, .-&'()*!+*, +!,#$%&
(3.31)

Where, the εr,PDMS and εr,Substrate are the relative permittivity values of PDMS and the

78
substrate, respectively and dPDMS and dsubstrate are the thicknesses of PDMS and the
substrate film (the ‘yellow’ layer in Fig. 3.1).

Table 3.2. Parameters for ‘top tribo-layer’ [42].

Substrate (PTFE, PET, Paper, Bakelite, Contact Material (PDMS)
Neoprene rubber, PVDF)
Thickness = 100µm Thickness = 20$Y

Relative Permittivity ( #%1 ) varying Relative Permittivity (#%2 = 2.7)

(depending on type of substrate)
Size: 2cm × 2cm Size: 2cm × 2cm

Table 3.3. Substrate relative material permittivity εr, ‘top tribo-layer’ relative equivalent permittivities
ε(r,eq) and ‘top tribo-layer’ absolute equivalent permittivities εeq [42].
Substrate Materials #% of substrate #%,S[ of top #S[ of top tribo-layer
materials tribo-layer × 10:12 F/m
PTFE (Teflon) 2.1 2.2 19.3
PET 3.3 3.2 28.2
Paper 3.7 3.5 30.9
Bakelite 4.9 4.3 38.3
Neoprene Rubber 6.7 5.4 47.7
$ − &ℎ()* PVDF 7.5 5.8 51.3

Using the relative permittivity for PDMS (Table 3.2) together with the substrate relative
permittivity values in Table 3.3 (first column), the equivalent relative permittivity of
the top tribo-layer was calculated using Eq. 3.31 for a range of substrate materials
(PTFE, PET, Paper, Bakelite, Neoprene rubber and $ − &ℎ()* PVDF). These are listed
as the second column in Table 3.3. The third column in Table 3.3 gives the equivalent
absolute permittivity of the ‘top tribo-layer’. These materials represent a relatively wide
range of permittivity values. Therefore, the relationship between substrate permittivity
and TENG output (i.e. open circuit voltage, short circuit charge density and short circuit
current density) can be summarized by simulating TENGs having these substrate
materials. This essentially allows determination of an optimum substrate material. In
the first part of the simulation work, the effect of substrate material relative permittivity
on TENG output performance was studied using the same dimensions as in previous
79
literature [26], including the nominal contact area of 5cm × 5cm and the thickness of
the interface pairs (PDMS and PET) of 20 and 200μm. This was done to allow direct
comparison with the result in [26] where a PET substrate was used. Subsequently, the
thicknesses of the PDMS and PET were altered to 20μm and 127μm and the nominal
contact area was altered to 2cm × 2cm (see Table 3.4) in order to match the devices
fabricated in this work. Moreover, the thickness of the substrates was fixed at 100μm.
The relationship between substrate permittivity and TENG output was simulated by
these parameters to compare the simulation and experimental results. Note, that the
substrate is implanted under the negative triboelectric interface (PDMS). Based on
these parameters, the open circuit voltage, short circuit current and charge have been
simulated and results are discussed in Section 3.4.2.

Table 3.4: Comparison of fabricated TENG devices and test parameters [42].
Device 1 Device 2
Sample Size = 2 cm × 2cm Sample Size = 2 cm × 2cm
Negative Contact Material – PDMS (20"# Negative Contact Material – PDMS (20"#
thickness, $IJKL = 2.7) thickness, $IJKL = 2.7)
Positive Contact Material – ITO coated PET Positive Contact Material – ITO coated PET film
film (127"# thickness, $IMN = 3.3) (127"# thickness, $IMN = 3.3)
Substrate Material – PET film Substrate Material – PTFE film (100"#
(100"# thickness, $IMN = 3.3) thickness, $INOM = 2.1)
4Hz Operation frequency 4Hz Operation frequency
8N applied force 8N applied force
1mm separation distance 1mm separation distance

B．Fabrication

For the experimental aspect of the work, two types of devices were fabricated based on
two substrate materials having sufficiently different permittivity values (PTFE and PET
- see Table 3.4 for comparison). The overall approach to device structure and fabrication
is somewhat similar to that employed by Dharmasena et al. [59] with some differences.
The bottom half of the TENG was fabricated simply as ITO coated PET film (sheet
resistivity = 60Ω/sq, thickness 0.127mm, Sigma Aldrich, UK). The top half consisted
of PDMS, the substrate material (either PET or PTFE) and a gold coating as the

80
electrode. PDMS is the negative contact layer and PET is the positive contact layer. To
create the PDMS layer, PDMS cross-linker was mixed with its elastomer by a 1:10 ratio
and degassed under vacuum for 45 min. Subsequently, the PDMS was spin-coated on a
silicon wafer symmetrically at 3800 rpm for 60 seconds to obtain a 20μm thickness. In
this experiment, the negative contact material (PDMS) and the positive contact layer
material (PET) were fixed in order to ensure the same tribo-charge density generation
for each device. A substrate sheet (100 μm thickness of either PET or PTFE) with
2cm×2cm size was attached on the previous PDMS layer and cured at 70℃ for 2hrs.
Finally, the gold film was deposited. The fabrication steps for the top triboelectric layer
are shown Fig. 3.6. A layer of Kapton tape was required on each electrode for sufficient
isolation. The mechanical energy supply for the TENGs was supplied by an oscillating
force (maximum 8 N with frequency 4 Hz) using a TIRA shaker. The separation
distance was fixed at 1mm and the open circuit voltage was recorded by an oscilloscope
(KETSIGHT, MSO-X 4154A). The oscilloscope was connected with a voltage
separation circuit comprising a R1=1GOhm and a R2=10MOhm resistor as shown
schematically in Fig. 3.7. This was done to ensure that the impedance of the voltage
meter setup was much larger than the TENG internal impedance.

Fig. 3.6. Fabrication steps for the ‘top tribo-layer’ [42].

81
Fig. 3.7. Schematic of equivalent circuit for open circuit voltage measurements [49].

3.4.2 Results and discussion

A. Simulation results
Fig. 3.8(a) presents the device’s outputs (open circuit voltage vs separation distance)
for device dimensions of 2cm × 2cm. Fig. 3.8(b-d) reports the device output (open
circuit voltage vs separation distance, short circuit charge density vs time and short
circuit current density vs time) for a device having the same dimensions as the
experimental devices (i.e. 2 cm × 2 cm). Fig. 3.8(e-f) then plots these outputs (2 cm ×
2 cm area) explicitly against relative permittivity. Note, the results in Fig. 3.8(c-f) are
for a 1 mm separation distance. It is clear from Fig. 3.8(a) that open circuit voltage
increases significantly as the relative permittivity of the ‘top tribo-layer’ is reduced.
Unsurprisingly, the trend in Figs. 3.8(a) and 3.8(b) is similar as only the device area and
layer thicknesses have changed. In Fig. 3.8(e), the open circuit voltage increases from
1032 V to 1643 V (i.e. 1.6 times), when the substrate material changes from PVDF to
PTFE (i.e. relative permittivity of substrate reduced from 7.5 to 2.1). The improvement
with PTFE is 1.23 times the output reported in [26] with PET (εr = 3.3) as the substrate
material (the result from [26] is identical to the orange line in Fig. 3.8(a)). Hence, open
circuit voltage can be enhanced by embedding a substrate with lower relative
permittivity than the original triboelectric contact material to reduce the relative
equivalent permittivity of the overall top triboelectric layer in Fig. 3.1. Although Fig.
3.8(c-d) shows short circuit charge and current density declining somewhat with
increasing relative permittivity, Fig. 3.8(f) shows that the change is insignificant
relative to the magnitudes of these outputs. Over the range of relative permittivity, they

are approximately fixed at about 36.2 μC/m2 and 229.2 μA/m2 respectively. Hence, by

choosing PTFE as the optimum substrate material, we can obtain useful increases in

82
 open circuit voltage while not affecting current and charge density very much. In
summary, the simulation results have indicated that open circuit voltage can be
significantly improved, while short circuit charge density and short circuit current
density are only very slightly increased by implanting a low permittivity material under
the negative triboelectric interface.
(a) (b)

(c) (d)

(e) (f)

0 − 23456

Fig. 3.8: Simulation results: (a) Open circuit voltage vs separation distance with 5cm×5cm areas and

83
200μm thickness substrate, (b) Open circuit voltage vs separation distance with 2cm×2cm areas and

100μm thickness substrate, (c) Short circuit charge density vs time, and (d) Short circuit current density

vs time with substrate materials having a range of permittivity values (at 1mm separation distance for c

and d), (e)–(f) Open circuit voltage, short circuit current density and short circuit charge density versus

relative permittivity for different substrate materials (at 1mm separation) [42].

B. Experimental results
In order to provide a preliminary check on the validity of the simulation predictions,
two materials from Table 3.4 (PET and PTFE – the best performing substrate materials
in Fig. 3.8(e)) were chosen as the substrate material and two corresponding TENG
devices were fabricated (described above) and tested. The devices were fabricated
identically – the only difference being the relative permittivity of the substrate material
reduces from 3.3 to 2.1 in going from a PET to PTFE substrate. Identical interface
materials ensure that both devices generate the same tribo-charge density. Test
conditions were also identical for both devices: a frequency of 4 Hz, a max load of 8 N
and a 1 mm separation distance. The substrate thickness was 100 μm for both devices.
A 5 min pre-charge time was employed before each set of measurements to ensure
equilibrium of tribo-charge generation. Table 3.4 summarizes the device and test
specifications. Fig. 3.9 plots the output open circuit voltage signal from both devices.
The max output voltage increased from 29.5 to 39.4 V (i.e. 1.3 times) and the peak-to-
peak value from 46.5 to 61.4 V (i.e. 1.3 times). These increases are very much in line
with the simulation prediction from Fig. 3(e) where the simulated open circuit voltage
increased by 1.3 times between PET and PTFE. Note, that the voltage magnitudes
between simulation (Fig. 3e) and experiment (Fig. 6) are quite different however. A key
reason for this is likely to be because the simulation results are based on a perfect
contact interface; whereas, the amount of ‘real contact area’ in the experimental device
(at these loads) is likely to be only a fraction of the nominal contact area. A lower
experimental ‘real contact area’ would be likely to generate a reduced tribo-charge
density at the interface and therefore, a reduced open circuit voltage output.
Nevertheless, contrasting the relative voltage increase represents a valid comparison

84
and results suggest that the experimental check is in-line with the simulation prediction
(as both predict an increase in open circuit voltage of roughly 1.3 for the move from
PET to PTFE). Finally, the experimental results showed no appreciable difference in
short circuit current density in moving from PET to PTFE substrates as predicted by
the simulation result in Fig. 3.8(f).

Fig. 3.9: Comparison of open circuit voltage signals for the devices with PTFE and PET substrates [42].

3.4.3 Application in a self-powered system by integration with a stretchable

photodetector.
In this section, the fabricated CS-TENG is used to power an energy autonomous UV
photodetection sensor. Before the demo, the high flexibility/conformability of the
fabricated TENGs and photodetectors (PDs) were tested by placing them over human
fingers and glass vials in Fig. 3.10. This qualitatively confirms the suitability of the
fabricated devices for wearable applications.

(a) (b)

Fig. 3.10: Bending stability test for TENG devices: (a) Optical image while convex bending, (b) Optical
image while concave bending [50].

85
Figure 0.1Fig. 3.10: Bending stability test for TENG devices: (a) Optical image while convex bending, (b) Optical
image while concave bending [50].

Accordingly, a self-powered UV PD concept was implemented by integrating a PD with

the TENG through a rectifier bridge circuit. It was operated by applying an 8N force at
4Hz. A LED was connected in series with the PD, in other words, one terminal of the
PD was connected to the TENG through the bridge rectifier circuit and the other end
was connected to the LED light. The setup and the working demo of the self-powered
PD are given in Fig. 3.11a. Without UV radiation, the resistance of the UV detector
remains high so there is no current flow. As a result, the LED was in an ‘off’ state (Fig.
3.11b(iii)). When PD2 was exposed to UV radiation, the resistance of the PD decreased,
allowing the current to flow through the PD and hence the LED started glowing, as
shown in Fig. 3.11b(iv). Further, the harsh deformations (stretching, bending, and
twisting) do not affect the ability of PD to detect the UV radiation (Fig. 3.11b(v and
vi)). Thus, the presented self-powered stretchable PD can be attached to any curvilinear
surfaces, such as fabric and skin, for real-time detection of UV radiation for health and
environmental monitoring.

(b)
(iv)

(i)
(a)
(v)
Equivalent circuit

(ii)

(vi)

(iii)

Fig. 3.11: (a) Equivalent circuit for TENG-PD based self-powered system, (b) Integration of stretchable
PD with TENG: Output voltage of TENG at different frequency across PD irradiated with and without
UV light and their respective responsivity at generated voltage. Photographic demo showing the
connection of LED in series with self-powered photodetector for real-time UV detection [50]. (Courtesy:
the photodetector was designed and fabricated by Dr. Yogeenth Kumaresan)

86
3.4.4 Summary on low permittivity substrate work
Materials with strong triboelectric effect and low permittivity are among the
requirements for improving triboelectric nanogenerator (TENG) performance. These
features are not always possible in the same material, so this work has outlined the idea
of optimising a low permittivity substrate material underneath the tribo-contact layer
so that both optimum triboelectric effect and low permittivity can be incorporated in
the same TENG. Results (simulated using a distance-dependent electric field model)
show that open circuit voltage increases with reducing substrate permittivity. Going
from PVDF (εr = 7.5) to PTFE (εr = 2.1), open circuit voltage increased by a factor of
1.6. Therefore, PTFE was selected as a suitable substrate material. Therefore,
simulation results show that low permittivity can be used to boast TENG performance.
A preliminary check on the simulation results was provided by fabricating and testing
two TENG devices having different substrate permittivity values. PET and PTFE
substrates of 100 μm thickness were incorporated in otherwise identical devices. In
moving from PET to PTFE substrates (εr = 3.3 to εr = 2.1), the experiments predicted
almost the same increase in open circuit voltage as the simulations (roughly 1.3 times).
Experiments also confirmed little increase in short circuit current density with substrate
permittivity as predicted by the simulations. The conclusion is that low permittivity
substrates can be used to enhance the open circuit voltage of TENGs. Current work is
concentrating on looking at this effect over a wider range of substrate materials. In
future, we will focus on demonstrating the further significant gains predicted by further
reducing the substrate thickness. Finally, TENGs are currently being considered as a
viable autonomous energy supply source for a host of self-powered sensor technologies
and the advances discussed here in terms of enhancing output will be directly useful in
those sensor applications.

3.5 Conclusions
This chapter initiates the task of attempting to optimize the output performance of the
CS-TENG. First, the evolution of the key models describing CS-TENG operation were
analyzed as a means of discovering the key parameters governing CS-TENG output.
Models have evolved from the simple infinite parallel plate capacitator model of Niu et
87
al. [57] to the distance-dependent electric field (DDEF) model of Dharmasena et al. [59]
which accounts for the important reality that TENGs are based on finitely sized plates.
From the DDEF model, it can be noted that the key parameters for designing a CS-
TENG are tribo-charge density, separation distance, thickness of the tribo-layers and
material permittivity. The DDEF simulation results show that the output performance
increases with increasing separation distance (up to a saturation distance) and
decreasing permittivity of the tribo-materials. Analysis of the model also indicates that
decreasing the thickness of the tribo-layers is a potential route to boosting CS-TENG
performance. The load dependent model (LDM) by Xu et al. [147] then added
consideration of surface roughness resulting in the contact force dependent TENG
results that have been observed in experiments. The numerical studies for investigating
the mechanism of tribo-charge transfer by energy bands for both metal-dielectric and
dielectric-dielectric mode material selection were also discussed. The equations for two
mode types have been calculated to obtain the key parameters governing tribo-charge
transfer. These suggest that the important parameters for tribo-charge transfer are:
permittivity of contact materials, material surface states and the differences in material
work function. Therefore, among the factors that confer higher TENG performance are
materials with a strong triboelectric effect together with low permittivity, but it is
challenging to optimize both within a single material. This chapter presented a solution
to this challenge by optimizing a low permittivity substrate beneath the tribo-contact
layer. Using simulations based on the DDFE model, an optimization scheme has been
presented allowing prediction of TENG output for substrates having varying
permittivities. Results were simulated over a range of substrate permittivities. The open
circuit voltage was found to increase by a factor of 1.6 in moving from PVDF to the
lower permittivity PTFE. Two TENG devices have been fabricated with 100µm PET
and PTFE substrates to compare performance. The experiments confirm that lowering
the substrate dielectric constant (i.e. PET to PTFE) raises the open circuit voltage in
line with simulation predictions. Finally, an application using the same CS-TENG
device to power a LED and switching by a photodetector has been demonstrated.

88
Chapter IV.

Origin of the contact force-dependent response of triboelectric

nanogenerators
4.1 Introduction
Chapter II has reviewed several methods to optimize triboelectric nanogenerator
performance. Most of the focus has been on optimizing the material selections [43] and
engineering the contact surfaces [137, 206] and most of the work conducted to date has
involved relatively low contact forces/pressures (such as with wearables). Output
performance depends on a number of factors: materials, contact pressure, surface
topography, frequency and separation distance. Surprisingly, the role of some of these
aspects is still rather poorly characterized and understood in TENG design. One such
example is the role of contact pressure. This chapter investigates exactly why TENG
output increases with contact pressure. For the first time, the work characterizes both
the electrical and real contact area response for a TENG based on random rough
surfaces in the high contact pressure regime as well as at low pressures. Surface
roughness is known to effect aspects such as the frictional response of surfaces [134],
the failure behaviour of materials [135] or the contact stiffness of interfaces [136], but
here, we demonstrate how it governs the contact pressure dependent output of TENGs
via its influence on the real contact area. The results lead to interesting implications
including the ability to use large contact pressures to boost performance without the
need for costly surface engineering. A number of publications (Section 2.4) have
pointed to the contact pressure dependent response, but its origin is yet to be established.
Herein, this chapter investigates the physical mechanism underpinning this behaviour
via customised experiments designed to probe the behaviour.
This chapter primarily discusses the following points:
• Origin of the contact force-dependent performance of TENGs
• Concurrent real contact area & electrical output measurement over wide contact
pressure range
• Implications for definition & measurement of tribo-charge density and
• Applications of load-dependent behaviour

89
4.2 Fabrication and experiments
Device fabrication and surface characterization is described in Section 4.2.1. Then two
separate tests are described: one to measure TENG electrical output versus contact force
(Section 4.2.2) and one to measure TENG real contact area versus contact force
(Section 4.2.3).

4.2.1 TENG fabrication and surface characterization

The device used for this study was a conductor-to-dielectric TENG in vertical contact
separation mode. Fig. 4.1 shows a schematic. The TENG tribo-contact layers were
copper sheet (100% Cu, thickness 50μm) in contact with polyethylene terephthalate
(PET) sheet (thickness 127μm). The PET is coated with indium tin oxide (ITO)
(resistivity 60 Ω/sq, thickness 0.130μm, Sigma Aldrich, UK) to form the electrode on
the PET side. Device dimensions (active tribo-contact surface) were 2.5 cm × 2.5 cm.
Since the study is focused on the role of TENG real contact area, some additional
measures were taken to ensure uniform contact conformity between the upper and lower
tribo-contact layers. The upper and lower parts of the TENG were both fixed to optical
glass plates (using double sided tape). Float glass plates possess both low flatness and
low surface roughness (<0.05μm). This measure is designed to eliminate any unwanted
perturbances arising from substrate roughness and facilitate parallel contact of the tribo-
layers. In addition, the lead wires were attached to the electrodes outside the active
device area to avoid a bulge affecting the contact area.

Fig. 4.1. Schematic of the contact separation mode TENG having copper (Cu) in contact with
polyethylene terephthalate (PET) [152].

The surface topography of both the Cu and PET surfaces was measured prior to testing

90
(i.e. pristine surfaces were measured). The Cu surface was scanned using an optical 3D
surface profilometer (Alicona InfiniteFocus) and the PET was scanned using an atomic
force microscope (AFM) (Dimension Icon, Bruker). The roughness scans for both Cu
and PET are shown in Fig. 4.2 and Table 1 reports center-line average roughness, Root
Mean Square (RMS) roughness and RMS surface gradient (see Appendix A for
explanation of these parameters). Note that the Cu surface is nearly three orders of
magnitude rougher than the PET. PET film is transparent, and its root mean square (rms)
roughness is in the nano-scale. These two factors make it nearly impossible for the op-
tical profiler to determine accurate roughness measurements. Hence an AFM was used
to measure the PET surface.

Fig. 4.2. TENG surface topography (a) Cu foil and (b) PET film. Note: the z-axis is in micrometers for
Cu and nanometer for PET due to the roughness difference [152].

Table 4.1: Areal surface topography parameters of TENG tribo-contact surfaces (see Appendix A for
information on parameters used) [152].
Contact material Center line average (+7 ) RMS roughness (+* ) RMS surface gradient
(µm) (μm)
Cu foil 4.8 6.1 0.84
PET film 0.0067 0.0095 0.17

4.2.2 TENG Electrical measurements

The setup in Fig. 4.1 (TENG with glass backing plates) was then installed in a custom
designed self-aligning test rig as shown in Fig. 4.3a. Fig. 4.3b shows the key detail via
a sectional view. The device in Fig. 4.1. (TENG (1) with glass backing plates (2)) is

91
attached to upper and lower smooth ground metal platforms (3 and 4). Again, because
the focus is on contact area, it is critical that the rig is designed to enable parallel contact
of the upper and lower TENG tribo-contact surfaces. Indeed, Hong et al (2018) [207]
found that open circuit voltage dropped by nearly 75% in their study for a misalignment
of only 1◦. In the rig in this study, rotation of the upper platform (3) is fixed, but the
lower platform (4) is free to rotate on a spherical bearing ball (5). Prior to testing, a pre-
load is applied to make contact and self-align the surfaces. The alignment is then
‘locked-in’ by tightening the lock screws (6).

Fig. 4.3. TENG self-aligning test setup: (a) Photo of test rig in mechanical test machine and (b) Sectional
view of the test rig. The numbered parts are (1) TENG device (Cu and PET), (2) Glass plates, (3) Upper
rig platform, (4) Lower rig platform, (5) Spherical bearing ball and (6) Locking screws [152].

Mechanical oscillation was supplied at 0.5 Hz via a mechanical testing frame (Instron
3367, USA) for contact loads ranging from 20 to 930 N. Note, a low operation
frequency ensures the load cell can correctly resolve the peak contact force. This is an
important consideration since we are studying the effect of contact force. At high
frequencies, it would be important to ensure a given load cell and data acquisition
system has the temporal resolution to capture the peak contact force. A key element of
the study to be able to characterize the force-dependent response over a sufficiency
wide range of applied forces: specifically, from very low forces up to forces required
for potential saturation of TENG electrical output. Therefore, for low forces (2–10N),
mechanical oscillation was supplied by an electrodynamic shaker system (TIRA, TV
50018, Germany) with a 20 N load cell capacity.

92
Separation distance was fixed at 2 mm and the open circuit voltage was recorded by an
oscilloscope (MSO-X 4154 A, KEYSIGHT, USA). The oscilloscope was connected
with an operational amplifier and a voltage divider circuit (resistances of 1 G Ω and 6
G Ω) to ensure that the impedance of the voltage meter setup was much larger than the
TENG internal impedance in Fig. 4.4(a) [141]. Details of the short circuit current
measurement approach are given in Fig. 4.4(b). Both tribo-contact surfaces were
cleaned prior to testing (using acetone and isopropanol initially, then rinsing in RO
water and finally blow dried using a nitrogen gun). During the test, the contact load was
progressively increased and, at each load value, TENG open circuit voltage and short
circuit current was measured after a TENG operation period of 250 cycles in order to
ensure equilibrium of tribo-charge generation.

(a) (b) 100KΩ

+V

Electrode A - 100Ω
osc
LMC
LMC 6001

Electrode B +
-V

Fig. 4.4: Schematic of equivalent circuit for (a) open circuit voltage measurements, and (b) short circuit
current measurements [152].

4.2.3 TENG Contact area measurements

Pressure sensitive film (SPF-A, Pressure Sensors Inc., USA) was used to determine the
TENG real contact area. The contact area measurement technique is outlined
schematically in Fig. 4.5. The film is placed between the TENG tribo-contact layers
(Fig. 4.5a) and contact force is applied via the test rig (Fig. 4.5b). The pressure sensitive
film consists of two PET layers: a donor layer and a receiver layer. Both donor and
receiver films where cut to 3 cm × 3 cm (larger than the 2.5 cm × 2.5 cm TENG area to
avoid edge effects). The donor film is coated with a thin layer of micro-capsules and a
colour developing layer is coated onto the receiver. Once the local contact pressure
exceeds a threshold value, the micro-capsules at these locations break and the capsuled
ink transfers to the receiver surface where it reacts with the developer layer to produce

93
red spots in areas of contact (see Fig. 4.5c). A sample of a tested film is shown in Fig.
4.5d. The lateral resolution of the measurement (2.6μm) is determined by the capsule
size (specifications for the pressure sensitive film are given in [133]). Following testing,
the film is scanned (with 1200 dpi resolution) to produce a digital image for analysis.
A MATLAB code is then used to threshold the image to generate binary images and the
total area of contact is calculated. The contact area test had to be carried out separately
to the electrical measurement test (as the pressure sensitive film will grossly alter the
electrical properties of the device). The surfaces were cleaned and aligned (for
parallelism) prior to testing. Load was then increased monotonically up to the same
load points as for the electrical measurements (i.e. there is no oscillation in the contact
area test). At each load point, the pressure sensitive film is removed and a new one is
inserted before proceeding to the next load point.

Fig. 4.5: Real contact area measurement using pressure sensitive film: (a) Pressure sensitive film
sandwiched between tribo-layers, (b) Compression of layers under contact force, (c) Pressure sensitive
film operation and (d) Sample tested film showing areas of contact in red/pink [152].

4.3 Results: electrical output and contact area

Fig. 4.6(a) plots the TENG open circuit voltage signals used for device characterization.
A signal is shown for 12 contact forces between 20 and 930 N corresponding to nominal
contact pressures of 32–1488 kPa. Note that, since the nominal contact area varies in
the TENG literature, it is necessary to use nominal contact pressure (i.e. force/nominal
area) to enable comparison. For a given contact force, the peak voltages in Fig. 4.6(a)
are almost constant indicating equilibrium of TENG operation. The peak positive
voltages are then taken from Fig. 4.6(a) and plotted (together with short circuit current)
against contact pressure (and force) in Fig. 4.6(b). In Fig. 4.6(b), open circuit voltage

94
increases with contact pressure and the upper end of the pressure range here is also
sufficient to saturate the open circuit voltage: i.e. beyond about 1176 kPa (753 N), the
voltage levels off at around 88 V. This is in-line with previous studies [137, 138, 141,
142] who have noted a similar response. Fig. 4.6(b) also indicates a very similar
response for short circuit current. What has not, so far, been elucidated very clearly is
exactly why electrical output increases with contact pressure. To answer this question,
we turn our attention to the real contact area at the interface.

Fig. 4.6. (a) Sample of TENG open circuit voltage signals for increasing contact force and (b) Peak open
circuit voltage and peak short circuit current versus nominal contact pressure (and force) [152].

Fig. 4.7(a) shows a sample of the pressure sensitive film results for the 32–1488 kPa
contact pressure range. The images on the left are the as-scanned images with the pink
spots representing real contact at the TENG interface. The images on the right are the
binary images with white representing real contact. The total real contact area Ar is
simply the sum of the white areas and the contact area ratio is defined as (Ar/ An) where
An is the nominal device contact area of 2.5×2.5 = 6.25 cm2. Note from Fig. 4.7(a) that
contact is spread relatively evenly over the nominal area and not biased towards any
particular edge or corner. This is a good indication that the careful measures taken to
ensure parallelism of the tribo-contact surfaces in Section 4.1.2 (i.e. glass backing plates
and self-aligning contact) have been successful. When the contact pressure is low (32
kPa case in Fig. 4.7(a), only a few points evenly distributed over the nominal contact
area are found to be in solid contact. As the pressure is increased, the area of solid

95
contact grows. The pressure sensitive film result for the maximum contact pressure case
in Fig. 4.7(a) (1488 kPa) is notably different to the 32 kPa result – here a rather large
amount of area is in contact. The increase is quantified in Fig. 4.7(b) where contact area
ratio is plotted against contact pressure. Notice just how small the real contact area is
at low contact pressures: at 32 kPa, the contact area ratio is only 0.0029 (i.e. 0.29%)
and, at 99 kPa, only 0.073 (i.e. 7.3%). This is important to bear in mind when we
consider that the vast majority of published work on TENGs has been carried out at
relatively low contact pressures.

Fig. 4.7: (a) Real contact area as detected by pressure sensitive film with increasing contact pressure
showing as-scanned film with pink indicating solid contact (left) and binarized image with white
indicating solid contact (right) and (b) Contact area ratio (Ar/An) versus nominal contact pressure (and
force). Note: the nominal contact patch size in (a) is 2.5 cm × 2.5 cm [152].

Table 4.2 compares the nominal contact pressure ranges used across a representative
selection of studies on contact-separation mode TENGs (with both engineered and non-
engineered surface topographies). All of the previous studies in Table 4.2 were
conducted at contact pressures below 200 kPa. This means that only the first two cases
(32 & 99 kPa) in Fig. 4.7(a) are within this range and we can see from the pressure
sensitive film that contact is still rather sparse at these pressures (for the material pair
here). Indeed, many of the prevalent TENG applications are only capable of applying
very small contact pressures (normal pressures in worn clothing are generally less than

96
10 kPa [208, 209]) and, although reasonable to a tribologist, it is likely that the TENG
community will not have been aware of just how little true contact area is generated
under these pressures.
In this work, we go beyond the 0–200 kPa range depicted in Table 4.2 and characterize
performance up to a contact pressure of 1488 kPa. To our knowledge, this is the first
time a TENG has been characterized over such a wide pressure range covering both
low and high pressures. Referring back to Fig. 4.7, we see that contact area grows
approximately linearly in the lower pressure regime and then tapers off at higher
pressures. Beyond about 1176 kPa (735 N), the contact area ratio saturates at about 0.82.
This kind of general evolution of real contact area versus contact pressure has been well
observed in the tribology literature [210, 211] and results essentially from the inevitable
presence of roughness on even the smoothest fabricated surfaces (i.e. only the extreme
peaks of the roughness make contact at first, more peaks then come into contact as the
contact pressure is increased until the number of new contact patches and the real
contact area saturates – see the roughness scans in Fig. 4.2). However, what is
particularly interesting here is the comparison of the contact area measurement with the
open circuit voltage and short circuit current measurement.

97
Table 4.2: Typical nominal contact pressures (force/area) used in TENG studies compared with the
present work (final row) [152].
Contact Materials Surface Contact Nominal contact Nominal contact Ref.
2
topography force (N) area (cm ) pressure (kPa)
PDMS & Al/Ag Engineered 0.5 to 7 2.9 1.7 to 24 [138]
CNF
Ag & PDMS Engineered 4 to 60 4.0 10 to 150 [137,
138]
Ag & PDMS Engineered 200 to 700 36.0 56 to 194 [140]
PTFE & C-PUF Engineered 1 to 12 36.0 0.3 to 3.3 [141]
Cu & PDMS Engineered 18 to 90 9.0 20 to 90 [142]
Al & PDMS Non-Engineered 40 to 220 25.0 16 to 88 [212]
Glass & PDMS Non-Engineered 10 to 50 25.0 4 to 20 [213]
PET & PDMS Non-Engineered 20 25.0 8 [59]
PI & Nylon Non-Engineered 30 to 60 6.25 48 to 96 [214]
Al & PDMS Non-Engineered 20 to 40 2.25 89 to 178 [215]
PET & PDMS Non-Engineered 4.3 to 5.1 3.75 11 to 13.6 [139]
Al & PVDF Non-Engineered 50 14.44 34.6 [216]
Cu & PET Non-Engineered 2 to 930 6.25 3.2 to 1488 [133]

Fig. 4.8(a) plots both open circuit voltage and contact area ratio against contact pressure
in the same graph. The max open circuit voltage increases from 14.9 V to 87.5 V as the
contact area ratio increases from 0.0025 to 0.82. What is particularly interesting are the
similarities in form between the two curves. In particular, both open circuit voltage and
contact area ratio saturate at about the same value of contact pressure (i.e. around 1176
kPa). A similar trend is seen in Fig. 4.8(b) for short circuit current. This essentially
suggests that the electrical output is governed by the real contact area. To gain an insight
into how this happens, we recall that the open circuit voltage Voc for the device (parallel
plat capacitor model) is given by:
)% .(()
,<4 = E 9 +*
(4.1)

where, x(t) is separation distance, εo is the permittivity of air and An is the nominal
contact area. Now, since these parameters are constant in our experiment, we can
conclude that the total tribo-charge QT is likely increasing with contact pressure.
Assuming tribo-charges can only transfer through the real contact area Ar, and defining
σT as the areal density of tribo-charge transfer through Ar, we can write:

98
6! = !! F% (4.2)
Taking σT as a constant for the material pair, Eqs. (1) and (2) suggest why Voc is
dependent on real contact area. The assumption that tribo-charges can only transfer at
areas of real contact is reasonable as it has been demonstrated in recent work [63, 144]
that contact electrification (via electron tunneling) requires the interatomic distance
between atoms across an interface to be within the equilibrium bond distance (i.e.
within the repulsive regime). Also, because we are dealing with dielectric materials,
this charge should remain localized in the real contact areas. We established in Fig. 4.7
that real contact area in the TENG is highly contact pressure dependent. Thus, taking
these two pieces of information together, it should follow that total tribo-charge, and
hence TENG electrical output, should be dependent on contact pressure via the real
contact area. Indeed, as we have seen, the results in Fig. 4.8 strongly point in this
direction as Voc and Isc appear to follow the evolution of real contact area with pressure.
Eq. (4.2), of course, assumes that tribo-charges are uniformly distributed and constant
over areas of real contact. In fact, the distribution of charges within the real contact area
is somewhat unknown and requires significant further study. However, a constant tribo-
charge density has generally been assumed in the literature and, in any case, even if it
is not constant, the σT in Eq. (4.2) can be considered as the net or equivalent value
leading to the same total charge. The blue hatched regions in Fig. 4.8 encompasse the
pressure range (0–200 kPa) from previous literature as depicted in Table 4.2. We can
see that a much wider pressure range has been investigated in the present work
revealing significant scope for increasing output at higher pressures. For example, at
200 kPa (the max pressure in Table 4.2), the contact area ratio is still less than 0.2, but
it reaches 0.82 beyond about 1168 kPa. Note: we have observed gains of 5.8 and 4.9
for VOC and ISC, respectively as Ar was increased from 0.25% to 82%. It is difficult to
compare this directly with the effect of other TENG parameters due to the variety of
material pairs etc. used across the literature. However, to provide some indicative
perspective: It is less than the gains of 50 and 8.5 reported in [59] for VOC and ISC as
separation distance was varied from 0.1 to 1 mm. On the other hand, the VOC gain of
5.8 is comparable with VOC gains of 2.6–8 reported with changes in tribo-layer thickness
99
(0.02–2 mm) in [125, 126], while the ISC gain of 4.9 is of a similar order to ISC gains of
2.4–6 reported with change in frequency (0.5–5 Hz) in [59, 127]. A detailed parametric
study would be required to determine a rank order of the sensitivity of TENG output to
the different parameters.

Fig. 4.8: (a) Open circuit voltage and contact area ratio (Ar/An) versus nominal contact pressure and (b)
Short circuit current and contact area ratio (Ar/An) versus nominal contact pressure. The blue hatched
regions represent the pressure range in previous literature (as depicted in Table 4.2) [152].

To provide a visual demonstration of the effect of increasing the real contact area (via
the contact pressure), 21 LEDs forming the word ‘BEST’ were connected in series with
the TENG as the power source (Fig. 4.9(b)). Contact pressure was then increased in the
same steps as before and LED performance was imaged via a digital camera at each
step. Fig. 4.9(a) illustrates how the LEDs get progressively brighter as the contact
pressure (and contact area ratio) increase. Contact pressure increases the real contact
area meaning that total charge can increase: this increases voltage and current, and
hence, the LEDs become progressively brighter until saturating at points 3 and 4 in Fig.
4.9(a).

100
Fig. 4.9: (a) LED array performance versus nominal contact pressure and force (and contact area ratio)
and (b) representative energy harvesting circuit. The LED array represents the word “BEST” – the LEDs
become progressively brighter as the contact pressure (and contact area) is increased [152].

4.4 Implications and applications of pressure and area dependence

There are a number of important implications that follow from the idea that TENG
electrical performance is dependent on a contact pressure-dependent real contact area.
The first is that device designers need to carefully account for the real contact area
dependence over the design range of contact pressures envisaged for a given device:
failure to account for this is likely to result in inaccurate predictions of device
performance. While the early TENG models [57, 59] did not account for contact
pressure dependence, a recent model by Yang et al (2020) [147] accounts for the effect
by including the rough surface contact mechanics governing the behavior in a unified
TENG model. Material properties, surface topography and contact pressure all
influence the real contact area and the model in Xu et al (2020) [147] uses input from
each of these to predict real contact area, and by extension, electrical output.
The continued boosting of TENG output at higher pressures characterized in the present
work also indicates possible new application areas for TENGs – what we might call
‘high pressure applications’. In fact, there are a number of sources of high-pressure
cyclic loading in the world around us. Table 4.3 summarizes the typical contact
pressures available with a selection of these sources. A particularly suitable and

101
abundant source is the vehicle wheel-on-surface interaction. For example, in trucks, a
contact pressure in excess of 1700 kPa is available at the tyre-road interface (Table 4.3).
Here, TENGs positioned either in the tyre or on the road surface could be used to power
a variety of sensors for integration with either the truck’s electronic systems or the road
infrastructure (e.g. traffic light sensors, warning lights etc.). Similar possibilities occur
with automobiles, locomotives and construction & agriculture-based vehicles. Another
suitable source is the repeating pressure available from sea waves. Here TENGs can be
adapted to take advantage of the higher pressures available for the purpose of generating
electricity. For example, pressures in excess of 100 kPa are available from wave
interaction with a stationary surface and pressures of almost 9000 kPa have been
observed for waves impacting a moving ship hull (Table 4.3).

Table 4.3: Typical contact pressures available in possible high load applications for TENGs [152].

High pressure application Contact pressure range Ref

Truck tire vs road 344 to 1723 kPa (range in central 8 cm) [217]

Locomotive wheel vs rail 700 to 1600 MPa (Peak), 117 to 1167 MPa (Mean) [218]

Automobile tire vs road 172 to 275 kPa (mean) [219]

Wave vs tilted slope (1:4) 50 kPa (mean pressure at max peak wave loading) [220]

Wave vs offshore vertical wall 110 kPa (mean pressure at max peak wave loading) [221]

Wave vs moving ship hull 2068 to 8963 kPa (mean pressure at peak wave [222,

(‘slamming’) loading) 223]

Previous literature has focused on boosting TENG output by developing highly

engineered surface topographies. Indeed, the present work helps to explain why so
many authors [142, 206, 216, 224-226] have noted improvements in TENG
performance based on modifications to surface topography – i.e. they were most likely
increasing the real contact area. An interesting point to add to this is that, achieving a
contact area ratio greater than unity is possible – imagine a hard-sawtooth type surface
in contact with a soft counter-surface. The sawtooth surface clearly has an area greater
than the nominal area and this can become the interfacial contact area when pressed
102
into a sufficiently soft counter-surface. However, the present work demonstrates that
we can also boost TENG performance with simple non-engineered surfaces. In
applications where sufficient pressure is available (Table 4.3), this is a far more cost-
effective route. Where contact pressure is limited, softer materials can be easily be
chosen so that contact area saturation occurs at lower pressures.
Another important implication relates to the definition and measurement of tribo-charge
density. The fact that Ar is generally less than An (and especially so at low contact
pressures) highlights an important distinction between the tribo-charge density as
defined in Eq. (4.2) (i.e. QT/Ar) and the conventional definition as QT/An. It means that
tribo-charge density defined as QT/An cannot be considered to be a fixed quantity (for
a material pair), but will depend on the contact pressure and even the surface roughness
(via the QT generated by the real contact area which is both pressure and roughness
dependent). Therefore, there is a need for clarity in the measurement of tribo-charge
density essentially to account for the important distinction between the contact pressure
dependent Ar and the nominal device area An. Current literature does not account for
this and there is, therefore, a strong need for further research focused on the
measurement of tribo-charge density. Another important consideration here might be
whether the interface materials are insulators or conductors. While tribo-charges will
remain localized at areas of solid contact for dielectric materials, tribo-charges for a
conductor (i.e. in a conductor-to-dielectric TENG) can be expected to distribute over
the entire nominal device area. Yet another consideration is how the distribution of real
contact area effects output – i.e. would identical real contact areas with differing spatial
distribution produce similar electrical output? This question also requires further
attention.
Finally, the results demonstrated in Section 4.2 mean that TENGs can be used as
effective load and pressure sensors. A careful calibration of output voltage or current
versus contact force (and operation within the elastic regime) is essentially all that
would be required for a TENG system to report loads and/or pressures. Example
applications include pressure/force measurement in vehicle weigh stations, road tyres,
shoes and at contact interfaces in vibrating machinery.
103
4.5 Conclusions
This chapter has investigated the origin of the contact force-dependent performance of
TENGs. Experiments were carried out where open circuit voltage and short circuit
current were measured together with real contact area under increasing contact
force/pressure with identical test conditions for a TENG involving two nominally flat
surfaces (PET and Cu) in contact separation mode. Digitized pressure sensitive film
was used to directly measure the real contact area – to the authors knowledge, this is
the first-time both real contact area and electrical output have been measured between
two nominally flat surfaces in a TENG device. Special precautions such as ultra-flat
backing plates and a self-alignment facility were shown to ensure the required contact
conformity at the interface. Results show that open circuit voltage, short circuit current
and real contact area all increase with contact pressure. For low contact pressures, the
amount of real contact area is tiny. For example, at 16 kPa (10 N on 25 × 25 mm2) the
percentage of real contact was only 0.25% of the active device area. Voltage, current
and contact area increase roughly linearly at first and later saturate to a converged value
at higher contact pressures. Real contact area at saturation (beyond about 1168 kPa)
was about 82% - this represents an enormous increase in contact area over the pressure
range investigated. Critically, both open circuit voltage and short circuit current
saturated at almost the same contact pressure as the real contact area suggesting that
electrical output follows the evolution of the real contact area. Recourse to simple
TENG theory reveals that, since the key TENG parameters (maximum separation
distance, material properties etc.) where constant during the experiment, the total tribo-
charge at the interface must have been increasing with contact pressure. The most likely
explanation; therefore, is that tribo-charges require real (or solid) contact in order to
transfer at an interface (a valid assumption given that recent work suggests that contact
electrification can only occur for atom pairs within the equilibrium interatomic bond
distance). Therefore, an increasing total tribo-charge (and hence electrical output) is to
be expected if real contact area increases with contact pressure. The reason why real
contact area increases with contact pressure is simply because nominally flat surfaces,
however smooth, inevitably contain random multi-scale surface roughness and the

104
mechanics of rough surface contact essentially leads to a contact pressure-dependent
contact area. Material properties, surface topography and contact pressure all influence
the contact area response.
There are several implications of these observations for TENGs. The first is that TENG
designers will need to account for pressure dependence over the design range of contact
pressures experienced by a device – otherwise predictions of output performance may
be inaccurate. This is especially important as most TENG devices operate in the low
contact pressure regime where contact area (as we see here) can be a surprisingly small
fraction of the nominal area even for very smooth surfaces (and is also very sensitive
to contact pressure in this regime). The chapter also essentially explains why so many
authors have been able to boost TENG performance by designing novel surface
topographies – in most instances, the increase in output is likely due to an increase in
real contact area. The distinction between real and nominal contact area leads to another
important point relating to tribo-charge density and its measurement. There now needs
to be clarity on whether tribo-charge density is calculated based on the real contact area
or the nominal active device area and the approach needs to be justifiable for a given
application. Pressure-dependence also means that (after a simple calibration) TENGs
can be used as force and pressure sensors (e.g. pressure/ force measurement in vehicle
weigh stations, road tyres, shoes or even at contact interfaces in vibrating machinery).
Finally, unlike in previous studies, the work characterizes the increasing TENG
output well into the high contact pressure regime (up to 1488 kPa). This leads to the
ability to use large contact pressures to boost performance without the need for costly
surface engineering. In addition, the results demonstrated that a CS-TENG based self-
powered pressure sensor can detect a wide range of pressure with high sensitivity, by
using the effect of changing Ar between the interface pair with the increase of contact
pressure. Note, the pressure sensitive film was used to quantify the Ar/An ratio. The
findings in this chapter were the impetus to developing the wide range flexible self-
powered pressure sensor in Chapter VI. The key outcome of this chapter is illustrated
graphically in Fig. 4.10.

105
Fig. 4.10: Overall view of the relationship between CS-TENG output and contact pressure [152].

106
Chapter V

Ferroelectric-Assisted High-Performance Triboelectric

Nanogenerators based on Electrospun P(VDF-TrFE) Composite

Nanofibers with Barium Titanate Nanofillers

5.1 introduction
A sustainable future of wearable electronic systems depend on assistance from energy
harvester and storage devices. Energy harvesters with flexible and/or stretchable form
factors are key to the effective use of autonomous systems and have found application
in a wide spectrum of areas such as: internet-of-things (IoT), wearable electronics,
rehabilitation, respiration monitoring, actuation and motion sensing, air purification and
gas sensing, and implantable biomedical devices [111, 150-153, 227-231]. As we have
seen, TENGs are flexible, efficient, and cost-effective energy harvesters. Chapter II
illustrated how TENGs have been explored in a wide range of applications, including
self-powered wearable electronics/sensors, powering of flexible electronics and even
for harvesting ocean wave/wind energy.
Chapter II has also reviewed the application of ferroelectric materials to enhance TENG
performance. Firstly, the polar β-phase of ferroelectric (FE) (e.g. polymers
polyvinylidene (PVDF) and copolymer poly(vinylidene fluoride-co-trifluoroethylene)
(P(VDF-TrFE))) have received particular attention due to their permanent polarization
of the tribo-contact surfaces. Secondly, to promote β-phase formation, electrospinning
is ideal as it would seem to combine uniaxial stretching and electric field poling in a
single step [108, 175, 232]. The presence of aligned dipoles in the β-phase of FE
polymers results in the enhancement of the spontaneous polarization and hence the
surface charge density [233] . High crystallinity can result in a higher dipole alignment
in FE polymers. Thus, because PVDF can be treated with lower crystallinity than its
copolymer under the same fabrication conditions, β-phase P(VDF-TrFE) was selected
as the tribo-negative material in the following experiments. Note, the crystallinity of β-

107
phase P(VDF-TrFE) could be achieved at 90% by suitable pretreatment. Thirdly, the
addition of filler material (e.g. BTO with formula BaTiO3) into the FE polymer matrix
can enhance the formation of a highly polar crystalline β-phase, because of the
electrostatic interactions between the nanostructures and the highly aligned dipoles of
the FE polymer chains [126, 128, 129, 170, 171] . Finally, as TENG output increases
with real contact area (Chapter IV), there is another potential advantage to having a
fibrous mat as one of the tribo-surfaces; namely, in certain cases, the flexible fibrous
mat may be better able to conform to the topography of the opposing tribo-contact
surface.
Therefore, this thesis develops high-performance ferroelectric-assisted TENGs using
electrospun fibrous surfaces based on P(VDF-TrFE) with dispersed BaTiO3 (BTO)
nanofillers in either cubic (CBTO) or tetragonal (TBTO) form. TENGs with three types
of tribo-negative surface (pristine P(VDF-TrFE), P(VDF-TrFE)/CBTO and P(VDF-
TrFE)/TBTO) in contact with PET were investigated and the effect on output is
explored. TENGs with P(VDF-TrFE)/BTO have been explored before [126, 128], but
only using a spin coated film. Here we explore the potential additional boost to output
in using electrospinning to develop a fibrous P(VDF-TrFE)/BTO mat. It is hypothesized
that electrospinning might increase dipole alignment due to high applied voltages, but
also encourage the formation of a highly oriented crystalline β-phase via uniaxial
stretching. The mechanism by which BTO fillers boost output is also explored
particularly by comparing the outcomes with CBTO and TBTO. Finally, the
ferroelectric-assisted TENG was integrated with a flexible graphene electrode-based
supercapacitor to produce a rapid self-charging system. The results demonstrated in this
chapter show that this technology can be valuable particularly in wearable applications.
5.2 Experiments
5.2.1 Preparation of Materials
The fibrous (electronegative) surfaces were electrospun from three different solutions:
pure P(VDF-TrFE) (Piezotech, France), P(VDF-TrFE) mixed with cubic BaTiO3
(CBTO, Sigma-Aldrich, UK), and tetragonal BaTiO3 (TBTO, Sigma-Aldrich, UK)
powders. The fabrication steps for the P(VDF-TrFE), P(VDF-TrFE)/CBTO and

108
P(VDF-TrFE)/TBTO solutions are illustrated in Fig 5.1. A 15 wt% P(VDF-TrFE)
solution was prepared by dissolving P(VDF-TrFE) powder in a DMF and acetone
(Sigma-Aldrich, UK) mixed solvent (1:1 ratio) for 10h under stirring by a magnetic stir

bar at 300rpm and 25℃ (see Fig. 5.1(a)). The resulting solution was stirred until the

solution was homogeneous and completely transparent. P(VDF-TrFE) and BTO

composite solutions were prepared by mixing CBTO or TBTO with the resulting
P(VDF-TrFE) solution at 3 wt% (Fig. 5.1(b-c)). Note, 3 wt% was chosen as this led to
optimum voltage and current output (See Fig. 5.2). The nanocomposite solutions were

then stirred at 300rpm at 25℃ for 4hr to disperse the CBTO or TBTO in the P(VDF-

TrFE) solution uniformly and form the P(VDF-TrFE)/CBTO and P(VDF-TrFE)/TBTO

slurry.

Fig. 5.1: Fabrication procedures for (a) P(VDF-TrFE), (b) P(VDF-TrFE)/CBTO and (c) P(VDF-
TrFE)/TBTO NF mats by electrospinning [49].

109
Fig. 5.2: TENG performance versus wt% of TBTO fillers in the tribo-negative P(VDF-TrFE)/TBTO layer
showing (a) open circuit voltage and (b) short circuit current density [49].

5.2.2 Electrospinning
Electrospinning of the NFs was conducted using an electrospinning machine. The
prepared solutions were filled into a 10mL syringe and connected to a spinneret with a

21G-gauge needle. After that, the electrospinning of solution was executed at a high

voltage of 17kV and a flow rate of 0.5 mL/hr with a 15 cm distance between the needle
and drum collector, as shown in Fig. 5.1(a-c). The electrospun NFs were collected by a
drum collector, which was covered by an aluminum foil (Sigma-Aldrich, UK), with a
2 kV voltage rotating at 1500rpm. All three polymer solutions (P(VDF-TrFE), P(VDF-
TrFE)/CBTO and P(VDF-TrFE)/TBTO NFs) were electrospun using the same
fabrication parameters at room temperature and humidity. Note, achieving optimum
electrospinning parameters for this application is not straightforward as formation of β-
phase P(VDF-TrFE) (critical for enabling higher polarization) is sensitive to several
parameters including voltage, concentration of P(VDF-TrFE), viscosity and distance
between needle and collector. To have NF mats with similar thicknesses, each of the
three NF mats were collected on the Al foil for around 2 hrs which produced nanofiber
mats with thickness of about 8 µm. The electrospun NF mats covering the Al foil were
then cut into small pieces (4 × 2.5 cm2) as shown in Fig. 5.3 and used as the negative
tribo-layer. A fixed ITO coated Polyethylene terephthalate (PET) film (4 × 2.5 cm2,
Sigma-Aldrich, UK) with a resistivity of 60 Ω/Uq and 0.127 mm thickness was used
as the positive tribo-layer for each of the three sample types (to keep the counter-surface

110
material constant). To avoid a bulge from the electrode lead wires interfering with the
contact area, the interface pair positioning was mismatched at a fixed nominal contact
area of 2.5 × 2.5 cm2 during the measurements (i.e., to allow a free region for
attachment of the wires). To measure the pure piezoelectric response, the P(VDF-
TrFE)/TBTO nanofiber-based mat was sandwiched between two Al electrodes and
encapsulated with a polyimide tape as shown in Fig. 5.4.

(a) (b)

PVDF-TrFE/TBTO PVDF-TrFE/CBTO

(c) (d) (e)

PVDF-TrFE PET

Supercapacitor

Fig. 5.3: indicates the purely piezoelectric output when a piezoelectric nanogenerators is fabricated for
the PVDF-TrFE/TBTO case [49].

Fig. 5.4: (a) Schematic of equivalent piezoelectric nanogenerator (i.e. piezo only) by PVDF-TrFE/TBTO
NFs, (b) output voltage signal with varying applied contact forces and (b) output voltage with varying
operation frequencies [49].

111
5.2.3 Characterization and testing
X-ray diffraction (XRD), employing a PANalytical X‘Pert Pro diffractometer with a Cu
Kα source (λ = 1.541 Å) was used to characterize crystallinity and the * − phase in
the P(VDF-TrFE) matrix for each of the three varieties of electrospun NFs. The surface
topography of electrospun NFs and PET surface was measured prior to testing by an
optical 3D surface profilometer (Alicona, InfiniteFocus) and an atomic force
microscope (AFM) (Dimension Icon, Bruker). A scanning electron microscopy (SEM)
was used to describe the surface structure and morphology of the electrospun NF mats.
The rig used for TENG testing (normal contact separation mode) is the same as
explained elsewhere [147] . Prior to testing, a pre-load was applied to make contact and
to self-align the surfaces. The alignment is then ‘locked-in’ by tightening lock screws.
The mechanical oscillation for the TENGs was supplied by an oscillating force (8N
with 4Hz frequency) using an electrodynamic shaker system (TIRA, TV 50018,
Germany). This produces a corresponding nominal contact pressure of 12.8kPa.
Mechanical oscillation was configured with a 2 mm max separation distance (i.e., peak-
to-peak amplitude of 2 mm). A 5 min pre-charge time is employed before each set of
measurements to ensure equilibrium of tribo-charge generation. The output voltage was
recorded by an oscilloscope (MSO-X 4154A, KEYSIGHT, USA) connected to an
operational amplifier and a voltage divider circuit to ensure that the impedance of the
voltage meter setup was much larger than the TENG internal impedance (See Fig. 4.4a
for details). Simultaneously, the oscilloscope was also coupled with an inverting current
to voltage converter which consisted of a 100kΩ feedback resistor and a LMC6001
operational amplifier (ultra-input current amplifier) to measure the output current in Fig.
4.4b. When comparing the three TENG devices with the different tribo-negative
surfaces (and when testing with different resistive loads) all test parameters (contact
pressure, frequency, separation distance etc.) were maintained constant.

5.2.4 Graphene supercapacitor

To demonstrate the self-charging capability of the developed Fe-TENGs, a flexible
supercapacitor (FSC) was fabricated using multilayer graphene sheets (GS, Graphene

112
Supermarket, USA) based on a previous report [234]. In a typical process, the GS film
having an area of 3 cm2 was used as the current collector and electrode material. At first,
the GS film was pasted on flexible polyethylene terephthalate (PET) substrates by
polyurethane (PU) resin as the binder. A conducting metal wire was fixed on top of the
graphene sheet by silver (Ag) conductive epoxy (RS components, 1863600) for the

external connection and insulated with PU coating and dried for 2 hrs at 80℃ in a hot

air oven. Polyvinyl alcohol (PVA)-phosphoric acid (H3PO4) based gel electrolyte was
used for the FSC fabrication. The preparation of the gel electrolyte was reported in
previous work [235]. Herein, first we dissolved the PVA (10 wt.%) in hot water with
continuous stirring followed by the addition of H3PO4. A polyester/cellulose blend
(Techni Cloth, TX 612) was used as a separator. A schematic representation of the
fabricated FSC is shown in Fig. 5.12(a). To charge the supercapacitor, the Fe-TENGs
and supercapacitor were coupled by a rectifier circuit which consists of four diodes as
shown in Fig. 5.12(b).
5.3. Result and Discussion
5.3.1 TENG performance
Fig. 5.5(a-c) illustrates the structure of the three TENG devices under investigation
having the tribo-negative material as pristine P(VDF-TrFE), P(VDF-TrFE)/CBTO
composite or P(VDF-TrFE)/TBTO composite (each in contact with PET). Device
output is characterized in Fig. 5.5(d-i). Fig. 5.5(d) and Fig. 5.5(e) compare voltages and
current densities for each of the three devices at open circuit and short circuit conditions,
respectively. Both outputs increase as we move from the pristine P(VDF-TrFE) case to
P(VDF-TrFE)/CBTO and to P(VDF-TrFE)/TBTO. The peak positive open circuit
voltage Voc increased from 225 V (pristine) to 293V (P(VDF-TrFE)/CBTO) and to 315
V (P(VDF-TrFE)/TBTO), i.e., 1.4 times increase from pristine to P(VDF-TrFE)/TBTO.
These voltages are comparable with previous work (Table 2.4). TENGs are high voltage
but low current devices and a critical requirement therefore is to boost the current
density. In Fig 5.5e, the peak short circuit current density was about 1.95 $A/cm2 for
the pristine case, 5.5 $A/cm2 for P(VDF-TrFE)/CBTO and a maximum of 6.7 $A/cm2
for P(VDF-TrFE)/TBTO representing an increase of 3.4 times over the pristine case.

113
 (j) (k)

Fig. 5.5: Characterization of device performance. (a-c) illustrate Fe-TENG device structure having
negative tribo-surface as: (a) P(VDF-TrFE), (b) P(VDF-TrFE)/CBTO and (c) P(VDF-TrFE)/TBTO (with

114
PET at the positive tribo-layer in each case). (a-c) also illustrate the role of ferroelectric polarization in
affecting the triboelectric surface potential and charge density by indicating how the bias of the bound
charges and the level of tribocharges increases as we move to P(VDF-TrFE)/TBTO (Note: XC is degree
of crystallinity & σT is tribocharge density). Electrical output is characterized in (d-i) for each of the three
device types as follows: (d) open circuit voltage, (e) short circuit current density, (f) output power density
and (g-i) output current and voltage versus resistive load for (g) pristine P(VDF-TrFE) & PET, (h)
P(VDF-TrFE)/CBTO & PET and (i) P(VDF-TrFE)/TBTO & PET. Contact pressure was fixed at 12.8
kPa. (j) transferred charge for three devices. (k) the stability of P(VDF-TrFE)/TBTO & PET device [49].

Additionally, a similar trend is seen in Fig. 5.5(j) for the transferred charges. The
transferred charges of three devices have been measured as 35nC, 56nC and 74nC by
an electrometer (Keithley 6517B, USA). The transferred charges result is illustrated in
Fig. 5.5(j). The resulting boost in the transferred charges by mixing BTO particles into
PVDF-TrFE can be attributed to higher crystallinity (and amount) of β-phase and
increased permittivity as outlined in the paper. Note, the variation of transferred charges
here coincide with trends for open circuit voltage and short circuit current in Fig. 5(d-
e). The comparison of our result with previous work on similar materials is given in
Table 2.4. For example, with a PVDF-TBTO based device in contact with PA6 in [126]
and a P(VDF-TrFE)/TBTO based device in contact with Aluminum in [128]. The
former achieved a maximum current density of 3.2 $A/cm2 at a contact pressure of 250
kPa and the latter reported a maximum of 1.25 $A/cm2 at a contact pressure of 98 kPa.
The present work has produced significantly higher current density (6.7 $A/cm2) at
much lower contact pressure of only 12.8 kPa. Interestingly, these gains are possible
with the counter surface (PET), which is far less tribo-positive than the tribo-positive
materials used in above mentioned previous works. In fact, 6.7 $A/cm2 is among the
highest values in Table 2.4 where the present work is compared to related work in the
literature (i.e. where similar materials were used). The contact pressure is an important
factor here as TENG output increases significantly with contact pressure. Therefore,
this marked boosting of the current density could be largely attributed to the effect of
the electrospinning process used here as previous works used spin-coated films with
[128] and without [126] electrical poling. Specifically, the electrospinning here used a
higher voltage (17 kV) compared to the poling used in [128] (8 kV). To illustrate this,
we have compared our electrospun results to those from equivalent DC poled spin

115
coated films. Four devices each having non-poled or poled PVDF-TrFE and non-poled
or poled PVDF-TrFE/TBTO as the negative tribo-layer (with a fixed PET at the tribo-
positive layer) have been fabricated for comparison of output from DC poled and
electrospun devices. To execute the polling process, high electric field (25MV/m) is
applied on the spin-coated PVDF-TrFE and PVDF-TrFE/TBTO films. Accordingly, the
dipoles are aligned with the direction of the external electric field. The increased dipole
alignment (i.e. sum of the aligned dipoles) delivers a high surface potential in the films
and attracts more surface tribo-charges to maintain the new equilibrium, resulting in
enhanced surface tribo-charge density [1].
Fig. 5.6 shows the result of the comparison. For pristine P(VDF-TrFE), output voltage
increased 1.5 times going from non-poled to poled and by 2.1 times going from poled
to electrospun (Fig. 5.6(a)). The output voltage of the device fabricated by electrospun
P(VDF-TrFE)/TBTO is 2.1 times higher than the equivalent poled PVDF-TrFE/TBTO
device in Fig. 5.6(b). The higher voltage in electrospinning is likely to encourage
enhanced dipole alignment in the P(VDF-TrFE) matrix. Also, uniaxial fiber stretching
during the electrospinning process is likely to result in more highly aligned, more
crystalline P(VDF-TrFE). These effects are likely to confer higher surface polarity (i.e.,
higher surface potential) and enable a further boosting of triboelectric charge transfer
leading to the higher currents observed. Electrospinning also has the advantage of
eliminating the extra poling step as this is now a seamless part of the process. Output
power is plotted against resistive load in Fig. 5.5(f) and the corresponding plots of
voltage and current versus resistance are shown in Fig. 5.5(g-i). Energy harvesting
devices such as TENGs generally present high internal impedance to load circuits and
in general, the maximum output power is obtained when the load impedance is matched
to that of the harvester. Across the three TENG device variants here, the optimum power
occurs in the resistance range of 100 to 500 MΩ. Following the voltage and current
trend, peak power increased from about 0.75 W/m2 for the pristine P(VDF-TrFE) to
about 2 W/m2 for P(VDF-TrFE)/CBTO and 2.75 W/m2 for P(VDF-TrFE)/TBTO - thus
representing 3.2 times increase over the pristine case. The developed Fe-TENG retained
approximately 95% of the original output voltage in Fig. 5.5(i) after 10000 cycles of
116
repeated vibration, as shown in Fig. 5.5(k) – thus the device exhibits good stability.

Fig. 5.6: Comparison of output voltage for the device with (a) PVDF-TrFE and (b) PVDF-TrFE/TBTO
by non-poling, poling and electrospinning processes [49].

Analyzing the intrinsic properties of the three negative tribo-layer materials can help
trace the source of the output differences between the three cases and improve our
understanding of the underlying behavior. It has been shown that the intrinsic properties
of P(VDF-TrFE) NFs govern the output performance of TENGs. This is because the
surface potential of the contact layer is correlated with the degree of crystallinity of the
*-phase of the P(VDF-TrFE) NFs [174, 216, 236]. To probe the crystallinity of the NFs,
X-ray diffraction (XRD) measurements were performed on all three categories of
P(VDF-TrFE) and its composite NF samples (results are shown in Fig. 5.7(a)). From
the XRD data, we can see a dominant peak centered at 2v=19.96° for all three samples,
corresponding to the reflection of the crystalline *-phase (110/200) of P(VDF-TrFE).
It is important to note that none of the samples showed the non-polar α phase (present
at 2θ = 18.3). These results also suggest good dispersion of BTO nanoparticles in the
P(VDF-TrFE) matrix. Among the three phases of P(VDF-TrFE) namely, alpha, beta and
gamma, the β-phase can generate highest surface tribo-charge density because it
provides the highest dipolar moment per unit cell (i.e. surface dipole moment) [237,
238]. The characteristic diffraction peak of BTO nanoparticles at 2v=22.2° is also
apparent in Fig. 5.7(a) for the P(VDF-TrFE) composite NFs (in blue for TBTO and
black for CBTO). In theory, the XRD diffraction of the tetragonal BTO show peak

117
splitting at 22.2° corresponding to the (hkl) Miller index (001) and (100), whereas cubic
BTO would have a single peak corresponding to (002) [239]. The P(VDF-TrFE)/TBTO
NFs diffraction pattern (blue line) exhibits typical characteristics of tetragonal BTO
(peak splitting at 22.2 degree) with a single peak at 2v = 22.2° for the CBTO sample.
The cubic phase of BTO belongs to the centrosymmetric crystal class m3m and thus, is
non-piezoelectric [240].

(a) (b) 27
24
PVDF-TrFE
PVDF-TrFE/CBTO

Dielectric constant
PVDF-TrFE/TBTO
21
18
Intensity (a.u.)

15
12
9
6
3
0
0.0 0.5 1.0
Two theta (degree) Frequency (MHz)
(c) 0.5 (d)
PVDF-TrFE/CBTO
PVDF-TrFE/TBTO
0.4
Absorbance [a.u.]

1400 (b)

0.3
850 (b)

1280 (b)
0.2
763 (a )

0.1

0.0

300 600 900 1200 1500

Wavenumber [cm-1]

Fig. 5.7: Materials characterization for P(VDF-TrFE) and P(VDF-TrFE)/BTO: (a) X-ray diffraction
(XRD) results of P(VDF-TrFE) (red line), P(VDF-TrFE)/CBTO (black line) and P(VDF-TrFE)/TBTO
(blue line) with crystalline β-phase (110/200) and XRD patterns of the cubic and tetragonal BTO
nanoparticles (001/100); (b) as-measured dielectric properties of the nano-fibrous mats for: P(VDF-
TrFE), P(VDF-TrFE)/CBTO, and P(VDF-TrFE)/ TBTO. Note: there is little difference in permittivity
between CBTO and TBTO nanofillers meaning that permittivity cannot explain the TENG output boost
with TBTO (c) : FTIR spectra of P(VDF-TrFE)/BTO composite nanofibers with P(VDF-TrFE)/CBTO
(red) and P(VDF-TrFE)/TBTO (blue) and (d) as-measured tangent loss of the nanofibrous mats for:
P(VDF-TrFE), P(VDF-TrFE)/CBTO and P(VDF-TrFE)/TBTO [49].

Exploring both CBTO and TBTO in the same work allows us to probe the underlying
behavior further. We noted from Fig. 5.5, that voltage, current and power increase
progressively as we move from pristine P(VDF-TrFE) to P(VDF-TrFE)/CBTO to
P(VDF-TrFE)/TBTO. For example, the current density increases 2.8 times in going

118
from the pristine case to P(VDF-TrFE)/CBTO. If we then go to P(VDF-TrFE)/TBTO,
we notice a further boosting of output to 3.4 times the pristine case. To explain the
increases attributable to the BTO nanofillers, there are two key mechanisms: the first
(well outlined in [128] and [126]) is that the addition of high permittivity fillers (i.e. the
CBTO) dramatically increases the permittivity of the resulting P(VDF-TrFE)/BTO
matrix (for example, a factor of 6 to 7 increase in permittivity is reported after addition
of CBTO nanofillers to P(VDF-TrFE) [241]). Higher permittivity contact materials
increase TENG capacitance, thereby increasing the ability to induce and store more
tribo-charge and drive higher currents. Permittivity measurements for the three nano-
fibrous mat types in this work are given in Fig. 5.7(b). The results show the expected
boost in permittivity due to the BTO fillers, and also confirm that CBTO and TBTO
result in near identical permittivity values. Therefore, to explain the higher TENG
output with TBTO (compared to CBTO), we turn to the second reason and the one we
would like to focus on more in this paper.
In a recent work on piezoelectric sensors, the addition of CBTO nanofillers into a
P(VDF-TrFE) matrix has been noted to significantly boost *-phase crystallinity by
becoming advantageous nucleation sites for the crystalline β-phase [242]. This
therefore increased residual polarity and piezoelectric coefficient. Here we explore this
role of the BTO nanofillers in increasing crystallinity and how this might be
advantageous in the TENG context. The crystalline quality is a parameter that can be
determined by calculating the full width at half maximum (FWHM) [243]. Note, a
smaller value of FWHM corresponds to the higher β-phase crystallinity. We calculated
the FWHM values for pristine P(VDF-TrFE), P(VDF-TrFE)/CBTO and P(VDF-
TrFE)/TBTO NFs and these were 0.95, 0.89 and 0.81, respectively. Based on these
results, the P(VDF-TrFE)/TBTO NFs have highest crystallinity among the three
electrospun materials. We also noted via FTIR Spectroscopy analysis of the P(VDF-
TrFE)/BTO that the amount of β-phase was higher for TBTO nanofillers (67%) than for
CBTO (54%). Fig. 5.7(c) shows the absorbance spectra for the P(VDF-TrFE)/CBTO
and P(VDF-TrFE)/TBTO composite mats obtained by FTIR Spectroscopy. We note the
peaks at 850 cm-1, 1280 cm-1 and 1400 cm-1 corresponding to the *-phase and the x-
119
 E
phase peak at 763 cm-1. Using the Beer-Lambert law F(*)=1.2]E P6E , (F^ and F_ are
Q P

x-phase and *-phase absorption) the amount of β-phase, F(*), for the CBTO and
TBTO composites was determined at 54% and 67%, [Link] existence of *-
phase with higher crystallinity and in higher proportions will improve dipole alignment
and polarization in the P(VDF-TrFE)/BTO matrix leading to an increase in the surface
potential of the tribo-negative contact material. Estimates of surface potential outlined
in Fig. 5.8 follow this trend. Estimated surface potential for pristine P(VDF-TrFE) was
133 V and this increased to 189 V with addition of CBTO fillers and was further boosted
to 214 V for TBTO fillers.
Fig. 5.8 shows simulated estimates of surface potentials for the TENG devices having
each of the three types of tribo-negative nano-fibrous mats (PVDF-TrFE, PVDF-
TrFE/CBTO and PVDF-TrFE/TBTO). The finite element multiphysics software
COMSOL was used to estimate the surface potential distributions given the key
material and surface properties of the tribo-layers and the TENG dimensions and
separation distance. The negative tribo-layer consisted of 8 µm thickness of PVDF-
TrFE, PVDF-TrFE/CBTO or PVDF-TrFE/TBTO, while the positive tribo-layer was a
127 µm thick PET sheet. The lateral dimensions were 2.5 × 2.5cm2 and the max
separation distance was 2 mm. The key triboelectric parameter required by the
simulation is the tribo-charge density at the interface. This was estimated from the
experiments as follows. The voltage drop V across the TENG is:
)R )R :)% )R
,=E R1 + 1(3) + E R2 , (5.1)
9 +" E 9 +! 9 +$

were d1, x(t) and d2 are the tribo-positive dielectric thickness, air gap thickness and
tribo-negative layer thickness, respectively; ε1, ε0 and ε2 are the dielectric constants; An
is the nominal device area and QI and QT are the total induced charge (on the electrodes)
and the total tribo-charge at the interface, respectively. Therefore, the induced charge is
related to the tribo-charge in short circuit mode (i.e. V = 0) as:
)R ) 1
= E% × 1 0 1 0 (5.2)
E9 9 16 ! " 6 ! $
1" ;()) 1$ ;())

When: #$ < #1 ≈ #2 and R1 ≈ R2 ≪ 1(3) (which is the case here), Eq. (5.2)
approximates to:
120
61 ≈ 62 (5.3)
Therefore, we can use the induced charge to approximate the tribo-charge. The induced
charge was calculated from the experimental short circuit current ISC using:
) , 7(()
753 = (, R 6.(())
*
$
(5.4)
*

Since COMSOL assumes complete contact over the full device area, the tribo-charge
density for the COMSOL simulation was calculated in the average sense as:
"
!! = #= (5.5)
>

This led to tribo-charge densities of 0.53, 1.4 and 1.7 μC/cm2 (for PVDF-TrFE, PVDF-
TrFE/CBTO and PVDF-TrFE/TBTO, respectively). With the σT estimated from
experiment, the COMSOL simulated potential distributions are given in Fig, 5.8 for
PVDF-TrFE (left), PVDF-TrFE/CBTO (middle) and PVDF-TrFE/TBTO (right). The
estimations here confirm that surface potential is increased by the addition of CBTO
fillers (by 42.1%) and further increased by the addition of the TBTO fillers (a further
13.2%). As noted, the extra output available with the TBTO nanofillers (compared with
CBTO nanofillers) seems to be entirely attributable to the crystallinity effect (as both
P(VDF-TrFE)/CBTO and P(VDF-TrFE)/TBTO produce very similar permittivity – Fig.
5.7(b)). We can thus conclude that, in both cases (CBTO and TBTO nanofillers), the
TENG enhancement mechanism is a combination of enhancing both permittivity and
β-phase crystallinity.

Surface: Electric potential (V) Surface: Electric potential (V) Surface: Electric potential (V)

PET PET PET

141V 223V 252V

133V 189V 214V

PVDF-TrFE PVDF-TrFE/CBTO PVDF-TrFE/TBTO

Fig. 5.8: COMSOL simulated potential distributions for PVDF-TrFE (left), PVDF-TrFE/CBTO (middle)
and PVDF-TrFE/TBTO (right) devices. Surface potential is clearly highest for the TBTO case [49].

The loss tangent for P(VDF-TrFE)/TBTO, P(VDF-TrFE)/CBTO and P(VDF-TrFE) has

1
been calculated by using: tanδ = 2BL4 . Note, the <Y and ~Y were tested using a
S `S

121
LCR meter (Keysight E4980, UK). They are the parallel mode capacitance and
resistance of the sample as illustrated in Fig. 5.7d. The loss tangent is lower in the
P(VDF-TrFE)/BTO composites compared with the pristine P(VDF-TrFE), especially at
the higher frequencies in Fig. 5.7(d), because of the glass transition relaxation of the
PVDF-TrFE polymer matrix [244, 245]. However, the loss tangent of P(VDF-TrFE)/
TBTO and P(VDF-TrFE)/ CBTO result in near identical values.

To fully understand how enhanced residual polarization of the P(VDF-TrFE) and its
composites boosts TENG output, it is useful to refer to the overlapped electron-cloud
(OEC) model [157, 246]. This model describes the process of intrinsic electron transfer
between two contact materials. Using the OEC model, Fig. 5.9 schematically illustrates
the electron transfer process when each of the three tribo-negative materials are pressed
into contact with the tribo-positive PET. In Fig. 5.9, EF, P(VDF-TrFE), EF, P(VDF-TrFE)/CBTO and
EF, P(VDF-TrFE)/TBTO represent the Fermi levels of P(VDF-TrFE), P(VDF-TrFE)/CBTO
and P(VDF-TrFE)/TBTO, respectively and the greatest energy level difference between
these surfaces and PET are represented by ∆", ∆"′ and ∆"′′, respectively.
P(VDF-TrFE) and its composites play the role of the tribo-negative material, accepting
electrons from PET during contact electrification. The combination of polarization due
to electrospinning and due to enhancements in crystallinity is likely to lead to high
dipole alignment resulting in high surface potential. Simultaneously, the presence of
bound charges, generated by aligned dipoles, could impact the potential energy [216].
As confirmed by the XRD scans, the P(VDF-TrFE)/TBTO sample has shown higher
crystallinity than P(VDF-TrFE)/CBTO and P(VDF-TrFE)/CBTO has shown better
crystallinity than pristine P(VDF-TrFE). Accordingly, as shown in Fig. 5.9, the Fermi
levels should be in order of EF, P(VDF-TrFE) > EF, P(VDF-TrFE)/CBTO > EF, P(VDF-TrFE)/TBTO. As
shown in Fig. 5.9(a), when the PET is not in contact with pristine P(VDF-TrFE) NFs,
the electron clouds of the materials do not overlap and there is negligible electron
transfer between the two atoms. When contact occurs, the potential barrier between
them (ΔE) decreases, and the electrons can transfer between the two atoms until their
Fermi levels are in equilibrium (i.e. until "′M,GN! = "′M,GWHM:!%MN ). After
122
incorporating the CBTO/TBTO nanoparticles into the P(VDF-TrFE) matrix, the Fermi
level of the negative tribo-layer can be expected to shift downwards due to enhanced
surface polarization. As the Fermi level of the PET sheet ("M,GN! ) is fixed in each case,
the energy difference is in the following order: ∆" < ∆"′ < ∆"′′. A greater difference
in energy levels means more empty surface states of the negative tribo-layer can be
filled with electrons from the PET surface (due to the increased differences in work
function between the contact materials) and hence higher tribo-charge transfer. After
separating the interface pair, the tribo-charge density of a dielectric-dielectric TENG is
given by [139]:
S (Z$ :Z" )
!! = " " $8$ ;! (5.6)
6 6
D." D.$ 1!

Where, P1,2 and h51,52 are the work function and average surface states of the
positive and negative tribo-layer, respectively, g is the elementary charge, #$ is air
permittivity, and 1$ is the separation distance. Due to the increased differences in
work function and surface states of the negative tribo-layer (i.e., h52 ), Eq. 5.6 shows
how the tribo-charge density can clearly be boosted. It is important to note that TBTO
has a higher dipole moment than CBTO which is non-piezoelectric. Thus, it can be
expected to promote higher crystallinity and better dipole alignment similarly to how
high dipole solvents have been shown to increase crystallinity and dipole alignment for
P(VDF-TrFE) [236]. Thus, in the case of P(VDF-TrFE)-TBTO, the Fermi level of the
negative triboelectric material can be expected to further shift to a lower level due to a
higher surface potential. Thus, this case is expected to show the largest transfer of
electrons between the two materials. This appears to be borne out in the TENG
electrical output (Fig. 5.5) and the tribo-charge transfer result in Fig. 5.5(j) where the
TBTO composite gives the highest output. The effect of polarized dipoles in affecting
the output performance of Fe-TENGs is illustrated schematically in Fig. 5.5.

123
Fig. 5.9: (a) Overlapped electron cloud (OEC) model of electron transfer between interface pairs for (a)
P(VDF-TrFE) & PET, (b) P(VDF-TrFE)/CBTO & PET and (c) P(VDF-TrFE)/TBTO & PET [49].

To investigate the contribution of conventional piezoelectricity to the TENG output, we

tested a P(VDF-TrFE)/TBTO device in piezoelectric mode only. The result is shown in
Fig. 5.4. Here we see that output voltage (1.25 V) at max contact force (8 N) is less than
1% of the corresponding output (315 V) in TENG mode (see Fig. 5.5d). Thus, the
TENG devices in this work are referred to as ferroelectric-assisted TENGs rather than
hybrid TENGs as the piezoelectric contribution can be considered relatively
insignificant. All the above effects (increased permittivity and higher surface
polarization) essentially combine in the case of the electrospun P(VDF-TrFE)-TBTO
NFs to boost tribo-charge density and deliver the best output performance in the Fe-
TENGs device (Fig. 5.5).

124
5.3.2 Pressure sensitivity, topography and contact area
We turn now to some other key factors that affect the output of these ferroelectric-
assisted NF based TENGs. We know that TENG output is generally sensitive to contact
pressure for TENGs having solid rough surfaces [133, 147]. Fig. 5.10 shows that this is
also the case for the fibrous P(VDF-TrFE) matrix in contact with PET here. At this low-
pressure range (3.2 to 12.8 kPa), we observe a roughly linear increase of voltage with
pressure. The calculated pressure sensitivities (S) for pristine P(VDF-TrFE), P(VDF-
TrFE)/CBTO and P(VDF-TrFE)/TBTO are 27, 28, and 32 V/kPa, respectively.
Interestingly, the obtained sensitivity value for P(VDF-TrFE)/TBTO (S = 32 V/kPa) is
higher than previously reported for TENG-based pressure sensors [77]. As expected,
the P(VDF-TrFE)/TBTO plot shows higher voltages than P(VDF-TrFE)/CBTO and the
P(VDF-TrFE)-CBTO plot gives higher voltages than the pristine P(VDF-TrFE) case.

Fig. 5.10: Open circuit voltage versus nominal contact pressure [49].

We know that contact area also affects TENG output (Chapter IV) via the contact
pressure and, although the contact pressure was fixed at 12.8 kPa for the electrical
results in Fig. 5.5, it is still possible that differences in the topography or elastic modulus
of the fibrous (electrospun) surfaces could be responsible for the differences in output
already discussed. To evaluate this prospect, we imaged the topography of each surface
type and used the boundary element method to estimate the real contact area. The
topographies of all three P(VDF-TrFE) based nanofiber mats were measured using an
optical profilometer (Infinite Focus, Alicona) prior to testing and the surface roughness

125
of the PET film was measured by Atomic Force Microscopy (AFM, Bruker, US). The
roughness height contours of pure P(VDF-TrFE), P(VDF-TrFE)/CBTO and P(VDF-
TrFE)/TBTO NF mats are shown in Fig. 5.11(a-c), respectively. The root mean square
roughness (Sq) of the three types of P(VDF-TrFE) mats and the PET film are given in
Table 5.1. It is nearly impossible to optically measure the real contact area between
these two thin layers using the current test rig, thus, the boundary element method
(BEM) [247] is utilized to simulate the linear elastic quasi-static contact between PET
and each of the three types of P(VDF-TrFE) based NF mat. Since the sampling area
(0.5 × 0.7 mm2) measured by the optical profilometer is much smaller than the
nominal contact area of the TENG devices (25 × 25 mm2), the periodic assumption is
adopted in the BEM assuming that the sampling area is periodic along two orthogonal
scanning directions. Since the Sq of the PET film is approximately 143 times smaller
than the average Sq of all three P(VDF-TrFE) NF mats (see Table 5.1), the rough
surface contact simulation can be simplified to a rough-on-flat scenario. Note, the
roughness of the positive tribo-layer was fixed by reusing the PET sheet. The surface
morphology of the electrospun nanofibers for the three cases of tribo-negative layers
were characterized using the SEM and are shown in Fig. 5.11(d-f). Each of the three
types of NF mats exhibited a completely fibrous structure with uniform distribution of
nanofibers and an average fiber diameter of 250 nm. As shown in Fig. 5.11(e) and Fig.
5.11(f), the BTO nanoparticles are inserted into the nanofibers uniformly and particles
can be seen on the surface of the fibers. We see from Table 5.1 that the mats with the
BTO particles have higher Sq roughness (1.7 and 1.8 µm) than the pristine case (0.57
µm).

126
Fig. 5.11: Surface topography and contact analysis: surface height contour plots of (a) P(VDF-TrFE), (b)
P(VDF-TrFE)/CBTO and (c) P(VDF-TrFE)/TBTO. SEM images of nanofiber surfaces for (d) P(VDF-
TrFE), (e) P(VDF-TrFE)/CBTO and (f) P(VDF-TrFE)/TBTO (inset images at higher magnification).
BEM prediction of real contact area (in black) for: (g) P(VDF-TrFE) & PET, (h) P(VDF-TrFE)/CBTO
& PET and (i) P(VDF-TrFE)/TBTO & PET. Note how the predicted contact area follows the peaks of
the fibers in the fiber orientation [49].

The BEM results for real contact area on the Fe-TENGs interfaces (at 12.8 kPa) are
illustrated in Fig. 5.11(g-i) in the form of binary images where black pixels represent
contact [Link] binary images indicate that: (1) the contact commonly occurs on the
higher peaks of the PVDF mats and its composites and, (2) the contact area distributions
reflect the orientations of the clusters of high peaks on the rough surfaces. The total real
contact area (F% ) is simply the sum of the black areas and the contact area ratio is
defined as F% /Fa where Fa is the (total) sampling area of the rough surface. The
BEM calculation for each device is repeated three times using three different rough
surface data measurements at different locations over the same P(VDF-TrFE) based NF
mats. The BEM predicts a contact area ratio of 10.8% for P(VDF-TrFE), 6.7% for
P(VDF-TrFE)/CBTO and 5.6% for P(VDF-TrFE)/TBTO. This corresponds with the
surface roughness trend – i.e., the rougher the surfaces, the lesser the contact area.

127
Considering the electrical outputs now, we note that, although the contact area
prediction is highest for pristine P(VDF-TrFE), electrical output is lowest for this case.
Also, electrical output is significantly higher for P(VDF-TrFE)-TBTO compared to
P(VDF-TrFE) but this is not the case for the predicted real contact areas which were
similar (TBTO actually predicting a slightly lower contact area). Thus, we can rule out
contact area as an explanation for the trends already observed in the results based on all
devices based on fibrous electro-negative material.

Table 5.1: Areal root mean square surface roughness of the tribo-contact surfaces and predicted contact
area percentage (BEM) of the Fe-TENGs tribo-contact pair [49].
Negative tribo-layer RMS Roughness Average contact area Positive tribo-layer (PET
(Sq) ("#) percentage (Ar/An sheet) RMS Roughness (Sq)
×100) (+#)
P(VDF-TrFE) 0.57 10.8% 9.5
P(VDF-TrFE)/CBTO 1.7 6.7% 9.5
P(VDF-TrFE)/TBTO 1.8 5.6% 9.5

5.3.3 Ferroelectric-assisted TENG powered high performance self-charging

system
Finally, the ability of the best performing Fe-TENGs (i.e., P(VDF-TrFE)/TBTO & PET)
to charge an integrated supercapacitor was investigated. The purpose of demonstrating
charging of the supercapacitor with the Fe-TENG device is to show its self-charging
ability for applications like wearable electronics. In such applications, a supercapacitor
is more suitable compared to a common electrolytic capacitor due to its very high
energy density, flexibility and ease of integration. A graphene sheet (GS) based flexible
supercapacitor (FSC) [234] with PVA-H3PO4 gel electrolyte was used as shown in Fig.
5.12(a). The electronic circuit which connects it with the Fe-TENGs via a rectifier is
shown in Fig. 5.12(b). In the self-charging system, the Fe-TENGs will generate the free
charges which are stored in the graphene sheet supercapacitor (GSC) through the
rectifier. The purpose of the rectifier is to convert the alternating current (AC) output
of the Fe-TENGs (see Fig. 5.5(d-e)) to direct current (DC) and hence provide a
continuous electrical power supply to store the energy into the GSC.
Fig. 5.12(c) shows rapid charging (~5 min) of the GSC with the Fe-TENGs under
128
different applied forces with a fixed vibration time (5 min) and frequency (8Hz)
followed by a slower self-discharging. It is evident from Fig. 5.12(c) that with increase
in contact forces between 2 to 8N (i.e., 3.2 to 12.8 kPa), the maximum output voltage
of the SC also increased roughly linearly. This is further evident from Fig. 5.12(d)
which shows the variation in maximum output voltage with nominal contact pressures
(i.e., contact force/nominal area). It was observed that the voltage could be improved
from 0.75V to 1V, when the contact pressure increased from 3.2 to 12.8 kPa. A similar
observation can also be seen in Fig. 5.12(e) and Fig. 5.12(f), when the operation
frequencies varied between 2 to 10 Hz at a fixed time (5 min) and contact pressure 12.8
kPa (8N) – i.e., the output voltage also increases roughly linearly with frequency. This
is due to the increased output current with increasing TENG operating frequency [59]
which can be calculated as:
,)
7= (5.7)
,(

Where, Q is output charge and t is generation period. To increase confidence in the

results, three GSC devices with near identical specifications were fabricated and the
experiments repeated, with results shown in Fig. 5.12(d) and Fig. 5.12(f). Note that the
error in Fig. 5.12(d) and Fig. 5.12(f) is due to expected variation between the three
separately fabricated GSC devices rather than instability in a single device.

129
Fig. 5.12. (a) Schematic of graphene electrode based flexible supercapacitor (SC), (b) energy harvesting
circuit for self-charging system (i.e. integration with the Fe-TENGs), (c) charging and discharging
voltage of SC at different contact forces (frequency of 8Hz), (d) the maximum value of voltage charge
into the SC versus nominal contact pressure, (e) charging and discharging voltage of SC by different
frequencies (contact pressure 12.8 kPa), (f) the maximum value of voltage charge into the SC versus
vibration frequencies, (g) representative energy harvesting circuit of Fe-TENGs for lighting LEDs and
(h) photograph of lighted LED array representing the word ‘‘BEST’’ [49].

Table 5.2 summarizes the performance of the present device with the related TENG
based self-charging supercapacitor systems in the literature. In the present work, we
could charge the GSC to 1.25V voltage within the short time period of about 5 min.
This is a considerably lower charging time then the other TENG based supercapacitor

130
systems in Table 5.2 which vary from 20 to 180 min, thus showing the compatibility of
presented devices as an excellent energy harvesting and storage system. For instance,
the Polypyrrole paper-based SC charged by a TENG interface pair of PPy/Cellulose
paper and Nitrtocellulose membrane exhibited a maximum charged voltage value of
1.5V in 167 min at a frequency of 10 Hz (similar to this work) and under much higher
contact pressure of 30kPa [198]. In the present work, we were able to charge the SC in
a considerably lower charging time with much lower contact pressure, thus indicating
the high efficiency of the self-charging system. The rapid charge time is due to high
electron mobility in the graphene electrodes which reduce SC impedance as well as the
high current being produced by the Fe-TENGs. However, during the self-discharging
of the GSC, a rapid drop in voltage to about 0.4V, followed by a slow decrease, is
apparent (Fig. 5.12(c) and Fig. 5.12(e)). This high drift in potential or fast discharging
indicates the low efficiency of charge storing capacity of the graphene electrode. This
is because of the low electroactive surface area, low electrochemical activity and
hydrophobicity of graphene electrodes [234]. However, this may still be sufficient for
wearable applications as often low power sensors (nanowatt to microwatt power range
[200-203]) are needed to be powered. The charging of three series connected GSCs to
a maximum 2.3V voltage with the presented Fe-TENG devices has also been
demonstrated. To demonstrate the high energy harvesting efficiency of the Fe-TENGs
supercapacitor system, we have lighted 40 LEDs connected in series forming the word
‘BEST’ as shown in Fig 5.12g. In this case, the applied contact pressure was 0.8 kPa at
a frequency of 8 Hz. The LEDs are bright even at this low contact pressure. Hence the
Fe-TENGs/supercapacitor system produces usable output even at low contact pressures
such as those available in wearable applications.

131
 Table 5.2: Typical charging performance and charging period for TENG based self-charging systems in
the literature as compared with the present work (final row) [49].
TENG interface pair SC electrode & electrolytes Pressure Frequency Period ,TUV,XY Ref.
(kPa) (Hz) (min) (V)
C4F8 & ITO Electrode: CNT/paper - 8 180 0.9 [197]
Electrolytes: PVA/H3PO4
PTFE & ITO Electrode: Cu sheet 40 2 20 0.06 [248]
Electrolytes: [BMIM] [BF4]/ ETPTA
LIG & PTFE Electrode: LIG - - 117 3 [196]
Electrolytes: PVA/H2PO4
PTFE & Al Electrode: peptide-Co9S8 - 1 167 1.78 [249]
Electrolytes: KOH/PVA
PVDF-HFP & steel Electrode: CoFe2O4 and Fe-RGO 44 3.8 32 1.6 [250]
3+
Electrolytes: Fe -PAA/KOH
PPy/Cellulose paper & Electrode: Polypyrrole 30 10 167 1.5 [198]
Nitrocellulose Membrane Electrolytes: PVA/H3PO4
P(VDF-TrFE)/TBTO NFs & Electrode: Graphene 12.8 10 5 1.25 This
PET Electrolytes: PVA/H3PO4 work

5.4 Conclusions
In this chapter, novel ferroelectric-assisted TENGs (Fe-TENGs) were fabricated
whereby the tribo-negative surface is based on electrospun P(VDF-TrFE) composite
nano-fibrous mats (having the advantageous *-phase) with dispersed BaTiO3 (BTO)
nanofillers in either cubic (CBTO) or tetragonal (TBTO) form. The approach used
electrospinning to significantly increase triboelectric output by boosting surface
polarization. TENGs with three types of tribo-negative surface were investigated:
pristine P(VDF-TrFE), P(VDF-TrFE)/CBTO and P(VDF-TrFE)/TBTO). The tribo-
positive surface was smooth PET. Open circuit voltage and short circuit current density
increased progressively from the pristine P(VDF-TrFE) case (225 V & 1.9 µA/cm2) to
P(VDF-TrFE)/CBTO (293 V & 5.5 µA/cm2) and reached a maximum for P(VDF-
TrFE)/TBTO (315 V & 6.7 µA/cm2), representing 1.4 and 3.4 times increase in voltage
and current density, respectively. The corresponding optimum power densities were
0.75, 2 and 2.75 W/m2, respectively. Accounting for contact pressure, max output (Voc
= 315 V & Jsc = 6.7 µA/cm2) is significantly higher than for TENGs having spin-coated
P(VDF-TrFE)/BTO.
The gain (versus spin-coating and poling) is likely due to enhanced dipole alignment in
132
the P(VDF-TrFE) matrix owing to the higher applied voltage during electrospinning.
Also, uniaxial fiber stretching during electrospinning is likely to produce highly aligned,
more crystalline P(VDF-TrFE). These effects are likely to confer higher surface polarity
(i.e., higher surface potential). The increase in electrical output with the addition of the
BTO particles is partly due to these high permittivity particles boosting the permittivity
of the P(VDF-TrFE) as has been noted in previous literature. However, we found a
significant output increase in moving from P(VDF-TrFE)/CBTO to P(VDF-
TrFE)/TBTO even though these two materials have near identical permittivity. XRD
and FTIR Spectroscopy analysis have revealed that TENG output is boosted by
enhancing the amount and crystallinity of the *-phase of the P(VDF-TrFE) matrix.
Higher crystallinity in the *-phase is likely to further increase dipole alignment and
further boost surface potential. The overlapped electron cloud (OEC) model explains
how this can then boost TENG performance: higher surface potential increases the
energy gap between Fermi levels and thereby increases the amount of tribo-charge
transfer. Thus, this work has shown that the addition of BTO particles boosts TENG
output, not just by enhancing permittivity, but by enhancing the amount and
crystallinity of the P(VDF-TrFE)/BTO matrix – hence the higher output with TBTO.
Finally, the chapter developed an efficient self-charging system by integrating the Fe-
TENGs with a graphene electrode based flexible supercapacitor. The system is capable
of charging to 1.25 V in the notably short time of 5 mins due to high electron mobility
in the graphene electrodes and high current output from the Fe-TENGs. This technology
can have significant application in wearable electronics where higher output currents
and more efficient charging are required from devices.

133
Chapter VI.

A wide range self-powered flexible pressure sensor based on the

triboelectric nanogenerator

6.1 Introduction
From Chapter II, it is clear that one of the major applications of triboelectric
nanogenerators (TENGs) is their application as self-powered sensors. Flexible sensors
and sensing systems are widely used in numerous applications of information and
communication technology (ICT) such as health monitoring [251], electronic skin (e-
skin) [252-254], robotics [253, 254], and sensor networks for Internet-of-Things (IoT)
[255] etc. to simultaneously record the wide variety of physical, chemical and
biological parameters [238, 256-258]. Among the various sensors, the flexible pressure
sensors have attracted huge attention since the contact force or pressure forms an
important controlling parameter in numerous applications including for the human-
machine interface, and controlled robotic movements [160, 255, 259]. As a result,
highly sensitive flexible pressure sensors based on capacitive [260, 261], piezoelectric
[262], piezoresistive [263] and triboelectric [160] etc. transduction methods have been
reported. Most of these sensors are reported for robotics applications and have been
developed for tasks requiring pressure measurement in the range (0.1-100kPa) typically
experienced by humans in daily tasks [264]. However, there are many other applications,
both in the robotics field and outside, where wide-range pressure sensors are needed.
For example, the pressure sensors in the sole of a humanoid robot or in the soles of a
person’s shoes with walking problems should withstand the body weight. Flexible
pressure sensors capable of working over a wide range (i.e. low (1-10kPa), medium
(10-100kPa) and high (>100kPa)) for static and dynamic pressures for applications,
ranging from intra-body pressures to deep sea [265, 266] are much needed. Further,
self-powering of sensors is desirable for continuous operation [267]. High precision,
low maintenance, self-powering and wide sensing range are some of the key
requirements for the next generation sensors [268, 269].
134
 Chapter III discussed the contact force dependent behaviour often observed with
TENGSs (where output increases with contact force). For instance, some results in the
literature simply show a linear variation of electrical output with contact force [139],
while others also include a later-stage saturation occurring at higher forces [142]. In
Chapter IV, the origin of this behaviour was explored via concurrent measurement of
both electrical output and real contact area (Ar) in a TENG (from 3.2kPa to 1.488MPa).
The Ar was notably small at low pressures (0.25% at 16 kPa), but increased 328-fold to
as much as 82% when it saturated beyond about 1.12 MPa pressure. This means that
the magnitude of electrical output signals from TENGs can be correlated with the
external pressure leading to TENG based pressure sensors with the additional feature
of being ‘self-powered’ [43, 270]. These results suggest that the contact pressure
dependence of the real contact area in TENGs can be used as a basis to develop a wide
range pressure sensor (up to ~1200 kPa).
In this regard, this chapter has demonstrated TENG based wide-range self-powered
sensors by exploiting this real-contact area dependence. A CS-TENG was fabricated
using Cu foil and PET film as the general class of non-engineered triboelectric materials
(Fig.4.1). The sensitivities at different pressure ranges were calculated from voltage (V)
versus nominal pressure plots. Further, we propose that such a device is ideal for self-
powered flexible pressure sensing for applications such as detecting the pressure in
wearables, e-Skin for soles of humanoid robots to help them walk and stand, and for
sensing the impact force for water wave impacts. In addition to this, we demonstrate
the potential application of the self-powered pressure sensing technology in two
example applications. The fabricated CS-TENG is firstly employed to monitor human
body movements including finger tapping and to collect users gait information, and,
secondly, for impact force sensing for free-falling objects.

6.2 Characterization of self-powered pressure sensor

Fig.6.1a schematically shows the working principle of a CS-TENG. The obtained Voc,
collected by an oscilloscope, is used as the sensing signal for the applied dynamic
pressures. Fig. 6.1b plots the output Voc signal from the CS-TENG, corresponding to a

135
nominal contact pressure of 1176 kPa ( 735N on an area of 6.25cm2) at a frequency of
0.5Hz. For a given contact force, the peak voltage signal in Fig. 6.1(b) is almost
constant indicating equilibrium of TENG operation. The peak positive voltages at
different contact pressures are then taken from the Voc signals and plotted against
applied contact pressure in Fig. 6.1(c-e). Note, the nominal contact pressure has been
separated as three segments, comprising low (1-10kPa), medium-high (10-500kPa) and
super high-pressure range (>500kPa). The obtained sensitivities for each of these three
pressure regimes are 3.16, 0.023 and 0.031 V∙kPa-1, respectively. To the best of the
authors knowledge, this is the first time a TENG as a self-powered pressure sensor has
been characterized over such a wide pressure range covering both low and super high
pressures. Several suitable low to super-high pressure application areas are proposed
incorporating TENG based self-powered sensors. Fig. 6.1 also indicates the typical
contact pressure available for various applications, including medical diagnosis,
road/traffic monitoring and the deep undersea environment. One potential application
could be for accurate pressure measurement in ocean and maritime research. The
developed pressure sensor can be placed on human-like robotic structures to collect
vital information, which could help researchers measure and record deep-sea pressures
(up to ~100 m below the sea surface which corresponds to 1.2MPa) [219]. It can also
sense a wide range under sea forces to give haptic feedback to control the robot’s
activity such as grasping objects.

136
 Releasing
(a)
(a) Pressed
Pressed Releasing (b)
100
80
+- -+-+- +- -+- +- -+-+- +- -+-+- +- -+-+- +- -+- +- -+-+ +++++++ ++ ++ ++ + + + + ++ ++ ++ ++++++ + + +
+++++++++++++++++
V VV
- ------- -- -- -- -- -- -- -- -- --------- -- -- -- -- -- -- - - - -
------------------------ 60
V

Voltage (V)
40
20
Pressing Released
Released
0
Pressing +++++++++++++++++++
+ + + + ++ ++ +++++++++++++++++++ ++ + + + V +++++++++++++++++
-------------------------
V -20
------- -- -- -- -- -- -- -- --------- -- -- -- -- -- -- -- -- - - -
V ------------------------ -40
V
0 5 10 15 20
Time (s)

Low pressure Medium-high pressure Super-high pressure

Intra-body pressures, intracranial Traffic counter, Blood or pulse Railway pressure, deep
pressure, intraocular pressure pressures undersea pressure

1kPa 10kPa 500kPa

(c) LowLowpressure
pressure (d) 72 Medium-high
Medium-highpressure
pressure
(e) 88 Super high pressure Ultra high pressure
40
86
Voltage (V)

30
Voltage (V)

Voltage (V)
68 84
82
20
80
64
10 Sensitivity=3.16V!kPa-1 Sensitivity=0.023V!kPa-1 78 Sensitivity=0.031V!kPa-1
76
60
0 74
0 4 8 12 0 100 200 300 400 500 800 900 1000 1100 1200
Nominal contact pressure (kPa) Nominal contact pressure (kPa) Nominal contact pressure (kPa)

Fig. 6.1: Characterization of TENG based self-powered pressure sensor and its potential applications: (a)
working principles of a CS-TENG, (b) a sample of TENG Voc signal under the contact pressure of
1176kPa (735N). (c-e) Voltage output of TENG based self-powered pressure sensor for: (c) low, (d)
medium-high pressure and (e) Super high-pressure ranges [27].

6.3 Applications
The fabricated TENG device shows a high response and wide range of pressure
detection capability, which can be applied to detect human body movements and impact
forces. Cu and PET tribo-contact layers (i.e., same as the device in Fig. 4.1) separated
by a gap of 5 mm using a spacer is used to demonstrate the multi-sensing capabilities
of the TENG device.

6.3.1 Monitoring of human body movements

The TENG device is demonstrated for human body movement monitoring. As different
body parts exert a wide range of pressures, the device here is ideal for recording
different physiological signal characteristics.
1) Finger tapping
Fig. 6.2 shows the output voltage signal of the TENG device after being pressed by
different numbers of fingers. When the device was pressed by one, two and three fingers,
the peak positive output voltage of the TENG was increased from 50 to 65, and to 90V,

137
respectively (Fig. 6.2). The enhancement in the output voltage is due to the increase in
the contact area between the active TENG layers when pressed by increasing numbers
of fingers (contact force was kept constant). The electrical output of the self-powered
sensor also exhibits a high stability which is clearly evident from the uniform positive
voltage output peak in Fig. 6.2. Therein, the positive voltage peaks correspond to the
status of pressing the device, and the subsequent negative voltage peaks are
corresponding to the releasing state of the device.

Fig. 6.2: Voltage signal of the pressure sensing TENG when pressed using different numbers of fingers.

2) Gait information collection

The fabricated TENG based self-powered sensor was integrated under the shoes of
different users to collect their gait information. Through the TENG output voltage
signal, the gait of the users can be analyzed. According to the working mechanism of
TENGs, the positive voltage peaks are when the tribo-layers come into contact due to
the applied force on the device and the negative voltage peaks can be detected when
the interface pair is releasing. Thus, some preliminary gait characterizations can be
obtained by monitoring the output voltage features, including amplitudes of output
voltage peaks, the frequency of cycles etc. (Fig. (6.3-6.4)). A variety of people’s
activities can be monitored by these detected gait features. For example, the frequencies
of the output voltage signals are 1Hz, 3Hz and 4Hz which corresponds to walking,
jogging, and running, respectively (Fig. 6.3(a-c)). Additionally, there is a 20V negative
voltage peaks and a 157V positive voltage peak in Fig. 6.3(d). The negative voltage

138
peak indicates the people jumping up, and the positive voltage peak corresponds to feet
impacting with the ground. Note, the gait information in Fig. 6.3 was collected from a
man weighing 80 kg (corresponding to a 1280kPa pressure). Fig. 6.4 indicates a very
similar response for the gait information extracted from a 45 kg female (corresponding
to a 720 kPa pressure). Comparing the output voltage amplitude in Fig. 6.3 and 6.4, the
output under four types of gait conditions from the man is higher than from the woman.
This is, of course, because the higher the weight of the person, the greater the contact
pressure on the TENG device. The output differences in Fig. 6.3 and 6.4, clearly
indicate that this self-powered pressure sensor can detect a wide range of pressures.

(a) (b)

(c) (d)

Fig. 6.3: Gait information from a man of 80kg weight collected by the pressure sensing TENG (a)
walking, (b) jogging, (c) running and (d) jumping.

139
 (a) (b)

(c) (d)

Fig. 6.4: Gait information from a woman of 45kg weight collected by the pressure sensing TENG (a)
walking, (b) jogging, (c) running and (d) jumping.

6.3.2 TENG as self-powered impact force sensor for falling objects

The fabricated TENG device also has potential to be used as a self-powered impact
force sensor. This is demonstrated by free-falling an object of a fixed mass over the
device from different heights, as illustrated in Fig. 6.5(a). The response of the TENG
based self-powered impact force sensor is essentially given by the output voltage of the
TENG. TENG output voltage exhibits a number of peaks because of multiple
rebounding of the object. The high amplitude first peak in Fig. 6.5(b) indicates the first
direct impact of the object when dropped from a chosen height and the subsequent
voltage peaks are corresponding to the rebounding (i.e., object jumps on the device)
motion of the object. The rebounded height of the object is lower than the pre-
established height (due to the energy dissipation). Thus, the values of the voltage peaks
are lower for subsequent peaks. Fig. 6.5(b) plots the TENG output voltage signals used
for observing the impact force sensor characterization. When the object is dropped from
a 5cm height, the value of the first positive output voltage peak is 117V. The peak value
then increases to 183V, when the preestablished height is adjusted to 14cm. Thus, the
output voltage of the TENG is (as expected) related to the dropping height of the object.

140
The positive voltages for the first peak are then taken from Fig. 6.5(b) and plotted
against dropping height in Fig. 6.5(c). As the dropping height increases (5cm to 14cm),
a highly linear increase of voltage with dropping height is observed. Note, the
calculated height sensitivity (S) is approximated to 7V/cm. According to the law of
conservation of energy, the impact velocity and impact force can be calculated by using
the dropping height. Therefore, this sensor not only can detect the falling height of the
object, but its impact velocity and impact force as depicted in Fig. 6.5(d&e),
respectively.
Due to the free-falling motion of the object, the potential energy from the gravity of the
object is converted to kinetic energy just before the impact as:
1 1
Work done by gravity = 2
Y:bcdefg 2 − 2 Y:bhbgbei 2 (6.1)
Work done of gravity = Y=ℎ (6.2)

Where, m is the mass of the object, :bhbgbei is the initial velocity of the object, :bcdefg

is object impact velocity, and h is the preestablished height. Due to the free-falling
motion, the initial velocity of the object is zero. Thus, the impact velocity can be
calculated by:

:bcdefg = á2=ℎ (6.3)

According to Eq. (6.3), the impact velocity of the object from a height of 5cm, 10.5cm,
13cm and 14cm was calculated. The increase of impact velocity is quantified in Fig.
6.5(d) where impact velocity is plotted against height. The obtained sensitivity between
the impact velocity and height is 0.075m/s*cm. Simultaneously, the sensitivity (S’)
between output voltage (Vo) and impact velocity can be calculated, depending on the
regression equation of Vo vs. h and v_impact vs. h. S’ is the proportionality constant
having the value of 92.8m(s∙V)-1.
Additionally, the impact force is related to the impact energy and deformation of the
impacted object. The impact energy is equal to the work done by gravity from the
dropping point to the TENG top surface or the kinetic energy just before the impact.
Thus, according to the energy conservation, the impact force can be calculated by the
141
work-energy theorem [38], as
"
l&%m ,&aS &L jkYA3( L&%3S k7-ZS+/) $ k?Q
àjkYA3( = =$ = (6.4)
, , ,

Where, d is the distance of impacted object deformation. Hence, in this experiment, d

should approximate to the separation distance of the TENG. The calculated impact
force was calculated by the work done by gravity. The calculated impact forces are
plotted against the dropping height in Fig. 6.5(e). For this height range (5cm to 14cm),
we observe a linear increase of impact force with dropping height. The extracted
sensitivity of 0.4N/cm is 4 times higher than the previously reported TENG based
impact force sensors [271]. The sensitivity (S’’) between output voltage (Vo) and impact
force can be recalculated by the regression equation of Vo versus h and F_impact versus
h. S’’ is the proportionality constant having a value of 17.4V/N.
Furthermore, we demonstrate wireless transfer of the output voltage signal from the
TENG. Towards this aim, the AC signal was converted to DC using a microcontroller
circuit (FireBeetle ESP32-E) and sent via the Bluetooth Classic protocol (2.4 GHz) to
a mobile device. As a proof-of-concept, the finger tapping signals are transferred to a
tablet.

(a) 20g
20g Weight
Weight
20g
Weight

20g 13cm 14cm

Weight
10.5cm
5cm

V V V V

(b) (i) (ii) (iii) (iv)

Height = 5cm Height = 10.5cm Height = 13cm Height = 14cm

(c) (d) (e)

Sensitivity=0.075m/s*cm
!! =0.99361

Sensitivity=6.96V/cm Sensitivity=0.4N/cm
!! =0.99876 !! =1

Fig. 6.5: (a) Schematic of free-falling object onto TENG impact sensor from different heights of 5cm,
10.5cm, 13cm, 14cm; (b) output voltage peaks of TENG impact sensor showing the consecutive
rebounding of the object from different heights shown in (a); (c) represents the results of output voltage
against heights; (d) depicts impact velocity versus fall height and (e) shows the results for impact force
against heights.

142
6.4 Conclusion
The results demonstrate that a CS-TENG based self-powered pressure sensor can detect
a wide range of pressures with high sensitivity up to considerably higher pressures than
previously considered. To achieve this, the sensor relies on the contact pressure
dependence of the real contact area Ar in TENGs. This leads directly to a contact
pressure dependent electrical output that can be utilized for pressure sensing. In order
to explain the effect, pressure sensitive film was used to quantify the Ar/An ratio over a
wide range of pressures. The detected pressure sensing range was separated into three
regimes comprising: low, medium-high and super high pressure covering a range from
3.2 to 1176 kPa. The associated sensor sensitivities for the three pressure regimes were
determined at 3.16V∙kPa-1, 0.023 V∙kPa-1 and 0.031 V∙kPa-1, respectively.
Simultaneously, it is proposed that the TENG based sensor has high potential to be used
for ocean and robotics research due to its ability to detect high pressures. Finally, we
have demonstrated how the fabricated pressure sensing CS-TENG can be applied to a
variety of applications. To provide relevant examples, the chapter shows how the sensor
can be used to detect finger tapping, collect user gait information and detect impact
forces.

143
Chapter VII.

Conclusion and future perspectives

7.1 Conclusions
The overall aim of this PhD research was to boost the performance of contact and
separation mode triboelectric nanogenerators (CS-TENGs) via parameters that can
have a significant effect on output. An important goal was also to study how the physics
of these parameters effects output. The application of these high performance TENGs
in self-powered sensors/systems and self-charging systems was also explored.
TENGs as newly emerging renewable energy harvesters are ideal as the energy supply
in autonomous systems and have found application in a wide spectrum of areas such as:
internet-of-things (IoT), powering wearable electronics, rehabilitation, respiration
monitoring, powering or self-powered actuation and motion sensing, air purification
and gas sensing, and implantable biomedical devices. Output power of TENGs is still
comparatively low and any increase can have significant implications. This thesis has
put forward three key approaches to boosting output: boosting output via control of
substrate permittivity, via control of contact area and via use of permanently polarised
optimised composite ferroelectric materials.
The theories of CS-TENGS (e.g. device structure and tribo-charge transfer) were
studied by simulation and numerical study in Chapter III. These simulations showed
that, among the factors that confer higher performance are: materials with a strong
triboelectric effect together with low permittivity. However, it is challenging to
optimize both within a single material. Therefore, Chapter III presented a solution to
this challenge by optimizing a low permittivity substrate beneath the tribo-contact layer.
Results were simulated over a range of substrate permittivity. The open circuit voltage
was found to increase by a factor of 1.6 in moving from PVDF to the lower permittivity
PTFE. Two TENG devices have been fabricated with 100 μm PET and PTFE substrates
to compare performance. The experiments confirm that lowering the substrate dielectric
constant (i.e. PET to PTFE) raises the open circuit voltage in line with simulation
predictions.
144
In Chapter IV, attention turned to contact area. The review in Chapter II indicated that
TENG output tends to be contact force dependent, but its origin had yet to be
established. Herein, this thesis shows that the origin lies in the real contact area Ar,
probed with novel experiments specifically designed for this purpose. The open circuit
voltage Voc, short circuit current Isc and Ar for a TENG, having two nominally flat tribo-
contact surfaces, were found to increase with contact force/pressure. The Ar was notably
small at low pressures (0.25% at 16 kPa) that are typically experienced in wearable
applications. However, it increased 328 fold to as much as 82% when it saturated
beyond about 1.12 MPa pressure - achievable for something like impact with ocean
waves. Critically, Voc and Isc saturated at the same contact pressure as Ar suggesting that
electrical output follows the evolution of the Ar. Assuming that tribo-charges can only
transfer across the interface at areas of real contact, it follows that an increasing Ar with
contact pressure should produce a corresponding increase in the electrical output. These
results underline the importance of accounting for real contact area in TENG design to
boost their performance, the distinction between real and nominal contact area in tribo-
charge density definition, and the possibility of using TENGs as a self-powered
pressure/load sensors. Crucially, the results indicate that the large contact pressures,
readily available in applications such as road-tyre contact and wave energy, alone could
be enough to boost the performance, thus avoiding the need for costly surface
engineering to increase Ar.
Chapter V focused on boosting output via careful optimisation of composite
ferroelectric materials in fiborous matt form. This entailed the development of high-
performance ferroelectric-assisted TENGs using electrospun fibrous surfaces based on
P(VDF-TrFE) with dispersed BaTiO3 (BTO) nanofillers in either cubic (CBTO) or
tetragonal (TBTO) form. TENGs with three types of tribo-negative surface (pristine
P(VDF-TrFE), P(VDF-TrFE)/CBTO and P(VDF-TrFE)/TBTO) in contact with PET
were investigated and output increased progressively from pristine (0.75 W/m2) to
CBTO (2 W/m2) and to TBTO (2.75 W/m2). Accounting for contact pressure, the max
output (Voc = 315 V & Jsc = 6.7 μA/cm2) was significantly higher than for TENGs
having spin-coated P(VDF-TrFE)/BTO. It is hypothesized that electrospinning
145
increases dipole alignment due to high applied voltages, but also aids the formation of
a highly oriented crystalline β-phase via uniaxial stretching. Essentially, tribo-charge
transfer is boosted due to increased surface potential owing to enhanced ferroelectric
polarization. P(VDF-TrFE)/TBTO produced higher output than P(VDF-TrFE)/CBTO
even though permittivity is nearly identical. Thus, it is shown that BTO fillers boost
output, not just by increasing permittivity, but also by enhancing the crystallinity and
amount of the β-phase (as TBTO produced a more crystalline β-phase present in greater
amounts). Finally, the ferroelectric-assisted TENG was integrated with a flexible
graphene electrode-based supercapacitor to produce a self-charging system capable of
charging to 1.25 V in just 5 min. These results demonstrate that this technology can be
valuable in wearable applications where higher power output, more efficient charging
and flexibility are paramount.
In Chapter VI, a CS-TENG-based self-powered pressure sensor has been demonstrated
to detect a wide range of pressures with high sensitivity up to considerably higher
pressures than previously considered. To achieve this, the sensor relies on the contact
pressure dependence of the real contact area Ar in TENGs. In order to explain the effect,
pressure sensitive film was used to quantify the Ar/An ratio over a wide range of
pressures. The detected pressure sensing range was separated into three regimes
comprising: low, medium-high and super high pressure covering a range from 3.2 to
1176 kPa. The associated sensor sensitivities for the three pressure regimes were
determined at 3.16V∙kPa-1, 0.023 V∙kPa-1 and 0.031 V∙kPa-1, respectively.
Simultaneously, we propose that the TENG based sensor has high potential to be used
for ocean and robotics research due to its ability to detect high pressures. Finally, we
have demonstrated how the fabricated pressure sensing CS-TENG can be applied to a
variety of applications. To provide relevant examples, the chapter shows how the sensor
can be used to detect finger tapping, collect user gait information and detect impact
forces.
In summary, this work has demonstrated how CS-TENG output can be boosted via three
distinct approaches: by introduction of a low permittivity substrate, by increase of the
real contact area and via use of highly optimised (and polarised) ferroelectric
146
composites (in this case, P(VDF-TrFE) with barium titanate (BaTiO3) fillers). The
physical understanding developed in relation to these will be valuable in helping to
optimise performance in future TENG design work. One avenue for future work would
be to attempt to combine each of these optimisations into a single TENG design. Further
possibilities for future work are discussed next.

7.2 Future perspectives

The emphasis of this multidisciplinary study was to provide methods to optimize
contact and separation mode triboelectric nanogenerator (CS-TENG) performance.
However, there are several key areas where additional studies are required to address,
improve and further develop the conclusions in this thesis. An effective method to
measure the real contact area (Ar) of CS-TENG by pressure sensitive film has been
demonstrated in Chapter VI. However, it is believed that the electrical outputs of CS-
TENGs can be significantly improved by using micro/nano-structured surfaces for both
mating surfaces. Thus, a competitive CS-TENG fabrication method to produce
micro/nano-structured surface has already been applied in Chapter V. However, there
are many gaps between the nanofibers on the fibrous mat surface. If there is an efficient
material surface design to match with these gaps, then the Ar between the interface pair
could be greatly increased. Therefore, use of a ZnO nanowire surface to contact with
the fibrous mat surface is planned.
The electrospun fibrous mat is more suitable for designing a wearable TENG, because
it not only possesses high triboelectric properties, but also good breathability. However,
if a polymer film substrate or metal layer electrode is adopted, it will decrease the
breathability. Therefore, if the fibrous mat can combine with a fabric film, their
wearability will be improved significantly.
One of the applications in TENG based self-charging system has been highlighted in
Chapter V. However, due to the low output current of TENG devices, it will take a long
time to charge high performance energy storage devices (e.g. lithium battery). A
potential method to negotiate this limit is to use a step-down transformer to convert it
as high output current and low output voltage. Therefore, a wearable step-down

147
transformer circuit is planned for future work. Simultaneously, since TENGs depend on
mechanical energy input, this places a limitation on its applications. In light of this,
there is potential to design hybrid nanogenerators where non-mechanical energy input
can also be relied on (e.g. integrating solar cell and TENG or PyENG and TENG).

148
Appendix A: Surface Roughness Metrology
The definitions, as well as the computing methods relating to the surface metrology
parameters used in Table. 4.1 are given here. All of these parameters are easily
calculated from surface x, y, z data obtained from surface profilometers at sufficient
resolution.
The roughness height, ℎ, is leveled so that mean(ℎ) = 0.
Center line average (CLA) ;A :
a; a[
1
;A = á|ℎ| = â ä ä |ℎjo |
A. An
jp1 op1

Root mean square (RMS) roughness ;[ :

a; a[
1 2
;[ = á〈|ℎ|2 〉 = â ä ä ℎjo
A. An
jp1 op1

RMS surface gradient ;,[ :

A continuous form of RMS surface gradient is

êℎ êℎ 2
;,[ = á〈|∇ℎ|2 〉 = è〈] + _ 〉
ê1 êa

a; :1 a[ :1 2
1 êℎ êℎ
≈â ä ä ë] _ + ] _ í
(A. − 2)(An − 2) ê1 jo êa jo
jp2 op2

where the partial derivatives are calculated using the central differential scheme:
êℎ ℎj61,o − ℎj:1,o êℎ ℎj,o61 − ℎj,o:1
] _ = ,] _ = ,
ê1 jo 2Δ. êa jo 2Δn

where Δ. and Δn are the sampling resolution in the 1 and a directions,

respectively.

149
Appendix B: MATLAB code
function dU_dt = myodefunction(t,U)
x1=0.0002;%thickness of surface1
x2=0.00022;%thickness of surface2
w=0.05;%wide of device
e1=28.674*10.^-12; %permmitity of surface1
e2=29.205*10.^-12;%permmitity of surface2
z=linspace(22,0.01,170);%distance
T=40.7;%density of tribo-charge
H=1*(10.^-3);
f=1; %operational frequency
z=(H.*(sin((((2*pi)*f).*t)+((3*pi)/2))))+H;
R=10.^6;
M1=(1/e1).*((((x1+z).*(atan((w)./((2.*(x1+z)).*(sqrt((4.*((x1+z).^2))+2))))))-
((w/2).*(log((sqrt((4.*((x1+z).^2))+2)+1)./(sqrt((4.*((x1+z).^2))+2)-1)))))-
(((x1).*(atan((w)./((2.*(x1)).*(sqrt((4.*((x1).^2))+2))))))-
((w/2).*(log((sqrt((4.*((x1).^2))+2)+1)./(sqrt((4.*((x1).^2))+2)-1))))));
M2=(1/e2).*((((x2+z).*(atan((w)./((2.*(x2+z)).*(sqrt((4.*((x2+z).^2))+2))))))-
((w/2).*(log((sqrt((4.*((x2+z).^2))+2)+1)./(sqrt((4.*((x2+z).^2))+2)-1)))))-
(((x2).*(atanh((w)./((2.*(x2)).*(sqrt((4.*((x2).^2))+2))))))-
((w/2).*(log((sqrt((4.*((x2).^2))+2)+1)./(sqrt((4.*((x2).^2))+2)-1))))));
M3=((1/e1)+(1/e2)).*((((x1+x2+z).*(atan((w)./((2.*(x1+x2+z)).*(sqrt((4.*((x1+x2+z).^2))+2))))))-
((w/2).*(log((sqrt((4.*((x1+x2+z).^2))+2)+1)./(sqrt((4.*((x1+x2+z).^2))+2)-1)))))-
(((0).*(atanh((w)./((2.*(0)).*(sqrt((4.*((0).^2))+2))))))-
((w/2).*(log((sqrt((4.*((0).^2))+2)+1)./(sqrt((4.*((0).^2))+2)-1))))));
M4=((T./pi).*(M1+M2));
P=(M3./pi);
dU_dt=[(1/((w.^2)*(R))).*((M4)-(U.*P))];
end

150
References:
[1] T. Ahmad and D. Zhang, "A critical review of comparative global historical energy consumption
and future demand: The story told so far," Energy Reports, vol. 6, pp. 1973-1991, 2020.
[2] J. Broderick, K. Anderson, C. Jones, and J. Watson, "A review of research relevant to new build
nuclear power plants in the UK," 2013.
[3] G. Crabtree, "Energy: the Next Fifty Years," 2014.
[4] S. L. Montgomery. "Biden taps the Strategic Petroleum Reserve – What is it? Where did it come
from? And does the US still need it?" The conversation. [Link]
the-strategic-petroleum-reserve-what-is-it-where-did-it-come-from-and-does-the-us-still-need-
it-172473 (accessed.
[5] "Proved Reserves of Crude Oil and Natural Gas in the United States, Year-End 2019." eia U.S
Energy Information Administration. [Link]
(accessed.
[6] J. Jurasz, F. Canales, A. Kies, M. Guezgouz, and A. Beluco, "A review on the complementarity
of renewable energy sources: Concept, metrics, application and future research directions,"
Solar Energy, vol. 195, pp. 703-724, 2020.
[7] J. Iannacci, "Microsystem based Energy Harvesting (EH-MEMS): Powering pervasivity of the
Internet of Things (IoT)–A review with focus on mechanical vibrations," Journal of King Saud
University-Science, vol. 31, no. 1, pp. 66-74, 2019.
[8] L. Mateu and F. Moll, "Review of energy harvesting techniques and applications for
microelectronics," in VLSI Circuits and Systems II, 2005, vol. 5837: International Society for
Optics and Photonics, pp. 359-373.
[9] A. S. Algamili et al., "A review of actuation and sensing mechanisms in MEMS-based sensor
devices," Nanoscale research letters, vol. 16, no. 1, pp. 1-21, 2021.
[10] L. A. Villarruel Mendoza, N. A. Scilletta, M. G. Bellino, M. F. Desimone, and P. N. Catalano,
"Recent advances in micro-electro-mechanical devices for controlled drug release applications,"
Frontiers in Bioengineering and Biotechnology, vol. 8, p. 827, 2020.
[11] A. Dehghani-Sanij, E. Tharumalingam, M. Dusseault, and R. Fraser, "Study of energy storage
systems and environmental challenges of batteries," Renewable and Sustainable Energy
Reviews, vol. 104, pp. 192-208, 2019.
[12] Z. L. Wang, L. Lin, J. Chen, S. Niu, and Y. Zi, Triboelectric nanogenerators. Springer, 2016.
[13] G. Zhu, B. Peng, J. Chen, Q. Jing, and Z. L. Wang, "Triboelectric nanogenerators as a new
energy technology: from fundamentals, devices, to applications," Nano Energy, vol. 14, pp. 126-
138, 2015.
[14] B. Bera, "Literature review on triboelectric nanogenerator," Imperial Journal of
Interdisciplinary Research (IJIR), vol. 2, no. 10, pp. 1263-1271, 2016.
[15] S. Pan and Z. Zhang, "Fundamental theories and basic principles of triboelectric effect: A
review," Friction, vol. 7, no. 1, pp. 2-17, 2019.
[16] S. P. Sukhatme and J. Nayak, Solar energy. McGraw-Hill Education, 2017.
[17] H. J. Goldsmid, Introduction to thermoelectricity. Springer, 2010.
[18] K. Liu et al., "Thermal–electric nanogenerator based on the electrokinetic effect in porous
carbon film," Advanced Energy Materials, vol. 8, no. 13, p. 1702481, 2018.
[19] B. Bera and M. D. Sarkar, "Piezoelectricity in PVDF and PVDF based piezoelectric
nanogenerator: A concept," IOSR J. Appl. Phys, vol. 9, no. 3, pp. 95-99, 2017.

151
[20] S. Wang, S. Niu, J. Yang, L. Lin, and Z. L. Wang, "Quantitative measurements of vibration
amplitude using a contact-mode freestanding triboelectric nanogenerator," ACS nano, vol. 8, no.
12, pp. 12004-12013, 2014.
[21] P. Bai et al., "Integrated multilayered triboelectric nanogenerator for harvesting biomechanical
energy from human motions," ACS nano, vol. 7, no. 4, pp. 3713-3719, 2013.
[22] C. Rodrigues et al., "Integrated study of triboelectric nanogenerator for ocean wave energy
harvesting: Performance assessment in realistic sea conditions," Nano Energy, vol. 84, p.
105890, 2021.
[23] Y. Xie et al., "Rotary triboelectric nanogenerator based on a hybridized mechanism for
harvesting wind energy," ACS nano, vol. 7, no. 8, pp. 7119-7125, 2013.
[24] Z. L. Wang, "Triboelectric nanogenerators as new energy technology for self-powered systems
and as active mechanical and chemical sensors," ACS nano, vol. 7, no. 11, pp. 9533-9557, 2013.
[25] J. Luo and Z. L. Wang, "Recent advances in triboelectric nanogenerator based self-charging
power systems," Energy Storage Materials, vol. 23, pp. 617-628, 2019.
[26] P. Gipe, "Wind power," Wind Engineering, vol. 28, no. 5, pp. 629-631, 2004.
[27] A. de O Falcão, "Control of an oscillating-water-column wave power plant for maximum energy
production," Applied Ocean Research, vol. 24, no. 2, pp. 73-82, 2002.
[28] F. Verástegui, Á. Lorca, M. Negrete-Pincetic, and D. Olivares, "Firewood heat electrification
impacts in the Chilean power system," Energy Policy, vol. 144, p. 111702, 2020.
[29] R. Fuller and L. Aye, "Human and animal power–The forgotten renewables," Renewable energy,
vol. 48, pp. 326-332, 2012.
[30] H. Ritchie and M. Roser, "Fossil fuels," Our world in data, 2017.
[31] M. H. Dickson and M. Fanelli, Geothermal energy. Wiley, 2004.
[32] P. P. Edwards, V. Kuznetsov, and W. I. David, "Hydrogen energy," Philosophical Transactions
of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 365, no. 1853,
pp. 1043-1056, 2007.
[33] S. V. Vassilev, C. G. Vassileva, and V. S. Vassilev, "Advantages and disadvantages of
composition and properties of biomass in comparison with coal: An overview," Fuel, vol. 158,
pp. 330-350, 2015.
[34] D. Bodansky, Nuclear energy: principles, practices, and prospects. Springer Science &
Business Media, 2007.
[35] R. J. Pearson, A. B. Antoniazzi, and W. J. Nuttall, "Tritium supply and use: a key issue for the
development of nuclear fusion energy," Fusion Engineering and Design, vol. 136, pp. 1140-
1148, 2018.
[36] A. Gorlov, "Tidal energy," Elements of Physical Oceanography, pp. 103-108, 2001.
[37] K. A. Kvenvolden and T. D. Lorenson, "The global occurrence of natural gas hydrate," Natural
Gas Hydrates: Occurrence, Distribution, and Detection: Occurrence, Distribution, and
Detection, vol. 124, pp. 3-18, 2001.
[38] L. Lakatos, G. Hevessy, and J. Kovács, "Advantages and disadvantages of solar energy and
wind-power utilization," World Futures, vol. 67, no. 6, pp. 395-408, 2011.
[39] I. Yüksel, "Hydropower for sustainable water and energy development," Renewable and
Sustainable Energy Reviews, vol. 14, no. 1, pp. 462-469, 2010.
[40] N. Sriram and M. Shahidehpour, "Renewable biomass energy," in IEEE Power Engineering
Society General Meeting, 2005, 2005: IEEE, pp. 612-617.

152
[41] P. Takahashi and A. Trenka, "Ocean thermal energy conversion," 1996.
[42] "Questions and Answers - Making our energy system fit for our climate targets." (accessed.
[43] G. Min, L. Manjakkal, D. M. Mulvihill, and R. S. Dahiya, "Triboelectric nanogenerator with
enhanced performance via an optimized low permittivity substrate," IEEE Sensors Journal, vol.
20, no. 13, pp. 6856-6862, 2019.
[44] M. Sarno, "Nanotechnology in energy storage: The supercapacitors," in Studies in Surface
Science and Catalysis, vol. 179: Elsevier, 2020, pp. 431-458.
[45] H. Xue et al., "A wearable pyroelectric nanogenerator and self-powered breathing sensor," Nano
Energy, vol. 38, pp. 147-154, 2017.
[46] H. Ryu and S. W. Kim, "Emerging pyroelectric nanogenerators to convert thermal energy into
electrical energy," Small, vol. 17, no. 9, p. 1903469, 2021.
[47] J. Briscoe and S. Dunn, "Piezoelectric nanogenerators–a review of nanostructured piezoelectric
energy harvesters," Nano Energy, vol. 14, pp. 15-29, 2015.
[48] Z. Zhang et al., "Deep learning-enabled triboelectric smart socks for IoT-based gait analysis and
VR applications," npj Flexible Electronics, vol. 4, no. 1, pp. 1-12, 2020.
[49] Z. Zhou et al., "Wireless self-powered sensor networks driven by triboelectric nanogenerator
for in-situ real time survey of environmental monitoring," Nano Energy, vol. 53, pp. 501-507,
2018.
[50] G. Min et al., "Ferroelectric-assisted high-performance triboelectric nanogenerators based on
electrospun P (VDF-TrFE) composite nanofibers with barium titanate nanofillers," Nano Energy,
vol. 90, p. 106600, 2021.
[51] Y. Kumaresan, G. Min, A. S. Dahiya, A. Ejaz, D. Shakthivel, and R. Dahiya, "Kirigami and
Mogul‐Patterned Ultra‐Stretchable High‐Performance ZnO Nanowires‐Based Photodetector,"
Advanced Materials Technologies, p. 2100804, 2021.
[52] X. Pu et al., "A self‐charging power unit by integration of a textile triboelectric nanogenerator
and a flexible lithium‐ion battery for wearable electronics," Advanced Materials, vol. 27, no.
15, pp. 2472-2478, 2015.
[53] C. Zhang, W. Fan, S. Wang, Q. Wang, Y. Zhang, and K. Dong, "Recent Progress of Wearable
Piezoelectric Nanogenerators," ACS Applied Electronic Materials, 2021.
[54] X.-S. Zhang, M. Han, B. Kim, J.-F. Bao, J. Brugger, and H. Zhang, "All-in-one self-powered
flexible microsystems based on triboelectric nanogenerators," Nano Energy, vol. 47, pp. 410-
426, 2018.
[55] L. Xu et al., "Coupled triboelectric nanogenerator networks for efficient water wave energy
harvesting," ACS nano, vol. 12, no. 2, pp. 1849-1858, 2018.
[56] S. Naik, R. Mukherjee, and B. Chaudhuri, "Triboelectrification: A review of experimental and
mechanistic modeling approaches with a special focus on pharmaceutical powders,"
International journal of pharmaceutics, vol. 510, no. 1, pp. 375-385, 2016.
[57] S. Niu et al., "Theoretical study of contact-mode triboelectric nanogenerators as an effective
power source," Energy & Environmental Science, vol. 6, no. 12, pp. 3576-3583, 2013.
[58] Y. S. Zhou et al., "Manipulating nanoscale contact electrification by an applied electric field,"
Nano letters, vol. 14, no. 3, pp. 1567-1572, 2014.
[59] R. D. I. G. Dharmasena et al., "Triboelectric nanogenerators: providing a fundamental
framework," Energy & Environmental Science, vol. 10, no. 8, pp. 1801-1811, 2017.
[60] S. Matsusaka, H. Maruyama, T. Matsuyama, and M. Ghadiri, "Triboelectric charging of

153
 powders: A review," Chemical Engineering Science, vol. 65, no. 22, pp. 5781-5807, 2010.
[61] J. Wu, X. Wang, H. Li, F. Wang, W. Yang, and Y. Hu, "Insights into the mechanism of metal-
polymer contact electrification for triboelectric nanogenerator via first-principles
investigations," Nano Energy, vol. 48, pp. 607-616, 2018.
[62] J. Wu, X. Wang, H. Li, F. Wang, and Y. Hu, "First-principles investigations on the contact
electrification mechanism between metal and amorphous polymers for triboelectric
nanogenerators," Nano Energy, vol. 63, p. 103864, 2019.
[63] Z. L. Wang and A. C. Wang, "On the origin of contact-electrification," Materials Today, vol. 30,
pp. 34-51, 2019.
[64] F. A. Furfari, "A history of the Van de Graaff generator," IEEE Industry Applications Magazine,
vol. 11, no. 1, pp. 10-14, 2005.
[65] R. Herb, "Van de Graaff generators," in Nuclear Instrumentation I/Instrumentelle Hilfsmittel
der Kernphysik I: Springer, 1959, pp. 64-104.
[66] F.-R. Fan, Z.-Q. Tian, and Z. L. Wang, "Flexible triboelectric generator," Nano energy, vol. 1,
no. 2, pp. 328-334, 2012.
[67] R. Dharmasena, K. Jayawardena, C. Mills, R. Dorey, and S. Silva, "A unified theoretical model
for Triboelectric Nanogenerators," Nano Energy, vol. 48, pp. 391-400, 2018.
[68] H. Zhang, L. Yao, L. Quan, and X. Zheng, "Theories for triboelectric nanogenerators: A
comprehensive review," Nanotechnology Reviews, vol. 9, no. 1, pp. 610-625, 2020.
[69] Y. Li et al., "Contribution of Ferromagnetic Medium to the Output of Triboelectric
Nanogenerators Derived from Maxwell's Equations," Advanced Energy Materials, p. 2003921,
2021.
[70] Z. L. Wang, "From contact-electrification to triboelectric nanogenerators," Reports on Progress
in Physics, 2021.
[71] S. Niu et al., "Theory of sliding‐mode triboelectric nanogenerators," Advanced materials, vol.
25, no. 43, pp. 6184-6193, 2013.
[72] J. Shao et al., "Studying about applied force and the output performance of sliding-mode
triboelectric nanogenerators," Nano Energy, vol. 48, pp. 292-300, 2018.
[73] Y. Mao, D. Geng, E. Liang, and X. Wang, "Single-electrode triboelectric nanogenerator for
scavenging friction energy from rolling tires," Nano Energy, vol. 15, pp. 227-234, 2015.
[74] Y. Li, G. Cheng, Z.-H. Lin, J. Yang, L. Lin, and Z. L. Wang, "Single-electrode-based rotationary
triboelectric nanogenerator and its applications as self-powered contact area and eccentric angle
sensors," Nano Energy, vol. 11, pp. 323-332, 2015.
[75] Q. Liang et al., "Highly transparent triboelectric nanogenerator for harvesting water-related
energy reinforced by antireflection coating," Scientific reports, vol. 5, no. 1, pp. 1-7, 2015.
[76] S. A. Shankaregowda et al., "Single-electrode triboelectric nanogenerator based on economical
graphite coated paper for harvesting waste environmental energy," Nano Energy, vol. 66, p.
104141, 2019.
[77] X. Pu et al., "Ultrastretchable, transparent triboelectric nanogenerator as electronic skin for
biomechanical energy harvesting and tactile sensing," Science advances, vol. 3, no. 5, p.
e1700015, 2017.
[78] N. Zhang, C. Tao, X. Fan, and J. Chen, "Progress in triboelectric nanogenerators as self-powered
smart sensors," Journal of Materials Research, vol. 32, no. 9, pp. 1628-1646, 2017.
[79] H. Zhang et al., "A theoretical approach for optimizing sliding-mode triboelectric nanogenerator

154
 based on multi-parameter analysis," Nano Energy, vol. 61, pp. 442-453, 2019.
[80] S. Niu et al., "Theoretical investigation and structural optimization of single‐electrode
triboelectric nanogenerators," Advanced Functional Materials, vol. 24, no. 22, pp. 3332-3340,
2014.
[81] S. Niu et al., "Theory of freestanding triboelectric-layer-based nanogenerators," Nano Energy,
vol. 12, pp. 760-774, 2015.
[82] F. Peng et al., "Facile fabrication of triboelectric nanogenerator based on low-cost thermoplastic
polymeric fabrics for large-area energy harvesting and self-powered sensing," Nano Energy, vol.
65, p. 104068, 2019.
[83] L. Jin et al., "Manipulating relative permittivity for high-performance wearable triboelectric
nanogenerators," Nano Letters, vol. 20, no. 9, pp. 6404-6411, 2020.
[84] Y. Yu, Z. Li, Y. Wang, S. Gong, and X. Wang, "Sequential infiltration synthesis of doped
polymer films with tunable electrical properties for efficient triboelectric nanogenerator
development," Advanced Materials, vol. 27, no. 33, pp. 4938-4944, 2015.
[85] H. Zou et al., "Quantifying the triboelectric series," Nature communications, vol. 10, no. 1, pp.
1-9, 2019.
[86] A. Chen, C. Zhang, G. Zhu, and Z. L. Wang, "Polymer materials for high‐performance
triboelectric Nanogenerators," Advanced Science, vol. 7, no. 14, p. 2000186, 2020.
[87] V. F. Cardoso, D. M. Correia, C. Ribeiro, M. M. Fernandes, and S. Lanceros-Méndez,
"Fluorinated polymers as smart materials for advanced biomedical applications," Polymers, vol.
10, no. 2, p. 161, 2018.
[88] B. Dudem, D. H. Kim, A. R. Mule, and J. S. Yu, "Enhanced performance of microarchitectured
PTFE-based triboelectric nanogenerator via simple thermal imprinting lithography for self-
powered electronics," ACS applied materials & interfaces, vol. 10, no. 28, pp. 24181-24192,
2018.
[89] Y. Yang et al., "Single-electrode-based sliding triboelectric nanogenerator for self-powered
displacement vector sensor system," Acs Nano, vol. 7, no. 8, pp. 7342-7351, 2013.
[90] T. Prada et al., "Enhancement of output power density in a modified polytetrafluoroethylene
surface using a sequential O2/Ar plasma etching for triboelectric nanogenerator applications,"
Nano Research, pp. 1-8, 2021.
[91] Y. Yang, Y. S. Zhou, H. Zhang, Y. Liu, S. Lee, and Z. L. Wang, "A single‐electrode based
triboelectric nanogenerator as self‐powered tracking system," Advanced Materials, vol. 25, no.
45, pp. 6594-6601, 2013.
[92] Z. Zhang, Y. Xu, D. Wang, H. Yang, J. Guo, and L.-S. Turng, "Enhanced performance of an
expanded polytetrafluoroethylene-based triboelectric nanogenerator for energy harvesting,"
Nano Energy, vol. 60, pp. 903-911, 2019.
[93] Y. Zhou, W. Deng, J. Xu, and J. Chen, "Engineering materials at the nanoscale for triboelectric
nanogenerators," Cell Reports Physical Science, p. 100142, 2020.
[94] P. Jiang et al., "Signal output of triboelectric nanogenerator at oil–water–solid multiphase
interfaces and its application for dual‐signal chemical sensing," Advanced Materials, vol. 31,
no. 39, p. 1902793, 2019.
[95] D. Graiver, K. Farminer, and R. Narayan, "A review of the fate and effects of silicones in the
environment," Journal of Polymers and the Environment, vol. 11, no. 4, pp. 129-136, 2003.
[96] M. P. Wolf, G. B. Salieb-Beugelaar, and P. Hunziker, "PDMS with designer functionalities—

155
 Properties, modifications strategies, and applications," Progress in Polymer Science, vol. 83, pp.
97-134, 2018.
[97] H. G. Menge, J. O. Kim, and Y. T. Park, "Enhanced Triboelectric Performance of Modified
PDMS Nanocomposite Multilayered Nanogenerators," Materials, vol. 13, no. 18, p. 4156, 2020.
[98] D. Tantraviwat et al., "Highly dispersed porous polydimethylsiloxane for boosting power-
generating performance of triboelectric nanogenerators," Nano Energy, vol. 67, p. 104214, 2020.
[99] X. He, H. Guo, X. Yue, J. Gao, Y. Xi, and C. Hu, "Improving energy conversion efficiency for
triboelectric nanogenerator with capacitor structure by maximizing surface charge density,"
Nanoscale, vol. 7, no. 5, pp. 1896-1903, 2015.
[100] S. Y. Shin, B. Saravanakumar, A. Ramadoss, and S. J. Kim, "Fabrication of PDMS‐based
triboelectric nanogenerator for self‐sustained power source application," International Journal
of Energy Research, vol. 40, no. 3, pp. 288-297, 2016.
[101] X.-Z. Jiang, Y.-J. Sun, Z. Fan, and T.-Y. Zhang, "Integrated flexible, waterproof, transparent,
and self-powered tactile sensing panel," ACS nano, vol. 10, no. 8, pp. 7696-7704, 2016.
[102] W. He, M. Sohn, R. Ma, and D. J. Kang, "Flexible single-electrode triboelectric nanogenerators
with MXene/PDMS composite film for biomechanical motion sensors," Nano Energy, vol. 78,
p. 105383, 2020.
[103] T. Otitoju, A. Ahmad, and B. Ooi, "Polyvinylidene fluoride (PVDF) membrane for oil rejection
from oily wastewater: A performance review," Journal of Water Process Engineering, vol. 14,
pp. 41-59, 2016.
[104] Y. Lu, Y. Jiang, Z. Lou, R. Shi, D. Chen, and G. Shen, "Wearable supercapacitor self-charged
by P (VDF-TrFE) piezoelectric separator," Progress in Natural Science: Materials International,
vol. 30, no. 2, pp. 174-179, 2020.
[105] R. Liu et al., "Poly (vinylidene fluoride) separators for next‐generation lithium based batteries,"
Nano Select, 2021.
[106] C. Baur, D. J. Apo, D. Maurya, S. Priya, and W. Voit, "Advances in piezoelectric polymer
composites for vibrational energy harvesting," in Polymer composites for energy harvesting,
conversion, and storage: ACS Publications, 2014, pp. 1-27.
[107] Q. Jing and S. Kar-Narayan, "Nanostructured polymer-based piezoelectric and triboelectric
materials and devices for energy harvesting applications," Journal of Physics D: Applied
Physics, vol. 51, no. 30, p. 303001, 2018.
[108] H.-Y. Mi et al., "High-performance flexible triboelectric nanogenerator based on porous
aerogels and electrospun nanofibers for energy harvesting and sensitive self-powered sensing,"
Nano Energy, vol. 48, pp. 327-336, 2018.
[109] J. Kim et al., "High‐performance piezoelectric, pyroelectric, and triboelectric nanogenerators
based on P (VDF‐TrFE) with controlled crystallinity and dipole alignment," Advanced
Functional Materials, vol. 27, no. 22, p. 1700702, 2017.
[110] A. Šutka, K. Mālnieks, L. Lapčinskis, M. Timusk, K. Pudzs, and M. Rutkis, "Matching the
directions of electric fields from triboelectric and ferroelectric charges in nanogenerator devices
for boosted performance," Iscience, vol. 23, no. 4, p. 101011, 2020.
[111] T. Huang et al., "“Self‐Matched” Tribo/Piezoelectric Nanogenerators Using Vapor‐Induced
Phase‐Separated Poly (vinylidene fluoride) and Recombinant Spider Silk," Advanced Materials,
vol. 32, no. 10, p. 1907336, 2020.
[112] J.-H. Zhang, Y. Li, J. Du, X. Hao, and H. Huang, "A high-power wearable triboelectric

156
 nanogenerator prepared from self-assembled electrospun poly (vinylidene fluoride) fibers with
a heart-like structure," Journal of Materials Chemistry A, vol. 7, no. 19, pp. 11724-11733, 2019.
[113] G. Kalimuldina et al., "A review of piezoelectric PVDF film by electrospinning and its
applications," Sensors, vol. 20, no. 18, p. 5214, 2020.
[114] G. L. Robertson, Food packaging: principles and practice. CRC press, 2005.
[115] J. Jog, "Crystallization of polyethyleneterephthalate," Journal of Macromolecular Science, Part
C: Polymer Reviews, vol. 35, no. 3, pp. 531-553, 1995.
[116] L. Cheng, Q. Xu, Y. Zheng, X. Jia, and Y. Qin, "A self-improving triboelectric nanogenerator
with improved charge density and increased charge accumulation speed," Nature
communications, vol. 9, no. 1, pp. 1-8, 2018.
[117] J. Xiong et al., "Skin-touch-actuated textile-based triboelectric nanogenerator with black
phosphorus for durable biomechanical energy harvesting," Nature communications, vol. 9, no.
1, pp. 1-9, 2018.
[118] F. R. Fan et al., "Highly transparent and flexible triboelectric nanogenerators: performance
improvements and fundamental mechanisms," Journal of Materials Chemistry A, vol. 2, no. 33,
pp. 13219-13225, 2014.
[119] C. Liu, J. Li, L. Che, S. Chen, Z. Wang, and X. Zhou, "Toward large-scale fabrication of
triboelectric nanogenerator (TENG) with silk-fibroin patches film via spray-coating process,"
Nano Energy, vol. 41, pp. 359-366, 2017.
[120] K. Xia et al., "A triboelectric nanogenerator based on cosmetic fixing powder for mechanical
energy harvesting," Microsystems & nanoengineering, vol. 5, no. 1, pp. 1-9, 2019.
[121] H. Wang, L. Xu, and Z. Wang, "Advances of High-Performance Triboelectric Nanogenerators
for Blue Energy Harvesting," Nanoenergy Advances, vol. 1, no. 1, pp. 32-57, 2021.
[122] F.-R. Fan, L. Lin, G. Zhu, W. Wu, R. Zhang, and Z. L. Wang, "Transparent triboelectric
nanogenerators and self-powered pressure sensors based on micropatterned plastic films," Nano
letters, vol. 12, no. 6, pp. 3109-3114, 2012.
[123] D. W. Kim, J. H. Lee, J. K. Kim, and U. Jeong, "Material aspects of triboelectric energy
generation and sensors," NPG Asia Materials, vol. 12, no. 1, pp. 1-17, 2020.
[124] M. Akdere and T. Schneiders, "Modeling of the electrospinning process," in Advances in
Modeling and Simulation in Textile Engineering: Elsevier, 2021, pp. 237-253.
[125] Z. Wang et al., "Two voltages in contact-separation triboelectric nanogenerator: from
asymmetry to symmetry for maximum output," Nano Energy, vol. 69, p. 104452, 2020.
[126] X. Tao et al., "Significantly enhanced performance of triboelectric nanogenerator by
incorporating BaTiO3 nanoparticles in poly (vinylidene fluoride) film," physica status solidi (a),
vol. 216, no. 7, p. 1900068, 2019.
[127] A. Ahmed et al., "Triboelectric nanogenerator versus piezoelectric generator at low frequency
(< 4 Hz): a quantitative comparison," Iscience, vol. 23, no. 7, p. 101286, 2020.
[128] Y. Park et al., "Ferroelectric multilayer nanocomposites with polarization and stress
concentration structures for enhanced triboelectric performances," ACS nano, vol. 14, no. 6, pp.
7101-7110, 2020.
[129] S. Cheon et al., "High‐performance triboelectric nanogenerators based on electrospun
polyvinylidene fluoride–silver nanowire composite nanofibers," Advanced Functional
Materials, vol. 28, no. 2, p. 1703778, 2018.
[130] Y. Zi, H. Guo, Z. Wen, M.-H. Yeh, C. Hu, and Z. L. Wang, "Harvesting low-frequency (< 5 Hz)

157
 irregular mechanical energy: a possible killer application of triboelectric nanogenerator," ACS
nano, vol. 10, no. 4, pp. 4797-4805, 2016.
[131] Z. Zhang, X. Sun, Y. Chen, D. K. Debeli, and J. Guo, "Comprehensive dependence of
triboelectric nanogenerator on dielectric thickness and external impact for high electric outputs,"
Journal of Applied Physics, vol. 124, no. 4, p. 045106, 2018.
[132] A. Gomes, C. Rodrigues, A. Pereira, and J. Ventura, "Influence of thickness and contact area on
the performance of pdms-based triboelectric nanogenerators," arXiv preprint arXiv:1803.10070,
2018.
[133] G. Min, Y. Xu, P. Cochran, N. Gadegaard, D. M. Mulvihill, and R. Dahiya, "Origin of the contact
force-dependent response of triboelectric nanogenerators," Nano Energy, vol. 83, p. 105829,
2021/05/01/ 2021, doi: [Link]
[134] D. M. Mulvihill and M. P. Sutcliffe, "Effect of tool surface topography on friction with carbon
fibre tows for composite fabric forming," Composites Part A: Applied Science and
Manufacturing, vol. 93, pp. 199-206, 2017.
[135] S. Sneddon, Y. Xu, M. Dixon, D. Rugg, P. Li, and D. M. Mulvihill, "Sensitivity of material
failure to surface roughness: a study on titanium alloys Ti64 and Ti407," Materials & Design,
vol. 200, p. 109438, 2021.
[136] D. Mulvihill, H. Brunskill, M. Kartal, R. Dwyer-Joyce, and D. Nowell, "A comparison of
contact stiffness measurements obtained by the digital image correlation and ultrasound
techniques," Experimental Mechanics, vol. 53, no. 7, pp. 1245-1263, 2013.
[137] M.-L. Seol, S.-H. Lee, J.-W. Han, D. Kim, G.-H. Cho, and Y.-K. Choi, "Impact of contact
pressure on output voltage of triboelectric nanogenerator based on deformation of interfacial
structures," Nano Energy, vol. 17, pp. 63-71, 2015.
[138] A. I. Uddin, P. S. Kumar, K. Hassan, and H. C. Kim, "Enhanced sensing performance of
bimetallic Al/Ag-CNF network and porous PDMS-based triboelectric acetylene gas sensors in
a high humidity atmosphere," Sensors and Actuators B: Chemical, vol. 258, pp. 857-869, 2018.
[139] P. Vasandani, Z.-H. Mao, W. Jia, and M. Sun, "Relationship between triboelectric charge and
contact force for two triboelectric layers," Journal of Electrostatics, vol. 90, pp. 147-152, 2017.
[140] J. Song, L. Gao, X. Tao, and L. Li, "Ultra-Flexible and large-area textile-based triboelectric
nanogenerators with a sandpaper-induced surface microstructure," Materials, vol. 11, no. 11, p.
2120, 2018.
[141] H. Zhang, Y. Lu, A. Ghaffarinejad, and P. Basset, "Progressive contact-separate triboelectric
nanogenerator based on conductive polyurethane foam regulated with a Bennet doubler
conditioning circuit," Nano energy, vol. 51, pp. 10-18, 2018.
[142] X.-W. Zhang et al., "High-performance triboelectric nanogenerator with double-surface shape-
complementary microstructures prepared by using simple sandpaper templates," ACS
Sustainable Chemistry & Engineering, vol. 6, no. 2, pp. 2283-2291, 2018.
[143] M.-L. Seol, J.-W. Han, D.-I. Moon, and M. Meyyappan, "Hysteretic behavior of contact force
response in triboelectric nanogenerator," Nano Energy, vol. 32, pp. 408-413, 2017.
[144] S. Li, Y. Zhou, Y. Zi, G. Zhang, and Z. L. Wang, "Excluding contact electrification in surface
potential measurement using kelvin probe force microscopy," ACS nano, vol. 10, no. 2, pp.
2528-2535, 2016.
[145] Y. S. Zhou et al., "In situ quantitative study of nanoscale triboelectrification and patterning,"
Nano letters, vol. 13, no. 6, pp. 2771-2776, 2013.

158
[146] W. Yang, X. Wang, H. Li, J. Wu, and Y. Hu, "Comprehensive contact analysis for vertical-
contact-mode triboelectric nanogenerators with micro-/nano-textured surfaces," Nano Energy,
vol. 51, pp. 241-249, 2018.
[147] Y. Xu, G. Min, N. Gadegaard, R. Dahiya, and D. M. Mulvihill, "A unified contact force-
dependent model for triboelectric nanogenerators accounting for surface roughness," Nano
Energy, vol. 76, p. 105067, 2020.
[148] Y. Nurmakanov, G. Kalimuldina, G. Nauryzbayev, D. Adair, and Z. Bakenov, "Structural and
Chemical Modifications Towards High-Performance of Triboelectric Nanogenerators,"
Nanoscale Research Letters, vol. 16, no. 1, pp. 1-27, 2021.
[149] C. Bu, F. Li, K. Yin, J. Pang, L. Wang, and K. Wang, "Research progress and prospect of
triboelectric nanogenerators as self-powered human body sensors," ACS Applied Electronic
Materials, vol. 2, no. 4, pp. 863-878, 2020.
[150] J.-H. Seo, K. Zhang, M. Kim, W. Zhou, and Z. Ma, "High-performance flexible BiCMOS
electronics based on single-crystal Si nanomembrane," NPJ Flexible Electronics, vol. 1, no. 1,
pp. 1-7, 2017.
[151] Y. Su et al., "Alveolus-inspired active membrane sensors for self-powered wearable chemical
sensing and breath analysis," ACS nano, vol. 14, no. 5, pp. 6067-6075, 2020.
[152] Y. Su et al., "A wireless energy transmission enabled wearable active acetone biosensor for non-
invasive prediabetes diagnosis," Nano Energy, vol. 74, p. 104941, 2020.
[153] Y. Su et al., "Self-powered room temperature NO2 detection driven by triboelectric
nanogenerator under UV illumination," Nano Energy, vol. 47, pp. 316-324, 2018.
[154] S. Wang et al., "An integrated flexible self-powered wearable respiration sensor," Nano Energy,
vol. 63, p. 103829, 2019.
[155] J. Chen et al., "Enhancing performance of triboelectric nanogenerator by filling high dielectric
nanoparticles into sponge PDMS film," ACS applied materials & interfaces, vol. 8, no. 1, pp.
736-744, 2016.
[156] Y. J. Kim, J. Lee, S. Park, C. Park, C. Park, and H.-J. Choi, "Effect of the relative permittivity
of oxides on the performance of triboelectric nanogenerators," RSC advances, vol. 7, no. 78, pp.
49368-49373, 2017.
[157] J. S. C. Koay, W. C. Gan, A. E. Soh, J. Y. Cheong, K. C. Aw, and T. S. Velayutham, "An
overlapped electron-cloud model for the contact electrification in piezo-assisted triboelectric
nanogenerators via control of piezoelectric polarization," Journal of Materials Chemistry A, vol.
8, no. 48, pp. 25857-25866, 2020.
[158] H. Oh et al., "Hard coating films of fluorine-containing ladder-like structured
polysilsesquioxane as negative triboelectric materials for high-performance triboelectric
generators," Extreme Mechanics Letters, p. 101533, 2021.
[159] H. S. Nalwa, Ferroelectric polymers: chemistry: physics, and applications. CRC Press, 1995.
[160] S. Wang, L. Lin, and Z. L. Wang, "Triboelectric nanogenerators as self-powered active sensors,"
Nano Energy, vol. 11, pp. 436-462, 2015.
[161] J. Chen, H. Guo, C. Hu, and Z. L. Wang, "Robust triboelectric nanogenerator achieved by
centrifugal force induced automatic working mode transition," Advanced Energy Materials, vol.
10, no. 23, p. 2000886, 2020.
[162] K. Y. Lee et al., "Controllable charge transfer by ferroelectric polarization mediated
triboelectricity," Advanced Functional Materials, vol. 26, no. 18, pp. 3067-3073, 2016.

159
[163] W. Seung et al., "Boosting power‐generating performance of triboelectric nanogenerators via
artificial control of ferroelectric polarization and dielectric properties," Advanced Energy
Materials, vol. 7, no. 2, p. 1600988, 2017.
[164] Z. Wang et al., "Ultrahigh electricity generation from low-frequency mechanical energy by
efficient energy management," Joule, vol. 5, no. 2, pp. 441-455, 2021.
[165] H. Niu, H. Zhou, and H. Wang, "Electrospinning: an advanced nanofiber production
technology," Energy Harvesting Properties of Electrospun Nanofibers, pp. 1-44, 2020.
[166] K. Gao, X. Hu, C. Dai, and T. Yi, "Crystal structures of electrospun PVDF membranes and its
separator application for rechargeable lithium metal cells," Materials Science and Engineering:
B, vol. 131, no. 1-3, pp. 100-105, 2006.
[167] Z. He, F. Rault, M. Lewandowski, E. Mohsenzadeh, and F. Salaün, "Electrospun PVDF
nanofibers for piezoelectric applications: A review of the influence of electrospinning
parameters on the β phase and crystallinity enhancement," Polymers, vol. 13, no. 2, p. 174, 2021.
[168] S. M. Damaraju, S. Wu, M. Jaffe, and T. L. Arinzeh, "Structural changes in PVDF fibers due to
electrospinning and its effect on biological function," Biomedical Materials, vol. 8, no. 4, p.
045007, 2013.
[169] H. Jiyong, Z. Yinda, Z. Hele, G. Yuanyuan, and Y. Xudong, "Mixed effect of main
electrospinning parameters on the β-phase crystallinity of electrospun PVDF nanofibers," Smart
Materials and Structures, vol. 26, no. 8, p. 085019, 2017.
[170] X. Ma et al., "Regulating the output performance of triboelectric nanogenerator by using P
(VDF-TrFE) Langmuir monolayers," Nano Energy, vol. 66, p. 104090, 2019.
[171] Y. Su et al., "Piezoelectric fiber composites with polydopamine interfacial layer for self-
powered wearable biomonitoring," Nano Energy, vol. 89, p. 106321, 2021.
[172] H.-S. Kim and I.-K. Park, "Enhanced output power from triboelectric nanogenerators based on
electrospun Eu-doped polyvinylidene fluoride nanofibers," Journal of Physics and Chemistry
of Solids, vol. 117, pp. 188-193, 2018/06/01/ 2018, doi:
[Link]
[173] Y. Wu, J. Qu, W. A. Daoud, L. Wang, and T. Qi, "Flexible composite-nanofiber based piezo-
triboelectric nanogenerators for wearable electronics," Journal of Materials Chemistry A, vol.
7, no. 21, pp. 13347-13355, 2019.
[174] S. Cheon et al., "High-Performance Triboelectric Nanogenerators Based on Electrospun
Polyvinylidene Fluoride–Silver Nanowire Composite Nanofibers," Advanced Functional
Materials, [Link] vol. 28, no. 2, p. 1703778, 2018/01/01 2018,
doi: [Link]
[175] Y. Guo et al., "All-fiber hybrid piezoelectric-enhanced triboelectric nanogenerator for wearable
gesture monitoring," Nano Energy, vol. 48, pp. 152-160, 2018.
[176] T. Bhatta et al., "High-performance triboelectric nanogenerator based on MXene functionalized
polyvinylidene fluoride composite nanofibers," Nano Energy, vol. 81, p. 105670, 2021/03/01/
2021, doi: [Link]
[177] L. Shi et al., "High-performance triboelectric nanogenerator based on electrospun PVDF-
graphene nanosheet composite nanofibers for energy harvesting," Nano Energy, vol. 80, p.
105599, 2021/02/01/ 2021, doi: [Link]
[178] S. M. S. Rana et al., "Electrospun PVDF-TrFE/MXene Nanofiber Mat-Based Triboelectric
Nanogenerator for Smart Home Appliances," ACS Applied Materials & Interfaces, vol. 13, no.

160
 4, pp. 4955-4967, 2021/02/03 2021, doi: 10.1021/acsami.0c17512.
[179] D. Ali, B. Yu, X. Duan, H. Yu, and M. Zhu, "Enhancement of output performance through post-
poling technique on BaTiO3/PDMS-based triboelectric nanogenerator," Nanotechnology, vol.
28, no. 7, p. 075203, 2017/01/13 2017, doi: 10.1088/1361-6528/aa52b7.
[180] H. Patnam, B. Dudem, S. A. Graham, and J. S. Yu, "High-performance and robust triboelectric
nanogenerators based on optimal microstructured poly(vinyl alcohol) and poly(vinylidene
fluoride) polymers for self-powered electronic applications," Energy, vol. 223, p. 120031,
2021/05/15/ 2021, doi: [Link]
[181] Y. Z. Liu, H. Zhang, J. X. Yu, Z. Y. Huang, C. Wang, and Y. Sun, "Ferroelectric P(VDF-
TrFE)/POSS nanocomposite films: compatibility, piezoelectricity, energy harvesting
performance, and mechanical and atomic oxygen erosion," RSC Advances,
10.1039/D0RA01769H vol. 10, no. 29, pp. 17377-17386, 2020, doi: 10.1039/D0RA01769H.
[182] X. Ma et al., "Regulating the output performance of triboelectric nanogenerator by using
P(VDF-TrFE) Langmuir monolayers," Nano Energy, vol. 66, p. 104090, 2019/12/01/ 2019, doi:
[Link]
[183] J. Chen et al., "Self-powered triboelectric micro liquid/gas flow sensor for microfluidics," ACS
nano, vol. 10, no. 8, pp. 8104-8112, 2016.
[184] C. B. Han et al., "Self-powered velocity and trajectory tracking sensor array made of planar
triboelectric nanogenerator pixels," Nano Energy, vol. 9, pp. 325-333, 2014.
[185] Q. Jing et al., "Self-powered triboelectric velocity sensor for dual-mode sensing of rectified
linear and rotary motions," Nano Energy, vol. 10, pp. 305-312, 2014.
[186] T. K. Phan et al., "A self-powered and low pressure loss gas flowmeter based on fluid-elastic
flutter driven triboelectric nanogenerator," Sensors, vol. 20, no. 3, p. 729, 2020.
[187] S. Shen, X. Xiao, and J. Chen, "Wearable triboelectric nanogenerators for heart rate
monitoring," Chemical Communications, 2021.
[188] B. Wang et al., "New Hydrophobic Organic Coating Based Triboelectric Nanogenerator for
Efficient and Stable Hydropower Harvesting," ACS Applied Materials & Interfaces, vol. 12, no.
28, pp. 31351-31359, 2020.
[189] M. Xu et al., "A highly-sensitive wave sensor based on liquid-solid interfacing triboelectric
nanogenerator for smart marine equipment," Nano Energy, vol. 57, pp. 574-580, 2019.
[190] B. Zhang et al., "Self-powered acceleration sensor based on liquid metal triboelectric
nanogenerator for vibration monitoring," ACS nano, vol. 11, no. 7, pp. 7440-7446, 2017.
[191] C. Zhang et al., "Harvesting Wind Energy by a Triboelectric Nanogenerator for an Intelligent
High-Speed Train System," ACS Energy Letters, vol. 6, no. 4, pp. 1490-1499, 2021.
[192] K. Parida, V. Bhavanasi, V. Kumar, R. Bendi, and P. S. Lee, "Self-powered pressure sensor for
ultra-wide range pressure detection," Nano Research, vol. 10, no. 10, pp. 3557-3570, 2017.
[193] X. Xia, Q. Liu, Y. Zhu, and Y. Zi, "Recent advances of triboelectric nanogenerator based
applications in biomedical systems," EcoMat, vol. 2, no. 4, p. e12049, 2020.
[194] J. Shao, T. Jiang, and Z. Wang, "Theoretical foundations of triboelectric nanogenerators
(TENGs)," Science China Technological Sciences, vol. 63, no. 7, pp. 1087-1109, 2020.
[195] Z. L. Wang, "On the expanded Maxwell’s equations for moving charged media system–General
theory, mathematical solutions and applications in TENG," Materials Today, 2021.
[196] J. Luo et al., "Integration of micro-supercapacitors with triboelectric nanogenerators for a
flexible self-charging power unit," Nano Research, vol. 8, no. 12, pp. 3934-3943, 2015.

161
[197] Y. Song et al., "Integrated self-charging power unit with flexible supercapacitor and triboelectric
nanogenerator," Journal of Materials Chemistry A, vol. 4, no. 37, pp. 14298-14306, 2016.
[198] X. Shi, S. Chen, H. Zhang, J. Jiang, Z. Ma, and S. Gong, "Portable self-charging power system
via integration of a flexible paper-based triboelectric nanogenerator and supercapacitor," ACS
Sustainable Chemistry & Engineering, vol. 7, no. 22, pp. 18657-18666, 2019.
[199] M. Parvez Mahmud, N. Huda, S. H. Farjana, M. Asadnia, and C. Lang, "Recent advances in
nanogenerator‐driven self‐powered implantable biomedical devices," Advanced Energy
Materials, vol. 8, no. 2, p. 1701210, 2018.
[200] C. G. Núñez, L. Manjakkal, and R. Dahiya, "Energy autonomous electronic skin," npj Flexible
Electronics, vol. 3, no. 1, pp. 1-24, 2019.
[201] H. Wang and P. P. Mercier, "Near-Zero-Power Temperature Sensing via Tunneling Currents
Through Complementary Metal-Oxide-Semiconductor Transistors," Scientific Reports, vol. 7,
no. 1, p. 4427, 2017/06/30 2017, doi: 10.1038/s41598-017-04705-6.
[202] T. Tekeste, H. Saleh, B. Mohammad, A. Khandoker, H. Jelinek, and M. Ismail, "A Nanowatt
Real-Time Cardiac Autonomic Neuropathy Detector," IEEE Transactions on Biomedical
Circuits and Systems, vol. 12, no. 4, pp. 739-750, 2018, doi: 10.1109/TBCAS.2018.2833624.
[203] P. Escobedo, M. Bhattacharjee, F. Nikbakhtnasrabadi, and R. Dahiya, "Smart Bandage with
Wireless Strain and Temperature Sensors and Battery-less NFC Tag," IEEE Internet of Things
Journal, 2020.
[204] Y. Wang, Y. Yang, and Z. L. Wang, "Triboelectric nanogenerators as flexible power sources,"
npj Flexible Electronics, vol. 1, no. 1, pp. 1-10, 2017.
[205] D. Halliday, R. Resnick, and J. Walker, Fundamentals of physics. John Wiley & Sons, 2013.
[206] S.-J. Park, M.-L. Seol, S.-B. Jeon, D. Kim, D. Lee, and Y.-K. Choi, "Surface engineering of
triboelectric nanogenerator with an electrodeposited gold nanoflower structure," Scientific
reports, vol. 5, no. 1, pp. 1-7, 2015.
[207] D. Hong, Y.-M. Choi, and J. Jeong, "Test bed for contact-mode triboelectric nanogenerator,"
Review of Scientific Instruments, vol. 89, no. 6, p. 065110, 2018.
[208] Y. Xiong and X. Tao, "Compression garments for medical therapy and sports," Polymers, vol.
10, no. 6, p. 663, 2018.
[209] F. You, J. Wang, X. Luo, Y. Li, and X. Zhang, "Garment's pressure sensation (1): subjective
assessment and predictability for the sensation," International Journal of Clothing Science and
Technology, 2002.
[210] M. H. Müser et al., "Meeting the contact-mechanics challenge," Tribology Letters, vol. 65, no.
4, pp. 1-18, 2017.
[211] L. Helseth, "Optical force sensing principle based on transparent elastomer with a rough
surface," Sensors and Actuators A: Physical, vol. 263, pp. 667-676, 2017.
[212] L. Dhakar et al., "Large scale triboelectric nanogenerator and self-powered pressure sensor array
using low cost roll-to-roll UV embossing," Scientific reports, vol. 6, no. 1, pp. 1-10, 2016.
[213] J. Chen et al., "Transparent triboelectric generators based on glass and polydimethylsiloxane,"
Nano Energy, vol. 30, pp. 235-241, 2016.
[214] Z. Saadatania, S. G. Mosanenzadeh, E. Esmailzadeh, and H. Naguib, "Performance-enhanced
triboelectric nanogenerator using polyimide aerogel for energy harvesting and sensing," in
Smart Materials and Nondestructive Evaluation for Energy Systems IV, 2018, vol. 10601:
International Society for Optics and Photonics, p. 1060105.

162
[215] J. Peng et al., "A composite generator film impregnated with cellulose nanocrystals for
enhanced triboelectric performance," Nanoscale, vol. 9, no. 4, pp. 1428-1433, 2017.
[216] P. Bai et al., "Dipole-moment-induced effect on contact electrification for triboelectric
nanogenerators," Nano Research, vol. 7, no. 7, pp. 990-997, 2014.
[217] H. Chen et al., "Effects of truck tire pressures on pavements," 1987.
[218] L. Zhou, H. Brunskill, M. Pletz, W. Daves, S. Scheriau, and R. Lewis, "Real-time measurement
of dynamic wheel-rail contacts using ultrasonic reflectometry," Journal of Tribology, vol. 141,
no. 6, p. 061401, 2019.
[219] J. G. Rose and T. E. Guenther, "Vehicle Tire-Pavement Interfacial Surface Pressure
Measurements and Assessments," 2009.
[220] A. Führböter, "Model and prototype tests for wave impact and run-up on a uniform 1: 4 slope,"
Coastal engineering, vol. 10, no. 1, pp. 49-84, 1986.
[221] D. Kisacik, "Loading conditions due to violent wave impacts on coastal structures with
cantilever surfaces," Ghent University, 2012.
[222] J.-R. Henry and F. Bailey, "SLAMMING OF SHIPS: A CRITICAL REVIEW OF THE
CURRENT STATE OF KNOWLEDGE," 1970.
[223] J. C. Daidola, V. Mishkevich, and A. Bromwell, "Hydrodynamic Impact on Displacement Ship
Hulls. An Assessment of the State of the Art. Bibliography," ROSENBLATT (M) AND SON
INC NEW YORK, 1995.
[224] J.-H. Yang, Y.-K. Kim, and J. Y. Lee, "Simplified process for manufacturing macroscale patterns
to enhance voltage generation by a triboelectric generator," Energies, vol. 8, no. 11, pp. 12729-
12740, 2015.
[225] G.-G. Cheng et al., "Effect of argon plasma treatment on the output performance of triboelectric
nanogenerator," Applied Surface Science, vol. 412, pp. 350-356, 2017.
[226] C. K. Jeong et al., "Topographically-designed triboelectric nanogenerator via block copolymer
self-assembly," Nano letters, vol. 14, no. 12, pp. 7031-7038, 2014.
[227] K. Dong, X. Peng, and Z. L. Wang, "Fiber/fabric‐based piezoelectric and triboelectric
nanogenerators for flexible/stretchable and wearable electronics and artificial intelligence,"
Advanced Materials, vol. 32, no. 5, p. 1902549, 2020.
[228] J. Chen et al., "Actuation and sensor integrated self-powered cantilever system based on TENG
technology," Nano Energy, vol. 64, p. 103920, 2019.
[229] H. Guo et al., "A highly efficient triboelectric negative air ion generator," Nature Sustainability,
vol. 4, no. 2, pp. 147-153, 2021.
[230] Y. Su et al., "Self‐Powered Respiration Monitoring Enabled By a Triboelectric Nanogenerator,"
Advanced Materials, vol. 33, no. 35, p. 2101262, 2021.
[231] Y. Su et al., "Muscle fibers inspired high‐performance piezoelectric textiles for wearable
physiological monitoring," Advanced Functional Materials, vol. 31, no. 19, p. 2010962, 2021.
[232] S. Ma et al., "Highly oriented electrospun P (VDF‐TrFE) fibers via mechanical stretching for
wearable motion sensing," Advanced Materials Technologies, vol. 3, no. 7, p. 1800033, 2018.
[233] Y. Liu, H. Zhang, J. Yu, Z. Huang, C. Wang, and Y. Sun, "Ferroelectric P (VDF-TrFE)/POSS
nanocomposite films: Compatibility, piezoelectricity, energy harvesting performance, and
mechanical and atomic oxygen erosion," RSC Advances, vol. 10, no. 29, pp. 17377-17386, 2020.
[234] L. Manjakkal, W. T. Navaraj, C. G. Núñez, and R. Dahiya, "Graphene–graphite polyurethane
composite based high‐energy density flexible supercapacitors," Advanced Science, vol. 6, no. 7,

163
 p. 1802251, 2019.
[235] A. Pullanchiyodan, L. Manjakkal, S. Dervin, D. Shakthivel, and R. Dahiya, "Metal coated
conductive fabrics with graphite electrodes and biocompatible gel electrolyte for wearable
supercapacitors," Advanced Materials Technologies, vol. 5, no. 5, p. 1901107, 2020.
[236] J. Kim et al., "High-Performance Piezoelectric, Pyroelectric, and Triboelectric Nanogenerators
Based on P(VDF-TrFE) with Controlled Crystallinity and Dipole Alignment," vol. 27, no. 22,
p. 1700702, 2017, doi: [Link]
[237] P. Martins, A. Lopes, and S. Lanceros-Mendez, "Electroactive phases of poly (vinylidene
fluoride): Determination, processing and applications," Progress in polymer science, vol. 39,
no. 4, pp. 683-706, 2014.
[238] E. S. Hosseini, L. Manjakkal, D. Shakthivel, and R. Dahiya, "Glycine–chitosan-based flexible
biodegradable piezoelectric pressure sensor," ACS applied materials & interfaces, vol. 12, no.
8, pp. 9008-9016, 2020.
[239] D. Yoon, "Tetragonality of barium titanate powder for a ceramic capacitor application," Journal
of Ceramic Processing Research, vol. 7, no. 4, p. 343, 2006.
[240] M. Acosta et al., "BaTiO3-based piezoelectrics: Fundamentals, current status, and
perspectives," Applied Physics Reviews, vol. 4, no. 4, p. 041305, 2017.
[241] J. Nunes-Pereira et al., "Energy harvesting performance of BaTiO3/poly (vinylidene fluoride–
trifluoroethylene) spin coated nanocomposites," Composites Part B: Engineering, vol. 72, pp.
130-136, 2015.
[242] X. Hu et al., "Improved Piezoelectric Sensing Performance of P (VDF–TrFE) Nanofibers by
Utilizing BTO Nanoparticles and Penetrated Electrodes," ACS applied materials & interfaces,
vol. 11, no. 7, pp. 7379-7386, 2019.
[243] A. P. Indolia and M. Gaur, "Investigation of structural and thermal characteristics of PVDF/ZnO
nanocomposites," Journal of thermal analysis and calorimetry, vol. 113, no. 2, pp. 821-830,
2013.
[244] K. Yu, H. Wang, Y. Zhou, Y. Bai, and Y. Niu, "Enhanced dielectric properties of
BaTiO3/poly(vinylidene fluoride) nanocomposites for energy storage applications," Journal of
Applied Physics, vol. 113, no. 3, p. 034105, 2013/01/21 2013, doi: 10.1063/1.4776740.
[245] S. Gupta, D. Shakthivel, L. Lorenzelli, and R. Dahiya, "Temperature Compensated Tactile
Sensing Using MOSFET With P(VDF-TrFE)/BaTiO₃ Capacitor as Extended
Gate," IEEE Sensors Journal, vol. 19, no. 2, pp. 435-442, 2019, doi:
10.1109/JSEN.2018.2876678.
[246] C. Xu et al., "On the electron‐transfer mechanism in the contact‐electrification effect,"
Advanced materials, vol. 30, no. 15, p. 1706790, 2018.
[247] I. Polonsky and L. Keer, "A numerical method for solving rough contact problems based on the
multi-level multi-summation and conjugate gradient techniques," Wear, vol. 231, no. 2, pp. 206-
219, 1999.
[248] J. Zou et al., "Coupled Supercapacitor and Triboelectric Nanogenerator Boost Biomimetic
Pressure Sensor," vol. 8, no. 10, p. 1702671, 2018, doi:
[Link]
[249] W. Xiong et al., "A wearable system based on core-shell structured peptide-Co9S8
supercapacitor and triboelectric nanogenerator," Nano Energy, vol. 66, p. 104149, 2019/12/01/
2019, doi: [Link]

164
[250] A. Maitra et al., "Triboelectric Nanogenerator Driven Self-Charging and Self-Healing Flexible
Asymmetric Supercapacitor Power Cell for Direct Power Generation," ACS Applied Materials
& Interfaces, vol. 11, no. 5, pp. 5022-5036, 2019/02/06 2019, doi: 10.1021/acsami.8b19044.
[251] J. McWhorter, L. Brown, and L. Khansa, "A wearable health monitoring system for
posttraumatic stress disorder," Biologically inspired cognitive architectures, vol. 22, pp. 44-50,
2017.
[252] R. Dahiya et al., "Large-area soft e-skin: The challenges beyond sensor designs," Proceedings
of the IEEE, vol. 107, no. 10, pp. 2016-2033, 2019.
[253] W. Taube Navaraj et al., "Nanowire FET based neural element for robotic tactile sensing skin,"
Frontiers in neuroscience, vol. 11, p. 501, 2017.
[254] C. G. Núñez, W. T. Navaraj, E. O. Polat, and R. Dahiya, "Energy‐autonomous, flexible, and
transparent tactile skin," Advanced Functional Materials, vol. 27, no. 18, p. 1606287, 2017.
[255] Y. Zhou, M. Shen, X. Cui, Y. Shao, L. Li, and Y. Zhang, "Triboelectric nanogenerator based
self-powered sensor for artificial intelligence," Nano Energy, p. 105887, 2021.
[256] M. Soni, M. Bhattacharjee, M. Ntagios, and R. Dahiya, "Printed temperature sensor based on
PEDOT: PSS-graphene oxide composite," IEEE Sensors Journal, vol. 20, no. 14, pp. 7525-
7531, 2020.
[257] L. Manjakkal, S. Dervin, and R. Dahiya, "Flexible potentiometric pH sensors for wearable
systems," RSC Advances, vol. 10, no. 15, pp. 8594-8617, 2020.
[258] Y. Su et al., "Self-powered humidity sensor based on triboelectric nanogenerator," in 2017
IEEE SENSORS, 2017: IEEE, pp. 1-3.
[259] Y. Kumaresan, O. Ozioko, and R. Dahiya, "Effect of Dielectric and Stiffness of Soft Material
between the Electrodes of a Capacitive Pressure Sensor on its Performance," in 2020 IEEE
International Conference on Flexible and Printable Sensors and Systems (FLEPS), 2020: IEEE,
pp. 1-4.
[260] M. Ntagios, H. Nassar, A. Pullanchiyodan, W. T. Navaraj, and R. Dahiya, "Robotic hands with
intrinsic tactile sensing via 3D printed soft pressure sensors," Advanced Intelligent Systems, vol.
2, no. 6, p. 1900080, 2020.
[261] Y. Kumaresan, O. Ozioko, and R. Dahiya, "Multifunctional electronic skin with a stack of
temperature and pressure sensor arrays," IEEE Sensors Journal, 2021.
[262] W. Navaraj and R. Dahiya, "Fingerprint‐enhanced capacitive‐piezoelectric flexible sensing skin
to discriminate static and dynamic tactile stimuli," Advanced Intelligent Systems, vol. 1, no. 7,
p. 1900051, 2019.
[263] O. Ozioko, P. Karipoth, P. Escobedo, M. Ntagios, A. Pullanchiyodan, and R. Dahiya, "SensAct:
The Soft and Squishy Tactile Sensor with Integrated Flexible Actuator," Advanced Intelligent
Systems, vol. 3, no. 3, p. 1900145, 2021.
[264] R. S. Dahiya, P. Mittendorfer, M. Valle, G. Cheng, and V. J. Lumelsky, "Directions toward
effective utilization of tactile skin: A review," IEEE Sensors Journal, vol. 13, no. 11, pp. 4121-
4138, 2013.
[265] Y. Zang, F. Zhang, C.-a. Di, and D. Zhu, "Advances of flexible pressure sensors toward artificial
intelligence and health care applications," Materials Horizons, vol. 2, no. 2, pp. 140-156, 2015.
[266] M. Li, J. Liang, X. Wang, and M. Zhang, "Ultra-sensitive flexible pressure sensor based on
microstructured electrode," Sensors, vol. 20, no. 2, p. 371, 2020.
[267] J. A. Rogers, "A diverse printed future," Nature, vol. 468, no. 7321, pp. 177-178, 2010.

165
[268] D.-H. Kim et al., "Epidermal electronics," science, vol. 333, no. 6044, pp. 838-843, 2011.
[269] S. Xu, Y. Qin, C. Xu, Y. Wei, R. Yang, and Z. L. Wang, "Self-powered nanowire devices," Nature
nanotechnology, vol. 5, no. 5, pp. 366-373, 2010.
[270] Z. L. Wang, L. Lin, J. Chen, S. Niu, and Y. Zi, "Triboelectrification," in Triboelectric
Nanogenerators: Springer, 2016, pp. 1-19.
[271] A. K. Kasi, J. K. Kasi, M. Uddin, and M. Bokhari, "Triboelectric nanogenerator as self-powered
impact force sensor for falling object," Current Applied Physics, vol. 20, no. 1, pp. 137-144,
2020.

166