Applied Materials Today 18 (2020) 100496

Contents lists available at ScienceDirect

Applied Materials Today

journal homepage: www.elsevier.com/locate/apmt

Surface charge density of triboelectric nanogenerators: Theoretical boundary and

optimization methodology
Chunlei Zhang a,b,1 , Linglin Zhou a,b,1 , Ping Cheng a,b,1 , Xing Yin a,b , Di Liu a,b , Xinyuan Li a,b ,
Hengyu Guo a,b , Zhong Lin Wang a,b,c,∗ , Jie Wang a,b,∗
a
Beijing Institute of Nanoenergy and Nanosystems, Chinese Academy of Sciences, Beijing 100083, China
b
College of Nanoscience and Technology, University of Chinese Academy of Sciences, Beijing 100049, China
c
School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332, USA

a r t i c l e i n f o a b s t r a c t

Article history: Although high charge densities of triboelectric nanogenerators (TENG) were achieved by working in high
Received 14 August 2019 vacuum or charge pumping techniques in atmosphere, owing to their complex structure and/or stability
Received in revised form issues, it still remains a great challenge and necessity to directly obtain the high charge density directly
26 September 2019
through triboelectrification effect in atmosphere. Here, a basic theory about the limitation factors of
Accepted 2 November 2019
surface charge density is comprehensively rebuilt through analytical mathematical derivation of the lim-
itation equation. As a result, high surface charge density can be obtained directly by a new optimization
Keywords:
methodology, i.e. using thin dielectric layer, which is demonstrated by the designed contact-separation
Triboelectric nanogenerators
Mechanical energy harvesting
model TENG and sliding model TENG. In addition, the theoretical models of charge decay and charge
Charge decay accumulation during triboelectrification process were built. This work provides not only a new facile
Charge accumulation and universal optimization methodology for TENG, but also a new insight in the triboelectrification pro-
Thin dielectric layer cess, both of which will prompt the applications of TENG ranging from powering electronic devices to
harvesting large-scale blue energy.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction energy from different types of motion, like body motions, wind,
and ocean waves [10–14]. Abundant researches on TENG have well
With a rapid development of small electronic devices in the confirmed its potential of wide application in not only powering
past several decades, sustainably powering these uncountable small electronic devices as self-powered systems and working as
devices with pollution-free energy sources has been becoming active sensors, but also harvesting water wave energy in large-
a severe issue [1,2]. To obtain a clean and sustainable power scale blue energy [15–19]. As an energy harvesting device, the
source, harvesting energy from ambient environment has been further extensive application of TENGs needs to enhance its output
proven to be a promising strategy. As a new energy harvesting power density, which has been demonstrated to be quadratically
technology invented in 2012, triboelectric nanogenerator (TENG) related to its surface charge density (␴sc ) [20,21]. Hence, lots of
has been demonstrated to efficiently harvest ubiquitous and con- efforts have been focused on improving ␴sc by means of mate-
stantly available mechanical energy and convert it into electricity. rial choice, structural optimization, artificial ion injection and so
The working principle of TENG is based on the coupling effect of on, where ␴sc was improved from 30 ␮C m-2 to 250 ␮C m-2
triboelectrification and electrostatic induction, and the fundamen- [20,22–27]. Nevertheless, their outputs were limited by the air
tal theory lies in Maxwell’s displacement current and change in avalanche breakdown according to Paschen’s law [28]. To avoid
surface polarization [3–9]. Based on the basic operating princi- the air breakdown, the TENG was operated in vacuum environ-
ple, various TENGs have been developed to harvest mechanical ment and a high charge density of 1003 ␮C m-2 was achieved
[29]. Whereas, in most practical situations, TENGs are operating
in the air atmosphere, where the maximum surface charge den-
sity ( m ) has been still restricted because of the air breakdown
∗ Corresponding authors at: Beijing Institute of Nanoenergy and Nanosystems,
effect. Very recently, high surface charge densities in atmosphere
Chinese Academy of Sciences, Beijing 100083, China.
E-mail addresses: [email protected] (Z.L. Wang),
were achieved by an external-charge TENG pumper and a self-
[email protected] (J. Wang). charge excitation TENG system, respectively [30–32]. However,
1
These authors contributed equally to this work. owing to complex circuit structure and stable issue of these sys-

https://doi.org/10.1016/j.apmt.2019.100496
2352-9407/© 2019 Elsevier Ltd. All rights reserved.
2 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496

tems, it still remains a great challenge and necessity to directly short-circuit condition can be derived as follow (Supplementary
obtain the high output charge density only through triboelectrifi- Note 2) [26]:
cation effect and electrostatic induction effect in atmosphere until
0 dx
now. Vgap = (1)
ε0 (d + εr x)
In this work, a high surface charge density is realized in atmo-
sphere by using an thin film as TENG dielectric layer. Firstly, we where ␴0 is the surface charge density of dielectric layer, d is the
systematically and precisely analyzed the effects of dielectric thick- thickness of dielectric layer, x represents the gap distance, ␧0 and
ness (d) and separation distance (xSep ) on surface charge density ␧r are the vacuum permittivity (8.85 × 10-12 F m-1 ) and relative
of dielectric layer for the contact-separation TENG (CS-TENG) by permittivity of dielectric layer, respectively. Theoretical analysis
mathematical derivation of the limitation equation. It is the first based on air breakdown voltage (Vb ) described by the Paschen’s law
time to point out the available range of Paschen’s law in TENG and the Eq. (1) can derive the relationship between the thickness
field. Though reducing the d, we obtain an high surface charge (d) of the dielectric film on the maximum surface charge density
density of 1090 ␮C m-2 in atmosphere through ion injection on ( m ), named the limitation equation (Supplementary Note 3):
an thin 6 ␮m Kapton film, further demonstrate the suitability of  BPε (d + ε x) 
avalanche breakdown model in thin dielectric layer and establish 0 r
m = (2)
a charge interface recombination model to explain the follow- (ln (Px) + C)d min
ing charge decay process. Furthermore, we attain stable values where B and C are the constants determined by the composition
of 418 and 740 ␮C m-2 by using fluorinated ethylene propy- and the pressure of the gas. For air at standard atmospheric pressure
lene (FEP) film with a thickness of 15 ␮m via the accumulation (P) of 101 kPa, they are 273.75 V Pa-1 m-1 and 1.08, respectively. It is
process of the triboelectric charge by contact-separation model found that the  m will increase with the decreasing of the d (Fig. 1a),
and sliding model TENG, respectively. This work not only reach which has been confirmed by previous reports [26]. For the CS-
a new milestone of triboelectric charge density in atmosphere TENG based on Cu and polytetrafluoroethylene (PTFE), when the
and a new paradigm shift to improve the outputs of TENGs, but thickness of PTFE decreased from 600 to 200 ␮m in air, the charge
also provide some new insights into the triboelectrification pro- density increased from 90 to 120 ␮C m−2 25 . The value was further
cess. enhanced to 218.64 ␮C m−2 at fluorinated ethylene propylene (FEP)
with thickness of 50 ␮m [20]. Additionally, the ion injection method
for the introduction of surface charges on the FEP with different
2. Results thickness was also carried out to obtain the corresponding  m . With
the thickness of FEP decreasing from 125 to 50 ␮m, the charge den-
2.1. Limitation factors of surface charge density sity improved from 160 to 240 ␮C m−2 22 . It can be found that both
of the experimental values obtained from triboelectrification pro-
When the electrode and dielectric layer of CS-TENG sepa- cess and ion injection method agree well with the theoretical curve
rate, an extremely high electrostatic field will be built between when the d exceeds 50 ␮m, which clearly illustrates that a thinner
their surfaces with opposite triboelectric charges, simulated by dielectric film is favorable for enhancing the surface charge density.
finite-element method (Fig. S1, 2 and Note 1). Through the clas- However, the relationship between the  m and d of less than 50 ␮m
sical electrodynamics derivation and by ignoring edge effect, was not confirmed through experiment to date. Besides, previous
the drop of voltage between the contact surfaces under the work focused on enhancing surface charges has verified that  m is

Fig. 1. Limitation factors of surface charge density. a) Theoretical analysis on the relationship between the maximum surface charge density ( m ) and the thickness (d) of the
FEP film in atmosphere, with the experimental data points of  m obtained for the d exceeding 50 ␮m.; b) Air breakdown voltage based on the Paschen’s law and gap voltage
of TENG with various charge densities, wherein PTFE film with thickness of 200 ␮m is used as the triboelectric material. c) Air breakdown voltage based on the Paschen’s
law at atmosphere and gap voltage of TENG between copper and an thin Kapton film of 6 ␮m with various charge densities. d) The simplified schematic diagram of charge
accumulation and attenuation of TENGs. e) Comparison of the surface charge density measured in this work with that in previous works through the triboelectrification and
the ion injection respectively, * means the value obtained by siding model TENG.
 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496 3

limited by air breakdown, which is almost inevitable in contact- where B and C are the constants and which values are the same as
separation TENG operating in atmosphere [28]. Recently, a high the ones inEq. (2). It’s apparent that Vb must be larger than zero,
charge density in high vacuum is boosted to 660 ␮C m-2 with a Cu therefore,
and PTFE [29], because Vgap in high vacuum (such as the pressure
ln(Px) + C > 0 (4)
of 10-6 torr) is smaller than Vb with a general separation distance
(1 cm), thereby avoiding the limitation of air breakdown (Fig. 1b). Then, we can get the theoretical domain area of x >3.34 ␮m,
Whereas,  m of the TENG has been still limited to ∼250 ␮C m-2 in which is consistent with the basic principle of Paschen’s law. It
the air atmosphere. was established on the theory of Townsend discharge, in which
To avoid air breakdown, the Vgap must be smaller than Vb at any it’s a result of avalanche multiplication that permits electrical con-
separation distance. Even if the surface charge density on the thin duction through the gas. The avalanche multiplication is a gas
dielectric film (Kapton film, 6 ␮m) is up to an high value of 1716 ␮C ionization process, wherein the free electrons are accelerated and
m-2 , the Vgap is still not enough to break down the air between the then collided with gas molecules. So there will be a section of
gaps in theory (Fig. 1c). Therefore, the performance of TENGs at dif- distance for the free electrons acceleration and a chain reaction
ferent thickness of Kapton and FEP film below 50 ␮m was studied in micron dimension. Through the above theoretical derivation,
and the results are presented in the following sections. Two points avalanche breakdown occurred where the gap distance (xgap ) must
are of particular interest here. One is that an high surface charge be greater than 3.34 ␮m.
density in atmosphere can be obtained through ion injection on It’s obvious to note that  m depends on the x and d from lim-
thin dielectric films, but it decayed gradually. A charge interface itation equation and then their three-dimension diagram can be
recombination model is established to explain the surface charge plotted (Fig. 2a). Also the theoretical curve of  m for different thick-
decay process, where the surface charges may diffuse to atmo- ness of thin FEP films below 50 ␮m is plotted based on the analysis
sphere or the back electrode then result in a recombination with above (Fig. 2b). To further obtain the quantitate mathematic rela-
the opposite charges (Fig. 1d (i)). On the other hand, a stable and tionship, we define the variable part of the limitation equation as a
high surface charge density can be achieved in atmosphere by using function f(x,d):
thin dielectric film based on triboelectrification process and charge
accumulation process (Fig. 1d (ii)). Then, a stable maximized out- d + εr x
f (x, d) = (5)
put of TENG can be obtained when a dynamic equilibrium between d(ln(Px) + C)
charge accumulation and decay is built (Fig. 1d (iii)). Therefore,
The significant stages in the development of TENG in air atmo-
sphere are summarized (Fig. 1e). With the triboelectrification m = BPε0 f (x, d)min (6)
between Ag and PTFE, the surface charge density of TENG is around
In order to obtain the minimum value of f (x, d), it is necessary
30 ␮C m−2 at the beginning [18]. By introducing the electric double-
to take its partial derivatives for x and d respectively as following:
layer effect and adding a charge transport layer, it can reach to 110
␮C m-2 19 . The value increases to 218.64 ␮C m−2 when liquid metal ∂f εr (ln (Px) + C) − (d + εr x) 1x
gallium replaced Cu as triboelectric electrode, due to better contact = 2
(7)
∂x d(ln(Px) + C)
intimacy [20]. Inspired by liquid metal, we designed a soft mate-
rial and fragmental contacting structure TENG to further enhance ∂f εr x 1
=− · (8)
the surface charge density. The value is improved to 250 ␮C m−2 , ∂d ln(Px) + C d2
approaching the limitation of air breakdown. In this work, a stable
These equations indicate that the function f(x,d) increases first
and high surface charge density of 418 ␮C m−2 and 735 ␮C m−2 is
and then decreases in direction x, and monotonically decreases
achieved in atmosphere by using FEP film with thickness of 15 ␮m
with d.
via triboelectrification and surface charge accumulation process in
Then, set the ∂f /∂x =0, and we can get the following equation:
contact-separate model TENG and sliding model TENG. Compared
with that of the beginning TENG, the output triboelectric charge εr xE (ln(PxE ) + C − 1) = d (9)
density has been boosted by 14-fold. Besides, using an thin Kap-
ton film with the thickness of 6 ␮m, an high surface charge density The xE is the gap distance in the case of ∂f /∂x = 0, where the air
of 1090 ␮C m−2 in atmosphere is realized through ion injection breakdown is most likely to take place. From this equation, it can be
technique, which is 4 times as high as the value of 240 ␮C m−2 in found that xE must be larger than 9.09 ␮m to satisfy d > 0 (Fig. 2b).
previous work on the FEP film of 50 ␮m. The results indicate that In fact, the gap distance is generally larger than 1 cm in practical
a thinner dielectric film is favorable for achieving higher surface operation of CS-TENG.
charge densities, which provides an optimized direction for design- According to the Eqs. (8) and (9), the maximum surface charge
ing the device of TENG and enhancing the surface charge density of density can be obtained by the maximum surface charge density
TENG. Eq. (10) or equation set (11, 12), while separation distance xSep is
smaller or larger than xE , respectively, (Supplementary Note 4):
2.2. Theoretical maximum surface charge density BPε0 (d + εr xsep )
m = xsep ∈ (3.34, xE ) (10)
(ln(Pxsep ) + C)d
In previous work, the existence of air breakdown effect had
BPε0
been confirmed repeatedly. Once air breakdown occurs, immedi- m = (11)
ate surface discharge on the dielectric layer will happen, resulting ln(PxE ) + C − 1
in a sharp decrease of dielectric surface charge density. Therefore, εr exp(BPε0 /m + ln(Bε0 /m ) + 1 − C) = d xsep ∈ (xE , +∞) (12)
it’s a dominant limitation of the output performance of TENG. The
threshold voltage of air breakdown between two parallel plates fol- With a certain d,  m shows a decreasing trend with the sepa-
lows the Paschen’s law 24, 28 (Supplementary Note 3), which can be ration distance of CS-TENG increasing from 3.34 ␮m to xE . We can
simplified as: see that the xE is only 24 ␮m with the general FEP film of 50 ␮m.
(Fig. 2c) For kapton dielectric layer, the relationship of  m between
BPx xE and d is shown in Supplementary Fig. 3. Hence, a higher theoreti-
Vb = (3)
ln (Px) + C cal value of  m can be achieved when the micro separation distance
4 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496

Fig. 2. Theoretical analysis of the performance of TENGs based on Paschen’s law. a) The relationship plot among surface charge density, gap distance, and thickness of
dielectric layer. b) The theoretical maximum surface charge density allowed without air breakdown of FEP films with thickness of 12.5, 15, 30 and 50 ␮m in various gap
distances. c) The relationship between the  m of dielectric layer FEP with different thickness (12.5, 15, 30, and 50 ␮m) and the separation distance of TENG. d) The relationship
between the dielectric layer thickness and the gap distance in the condition of ∂( 0 )max /∂x = 0, and the relationship of theoretical maximum surface charge density and the
thickness of dielectric layer.

smaller than xE calculated from the Eq. (10). And as CS-TENG work- maximum surface charge density. With the thickness of FEP (and
ing with a micro dimension separation distance, the thin film is Kapton) decreasing, the  m has a tendency to increase, which val-
prerequisite to realize the approximately full transfer of electrons ues are consistent with the theoretical analysis (Fig. 3b, e). As a
from front electrode to back electrode [33] (Supplementary Note 5). result, the  m in air is up to an high value of 1090 ␮C m-2 with
Once the separation distance is larger than the xE , the occurrence an thin Kapton film of 6 ␮m, and the value of 575 ␮C m-2 for FEP
of air breakdown will result in a constant  m which is determined film is achieved at the thickness of 12.5 ␮m (Figs. S4 and 5). It can
by the d, which value can be calculated by Eq. (12). be found that the experimental results by ions injection method
Generally, after the d is given, the xE and  m can be calculated match well with theoretical curve, except for the relatively thinner
from the Eqs. (9) and (12). According to the Eq. (12), it reveals that Kapton films of 6 ␮m. This inconsistency may be due to the surface
 m increases with the decreasing of d (Fig. 2d). It provides us an defects increasing with the thickness decreasing, which may result
accurate way to estimate the theoretical limitation for different in a higher partial charge density compared to the theoretical value,
dielectric layer. Therefore, thinning the dielectric layer is an opti- or the partial micro discharging resulted from the charge diffusion
mizing method to enhance the output performance of CE-TENG. to the back electrode, which is easier for the thin dielectric film.
Of course, the  m will be limited by dielectric layer breakdown or It should be noted that even though a high charge density can
quantum tunneling effect in the case of thin dielectric layer. be obtained using thinner Kapton or FEP films, a charge decreasing
similar to exponential decay can be observed under high charge
density after air breakdown, and then the decay in charge density
2.3. Achieving maximum surface charge density via. ion injection exhibits slowdown trend with time increasing (Fig. 3c, f and Supple-
method mentary Fig. 6). To explain the phenomenon of surface charge decay
with time in atmosphere, the decay equation based on charge inter-
The validity of relationship of  m and the d is confirmed in the face recombination model was established. Firstly, we assume that
case of d > 50 ␮m. Considering the features of triboelectrification the amount of initial charges on the dielectric layer surface after
and store charge are the key important factors for achieving high air breakdown is QI . Due to the large difference of charge polarity
surface charge density of TENG, we choose the Kapton with high between the dielectric layer surface and air, the charges on dielec-
dielectric constant and FEP with high triboelectrification capacity tric layer surface will spontaneously recombine with the particles
to better study the relationship between  m and the d with the d charged oppositely in air. For the apparent decay of surface charges,
less than 50 ␮m. it’s rational to ignore the effect of triboelectrification, especially
We adopted the method of injecting enough negative ions onto during the short time after ion injection. At the time of considering
the FEP (and Kapton) films surface by an air-ionization-gun when the effect of charges produced by triboelectrification, the corre-
the two electrodes of TENG were separated. The process of ion sponding amount of surface charges is defined as QT . Hence, only
injection was carried out for several times to realize a high enough the process of charges decay during the time that surface charges
surface charge density that can result in air breakdown. Then a decreased from QI to QT should be concerned about, wherein the
sharp decrease of charge density is observed when the electrodes accumulative decayed charges (QD ) can be written as:
move to contact, indicating air breakdown occurred (Fig. 3a, d).
Due to the instantaneity of the air breakdown, the charge density
in the first cycle after the sharp decrease could be regarded as the QD = QI − QT (13)
 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496 5

Fig. 3. Output performance of TENG with thin dielectric layer via ion injection method. a) The surface charge density before and after the ions injection during the separation
process showing the occurring of air break down; b) Comparison between the charge density of FEP films (12.5, 15, 30 and 50 ␮m) after ion injection in first cycle and
theoretical values. c) The charge decay process on FEP film of 12.5 ␮m after ions injection. d) The surface charge density before and after the ions injection during the
separation process showing the occurring of air break down; e) Comparison between the charge density of Kapton film (6, 12.5, 25, and 30 ␮m) after ion injection in first
cycle and theoretical calculation. f) The charge decay process on Kapton film of 12.5 ␮m after ions injection. R-Square (COD), coefficient of determination, stands for the
proportion of the variance in the dependent variable that is predictable from the independent variable.

To describe the average existing time of dielectric surface surface charge density of 418 ␮C m-2 was achieved in atmosphere
charges, it is called the lifetime displayed by , then the decay by using FEP film with thickness of 15 ␮m (Fig. 4d). Compared
equation can be written as (Supplementary Note 7): with the significant charge decay via ion injection method, a stable
  surface charge density of different thickness of dielectric films is
(t) = D exp − t/ + T (14)
obtained though triboelectrification process. The theory analysis is
To verify our assumption, the function was applied to fit the carried out to elaborate the phenomenon as following. It’s widely
experiment values of surface charge density. It should be noted acknowledged that it takes some time for charges distributing on
that the dielectric constant of FEP (2.1) is almost the half of Kapton the surface of dielectric films and then accumulating to saturation
film (4), thus Kapton film with 25 ␮m and FEP film with 12.5 ␮m step by step. Hence, it’s rational for us to assume that there are some
have a similar theoretical  m (Supplementary Note 8). Both of their charges QA transferred from triboelectric electrode to dielectric
charge decay process shows great consistency with the charge layer in every cycle, and at the same time, some charges QD would
interface recombination model during the time until triboelectri- decay as described before during the separation process in the air.
fication effect working. The higher  T of FEP means that FEP has a As a consequence, the effective increasing amount of charges QE
better performance of triboelectrification than that of Kapton. This in every cycle can be defined as Eq. (15):
may be a new method to characterize the triboelectrification ability
QE = QA − QD (15)
of dielectric material.
When the QE is equal to zero, the surface charge density
2.4. Achieving high surface charge density via triboelectrification will be saturated. Then, the dynamic equilibrium between charges
accumulating and decaying will be established at the same time,
Except introducing the charge by using artificial injection resulting in the steady maximized output of TENGs. Therefore,
method, the relationship between the surface charge density and the obtained stable charge density through triboelectrification is
dielectric layer thickness below 50 ␮m was also investigated via derived from the dynamic equilibrium between charge accumula-
triboelectrification process, which can be used to study the charge tion and decay process.
accumulation process. In this experiment, the mixture of silicon The surface charge density of Kapton increased to the maxi-
rubber and carbon black instead of the cooper electrode was used as mum value with the thickness decrease from 30 to 25 ␮m, and
triboelectric electrode for improving contact intimacy [27], mean- then declined with the thickness further decrease from 25 to 6 ␮m
while FEP films with the thicknesses of 12.5, 15, 30, and 50 ␮m (Fig. 4b). Though their surface charge density is far from the theo-
and Kapton films with the thicknesses of 6, 12.5, 15, and 30 ␮m retical value due to its ordinary triboelectrification performance, it
were served as dielectric layer, respectively. When Kapton film con- has an improvement than previous work23 . The similar tendency
tacted with triboelectric electrode, the surface charges generated was observed by using different thickness of FEP films (Fig. 4e). It
and the amount of charges accumulated while the triboelectrifi- can be found that the measured datum of FEP (50, 30 and 15 ␮m) is
cation process carried out repeatedly (Fig. 4a). With the further extremely close to the theoretical value, while the FEP of 12.5 ␮m
triboelectrification, the charge density gradually tends to be stable shows colossal discrepancies with simulated points (Fig. 4c, f).
and finally reached 303 ␮C m-2 (Fig. S6). Similarly, a stable and high The unusual phenomenon may be derived from materials defect
6 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496

Fig. 4. Output performance of TENG with different thickness of dielectric layer via triboelectrification at frequency of 0.4 Hz. a) Charge accumulation process of Kapton with
thickness of 25 ␮m. b) Surface charge density of Kapton films with thickness of 6, 12.5, 25, and 30 ␮m via the triboelectrification. c) Theoretical relationship between  m and
the thickness of Kapton film, with the points of the experimental values. d) Charge accumulation process of FEP with thickness of 30 ␮m. e) Surface charge density of FEP films
with thickness of 12.5, 15, 30, 50 ␮m via the triboelectrification. f) Theoretical relationship between  m and the thickness of FEP film, with the points of the experimental
values.

on thinner film surface, which would lead to the imperfect con- in CS-TENG, resulting a faster charge accumulation and a higher
tact between the triboelectric electrode and thin dielectric film, equilibrium surface charge density according to formula (15).
then resulting in the vastly weaken of output performance dur-
ing triboelectrification process (Figs. S8 and 9). Of course, the 2.5. Application demonstration of TENG with thin dielectric layer
exact mechanism needs to be investigated in the future work. On
the other hand, the decay process will enhance on the thin films To further demonstrate the effectivity of adopting thin films,
as discussed before. These two hypotheses are confirmed by the contrast experiments were conducted through driving various
phenomena that an high charge density of 1003 ␮C m-2 could electronics by the CS-TENGs with thin FEP films (15 ␮m) and
be achieved through the charge accumulation on the FEP film of conventional FEP films (200 ␮m), named TENG15 and TENG200 ,
200 ␮m in in high vacuum (10-6 torr) without the charge decay respectively. The simplified self-powered system circuit schematic
caused by the charge recombination in the air and without consid- consists of TENG, a full-wave rectifier and an energy storage unit
ering the defects in thick dielectric layer [25] (Fig. S10). (a capacitor) (Fig. 5a). When the switch K1 is on and K2 is off, the
From the Fig. 2c, it notes that the  m can reach an high theoret- TENGs were operated at 2 Hz to charge a capacitor of 10 ␮F. By
ical value with a micro separation distance, which is much higher using TENG15 , it only takes 41 s to make the voltage of capacitor
than that with a centimeter separation distance [30]. To demon- increase by 9.5 V. As for the TENG200 , it takes approximately 3-fold
strate the theoretical deduction, we design a sliding model TENG time (128 s) for charging the same capacitor (Fig. 5b).
(Supplementary Note 8) [34]. A high and stable surface charge den- When the switch K1 is off and K2 is on, a capacitor of 47 ␮F was
sity of 735 ␮Cm−2 is achieved with an thin FEP film of 15 ␮m (Fig. discharging to drive the electronic watch alone, and the equivalent
S11 and 12), which is much higher than that in CS-TENG. From galvanostatic discharging current (Ieg ) is 0.61 ␮A (Supplementary
the Eqs. (9), (10) and Fig. 2d, it can be figured out that a CS-TENG Note 9). When both switch K1 and K2 are on and the TENGs were
with a FEP film (15 ␮m) whose maximum surface charge density operated at a low frequency of 3 Hz, the voltage of capacitor is
is 507 ␮Cm−2 when the separation distance is over 14.7 ␮m. There increasing over the time, which means that TENGs can generate
are four reasons why the maximum surface charge density in slid- more energy than the watch consumes. It can sustainably drive
ing model TENG is higher than that in CS-TENG. Firstly, if the gap electronic watch and charge the capacitor (Supplementary Movie
distance between the Cu electrode and dielectric layer is smaller 1), but the charging current of capacitor with TENG15 is 1.74 ␮A,
than 14.7 ␮m, the high charge density can be attained with almost and that is only 0.78 ␮A for TENG200 (Fig. 5c).
no air breakdown. Furtherly, if the gap distance is smaller than To further verify the effectivity of thin film TENG, a calculator
6.2 ␮m, which is entirely possible in sliding model TENG, there whose energy consumption is higher than that of electronic watch
will be no air breakdown even with a high surface charge density was driven by the TENGs with the same method, where its Ieg is
of 735 ␮C m-2 , according to the Eq. (10) and Fig. 2c. In addition, 1.41 ␮A. The voltage of capacitor declined when the TENG200 was
due to the electrostatic shielding of the back electrode, there is working, indicating that TENG200 couldn’t meet the energy con-
almost no air breakdown between the edges of sliding electrode and sumption of calculator. By contrast, The TENG15 can power the
dielectric layer. More importantly, the S-TENG enhances the tribo- calculator and charge the capacitor simultaneously and sustain-
electrification effect between the Cu electrode and dielectric layer, ably, while the charging current is 1.36 ␮A (Fig. 5d, Supplementary
so there will be more charges transferred in every cycle than that Movie 2).
 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496 7

Fig. 5. Application of the TENG with thin dielectric layer to drive electronics devices. a) Circuit schematic of the self-powered system consisting of the TENG and a capacitor.
b) Comparison of the charging curves of the capacitor of 10 ␮F by operating the TENG15 and TENG200 at 2 Hz. c) An electronic watch is driven and a capacitor of 47 ␮F is
charged simultaneously by TENG15 and TENG200 at 3 Hz. d) A calculator is driven and a capacitor of 100 ␮F is charged simultaneously by TENG15 but the calculator is driven
by TENG200 and the capacitor.

3. Conclusions ever, further decreasing the thickness of dielectric layer will cause
the decline of  m , which may be arise from materials defect on thin-
In this work, we comprehensively rebuilt the basic theory ner film surface. Future work will involve the material optimization
about the limiting factors of maximum surface charge density ( m ) to further enhance the surface charge and output performance of
through mathematical derivation of its limitation equation. It was TENG.
first to point out the dead zone of Paschen’s law in the output per- In summary, our study firstly points out the available range of
formance description of TENG, in which the air gap should be over Paschen’s law in TENG field, and then points to effective approaches
3.34 ␮m. Besides, within the restricted x>9.09 ␮m, the relationship for enhancing the  m of TENGs with micro-nano separation dis-
between the  m and the d described by the Eqs. (11) and (12) is tance or thin films. An high surface charge density was directly
workable. obtained through triboelectrification effect in atmosphere. Addi-
Through the above mathematical derivation, the quantitative tionally, the revealed charges accumulation and decay processes
relationship  m and xSep is first proposed. With a certain d, the  m will benefit for the fundamental understanding of the triboelec-
increases with the xSep decreasing in the range from 3.34 ␮m to the trification effect lasting for thousands of years. Our findings not
distance(xE ) where air breakdown occurs. Therefore, decreasing the only establish new optimization methodologies for TENG, but also
xSep to microscale, the  m will be further largely improved, which provide a new perspective of triboelectrification process and tribo-
is verified by the high triboelectric charge density of 735 ␮C m-2 electric nanogenerators.
attained in sliding model TENG with a FEP of 15 ␮m. Furthermore,
according to the Eq. (10), it can be figured out that the  m can be up
to 1100 ␮C m-2 when the xSep is less than 4.87 ␮m, which is close 4. Experimental section
to the FEP film dielectric breakdown25 . These results demonstrate
that a higher  m can be achieved in micro-nano separation distance, 4.1. Fabrication of the TENGs
which is also beneficial to the miniaturization and integration of
TENG. TENGs used in ion injection: a piece of foam was attached to
Through the mathematical derivation, the quantitative relation- an acrylic substrate, then a Cu film working as an electrode was
ship of  m and the d is also deduced, which is that  m increases adhered to the foam. As for the triboelectric dielectric material, a
with the decreasing of d. By ion injection method, it’s demonstrated FEP (Kapton) film deposited with a Cu layer via vacuum magnetron
that the relationship is also suitable for the dielectric layer thinner sputtering on the bottom was adhered to another acrylic substrate
than 50 ␮m. And through a charge accumulation process caused (Fig. S13a). TENGs used in triboelectrification: a piece of sili-
by triboelectrification, the charge densities of FEP films (50, 30, con rubber-carbon electrode was attached to an acrylic substrate,
and 15 ␮m) are extremely close to the theoretical value, further which detailed preparation method is presented in Supplemen-
confirming the theory. Besides, the validity of thin film also com- tary Note 10. As for the triboelectric material, a FEP (Kapton) film
pletely verified in the self-charge excitation TENG system, in which deposited with a Cu layer via vacuum magnetron sputtering on the
a high charge density of 720 ␮C m-2 was attained with a Kapton film bottom was adhered to another acrylic substrate (Fig. S13b). Muti-
of 9 ␮m28 . Theoretically, the ␴m can reach 1100 ␮C m-2 with the layered TENGs: Kapton film with thickness of 130 ␮m was folded
ultalthin film FEP of 5.32 ␮m calculated from the Eq. (12). How- into zigzag with 5 layers, acting as the substrate. Then a Cu film with
8 C. Zhang, L. Zhou, P. Cheng et al. / Applied Materials Today 18 (2020) 100496

an area of 3 × 5 cm2 is adhere to the both sides of Kapton film. Then [3] F. Fan, Z. Tian, Z.L. Wang, Nano Energy 1 (2012) 328–334.
FEP (15 or 200 ␮m) layers were attached on both sides of Cu layer [4] Z.L. Wang, Faraday Discuss. 176 (2012) 447–458.
[5] Z.L. Wang, Mater. Today 20 (2017) 74–82.
every other layer. The small dimension TENG with size of 2 × 2 cm2 [6] H. Zou, Y. Zhang, L. Guo, P. Wang, X. He, G. Dai, H. Zheng, C. Chen, A.C. Wang,
used for studying size effects were fabricated by the same method C. Xu, Z.L. Wang, Nat. Commun. 10 (2019) 1427.
(Fig. S14). [7] C. Wu, A.C. Wang, W. Ding, H. Guo, Z.L. Wang,Adv. Energy Mater. 9 (2019),
1802906.
[8] S. Pan, Z. Zhang,J. Appl. Phys. 122 (2017), 144302.
4.2. Characterization [9] S. Pan, Z. Zhang, Friction 7 (2019) 2–17.
[10] Q. Zheng, Y. Zou, Y. Zhang, Z. Liu, B. Shi, X. Wang, Y. Jin, H,Sci. Adv. 2 (2016),
e1501478.
The short-circuit current and transferred charges were mea- [11] J. Bae, J. Lee, S. Kim, J. Ha, B.S. Lee, Y. Park, C. Choong, J.B. Kim, Z.L. Wang, H.Y.
sured by a programmable electrometer (Keithley model 6514). The Kim, J.J. Park, U.I. Chung, Nat. Commun. 5 (2014) 4929.
contact-separation process was implemented by a linear motor [12] G. Zhu, J. Chen, T. Zhang, Q. Jing, Z.L. Wang, Nat. Commun. 5 (2014) 3426.
[13] Z.L. Wang, Nature 542 (2017) 159–160.
(TSMV120-1S). Ionized-air-injection for the introduction of the
[14] Y. Zi, L. Lin, J. Wang, S. Wang, J. Chen, X. Fan, P.K. Yang, F. Yi, Z.L. Wang, Adv.
surface charges was achieved by an air-ionization gun (Milty Zero- Mater. 27 (2015) 2340–2347.
stat 3). A potentiostat (Bio-Logic VSP-300, France) was utilized to [15] Z. Wen, M.H. Yeh, H. Guo, J. Wang, Y. Zi, W. Xu, J. Deng, L. Zhu, X. Wang, C. Hu,
L. Zhu, X. Sun, Z.L. Wang,Sci. Adv. 2 (2016), e1600097.
test the capacitance of the capacitor and the charging/discharging
[16] F. Yi, X. Wang, S. Niu, S. Li, Y. Yin, K. Dai, G. Zhang, L. Lin, Z. Wen, H. Guo, J.
curves of the self-charging power system. The microstructures of Wang, M.H. Yeh, Y. Zi, Q. Liao, Z. You, Y. Zhang, Z.L. Wang,Sci. Adv. 2 (2016),
samples were conducted by a cold field emission scanning electron eaat2516.
microscope (FESEM, HITACHI SU8200) with the working voltage [17] X. Pu, H. Guo, J. Chen, X. Wang, Y. Xi, C. Hu, Z.L. Wang,Sci. Adv. 3 (2017),
e1700694.
and current of 5 kV and 10 ␮A, respectively. [18] H. Guo, X. Pu, J. Chen, Y. Meng, M.H. Yeh, G. Liu, Q. Tang, B. Cheng, D. Liu, S. Qi,
C. Wu, C. Hu, J. Wang, Z.L. Wang,Sci. Robot. 3 (2018), eaat2516.
Declaration of Competing Interest [19] J. Wang, X. Li, S. Wang, Z. Li, L. Zheng, F. Yi, S. Li, Z.L. Wang, Adv. Mater. 27
(2015) 4830–4836.
[20] Y. Zi, S. Niu, J. Wang, Z. Wen, W. Tang, Z.L. Wang, Nat. Commun. 6 (2015) 8376.
The authors declared that they have no conflicts of interest to [21] J. Peng, S.D. Kang, G.J. Snyder,Sci. Adv. 3 (2017), eaap8576.
this work. [22] J. Zhong, Q. Zhong, F. Fan, Y. Zhang, S. Wang, B. Hu, Z.L. Wang, J. Zhou, Nano
Energy 2 (2013) 491–497.
[23] N. Cui, L. Gu, Y. Lei, J. Liu, Y. Qin, X. Ma, Y. Hao, Z.L, ACS Nano 10 (2016)
Acknowledgments 6131–6138.
[24] J. Wang, Z. Wen, Y. Zi, P. Zhou, J. Lin, H. Guo, Y. Xu, Z.L. Wang, Adv. Funct.
Mater. 26 (2016) 1070–1076.
Research supported by the National Key R & D Project [25] S. Wang, Y. Zi, Y. Zhou, S. Li, F. Fan, L. Lin, Z.L. Wang, J. Mater. Chem. A 4 (2016)
from Minister of Science and Technology (2016YFA0202704), 3728–3734.
National Natural Science Foundation of China (Grant No.61774016, [26] S. Wang, Y. Xie, S. Niu, L. Lin, C. Liu, Y. Zhou, Z.L. Wang, Adv. Mater. 26 (2014)
6720–6728.
21773009, 51432005, 5151101243, 51561145021), Beijing Munic- [27] J. Wang, S. Li, F. Yi, Y. Zi, J. Lin, X. Wang, Y. Xu, Z.L. Wang, Nat. Commun. 7
ipal Science & Technology Commission (Z171100000317001, (2016) 12744.
Z171100002017017, Y3993113DF). [28] Y. Zi, C. Wu, W. Ding, Z.L. Wang,Adv. Funct. Mater. 27 (2017), 1700049.
[29] J. Wang, C. Wu, Y. Dai, Z. Zhao, A. Wang, T. Zhang, Z.L. Wang, Nat. Commun. 8
(2017) 88.
Appendix A. Supplementary data [30] L. Xu, T.Z. Bu, X.D. Yang, C. Zhang, Z.L. Wang, Nano Energy 49 (2018)
625–633.
[31] L. Cheng, Q. Xu, Y. Zheng, X. Jia, Y. Qin, Nat. Commun. 9 (2018) 3773.
Supplementary material related to this article can be found,
[32] W.Z. Liu, G. Wang, G. Wang, J. Liu, X. Chen, Y. Pu, X. Xi, H. Wang, C. Guo, Z. Hu,
in the online version, at doi:https://doi.org/10.1016/j.apmt.2019. L. Wang, Nat. Commun. 10 (2019) 1426.
100496. [33] K.T.A.L. Burm, Contrib. Plasma Phys. 47 (2007) 177–182.
[34] S. Niu, S. Wang, L. Lin, Y. Liu, Y. Zhou, Y. Hu, Z.L. Wang, Energy Environ. Sci. 6
(2013) 3576.
References

[1] S. Chu, A. Majumdar, Nature 488 (2012) 294–303.

[2] J. Chen, Y. Huang, N. Zhang, H. Zou, R. Liu, C. Tao, X. Fan, Z.L. Wang, Nat. Energy
1 (2016) 16138.