The Impact of Pulse Charging Parameters on the Life Cycle of Lithium-Ion Polymer Batteries

Abstract

The pulse charging algorithm is seen as a promising battery charging technique to satisfy the needs of electronic device consumers to have fast charging and increased battery charge and energy efficiencies. However, to get the benefits of pulse charging, the pulse charge current parameters have to be chosen carefully to ensure optimal battery performance and also extend the life cycle of the battery. The impact of pulse charge current factors on the life cycle and battery characteristics are seldom investigated. This paper seeks to evaluate the impact of pulse charge current factors, such as frequency and duty cycle, on the life cycle and impedance parameters of lithium-ion polymer batteries (LiPo) while using a design of experiments approach, Taguchi orthogonal arrays. The results are compared with the benchmark constant current-constant voltage (CC-CV) charging algorithm and it is observed that by using a pulse charger at optimal parameters, the cycle life of a LiPo battery can be increased by as much as 100 cycles. It is also determined that the duty cycle of the pulse charge current has the most impact on the cycle life of the battery. The battery impedance characteristics were also examined by using non-destructive techniques, such as electrochemical impedance spectroscopy, and it was determined that the ambient temperature at which the battery was charged had the most effect on the battery impedance parameters.

Full text

energies
Article
The Impact of Pulse Charging Parameters on the Life
Cycle of Lithium-Ion Polymer Batteries
J. M. Amanor-Boadu 1,*, A. Guiseppi-Elie 2ID and E. Sánchez-Sinencio 1
1Department of Electrical and Computer Engineering, Texas A&M University, College Station,
Texas 77840, USA; sanchez@ece.tamu.edu
2
EnMed Working Group and Department of Biomedical Engineering, Texas A&M University, College Station,
Texas 77840, USA; guiseppi@tamu.edu
*Correspondence: judyboadu@tamu.edu; Tel.: +01-979-845-9583
Received: 2 August 2018; Accepted: 15 August 2018; Published: 18 August 2018


Abstract:
The pulse charging algorithm is seen as a promising battery charging technique to satisfy the
needs of electronic device consumers to have fast charging and increased battery charge and energy
efficiencies. However, to get the benefits of pulse charging, the pulse charge current parameters
have to be chosen carefully to ensure optimal battery performance and also extend the life cycle of
the battery. The impact of pulse charge current factors on the life cycle and battery characteristics
are seldom investigated. This paper seeks to evaluate the impact of pulse charge current factors,
such as frequency and duty cycle, on the life cycle and impedance parameters of lithium-ion polymer
batteries (LiPo) while using a design of experiments approach, Taguchi orthogonal arrays. The results
are compared with the benchmark constant current-constant voltage (CC-CV) charging algorithm
and it is observed that by using a pulse charger at optimal parameters, the cycle life of a LiPo battery
can be increased by as much as 100 cycles. It is also determined that the duty cycle of the pulse charge
current has the most impact on the cycle life of the battery. The battery impedance characteristics were
also examined by using non-destructive techniques, such as electrochemical impedance spectroscopy,
and it was determined that the ambient temperature at which the battery was charged had the most
effect on the battery impedance parameters.
Keywords:
Li-ion polymer battery; pulse charging; constant current constant voltage; electrochemical
impedance spectroscopy; battery impedance; life cycle; Taguchi orthogonal arrays; design
of experiments
1. Introduction
Lithium-ion (Li-ion) batteries have continued to increase their market share in the rechargeable
battery market over the past few years [
1
,
2
]. They are projected to overtake the lead-acid battery
market share by 2024 [
3
]. This is primarily driven by their superior qualities of high energy density, low
self-discharge, and lack of memory effect [
4
]. They can be found in numerous devices and applications,
ranging from electric vehicles to the Internet of Things (IoT) devices. Lithium-ion polymer (LiPo)
batteries are analogous to their Li-ion counterparts, except for the use of a gel-like, polymer electrolyte,
which serves to improve ion mobility [
5
]. They are commonly packaged in a pouch format that
allows for the LiPo battery to be molded into many conformable form factors. LiPo batteries have
the consumer-held, market-driven advantages of being lightweight and having good mechanical
strength [4].
Fast charging, increased battery runtime, and increased battery charge and energy efficiencies are
important characteristics that consumers of portable electronic applications desire. Though battery
runtime can be improved by using various hardware and software techniques [
6
–
8
], these desirable
Energies 2018,11, 2162; doi:10.3390/en11082162 www.mdpi.com/journal/energies

Energies 2018,11, 2162 2 of 15
characteristics are impacted by the type of charging algorithm used, and as such, many charging
algorithms have been proposed to achieve desirable and optimal output performance metrics. Of these
charging algorithms, pulse charging is the most propitious [
9
–
13
]. Pulse charging is the application
of carefully controlled charge current pulses into the battery. Figure 1schematically illustrates the
typical LiPo battery undergoing a charging process and shows the movement of the lithium ions (Li
+
),
as they move from the positive electrode, cathode, through the separator to intercalate in the negative
electrode, anode. During charging, Li
+
arriving at the anode desolvate and diffuse within to become
intercalated into the anode material. Issues that are related to desolvation dynamics, space-charge
layers, and diffusion within the anode material may produce inefficiencies in the intercalation and
result in accumulation of Li at the anode-electrolyte interface. If this is not abated, deposits grow
into dendrite like structures [
14
] with the potential to penetrate the separator, leading to a thermal
runaway [
15
]. To prevent this problem, the constant current-constant voltage (CC-CV) charging
algorithm, which is considered as the benchmark charging algorithm, is used in many LiPo battery
powered devices [
16
]. The CC-CV charging algorithm is shown in Figure 2. There are two major
charging phases: constant current (CC) charge phase and the constant voltage (CV) charge phase.
In the CC charge phase, a constant charge current (
ICH ARGE
) is used to initially charge the battery.
The battery voltage quickly rises in this region, as the Li
+
can freely intercalate in the anode. When the
voltage is approaching full charge (
Vf ull
), the CC charge phase transitions to the CV charge phase,
where the result is a decreasing charge current that prevents damage to the battery, ensures safety of
the user, and allows time for the Li
+
to properly intercalate. However, this CV region prolongs charge
time due to the decreasing charge current used. Pulse charging does not have this drawback.
Energies 2018, 11, x FOR PEER REVIEW 2 of 15
desirable characteristics are impacted by the type of charging algorithm used, and as such, many
charging algorithms have been proposed to achieve desirable and optimal output performance
metrics. Of these charging algorithms, pulse charging is the most propitious [9–13]. Pulse charging is
the application of carefully controlled charge current pulses into the battery. Figure 1 schematically
illustrates the typical LiPo battery undergoing a charging process and shows the movement of the
lithium ions (Li+), as they move from the positive electrode, cathode, through the separator to
intercalate in the negative electrode, anode. During charging, Li+ arriving at the anode desolvate and
diffuse within to become intercalated into the anode material. Issues that are related to desolvation
dynamics, space-charge layers, and diffusion within the anode material may produce inefficiencies
in the intercalation and result in accumulation of Li at the anode-electrolyte interface. If this is not
abated, deposits grow into dendrite like structures [14] with the potential to penetrate the separator,
leading to a thermal runaway [15]. To prevent this problem, the constant current-constant voltage
(CC-CV) charging algorithm, which is considered as the benchmark charging algorithm, is used in
many LiPo battery powered devices [16]. The CC-CV charging algorithm is shown in Figure 2. There
are two major charging phases: constant current (CC) charge phase and the constant voltage (CV)
charge phase. In the CC charge phase, a constant charge current () is used to initially charge
the battery. The battery voltage quickly rises in this region, as the Li+ can freely intercalate in the
anode. When the voltage is approaching full charge (), the CC charge phase transitions to the CV
charge phase, where the result is a decreasing charge current that prevents damage to the battery,
ensures safety of the user, and allows time for the Li+ to properly intercalate. However, this CV region
prolongs charge time due to the decreasing charge current used. Pulse charging does not have this
drawback.
Figure 1. Schematic illustration of the charging process in a lithium-ion polymer (LiPo) battery.
Figure 2. Constant Current-Constant Voltage charging algorithm.
Figure 1. Schematic illustration of the charging process in a lithium-ion polymer (LiPo) battery.
Energies 2018, 11, x FOR PEER REVIEW 2 of 15
desirable characteristics are impacted by the type of charging algorithm used, and as such, many
charging algorithms have been proposed to achieve desirable and optimal output performance
metrics. Of these charging algorithms, pulse charging is the most propitious [9–13]. Pulse charging is
the application of carefully controlled charge current pulses into the battery. Figure 1 schematically
illustrates the typical LiPo battery undergoing a charging process and shows the movement of the
lithium ions (Li+), as they move from the positive electrode, cathode, through the separator to
intercalate in the negative electrode, anode. During charging, Li+ arriving at the anode desolvate and
diffuse within to become intercalated into the anode material. Issues that are related to desolvation
dynamics, space-charge layers, and diffusion within the anode material may produce inefficiencies
in the intercalation and result in accumulation of Li at the anode-electrolyte interface. If this is not
abated, deposits grow into dendrite like structures [14] with the potential to penetrate the separator,
leading to a thermal runaway [15]. To prevent this problem, the constant current-constant voltage
(CC-CV) charging algorithm, which is considered as the benchmark charging algorithm, is used in
many LiPo battery powered devices [16]. The CC-CV charging algorithm is shown in Figure 2. There
are two major charging phases: constant current (CC) charge phase and the constant voltage (CV)
charge phase. In the CC charge phase, a constant charge current () is used to initially charge
the battery. The battery voltage quickly rises in this region, as the Li+ can freely intercalate in the
anode. When the voltage is approaching full charge (), the CC charge phase transitions to the CV
charge phase, where the result is a decreasing charge current that prevents damage to the battery,
ensures safety of the user, and allows time for the Li+ to properly intercalate. However, this CV region
prolongs charge time due to the decreasing charge current used. Pulse charging does not have this
drawback.
Figure 1. Schematic illustration of the charging process in a lithium-ion polymer (LiPo) battery.
Figure 2. Constant Current-Constant Voltage charging algorithm.
Figure 2. Constant Current-Constant Voltage charging algorithm.

Energies 2018,11, 2162 3 of 15
The use of pulse charging has been shown to increase battery charge and energy efficiencies
and to also reduce charge time [
9
,
17
,
18
]. It also has the added benefit of improved safety, since the
relaxation times in between charge current pulses allows time for the Li
+
to successfully intercalate
and helps to prevent dendrite formation [14].
The challenge in the development and deployment of pulse charging algorithms is robustly
finding and defining the conditions that produce the most efficient battery performance. Pulse charge
currents have the different attributes, which can be at different levels and depending on the selection
of these attributes and their levels, the battery output performance metrics, such as the battery charge
and energy efficiencies, life cycle, and charge time, will differ. Pulse charging has been shown to have
superior performances but the impact of these factors and their levels on the cycle life of the battery
and its impedance parameters are seldom investigated. This paper seeks to address this shortcoming
by using a design of experiments (DoE) approach, Taguchi orthogonal array (OA), and non-destructive
techniques, such as electrochemical impedance spectroscopy, in order to determine the impact of pulse
charge current factors and levels on battery cycle life and impedance parameters. This work also
addresses the question of what factors have the most significant impact on the battery cycle life.
This paper is organized as follows: Section 2provides a brief background on pulse charging and
Taguchi OA. Section 3reviews different techniques for evaluating the impact of charging parameters
on battery characteristics. The design procedure is detailed in Section 4and the experimental results
are presented in Section 5. Conclusions are drawn in Section 6.
2. Background on Pulse Charging and Taguchi Orthogonal Array (OA)
2.1. Pulse Charging
Pulse charging involves using carefully selected and controlled charge current pulses to charge a
battery. Figure 3a shows a macromodel of a pulse charger where by controlling the switch, S
w
, charge
current, I
c
, is pulsed into the battery. Every pulse charge current that is applied to the battery has the
following factors or attributes: peak current amplitude,
Ipk
, a duty cycle,
D=ton /Tp
, and frequency, f,
as shown in Figure 3b.
Energies 2018, 11, x FOR PEER REVIEW 3 of 15
The use of pulse charging has been shown to increase battery charge and energy efficiencies and
to also reduce charge time [9,17,18]. It also has the added benefit of improved safety, since the
relaxation times in between charge current pulses allows time for the Li+ to successfully intercalate
and helps to prevent dendrite formation [14].
The challenge in the development and deployment of pulse charging algorithms is robustly
finding and defining the conditions that produce the most efficient battery performance. Pulse charge
currents have the different attributes, which can be at different levels and depending on the selection
of these attributes and their levels, the battery output performance metrics, such as the battery charge
and energy efficiencies, life cycle, and charge time, will differ. Pulse charging has been shown to have
superior performances but the impact of these factors and their levels on the cycle life of the battery
and its impedance parameters are seldom investigated. This paper seeks to address this shortcoming
by using a design of experiments (DoE) approach, Taguchi orthogonal array (OA), and non-
destructive techniques, such as electrochemical impedance spectroscopy, in order to determine the
impact of pulse charge current factors and levels on battery cycle life and impedance parameters.
This work also addresses the question of what factors have the most significant impact on the battery
cycle life.
This paper is organized as follows: Section 2 provides a brief background on pulse charging and
Taguchi OA. Section 3 reviews different techniques for evaluating the impact of charging parameters
on battery characteristics. The design procedure is detailed in Section 4 and the experimental results
are presented in Section 5. Conclusions are drawn in Section 6.
2. Background on Pulse Charging and Taguchi Orthogonal Array (OA)
2.1. Pulse Charging
Pulse charging involves using carefully selected and controlled charge current pulses to charge
a battery. Figure 3a shows a macromodel of a pulse charger where by controlling the switch, Sw,
charge current, Ic, is pulsed into the battery. Every pulse charge current that is applied to the battery
has the following factors or attributes: peak current amplitude, , a duty cycle, = 
⁄, and
frequency, f, as shown in Figure 3b.
(a) (b)
Figure 3. (a) Pulse charging micromodel; and, (b) pulse waveform.
The selection of pulse attributes and their different levels also impacts the charge time, battery
life cycle, and battery impedance parameters. The LiPo battery, which can be ideally modelled as a
series resistance, Rs, which represents the resistance of the electrolyte and other contact resistances,
in series with a parallel combination of the charge transfer resistance, Rct, which represents the
resistance of the kinetics of the Faradaic processes in the battery, and Cdl, which represents the
charged areas between the electrode and electrolyte [19], is shown in Figure 4a. By using
electrochemical impedance spectroscopy (EIS), the battery impedance parameters can be extracted
from the Nyquist plot that is shown in Figure 4b. During pulse charging, the Li+ are able to intercalate
properly in the anode, thereby reducing polarization voltage [20]. This, in turn, reduces energy losses
in the battery, increasing the power transfer rate and improving charge and energy efficiencies. For
a fixed pulse charge current amplitude, lower duty cycles result in longer charge times, and higher
Figure 3. (a) Pulse charging micromodel; and (b) pulse waveform.
The selection of pulse attributes and their different levels also impacts the charge time, battery life
cycle, and battery impedance parameters. The LiPo battery, which can be ideally modelled as a series
resistance, R
s
, which represents the resistance of the electrolyte and other contact resistances, in series
with a parallel combination of the charge transfer resistance, R
ct
, which represents the resistance
of the kinetics of the Faradaic processes in the battery, and C
dl
, which represents the charged areas
between the electrode and electrolyte [
19
], is shown in Figure 4a. By using electrochemical impedance
spectroscopy (EIS), the battery impedance parameters can be extracted from the Nyquist plot that
is shown in Figure 4b. During pulse charging, the Li
+
are able to intercalate properly in the anode,
thereby reducing polarization voltage [
20
]. This, in turn, reduces energy losses in the battery, increasing
the power transfer rate and improving charge and energy efficiencies. For a fixed pulse charge current

Energies 2018,11, 2162 4 of 15
amplitude, lower duty cycles result in longer charge times, and higher duty cycles result in shorter
charge times, but the battery is more susceptible to overcharging and overvoltage, two conditions that
can be detrimental to the life cycle of the battery. For a fixed duty cycle, higher current amplitudes,
resulting in higher average charge currents, will result in faster charge times, but the tendency of
the battery to overcharge is high. The mathematical relation between
Ipk
,
D
, and the average charge
current Iavg can be expressed as:
Iavg =Ipk ·D(1)
Energies 2018, 11, x FOR PEER REVIEW 4 of 15
duty cycles result in shorter charge times, but the battery is more susceptible to overcharging and
overvoltage, two conditions that can be detrimental to the life cycle of the battery. For a fixed duty
cycle, higher current amplitudes, resulting in higher average charge currents, will result in faster
charge times, but the tendency of the battery to overcharge is high. The mathematical relation
between , , and the average charge current  can be expressed as:
=∙
(1)
(a) (b)
Figure 4. (a) LiPo battery equivalent circuit; and (b) Nyquist plot.
In terms frequency, [17] and [18] determined that pulsing at the frequency at which the battery
impedance is minimum  resulted in faster charge times. These previous works determined the
optimal pulse charge current factors that will result in a reduced charge time and increased battery
charge and energy efficiencies. Short term improvements in efficiencies and charge times have been
recorded for pulse chargers [10,17,18,21], but it is important to determine how the pulse charge
current attributes or factors affect the battery in the long term. This will impact the overall cost of a
system, as if the proposed pulse charging algorithms reduces the life cycle of batteries or increases
battery impedance parameter values, frequent replacements will be needed and that will impact the
long-term cost of the battery powered application. Other works, [12] and [13], have investigated the
impact of pulse charging on the life cycle of Li-ion batteries and found that pulse charging extended
the life of Li-ion batteries when compared with dc charging protocols. These previous works,
however, do not provide details regarding which factors of the pulse charge current have the most
significant impact on the battery cycle life. By using a design of experiments approach, the impact of
pulse charge current factors on battery cycle life and impedance parameters can be determined and
evaluated.
2.2. Taguchi Orthogonal Arrays
A typical system undergoing an experiment is shown in Figure 5. Every experiment will have
input factors, output responses, control factors, and uncontrollable factors. Input factors are variables
that can be controlled to impact the output performance of a system. It is sometimes needed to
determine how the output responses are impacted by the input, control factor levels, and
uncontrollable factors. Designs of experiments are therefore methods that are used to acquire
predictive information about systems. They can be used to optimize the output responses of a system
by determining which input factors and their control factors impact the output responses the most.
Taguchi OA are one of such methods that seeks to reduce the number of experiments without
affecting the outcome of the system. Its design consists of balanced experimental sets where different
input factors are at different levels. It can be used to evaluate the dependence of a certain output
response on the input factor and input factor levels. It can also be designed to include uncontrollable
factors, such as disturbances or tolerances. A typical Taguchi OA is shown in Table 1, where F1 to Fn
are the input factor at different levels and O are the output responses. This table can also be expressed
Figure 4. (a) LiPo battery equivalent circuit; and (b) Nyquist plot.
In terms frequency, References [
17
,
18
] determined that pulsing at the frequency at which the
battery impedance is minimum
fzmin
resulted in faster charge times. These previous works determined
the optimal pulse charge current factors that will result in a reduced charge time and increased battery
charge and energy efficiencies. Short term improvements in efficiencies and charge times have been
recorded for pulse chargers [
10
,
17
,
18
,
21
], but it is important to determine how the pulse charge current
attributes or factors affect the battery in the long term. This will impact the overall cost of a system,
as if the proposed pulse charging algorithms reduces the life cycle of batteries or increases battery
impedance parameter values, frequent replacements will be needed and that will impact the long-term
cost of the battery powered application. Other works, References [
12
,
13
] have investigated the impact
of pulse charging on the life cycle of Li-ion batteries and found that pulse charging extended the life of
Li-ion batteries when compared with dc charging protocols. These previous works, however, do not
provide details regarding which factors of the pulse charge current have the most significant impact on
the battery cycle life. By using a design of experiments approach, the impact of pulse charge current
factors on battery cycle life and impedance parameters can be determined and evaluated.
2.2. Taguchi Orthogonal Arrays
A typical system undergoing an experiment is shown in Figure 5. Every experiment will have
input factors, output responses, control factors, and uncontrollable factors. Input factors are variables
that can be controlled to impact the output performance of a system. It is sometimes needed to
determine how the output responses are impacted by the input, control factor levels, and uncontrollable
factors. Designs of experiments are therefore methods that are used to acquire predictive information
about systems. They can be used to optimize the output responses of a system by determining which
input factors and their control factors impact the output responses the most. Taguchi OA are one of
such methods that seeks to reduce the number of experiments without affecting the outcome of the
system. Its design consists of balanced experimental sets where different input factors are at different
levels. It can be used to evaluate the dependence of a certain output response on the input factor and
input factor levels. It can also be designed to include uncontrollable factors, such as disturbances or

Energies 2018,11, 2162 5 of 15
tolerances. A typical Taguchi OA is shown in Table 1, where F
1
to F
n
are the input factor at different
levels and Oare the output responses. This table can also be expressed as
LyAx
, where yis the number
of experiments, Ais the levels of each factor, xis the number of factors, and Lis for Latin square.
The designed array is always orthogonal, indicating that the number of factor levels in each column
occurs the same number of times, and therefore it is balanced. Taguchi OAs have been applied in
the field of industrial applications [
22
–
25
], communication [
26
], agriculture [
27
], and battery charging
systems [18,28–30]. Different OA sizes have been described in [31].
Energies 2018, 11, x FOR PEER REVIEW 5 of 15
as , where y is the number of experiments, A is the levels of each factor, x is the number of
factors, and L is for Latin square. The designed array is always orthogonal, indicating that the number
of factor levels in each column occurs the same number of times, and therefore it is balanced. Taguchi
OAs have been applied in the field of industrial applications [22–25], communication [26], agriculture
[27], and battery charging systems [18], [28–30]. Different OA sizes have been described in [31].
Figure 5. A system undergoing an experiment.
Table 1. A typical Taguchi orthogonal array (OA).

Factors






1



2





By applying statistical analysis methods, such as analysis of means (ANOM) and analysis of
variance (ANOVA), the output responses of the Taguchi OA can be analyzed. Genichi Taguchi, who
is the pioneer of Taguchi OAs [32], recommended the use of signal-to-noise (S/N) ratios to evaluate
the output responses from the experimental design. They are expressed as:

=
⎩
⎪
⎪
⎨
⎪
⎪
⎧−10log



−ℎ−
−10log
−ℎ−
−10log

ℎ
(2)
where =∑  and =(−1)∑−
; and,  and  refer to the output response in
in the  row and  column.
In this work, Taguchi OA is used to determine the effect of pulse charge current factors on the
cycle life and the impedance parameters of the battery. The input is the pulse charge current and the
control factors to be evaluated will be duty cycle, frequency, and the ambient temperature, T, at which
the battery is charged. The output responses were battery cycle life, Rs, and Rct. Battery cycle life in
this work is defined as the number of charge and discharge cycles that the LiPo battery can undergo
before its capacity reaches 80% of its original capacity. The battery Rs and Rct are also evaluated after
the aging process.
Figure 5. A system undergoing an experiment.
Table 1. A typical Taguchi orthogonal array (OA).
LFactors O
F1F2······ FnOa······ Om
1
An
Oa1Om1
2Oa2······ Om2
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
y Oay Omy
By applying statistical analysis methods, such as analysis of means (ANOM) and analysis of
variance (ANOVA), the output responses of the Taguchi OA can be analyzed. Genichi Taguchi, who is
the pioneer of Taguchi OAs [
32
], recommended the use of signal-to-noise (S/N) ratios to evaluate the
output responses from the experimental design. They are expressed as:
S
N=















−10 logn−1∑
i
Oij
−2larger −the −better
−10 logn−1∑
i
Oij 2smaller −the −better
−10 logO2
v2Nominal is the best
(2)
where O=n−1∑
i
Oij and v2=(n−1)−1∑
iOij −O2; and, iand jrefer to the output response in the
ith row and jth column.
In this work, Taguchi OA is used to determine the effect of pulse charge current factors on the
cycle life and the impedance parameters of the battery. The input is the pulse charge current and the
control factors to be evaluated will be duty cycle, frequency, and the ambient temperature, T, at which
the battery is charged. The output responses were battery cycle life, R
s
, and R
ct
. Battery cycle life in
this work is defined as the number of charge and discharge cycles that the LiPo battery can undergo
before its capacity reaches 80% of its original capacity. The battery R
s
and R
ct
are also evaluated after
the aging process.

Energies 2018,11, 2162 6 of 15
3. Techniques for Evaluation of Impact of Charging Parameters on Battery Characteristics
To determine the impact of charging factors on the cycle life and the impedance parameters of the
battery, it is important to use techniques that give relevant information about the battery characteristics.
These techniques can be classified as destructive and non-destructive [33].
3.1. Destructive Techniques
Destructive techniques involve disassembling the battery to investigate the current state of its internal
components, in terms of structure, chemical composition, and morphology. These techniques can be used
to determine how different charging methods have impacted electrode morphology and structure [
34
,
35
],
dendrite depositions [
36
,
37
], separator [
38
], and electrolyte degradation [
39
,
40
]. Investigation of these battery
internal components can be performed while using scanning electron microscopy (SEM) [
33
–
35
,
37
,
41
],
atomic force microscopy (AFM) [
39
,
42
], X-ray diffraction (XRD) [
41
,
43
–
45
], Raman microscopy [
39
,
43
],
and electron probe microscopic analysis (EPMA) [46], of which SEM is the most popular.
3.2. Non-Destructive Techniques
Non-destructive techniques do not involve the disassembling of the battery and have been known
to be the popular ways for determining the battery impedance and performance across its life cycle.
These techniques include cyclic voltammetry [
44
,
46
,
47
], impedance measurements, charge-discharge
tests, and impedance spectroscopy [
48
], of which impedance spectroscopy is the most popular [
49
].
In this work, a combination of different non-destructive techniques will be used to characterize the
battery aging, in effect its cycle life and the impact on battery impedance parameters.
By using the Taguchi OA design method that is described in Section 2in combination with
regression analysis, the factors and factor levels that have the greatest impact on the cycle life and
impedance parameters of the battery can be determined. All of the results are compared with the
benchmark CC-CV charging algorithm.
4. Design Procedure
Figure 6shows the flow chart of the design procedure to determine the impact of pulse charge
current factors on battery cycle life and impedance parameters. Before the Taguchi OA is designed and
experiments conducted, the input factors and their levels, output responses, and uncontrollable factors
should be defined. The number of factors and factor levels will indicate the size of the OA.
Energies 2018, 11, x FOR PEER REVIEW 6 of 15
3. Techniques for Evaluation of Impact of Charging Parameters on Battery Characteristics
To determine the impact of charging factors on the cycle life and the impedance parameters of
the battery, it is important to use techniques that give relevant information about the battery
characteristics. These techniques can be classified as destructive and non-destructive [33].
3.1. Destructive Techniques
Destructive techniques involve disassembling the battery to investigate the current state of its
internal components, in terms of structure, chemical composition, and morphology. These techniques
can be used to determine how different charging methods have impacted electrode morphology and
structure [34,35], dendrite depositions [36,37], separator [38], and electrolyte degradation [39,40].
Investigation of these battery internal components can be performed while using scanning electron
microscopy (SEM) [33–35,37,41], atomic force microscopy (AFM) [39,42], X-ray diffraction (XRD)
[41,43–45], Raman microscopy [39,43], and electron probe microscopic analysis (EPMA) [46], of which
SEM is the most popular.
3.2. Non-Destructive Techniques
Non-destructive techniques do not involve the disassembling of the battery and have been
known to be the popular ways for determining the battery impedance and performance across its life
cycle. These techniques include cyclic voltammetry [44,46,47], impedance measurements, charge-
discharge tests, and impedance spectroscopy [48], of which impedance spectroscopy is the most
popular [49]. In this work, a combination of different non-destructive techniques will be used to
characterize the battery aging, in effect its cycle life and the impact on battery impedance parameters.
By using the Taguchi OA design method that is described in Section 2 in combination with
regression analysis, the factors and factor levels that have the greatest impact on the cycle life and
impedance parameters of the battery can be determined. All of the results are compared with the
benchmark CC-CV charging algorithm.
4. Design Procedure
Figure 6 shows the flow chart of the design procedure to determine the impact of pulse charge
current factors on battery cycle life and impedance parameters. Before the Taguchi OA is designed
and experiments conducted, the input factors and their levels, output responses, and uncontrollable
factors should be defined. The number of factors and factor levels will indicate the size of the OA.
Figure 6. Design procedure for determining the impact of pulse current factors on LiPo battery.
Figure 6. Design procedure for determining the impact of pulse current factors on LiPo battery.
Input factors and output responses definition: The input factors chosen are the duty cycle and
frequency of the pulse charge current, and the ambient temperature at which the battery is charged at.

Energies 2018,11, 2162 7 of 15
Three levels of duty cycle levels are chosen: 20%, 50%, and 80%. The impedance of the battery changes
with frequency, as shown in Figure 7. At much lower frequencies, the impedance is much higher due
to mass transport effects, and at high frequencies, the impedance is mainly due to the inductance of
the battery, which is as a result of the electrodes [
50
]. Six levels of frequency are chosen to cover the
frequency range of interest: 0.1 kHz, 1 kHz, 6 kHz,
fzmin
, 50 kHz, and 100 kHz.
fzmin
was included
to investigate the impact it has on the battery characteristics. The ambient temperature chosen were
0
◦
C, 23
◦
C and 45
◦
C. The amplitude is not chosen as a factor so as to have a fair comparison with the
CC-CV charging algorithm. Table 2shows the input factors and their levels. The uncontrollable factors
can be determined as the manufacturing variabilities on the batteries under test.
Energies 2018, 11, x FOR PEER REVIEW 7 of 15
Input factors and output responses definition: The input factors chosen are the duty cycle and
frequency of the pulse charge current, and the ambient temperature at which the battery is charged
at. Three levels of duty cycle levels are chosen: 20%, 50%, and 80%. The impedance of the battery
changes with frequency, as shown in Figure 7. At much lower frequencies, the impedance is much
higher due to mass transport effects, and at high frequencies, the impedance is mainly due to the
inductance of the battery, which is as a result of the electrodes [50]. Six levels of frequency are chosen
to cover the frequency range of interest: 0.1 kHz, 1 kHz, 6 kHz, , 50 kHz, and 100 kHz. 
was included to investigate the impact it has on the battery characteristics. The ambient temperature
chosen were 0 °C, 23 °C and 45 °C. The amplitude is not chosen as a factor so as to have a fair
comparison with the CC-CV charging algorithm. Table 2 shows the input factors and their levels. The
uncontrollable factors can be determined as the manufacturing variabilities on the batteries under
test.
Figure 7. Bode magnitude and phase plot for equivalent battery circuit model.
Table 2. Control factors for Taguchi OA design.
Control factors Factor levels
1 2 3 4 5 6
D 0.2 0.5 0.8 - - -
T (°C) 0 23 45 - - -
f (kHz) 0.1 1 6

50 100
The output responses are chosen to be battery cycle life, Rs, and Rct. Cdl is not chosen as an output
response since the magnitude of Rs, and Rct is much larger and will have more impact on the battery
impedance, which can also impact the battery aging.
OA size selection: From Table 2, there are three control factors with two factors at three levels,
and one factor at six levels. Hence, the OA to be designed has to be mixed level. An OA (3×6)
was chosen. The design of this OA ensured that all input factors and their levels appeared an equal
number of times. This OA is shown in Table 3.
Figure 7. Bode magnitude and phase plot for equivalent battery circuit model.
Table 2. Control factors for Taguchi OA design.
Control factors
Factor levels
1 2 3 4 5 6
D0.2 0.5 0.8 - - -
T(◦C) 0 23 45 - - -
f(kHz) 0.1 1 6 fzmin 50 100
The output responses are chosen to be battery cycle life, R
s
,and R
ct
.C
dl
is not chosen as an output
response since the magnitude of R
s,
and R
ct
is much larger and will have more impact on the battery
impedance, which can also impact the battery aging.
OA size selection: From Table 2, there are three control factors with two factors at three levels,
and one factor at six levels. Hence, the OA to be designed has to be mixed level. An OA
L3632×61
was chosen. The design of this OA ensured that all input factors and their levels appeared an equal
number of times. This OA is shown in Table 3.

Energies 2018,11, 2162 8 of 15
Table 3. Taguchi orthogonal array.
Factors Factors
No. T D f No. T D f
1 23 0.2 12k19 0 0.8 12k
2 45 0.8 100 20 0 0.5 6k
3 45 0.2 50k21 0 0.2 1k
4 45 0.2 6k22 0 0.2 100
5 45 0.5 100k23 23 0.2 100k
6 0 0.8 12k24 0 0.5 6k
7 0 0.2 100 25 45 0.5 100k
8 45 0.8 1k26 23 0.2 100k
9 23 0.5 1k27 23 0.8 50k
10 0 0.2 1k28 45 0.2 50k
11 0 0.5 50k29 23 0.8 6k
12 45 0.5 12k30 23 0.5 100
13 23 0.8 50k31 0 0.5 50k
14 45 0.5 12k32 0 0.8 100k
15 45 0.8 1k33 45 0.2 6k
16 45 0.8 100 34 23 0.5 1k
17 23 0.5 100 35 23 0.8 6k
18 23 0.2 12k36 0 0.8 100k
Experiments and response analysis: Experiments are conducted according to Table 3and responses
recorded and procedure is repeated until the battery ages significantly. At the end of the experiments,
the responses are then analyzed while using the Taguchi S/N ratios, ANOM, and ANOVA. From these
results, the factors that impact battery cycle life and impedance parameters can be determined.
By decomposing the variance and computing the sum of squares, the effects of any of the factors can
be determined from the sum of squares, which can be expressed as:
SSFn=∑
Ln


nLn·

1
nLn
∑
iLn
OiLnj−O

2

(3)
where
Oj
the output response,
nLn
is the number of output responses corresponding to level
n
, and
iLn
denotes the rows corresponding to level nresponses only.
5. Experimental Results
Figure 8shows the experimental test setup for running the experiments, according to Table 3.
New 600 mAh 3.7 V LiPo batteries were first characterized at 0% SoC (state of charge) by using EIS
to determine the battery impedance parameters. Battery impedance characteristics were measured
by performing an EIS using a Versastat Potentiostat [
51
] and the Nyquist plots analyzed by using the
Simplex method [
52
] for curve fitting to obtain battery impedance parameters. The
fzmin
of the batteries
was determined to be 12 kHz. Battery capacity measurements were also performed. The batteries
were then placed in a temperature chamber, TestEquity model 107 [
53
], and experiments conducted
according to Table 3. The batteries were subjected to a charge rate of 0.5C. After charging, when the
batteries reached a state of charge (SoC) of 100%, the batteries were allowed to rest for one hour before
being discharged at 0.5C until a SoC of 0% was attained. EIS measurements were then performed at
this SoC. Both pulse and CC-CV charging algorithms used the same charge and discharge rate of 0.5C.

Energies 2018,11, 2162 9 of 15
Energies 2018, 11, x FOR PEER REVIEW 9 of 15
Figure 8. Experimental setup for determining impact of charge algorithms on battery.
After about 400 cycles, the final EIS was performed and the battery impedance parameters
extracted. All the EIS measurements were performed at 0% SoC. The battery cycle life was estimated
from the recorded capacity measurements by using regression analysis and the cycle number at
which the battery capacity is 80% of the original capacity recorded. The output responses are shown
in Table 4.
Table 4. Output responses.
Factors Factors
No. Cycle
Life


No. Cycle
Life


1 583 0.8604 0.08091 19 873 0.9420 0.16680
2 370 0.8680 0.09468 20 308 0.9100 0.15500
3 239 0.9200 0.09845 21 438 0.9020 0.10390
4 421 0.9200 0.09845 22 360 0.9020 0.10390
5 763 0.8590 0.12000 23 295 0.8604 0.08091
6 446 0.9420 0.16680 24 214 0.9100 0.15500
7 213 0.9020 0.10390 25 755 0.8590 0.12000
8 258 0.8680 0.09468 26 250 0.8604 0.08091
9 545 0.8589 0.06816 27 486 0.9000 0.08174
10 303 0.9020 0.10390 28 408 0.9200 0.09845
11 512 0.9100 0.15500 29 377 0.9000 0.08174
12 388 0.8590 0.12000 30 828 0.8589 0.06816
13 309 0.9000 0.08174 31 871 0.9100 0.15500
14 311 0.8590 0.12000 32 401 0.9420 0.16680
15 215 0.8680 0.09468 33 593 0.9200 0.09845
16 268 0.8680 0.09468 34 396 0.8589 0.06816
17 434 0.8589 0.06816 35 393 0.9000 0.08174
18 456 0.8604 0.08091 36 408 0.9420 0.16680
Figure 9 shows the main effects plot and the S/N ratio, the larger-the-better, respectively, of the
LiPo battery cycle life. It can be seen from Figure 9 that the duty cycle of the pulse charge current has
a great impact on the battery cycle life and pulsing at 50% resulted in the longest cycle life. The
temperature at which the battery is charged also had an impact on the cycle life, and from Figure 9,
charging at 45 °C produced the shortest cycle life. Charging at  also impacts the battery, as it is
at this frequency that the battery impedance is minimum, and hence less heating losses. Charging at
20% duty cycle produced the shortest cycle life due to the higher peak current used at this duty cycle
to maintain the same average charge current, according to Equation (1). Higher peak currents can be
Figure 8. Experimental setup for determining impact of charge algorithms on battery.
After about 400 cycles, the final EIS was performed and the battery impedance parameters extracted.
All the EIS measurements were performed at 0% SoC. The battery cycle life was estimated from the
recorded capacity measurements by using regression analysis and the cycle number at which the battery
capacity is 80% of the original capacity recorded. The output responses are shown in Table 4.
Table 4. Output responses.
Factors Factors
No. Cycle Life RsRct No. Cycle Life RsRct
1 583 0.8604
0.08091
19 873 0.9420
0.16680
2 370 0.8680
0.09468
20 308 0.9100
0.15500
3 239 0.9200
0.09845
21 438 0.9020
0.10390
4 421 0.9200
0.09845
22 360 0.9020
0.10390
5 763 0.8590
0.12000
23 295 0.8604
0.08091
6 446 0.9420
0.16680
24 214 0.9100
0.15500
7 213 0.9020
0.10390
25 755 0.8590
0.12000
8 258 0.8680
0.09468
26 250 0.8604
0.08091
9 545 0.8589
0.06816
27 486 0.9000
0.08174
10 303 0.9020
0.10390
28 408 0.9200
0.09845
11 512 0.9100
0.15500
29 377 0.9000
0.08174
12 388 0.8590
0.12000
30 828 0.8589
0.06816
13 309 0.9000
0.08174
31 871 0.9100
0.15500
14 311 0.8590
0.12000
32 401 0.9420
0.16680
15 215 0.8680
0.09468
33 593 0.9200
0.09845
16 268 0.8680
0.09468
34 396 0.8589
0.06816
17 434 0.8589
0.06816
35 393 0.9000
0.08174
18 456 0.8604
0.08091
36 408 0.9420
0.16680
Figure 9shows the main effects plot and the S/N ratio, the larger-the-better, respectively, of the
LiPo battery cycle life. It can be seen from Figure 9that the duty cycle of the pulse charge current
has a great impact on the battery cycle life and pulsing at 50% resulted in the longest cycle life.
The temperature at which the battery is charged also had an impact on the cycle life, and from Figure 9,
charging at 45
◦
C produced the shortest cycle life. Charging at
fzmin
also impacts the battery, as it is
at this frequency that the battery impedance is minimum, and hence less heating losses. Charging at
20% duty cycle produced the shortest cycle life due to the higher peak current used at this duty cycle
to maintain the same average charge current, according to Equation (1). Higher peak currents can be
detrimental to the battery life cycle. The factor levels that resulted in a longer battery cycle life were
f=
12
kHz (fZmin)
,
D=
0.5, and T= 23
◦
C. By computing the ANOVA from the responses in Table 4,
the ANOVA table can be obtained, as shown in Table 5. From this table, it is seen that duty cycle had

Energies 2018,11, 2162 10 of 15
the largest impact on the cycle life, contributing 57.5%, as already determined from Figure 9. The
next largest contribution was attributed to the frequency at which the pulse charge current operated,
which contributed 39.8% to the total sum of squares.
Energies 2018, 11, x FOR PEER REVIEW 10 of 15
detrimental to the battery life cycle. The factor levels that resulted in a longer battery cycle life
were=12(), =0.5, and T = 23 °C. By computing the ANOVA from the responses in
Table 4, the ANOVA table can be obtained, as shown in Table 5. From this table, it is seen that duty
cycle had the largest impact on the cycle life, contributing 57.5%, as already determined from Figure
9. The next largest contribution was attributed to the frequency at which the pulse charge current
operated, which contributed 39.8% to the total sum of squares.
(a) (b)
Figure 9. (a) Main effects plot for battery cycle life; (b) Main effects plot of SN ratios for battery cycle
life.
Table 5. Analysis of variance (ANOVA) table for output responses.
Cycle Life
Factor Degrees of Freedom Sum of Squares Mean Square F
Duty Cycle 2 152562 76281 2.41
Frequency 5 105460 21092 0.58
Temperature 2 7221 3611 0.1
Total 265243
RS
Factor Degrees of Freedom Sum of Squares Mean Square F
Duty Cycle 2 0.004654 0.002327 3.02
Frequency 5 0.007104 0.001421 1.85
Temperature 2 0.013494 0.006747 13.41
Total 0.025252
Rct
Factor Degrees of Freedom Sum of Squares Mean Square F
Duty Cycle 2 0.003193 0.001596 1.59
Frequency 5 0.007084 0.001417 1.45
Temperature 2 0.025525 0.012762 38.84
Total 0.035802
Evaluating the impact of pulse charging parameters on the battery impedance parameters, Rs,
and Rct, the main effects plots and the S/N ratio, smaller the better, plots are shown in Figures 10 and
11. From the ANOVA table, the frequency at which the pulse charger operates had the second largest
effect on the Rs, and Rct values. The ambient temperature had the greatest impact on the impedance
parameter values, contributing 53.4% and 71.3% to the total sum of squares for Rs, and Rct,
respectively. This is as expected, as temperature causes the accelerated aging of the battery [54]. The
degradation in Rs and Rct may be due to electrolyte decomposition, increase in surface film growth,
reduction in active materials [44,46,49,55], and other effects that are produced due to various side
Figure 9.
(
a
) Main effects plot for battery cycle life; (
b
) Main effects plot of SN ratios for battery
cycle life.
Table 5. Analysis of variance (ANOVA) table for output responses.
Cycle Life
Factor Degrees of Freedom Sum of Squares Mean Square F
Duty Cycle 2 152,562 76,281 2.41
Frequency 5 105,460 21,092 0.58
Temperature 2 7221 3611 0.1
Total 265,243
RS
Factor Degrees of Freedom Sum of Squares Mean Square F
Duty Cycle 2 0.004654 0.002327 3.02
Frequency 5 0.007104 0.001421 1.85
Temperature 2 0.013494 0.006747 13.41
Total 0.025252
Rct
Factor Degrees of Freedom Sum of Squares Mean Square F
Duty Cycle 2 0.003193 0.001596 1.59
Frequency 5 0.007084 0.001417 1.45
Temperature 2 0.025525 0.012762 38.84
Total 0.035802
Evaluating the impact of pulse charging parameters on the battery impedance parameters, R
s,
and R
ct
, the main effects plots and the S/N ratio, smaller the better, plots are shown in Figures 10 and 11.
From the ANOVA table, the frequency at which the pulse charger operates had the second largest
effect on the R
s,
and R
ct
values. The ambient temperature had the greatest impact on the impedance
parameter values, contributing 53.4% and 71.3% to the total sum of squares for R
s,
and R
ct
,respectively.
This is as expected, as temperature causes the accelerated aging of the battery [
54
]. The degradation in
R
s
and R
ct
may be due to electrolyte decomposition, increase in surface film growth, reduction in active
materials [
44
,
46
,
49
,
55
], and other effects that are produced due to various side chemical reactions that
usually occur at the electrode/electrolyte interface [
49
]. These result in different degradations at the
anode and cathode. Cathode degradation can arise from structural changes that occur during charge
and discharge cycles while anode degradation can arise from increase in surface film growth (SEI) [
54
].
As a result of these degradations, Rsand Rct increase as the battery ages, as shown in Figure 12.

Energies 2018,11, 2162 11 of 15
Energies 2018, 11, x FOR PEER REVIEW 11 of 15
chemical reactions that usually occur at the electrode/electrolyte interface [49]. These result in
different degradations at the anode and cathode. Cathode degradation can arise from structural
changes that occur during charge and discharge cycles while anode degradation can arise from
increase in surface film growth (SEI) [54]. As a result of these degradations, Rs and Rct increase as the
battery ages, as shown in Figure 12.
(a) (b)
Figure 10. (a) Main effects plots for Rs; (b) Main effects plots of SN for Rs.
(a) (b)
Figure 11. (a) Main effects plot for Rct; (b) Main effects plots of SN for Rct.
(a) (b) (c)
Figure 12. Trend of Rs, and Rct vs charge/discharge cycles, (a) at 0 oC; (b) at 23 oC; (c) at 45 oC.
From Figure 12, batteries that were subjected to pulse charging were compared with the
benchmark CC-CV charging algorithm across different ambient temperatures. It can be seen that, at
room temperature, the + of the battery pulsed at 50% duty cycle was lower across the charge
and discharge cycles when compared to the battery subjected to the CC-CV charging algorithm.
By using the accelerated aging effect [56], the cycle life of batteries subjected to pulsing at optimal
values obtained from [18], i.e., =12.6() and =0.5, was compared to the benchmark
CC-CV charging algorithm and plotted in Figure 13. It was observed that the pulse charged batteries
attained more than 100 extra cycles when compared to the CC-CV charged batteries. This is due to
the rest periods in the pulses that allows for the effective intercalation of the Li+ before the next
application of the pulse, thereby preventing capacity fading due to the buildup of lithium deposits.
Figure 10. (a) Main effects plots for Rs; (b) Main effects plots of SN for Rs.
Energies 2018, 11, x FOR PEER REVIEW 11 of 15
chemical reactions that usually occur at the electrode/electrolyte interface [49]. These result in
different degradations at the anode and cathode. Cathode degradation can arise from structural
changes that occur during charge and discharge cycles while anode degradation can arise from
increase in surface film growth (SEI) [54]. As a result of these degradations, Rs and Rct increase as the
battery ages, as shown in Figure 12.
(a) (b)
Figure 10. (a) Main effects plots for Rs; (b) Main effects plots of SN for Rs.
(a) (b)
Figure 11. (a) Main effects plot for Rct; (b) Main effects plots of SN for Rct.
(a) (b) (c)
Figure 12. Trend of Rs, and Rct vs charge/discharge cycles, (a) at 0 oC; (b) at 23 oC; (c) at 45 oC.
From Figure 12, batteries that were subjected to pulse charging were compared with the
benchmark CC-CV charging algorithm across different ambient temperatures. It can be seen that, at
room temperature, the + of the battery pulsed at 50% duty cycle was lower across the charge
and discharge cycles when compared to the battery subjected to the CC-CV charging algorithm.
By using the accelerated aging effect [56], the cycle life of batteries subjected to pulsing at optimal
values obtained from [18], i.e., =12.6() and =0.5, was compared to the benchmark
CC-CV charging algorithm and plotted in Figure 13. It was observed that the pulse charged batteries
attained more than 100 extra cycles when compared to the CC-CV charged batteries. This is due to
the rest periods in the pulses that allows for the effective intercalation of the Li+ before the next
application of the pulse, thereby preventing capacity fading due to the buildup of lithium deposits.
Figure 11. (a) Main effects plot for Rct; (b) Main effects plots of SN for Rct.
Energies 2018, 11, x FOR PEER REVIEW 11 of 15
chemical reactions that usually occur at the electrode/electrolyte interface [49]. These result in
different degradations at the anode and cathode. Cathode degradation can arise from structural
changes that occur during charge and discharge cycles while anode degradation can arise from
increase in surface film growth (SEI) [54]. As a result of these degradations, Rs and Rct increase as the
battery ages, as shown in Figure 12.
(a) (b)
Figure 10. (a) Main effects plots for Rs; (b) Main effects plots of SN for Rs.
(a) (b)
Figure 11. (a) Main effects plot for Rct; (b) Main effects plots of SN for Rct.
(a) (b) (c)
Figure 12. Trend of Rs, and Rct vs charge/discharge cycles, (a) at 0 oC; (b) at 23 oC; (c) at 45 oC.
From Figure 12, batteries that were subjected to pulse charging were compared with the
benchmark CC-CV charging algorithm across different ambient temperatures. It can be seen that, at
room temperature, the + of the battery pulsed at 50% duty cycle was lower across the charge
and discharge cycles when compared to the battery subjected to the CC-CV charging algorithm.
By using the accelerated aging effect [56], the cycle life of batteries subjected to pulsing at optimal
values obtained from [18], i.e., =12.6() and =0.5, was compared to the benchmark
CC-CV charging algorithm and plotted in Figure 13. It was observed that the pulse charged batteries
attained more than 100 extra cycles when compared to the CC-CV charged batteries. This is due to
the rest periods in the pulses that allows for the effective intercalation of the Li+ before the next
application of the pulse, thereby preventing capacity fading due to the buildup of lithium deposits.
Figure 12. Trend of Rs,and Rct vs charge/discharge cycles, (a) at 0 oC; (b) at 23 oC; (c) at 45 oC.
From Figure 12, batteries that were subjected to pulse charging were compared with the
benchmark CC-CV charging algorithm across different ambient temperatures. It can be seen that,
at room temperature, the
Rs+Rct
of the battery pulsed at 50% duty cycle was lower across the charge
and discharge cycles when compared to the battery subjected to the CC-CV charging algorithm.
By using the accelerated aging effect [
56
], the cycle life of batteries subjected to pulsing at optimal
values obtained from [
18
], i.e.,
f=
12.6
kHz (fZmin)
and
D=
0.5, was compared to the benchmark
CC-CV charging algorithm and plotted in Figure 13. It was observed that the pulse charged batteries
attained more than 100 extra cycles when compared to the CC-CV charged batteries. This is due to the
rest periods in the pulses that allows for the effective intercalation of the Li
+
before the next application
of the pulse, thereby preventing capacity fading due to the buildup of lithium deposits.

Energies 2018,11, 2162 12 of 15
Energies 2018, 11, x FOR PEER REVIEW 12 of 15
Figure 13. Cycle life comparison of CC-CV charged battery and pulse charged battery.
6. Conclusions
The influence of pulse charging parameters and ambient temperature at which the battery is
charged on LiPo battery impedance parameters and life cycle has been determined. Battery
impedance parameter values increased as the number of charge and discharge cycles increased. This
is due to the various side chemical reactions that occur in the battery as it ages. The byproducts of
these reactions impact the battery impedance. Batteries that were subjected to pulse charging at 50%
duty cycle at room temperature had lower impedance values when compared to batteries subjected
to CC-CV at the same temperature. Results from the accelerated aging of the battery indicated that
pulse charging at the optimal values resulted in an increase of an additional 100 cycles when
compared to the benchmark method CC-CV. In terms of determining which factor and factor levels
of the pulse charge current affected the cycle life of the batteries, the Taguchi OA approach was used.
It was deduced that the duty cycle of the pulse charge current played a major role in battery cycle life
extension, followed by frequency at which the battery is charged. Ambient temperature impacted the
battery impedance parameters greatly.
This work has demonstrated the impact of pulse charge current factors and factor levels on the
cycle life of the battery. Accelerated life cycle tests have also proven that batteries subjected to the
pulse charging at the optimal parameters can result in longer cycle life when compared to batteries
that were subjected to the CC-CV charging algorithm.
Author Contributions: Conceptualization, J.M.A.-B. and A.G.-E.; Methodology, J.M.A.-B.; Software, J.M.A.-B.;
Validation, J.M.A.-B.; Formal Analysis, J.M.A.-B.; Investigation, J.M.A.-B.; Resources, A.G.-E. and E.S.-S.;
Writing-Original Draft Preparation, J.M.A.-B.; Writing-Review & Editing, J.M.A.-B., A.G.-E., and E.S.-S.;
Supervision, A.G.-E. and E.S.-S.
Funding: This research received no external funding.
Acknowledgments: The authors want to acknowledge Ankita Bhat of the Center for Bioelectronics, Biosensors,
and Biochips (C3B®) in the department of Biomedical Engineering at Texas A&M University, College Station,
TX,USA, for support in performing battery impedance measurements.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. PR Newswire Association LLC, Lithium-Ion Battery Market is Expected to Reach $46.21 Billion,
Worldwide, by 2022. Available online: https://www.prnewswire.com/news-releases/lithium-ion-battery-
market-is-expected-to-reach-4621-billion-worldwide-by-2022-575386231.html (accessed on 12 March 2018).
Figure 13. Cycle life comparison of CC-CV charged battery and pulse charged battery.
6. Conclusions
The influence of pulse charging parameters and ambient temperature at which the battery is
charged on LiPo battery impedance parameters and life cycle has been determined. Battery impedance
parameter values increased as the number of charge and discharge cycles increased. This is due to the
various side chemical reactions that occur in the battery as it ages. The byproducts of these reactions
impact the battery impedance. Batteries that were subjected to pulse charging at 50% duty cycle at
room temperature had lower impedance values when compared to batteries subjected to CC-CV at the
same temperature. Results from the accelerated aging of the battery indicated that pulse charging at
the optimal values resulted in an increase of an additional 100 cycles when compared to the benchmark
method CC-CV. In terms of determining which factor and factor levels of the pulse charge current
affected the cycle life of the batteries, the Taguchi OA approach was used. It was deduced that the
duty cycle of the pulse charge current played a major role in battery cycle life extension, followed
by frequency at which the battery is charged. Ambient temperature impacted the battery impedance
parameters greatly.
This work has demonstrated the impact of pulse charge current factors and factor levels on the
cycle life of the battery. Accelerated life cycle tests have also proven that batteries subjected to the
pulse charging at the optimal parameters can result in longer cycle life when compared to batteries
that were subjected to the CC-CV charging algorithm.
Author Contributions:
Conceptualization, J.M.A.-B. and A.G.-E.; Methodology, J.M.A.-B.; Software, J.M.A.-B.;
Validation, J.M.A.-B.; Formal Analysis, J.M.A.-B.; Investigation, J.M.A.-B.; Resources, A.G.-E. and E.S.-S.;
Writing-Original Draft Preparation, J.M.A.-B.; Writing-Review & Editing, J.M.A.-B., A.G.-E., and E.S.-S.;
Supervision, A.G.-E. and E.S.-S.
Funding: This research received no external funding.
Acknowledgments:
The authors want to acknowledge Ankita Bhat of the Center for Bioelectronics, Biosensors,
and Biochips (C3B
®
) in the department of Biomedical Engineering at Texas A&M University, College Station, TX,
USA, for support in performing battery impedance measurements.
Conflicts of Interest: The authors declare no conflict of interest.

Energies 2018,11, 2162 13 of 15
References
1.
PR Newswire Association LLC, Lithium-Ion Battery Market is Expected to Reach $46.21 Billion, Worldwide,
by 2022. Available online: https://www.prnewswire.com/news-releases/lithium-ion-battery-market-is-
expected-to-reach-4621-billion-worldwide- by-2022-575386231.html (accessed on 12 March 2018).
2.
Global Market for Lithium-Ion Batteries—Forecast, Trends & Opportunities 2014–2020. Available online:
https://www.researchandmarkets.com/reports/2904215/global-market-for-lithium-ion-batteries (accessed
on 12 March 2018).
3.
Grand View Research, Battery Market Size, Share | Industry Research Report, 2024. Available online:
https://www.grandviewresearch.com/industry-analysis/battery-marketvol (accessed on 23 March 2018).
4. Reddy, T.B. Linden’s Handbook of Batteries, 4th ed.; McGraw Hill Professional: New York, NY, USA, 2010.
5.
Meyer, W.H. Polymer electrolytes for lithium-ion batteries. Adv. Mater. (Deerfield Beach, Fla.)
1998
,10,
439–448. [CrossRef]
6.
Piqué, G.V.; Bergveld, H.J. State-of-the-Art of Integrated Switching Power Converters. In Analog Circuit
Design: Low Voltage Low Power; Short Range Wireless Front-Ends; Power Management and DC-DC; Steyaert, M.,
Roermund, A.v., Baschirotto, A., Eds.; Springer Netherlands: Dordrecht, The Netherlands, 2012; pp. 259–281.
7.
Iyenggar, A.; Tripathi, A.; Basarur, A.; Roy, I. Unified Power Management Framework for Portable Media
Devices. In Proceedings of the IEEE International Conference on Portable Information Devices, Orlando, FL,
USA, 25–29 May 2007; pp. 1–5.
8.
Lu, Y.H.; Simunic, T.; Micheli, G.d. Software controlled power management. Available online: http:
//seelab.ucsd.edu/papers/ylu_codes99.pdf (accessed on 1 April 2018).
9.
Yin, M.D.; Cho, J.; Park, D. Pulse-Based Fast Battery IoT Charger Using Dynamic Frequency and Duty
Control Techniques Based on Multi-Sensing of Polarization Curve. J. Energies 2016,9, 209. [CrossRef]
10.
Chen, L.-R. A Design of an Optimal Battery Pulse Charge System by Frequency-Varied Technique. IEEE TiE
2007,54, 398–405. [CrossRef]
11.
Purushothaman, B.K.; Landau, U. Rapid Charging of Lithium-Ion Batteries Using Pulsed Currents. JES
2006
,
153, A533. [CrossRef]
12.
Popov, B.N.; Haran, B.S.; Durairajan, A.; White, R.; Podrazhansky, Y.; Cope, R.C. Studies on capacity fade
of Li-ion cells cycled using pulse and DC charging protocols. In Proceedings of the Name of the 197th
Electrochemical Society meeting, Toronto, ON, Canada, 14–18 May 2000.
13.
Li, J.; Murphy, E.; Winnick, J.; Kohl, P.A. The effects of pulse charging on cycling characteristics of commercial
lithium-ion batteries. J. Power Sources 2001,102, 302–309. [CrossRef]
14.
Ely, D.R.; García, R.E. Heterogeneous Nucleation and Growth of Lithium Electrodeposits on Negative
Electrodes. JES 2013,160, A668. [CrossRef]
15.
Wang, Q.; Ping, P.; Zhao, X.; Chu, G.; Sun, J.; Chen, C. Thermal runaway caused fire and explosion of lithium
ion battery. J. Power Sources 2012,208, 210–224. [CrossRef]
16.
Shen, W.; Vo, T.T.; Kapoor, A. Charging algorithms of lithium-ion batteries: An overview. In Proceedings
of the 2012 7th IEEE Conference on Industrial Electronics and Applications (ICIEA), Singapore, Singapore,
18–20 July 2012; pp. 1567–1572.
17.
Boadu, J.M.A.; Abouzied, M.; Sanchez-Sinencio, E. An Efficient and Fast Li-ion Battery Charging System
Using Energy Harvesting or Conventional Sources. IEEE TIE 2018,65, 7383–7394.
18.
Boadu, J.M.A.; Guiseppi-Elie, A.; Sanchez-Sinencio, E. Search for Optimal Pulse Charging Parameters for
Li-ion Polymer Batteries Using Taguchi Orthogonal Arrays. IEEE TIE 2018,65, 8982–8992.
19. Randles, J.E.B. Kinetics of rapid electrode reactions. Discuss. Faraday Soc. 1947,1, 11–19. [CrossRef]
20.
Savoye, F.; Venet, P.; Millet, M.; Groot, J. Impact of Periodic Current Pulses on Li-Ion Battery Performance.
IEEE TIE 2012,59, 348–3488. [CrossRef]
21.
Chen, L.-R. Design of Duty-Varied Voltage Pulse Charger for Improving Li-Ion Battery-Charging Response.
IEEE TIE 2009,56, 480–487.
22.
Pundir, R.; Chary, G.H.V.C.; Dastidar, M.G. Application of Taguchi method for optimizing the process
parameters for the removal of copper and nickel by growing Aspergillus sp. Water Resour. Ind.
2016
.
[CrossRef]

Energies 2018,11, 2162 14 of 15
23.
Shahbazian, A.; Davood, A.; Dabirsiaghi, A. Application of Taguchi Method to Investigate the Effects of
Process Factors on the Production of Industrial Piroxicam Polymorphs and Optimization of Dissolution Rate
of Powder. Iran. J. Pharm. Res. IJPR 2016,15, 395. [PubMed]
24.
Pang, J.S.; Ansari, M.N.M.; Zaroog, O.S.; Ali, M.H.; Sapuan, S.M. Taguchi design optimization of machining
parameters on the CNC end milling process of halloysite nanotube with aluminium reinforced epoxy matrix
(HNT/Al/Ep) hybrid composite. HBRC J. 2014,10, 138–144. [CrossRef]
25.
Hsu, C.C.; Chien, C.S.; Wei, T.Y.; Li, Y.T.; Li, H.F. Taguchi DoE for ceramic substrate SMT defects improvement.
In Proceedings of the 2016 11th International Microsystems, Packaging, Assembly and Circuits Technology
Conference (IMPACT), Taipei, Taiwan, 26–28 October 2016; pp. 381–384.
26.
Dudhe, R.; Ayyalusamy, S.; Desai, T. Optimization of BAW Resonator for Wireless Applications using
Taguchi’s Orthogonal Array Method. Available online: https://www.researchgate.net/profile/Ravishankar_
Dudhe2/publication/303719442_Optimization_of_BAW_Resonator_for_Wireless_Applications_using_
Taguchi\T1\textquoterights_Orthogonal_Array_Method/links/574f1e0b08ae1880a820f7f4.pdf (accessed
on 23 January 2018).
27.
Deo, R.; Likhite, A.A.; Zanwar, D.R.; Majumdar, G. Design of experiments by Taguchi method in agricultural
research. J. Soils Crop. 2007,17, 320–325.
28.
Wang, S.C.; Chen, Y.L.; Liu, Y.H.; Huang, Y.S. A fast-charging pattern search for li-ion batteries with
fuzzy-logic-based Taguchi method. In Proceedings of the 2015 IEEE 10th Conference on Industrial Electronics
and Applications (ICIEA), Auckland, New Zealand, 15–17 June 2015; pp. 855–859.
29.
Liu, Y.H.; Luo, Y.F. Search for an Optimal Rapid-Charging Pattern for Li-Ion Batteries Using the Taguchi
Approach. IEEE TIE 2010,57, 3963–3971. [CrossRef]
30.
Wang, S.C.; Huang, J.W.; Liu, Y.H.; Hsieh, C.H. The implementation of consecutive orthogonal array method
on searching optimal five step charging pattern for Lithium-ion batteries. In Proceedings of the 2011 9th
World Congress on Intelligent Control and Automation, Taipei, Taiwan, 21–25 June 2011; pp. 358–363.
31.
Phadke, M.S. Quality Engineering Using Robust Design, 1st ed.; Prentice Hall: Upper Saddle River, NJ, USA,
1989.
32. Taguchi, G. Experimental Designs; Maruzen Publishing Company: Tokyo, Japan, 1976.
33.
Williard, N.; Sood, B.; Osterman, M.; Pecht, M. Disassembly methodology for conducting failure analysis on
lithium–ion batteries. J. Mater. Sci. Mater. Electron. 2011,22, 1616. [CrossRef]
34.
Scipioni, R.; Jørgensen, P.S.; Ngo, D.; Simonsen, S.B.; Liu, Z.; Yakal-Kremski, K.J.; Wang, H.; Hjelm, J.;
Norby, P.; Barnett, S.A.; Jensen, S.H. Electron microscopy investigations of changes in morphology and
conductivity of LiFePO4/C electrodes. J. Power Sources 2016,307, 259–269. [CrossRef]
35.
Freitag, S.; Berger, C.; Gelb, J.; Weisenberger, T.B.C. Li-Ion Battery Components—Cathode, Anode, Binder,
Separator—Imaged at Low Accelerating Voltages with ZEISS FE-SEMs, Zeiss White Paper, Carl Zeiss
Microscopy GmbH. Available online: https://p.widencdn.net/nih1gr/EN_wp_Battery_Low_Accelerating_
Voltages_ZEISS-FE_SEMs (accessed on 23 March 2018).
36.
Kong, L.; Xing, Y.; Pecht, M.G. In-Situ Observations of Lithium Dendrite Growth. IEEE Access
2018
,6,
8387–8393. [CrossRef]
37.
Orsini, F.; Pasquier, A.D.; Beaudoin, B.; Tarascon, J.M.; Trentin, M.; Langenhuizen, N.; de Beer, E.; Notten, P.
In situ Scanning Electron Microscopy (SEM) observation of interfaces within plastic lithium batteries.
J. Power Sources 1998,76, 19–29. [CrossRef]
38.
Chan, J.Y.; Kim, J.H.; Seunghoon, N.; Park, C.R.; Yang, S.J. Rational Design of Nanostructured Functional
Interlayer/Separator for Advanced Li–S Batteries. Adv. Funct. Mater. 2018, 1707411. [CrossRef]
39.
Jeong, S.; Inaba, M.; Iriyama, Y.; Abe, T.; Ogumi, Z. Surface film formation on a graphite negative electrode
in lithium-ion batteries: AFM study on the effects of co-solvents in ethylene carbonate-based solutions.
Electrochimica Acta 2002,47, 1975–1982. [CrossRef]
40.
Waag, W.; Käbitz, S.; Sauer, D.U. Experimental investigation of the lithium-ion battery impedance
characteristic at various conditions and aging states and its influence on the application. Appl. Energy
2013,102, 885–897. [CrossRef]
41.
Aurbach, D.; Zinigrad, E.; Cohen, Y.; Teller, H. A short review of failure mechanisms of lithium metal and
lithiated graphite anodes in liquid electrolyte solutions. Solid State Ion. 2002,148, 405–416. [CrossRef]

Energies 2018,11, 2162 15 of 15
42.
Jeong, S.; Inaba, M.; Mogi, R.; Iriyama, Y.; Abe, T.; Ogumi, Z. Surface Film Formation on a Graphite Negative
Electrode in Lithium-Ion Batteries: Atomic Force Microscopy Study on the Effects of Film-Forming Additives
in Propylene Carbonate Solutions. Langmuir 2001,17, 8281–8286. [CrossRef]
43.
Shim, J.; Kostecki, R.; Richardson, T.; Song, X.; Striebel, K.A. Electrochemical analysis for cycle performance
and capacity fading of a lithium-ion battery cycled at elevated temperature. J. Power Sources
2002
,112,
222–230. [CrossRef]
44.
Ramadass, P.; Durairajan, A.; Haran, B.S.; White, R.E.; Popov, B.N. Capacity fade studies on spinel based
Li-ion cells. In Proceedings of the Seventeenth Annual Battery Conference on Applications and Advances,
Long Beach, CA, USA, 18 January 2002; pp. 25–30.
45.
Itou, Y.; Ukyo, Y. Performance of LiNiCoO
2
materials for advanced lithium-ion batteries. J. Power Sources
2005,146, 39–44. [CrossRef]
46.
Zhang, D.; Haran, B.S.; Durairajan, A.; White, R.E.; Podrazhansky, Y.; Popov, B.N. Studies on capacity fade
of lithium-ion batteries. J. Power Sources 2000,91, 122–129. [CrossRef]
47.
Lundgren, C.A.; Xu, K.; Jow, T.R.; Allen, J.; Zhang, S.S. Springer Handbook of Electrochemical Energy; Springer:
Berlin Heidelberg, 2017; pp. 449–494. Available online: https://www.springer.com/us/book/9783662466568
(accessed on 20 March 2018).
48.
UTröltzsch; Kanoun, O.; Tränkler, H. Characterizing aging effects of lithium ion batteries by impedance
spectroscopy. Electrochimica Acta 2006,51, 1664–1672.
49.
Vetter, J.; Novák, P.; Wagner, M.R.; Veit, C.; Möller, K.; Besenhard, J.O.; Winter, M.; Wohlfahrt-Mehrens, M.;
Vogler, C.; Hammouche, A. Ageing mechanisms in lithium-ion batteries. J. Power Sources
2005
,147, 269–281.
[CrossRef]
50. Jossen, A. Fundamentals of battery dynamics. J. Power Sources 2006,154, 530–538. [CrossRef]
51.
Ametek Scientific Instruments, VersaSTAT 4|Potentiostat Galvanostat|Princeton Applied Research.
Available online: https://www.ameteksi.com/products/potentiostats/single-channel/versastat-series/
versastat-4 (accessed on 20 April 2018).
52.
Dincer, I. Comprehensive Energy Systems. Elsevier, Posted February 21, 2018. Available online: https://www.
sciencedirect.com/referencework/9780128149256/comprehensive-energy-systems (accessed on 2 March 2018).
53.
TestEquity, TestEquity 107 Benchtop Temperature Chamber (Environmental Chamber). Available online:
https://www.testequity.com/products/1592/ (accessed on 20 April 2018).
54.
Leng, F.; Tan, C.M.; Pecht, M. Effect of Temperature on the Aging rate of Li Ion Battery Operating above
Room Temperature. Sci. Rep. 2015,5, 12967. [CrossRef] [PubMed]
55.
Ning, G.; Haran, B.; Popov, B.N. Capacity fade study of lithium-ion batteries cycled at high discharge rates.
J. Power Sources 2003,117, 160–169. [CrossRef]
56.
Ecker, M.; Gerschler, J.B.; Vogel, J.; Käbitz, S.; Hust, F.; Dechent, P.; Sauer, D.U. Development of a lifetime
prediction model for lithium-ion batteries based on extended accelerated aging test data. J. Power Sources
2012,215, 248–257. [CrossRef]
©
2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).