RESEARCH ARTICLE

www.advenergymat.de

The Distance-Dependent Electric Field Theory for Sliding

Mode Triboelectric Nanogenerators
Rameesh L. Bulathsinghala, Jonathan H.B. Deane, and R.D. Ishara G. Dharmasena*

as energy harvesters and self-powered ac-

Triboelectric nanogenerators (TENGs) have demonstrated outstanding tive sensors in a wide range of applica-
potential as energy harvesters and sensors for future wearable electronics. tions including the Internet of Things
However, TENGs still require major improvements in their theory and (IoT), 5G, healthcare, personal electron-
optimization, especially for the sliding-mode designs. Addressing this gap, a ics, wearable technologies, transport, and
security. Some TENG designs have re-
novel theoretical model based on the distance-dependent electric field (DDEF) ported excellent output performances
theory for sliding mode TENGs is presented here. The model is used to with high instantaneous power den-
simulate the electrical outputs and impedance behaviour of a sliding mode sities (>500 W m2 ) and efficiencies
TENG, and the results are verified experimentally. The outcomes indicate that (>50%).[3,11] In addition, TENGs can be
compared to existing theoretical models, this new model provides higher constructed using a wide range com-
monly available materials at low cost, via
accuracy in representing experimental TENG. Next, all the primary
scalable manufacturing techniques.[1,7]
parameters (material, structural and motion parameters) which affect the Some of their unique advantages com-
sliding mode TENG are analyzed, uncovering new optimization strategies and pared to other mechanical energy har-
more comprehensive parametric analysis compared to previous models. More vesting technologies include the self-
importantly, the theoretical approach is equally applicable to sliding mode powered operation, flexibility, wearabil-
TENG as well as contact-separation mode TENG. This eliminates the need for ity, capability of efficiently operating us-
ing low motion frequencies such as hu-
bespoke capacitor models for each TENG type, leading to a universal man motion and ambient vibrations.[2,4,8]
theoretical platform for TENGs. The new model facilitates cross-comparison More recently, TENGs have been increas-
between different TENG working modes, uncovering a range of previously ingly used as effective energy harvesting
unreported output trends. Hence, this work significantly expands the sources for high entropy energy.[12,13]
understanding of TENGs, paving way to more efficient future device designs. Numerous TENG architectures were
demonstrated in literature on different
working modes including vertical con-
tact separation mode (VCSTENG), single
1. Introduction electrode mode (STENG), free-standing triboelectric layer mode
(FSTENG) and lateral sliding mode (LSTEG) TENG designs.[10]
Triboelectric Nanogenerators (TENGs) have emerged as an in- However, all these structures depend on either contact-separation
novative and sustainable mechanical energy harvesting technol- (contact-separation mode TENG) or sliding (sliding mode TENG)
ogy within the past decade.[1–10] They have been demonstrated movements between triboelectric surfaces to produce electrical
outputs. TENGs operate through a combination of static charg-
ing and electrostatic induction,[14] and their output generation is
R. L. Bulathsinghala, R. I. G. Dharmasena governed by Maxwell’s displacement current.[15] The contact sur-
Wolfson School of Mechanical, Electrical and Manufacturing Engineering
Loughborough University faces of a TENG gather static charges of opposite polarities due to
Loughborough, Leicestershire LE11 3TU, UK mutual contact, caused by the transfer of electrons,[16,17] ions[18]
E-mail: [email protected] or charged materials.[19] Once triboelectrically charged, relative
J. H. Deane movement of these surfaces induces a potential difference be-
Department of Mathematics tween their electrodes, creating electrical outputs.
University of Surrey
In order to design efficient TENGs, a deeper understanding is
Guildford GU2 7XH, UK
required on their working principles, the fundamental parame-
The ORCID identification number(s) for the author(s) of this article ters that control their outputs, as well as the techniques to bet-
can be found under https://doi.org/10.1002/aenm.202403853 ter simulate experimental TENGs and compare performances.[1]
© 2024 The Author(s). Advanced Energy Materials published by This has been recently highlighted as an urgent need to progress
Wiley-VCH GmbH. This is an open access article under the terms of the the TENG research field toward practical applications.[1,20–23] In
Creative Commons Attribution License, which permits use, distribution particular, sliding mode TENGs, which account for a large per-
and reproduction in any medium, provided the original work is properly
cited.
centage of all TENG designs, have been subjected to limited

DOI: 10.1002/aenm.202403853

Adv. Energy Mater. 2024, 2403853 2403853 (1 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

theoretical developments compared to contact-separation mode to derive the potential and the output characteristics of TENG
TENGs, which has impeded their progress.[20,22,24,25] electrodes.[22,33,34] However, a major drawback of this method is
There are several theoretical approaches to model and simulate its increased complexity and reliance on advanced calculations,
sliding mode TENG, which include analytical or numerical ap- making it less desirable in practical applications. Consequently,
proaches such as the classical (capacitor based) models[25] and 3D numerous approximations to its limits are necessary to sim-
electrical potential models,[26] as well as finite element models.[27] plify and solve this model. Although this model has presented
The earliest among these is the classical theoretical approach several dimensional analysis for TENG, applicability of this 3D
based on capacitor models which utilized bespoke configurations model into bridging the fundamental TENG parameters with
of capacitor elements and circuit elements to simulate the out- their output generation and impedance behavior has not been
puts of TENGs.[24,25] These models derived a relationship be- demonstrated for the sliding mode TENG.[26] Further, compre-
tween the voltage (V), charge (Q) and layer separation (x) of the hensive theoretical and empirical analysis on the output behavior
TENG (V-Q-x relationship), using different parallel-plate capac- of the sliding mode TENG under different conditions, the under-
itor circuits to simulate the output behavior of TENGs.[25,28,29] lying factors which dictate such behavior, and their optimization
This model provides simplicity, and, in most cases, it is conve- strategies have not yet been descriptively demonstrated with this
nient to be used in approximating TENG outputs. However, this approach.[26,35]
model assumes that the triboelectric contact surfaces contain in- This demands for new theoretical developments for sliding
finitely large surface area with respect to its relative motion pro- mode TENG which can a) build a theoretical platform to fully
file, thus, assumes uniform electric fields between their contact describe TENG working principles, b) more accurately represent
surfaces.[14,30] This makes the model too idealized to represent outputs of practical TENGs, c) provide detailed analysis on funda-
practical TENG designs, and the predictions of this model de- mental parameters affecting TENG outputs, d) optimize electri-
viate notably from real world TENGs, as demonstrated in sev- cal outputs and impedance of TENGs, and e) contain consistency
eral studies including our previous work.[14,30,31] Furthermore, and versatility to facilitate comparison between different (contact
the model is typically limited to planar geometries with simple separation and sliding) TENG working modes.
shapes. In addition, it is difficult to obtain a detailed understand- In our recent work, we introduced the distance-dependent
ing of the TENG working principles such as the dielectric po- electric field (DDEF) concept for contact-separation mode
larization and output induction using this model.[14,31] More im- TENGs, marking revolutionary improvements in their theory and
portantly, this model so far, has not developed a detailed analysis simulations.[14,30,31,33] Based on Maxwell’s equations, the DDEF
between the physical parameters of the TENG and their output model simulates the electric field behavior of finite TENG sur-
performance.[25] This has created a void in linking critical fac- faces and accurately predicts the voltage, current, charge, power
tors such as material, device and motion parameters of the TENG outputs, and impedance behavior of contact-separation mode
with their electrical outputs and impedance performance, which TENGs.[14,30,31] The DDEF model was the first model to fully de-
in turn, impedes their optimization process. Finally, the classical scribe TENG working principles including dielectric polarization
approach requires a bespoke theoretical model to be developed and output induction.[14] This model unified the TENG theory by
for each TENG working mode using different assumptions and creating a single platform to represent all TENG types which op-
configurations.[30,31] Therefore, the consistency and comparabil- erate using contact-separation movements, including the vertical
ity of these different models is poor, making it extremely diffi- contact separation TENG, free-standing mode TENG, single elec-
cult to compare the performance between various TENG working trode mode TENG and multi-layered TENG designs.[30] More-
modes.[14,30,31] over, the DDEF model successfully predicted the output trends
Next category is the finite element modelling, which provides of non-planar TENG architectures for the first time by simulating
improved accuracy especially for potential measurements, and the output trends of convex and concave surfaces.[14] This model
improved comparability between different working modes.[27,32] was consequently used to derive the TENG power transfer the-
However, this method requires specialized software packages ory and the impedance plots for contact-separation TENGs, and
and expertise, which increases the cost and time, and restricts its the resultant device optimization techniques were shown to in-
universal applicability.[27] Moreover, these methods create com- crease the power outputs by over two orders of magnitude, along
plications in simulating rate-dependent or movement-dependent with similar reductions in their internal impedance.[31] Further-
outputs such as the current and power of the TENG. Further, more, the DDEF model was used as the basis for numerous fun-
many of these finite element simulations depend on in-built elec- damental findings on the behavior and optimization of contact-
trostatic modules, therefore, it is difficult to obtain a detailed pic- separation mode TENGs including the intrinsic asymmetry of
ture of the underlying working mechanism of TENGs through current outputs,[36] asymmetric external charging[37] and the con-
them. Consequently, similar to classical models, there is a lack tact mode DC-TENG concept.[38]
of comprehensive studies on linking material, device and mo- Despite its exceptional benefits, DDEF model has been re-
tion parameters of the TENG into their output performance and stricted to contact-separation mode TENG types, and a theoretical
impedance behavior, which makes the device optimization pro- platform based on this concept for the sliding mode TENG has
cess challenging. not yet been realized. Recent literature has emphasized issues
More recently, several theoretical models were presented based such as the lack of a suitable coordinate system, as the root cause
on 3D electrical potential simulations,[22,33,34] showing better ac- for this limitation.[20,22,24] Developing a new model based on the
curacy and versatility compared to both the previous approaches. DDEF theory for sliding mode TENGs will be a key milestone in
This model considers the 3D arrangement of electric fields orig- the TENG theory, which provides the pathway not only to under-
inating from the triboelectric charges and the output charges stand, simulate and optimize these devices, but also to compare

Adv. Energy Mater. 2024, 2403853 2403853 (2 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

the performance between contact-separation and sliding modes shown in Figure 1a. The two surfaces (layer 1 and layer 2) are
TENG types. This will create a consistent, reliable and versatile placed in the x-y plane, and they have Length L, Width W. The
universal theory for TENG, which addresses the challenges (a) to top surface (layer 1) has a charge density of + 𝜎 and the sur-
(e) highlighted above. face charge density of bottom surface (layer 2) is − 𝜎. The non-
Bridging this gap, herein, we develop the first DDEF model overlapped length of the top layer is marked as distance x, which
based theoretical platform for sliding mode TENGs. The elec- varies between 0 (fully overlapped scenario) and L (fully non-
tric field behavior of finite TENG surfaces caused by triboelectric overlapped scenario). In practical terms, distance x for a TENG is
charges and output charges is modelled using the DDEF theory, often a function of time (t) (e.g.,: constant velocity or sinusoidal
with respect to their relative sliding movements. This facilitates movement) which can be denoted as x(t).
a simple yet more accurate method of simulating the voltage, In the overlapped region of the two surfaces, both positive
charge, current and power outputs of the sliding mode TENGs, and negative triboelectric charges are approximately on the same
and their impedance behavior. The simulated outputs are then plane. Therefore, we can reasonably assume that the positive and
verified using empirical results, proving the accuracy and con- negative electric fields in z direction in the overlapped area will
sistency of the new model. Next, a detailed study is conducted cancel each other out. Hence, the overall electric field perpen-
analyzing all the basic parameters that affect TENG outputs (ma- dicular to layer 1 depends on the surface area W*x(t), and the
terial, device and motion), revealing a unique set of output opti- distance from the charged sheet (z) at which the electric field is
mization strategies for sliding mode TENG. This creates the links considered. Using the DDEF model (equation (1), this electric
between TENG output trends observed under different operating field can be represented as
conditions to the underlying physical parameters governing such
behavior, which is highly desired for the field of sliding mode ⎛ ⎞
TENG. Finally, a theoretical and experimental comparison is car- 𝜎 ⎜ x (t) ∕W ⎟
ried out between equivalent contact-separation and sliding mode E(x(t),z) = arctan ⎜ √ ⎟
𝜋𝜀 ⎜ 2 (z∕W ) 4(z∕W )2 + (x (t) ∕W )2 + 1 ⎟
TENG devices to assess their relative output performances, cre- ⎝ ⎠
ating the long-awaited universal theoretical platform for TENG.
𝜎
= f (x (t) , z) (2)
𝜋𝜀
2. Derivation of the DDEF Theoretical Model for Equation (2) forms the basis for the DDEF model for sliding
Sliding Mode TENG mode TENGs.
When two triboelectric surfaces of a TENG contact or rub against Now let us consider the case of a lateral sliding mode TENG
each other, they become statically charged with opposite polari- (LSTENG) with similar dimensions and parameters to the previ-
ties. After a few initial contact-separation or sliding cycles, these ous example, as shown in Figure 1b. Herein, the thickness and
static charge levels reach a dynamic equilibrium (saturated state), the dielectric constant of the top triboelectric layer (layer 1) and
and in practical terms, this initial charging period is consid- bottom triboelectric layer (layer 2) are given by d1 and d2 and 𝜖 1
ered negligible.[24] Thereafter, the triboelectric charge densities and 𝜖 2 respectively, and the triboelectric charge density is 𝜎 T , with
on these surfaces do not change significantly for long periods two electrodes being placed on the reverse sides of the triboelec-
of time relative to the timescale of the operation of a TENG.[10] tric contact surfaces.
Hence, like all previous theoretical models, we assume triboelec-
tric charges at their saturated state, evenly distributed on the two
triboelectric contact surfaces. 2.1. Open Circuit Condition
Consider a statically charged rectangular surface with Length
L, width W (placed in the x-y axis plane), surface charge density In the open circuit condition (Figure 1c), there are no output
𝜎, located in a medium with dielectric constant 𝜖 (Figure S1, Sup- charges on the electrodes, therefore the potential of each elec-
porting Information). The DDEF theory uses Gauss’s law to de- trode depends on the non-overlapped surface area of the respec-
termine the electric field (Ez ) generated by this charged surface tive triboelectric surface. Therefore, using equation (2), the po-
in a direction perpendicular to its midpoint (along the z-axis in tential of the triboelectric layer 1 (Φ1 ) and triboelectric layer 2
this example), as explained in Note S1 (Supporting Information). (Φ2 ) can be evaluated as follows:
This distance-dependent electric field (Ez ) for this surface is rep-
[∞ ]
resented as, 𝜎T
Φ1 = ∫ f (x (t) , z) dz (3)
𝜋𝜀1 d1
⎛ ⎞
𝜎 ⎜ L∕W ⎟ 𝜎
Ez = arctan ⎜ √ ⎟ = 𝜋𝜀 f (z) (1) [∞ ]
𝜋𝜀 ⎜ 2 (z∕W) 4(z∕W) + (L∕W) + 1 ⎟ 𝜎T
⎝
2 2
⎠ Φ2 = − ∫ f (x (t) , z) dz (4)
𝜋𝜀2 d2

where the term z in equation (1) represents the distance from the
charged surface.[14,30,31] Equation (1) lays the foundation for the Therefore, the open circuit voltage (VOC ) of the TENG is given
DDEF model. by,
Let’s consider the generic case of two charged dielectric sur-
faces partially overlapping and moving in lateral sliding-mode as VOC = Φ1 − Φ2 (5)

Adv. Energy Mater. 2024, 2403853 2403853 (3 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 1. a–e) Derivation of the DDEF model for the sliding mode TENG, showing a) schematic of the generic case where two dielectric charged surfaces
slide on each other; b) schematic of the sliding mode TENG indicating triboelectrically charged dielectric surfaces, electrodes, and the device dimensions;
schematic of the c) open circuit state, d) short circuit state, and the e) load conditions, of the sliding mode TENG. f,g) Schematics of the experimental
setup used to characterise the TENG, demonstrating the configuration for the f) contact-separation mode, and g) sliding mode.

2.2. Short Circuit Condition the potential of the two electrodes can be expressed as
follows:
In the short circuit condition, the potential difference cre-
[∞ ] [ ]
ated by the non-overlapping areas of the triboelectric sur- 𝜎T 𝜎 (t ) ∞
faces cause a charge flow between the electrodes (Figure 1d). Φ1 = ∫ f (x (t) , z) dz − u ∫ f (x (t) , z) dz (6)
𝜋𝜀1 d1 𝜋𝜀1 0
The classical theoretical approach for sliding mode TENG
suggests that the output charges (referred to as transferred [∞ ] [ ]
charges “Q”) are distributed in areas of the electrodes where −𝜎T 𝜎 (t ) ∞
Φ2 = ∫ f (x (t) , z) dz + u ∫ f (x (t) , z) dz (7)
the triboelectric surfaces are overlapped as well as non- 𝜋𝜀2 d2 𝜋𝜀2 0
overlapped. However, in our approach, we suggest that the
output charges are distributed only within the non-overlapped where 𝜎 u (t) represents the output charge density (herein, output
regions of the electrodes for a typical TENG which con- charge density is represented as a function of time as it relies on
tains triboelectric layers with equal contact surface areas, and x(t)).
equal & opposite triboelectric charge densities (this is fur- As per the short circuit scenario, Φ1 − Φ2 = 0 therefore,
ther explained in Supporting Note 2 of the Supporting In-
formation Section). Therefore, the potential of each electrode [ ]
1
is affected by the triboelectric charges and output charges 𝜎T 𝜀1
∫d∞ f (x (t) , z) dz + 𝜀1 ∫d∞ f (x (t) , z) dz
1 2
𝜎u ( t ) = ( )
2
related to the non-overlapped region of the corresponding (8)
1 1
triboelectric layer. Using the DDEF equation (equation 2), 𝜀
+ 𝜀 ∫0 f x t , z) dz
∞
( ( )
1 2

Adv. Energy Mater. 2024, 2403853 2403853 (4 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

The short circuit current of the TENG can be evaluated by ob- According to TENG Power Transfer Theory, the effective capac-
taining the time derivative of the product between output charge itance of the system K(t)eff can be defined as:
density and the (non-overlapped) surface area of the TENG, since
the output charge distribution is restricted to the non-overlapped A (t ) 𝜋A (t)
region. Herein, it is important to note that this surface area is a K (t)eff = = ( ) (17)
G (t) 1
+ 1
∫0∞ f (x (t) , z) dz
time dependent function A(t) described by 𝜀 1 𝜀2

A (t) = x (t) ∗W (9)

Furthermore, for a sliding mode TENG which operates at a
motion frequency n (which is reasonably low (<1kHz) given the
Therefore, the short circuit current Isc (t) can be evaluated as:
nature of TENG applications), the magnitude of the internal
d [ ] impedance of the TENG at time t (ZS (t)) can be defined as,
Isc (t) = A (t) ∗𝜎U (t) (10)
dt
1
ZS (t) = (18)
2𝜋nK (t)eff
2.3. Load Conditions
Hence, according to the TENG Power Transfer Theory, the out-
Considering a load resistor R connected between the electrodes put current through the external load at time t, IL (t), can be as-
(Figure 1e) and using Ohm’s law, the output voltage of the TENG sessed using,
V(t) can be represented by

d [ ] Z S (t )
V (t) = Φ1 − Φ2 = R A (t) 𝜎U (t) (11) I L (t ) = I S (t ) (19)
dt Z S (t ) + Z L

Using equations (6) and (7), this can be expanded as: where the source current IS (t) is represented by the short circuit
( ) current (Isc (t)) of the TENG.
d [ ] 𝜎 (t ) 1 1 ∞
R A (t) 𝜎U (t) + U + ∫ f (x (t) , z) dz Therefore, the power output of the sliding mode TENG POut (t)
dt 𝜋 𝜀1 𝜀2 0 can be assessed as:
[ ]
𝜎 1 ∞ 1 ∞
− T ∫ f (x (t) , z) dz + ∫ f (x (t) , z) dz = 0 (12) POut (t) = IL (t)2 ZL (20)
𝜋 𝜀1 d1 𝜀2 d2

Equation (12) can be used to assess the output current I(t) Equations (17) – (20) forms the Power Transfer Theory for slid-
which is given by dtd [A(t)𝜎U (t)], and the power output of the TENG ing mode TENGs and lays the foundation for sliding mode TENG
P(t) can be assessed using, impedance plots that will be discussed later in section 4 of this
work.
P (t) = I2 (t) ∗R (13)

3. Experimental Section
2.4. TENG Power Transfer Theory for the Sliding Mode
3.1. Device Design
In previous work,[31] we introduced TENG Power Transfer The-
ory as a convenient and efficient method of analysing the power A lateral sliding mode TENG (LSTENG) was designed to evaluate
output behavior, impedance characteristics and the optimiza- the theoretical model. A nylon film (nylon 6,6, thickness 0.5 mm,
tion strategies of contact-separation mode TENGs. Using equa- dielectric constant 3.0, Goodfellow Cambridge Ltd. (UK)) was
tion (12), TENG Power Transfer Theory can be derived for the used as the positive triboelectric contact layer whereas a Poly-
sliding mode as follows. tetrafluoroethylene film (PTFE) (thickness 0.05 mm, dielectric
In equation (12), the load resistance R represents the external constant 2.0, Goodfellow Cambridge Ltd. (UK)) was used as the
impedance connected to the the system, which is defined as ZL negative triboelectric contact layer. Both these layers had similar
in the TENG Power Transfer equation. length and width of 5 cm x 5 cm. Copper (Cu) sheets (5 cm x 5 cm,
To simplify the derivation, we define: thickness 0.09 mm) were used as electrodes for both these layers.
( )∞ This LSTENG design was used for the DDEF model simulations
1 1 1 (as detailed in section 3.2), followed by experimental verifications
G (t ) = + ∫ f (x (t) , z) dz (14)
𝜋 𝜀1 𝜀2 0 (as detailed in section 3.3).
Next, a vertical contact-separation mode TENG (VCSTENG)
𝜎T ∞ with an identical device structure to the above LSTENG was de-
F1 ( t ) = ∫ f (x (t) , z) dz (15)
𝜋𝜀1 d1 signed. The theoretical and experimental output trends of the
LSTENG and the VCSTENG were compared with each other as
−𝜎T ∞ detailed in sections 3.2 and 3.3, to evaluate the relative output
F2 (t) = ∫ f (x (t) , z) dz (16)
𝜋𝜀2 d2 performance of the two operating modes.

Adv. Energy Mater. 2024, 2403853 2403853 (5 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

3.2. Theoretical Simulations amplitudes (0.5, 1, 3, 5, and 7.5mm), while the maximum contact
force was kept at 10 N. The motion cycle in each scenario started
The DDEF model derived in section 2 was used to simulate the from the fully contact/overlapped scenario and moved until their
short circuit current (ISC ) & charge (QSC ), open circuit voltage maximum displacement.
(VOC ), and power output behaviour of the LSTENG. Herein, the The electrical characterizations of TENG were conducted us-
triboelectric charge density (𝜎 T ) was approximated as 25 μC m2 ing a Keithley 6514B electrometer, where Qsc , Isc , Voc , and the
based on the experimental charge outputs of the LSTENG, using current through increasing loads were measured and the power
methods similar to those used in the previous work.[14,30,31,36,37] outputs were evaluated. Herein, the negative lead of the elec-
Next, this model was used to evaluate the impact of fundamental trometer was connected to the PTFE film, whereas the positive
TENG parameters (material, device and motion parameters) on lead was connected to the nylon film. The overlapped scenario
the LSTENG outputs, and to develop output optimization strate- of the TENG surfaces was considered as the reference measure-
gies. During this analysis, TENG impedance plots were used to ment point, and the measurement procedure followed identical
study both the electrical output and impedance variation trends. parameters to the theoretical simulations detailed in section 3.2,
For the convenience of the reader, an overview of the TENG resulting in similar output trends to the theoretical simulations.
impedance plot technique was provided in Note S3 (Support-
ing Information). It is important to note that, during the sim-
ulations, the starting position of the LSTENG movements repre- 4. Results and Discussion
sented their overlapped scenario, and the outputs were assessed When the PTFE and Nylon 6,6 triboelectric surfaces contact
with respect to this reference position. In other words, the initial each other, the PTFE surface is negatively charged and the Ny-
outputs at x(t) = 0 was set to zero. x(t) followed a sinusoidal move- lon 6,6 surface is positively charged.[42,43] As these charged sur-
ment profile with different amplitudes for different tests as spec- faces move relative to each other (either by sliding or contact-
ified in the results and discussion section, ranging from 0.1 mm separation), a potential difference is induced on their electrodes
up to 25 mm, and a starting phase angle of 270° (as specified in which drives the electrical outputs of the TENG.
sections 4.1 to 4.4). Hence, the triboelectric layers started their
motion from the fully overlapped scenario, travelled to one direc-
tion until their maximum amplitude, and returned to the over- 4.1. DDEF Model Simulations for LSTENG
lapped scenario to complete the motion cycle. Consequently, the
QSC and VOC outputs simulated in this study appear as positive First, our new theoretical model was used to simulate the out-
peaks, which start from zero, increase with x(t) up to its maxi- puts of the LSTENG described in section 3.1, under a sinusoidal
mum value, and returns to zero. Consequently, the current out- sliding movement of 1 Hz frequency, and amplitudes increasing
put, which represent the rate of change in charge, contain both from 0.5 to 25 mm (Figure 2a–f). The corresponding Qsc , Isc and
positive and negative peaks. Voc outputs are depicted in Figure 2a–c, respectively, all of which
Thereafter, a comparison was carried out between the simu- show a significant increase in the output levels against increasing
lations of DDEF model and classical theoretical model to deter- amplitude. To analyze further, the peak Qsc , Isc and Voc outputs
mine their relative accuracy. Finally, the DDEF theoretical plat- are shown in Figure 2d–f, respectively. When the amplitude was
form was applied to the sliding mode TENG and the contact- increased from 0.5 to 25 mm, the peak Qsc showed a correspond-
separation mode TENG ([14,30,31,36,37,39,40] ) to compare their relative ing increment from 1.13 to 61.99 nC, respectively, showing ap-
performances, considering identical LSTENG and VCSTENG proximately linear increment. A similar trend was observed for
structures described in section 3.1. peak Isc , with the corresponding outputs increasing from 3.64 to
195.38 nA, respectively. On the other hand, the DDEF model pre-
dicted that the Voc of the TENG would increase with a decreasing
3.3. Experimental Characterization rate, starting from 1.89 kV and systematically increasing up to
32.73 kV, respectively.
The LSTENG and VCSTENG described in section 3.1 were fabri- This output behavior can be explained using the DDEF theory.
cated and their outputs were evaluated to assess the DDEF model Here, the triboelectric surfaces of the LSTENG are initially fully
simulations described in section 3.2. The electrical characteriza- overlapped and in contact with each other. The sliding action of
tion of these TENG devices was conducted with a bespoke mo- LSTENG creates non-overlapped regions as well as an overlapped
tion control setup using methods similar to our previous stud- region (Figure S2, Supporting Information). In the overlapped re-
ies (Figure 1f,g).[30,31,36,37,41] Herein, the positive triboelectric layer gion, both positive and negative triboelectric charges are almost
(nylon) along with the Cu electrode were attached to the moving on the same plane, hence, this region does not contribute to the
stage of the setup. The negative triboelectric layer (PTFE) along potential difference between the electrodes. However, the non-
with the Cu electrode were attached to the stationary end which overlapped surface areas on the PTFE and Nylon 6,6 triboelec-
contained a load cell to measure the contact force applied during tric layers generate negative and positive electric fields, creating
the operation of the TENG. Two different configurations of the negative and positive potentials on respective electrodes. This po-
setup were developed to evaluate the performance of the TENG tential difference drives the outputs of the LSTENG. By increas-
in the contact-separation mode (Figure 1f) and the lateral sliding ing the amplitude of TENG layer movement, an increasing de-
mode (Figure 1g). The motion control system was used to pro- gree of non-overlapped surface area of the TENG is created, re-
vide relative movement between TENG contact surfaces using sulting in increased potential difference. This increases electri-
sinusoidal motion profiles with 1 Hz frequency and increasing cal outputs as demonstrated in Figure 2a–f, until they reach fully

Adv. Energy Mater. 2024, 2403853 2403853 (6 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 2. The DDEF model simulations depicting cyclic output response of the LSTENG showing a) QSC , b) ISC , and c) VOC plotted against time, and
the corresponding peak output values for d) QSC , e) ISC , and f) VOC , when subjected to a sinusoidal motion profile of 1 Hz frequency and increasing
amplitudes. The Comparison between the DDEF model simulations and experimental results for the LSTENG, depicting g) peak QSC and h) mean ISC ,
under sinusoidal movement of 1 Hz and increasing amplitudes. The average power output comparison depicting i) DDEF model predictions, and j)
experimental results for the LSTENG plotted against increasing load resistances, under sinusoidal movement of 1 Hz and increasing amplitudes. For
g), h) and j), error bars indicate the standard deviation of 10 readings.

Adv. Energy Mater. 2024, 2403853 2403853 (7 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

non-overlapped state (which occurs at 25 mm amplitude for this (Figure S5e–g, Supporting Information) were used to simulate
LSTENG). the output trends of the LSTENG (Note S5, Supporting Infor-
mation), to compare with the new DDEF model and the experi-
mental outputs. As shown in Figure S5a,b (Supporting Informa-
4.2. Experimental Verification of the Theoretical Model
tion) respectively, the peak Qsc and mean Isc outputs predicted
by the classical model show comparable trends to the DDEF
To assess the accuracy of simulations of the new DDEF model, a
model simulations. Furthermore, Figure S5e,f (Supporting In-
series of experiments were conducted as detailed in section 3.3,
formation) demonstrate the corresponding predictions from the
where the LSTENG was tested under a sinusoidal sliding move-
3D potential model, which suggest a relatively similar trend and
ment of 1 Hz frequency, 10 N maximum contact force, and am-
output values to the capacitor model and the DDEF model. The
plitudes increasing from 0.5 to 7.5 mm. The experimental results
simulated average power from the classical theoretical model
were compared with the corresponding simulation results as
(Figure S5d, Supporting Information) shows a different trend
demonstrated in Figure 2g–j. When the amplitude was increased
with an optimum load of 1 GΩ, which deviates from the DDEF
from 0.5 to 7.5 mm, the experimental Qsc outputs increased from
model and experimental outputs. On the other hand, the 3D po-
to 1.52 up to 20.23 nC, respectively, providing a close match
tential model predicts an average power output behavior similar
with the DDEF model simulations (Figure 2g). As shown in
to the DDEF model (Figure S5h, Supporting Information), with
Note S4 (Supporting Information), the experimental output trend
an optimum load of ≈100 GΩ. However, both the classical model
for peak Voc was comparable with DDEF model simulations,
(Figure S5c, Supporting Information) and the 3D potential model
showing an output increment against increasing amplitude, but
(Figure S5g, Supporting Information) predict that the Voc will in-
with a decreasing rate. We note that the mismatch between the
crease exponentially as the amplitude of the movement increases,
magnitude of the theoretical (Figure S4a, Supporting Informa-
which is a significant deviation from the DDEF model predic-
tion) and experimental (Figure S4b, Supporting Information) Voc
tions and the experimental output trends (Figure 2f; Figure S4,
values is a well-documented phenomenon that has been observed
Supporting Information). This provides strong evidence for the
for all previous TENG theoretical models,[14,24–26,31,33] which is at-
ability of the DDEF model to more accurately represent experi-
tributed to the high internal impedance of TENGs.[14,24,25,31,33]
mental LSTENG devices compared to these leading theoretical
The comparison between the DDEF model simulations and ex-
platforms.
perimental outcomes for the Isc is shown in Figure 2h. The mean
Isc of LSTENG corresponding to a half-cycle of its movement
(which captures positive peak of its alternating output current
4.3. Parametric Analysis and Device Optimization for Sliding
signal) was used for this analysis.[36,37] Herein, the experimen-
Mode TENG
tal Isc values match closely with the DDEF model simulations,
demonstrating a similar output trend. When the amplitude was
The new DDEF model was used to analyze the effect of funda-
increased from 0.5 up to 7.5 mm, the experimental Isc increased
mental TENG parameters (material, device, and motion param-
from 2.2 to ≈37 nA, whereas the DDEF model simulations sug-
eters) on electrical outputs of the LSTENG and to derive output
gested an Isc variation from 2.8 up to 41 nA, displaying a good
optimization strategies, as discussed in the following section. To
agreement.
understand the intrinsic impact of each fundamental TENG pa-
The DDEF model simulations and the experimental results for
rameter, when a particular parameter was systematically varied,
the output power of the LSTENG is demonstrated in Figure 2i,j,
all other parameters were kept constant at their base values de-
respectively. Herein, the average power output was plotted
scribed in section 3. The base motion profile used in this analy-
against increasing load resistances, under a sinusoidal move-
sis was a sinusoidal movement of 5mm amplitude and 1Hz fre-
ment of 1 Hz, and amplitudes increasing from 1 to 3 mm, 5 and
quency. For each parameter, the Qsc , Isc , Voc and maximum av-
7.5 mm. Looking into Figure 2i,j, both the DDEF model simu-
erage power were recorded (Figure 3–5). Next, TENG impedance
lations and experimental outputs demonstrate a similar trend,
plots along with the DDEF theory were used to analyze these out-
where the power increases gradually when the load was increased
put trends, providing detailed insights into the electrical output
from 100 to 10 GΩ. It is important to note that, within the
as well as the impedance characteristics of the LSTENG.
range of the loads experimentally tested in this study, the output
power kept increasing rather than showing the “increase-plateau-
decrease” pattern which is characteristic with TENGs. This is due
4.3.1. Material Parameters
to the high internal impedance of the LSTENG, causing the op-
timum load (load which provides maximum power) to be higher
As per the DDEF model, the triboelectric charge density and
than the loads tested here, resulting in the TENG not reaching
the dielectric constant are the two fundamental material prop-
its plateau power levels. A detailed analysis of this behavior is
erties that directly affect the output generation of a TENG.[14,30,31]
provided in section 4.3, under parametric analysis.
Figure 3a–f analyze the impact of increasing triboelectric charge
density (𝝈 T ) on the output generation of the LSTENG. When 𝝈 T
4.2.1. Comparison between the DDEF Model and Other Theoretical was increased from 1 to 10 μC m2 , 25, 50, and 100 μC m2 , the
Models Qsc increased from 0.49 to 4.89 nC, 12.23, 24.47, and 48.9 nC, re-
spectively, showing a linear increment (Figure 3a). As depicted
The classical theoretical model[25] (Figure S5a–d, Support- in Figure 3b,c, Voc and the Isc demonstrated a similar trend,
ing Information) and the 3D electrical potential model[26] showing a linear increment against increasing charge density.

Adv. Energy Mater. 2024, 2403853 2403853 (8 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 3. Theoretical analysis of the material parameters of LSTENG via DDEF Model. a–f) DDEF model simulations for the LSTENG outputs against
increasing 𝝈 T , showing a) QSC , b) VOC , c) ISC , d) Average power output. TENG impedance plots for e) 𝜎 T = 1 μC m2 , and f) 𝜎 T = 100 μC m2 . g–l) DDEF
model simulations for the LSTENG outputs against increasing dielectric constant of TENG layer 1 (𝜖 1 ), showing g) QSC , h) VOC , i) ISC , j) Average power
output. TENG impedance plots for e) 𝜖 1 =1, and f) 𝜖 1 =30.

Adv. Energy Mater. 2024, 2403853 2403853 (9 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 4. Theoretical analysis of the structural parameters of LSTENG via DDEF Model. a–f) DDEF model simulations for the LSTENG outputs against
increasing surface area of triboelectric layers, showing a) QSC , b) VOC , c) ISC , d) Average power output. TENG impedance plots for e) 10 mm × 10 mm,
and f) 500 mm × 500 mm surface areas. g–l) DDEF model simulations for the LSTENG outputs against increasing thickness of TENG layer 1 (d1 ),
showing g) QSC , h) VOC , i) ISC , j) Average power output. TENG impedance plots for e) d1 = 0.5 μm, and f) d1 =5000 μm.

Adv. Energy Mater. 2024, 2403853 2403853 (10 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 5. Theoretical analysis of the motion parameters of LSTENG via DDEF Model. a–f) DDEF model simulations for the LSTENG outputs against
increasing frequency (n), showing a) QSC , b) VOC , c) ISC , d) Average power output. TENG impedance plots for frequencies of e) 0.1 Hz, and f)100 Hz.
g–l) DDEF model simulations for the LSTENG outputs against increasing amplitude of its motion, showing g) QSC , h) VOC , i) ISC , j) Average power
output. TENG impedance plots for amplitudes of e) 0.1 mm, and f) 25 mm.

Adv. Energy Mater. 2024, 2403853 2403853 (11 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 3d shows the variation of the mean power against increas- dielectric constant. Both Isc (Figure 3i) and Qsc (Figure 3g) rep-
ing loads, when 𝝈 T was increased. For all 𝝈 T , the maximum aver- resent the short circuit condition, i.e., the minimum load sce-
age power was observed via the 100 GΩ load. Further, the power nario, therefore, these parameters do not show significant varia-
output demonstrated a quadratic relationship to 𝝈 T . For example, tions against increasing dielectric constant. On the other hand,
considering 100 GΩ, when 𝝈 T was increased from 1 to 10 μC m2 , the Voc (Figure 3h) represents the open circuit condition, i.e., the
25, 50, and 100 μC m2 , the average power increased from 0.089 maximum load scenario, therefore, it varies notably when the di-
to 8.91 μW, 55.7, 222.9, and 891.8 μW, respectively. electric constant is increased. Hence, the output trends shown in
To further analyse this scenario, TENG impedance plots cor- Figure 3g–i agree with our previous findings.
responding to 𝝈 T = 1 μC m2 and 𝝈 T = 100 μC m2 are shown in Looking at average power (Figure 3j), the overall maximum
Figure 3e,f, respectively. As detailed in Note S3 (Supporting In- power (maximum power among all load values) was observed at
formation), the impedance plots provide a convenient pathway to 𝝐 1 = 1, with 87.5 μW via a 1 TΩ load. For 𝝐 1 values of 3, 6, 10,
understand the outputs of LSTENGs. Here, the electrical output and 30, the maximum power was observed through 100 GΩ, with
generation of the LSTENG is represented by the current IS (t) (red gradually decreasing outputs of 55.7, 50.8, 48.1, and 44.9 μW, re-
line, linked to left-Y axis), and the impedance behaviour is repre- spectively. The reasons behind this power output behaviour can
sented by the impedance ratios linked to different loads (coloured be analysed using the TENG impedance plots Figure 3k,l, which
horizontal lines corresponding to each load, linked to right-Y represent 𝝐 1 = 1 and 𝝐 1 = 30 scenarios. Comparing Figure 3k,l,
axis). From Figure 3e,f, the current increases showing a linear the current shows only a marginal increment when 𝝐 1 is in-
proportionality to the triboelectric charge density (IS (t)∝𝝈 T ). This creased. However, looking at the impedance behaviour, the opti-
is consistent with equations (6) to (10) in section 2 that demon- mum load reduces significantly from 1 TΩ for 𝝐 1 = 1 (Figure 3k)
strate the linear proportionality of Voc , Isc and Qsc against 𝝈 T , to 100 GΩ for 𝝐 1 = 30 (Figure 3l), which is the more significant
thus, explaining the output trends depicted in Figure 3a–c. Sec- factor that dictate the power generation of the LSTENG. What this
ond, looking at the impedance behaviour, the optimum load of means is that when 𝝐 1 is increased, the power is generated by a
the LSTENG did not show any variation (100 GΩ), indicating relatively similar current but through a significantly lower opti-
that the impedance characteristics of the LSTENG was not af- mum load, causing a reduction in the maximum power output
fected by 𝝈 T . According to equation (20), this leads to a quadratic of the LSTENG (Figure 3j).
relationship between the output power and 𝝈 T (POut (t)∝𝝈 T 2 ).[31]
Looking at this in simple terms, triboelectric charges create elec-
tric fields that drive the operation of the TENG. A higher tribo- 4.3.2. Structural Parameters
electric charge density creates stronger electric fields that can in-
duce higher electrical outputs, resulting in the increasing output The triboelectric contact surface area (determined by L, W) and
trends depicted in Figure 3a–f. the thickness of TENG dielectric layers (d1 , d2 ) are the main struc-
Next, the LSTENG outputs against dielectric constant is sim- tural parameters that directly affect its outputs, according to the
ulated in Figure 3g–l. Here, the dielectric constant of the TENG DDEF model. Figure 4a–f demonstrate the DDEF model simula-
layer 1 (𝝐 1 ) was varied from 1 up to 3, 6, 10, and 30, whereas tions for the outputs of the LSTENG against increasing surface
the dielectric constant of the TENG layer 2 (𝝐 2 ) was maintained area. Herein, for the ease of calculations, L = W case (square
constant at 2.0. The resultant Qsc , Voc and the Isc outputs are TENG contact surfaces) was analyzed with 10 mm × 10 mm,
depicted in Figure 3g–i, respectively. Both Qsc (Figure 3g) and 50 mm × 50 mm, 100 mm × 100 mm, and 500 mm × 500 mm sur-
Isc (Figure 3i) showed initial minor increments when 𝝐 1 was in- face areas considered. Correspondingly, the peak Qsc increased
creased, but their outputs remained almost constant afterwards. from 2.4 to 12.2 nC, 24.6 and 123.5 nC (Figure 4a), and the
For example, Qsc corresponding to 𝝐 1 = 1 was 12.09 nC, which peak Isc increased from 7.6 to 38.8 nA, 77.9 and 390 nA, respec-
increased to 12.23 nC when 𝝐 1 = 3. When 𝝐 1 was further in- tively (Figure 4c). The Voc showed a similar increment from 6.34
creased to 6, 10, and 30, the corresponding Qsc values increased to 12.12 kV, 14.7 and 20.7 kV, respectively (Figure 4b). Further,
to 12.31, 12.36, and 12.41 nC. Isc outputs demonstrated a similar the maximum average power (Figure 4d) increased from 11.8 to
trend, with 38.6 nA when 𝝐 1 = 1, which increased to 38.9 nA at 55.7 μW, 167.5 and 701.7 μW respectively. Meanwhile, optimum
𝝐 1 = 3, and eventually up to 39.1 nA when 𝝐 1 was increased to 30. load for the LSTENGs shifted gradually from ≈1 TΩ for 10 cm
On the other hand, as depicted in Figure 3h, Voc significantly re- × 10 cm LSTENG down to 100 GΩ for the 500 cm × 500 cm
duced against increasing 𝝐 1 , showing an output of 21.57 kV at 𝝐 1 LSTENG, showing significant reduction in optimum load against
= 1, which reduced to 12.12, 9.75, 8.81, and 7.86 kV, when 𝝐 1 was increasing surface area.
increased to 3, 6, 10, and 30, respectively. The difference in the TENG impedance plots corresponding to 10 mm × 10 mm
response against increasing dielectric constant of the TENG be- and 500 mm × 500 mm cases are shown in Figure 4e,f respec-
tween short circuit outputs (Isc , Qsc ), and the open circuit outputs tively, which can shed light on these output trends. Comparing
(Voc ) can be explained using our previous study.[40] In our pre- Figure 4e,f, as the LSTENG surface area was increased, the cur-
vious work,[40] we conclusively demonstrated (both theoretically rent increased ≈7 to 390 nA. On the other hand, the optimum
and experimentally), that the variation of TENG electrical outputs load is reduced from close to 1 TΩ toward 100 GΩ, showing the
against increasing dielectric constant depends heavily on its ex- opposite trend. However, from equation 20, we know that the
ternal load. As per this study, under low external loads, both cur- power output of LSTENG is quadratically proportional to cur-
rent and voltage do not change significantly when the dielectric rent, and linearly proportional to the load. Consequently, the in-
constant of the TENG layer is increased, whereas at high external crement in current becomes the dominant factor here, result-
loads, both these outputs are significantly affected by increasing ing in increased maximum power output from the LSTENG.

Adv. Energy Mater. 2024, 2403853 2403853 (12 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

In simple terms, larger triboelectric surface areas result in a cies is 5 mm, so their maximum displacement is the same. This
higher number of triboelectric charges contributing to create the results in the peak Qsc and peak Voc remining the same for all
LSTENG outputs, resulting in the increased output levels shown frequencies (n), as both these parameters are independent of the
in Figure 4a–d. rate of movement (Figure 5a,b). This is also evident from equa-
Figure 4g–l demonstrate the DDEF simulations for VCSTENG tions (3) – (5) (for Voc ) and equations (6) – (8) (for Qsc ), which
outputs when its dielectric layer thickness was increased. Herein, elaborate the dependence of the outputs on the displacement x(t).
the thickness d1 was varied systematically from 0.5 μm up to On the other hand, peak Isc increased significantly from 3.8 to
50 000 μm, while maintaining d2 at 50 μm. When d1 was in- 38.9 nA, 194.3, 388.5, and 3885 nA, when the n was increased
creased from 0.5 to 5 μm, 50, 500, 5000, and 50 000 μm, the from 0.1 to 1 Hz, 5, 10, and 100 Hz, respectively (Figure 5c). Isc
peak Qsc varied from 12.47 to 12.46 nC, 12.44, 12.23, 10.76, and represents the rate of the flow of charge, which is directly depen-
8.19 nC (Figure 4g), the peak Voc varied from 12.35 to 12.34 kV, dent on the rate of movement of TENG layers (in this case n). As
12.32, 12.11, 10.66, and 8.11 kV (Figure 4h), and the peak Isc var- the amplitude is 5 mm for all frequencies, the maximum poten-
ied from 39.23 to 39.21 nA, 39.20, 38.85, 34.49, and 25.88 nA tial different and hence, the maximum number of charges that
(Figure 4i), respectively. Therefore, at lower d1 range (0.5 to flow between the electrodes is constant. However, at higher n,
500 μm), these outputs show minor reductions against increas- the potential differences are induced at a faster rate, hence, the
ing d1 , whereas when d1 reach significantly higher values (5000 to charges are made to travel faster between the two electrodes. This
50 000 μm), the output reductions become more visible. Consid- results in the increased Isc when n is increased (Figure 5c). The
ering the output power (Figure 4j), when d1 was increased from dependence of Isc on the frequency is also evident from equa-
0.5 to 5 μm, 50, 500, 5000, and 50 000 μm, the maximum average tion (10), which defines Isc as the time derivative between the
power of the LSTENG reduced from 57.4 to 57.3 μW, 57.2, 55.7, product of output charge 𝝈 U (t) and the area A(t). Consequently,
43.2, and to 24.8 μW. Interestingly, the optimum load of the slid- increasing n results in increased Isc .
ing mode TENG shows minimal variation and remains close to The average power output for the sliding mode TENG is shown
100 GΩ throughout the d1 values tested in this study. in Figure 5d. When n is increased from 0.1 to 1 Hz, 5, 10, and
The TENG impedance plots in Figure 4k,l, which represent d1 100 Hz, the average power increased accordingly, where the max-
= 0.5 μm and d1 = 50 000 μm cases, provide insights into this imum average power reached 5.5 to 55.7 μW, 266.4, 557.4, and
output behavior. In Figure 4k, the current reaches close to 40 nA, 5580 μW, respectively.
which reduces to ≈26 nA in Figure 4l, showing a reduction in TENG impedance plots for n = 0.1 Hz and n = 100 Hz are
current against increasing thickness. Further, the optimum load shown in Figure 5e,f, respectively which shows the reasons be-
does not significantly change. So, this results in the reduction of hind this power output behavior. Comparing Figure 5e,f, when
TENG outputs. the frequency was increased from 0.1 to 100 Hz, the current in-
The LSTENG output behavior against increasing thickness can creased proportionally from 3.8 to 3885 nA. On the other hand,
be analyzed using the DDEF theory[31] as follows. When d1 is the optimum load showed a proportional decrease, from close to
increased, the electric fields originating from triboelectric con- 1 TΩ down to ≈1 GΩ. Since the output power is linearly propor-
tact surfaces need to propagate for a longer distance through the tional to the load and quadratically proportional to the current,
dielectric layer to reach the electrodes. As the electric fields are the increment in the current becomes the dominant factor, and
distance-dependent, travelling larger distances would cause their the power output shows a linear proportionality to n.
magnitude to reduce, thus, inducing lower electrical outputs for Finally, LSTENG outputs trends against increasing motion
larger d1 values as observed in Figure 4g–l. amplitude are depicted in Figure 5g–l. When the amplitude was
increased from 0.1 to 1 mm, 5, 12.5, 20, and 25 mm, peak Qsc in-
creased from 0.2 to 2.3 nC, 12.2, 30.8, 49.6, and 61.9 nC, respec-
4.3.3. Motion Parameters tively (Figure 5g). The peak Voc increased from 0.43 to 3.4 kV,
12.1, 22.2, 29.1, and 32.7 kV (Figure 5h), whereas the peak Isc
The key motion parameters, as per the DDEF theory, include increased from 0.6 to 7.5 nA, 38.8, 97.5, 156.3, and 195.4 nA,
the motion rate (represented by the frequency (n) for sine move- respectively (Figure 5i). As discussed earlier, the electrical out-
ment), the motion amplitude, as well as their minimum separa- puts of the LSTENG are driven by the triboelectric charges on its
tion (contact/non-contact mode). The vast majority of LSTENG non-overlapping surface area. When the amplitude is increased,
operate based on the contact-sliding mode, where their min- increasingly larger proportion of the triboelectric charges con-
imum separation between sliding surfaces is practically zero, tribute toward the potential difference between the electrodes,
therefore, in this analysis, we will focus only on the motion fre- which can drive higher number of output charges. This results
quency and amplitude. in the increased electrical outputs observed in Figure 5g–i.
The variations in Qsc and Voc when n was varied from 0.1 to The average power outputs of the LSTENG are shown in
1 Hz, 5, 10, and 100 Hz are shown in Figure 5a,b, respectively. Figure 5j. When the amplitude was increased from 0.1 to 1 mm,
Here, the number of cycles for both these parameters change ac- 5, 12.5, 20, and 25 mm, the maximum average power increased
cording to the frequency, however, their peak outputs remained from 0.02 to 2.3 μW, 55.7, 305.7, 698.9, and to 1020 μW, re-
constant with Qsc of 12.23 nC and Voc of 12.1 kV. As per the DDEF spectively. Again, this behavior can be clarified using TENG
theory, both Qsc and peak Voc depend on the potential level differ- impedance plots depicted in Figure 5k,l, respectively. When the
ences between the electrodes, which in turn, depend on the rel- amplitude is increased from 0.1 mm (Figure 5k) up to 25 mm
ative displacement profile between the triboelectric contact sur- (Figure 5l), the current increased significantly from 0.6 nA up to
faces. Here, the amplitude of the movement for all the frequen- ≈196 nA. The optimum load, on the other hand, reduced from

Adv. Energy Mater. 2024, 2403853 2403853 (13 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

close to 1 TΩ to ≈100 GΩ. Here, the impact of the increment TENG layers come to a fully non-overlapped configuration, thus,
in current is more dominant, hence the output power of the the LSTENG reaches its maximum Voc at much lower amplitudes
LSTENG increases. compared to the VCSTENG (Figure 6a).
Following the above DDEF model analysis, a detailed experi- Figure 6b demonstrates the DDEF model simulations for peak
mental validation of these results is provided in Note S7 (Sup- Qsc . When the amplitude is increased, the peak Qsc of LSTENG
porting Information), which validate the proposed trends. demonstrates a corresponding liner increment and reaches an
Typically, most sliding mode TENG devices operate in the overall peak of ≈61.9 nC at 25 mm amplitude. The peak Qsc of the
contact-sliding state (where there is contact between the tribo- VCSTENG increases more rapidly and reaches ≈60 nC at ≈5 mm
electric surfaces during some parts of their relative motion cycle. amplitude, eventually reaching ≈61.7 nC. The experimental re-
However, recent developments have demonstrated the capability sults shown in Figure 2g and Figure S6a (Supporting Informa-
of sliding mode TENG operating in non-contact mode.[44] Sup- tion) agree with these predictions, showing a gradual increase of
porting Note 8 presents a detailed analysis of output generation Qsc for LSTENG, and a rapid increase for VCSTENG.
of the LSTENG operating in non-contact mode, for a range of gap This peak Qsc behavior is attributed to the effect of triboelec-
distance between the triboelectric surfaces. tric contact area. Both the VCSTENG and LSTENG contain iden-
tical triboelectric charge densities and triboelectric contact areas.
For the VCSTENG, its’ total surface area contributes to drive the
4.4. Sliding Mode versus Contact-Separation Mode charges between the electrodes throughout its entire motion cy-
cle, which helps to generate higher peak Qsc even at low motion
In our recent work,[30,31,38] we demonstrated the applicability of amplitudes. On the contrary, the Qsc of LSTENG is driven only by
the DDEF model for all TENG types that depend on the contact- its non-overlapped surface area, which is significantly smaller es-
separation movements. In the present work, a new DDEF model pecially during low motion amplitudes. Therefore, the peak Qsc
was derived to represent sliding mode TENGs. Both these models of LSTENG, especially at low amplitudes, are lower compared to
are derived based on the same basic equation (DDEF equation) the VCSTENG as shown in Figure 6b.
with high consistency, therefore, this facilitates the unique oppor- The mean Isc of the VCSTENG and LSTENG are depicted
tunity to conduct a comparison between the contact-separation in Figure 6c. Mean Isc is proportional to the peak Qsc of the
and sliding mode TENGs. Therefore, herein, the DDEF model TENG,[31] therefore, the Isc outputs of Figure 6c shows a similar
is used to simulate the outputs of the LSTENG and a VCSTENG trend to those of Figure 6b. This is also evident from the experi-
with identical parameters, to compare their output behaviour. An mental outputs for the LSTENG (Figure 2h) and the VCSTENG
experimental analysis is also done to verify these predictions. (Figure S6b, Supporting Information). On the other hand, peak
Figure 6a–f show the output trends of the VCSTENG and Isc trends shown in Figure 6d predict significantly higher outputs
LSTENG under a sinusoidal motion profile. First, the peak Voc for for the VCSTENG compared to the LSTENG. This means that the
the VCSTENG and LSTENG are depicted in Figure 6a. Consider- VCSTENG produces sharper (higher) current peaks compared to
ing the LSTENG, when amplitude is increased, Voc increases ac- the LSTENG. In our previous work,[31,36] we analyzed this output
cordingly and reaches an overall maximum of 32.7 kV at 25 mm generation behavior of the VCSTENG which is attributed to the
amplitude and plateaus afterwards. The Voc of VCSTENG also in- sharp gradient in their distance-dependent electric fields, allow-
creases but at a lower rate. Eventually, the VCSTENG reaches a ing them to generate high instantaneous current peaks even at
similar overall peak Voc of ≈32.7 kV, but, at a larger amplitude. low motion amplitudes. On the contrary, Isc of the LSTENG is
This Voc behaviour can be understood as follows. dictated by its non-overlapped surface area, hence, systematically
In both VCSTENG and LSTENG, Voc is affected by the over- increasing amplitude of the LSTENG causes gradually increasing
all electric field originating from the positive and negative tribo- Isc , resulting in lower peak Isc values compared to the VCSTENG
electric charges. The Voc of a TENG increases when each of the (Figure 6d).
two electrodes increase their respective potentials. However, the Furthermore, the mean power and peak power of these TENGs
electric field from the opposite TENG layer causes this potential are displayed in Figure 6e,f, respectively. In both cases, optimum
to decrease. Therefore, for Voc to reach its highest levels, each load of the VCSTENG is between 500 MΩ and 1 GΩ whereas
TENG layer needs to be free from the influence of electric field the LSTENG shows an optimum load of ≈100 GΩ. As a general
from its opposite TENG layer. trend, the VCSTENG provides higher power outputs via smaller
Now, consider the two TENG structures. In the VCSTENG, loads (typically <1 GΩ), whereas the LSTENG provides higher
the triboelectric contact layers are always parallel and overlapped. power outputs via larger loads (>1 GΩ). Furthermore, at lower
Hence, a significant separation between the TENG layers is re- amplitudes (e.g.,: 1 mm, 3 mm), the VCSTENG provides higher
quired (theoretically infinite/practically a large separation com- overall maximum power, whereas the LSTENG provides higher
pared to its dimensions) so that each electrode can be free from overall maximum power at higher amplitudes (e.g.,: 5 mm and
the influence of electric field from the opposite TENG layer.[14,31] above). A similar pattern for output power is observed through
Therefore, the maximum Voc for the VCSTENG can only be the experimental outputs for LSTENG (Figure 2j) and VCSTENG
achieved at relatively high amplitudes. On the other hand, the (Figure S6c, Supporting Information).
Voc of LSTENG is generated by creating a non-overlapped region For TENGs, it is well established that both current and power
of the triboelectric surfaces. In the non-overlapped region, the generation are affected significantly by their motion profile. Us-
electric fields from opposite TENG surfaces cannot significantly ing a constant velocity relative motion to analyze the output be-
impact the potential of the electrodes, enabling them to reach havior of the TENGs has previously been demonstrated as a
the maximum potential. When the amplitude is 25 mm, both the simpler approach to understand the nature of these devices,[31]

Adv. Energy Mater. 2024, 2403853 2403853 (14 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Figure 6. a–f) Comparison between the LSTENG (marked as SL) and the VCSTENG (marked as CS), subjected to a sinusoidal relative motion profile of
1 Hz frequency and increasing amplitudes, depicting a) peak Voc , b) peak Qsc , c) mean Isc , d) peak Isc , e) average power at increasing loads, and f) peak
power at increasing loads. h–j) Comparison between the LSTENG (marked as SL) and the VCSTENG (marked as CS), subjected to a range of constant
velocity motion profiles, depicting a) peak Isc , b) average power, and c) peak power outputs. Herein, the motion profiles represent relative TENG layer
movement from 0 to 50 mm.

compared to a sinusoidal motion which has a varying velocity area which increases gradually under a constant velocity profile,
profile. To this end, Figure 6g–i demonstrates the DDEF model resulting in lower Isc peaks (Figure 6g). On the other hand, the
simulations for the current and power outputs of the VCSTENG VCSTENG, which operates in a parallel and overlapped configu-
and LSTENG, subjected to constant velocity movements ranging ration, can create higher instantaneous Isc peaks (similar to sinu-
from 0.0001 to 100 ms−1 . In these simulations, the TENG layers soidal scenario discussed above), resulting in increased peak Isc
move starting from 0 up to 50 mm at each selected velocity. (Figure 6g).
Figure 6g compares the peak Isc between the VCSTENG and The average power and the peak power of the two TENG work-
LSTENG. Both the devices demonstrate an increase in peak Isc ing modes are shown in Figure 6h,i, respectively. Similar to the
against increasing velocity. However, the peak Isc of the VC- sinusoidal motion, at lower loads, the VCSTENG provides signif-
STENG is significantly higher than the comparable peak Isc of icantly higher power outputs for both average and peak power.
the LSTENG, considering the same velocity. As discussed above, At relatively large load values, the LSTENG provides marginally
the Isc of the LSTENG is driven by the non-overlapped surface higher power outputs compared to the VCSTENG. However, the

Adv. Energy Mater. 2024, 2403853 2403853 (15 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

overall maximum power for each of the velocities are higher for mental theoretical platform, this work provides in-depth insights
the VCSTENG compared to the equivalent LSTENG, which is a into the impact of fundamental TENG parameters on the output
different observation to the sinusoidal motion scenario. generation of sliding mode TENG, for the first time, covering the
For example, considering the average power (Figure 6h), the entire spectrum of material, device and motion parameters both
overall maximum power at 0.0001 ms−1 for the VCSTENG is theoretically and experimentally. This analysis discovers numer-
878 nW (via 100 TΩ) which reduces to 772 nW (via 100 TΩ) for ous previously unreported output trends and provides pathways
the LSTENG, and when the velocity is increased to 100 ms−1 , the to optimize these parameters to engineer sliding mode TENG
VCSTENG produces 0.87 W (via 100 MΩ) and the LSTENG pro- designs. Moreover, with the derivation of this model, TENG de-
duces 0.77 W (via 100 MΩ). Also, there is a noticeable difference signs which depend on contact-separation mode as well as sliding
in the output trends between the VCSTENG and LSTENG. For mode can now be represented using the same theoretical plat-
example, at 0.0001 ms−1 , the average power for the VCSTENG ap- form. This means that the DDEF model can now encompass all
pear to maximize between 10 to 100 TΩ, and for the LSTENG this possible TENG configurations in 3D space (x, y, and z axes), sig-
falls between 100–1000 TΩ. Considering 100 ms−1 case, the aver- nifying the unification of the TENG theory. Thus, the outcome
age power for the VCSTENG maximizes between 10 to 100 MΩ, of this study leads to the unique advantage of being able to use
and for the LSTENG this falls between 100–1000 MΩ. a single theoretical platform to compare different TENG types
Considering the peak power (Figure 6i), the VCSTENG pro- and to optimize them for future applications. These benefits will
vides significantly higher overall maximum power at notably simplify and significantly improve the design and optimization
lower optimum load values compared to the LSTENG. For ex- of real-life TENGs.
ample, considering the peak power at 0.0001 ms−1 (Figure 6i), While the DDEF model for the sliding mode TENG pro-
the overall maximum power of the VCSTENG is 2.08 μW via vides unique pathways to understand, simulate and optimize
10 TΩ whereas the LSTENG produces overall maximum power the behavior of these TENG designs, there are several chal-
of 1.04 μW via 100 TΩ, and considering 100 ms−1 , the VCSTENG lenges that this model needs to overcome to further improve
produces overall maximum power of 2.08 W via 10 MΩ and the its applicability to real world applications. First, the DDEF
corresponding value for the LSTENG is 1.04 W via 100 MΩ. model assumes that the triboelectric charges are evenly dis-
Consequently, under constant velocity motion profiles, the VC- tributed with opposite polarities in corresponding triboelec-
STENG provides relatively higher power outputs via lower opti- tric contact surfaces. Overall, this can be considered a fair as-
mum load values, with respect to an identical LSTENG. sumption at macroscopic scale in which most TENG devices
operate.[10,14,25,30,33,45,46] However, recent studies have shown that
5. Conclusions and Perspectives both triboelectric surfaces contain positive and negative charges,
distributed in nanoscopic mosaic-like regions.[19,47] Second, the
A novel comprehensive theoretical model for sliding mode TENG DDEF model assumes that the triboelectric charge density is at
was developed in this work using the DDEF theory. The new its saturated value, and constant throughout the operation of
DDEF model was used to simulate the output trends of a the TENG. While this can be considered a reasonable assump-
LSTENG describing current, charge, voltage and power output tion practically under typical environmental conditions, there
trends as well as its impedance behavior. The theoretical pre- are circumstances which challenge this assumption. According
dictions were verified using experimental results, proving the to empirical studies, triboelectric charging can be dynamic, and
accuracy and the effectiveness of this model. Consequently, a there can be time dependent[48,49] and condition dependent(e.g.,:
comparison was carried out between this model and the classi- temperature,[3,17,50] humidity)[51,52] variations. These variations
cal theoretical model for siding mode TENG, which proved the can currently be evaluated only using empirical methods, and fu-
better accuracy of the new DDEF model, and its improved capa- ture theoretical approaches are needed to accurately model these
bility to represent real-life sliding mode TENG. Next, a detailed variations, which can potentially be incorporated into the DDEF
analysis on the effect of material, device and motion parame- model.
ters on the output generation of the LSTENG was conducted us-
ing the DDEF model. Herein, the TENG power transfer theory
and TENG impedance plots were used to understand the fun-
damental factors which affect these output generation patterns, Supporting Information
and to derive a series of optimization strategies. Consequently,
Supporting Information is available from the Wiley Online Library or from
an in-depth analysis was carried out comparing the VCSTENG the author.
and LSTENG to evaluate their output generation capabilities,
impedance behavior and the corresponding governing factors.
The outcomes of this work add significant advancements and
insights into existing TENG theory. This marks the first time Acknowledgements
that the DDEF theory was successfully used to derive a theoreti-
cal model for the sliding mode TENG, surpassing the challenges R.L.B. acknowledges the support for the Ph.D. studentship from the Wolf-
which restricted its use only to the contact-separation mode. This son School of Mechanical, Electrical and Manufacturing Engineering,
Loughborough University. R.D.I.G.D. acknowledges the support from the
new theory provides improved accuracy along with simplicity Royal Academy of Engineering under the Research Fellowship scheme
compared to many existing theoretical approaches, and it can rep- (RF∖202021∖20∖252), the Royal Society NSFC International Exchanges
resent practical TENG designs with higher consistency, as evi- Cost Share grant (IEC∖NSFC∖223116), and the Loughborough University
dent from the theoretical and experimental results. As a funda- Institute of Advanced Studies (IAS).

Adv. Energy Mater. 2024, 2403853 2403853 (16 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH
 16146840, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aenm.202403853 by Lancaster University, Wiley Online Library on [20/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.advenergymat.de

Conflict of Interest [20] X.-W. Li, Y.-H. Wu, D.-L. Wen, Y. Chen, X.-S. Zhang, Appl. Phys. Lett.
2022, 121, 123904.
The authors declare no conflict of interest. [21] D. Zhao, X. Yu, J. Wang, Q. Gao, Z. Wang, T. Cheng, Z. L. Wang, Energy
Environ. Sci. 2022, 15, 3901.
[22] X. Guo, J. Shao, M. Willatzen, Y. Yang, Z. L. Wang, Nano Energy 2022,
92, 106762.
Data Availability Statement
[23] D. G. Dassanayaka, R. L. Bulathsinghala, B. T. R. Bulathsinghala, G.
The data that support the findings of this study are available from the Lanarolle, N. D. Wanasekara, R. Dharmasena, Nano Energy 2024,
corresponding author upon reasonable request. 109688.
[24] S. Niu, Z. L. Wang, Nano Energy 2015, 14, 161.
[25] S. Niu, Y. Liu, S. Wang, L. Lin, Y. S. Zhou, Y. Hu, Z. L. Wang, Adv.
Mater. 2013, 25, 6184.
Keywords [26] J. Shao, D. Liu, M. Willatzen, Z. L. Wang, Appl. Phys. Rev. 2020, 7.
DDEF model, energy harvesting, sliding mode TENG, TENG optimization, [27] J. Chen, J. Wang, W. Xuan, S. Dong, J. Luo, Sensors 2020, 20, 4838.
theoretical modelling, triboelectric nanogenerator [28] S. Niu, S. Wang, L. Lin, Y. Liu, Y. S. Zhou, Y. Hu, Z. L. Wang, Energy
Environ. Sci. 2013, 6, 3576.
Received: August 26, 2024 [29] S. Niu, Y. Liu, S. Wang, L. Lin, Y. S. Zhou, Y. Hu, Z. L. Wang, Adv. Funct.
Revised: November 18, 2024 Mater. 2014, 24, 3332.
Published online: [30] R. D. I. G. Dharmasena, K. D. G. I. Jayawardena, C. A. Mills, R. A.
Dorey, S. R. P. Silva, Nano Energy 2018, 48, 391.
[31] R. D. I. G. Dharmasena, J. H. B. Deane, S. R. P. Silva, Adv. Energy
Mater. 2018, 1802190.
[32] C. Callaty, I. Gonçalves, C. Rodrigues, J. Ventura, Curr. Appl. Phys.
[1] P. Basset, S. P. Beeby, C. Bowen, Z. J. Chew, A. Delbani, R. D. I. G.
2023, 50, 100.
Dharmasena, B. Dudem, F. R. Fan, D. Galayko, H. Guo, APL Mater.
[33] J. Shao, M. Willatzen, Z. L. Wang, J. Appl. Phys. 2020, 128, 111101.
2022, 10, 109201.
[34] J. Shao, M. Willatzen, Y. Shi, Z. L. Wang, Nano Energy 2019, 60,
[2] D. G. Dassanayaka, T. M. Alves, N. D. Wanasekara, I. G. Dharmasena,
630.
J. Ventura, Adv. Funct. Mater. 2022, 32, 2205438.
[35] J. Shao, T. Jiang, Z. Wang, Sci. China Technol. Sci. 2020, 63, 1087.
[3] C. Wu, A. C. Wang, W. Ding, H. Guo, Z. L. Wang, Adv. Energy Mater.
[36] R. Dharmasena, Nano Energy 2020, 76, 105045.
2019, 9, 1802906.
[37] R. Dharmasena, K. G. U. Wijayantha, Nano Energy 2021, 90, 106511.
[4] J. Li, C. Wu, I. Dharmasena, X. Ni, Z. Wang, H. Shen, S.-L. Huang, W.
[38] R. D. I. G. Dharmasena, H. M. Cronin, R. A. Dorey, S. R. P. Silva, Nano
Ding, Intell. Converg. Networks 2020, 1, 115.
Energy 2020, 75, 104887.
[5] R. D. I. G. Dharmasena, Triboelectric Self-Powered Energy Systems.
[39] J. H. B. Deane, R. D. I. G. Dharmasena, G. Gentile, Nano Energy 2018,
[Doctoral Dissertation], University of Surrey, 2019.
54, 39.
[6] W. Paosangthong, R. Torah, S. Beeby, Nano Energy 2019, 55, 401.
[40] R. L. Bulathsinghala, A. Ravichandran, H. Zhao, W. Ding, R.
[7] K. R. S. D. Gunawardhana, N. D. Wanasekara, R. D. I. G. Dharmasena,
Dharmasena, Nano Energy 2024, 123, 109383.
iScience 2020, 101360.
[41] K. R. S. Gunawardhana, N. D. Wanasekara, K. G. Wijayantha, R. D. I.
[8] R. D. I. G. Dharmasena, K. D. G. I. Jayawardena, Z. Saadi, X. Yao, R.
Dharmasena, ACS Appl. Electron. Mater. 2022, 4, 678.
M. I. Bandara, Y. Zhao, S. R. P. Silva, Proc. IEEE 2019, 107, 2118.
[42] A. F. Diaz, R. M. Felix-Navarro, J. Electrostat. 2004, 62, 277.
[9] C. Rodrigues, D. Nunes, D. Clemente, N. Mathias, J. M. Correia, P.
[43] H. Zou, Y. Zhang, L. Guo, P. Wang, X. He, G. Dai, H. Zheng, C. Chen,
Rosa-Santos, F. Taveira-Pinto, T. Morais, A. Pereira, J. Ventura, Energy
A. C. Wang, C. Xu, Z. L. Wang, Nat. Commun. 2019, 10, 1427.
Environ. Sci. 2020, 13, 2657.
[44] L. Long, W. Liu, Z. Wang, W. He, G. Li, Q. Tang, H. Guo, X. Pu, Y. Liu,
[10] R. D. I. G. Dharmasena, S. R. P. Silva, Nano Energy 2019, 62, 530.
C. Hu, Nat. Commun. 2021, 12, 4689.
[11] R. L. Bulathsinghala, W. Ding, R. Dharmasena, Nano Energy 2023,
[45] B. Yang, W. Zeng, Z.-H. Peng, S.-R. Liu, K. Chen, X.-M. Tao, Adv. Energy
108792.
Mater. 2016, 6, 1600505.
[12] J. Cui, H. Li, B. Chen, Z. L. Wang, Chem. Eng. J. 2024, 499, 156193.
[46] M. Khorsand, J. Tavakoli, K. Kamanya, Y. Tang, Nano Energy 2019, 66,
[13] H. Li, Y. Zhang, Z. Gao, L. Liang, X. Wang, X. Liu, Y. Wu, H. Zheng,
104115.
Nano Energy 2023, 109, 108266.
[47] T. A. L. Burgo, T. R. D. Ducati, K. R. Francisco, K. J. Clinckspoor, F.
[14] R. D. I. G. Dharmasena, K. D. G. I. Jayawardena, C. A. Mills, J. H. B.
Galembeck, S. E. Galembeck, Langmuir 2012, 28, 7407.
Deane, J. V. Anguita, R. A. Dorey, S. R. P. Silva, Energy Environ. Sci.
[48] H. T. Baytekin, B. Baytekin, S. Soh, B. A. Grzybowski, Angew. Chem.,
2017, 10, 1801.
Int. Ed. 2011, 50, 6766.
[15] Z. L. Wang, Mater. Today 2017, 20, 74.
[49] M. M. Apodaca, P. J. Wesson, K. J. M. Bishop, M. A. Ratner, B. A.
[16] C. Liu, A. J. Bard, Nat. Mater. 2008, 7, 505.
Grzybowski, Angew. Chemie 2010, 122, 958.
[17] S. Lin, L. Xu, C. Xu, X. Chen, A. C. Wang, B. Zhang, P. Lin, Y. Yang, H.
[50] C. X. Lu, C. B. Han, G. Q. Gu, J. Chen, Z. W. Yang, T. Jiang, C. He, Z.
Zhao, Z. L. Wang, Adv. Mater. 2019, 31, 1808197.
L. Wang, Adv. Eng. Mater. 2017, 19, 1700275.
[18] L. S. McCarty, G. M. Whitesides, Angew. Chem., Int. Ed. 2008, 47,
[51] T.-H. Chang, Y.-W. Peng, C.-H. Chen, T.-W. Chang, J.-M. Wu, J.-C.
2188.
Hwang, J.-Y. Gan, Z.-H. Lin, Nano Energy 2016, 21, 238.
[19] H. T. Baytekin, A. Z. Patashinski, M. Branicki, B. Baytekin, S. Soh, B.
[52] V. Nguyen, R. Yang, Nano Energy 2013, 2, 604.
A. Grzybowski, Science 2011, 1201512.

Adv. Energy Mater. 2024, 2403853 2403853 (17 of 17) © 2024 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbH