Melt Casting LiFeP O4: II. Particle Size Reduction and Electrochemical Evaluation

Article abstract of DOI:10.1149/1.3284506

LiFeP O₄ was prepared using a melt casting technique of Li ₂ C O₃ and FeP O₄ precursors at 1000°C. The product was characterized by X-ray diffraction and is of the olivine structure with a minor amount of Li₄ P₂ O₇ impurity. The synthesis, based on a molten procedure, provides a route to large-scale synthetic practices and reduced cost through the use of inexpensive precursors and short reaction times. The large particle sizes of the LiFeP O₄ crystals obtained from the melt casting were reduced to 200 nm by a series of successive milling techniques without affecting the purity of the sample. A subsequent carbon coating on this milled material with a variety of carbon precursors was capable of producing samples with high capacities and electrochemical results similar to that of commercial LiFeP O₄ powders. These results indicate that the melt casting procedure could be a competitive synthetic technique for the large-scale production of LiFeP O ₄.

Full text of DOI:10.1149/1.3284506

Journal of The Electrochemical Society, 157 ͑4͒ A463-A468 ͑2010͒
A463
0013-4651/2010/157͑4͒/A463/6/$28.00 © The Electrochemical Society
Melt Casting LiFePO4
II. Particle Size Reduction and Electrochemical Evaluation
,z
D. D. MacNeil,a, L. Devigne,a C. Michot,b I. Rodrigues,a G. Liang,b, and
*
*
M. Gauthierb
aDépartement de Chimie, Université de Montréal, Montréal, Québec H3T 1J4, Canada
bPhostech Lithium, Inc., St-Bruno de Montarville, Québec J3B 6V7, Canada
LiFePO4 was prepared using a melt casting technique of Li2CO3 and FePO4 precursors at 1000°C. The product was characterized
by X-ray diffraction and is of the olivine structure with a minor amount of Li4P2O7 impurity. The synthesis, based on a molten
procedure, provides a route to large-scale synthetic practices and reduced cost through the use of inexpensive precursors and short
reaction times. The large particle sizes of the LiFePO4 crystals obtained from the melt casting were reduced to 200 nm by a series
of successive milling techniques without affecting the purity of the sample. A subsequent carbon coating on this milled material
with a variety of carbon precursors was capable of producing samples with high capacities and electrochemical results similar to
that of commercial LiFePO4 powders. These results indicate that the melt casting procedure could be a competitive synthetic
technique for the large-scale production of LiFePO4.
© 2010 The Electrochemical Society. ͓DOI: 10.1149/1.3284506͔ All rights reserved.
Manuscript submitted October 15, 2009; revised manuscript received December 7, 2009. Published February 26, 2010.
samples with a variety of carbon precursors allows the achievement
of a high electrochemical capacity and stability with extended
elevated-temperature charge–discharge cycling.
LiFePO4 has received a large amount of attention as a positive
electrode material for lithium-ion batteries.1-4 It represents a low
cost, thermally stable, and environment-friendly substitute for
cobalt-based lithium metal oxides, which are currently used as cath-
odes in the batteries of portable computers and cellular phones. In
addition, the realization of high power batteries that have large ca-
pacity at high rates of charge and discharge is possible with small
particle, carbon-coated LiFePO4.5 These properties have led to its
consideration as the cathode material for the storage battery of
plug-in hybrid electric vehicles.6 The electrochemical performance
of this material is strongly dependent on the synthesis method. Thus
far, most commercial LiFePO4 products are synthesized by the use
of solid-state chemistry methods,7,8while a large amount of work is
devoted to the development of hydrothermally prepared LiFePO4
due to the smaller particle size obtained through this method.9,10
These two methods have drawbacks in that they either require a
lengthy process or costly precursors. In Part I, we introduced a
method of LiFePO4 synthesis that has the capability to provide large
amounts of sample from a variety of precursors and a short reaction
time.11,12 This melt casting procedure uses typical metallurgical syn-
thesis methods and can easily provide samples in kilogram quanti-
ties. In Part I of this two-part series, we described the synthesis and
structural characterization of LiFePO4 from melt casting and dem-
onstrated that pure LiFePO4 can be obtained from a variety of pre-
cursors. We were also able to demonstrate samples with metal sub-
stitutions, such as Mn and Mg.
Although the melt casting procedure does provide a pure, elec-
trochemically active material, the large particle size of the synthe-
sized sample results in inferior electrochemical properties, i.e., a low
capacity compared to its theoretical value as well as capacity loss
with extended charge–discharge cycling. Particle size reduction
helps reduce the mass transfer resistance within the LiFePO4 par-
ticle, while a carbon coating improves electrical conductivity. In the
present study, we detail the steps involved in the preparation of
LiFePO4 samples from a melt casting process with electrochemical
properties that are comparable to those of commercial solid-state
and hydrothermally prepared samples of LiFePO4. A continuous wet
milling procedure is required to efficiently lower the particle size of
the melt-cast samples to nanometer dimensions in which good elec-
trochemical properties are possible. Carbon coating the milled
Experimental
LiFePO4 samples were prepared by a melt casting process as
described in Ref. 12. Briefly, stoichoimetric amounts of
FePO4·2H2O ͑Buddenheim KG, Germany͒ and Li2CO3 ͑Limtech,
Québec͒ were combined with 25 mol % of graphite powder ͑Tim-
cal, Belgium͒. The mixture was then placed in a graphite crucible
and heated to 1000°C under an Ar flow for 1 h. After cooling to
room temperature, the sample was broken with a hammer. A rough
grinding was performed with a disk mill ͑Retsch DM 200͒. The
disk-milled material ͑particle size ϳ100 ␮m͒ was then placed in a
planetary mill ͑Fritsch͒ with a 5 times excess of isopropyl alcohol
͑IPA͒ for suspension. The planetary mill used a 250 mL Syalon
container with three 25 mm Syalon balls. A 90 min milling proce-
dure was performed. The sample was then collected and further
processed in a continuous-flow agitator bead mill ͑MiniFer by
Netszch͒. The milling procedure consisted of preparing a 20% ͑by
weight͒ suspension of LiFePO4 in IPA. This solution was then
passed through the mill containing 140 mL yttria-stabilized zirconia
grinding media operating at a speed of 1500 rpm. Two different
sized media were used for milling: ͑i͒ 0.7–0.9 mm beads ͑nanomill
no. 1͒ and ͑ii͒ 0.5–0.7 mm beads ͑nanomill no. 2 and no 3͒. The mill
ran continuously until a stabilized particle size was obtained. The
particle size was continuously monitored during the milling proce-
dure by the use of a Horiba particle size analyzer ͑LA 300͒. After
particle size stabilization, the sample was collected from the mill
and kept in suspension until further use.
Carbon-coated LiFePO4 samples were prepared by dissolving a
carbon precursor in either IPA or H2O. The precursors used in this
paper were ␤-lactose, L-ascorbic acid, poly͑maleic anhydride-alt-1-
octadecene͒ ͑PMAAO͒, salicylic acid, and hydroxyethyl cellulose.
The carbon precursor suspension was then added to the LiFePO4
suspension and the mixture was allowed to evaporate to dryness.
The powder was then heated to 700°C for 1 h under a N2 flow.
X-ray diffraction ͑XRD͒ was performed with a Bruker D8 Advance
equipped with a Cu X-ray tube and a diffracted-beam monochro-
mator. Scanning electron microscopy ͑SEM͒ micrographs were car-
ried out with a Hitachi S-4300 microscope. Chemical analysis was
performed with a Fisons Instruments ͑SPA, model EA1108͒ elemen-
tal analyzer to determine the C, H, N, and S concentration within the
sample.
Commercial carbon-coated LiFePO4 samples were obtained
from Phostech Lithium, Inc. Samples of both energy grade ͑P1͒ and
*
Electrochemical Society Active Member.
z E-mail: dean.macneil@umontreal.ca
Downloaded on 2014-10-30 to IP 128.114.34.22 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A464
Journal of The Electrochemical Society, 157 ͑4͒ A463-A468 ͑2010͒
Figure 2. SEM micrographs of LiFePO4 from a melt casting synthesis of
͑scale bar dimension indicated in brackets͒ ͑a͒ disk-milled product ͑20 ␮m͒,
͑b͒ planetary-milled product ͑20 ␮m͒, ͑c͒ nanomilled product with
0.7–0.9 mm grinding media ͑1 ␮m͒, and ͑d͒ nanomilled product with
0.5–0.7 mm grinding media ͑2 ␮m͒.
Figure 1. ͑Color online͒ XRD of LiFePO4 after ͑a͒ disk milling, ͑b͒ plan-
etary milling, and ͑c͒ nanomilling. Indicated peaks on scan ͑a͒ are from a
minor impurity of Li4P2O7.
sis using Scherrer’s formula gives an average crystallite size of
99 nm for the disk-milled sample, 98 nm for the planetary-milled
sample, and 60 nm for the nanomilled sample. Also, during the vari-
ous milling procedures there is no introduction of new impurity
species compared to the original synthesized material.
power grade ͑P2͒ LiFePO4 were provided and used as received. The
energy grade product, P1, was produced following a solid-state syn-
thesis method described in Ref. 14, while the power grade product,
P2, was produced via a hydrothermal synthesis method described in
Ref. 15. Electrochemical evaluation of the nanomilled and commer-
cial samples was performed by combining 11% of a conductive
carbon ͑Timcal EBN1010͒ and 7% polyvinylidene difluoride ͓3.3%
in n-methyl pyrrolidone ͑NMP͔͒ with the electrode powders. An
extra portion of NMP was added to the mixture to form a slurry,
which was then mixed overnight in a roller mill. The slurry was then
coated on a piece of carbon-coated Al foil. The electrode was dried
overnight at 60°C under vacuum. The next day, 13 mm diameter
disks were punched for cell assembly in standard 2032 coin-cell
hardware ͑Hohsen͒ with a single lithium-metal foil used as both
counter and reference electrodes. Cells were assembled in an argon-
filled glove box with 1 M LiClO4 in ethylene carbonate ͑EC͒/
dimethyl carbonate ͑1/1͒ ͑Ferro͒ or 1 M LiPF6 in EC/diethyl car-
bonate ͑3/1͒ ͑UBE͒ as the electrolyte. Electrochemical evaluations
were performed with a VMP electrochemical station ͑Biologic,
France͒. The cells were either evaluated for rate capability ͑with
LiClO4 electrolyte at room temperature͒ or for 60°C charge–
discharge cycle stability ͑with LiPF6 electrolyte͒. A sequence of
discharging the cell at a C/12 rate and then sequentially up to 10C
following a C/4 charge was used to test the rate capability of the
sample. The charge–discharge cycling stability tests were performed
by charging and discharging between 2.2 and 4 V at a rate of C/4 at
60°C. After 10 cycles at C/4, one C/8 charge–discharge cycle was
performed to test the low rate performance of the cell, after which
the C/4 sequence was repeated.
The size and morphology of the LiFePO4 particles were investi-
gated by SEM. Figure 2 shows the micrographs of the LiFePO4
sample after different milling procedures. The disk-milled product
͑Fig. 2a͒ contains large rectangular crystallites with well-defined and
smooth faces. As this product is milled in the planetary mill ͑Fig.
2b͒, the large crystallites get reduced and the morphology of the
crystals becomes more jagged and less defined. There is an overall
particle size reduction, but it comes mainly from the destruction/
milling of the larger particles while the small particles remain
largely untouched. During the nanomilling process ͑Fig. 2c and d͒
the large crystallites get reduced to smaller particles, whose final
particle size is dependent on the media size used during the milling
procedure. The final milled samples contain a few large particles of
1–2 ␮m and many smaller submicrometer-sized particles. The mor-
phology for the nanomilled samples is similar and consists of irregu-
larly shaped particles with sharp facets, which clearly demonstrates
the grinding process on the particles during the milling procedure.
Throughout the milling procedure, the particle size was continu-
ously monitored by laser diffraction. There was a clear decrease in
particle size with milling time until a point where the particle size
stabilized due to the loss of grinding capability of the media. At this
point, the milling operation was terminated and the sample was
collected. The particle size analysis of the final milled samples is
shown in Fig. 3. The particle size of the starting disk-milled sample
͑average particle size ϳ100 ␮m͒ was reduced initially to ϳ10 ␮m
through planetary milling and then to ϳ400 or ϳ200 nm during the
wet milling procedure, with the smaller sized media ͑0.5–0.7 mm͒
giving the smallest particles. Further particle size reduction is likely
possible with the use of smaller milling media. The nano- and
planetary-milled samples, which are represented in Fig. 3, were used
for further electrochemical evaluation. It is clear that as the sample
is nanomilled with smaller media ͑0.5–0.7 mm͒, the average particle
size is reduced and the particle size distribution significantly im-
proves. This results in a sample that is much more homogeneous.
Particle size reduction is an important characteristic for the achieve-
ment of good electrochemical performance from LiFePO4 as the
material has a very low intrinsic ionic conductivity.13
Results and Discussion
Figure 1 presents the XRD patterns of the LiFePO4 sample after
various milling procedures. The main reflections in all samples can
be indexed on an orthorhombic olivine-type structure with the Pnma
space group. There is a small amount of Li4P2O7 impurity in the
disk-milled sample. The sharp diffraction peaks in the disk-milled
sample are indicative of a highly crystalline product. In fact, the
sample after synthesis shows large green crystals that can extend
over several centimeters. It is clear from the width of the diffraction
peak and the loss of diffraction intensity in Fig. 1 that, through the
various milling procedures, the crystallite size diminishes. An analy-
Another important aspect for the development of LiFePO4 with
high power capabilities is the production of an effective carbon coat-
ing on the surface of the LiFePO4 particles. The carbon coating
Downloaded on 2014-10-30 to IP 128.114.34.22 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, 157 ͑4͒ A463-A468 ͑2010͒
A465
significantly improves the electrical conductivity of the material. In
this evaluation, several carbon precursors were tested. Table I pro-
vides a list of the precursors, the added concentration, and the re-
sultant carbon content of the sample after heating to 700°C for 1 h
under a N2 atmosphere. The morphologies of a select number of
samples are shown in Fig. 4; similar results are obtained from the
other samples but are omitted from Fig. 4 for clarity. In Fig. 4a and
b, nanomilled samples heated to 700°C without a carbon precursor
are shown. They clearly demonstrate that heating the nano-LiFePO4
material at an elevated temperature in the absence of a carbon pre-
cursor causes particle growth of a sample. The size of the heated
particles has increased almost 100 times from its precursor. There is
no doubt that the electrochemical characteristics of this particle
would be insufficient for application as a battery material. In com-
parison, the samples, shown in Fig. 4c-f have a particle size roughly
equal to that of the precursor LiFePO4 milled material shown in Fig.
2. The particle size deduced from Fig. 4 suggests that the process of
carbon coating LiFePO4 does not significantly change the particle
size of the underlying olivine material. A nano-LiFePO4 sample
coated with a carbon precursor inhibits the particle growth that nor-
mally occurs during high temperature annealing ͑Fig. 4a and b͒. In
addition, the process of carbon coating smooths out the morphology
of the nanomilled samples to produce particles with a much more
uniform circular shape. This variation in morphology is seen via a
change in the surface area of the material. The pre-carbon-coated
nanomilled LiFePO4 sample has a surface area of 12.2 m2/g, while
the carbon-coated material has a surface area of 8.2 m2/g.
Figure 3. ͑Color online͒ Particle size analysis of the various LiFePO4
samples under study: ͑a͒ Disk milled, ͑b͒ planetary milled, ͑c͒ nanomilled
no. 1, ͑d͒ nanomilled no. 2, and ͑e͒ nanomilled no. 3.
Table I. Summary of experimental results on melt-cast LiFePO4.
Rate capability
60°C cycling at C/4
͑capacity at indicated
discharge rate,
mAh/g͒
Discharge capacity
Cycle
Precursor
amount
͑%͒
͓Carbon͔
after heat
͑%͒
Calculated capacity
loss for 100
LFP milling
procedure
Carbon
precursor
Initial
͑mAh/g͒
40
͑mAh/g͒
C/12
C
4C
10C
20C
cycles ͑%͒
Planetary mill
␤-Lactose
L-Ascorbic
PMAAO
5
5
5
1.1
1.0
0.6
113
103
108
82
73
74
60
50
52
40
30
34
23
0
13
121
103
112
98
81
81
61
56
77
Hydroxyethyl
cellulose
Salicylic
5
5
0.7
0.9
108
58
71
29
47
18
30
9
0
0
111
53
91
32
56
81
Nano 1 ͑0.7–0.9 mm͒
Nano 2 ͑0.5–0.7 mm͒
Salicylic
PMAAO
Hydroxyethyl
cellulose
5
5
0.6
0.9
151
142
123
120
82
89
46
66
20
35
155
150
143
144
19
9
5
0.8
150
129
92
52
15
153
146
12
␤-Lactose
L-Ascorbic
PMAAO
5
5
5
1.3
1.1
0.9
152
154
149
138
140
134
116
116
109
90
83
78
58
35
34
156
149
154
148
144
148
14
11
11
Hydroxyethyl
cellulose
Salicylic
5
5
0.9
0.9
153
151
139
135
115
108
88
78
51
43
161
158
158
152
6
10
Nano 3 ͑0.5–0.7 mm͒
PMAAO
Salicylic
5
0.7
0.9
1.3
1.0
1.0
1.5
149
151
149
152
149
150
134
140
136
138
132
138
105
112
111
109
105
116
67
78
85
75
79
93
17
36
47
38
48
63
153
156
155
158
157
157
151
152
149
150
149
150
5
6
7
13
13
10
10
15
5
10
15
Hydroxyethyl
cellulose
5
1.4
150
138
110
78
25
158
150
13
10
15
5
10
15
1.8
2.4
1.1
2.0
3.0
151
156
156
151
142
138
142
143
138
126
112
121
122
116
103
88
98
101
90
65
68
65
59
47
156
155
159
158
153
153
152
152
153
152
9
7
10
7
␤-Lactose
77
2
Downloaded on 2014-10-30 to IP 128.114.34.22 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A466
Journal of The Electrochemical Society, 157 ͑4͒ A463-A468 ͑2010͒
Figure 6. ͑Color online͒ Capacity vs cycle number of PMAAO-coated
LiFePO4 samples from various milling procedures and carbon contents. Cy-
cling was performed at 60°C at a discharge rate of C/4 for 10 cycles, fol-
lowed by one cycle at C/8. Stars: planetary mill; cross: nanomill no. 1 dia-
monds: nanomill no. 2; squares: nanomill no. 3 5%, circles = nanomill no. 3,
10%; and triangles: nanomill no. 3, 15%.
Figure 4. SEM micrographs of carbon-coated LiFePO4 ͑scale bar indicated
in brackets͒: ͑a͒ Uncoated nanomilled no. 1 product heated to 700°C
͑10 ␮m͒, ͑b͒ uncoated nanomilled no. 2 product heated to 700°C ͑10 ␮m͒,
͑c͒ PMAAO-coated planetary-milled product ͑5 ␮m͒, ͑d͒ PMAAO-coated
cations. An application requiring high energy density but lower
power could utilize a larger particle size melt-cast material with a
lower carbon content, while a high power application would require
a smaller particle sized material with higher carbon content. This
optimization can be easily performed with the melt-cast LiFePO4
synthesis by the choice of a correct grinding media size for the
desired particle size of the product and carbon concentration for the
coating. Table I provides a summary of the rate capability results for
all materials and carbons tested in this evaluation. Overall, the trend
reported in Fig. 5 for the PMAAO-coated sample ͑increasing capac-
ity with smaller particle size and increasing carbon coating content͒
is followed by all carbon precursors investigated to date.
The 60°C discharge capacity with extended charge/discharge
cycles of the PMAAO-coated samples are shown in Fig. 6. The test
was run at a cycling rate of C/4. After 10 cycles one run at C/8 was
performed to test the effect of cycling lifetime on the low rate per-
formance of the cell. It should be emphasized that this test was
performed at 60°C, a temperature at which a poor LiFePO4 material
shows rapid capacity loss even though performance at room tem-
perature could appear adequate. In Fig. 6, it is clear that the
planetary-milled sample shows a significant drop in capacity
͑125 mAh/g͒ compared to the theoretical value of LiFePO4
͑170 mAh/g͒. Similarly, there is a large rate of capacity loss ͓77%
͑calculated͒ after 100 cycles͔ with increasing cycle number. This
rapid capacity loss with increasing charge/discharge cycles is likely
due to the cracking of the large particle ͑ϳ10 ␮m͒ during repeated
cycling. This cracking causes the destruction of the carbon coating,
which leads to a loss of capacity. Similarly, this cracking likely leads
to the loss of contact with the current collector, effectively eliminat-
ing the particle from possible electrochemical activity. In compari-
son, the nanomilled samples have capacities as high as 95% of the
theoretical capacity for LiFePO4 and exhibit excellent capacity re-
tention. The difference in capacity obtained by the two nanomilled
samples ͑0.5–0.7 mm media vs 0.7–0.9 mm media͒ is mainly due to
the particle size of the active material that is a result of the chosen
size of milling media, where the smaller particle sized material
͑smaller milling media͒ obtains a higher capacity. In addition, when
the carbon content of a particular milled sample is evaluated, the
optimal carbon precursor concentration can be determined. It is im-
portant to emphasize that the total weight of the carbon-coated
nanomilled no.
2 product ͑5 ␮m͒, ͑e͒ hydroxyethyl-cellulose-coated
planetary-milled product ͑5 ␮m͒, and ͑f͒ hydroxyethyl-cellulose-coated na-
nomilled no. 2 product ͑5 ␮m͒.
Figure 5 presents the rate capability of the PMAAO-coated
LiFePO4 material at various discharge rates, milling conditions, and
carbon coating amounts. It is clear that as the particle size of the
molten product is reduced, the capacity that is obtained at higher
discharge rates increases due to the improvement in electronic and
ionic conductivity within the particle. In addition, LiFePO4 particles
with the same primary particle size but with higher carbon coating
concentrations show larger capacities at higher discharge rates due
to the improvement of electrical conductivity. This demonstrates the
importance of the optimization of the particle size of the product for
its specific application, be it for energy density or for power appli-
Figure 5. ͑Color online͒ Rate capability of PMAAO-coated LiFePO4
samples from various milling conditions and carbon-coating content. Stars:
planetary mill; cross: nanomill no. 1 diamonds: nanomill no. 2, squares:
nanomill no. 3, 5%, circles: no. 3, 10%, and triangles: nanomill no. 3, 15%.
Downloaded on 2014-10-30 to IP 128.114.34.22 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, 157 ͑4͒ A463-A468 ͑2010͒
A467
Figure 7. ͑Color online͒ Comparison of commercial energy grade LiFePO4
͑P1, cross͒, power grade LiFePO4 ͑P2, squares͒, and nanomilled melt-cast
LiFePO4 samples ͑circles͒.
Figure 8. ͑Color online͒ Comparison of capacity vs cycle number of com-
mercial energy grade LiFePO4 ͑P1, cross͒, power grade LiFePO4 ͑P2, dia-
monds͒, and two nanomilled melt-cast LiFePO4 samples ͑circles, squares͒.
Cycling was performed at 60°C for 10 cycles at C/4, followed by one cycle
at C/8.
LiFePO4 sample is used to calculate the capacity. The optimal car-
bon precursor content is consistently 10% for the samples analyzed
here. The 10% precursor concentration leads to a carbon coating of
roughly 1–2% after treatment at 700°C for 1 h under a N2 atmo-
sphere. When the carbon precursor is lower or higher than 10%,
there is a drop in discharge capacity due to low conductivity ͑5%
added precursor͒ or large inactive carbon concentration ͑15% added
precursor͒. A summary of all the capacity retention results is pre-
sented in Table I.
trode films of high density for such applications as electric vehicles.
Ultimately, we are searching for methods to increase tap density and
lower the surface area of the carbon-coated LiFePO4 material, such
that high electrode density films could be produced from these melt-
cast materials.
Conclusion
Figure 7 compares the rate capability of two commercially avail-
able LiFePO4 products from Phostech Lithium to one of our best
performing nanomilled melt-cast LiFePO4 samples. All cells were
prepared and analyzed the same way with the same equipment. The
two commercial samples are optimized for specific applications, ei-
ther energy density ͑P1͒ or power density ͑P2͒. More information on
these products can be found elsewhere.14-16 It is clear from Fig. 7
that the nanomilled melt-cast sample compares very well to com-
mercial samples. These samples were prepared using laboratory
methods and were not optimized for rate capability. Higher capaci-
ties are likely when commercial cells and coating practices are used
for these samples. The coating methods used and the cell testing
methods demonstrate a baseline of performance for these materials.
Figure 8 compares the extended charge/discharge capacities of
the commercial samples with those of the best nanomilled melt-cast
samples. It is clear that the commercial samples have a significantly
improved capacity retention ability ͑1–2% calculated loss over
100 cycles͒ when compared to the nanomilled samples ͑ϳ7% cal-
culated loss over 100 cycles͒, but the nanomilled samples have
larger capacity numbers, which may be the result of a lower amount
of carbon coating. It should be emphasized that the milling proce-
dure was limited due to the available milling media size. Smaller
milling media should result in smaller sized LiFePO4 and improved
capacity at high discharge rates and perhaps improved capacity re-
tention ability.
The results in Fig. 7 and 8 clearly demonstrate that melt-cast,
milled, and carbon-coated LiFePO4 is capable of providing an elec-
trochemical performance similar to that of commercial material. The
results shown here are from experimental samples and have not been
optimized for powder processing; this is a very important part of the
technique to produce dense electrode films for commercial cells. For
example, the tap densities of commercial LiFePO4 powders are
1–1.1 g/mL for P1, and 0.6–0.7 g/mL for P2, while the experimen-
tal powders shown in Fig. 7 and 8 have tap densities of 0.8–1 g/mL.
Further work is required on the nanomilled samples to produce elec-
LiFePO4 was obtained from a molten process ͑1000°C͒ that has
shown, with the proper milling procedure, the ability to obtain high
capacities ͑95% of theoretical capacity͒ with excellent cyclability
and rate capability. The electrochemical characteristics compare fa-
vorably with those of commercially available material. It is crucial
that, with this synthetic approach, the optimal milling procedure is
obtained. We are continuing our investigation with smaller media
with the goal of decreasing particle size below 200 nm. With a sig-
nificantly lower reaction dwell time at high temperature, the ability
to use inexpensive precursors, and ease of scale-up compared to
other synthesis methods employed in the literature, it is believed that
the melt casting method to synthesize LiFePO4 will have a very
important role in the future development of LiFePO4 applications.
The method offers the promise of pure, low cost, and large-scale
sample production for use as high power hybrid electric vehicles and
plug-in hybrid electric vehicle applications. In addition, further work
is currently being performed on the development of LiMPO4, where
M = Mn,Co,Ni, or a combination thereof, as well as the optimiza-
tion of powder characteristics such that dense electrode films can be
produced.
Acknowledgments
The authors thank NSERC and Phostech Lithium, Inc. for fund-
ing this work under the auspices of the Industrial Research Chair
program. The authors also acknowledge the research group of Pro-
fessor Gregory Patience at École Polytechnique Montréal for the use
of their MiniFer mill.
Université de Montréal assisted in meeting the publication costs of this
article.
References
1. A. K. Padhi, K. S. Najundaswamy, and J. B. Goodenough, J. Electrochem. Soc.,
144, 1188 ͑1997͒.
2. A. Yamada, S. C. Chung, and K. Hinokuma, J. Electrochem. Soc., 148, A224
Downloaded on 2014-10-30 to IP 128.114.34.22 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A468
Journal of The Electrochemical Society, 157 ͑4͒ A463-A468 ͑2010͒
͑2001͒.
3248 ͑2007͒.
3. H. Huang, S. C. Yin, and L. F. Nazar, Electrochem. Solid-State Lett., 4, A170
͑2001͒.
4. K. Zaghib, K. Striebel, A. Guefi, J. Shim, M. Armand, and M. Gauthier, Electro-
chim. Acta, 50, 263 ͑2004͒.
11. M. Gauthier, C. Michot, N. Ravet, M. Duscheneau, J. Dufour, G. Liang, J.
Wontcheu, L. Gauthier, and D. D. MacNeil, J. Electrochem. Soc., 157, A453
͑2010͒.
12. L. Gauthier, M. Gauthier, D. Lavoie, C. Michot, and N. Ravet, U.S. Pat. 7,534,408
͑2009͒.
5. N. Ravet, Y. Chouinard, J. F. Magnan, S. Besner, M. Gauthier, and M. Armand, J.
Power Sources, 97–98, 503 ͑2001͒.
13. B. Ellis, P. S. Herle, Y. H. Rho, L. F. Nazar, R. Dunlap, L. K. Perry, and D. H.
Ryan, Faraday Discuss., 134, 119 ͑2007͒.
6. B. Kang and G. Ceder, Nature (London), 458, 190 ͑2009͒.
7. P. S. Herle, B. Ellis, N. Coombs, and L. F. Nazar, Nature Mater., 3, 147 ͑2004͒.
8. H. C. Kang, D. K. Jun, B. Jin, E. M. Jin, K. H. Park, H. B. Gu, and K. W. Kim, J.
Power Sources, 179, 340 ͑2008͒.
9. S. Yang, P. Y. Zavalij, and M. S. Whittingham, Electrochem. Commun., 3, 505
͑2001͒.
14. M. Armand, M. Gauthier, J. F. Magnan, and N. Ravet, WO 02/27823 A1 ͑2002͒.
15. G. Nuspl, C. Vogler, M. Eisgruber, L. Wimmer, and N. Schall, Abstract 293, in
12th International Meeting of Lithium Batteries, Nara, Japan ͑2004͒.
16. N. Ravet, C. Michot, G. Nuspl, G. Liang, and M. Gauthier, Abstract 225, The
Electrochemical Society Meeting Abstracts, Vol. 2006-02, Cancun, Mexico, Oct
29–Nov 3, 2006.
10. B. Ellis, W. H. Kan, W. R. M. Makahnouk, and L. F. Nazar, J. Mater. Chem., 17,
Downloaded on 2014-10-30 to IP 128.114.34.22 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).