Nature of insulating-phase transition and degradation of structure and electrochemical reactivity in an olivine-structured material, LiFePO4

Article abstract of DOI:10.1021/ic9009114

Synthesis time using microwave irradiation was varied to elucidate the electrochemical degradation mechanism of LiFePO⁴ related to the evolution of Fe²P. When the amount of Fe₂P was above a critical level, LiFePO₄

Full text of DOI:10.1021/ic9009114

Inorg. Chem. 2009, 48, 8271–8275 8271
DOI: 10.1021/ic9009114
Nature of Insulating-Phase Transition and Degradation of Structure and
Electrochemical Reactivity in an Olivine-Structured Material, LiFePO4
Min-Sang Song,†, Yong-Mook Kang,*,‡ Yong-Il Kim,§ Kyu-Sung Park, and Hyuk-Sang Kwon*,†
†Department of Materials Science and Engineering, Korea Advanced Institute of Science and Technology,
373-1 Guseong-Dong, Yuseong-gu, Daejeon, 305-701, Republic of Korea, ‡Division of Advanced Materials
Engineering, Kongju National University, 275 Budae-dong, Cheonan, Chungnam, Republic of Korea,
§Korea Research Institute of Standards and Science, P.O. Box 102, Yuseong, Daejeon, Republic of Korea, and
Battery Group, Emerging Center, Samsung Advanced Institute of Technology, San 14-1 Nongseo-dong,
Giheung-gu, Yongin-si, Gyeonggi-do, 446-712, Republic of Korea
Received May 11, 2009
Synthesis time using microwave irradiation was varied to elucidate the electrochemical degradation mechanism of
LiFePO4 related to the evolution of Fe2P. When the amount of Fe2P was above a critical level, LiFePO4 tended to
change into an insulating phase, Li4P2O7. The correlation between structural analysis and electrochemical analysis
attributed the initial degradation of LiFePO4 to the low electronic conductivity of Li4P2O7, whereas the deficiency of
P and O evolved by Li4P2O7 resulted in the cyclic degradation of LiFePO4. This kind of correlation between structure
and electrochemical performance in intercalation materials will significantly contribute to an explanation of their
degradation mechanism for their application.
Introduction
(LiFePO4/C composite),6-11 particle size reduction,5,12,13
and supervalent cation doping.14 Actually, the electrochemi-
cal enhancement coming from these strategies made LiFePO4
a feasible cathode material for commercialization.
LiFePO4, with an olivine structure, has been spotlighted as
a promising candidate cathode material due to its excellent
thermal stability, the low cost of its precursors, the high rever-
sibility of Li insertion/extraction, and a lack of toxicity.1-3
Although LiFePO4 has an inherent poor kinetic property
(σe=10-8 S cm-1 (low electronic conductivity), DLiþ=10-14
In the early stage of LiFePO4 research, Fe2P, formed in a
strong reductive atmosphere, was considered as a byproduct
which should be prevented from forming during the synthesis
of single-phase LiFePO4.12 However, Subramanya Herle
et al. demonstrated that metal phosphocarbides or Fe2P
(metallic compound, σ: 10-1 S cm-1 at rt) can contribute to
improving the room-temperature conductivity of LiFePO4/C
drastically to ∼10-2 S cm-1 thanks to its high electronic
conductivity.15 Recently, Xu et al. verified the positive effect
of Fe2P on the electrochemical performance of LiFePO4 by
showing that LiFePO4 including Fe2P has the best rate
capability in spite of having the largest particle size.16 How-
ever, our previous study demonstrated that the formation of
Fe2P can result in amphoteric effects on the electrochemical
performance of LiFePO4 depending on its quantity.17
cm2 s-1 at room temperature (rt; low Liþ diffusivity)),4,5
a
marvelous improvement in its kinetic property was achieved
during the past decade with the help of carbon coating
*Corresponding author. Tel.: þ82-16-257-9051 (Y.M.K.). Fax: þ82-41-
568-5776 (Y.M.K.). E-mail: dake1234@kongju.ac.kr (Y.M.K.), hskwon@
kaist.ac.kr (H.S.K.).
(1) Giorgetti, M.; Berretoni, M.; Scaccia, S.; Passerini, S. Inorg. Chem.
2006, 45, 2750.
(2) Tarascon, J. M.; Armand, M. Nature 2001, 414, 359.
(3) Yamada, A.; Chung, S. C.; Hinokuma, K. J. Electrochem. Soc. 2001,
148, 224.
(4) Chung, S. Y.; Chiang, Y. M. Electrochem. Solid-State Lett. 2003, 6, 278.
(5) Prosini, P. P.; Carewska, M.; Scaccia, S.; Wisniewski, P.; Pasquali, M.
Electrochim. Acta 2003, 48, 4205.
(6) Gabrisch, H.; Wilcox, J. D.; Doeff, M. M. Electrochem. Solid-State
Lett. 2006, 9, 360.
(7) Ravet, N.; Chouinard, Y.; Magnan, J. F.; Besner, S.; Gauthier, M.;
Armand, M. J. Power Sources 2001, 97-98, 503.
(8) Chung, H. T.; Jang, S. K.; Ryu, H. W.; Shim, K. B. Solid State
Commun. 2004, 131, 549.
::
(12) Arnold, G.; Garche, J.; Hemmer, R.; Strobele, S.; Vogler, C.;
Wohlfahrt-Mehrens, M. J. Power Sources 2003, 119-121, 247.
(13) Prosini, P. P.; Carewska, M.; Scaccia, S.; Wisniewski, P.; Passerini,
S.; Pasquali, M. J. Electrochem. Soc. 2002, 149, 886.
(14) Chung, S. Y.; Bloking, J. T.; Chiang, Y. M. Nat. Mater. 2002, 1, 123.
(15) Subramanya Herle, P.; Ellis, B.; Coombs, N.; Nazar, L. F. Nat.
Mater. 2004, 3, 147.
(16) Xu, Y.; Lu, Y.; Yan, L.; Yang, Z.; Yang, R. J. Power Sources 2006,
160, 570.
(17) Song, M. S.; Kim, D. Y.; Kang, Y. M.; Kim, Y. I.; Kwon, H. S.; Lee,
J. Y. J. Power Sources 2008, 108(1), 546.
(9) Huang, H.; Yin, S. C.; Nazar, L. F. Electrochem. Solid-State Lett.
2001, 4, 170.
(10) Chen, Z.; Dahn, J. R. J. Electrochem. Soc. 2002, 149, 1184.
(11) Franger, S.; Bourbon, C.; Le Cras, F. J. Electrochem. Soc. 2004, 151,
1024.
r
2009 American Chemical Society
Published on Web 08/05/2009
pubs.acs.org/IC

8272 Inorganic Chemistry, Vol. 48, No. 17, 2009
Song et al.
Generally, in previous reports, LiFePO4 including Fe2P
below the critical amount displayed an enhanced electroche-
mical performance, whereas Fe2P inclusion above the critical
amount seriously deteriorated the electrochemical perfor-
mance of LiFePO4. The positive effect of Fe2P below the
critical amount has been easily ascribed to Fe2P-induced
enhancement of electronic conductivity in LiFePO4 because
Fe2P is electronically conductive. The reason why LiFePO4
with Fe2P above the critical amount has to undergo a
significant degradation is assumed to be related to the Liþ
pathway being blocked by an excessive formation of Fe2P in
LiFePO4. Considering that the application (HEV, plug-in
HEV, and EV) of LiFePO4 requires high rate capability, the
detailed change of the LiFePO4 structure with the variation
of the Fe2P amount should be clarified in the first place for its
commercialization.
Figure 1. The initial charge/discharge curves of LiFePO4 synthesized by
microwave irradiation for 2 min (red-colored line) and 4 min (blue-
colored line). Inset figure consists of Nyquist plots for 2-min-irradiated
LiFePO4 (red-colored sphere) and 4-min-irradiated LiFePO4 (blue-
colored triangle).
So, in this paper, the effect of Fe2P formation on the local
structure of LiFePO4 was correlated with the electrochemical
performance of LiFePO4. For this, the local structure of
LiFePO4 was comprehensively analyzed with variation of the
Fe2P amount.
analyzer (SI1255, solartron instruments) connected to a poten-
tiostat/galvanostat (273A, EG&G Princeton Applied Re-
search). A small AC perturbation of 10 mV was applied with a
frequency sweep from 0.1 Hz to 100 kHz during the measure-
ments. The amount of iron dissolved from LiFePO4 to the
electrolyte in a cell was measured with an inductively coupled
plasma mass spectrometer (ICP-MASS, Elan 6100, Perkin-
Elmer).
Experimental Section
Synthesis. LiFePO4 (including carbon web) was synthesized
by ball-milling and subsequent microwave heating. A stoichio-
metric amount (1:1, molar ratio) of Li3PO4 (Aldrich) and
Fe3(PO4)2 8H2O (Kojundo) was ball-milled at a ball-to-powder
3
weight ratio of 8.10:1 with 5 wt % acetylene black for 30 min
under an Ar atmosphere using a vibrant-type mill (SPEX-
8000 mixer/mill), and then the ball-milled mixture was micro-
wave irradiated (750 W) for several minutes in a microwave
oven.
Results and Discussion
The initial electrochemical reactivity of LiFePO4 (includ-
ing carbon web) synthesized from microwave irradiation is
displayed in Figure 1. In the discharge curve for LiFePO4
prepared by microwave heating for 2 min, a long, flat voltage
plateau appears around 3.41 V, and then the voltage sharply
drops to the cutoff value (2.5 V), leading to a high discharge
capacity of 165 mAh g-1 (97.1% of the theoretical capacity of
LiFePO4). On the other hand, LiFePO4 obtained by micro-
wave heating for 4 min shows shorter voltage plateaus,
followed by a low discharge capacity (97 mAh g-1). An
EIS analysis was carried out to investigate the change of
charge-transfer resistance in LiFePO4 with the variation of
microwave heating time. As shown in the inset of Figure 1,
the longer the microwave heating time for the synthesis of
LiFePO4, the larger the charge-transfer resistance between
the electrolyte and LiFePO4 surface. Because the increase of
microwave heating time was correlated with the evolution of
Fe2P, an extremely conductive phase (Figure S1, Supporting
Information), a surge of charge-transfer resistance in
the 4 min sample may imply that there is another phase
transition explanatory of the electrochemical degradation of
LiFePO4.
In order to clarify the reason why the 4 min sample
encountered difficulty in the charge transfer leading to its
electrochemical degradation, the local structure (lattice para-
meter, space of Liþ diffusion path, occupation of Fe atoms in
the Li site) of LiFePO4 was probed using X-ray Rietveld
refinement. Figure 2a shows the refinement results for LiFe-
PO4 irradiated for 2 min.17 Herein, two phases are observed,
0.56 wt % Fe2P and 99.44 wt % LiFePO4, and the lattice
parameters of LiFePO4 from the refinement (Table S1,
Supporting Information) are almost equal to those reported
Cell Fabrication. LiFePO4 cathodes were manufactured by
casting, on an Al foil current collector, a N-methyl-2-pyrroli-
dene slurry composed of 72 wt % LiFePO4-C, 20 wt %
acetylene black, and 8 wt % polyvinylidene fluoride binder. A
total of 2016 coin-type cells were assembled in an Ar-filled
glovebox by stacking a microporous polypropylene separator
(Celgard 2400) containing liquid electrolyte (1 M LiClO4 in 1:1
EC/DMC) between the cathode and the lithium metal foil
anode.
Structural and Electrochemical Analyses. Two- and three-
phase Rietveld refinements for LiFePO4 (including carbon web)
were conducted using the General Structure Analysis System
program.18 The X-ray diffraction (XRD) data for Rietveld
refinement were measured from 15° to 120° at a step of 0.02°
using Cu KR radiation with a graphite monochromator in the
reflection geometry (Dmax2200 V, Rigaku). Si (NIST 640c)
powder was used as an external standard to correct the zero-
point shift for the measured diffraction data. The coordination
number of atoms around the Fe atom in LiFePO4 was evaluated
by extended X-ray absorption fine structure (EXAFS, R-XAS)
spectroscopy. Fe K-edge X-ray absorption spectra were
recorded in the transmission mode at room temperature using
the lab-scale EXAFS machine equipped with a 3 kW X-ray
generator, Mo target, W filament, and Ge(220) crystal. The
X-ray absorption spectra were Fourier-transformed using the
IFEFFIT program based on FEFF8 ab inito theory. An ex situ
XRD analysis (D/max-III C, Rigaku) of the LiFePO4 (inclu-
ding carbon web) cathode after cycling was conducted from
15° to 45° at a scan rate of 1° min-1 using Cu KR radiation.
Electrochemical impedance spectroscopy (EIS) analysis was
conducted at room temperature using a HF frequency response
(18) Larson, A. C.; Von Dreele, R. B. Los Alamos National Laboratory
Report LAUR; Los Alamos National Laboratory: Los Alamaos, NM, 1994; p 86.

Article
Inorganic Chemistry, Vol. 48, No. 17, 2009 8273
Figure 2. X-ray Rietveld refinement results for LiFePO4 synthesized by microwave irradiation for (a) 2 min and (b) 4 min. Plus (þ) marks represent the
observed intensities, and the solid line illustrates calculated ones. A difference (obsd - cald) plot is shown beneath. Tick marks above the difference data
indicate the reflection position. In 2-min-irradiated LiFePO4, the upper and lower tick marks above the difference data indicate the reflection position for
Fe2P and LiFePO4 phases, respectively, and in 4-min-irradiated LiFePO4, the upper, middle, and lower tick marks correspond to Li4P2O7, Fe2P, and
LiFePO4, respectively.
in the literature,19-22 implying that the structure of LiFePO4
irradiated for 2 min is almost perfect. X-ray Rietveld refine-
ment for LiFePO4 irradiated for 4 min discloses a phase
transition responsible for the electrochemical degradation
observed in Figure 1. The refinement results for LiFePO4
irradiated for 4 min are given in Figure 2b. Three phases
were identified: 8.92 wt % Li4P2O7, 12.65 wt % Fe2P, and
78.43 wt % LiFePO4. The drastic increment in the charge-
transfer resistance of the 4 min sample can be explained by
the formation of Li4P2O7 because it has very low electronic
conductivity (∼ 10-20 S/cm at rt) and lithium-ion conductiv-
ity (∼ 10-21 S/cm at rt) compared to those of LiFePO4
itself.23 Actually, it has been reported that the evolution of
phase transition to Li4P2O7 is not mandatory when the
amount of Fe2P is below 1 wt % of LiFePO4. The compar-
ison between 4-min-irradiated LiFePO4 and 2-min-irradiated
LiFePO4 under the refined lattice parameters made us believe
that, irrespective of the formation of Li4P2O7, the formation
of Fe2P hardly changes the one-dimensional structure of
LiFePO4 (Table S2, Supporting Information). The difference
between the lattice parameters of two samples is within the
error range. Furthermore, the unit cell volume of LiFePO4 is
˚
almost in accordance with that (291.1 A) of LiFePO4 with the
perfect olivine structure regardless of the Fe2P amount.25
Even if the excessive Fe2P accompanied by the formation of
Li4P2O7 deteriorates the charge-transfer resistance of our
LiFePO4, Fe atoms seemed to be perfectly ordered in the
structure of LiFePO4, and thus, Liþ motion through the one-
dimensional (1D) path in LiFePO4 looked like it was not
disturbed, irrespective of the amount of Fe2P formation. In
order to examine the effect of Fe2P formation on the 1D Liþ
diffusion path composed of the linear chains of edge-shared
LiO6 octahedra, a change in the Li-O interatomic distance
was also estimated. The distance between Li and O atoms in
LiFePO4 was also within the error range, proving that the
excessive Fe2P formation does not have a mal effect on Liþ
diffusion path in LiFePO4.
Li4P2O7 deteriorates the lithium-ion conductivity of Li1þX
-
Ti2-XAlX(PO4)3 with a NASICON framework.24 An abrupt
formation of an insulating phase, Li4P2O7, in LiFePO4
irradiated for 4 min seems to be attributed to the excessive
formation of Fe2P. LiFePO4 exposed to highly reductive
atmosphere has to undergo the reduction of Fe below þ2,
inducing the formation of Fe2P. Because Fe2P is an electro-
nically conductive phase, the formation of Fe2P below a
critical amount tends to help the electrochemical perfor-
mance of LiFePO4 become enhanced. However, an Fe2P
amount above the critical level is accompanied by a defi-
ciency of Fe and P atoms constituting the framework of
LiFePO4. Naturally, LiFePO4 should be changed into a
mixture of several phases according to the following reaction:
On the basis of the structural parameters obtained from
X-ray Rietveld refinement, it seemed that the local structure
of LiFePO4 is not influenced by Fe2P formation. However,
considering that the peak intensity in XRD patterns of
LiFePO4 is dependent on the amount of Fe2P, as shown in
Figure 2, the local structure of LiFePO4 needed to be probed
in detail. Therefore, an Fe-edge EXAFS analysis was con-
ducted to determine the coordination of atoms around the Fe
atom. Figure 3 displays the Fe-edge EXAFS spectra, where
the peak intensity is proportional to the coordination number
of atoms around the Fe atom. From the previous report, it
could be known that the first sphere, second sphere, and third
sphere correspond to the Fe-O bond, Fe-P bond, and Fe-
Fe bond, respectively. The EXAFS spectra for 2-min-irra-
diated LiFePO4 are very similar to those of LiFePO4 having
the perfect olivine structure.26 On the other hand, several
9
4LiFePO4 f Li4P2O7 þ 2Fe2P þ O2
2
Considering that the refinement simulation assuming three
phases, Li4P2O7, Fe2P, and LiFePO4, was appropriatefor not
the 2 min sample but the 4 min sample, we can know that the
::
::
(19) Andersson, A. S.; Kalska, B.; Haggstrom, L.; Thomas, J. O. Solid
State Ionics 2000, 130, 41.
(20) Cho, T. H.; Chung, H. T. J. Power Sources 2004, 133, 272.
(21) Prince, A. A. M.; Mylswamy, S.; Chan, T. S.; Liu, R. S.; Hannoyer,
B.; Jean, M.; Shen, C. H.; Huang, S. M.; Lee, J. F.; Wang, G. X. Solid State
Commun. 2004, 132, 455.
(22) Scarlet, N. V. Y.; Madsen, I. C.; Cranswick, L. M. D.; Lwin, T.;
Groleau, E.; Stephenson, G.; Aylmore, M.; Agron-Olshina, N. J. Appl.
Crystrallogr. 2002, 35, 383.
(23) Burmakin, E. I.; Shekhtman, G. S.; Aparina, E. R.; Korovenkova, E.
S. Sov. Electrochem. 1992, 28, 1014.
(24) Arbi, K.; Mandal, S.; Rojo, J. M.; Sanz, J. Chem. Mater. 2002, 14, 1091.
(25) Chen, J.; Whittingham, M. S. Electrochem. Commun. 2006, 8, 855.
(26) Deb, A.; Bergmann, U.; Cramer, S. P.; Cairns, E. J. Electrochim.
Acta 2005, 50, 5200.

8274 Inorganic Chemistry, Vol. 48, No. 17, 2009
Song et al.
serious changes are shown in the atomic coordination num-
bers of 4-min-irradiated LiFePO4. This phenomenon is
because the deficiencies of O, P, and Fe atoms are significant
in 4-min-irradiated LiFePO4, which has to undergo a phase
transition to a mixture of Fe2P and an insulating phase,
Li4P2O7. Because PO4 tetrahedra are fundamental building
blocks for the olivine structure of LiFePO4, a crucial struc-
tural instability can be induced by the deficiency of P and O
atoms. With the absence of PO4 tetrahedra, the structure of
LiFePO4 must collapse during the charge/discharge process.
Figure 4 shows XRD patterns of LiFePO4 before cycling and
after 50 cycles. Even after 50 cycles, the initial structure of
2-min-irradiated LiFePO4 was relatively well-maintained,
whereas that of the 4-min-irradiated sample underwent a
serious degradation. In accordance with EXAFS results, the
structural instability of LiFePO4 during cycling could be
attributed to the phase transition to the mixture of Fe2P and
an insulating phase, Li4P2O7. In Figure 5, it is shown how this
structural instability is correlated with the electrochemical
degradation of LiFePO4.
Figure 5a illustrates the electrochemical superiority of
2-min-irradiated LiFePO4 compared to the 4 min sample.
The 2 min sample featured a fabulous cyclic performance in
spite of the variation of the charge/discharge rate, while the
Figure 3. The radial distribution function obtained after Fourier transfor-
mation of k3χ(k) in the k3-weighted Fe-edge EXAFS spectrum. First sphere
(1.7-2.2 A), second sphere (2.7-3.2 A), and third sphere (3.7-4.2 A)
correspond to the Fe-O bond, Fe-P bond, and Fe-Fe bond, respectively.
˚
˚
˚
Figure 4. XRD patterns of LiFePO4 before and after cycling: (a) 2 min microwave-irradiated sample, (b) 4 min microwave-irradiated sample.
Figure 5. (a) The cyclic properties of LiFePO4 synthesized by microwave irradiation for 2 min (red-colored line) and 4 min (blue-colored line). Nyquist
plots for (b) 2-min-irradiated LiFePO4 and (c) 4-min-irradiated LiFePO4 before or after cycling.

Article
Inorganic Chemistry, Vol. 48, No. 17, 2009 8275
4 min sample exhibited poor cyclic performance even at a
very sluggish charge/discharge rate (C/10). This tendency in
the cyclic performance was consistent with those in the
charge/discharge capacity (Figure 1) and the rate capability
(Figure S2, Supporting Information). As shown in Figure 5b
and c, Nyquist plots obtained during the initial three cycles
may correlate the electrochemical performance of LiFePO4
with its local structure. The increase in the charge-transfer
resistance of LiFePO4 after cycling mainly comes from Fe
dissolution into the electrolyte because the Fe ion dissolved
from LiFePO4 goes to an anode and then increases the
interfacial impedance of the anode significantly by forming
an insulating film on the surface of the anode, as reported by
Amine et al.27 From ICP-MASS analysis (Table S3, Support-
ing Information), the increment of Fe dissolution is perfectly
in accordance with the above-mentioned structural instabil-
ity coming from the increased Fe2P formation followed by a
phase-transition to Li4P2O7. Herein, Fe2P itself is not the
fundamental reason for the increase in the charge-transfer
resistance of LiFePO4 during cycling, even if the amount of
Fe ions present in the electrolyte is proportional to the
amount of Fe2P. Considering that the local structure change
of LiFePO4 depends on the evolution of Fe2P, it seems to be
more reasonable that the electrochemical degradation indi-
cated by the increase of charge-transfer resistance is attrib-
uted to the deficiency of P and O evolved by the phase
transition into a mixture of Fe2P and Li4P2O7.
Conclusion
In conclusion, the evolution of Fe2P in LiFePO4 above a
critical concentration changed LiFePO4 into an insulating
phase, Li4P2O7. The insulating nature of Li4P2O7 leads to the
initial degradation of LiFePO4 in spite of the existence of a
conducting phase, Fe2P. The deficiency of P and O in the
LiFePO4 framework evolved by the phase transition to the
mixture of Fe2P and Li4P2O7 resulted in the structural instabi-
lity of LiFePO4, which deteriorated its cyclic performance. As a
result, the phase transition into Li4P2O7 induced by excessive
Fe2P formation brought out a dramatic degradation in the
electrochemical properties of LiFePO4, and this degradation
mechanism of LiFePO4 suggests the criteria for the synthesis of
high-performance LiFePO4 in a reductive atmosphere.
Acknowledgment. The authors acknowledge the Korea
Science and Engineering Foundation (KOSEF) grant
funded by the Korean government (MEST) (R01-2008-
000-11061-0). This research was also supported by the Basic
Science Research Program through the National Research
Foundation of Korea (NRF) funded by the Ministry of
Education, Science and Technology (2009-0072972) and a
grant (2009K000186) from the Center for Nanoscale Me-
chatronics and Manufacturing, one of the 21st Century
Frontier Research Programs, which are supported by the
Ministry of Education, Science and Technology, Korea.
Supporting Information Available: Additional figures and
tables. This material is available free of charge via the Internet
at http://pubs.acs.org.
(27) Amine, K.; Liu, J.; Belharouak, I. Electrochem. Commun. 2005, 7,
669.