Improving the rate capability of LiFePO4 electrode by controlling particle size distribution

Article abstract of DOI:10.1149/2.0621916jes

In this study, the rate performance of a LiFePO₄ (LFP) electrode has been enhanced by optimization of the particle size distribution of the LFP particles. Two LFP samples with different particle sizes (～50 and ～350 nm) are mixed with various ratios and the electrochemical performance has been evaluated. Reduction of the contact resistance and increase of the Li diffusion coefficient have been achieved. The electrode with a mixing ratio of 50:50 shows an improved initial capacity at C/10 and superior rate capability compared with the two pristine materials.

Full text of DOI:10.1149/2.0621916jes

A4128
Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
0
013-4651/2019/166(16)/A4128/8/$38.00 © The Electrochemical Society
Improving the Rate Capability of LiFePO Electrode by
4
Controlling Particle Size Distribution
1
,2,3
1,z
1
2,∗
Yin Zhang,
Jose A. Alarco,
Jawahar Y. Nerkar, Adam S. Best,
Graeme A. Snook,4 and Peter C. Talbot
,∗
1
1Institute for Future Environments and Science and Engineering Faculty, Queensland University of Technology (QUT),
Brisbane QLD 4001, Australia
2
CSIRO Manufacturing, Clayton South, Victoria 3169, Australia
CRRC Qingdao Sifang Rolling Stock Research Institute Co., Ltd., Qingdao 266031, China
CSIRO Mineral Resources, Clayton South, Victoria 3169, Australia
3
4
In this study, the rate performance of a LiFePO4 (LFP) electrode has been enhanced by optimization of the particle size distribution
of the LFP particles. Two LFP samples with different particle sizes (∼50 and ∼350 nm) are mixed with various ratios and the
electrochemical performance has been evaluated. Reduction of the contact resistance and increase of the Li diffusion coefficient have
been achieved. The electrode with a mixing ratio of 50:50 shows an improved initial capacity at C/10 and superior rate capability
compared with the two pristine materials.
©
2019 The Electrochemical Society. [DOI: 10.1149/2.0621916jes]
Manuscript submitted September 11, 2019; revised manuscript received November 11, 2019. Published December 12, 2019.
As one ofthe commercialcathode materials for rechargeableLi-ion
batteries (LIBs), LiFePO (LFP) offers many merits compared with
inated in nano LFP particles,17 resulting in better cycling performance.
By electrochemical impedance spectroscopy (EIS) and in-situ X-ray
4
18–20
conventional cathodes, such as environmental friendliness, low cost,
good safety, good cycling ability and a flat charge-discharge voltage
diffraction technology (XRD), the particle-by-particle
and hybrid
models have been confirmed for
20,21
(single-particle) phase-transition
+
plateau at relatively high potential of ∼3.45 V versus Li/Li . Al-
nano and micro LFP particles, respectively. In addition, a phase tran-
sition ahead of the charging and discharging processed of nano LFP
though LFP has had a wide-spread application in commercial LIBs in
1
,2
22
the last two decades, its application in the high-power scenario has
has been recently reported, which is related to the weaker memory
−
9
been somewhat limited by its poor intrinsic electronic (∼10 S/cm)
effect compared with micro LFP.
−
13
−16
2
3
and ionic conductivities (10 to 10 cm /s). Hence, considerable
efforts have been made to improve its rate capability. Olivine-type
LFP has an orthorhombic lattice structure with space group Pnma.4
The oxygen ions form strong covalent bonds with phosphorus ions
to form PO , which can stabilize the three-dimensional frame work
and provide safety and excellent cyclic performance. However, the
On the other hand, minimizing particle size can also bring prob-
lems. The tap density of nano-sized particles is generally lower than
that of micro-sized particles, which would decrease the energy density
of the cell. The nano-sized particles tending to agglomerate make the
homogeneous dispersion of the conductive carbon during electrode
3
4
−
5
23
fabrication very difficult. It should be noted that when the particle
strong covalent oxygen bonds also lead to low ionic diffusivity and
poor electronic conductivity. Therefore, modifications of LFP to im-
sizes are below 100 nm, the fraction of the material at the particle
surface increases adruptly, leading to lower stability due to the in-
6
14
prove the conductivity have drawn much attention. It has been reported
that by keeping the particles at nanoscale size, the rate performance
creased surface energy. More severe self-discharge has been found in
24
nano LFP of ∼25 nm compared with micro LFP of ∼2 μm. More-
7
–9
9
of LFP can be significantly improved. Kim et al. achieved LFP
nano-particles with a reversible capacity of 166 mAh/g and an excel-
over, impurity phases are reportedly easier to form during carbon coat-
25
ing on the surface of nano LFP, whose influence on electrochemical
8
26–29
lent rate capability of 50 mAh/g at 60C, while Bauer et al. achieved
properties is ambiguous in the literature.
The manufacturing cost
1
4,000 W/kg with 28% of the theoretical capacity preserved.
In order to understand the surprising improvement in kinetics of the
also increases with the reduction of particle size without sacrificing
phase purity. After weighing the pros and cons, an optimum particle
size for high-power applications has been suggested to be in the range
intrinsically insulating LFP material, the charge transport mechanism
and phase diagram of Li1-xFePO (0<x<1) have been extensively in-
vestigated. Due to the lack of continuous LiO octahedra in the direc-
tion of the a-axis and c-axis, lithium ions in the lattice of LFP can only
30
4
of 200–400 nm.
6
In this research, a new method to improve the rate performance
LFP electrode has been proposed. By making a mix of different par-
ticle sizes (ratio of nano and micro particles) of LFP with different Li
insertion/extraction mechanism in the electrode, the charge transfer
resistance can be significantly reduced leading to an enhancement of
high-power capability. The electrochemical performance of the op-
timized electrode has been carefully evaluated. The kinetics of the
mixed electrode has been discussed in detail.
10,11
migrate along the b-axis.
This one-dimensional diffusion channel,
which is easily blocked by crystal defects, has been blamed for the
low Li diffusion rate. Therefore, the ionic conductivity of LFP can be
enhanced by minimizing the particle size, as the reduced dimensions
of nano particles shorten the diffusion path of Li ions. Besides, the
Li migration through the channels in nano particles is easier than in
micro particles, because nano particles exhibit lower density of lattice
defects.1 It has been found that the solubility limit of Li in the LFP
structure is highly dependent on the particle size, where nano particles
2
13–15
Experimental
have higher solubility limit than micro particles.
The shrinking of
the miscibility gap has a strong influence on the phase transition of LFP
Synthesis.—The LFP samples with two different particle sizes
were synthesized using a solution-based method. Oxalic acid dihy-
drate (≥99%, Sigma-Aldrich) and Fe oxalate dihydrate (99%, Sigma-
Aldrich) were mixed in deionized water first. The molar ratios of oxalic
acid and Fe oxalate are 1.5:1 and 0.85:1 for the nano- and micro-sized
samples, respectively. 30 wt% H O was slowly added into the mix-
16
during charging and discharging. The result of Meethong et al. sug-
gested that the miscibility gap would completely disappear when the
particle size was below 15 nm at room temperature. It is believed that
the shrinking of the miscibility gap is responsible for the improvement
of the rate performance of LFP. In the meantime, the fracture caused
by the lattice mismatch of the two phases has been reported to be elim-
2
2
ture under magnetic stirring to dissolve Fe oxalate in accordance with
the following reaction,
∗
Electrochemical Society Member.
E-mail: jose.alarco@qut.edu.au
z
2FeC
2
O
4
·2H
2
O+C
2
H
2
O
4
·2H
2
O+2H
2
O
2
→Fe
2
(C
2
O
4
)3+8H
2
O+O
2
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
A4129
The temperature was controlled below 65°C during this process to
avoid the formation of impurities. Stoichiometric amounts of Li CO
PO (85 wt%) were added into the
solution. A small excess of Li (3 molar %) from the stoichiometric
amount, was added to the solution, which, from our experience, leads
to more reproducible, improved performance. 1% (based on the weight
of final product) of polyethylene glycol (PEG) was added to obtain
nano particles. The resultant clear green solution was drawn into a
reactor and reacted at 300°C for 1 hour under vacuum. The resultant
precursorswereringmilledandcalcinedat550°Cand710°Cfor1hour
light source. The power was controlled at 1mW to avoid damaging
the sample during measurement. The near-edge X-ray absorption fine
structure (NEXAFS) were obtained using soft X-ray absorption spec-
troscopy (sXAS) at the soft X-ray beamline of Australian Synchrotron
(AS). The NEXAFS spectra were simultaneously collected in total
electron yield (TEY), partial electron yield (PEY) and total fluores-
cence yield (TFY) modes with a step size of 0.1 eV. All the spectra
were normalized to I , which is a signal proportional to the X-ray
0
flux hitting the sample obtained with a gold mesh with about 90%
transmission, to get a flux independent measurement.
2
3
(
≥99%, Sigma-Aldrich) and H
3
4
4
under Ar atmosphere to form the nano and micro LiFePO particles.
The obtained powders were mixed with 8 wt% sucrose and calcined
at 710°C for 1 hour with Ar atmosphere for carbon coating.
Electrochemical testing.—The electrochemical properties of the
pristine materials and the blended samples were evaluated by con-
structing 2032 half cells. The slurry was prepared by mixing 90%
active material powder, 5% Super P, 5% PVDF (Arkema Kynar HSV
900 homopolymer) and NMP solvent (99.5%, Sigma-Aldrich). The
Characterization.—The structure and morphology of the powder
samples were characterized by X-ray diffraction (XRD), scanning
electron microscopy (SEM) and transmission electron microscopy
2
slurrywascastoncarboncoatedAlfoilwiththeloadingof ∼5mg/cm .
(
TEM). XRD was collected on a PANalytical X’Pert pro diffractome-
The 2032 coin cells were assembled inside a glove box with lithium
ter with Co-K radiation, over a 2θ range between 15° and 90° with
α
6
foil as anode and 1 M LiPF EC: DEC (1:1 by vol., Novolyte, BASF)
a 2θ step size of 0.017. The Rietveld refinement was conducted us-
ing HighscorePlus v4.8 software. The morphology and microstruc-
ture of the samples were investigated with a JEOL 7001 SEM and
JEOL 2100 TEM. The microstructure of the coated electrodes was
also characterized using SEM on polished cross-section embedded in
resin. The oxidation states on the surface of the samples was inves-
tigated with X-ray photoelectron spectroscopy (XPS, Kratos AXIS
Supra photoelectron spectrometer) using a focused monochromated
electrolyte. Galvanostatic and cyclic voltammetry testing were car-
ried out utilizing a battery test system (BioLogic, VMP-300) at room
temperature. Electrochemical impedance spectroscopy (EIS) was con-
ducted with a sine wave signal in the frequency range from 1MHz to
1 mHz and amplitude of 1 mV.
Results and Discussion
Al K
α
radiation (hν = 1486.6 eV). The spectra were calibrated using
Material characterization.—The XRD patterns and the Rietveld
refinements of the synthesized LFP samples are shown in Figure 1.
The patterns are entirely indexed as olivine LiFePO (ICDD: 98-016-
4
the C 1s peak at 285.0 eV. Carbon analyses were conducted using a
LECO TruMac CNS analyser. The furnace temperature for the sample
was 1300°C. Inductively coupled plasma (ICP) was used to analyze
the elemental compositions. 0.2g of each sample was digested in 4 mL
of a solution of 20 wt% of HCl and 20 wt% of HNO in deionized
3
water. Solid residues were filtered out and attributed to or associated
with undissolved carbon contents.
Particle surface characteristics were also examined with Raman
spectroscopy with a Renishaw inVia Microscope equipped with a
long working distance 50× objective lens and a 534 nm Ar+ laser
2282) with the space group Pnma, which confirms the phase purity
of the LFP samples. The crystallite sizes calculated with the Sherrer
equation are 65.0 nm and 42.3 nm for LFP samples calcined at 710 and
550°C respectively. The Rietveld refinement results are summarized
in Table I. As the existence of anti-site defect TMLi (transition metal
ions occupy Li site) has been reported present in olivine phosphates,3
the Li site has been left partially occupied with Fe ions while carrying
out the refinement. The lattice parameters of the as-prepared samples
–9
Figure 1. XRD and Rietveld refinement for the as-prepared LFP samples sintered at (a) 710 and (b) 550°C, respectively.
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A4130
Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
Table I. Rietveld refinement results for the as-prepared LFP
samples.
Table II. Mole ratios of lithium, iron and phosphorus obtained by
ICP and carbon content for the LFP samples.
Sinter temperature
710°C
550°C
Sample
Li:Fe:P
Carbon content/wt%
a/Å
b/Å
c/Å
α/°
β/°
10.3257
6.0062
4.6917
90
90
90
290.97
0.84
3.30
10.3190
6.0021
4.6919
90
90
90
290.597
1.98
2.23
LFP-350
LFP-50
1.06(5):1:0.97(3)
1.05(4):1:0.98(0)
1.91 ± 0.07
2.77 ± 0.03
times, estimated by approximating the particles to spheres, a much
thinner average carbon coating may be expected.
High-resolution XPS has been carried out to investigate the surface
state of the samples. As shown in the Fe-2p spectra in Figures 2c and
2f, the Fe ions on the surface exhibit mixed valence states for both
samples. Although the samples have slight Li excess, no peaks of im-
γ/°
3
Cell volume/ Å
FeLi/%
Rwp/%
RBragg/%
1.86
1.54
2+
purity phases are detected in the XRD patterns. The ratio of Fe and
3
+
Fe calculated from the peak area are about 50:50 (as summarized
in Table III), even though carbon is supposed to provide a protection
are in good agreement with previous studies,31 while the sample sin-
tered at 710°C exhibits slightly larger cell volume than that sintered
at 550°C. The sample sintered at 710°C has lower level of anti-site
defects suggesting a better crystallinity resulted from the higher tem-
perature calcination. The granular morphologies of the samples are
illustrated in Figure 2. Both samples have the granular morphology.
The average particle sizes are ∼350 and ∼50 nm for samples calci-
nated at 710 and 550°C respectively, indicating the sample sintered
at 710°C is polycrystalline. Both as-prepared LFP samples are well
crystallized with a coating layer of carbon (Figures 2b and 2d). The
as-prepared samples are labelled as LFP-350 and LFP-50 according
to their particle sizes respectively. The mole ratios of Li:Fe:P of the
samples obtained with ICP are shown in Table II. The ratio is typ-
3+
layer to avoid oxidation. The presence of Fe suggests the presence
of a Li-depletion layer around the particle in order to keep electri-
cal neutrality. Similar Li-depletion layer has also been observed in
carefully-prepared LFP, even when crystallized hydrothermally in
the presence of 3 times excess Li. The excess Li of the as prepared
samples may result in other defects preferably in the bulk, e.g. Li an-
tisites (Li on Fe sites), which is found to have low formation energy.33
29
Table III. Fitting results of the high-resolution XPS spectra for the
LFP samples.
2+
3+
Fe 2p
Fe 2p
32
ically within ± 4% of the expected 1:1:1. Both of the as prepared
samples show the slight Li excess, deliberately introduced. As summa-
rized in Table II, about 2.8% of carbon has been determined for LFP-
Sample
2p3/2
2p1/2
%
2p3/2
2p1/2
%
5
0, which is higher than the value for LFP-350 (about 1.9%). As the
LFP-350
LFP-50
710.8
710.7
724.2
724.0
49.8
48.8
713.2
713.2
726.9
726.8
50.2
51.2
2
LFP-50 has much larger surface area than LFP-350, about 49 ( = 7 )
Figure 2. SEM, TEM micrographs and high-resolution XPS spectra for (a-c) LFP-350 and (d-f) LFP-50.
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
A4131
Figure 3. NEXAFS for the LFP samples. (a) Fe L-edge and O K-edge for (b) LFP-350 and (c) LFP-50. The green, pink and blue lines in (b) and (c) are the O
K-edge collected with PEY, TEY and TFY mode, respectively.
The as-prepared samples are further investigated using
synchrotron-based sXAS. NEXAFS spectra are illustrated in
Figure 3. All the displayed spectra have been normalized to peak
maxima for comparison. The Fe L-edge spectra with TEY mode
are illustrated in Figure 3a. The Fe L-edge is considered stemming
from the dipole-allowed Fe-2p to Fe-3d transition, which probes
the unoccupied states of Fe-3d character. Therefore, the changes in
the peak shapes and energy positions can give an indication of the
oxidation state of the surface Fe. The Fe L-edge displays two groups
evidence of surface Li depletion, which is in good agreement with the
XPS results and earlier investigations.
44–46
Electrochemical performance of pristine materials.—The elec-
trochemical performance of the pristine LFP samples are evaluated
first. The charge and discharge behaviors of the two samples are illus-
trated in Figure 4. Both samples exhibit voltage plateaus at ∼3.45 V
at C/10 (1C = 170mA/g), suggesting the phase transition of LiFePO
4
and FePO during charge and discharge. LFP-50 shows a slightly
4
of peaks, labelled L
3
and L
2
, due to the core-hole spin-orbit-coupling
higher capacity (150 mAh/g) and lower polarization (∼49 mV) at
splitting of the 2p1/2 and 2p3/2 orbitals. As the two main peaks located
C/10 compared with LFP-350 (146 mAh/g and ∼64 mV). However,
LFP-50 starts to lose the initial flat voltage plateau and shows a slop-
ing voltage profile with the increase of the applied current when the
C-rate goes beyond C/5 (as shown in Figure 4c). Although a better
rate performance is generally expected for the samples with smaller
particle size, LFP-350 exhibits a higher capacity when the C-rate
is higher than 1C in this study (as shown in Figure 4a). Moreover,
a voltage overshoot has been noticed in the LFP-350 electrode at
the beginning of the charge and discharge in the whole measured
C-rate range in this study, and it gets more pronounced with in-
creasing C-rate, while no similar feature is observed for the LFP-50
samples.
2
+
3+
at 707.3 and 709.2 eV for Fe and Fe have different relative
34–36
intensity,
707/I709) can be associated to the oxidation state change of surface
Fe ions. According to Figure 3a, LFP-50 displays lower I707/I709 com-
the change of the intensity ratio of these two peaks
(
I
3
+
pared with LFP-350, suggesting more existence of Fe on LFP-50
particle surface. O K-edges of the as-prepared sample are illustrated
in Figures 3b and 3c. The O K-edge originates from the O-1s to O-2p
transition, while the pre-edge located at 531.7 eV is attributed to O-2p
weighting of states with predominantly Fe-3d character37,38 and the
main absorption is attributed to oxygen p character hybridized with
39,40
Fe 4s and 4p states.
Since the pre-edge peak has been reported to
be sensitive to the Li ion (de)intercalation,
41,42
the relatively higher
pre-edge intensity provides additional evidence for higher level of
surface Fe oxidation for LFP-50. By acquiring EY (TEY and PEY)
and TFY signals simultaneously, the information on the surface and
bulk can be obtained at the same time, as the fluorescent X-rays
and Auger electrons have different escape depths (∼3000 Å for
Electrochemical performance of mixed LFP electrodes.—The
Galvanostatic charge-discharge profiles of the pristine and mixed LFP
electrodes at C/10 are illustrated in Figure 5a. The mixed electrodes
are labelled with the ratio of LFP-50 and LFP-350. The mixed elec-
trodes exhibit voltage plateaus in between those of LFP-50 and LFP-
350 during charging, and higher than those of both pristine electrodes
during discharging. Among these, the electrode with a mix ratio of
50:50 shows the smallest polarization. Furthermore, the voltage over-
shoot for LFP-350 has been erased with the addition of LFP-50 par-
41
fluorescent X-rays and ∼50Å for Auger electrons). More surface
information can be obtained with the PEY mode because it is more
43
surface-sensitive compared with the TEY mode. Therefore, the
relatively higher pre-edge intensity in EY modes provides more
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A4132
Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
Figure 4. Galvanostatic charge and discharge of the pristine materials (a-b) LFP-350 and (c-d) LFP-50. (b) and (d) is the partial enlarged voltage plateau of (a)
and (c), respectively.
ticles. Most of the mixed electrodes show capacities between LFP-50
and LFP-350, while the electrode of 50:50 exhibits a specific capacity
boost (∼162 mAh/g).
lent circuit models. In this study, all the EIS spectra are acquired at
fully discharged state after the first three cycles at C/10. As displayed
in Figure 6a, all the spectra are composed of semicircles within the
high-frequency region and Warburg tails followed by sloping lines
within the low-frequency region. The impedance data are fitted with
the equivalent circuit illustrated in Figure 6c. The intersection with
A comparison of the rate capability has been conducted. The
voltage profiles at various C-rates are displayed in Supplementary
Figure 1 and the dependence of the specific discharge capacities on
discharge rate is depicted in Figure 5b. The addition of LFP-350 sig-
nificantly improves the rate performance of the LFP-50 dominated
electrodes with a slight initial capacity reduction. All the electrodes
with dominant LFP-350 show superior rate performance compared
with both pristine electrodes. In this research, the best rate perfor-
mance is achieved when LFP-50 and LFP-350 are blended 50:50 with
a specific capacity of 81 mAh/g at 10C. The large capacity drop
at 15C may be caused by the relatively high loading used in this
study and high equivalent series resistance resulting from the coin cell
configuration.
the real axis is the ohmic resistance of the cell (R
sum of the contributions from the current collectors, active material,
1
), which is the
47
electrolyte and separator. The resistors R
2 3
and R paralleled with
the constant phase element (CPE) account for the contact impedance
48
and charge transfer impedance, respectively. The ion diffusion in
49
w
the bulk olivine is described with the Warburg element (Z ). The
resistances obtained from the fittings are summarized in Table III. It
is worth noticing that both the contact resistance and charge transfer
resistance of LFP-350 are lower than those of LFP-50, indicating a
better packing and a more favorable surface for charge transfer has
been obtained with the LFP-350 samples. The mixing of the two LFP
particles brings down the contact resistance and the smallest contact
resistance is obtained with the mixing ratio of 50:50, suggesting the
best packing has been achieved, which is consistent with the results of
the density evaluation of the electrodes (summarized in Supplemen-
tary Table 1).
Comparison of kinetics for the electrodes.—Electrochemical
impedance spectroscopy (EIS) is an efficient tool for characteriza-
tion of electrochemical systems. The evaluations of the impedance
spectra, which contain information about the physico-chemical pro-
cesses inside the cells, are usually conducted by fitting with equiva-
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
A4133
Figure 5. (a) Galvanostatic charge-discharge profiles of the pristine and mixed electrodes at C/10. (b) Comparison of rate capabilities of the pristine and mixed
electrodes.
The diffusion coefficient of Li ion has been estimated with the
following equation,
where DLi is the chemical diffusion coefficient for Li ions, R is the
gas constant, T is the absolute temperature, A is the electrode area,
n is the number of electrons per molecule during oxidation, F is the
Faraday constant, C is the concentration of Li ion and σ is the Warburg
factor, which can be obtained from the slope of the real impedance vs
2
2
2
4
4
2 2
D
Li = R T /2A n F C σ
Figure 6. (a) EIS spectra, (b) plots of the real impedance as a function of the square root of angular frequency at low-frequency region and (c) the equivalent
circuit of the LFP electrodes.
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A4134
Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
Table IV. Resistances and Li diffusion coefficients obtained from
EIS results.
LFP-50 80:20 60:40 50:50 40:60 20:80 LFP-350
R1/ꢀ
R2/ꢀ
R3/ꢀ
DLi/ × 10
4.81
3.67
8.18
3.38 6.17 4.34 3.24 7.02 5.84
1.26 1.32 0.67 1.29 3.21 2.86
7.91 6.87 5.24 6.69 4.23 2.08
2.27 2.22 10.62 2.29 1.99 3.16
−
14
2
cm /s 1.25
−
1/2
ω
(ω is the angular frequency) in the Warburg region. The plots of
−1/2
the real impedance as a function of ω
and the linear fitting results
are illustrated in Figure 6b. The resulting Li diffusion coefficient are
summarized in Table IV. The DLi of pristine LFP-350 is higher than
thatofLFP-50, whichmayresultfromthebettercrystallinityandlower
level of anti-site defects (as shown in Table I) obtained with the higher
calcination temperature. Interestingly, by mixing the two pristine LFP
particles with the ratio of 50:50, the Li diffusion coefficient has been
improved by one order of magnitude. It indicates a modification on
the Li diffusion in LFP bulk has been achieved by particle mixing.
The increase of Li diffusion coefficient has been achieved along with
the decrease of the contact resistance, indicating the improvement of
the packing density facilitates the Li diffusion in the material. Such
improvement is not completely surprising since the Li diffusion is
an intricate process accompanied by the transport of the electronic
50
carriers to maintain electroneutrality.
Mechanism discussion.—A better rate capability has been
achieved with LFP-350 in this study, although the rate performance
is generally expected to be improved with particle size reduction. As
mentioned above, the thickness of carbon coating for LFP-50 is ex-
pected to be much thinner than that for LFP-350 due to the significantly
larger surface area. Therefore, it is likely that a more conductive and
homogeneously coated surface has been established on LFP-350 par-
ticles, resulting in the charge transfer resistance drop in the EIS and
better high rate performance. SEM micrographs of the polished cross-
sections of the electrodes made with the two different sizes of pristine
LFP and electrodes made with 50:50 mixture are displayed in Supple-
mentary Figure 2 and used to estimate the porosity. Lower porosity has
been achieved with the LFP-350 electrode, indicating a better pack-
ing compared with LFP-50 electrode. The result is consistent with the
density estimation (Supplementary Table S1) and the decrease of the
contact resistance can be attributed to the overall improved, electrical
connectivity. Furthermore, a lower level of anti-site defects for LFP-
Figure 7. Schematic of the LFP electrodes during charging. (a) LFP-50, (b)
LFP-350 and (c) mixed electrodes.
Therefore, by mixing the particles with different phase transi-
tion mechanisms, we can take the best of both sides. As the single-
phase Li insertion/extraction is faster than the two-phase Li inser-
tion/extraction, the particles with single-phase Li insertion can be the
reservoir of Li ion. Taking the charge process for explanation, as shown
in Figure 7c, the Li extraction at the beginning of charge would happen
in the small particles with single-phase insertion mechanism, which
eliminates the initial voltage overshoot. During the following charge,
charge transfer between the large and small particles would occur and
the two-phase Li extraction starts providing the flat voltage plateau.
In the meantime, the small particles which lost the Li ions to the large
particles can continuously get Li ions from the electrolyte. As a con-
sequence, the presence of the single-phase insertion particles in the
electrode can reduce the polarization caused by fast charge/discharge
and deliver improved rate performance.
350 compared with that of LFP-50 has also been confirmed with the
Rietveld refinements of XRD patterns. The anti-site defect has been
reported detrimental to the electrochemical performance by blocking
the one-dimensional Li diffusion channel.46,51–54 Consequently, from
all the combined factors just discussed, a better electrochemical per-
formance may have been achieved for LFP-350.
Conclusions
As shown in Supplementary Figure 2, the two kinds of particles
are uniformly distributed, and the porosity is decreased in the mixed
electrode, which suggests that a better contact among the particles
has been achieved, without limiting the access to the electrolyte. The
two as-prepared LFP particles exhibit different phase transitions dur-
ing charging and discharging, especially at high C-rates. Schematic
illustrations are shown in Figure 7. The sloping voltage profile of LFP
particles with ∼50 nm has been reported resulting from single-phase
In summary, the rate performance of the LFP electrode has been
improved by blending LFP particles of two different particle sizes and
Li insertion mechanisms. The optimization of the particle size distri-
bution offers better packing density and contact of the active material
particles. The improvement of packing results in better pathway for
electron transport and lower contact resistance. In the meantime, LFP
particles with single-phase Li insertion mechanism reduce the polar-
ization of the cell at high C-rate by acting as the reservoir of Li ions.
In this study, the electrode with a mixing ratio of 50:50 shows an im-
proved initial capacity at C/10 and superior rate capability compared
with the two pristine materials.
13
Li-insertion previously (as shown in Figure 7a), which is expected to
be favorable for rapid charge and discharge. However, the loss of the
voltage plateau sacrifices the energy density. Comparing with small
LFP particles, two-phase Li insertion is observed in larger LFP par-
ticles (∼350 nm in this study). The phase transition starts with a no-
ticeable voltage overshoot in the charge-discharge profile (Figure 7b).
The voltage overshoot has been reported as necessary to facilitate the
successive phase transition of large crystallites, as there is less specific
Acknowledgment
Yin Zhang would like to acknowledge CSIRO for the studentship,
andtheinvolvementofCRRCandtheRailManufacturingCooperative
Research Centre (funded jointly by participating rail organisations and
the Australian Federal Government’s Cooperative Research Centres
55
surface and fewer defects acting as nucleation sites. With the growth
of the new phase inside the particle, the cell voltage remains almost
independent with the intercalation of Li within the two-phase region.
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 166 (16) A4128-A4135 (2019)
A4135
Program). The experimental data reported in this paper were obtained
20. D. Li, X. Liu, and H. Zhou, Energy Technol., 2(6), 542 (2014).
2
1. T. Zhang, X. Liu, P. He, R. Peng, M. Wang, M. Han, and H. Zhou, J. Power Sources,
at the Central Analytical Research Facility operated by the Institute for
Future Environments, Queensland University of Technology (QUT),
Brisbane, Australia. The synchrotron based sXAS data reported in
this paper were obtained at the soft X-ray beamline of Australian Syn-
chrotron (AS). The authors would like to acknowledge Mark Quinlan
and Felix Lo, QUT, for the assistance with sample synthesis. The
authors would like to acknowledge Dr. Michael Jones, QUT, for the
assistance with the application for soft X-ray beamline at AS. The
authors are grateful to Prof. Gerbrand Ceder, UC Berkeley, for the
fruitful discussions on this topic at the 19th International Meeting on
Lithium Batteries (IMLB2018).
2
33, 299 (2013).
2. X. Guo, B. Song, G. Yu, X. Wu, X. Feng, Li, and Y. Chen, ACS Appl. Mater. Interfaces,
0(48), 41407 (2018).
23. C. M. Doherty, R. A. Caruso, andC. J. Drummond, Energy Environ. Sci., 3(6), (2010).
2
1
2
4. J. Wu, G. K. P. Dathar, C. Sun, M. G. Theivanayagam, D. Applestone,
A. G. Dylla, A. Manthiram, G. Henkelman, J. B. Goodenough, and K. J. Stevenson,
Nanotechnology, 24(42), 424009 (2013).
2
5. Y. Liu, J. Liu, J. Wang, M. N. Banis, B. Xiao, A. Lushington, W. Xiao, R. Li,
T. K. Sham, G. Liang, and X. Sun, Nature Communications, 9(1), 929 (2018).
6. Y.-H. Rho, L. F. Nazar, L. Perry, and D. Ryan, J. Electrochem. Soc., 154(4), A283
2
(2007).
2
2
7. S. Uchida, M. Yamagata, and M. Ishikawa, J. Power Sources, 243, 481 (2013).
8. E. Ershenko, A. Bobyl, M. Boiko, Y. Zubavichus, V. Runov, M. Trenikhin, and
M. Sharkov, Ionics, 23(9), 2293 (2017).
29. Y. Zhang, J. A. Alarco, A. S. Best, G. A. Snook, P. C. Talbot, and J. Y. Nerkar, RSC
Adv., 9(2), 1134 (2019).
0. W.-J. Zhang, J. Electrochem. Soc., 157(10), A1040 (2010).
1. J. J. Biendicho and A. R. West, Solid State Ionics, 203(1), 33 (2011).
ORCID
3
3
Yin Zhang https://orcid.org/0000-0003-4717-9899
Jose A. Alarco https://orcid.org/0000-0001-6345-071X
32. S. L. Bewlay, K. Konstantinov, G. X. Wang, S. X. Dou, and H. K. Liu, Mater. Lett.,
58(11), 1788 (2004).
3
3. K. Hoang and M. Johannes, Chem. Mater., 23(11), 3003 (2011).
4. S. Yang, D. Wang, G. Liang, Y. M. Yiu, J. Wang, L. Liu, X. Sun, and T.-K. Sham,
Energy Environ. Sci., 5(5), 7007 (2012).
3
References
35. X. Liu, D. Wang, G. Liu, V. Srinivasan, Z. Liu, Z. Hussain, and W. Yang, Nature
Communications, 4, 2568 (2013).
1
. J. D. Wilcox, M. M. Doeff, M. Marcinek, and R. Kostecki, J. Electrochem. Soc.,
54(5), A389 (2007).
36. T. K. Sham, Adv. Mater., 26(46), 7896 (2014).
37. R. Brydson, B. Williams, W. Engel, H. Sauer, E. Zeitler, and J. Thomas, Solid State
1
2
3
. L. Chen, Y.-Q. Yuan, X. Feng, and M.-W. Li, J. Power Sources, 214, 344 (2012).
. X.-C. Tang, L.-X. Li, Q.-L. Lai, X.-W. Song, and L.-H. Jiang, Electrochim. Acta,
Commun., 64(4), 609 (1987).
38. A. Augustsson, G. V. Zhuang, S. M. Butorin, J. M. Osorio-Guillen, C. L. Dong,
R. Ahuja, C. L. Chang, P. N. Ross, J. Nordgren, and J. H. Guo, The Journal of Chem-
ical Physics, 123(18), 184717 (2005).
39. P. Sarode, Journal of Physics F: Metal Physics, 17(7), 1605 (1987).
40. F. De Groot, M. Grioni, J. C. Fuggle, J. Ghijsen, G. A. Sawatzky, and H. Petersen,
Phys. Rev. B, 40(8), 5715 (1989).
41. J. McBreen, Journal of Solid State Electrochemistry, 13(7), 1051 (2009).
42. J. G. Lapping, S. A. Delp, J. L. Allen, J. L. Allen, J. W. Freeland, M. D. Johannes,
L. Hu, D. T. Tran, T. R. Jow, and J. Cabana, Chem. Mater., 30(6), 1898 (2018).
43. K. Nakanishi and T. Ohta, Surf Interface Anal, 44(6), 784 (2012).
44. A. S. Andersson, B. Kalska, L. Häggström, and J. O. Thomas, Solid State Ionics,
130(1), 41 (2000).
45. K. Zaghib, A. Mauger, F. Gendron, and C. M. Julien, Chem. Mater., 20(2), 462 (2008).
46. P. Axmann, C. Stinner, M. Wohlfahrt-Mehrens, A. Mauger, F. Gendron, and C. Julien,
Chem. Mater., 21(8), 1636 (2009).
54(8), 2329 (2009).
4
5
6
7
8
. S. Deng, H. Wang, H. Liu, J. Liu, and H. Yan, Nano-Micro Letters, 6(3), (2014).
. A. Yamada and S.-C. Chung, J. Electrochem. Soc., 148(8), A960 (2001).
. W.-J. Zhang, J. Power Sources, 196(6), 2962 (2011).
. S. Y. Chung, J. T. Bloking, and Y. M. Chiang, Nat. Mater., 1(2), 123 (2002).
. E. M. Bauer, C. Bellitto, M. Pasquali, P. P. Prosini, and G. Righini, Electrochem.
Solid-State Lett., 7(4), A85 (2004).
9. D.-H. Kim and J. Kim, Electrochem. Solid-State Lett., 9(9), A439 (2006).
1
0. D. Morgan, A. Van der Ven, and G. Ceder, Electrochem. Solid-State Lett., 7(2), A30
2004).
1. S.-I. Nishimura, G. Kobayashi, K. Ohoyama, R. Kanno, M. Yashima, and A. Yamada,
Nat. Mater., 7(9), 707 (2008).
2. R. Malik, D. Burch, M. Bazant, and G. Ceder, Nano Lett., 10(10), 4123 (2010).
3. P. Gibot, M. Casas-Cabanas, L. Laffont, S. Levasseur, P. Carlach, S. Hamelet,
J. M. Tarascon, and C. Masquelier, Nat. Mater., 7(9), 741 (2008).
(
1
1
1
47. D. Andre, M. Meiler, K. Steiner, C. Wimmer, T. Soczka-Guth, and D. U. Sauer, J.
1
1
1
1
1
1
4. G. Kobayashi, S.-I. Nishimura, M.-S. Park, R. Kanno, M. Yashima, T. Ida, and
A. Yamada, Adv. Funct. Mater., 19(3), 395 (2009).
5. Z. Li, J. Yang, C. Li, S. Wang, L. Zhang, K. Zhu, and X. Wang, Chem. Mater., 30(3),
Power Sources, 196(12), 5334 (2011).
48. Y. Zhu, Y. Xu, Y. Liu, C. Luo, and C. Wang, Nanoscale, 5(2), 780 (2013).
49. J. Gerschler and D. Sauer, Bruxelles, 2007.
50. N. Vicente, M. Haro, and G. Garcia-Belmonte, Chem. Commun., 54(9), 1025 (2018).
51. J. Chen and J. Graetz, ACS Appl. Mater. Interfaces, 3(5), 1380 (2011).
52. S.-Y. Chung, S.-Y. Choi, S. Lee, and Y. Ikuhara, Phys. Rev. Lett. 108, 195501 (2012).
53. Q. D. Truong, M. K. Devaraju, T. Tomai, and I. J. A. a. m. Honma, ACS. Appl. Mater.
Interfaces, 5(20), 9926 (2013).
54. J. J. Biendicho, K.-C. Hsiao, S. Hull, and A. R. West, Inorg. Chem., 56(6), 3657
(2017).
55. J. Jia, C. Tan, M. Liu, D. Li, and Y. Chen, ACS Appl. Mater. Interfaces, 9 (29), 24561
(2017).
874 (2018).
6. N. Meethong, H.-Y. S. Huang, W. C. Carter, and Y.-M. Chiang, Electrochem. Solid-
State Lett., 10(5), A134 (2007).
7. Y.-S. Yu, C. Kim, D. A. Shapiro, M. Farmand, D. Qian, T. Tyliszczak, A. D. Kilcoyne,
R. Celestre, S. Marchesini, and J. Joseph, Nano Lett., 15(7), 4282 (2015).
8. W. Dreyer, J. Jamnik, C. Guhlke, R. Huth, J. Moškon, and M. Gaberš cˇ ek, Nat. Mater.,
9(5), 448 (2010).
9. W. C. Chueh, F. El Gabaly, J. D. Sugar, N. C. Bartelt, A. H. McDaniel, K. R. Fenton,
K. R. Zavadil, T. Tyliszczak, W. Lai, and K. F. McCarty, Nano Lett., 13(3), 866 (2013).
Downloaded on 2019-12-13 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).