Article

pubs.acs.org/Langmuir

Optimization of LiFePO4 Nanoparticle Suspensions with

Polyethyleneimine for Aqueous Processing
Jianlin Li,*,† Beth L. Armstrong,† Jim Kiggans,† Claus Daniel,†,‡,§ and David L. Wood, III†
†
Materials Science and Technology Division and ‡Energy and Transportation Science Division, Oak Ridge National Laboratory,
Oak Ridge, Tennessee 37831-6083, United States
§
Bredesen Center for Interdisciplinary Research and Education, University of Tennessee, Knoxville, Tennessee 37996, United States

ABSTRACT: Addition of dispersants to aqueous based lithium-ion battery electrode

formulations containing LiFePO4 is critical to obtaining a stable suspension. The
resulting colloidal suspensions enable dramatically improved coating deposition when
processing electrodes. This research examines the colloidal chemistry modifications
based on polyethyleneimine (PEI) addition and dispersion characterization required
to produce high quality electrode formulations and coatings for LiFePO4 active
cathode material. The isoelectric point, a key parameter in characterizing colloidal
dispersion stability, of LiFePO4 and super P C45 were determined to be pH = 4.3 and
3.4, respectively. PEI, a cationic surfactant, was found to be an effective dispersant. It is
demonstrated that 1.0 wt % and 0.5 wt % PEI were required to stabilize the LiFePO4
and super P C45 suspension, respectively. LiFePO4 cathode suspensions with 1.5 wt % PEI demonstrated the best dispersibility
of all components, as evidenced by viscosity and agglomerate size of the suspensions and elemental distribution within dry
cathodes. The addition of PEI significantly improved the LiFePO4 performance.

■ INTRODUCTION
Rechargeable lithium-ion batteries have received extensive
Additionally, replacing PVDF with xanthan gum6 or carbox-
ymethyl cellulose (CMC)7 would reduce fluorine content in the
attention in the last two decades and have been used in electrodes, and the formation of LiF could be suppressed. The
portable electronic devices such as laptop computers, cellular overall process would become substantially more environ-
phones, and personal digital assistants. However, application to mentally benign; consequently, there is growing interest in
electric vehicles and grid energy storage is limited by fabricating composite cathodes through aqueous processing.8
performance and cost.1 The main components of battery costs However, replacing NMP with water creates problems with
are materials, labor, and overhead with the cost of materials and dispersion stability. Particles in water based dispersions can
associated processing making up over 80% of total costs of high agglomerate due to hydrogen bonding and strong electrostatic
power batteries.2 Thus, the key to reducing costs of lithium-ion forces.9 These driving forces are even more problematic for
batteries lies in achieving low cost materials and developing LiFePO4, since the material is optimized for improved
low cost material processing,3 which is especially true for the electrochemical performance by making nanoparticles (i.e., larger
cathode. Cathode materials and processing represent the surface area).
majority of the total cost of high power batteries.2 Agglomeration is caused by the interactions between colloidal
LiFePO4 is a promising cathode material for the next particles. These interactions include attractive and repulsive
generation of scalable lithium-ion batteries, which is ascribed to potentials, which are generated from van der Waals and
low price, good cycle life,4 safety,5 and low environmental impact Coulomb forces, respectively. Usually, the attractive potential is
(i.e., no toxic elements in the compound). For conventional dominant at greater distances between particles. The stability of
lithium-ion batteries, the manufacturing process of LiFePO4 the particles depends on the net potential generated between
cathodes involves a slurry processing in which LiFePO4 is the van der Waals and Coulomb forces. Therefore, to minimize
mixed with other additives in a solvent. Polyvinylidene fluoride agglomeration, the key is to increase the repulsive potential
(PVDF) and N-methyl-2-pyrrolidone (NMP) are the typical (i.e., increase the Coulomb force) between particles. The re-
binder and solvent, respectively. If the composite cathodes could pulsive potential depends on the particle surface charge, and it is
be processed through an aqueous system, in which the expensive measured indirectly. The measurement is known as the zeta
NMP (>$2.25/L when purchased in large quantity) is replaced potential, and it is mostly dependent on the surface chemistry
with deionized water ($0.015/L), the cost would be significantly of colloidal particles.10 The agglomeration can be mitigated or
reduced, and the process for recovery and treatment of NMP eliminated by adding a dispersant to the aqueous slurries, and
would be eliminated. Based on a recent Oak Ridge National
Laboratory (ORNL) processing cost study, this approach has a Received: December 30, 2011
potential to save up to 70% in the electrode processing steps Revised: January 30, 2012
translating to a 12% reduction of the overall battery pack cost. Published: January 31, 2012

© 2012 American Chemical Society 3783 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790
Langmuir Article

there has been some progress in this area.8c,g,h,11 Porcher et al.8h LiFePO4 cathodes was characterized by energy dispersive spectroscopy
chose a nonionic dispersant, Triton X-100, to improve the (EDS) mapping using scanning electron microscopy (SEM) (JEOL
dispersion of carbon black. Li et al. used the anionic dispersants 6500 FEG-SEM). The morphology of the dried LiFePO4 green tapes
poly(acrylic acid) (PAA-NH4)8c and poly(4-styrene sulfonic acid) with and without PEI additions were characterized by scanning electron
microscopy (SEM) (JEOL 6500 FEG-SEM).
(PSSA)8g to improve the dispersion of LiCoO2 and LiFePO4 Half cells were assembled with LiFePO4 and Li metal as the cathode
composite cathodes, respectively. However, the dispersant effect and anode, respectively, with Celgard 2325 as the separator and 1.2 M
on processing of the electrodes is rarely discussed. LiPF6 in ethylene carbonate/diethyl carbonate (3/7 wt ratio, Novolyte)
Our strategy to improve the dispersion of C-coated LiFePO4 as the electrolyte. The cells were cycled at 0.2 C/−0.2 C between 2.5
(LiFePO4 hereafter) suspensions is by a cationic dispersant, and 4.2 V with VSP potentiostats (BioLogic) at 25 °C.
polyethylenimine (PEI), which was selected based on the
isoelectric point (IEP) of each cathode component, determined
by the zeta potential measurement. The zeta potential of each
■ RESULTS AND DISCUSSION
Zeta Potential Analysis. Zeta potentials of the LiFePO4
cathode component with and without PEI was determined, and aqueous cathode suspension are shown in Figure 1 with and
the effect of PEI on the rheological properties of LiFePO4 sus- without PEI. From Figure 1a, it is seen that the isoelectric point
pensions and LiFePO4 cathode performance was also investigated.

■ EXPERIMENTAL SECTION
As received LiFePO4 (2−3 wt % C, P2, Phostech Lithium Inc.), Super
P C45 (Timcal) (C45 hereafter), xanthan gum binder (XG hereafter,
Nuts Online) and polyethyleneimine (PEI) (Mw = 25 000 g mol−1,
Sigma-Aldrich) were used. Super P carbon black is a common car-
bon black used in lithium-ion batteries. Super P 45 (BET surface
area ∼45 m2/g) shows lower surface area compared to Super P 65
(BET surface area ∼65 m2/g), which allows for easier dispersion in
water. XG is a cheap, nontoxic binder that has been demonstrated in
lithium-ion batteries.6
Zeta potential measurements were carried out on LiFePO4, C45, and
XG in dilute aqueous suspensions using electrophoretic light scattering
(Zetasizer 3000HS, Malvern Instruments). Dilute suspensions with a
solids volume fraction of ∼10−5 were prepared by dispersing the
desired amount of powders in KNO3 (0.001 mol L−1) in distilled water.
The suspension pH was adjusted between 2 and 10 using solutions of
HNO3 (1 mol/L) and NH4OH (1 mol/L), respectively. Dilute
suspensions were mixed by magnetic stirring, and pH was measured by
a pH meter. Both pH and zeta potential of the dilute suspensions were
measured at 25 °C immediately after the suspensions were prepared
and after the suspensions were stored for 24 h. A period of 24 h was
selected, since it represents a reasonable duration between preparation
and processing of electrode suspensions for lithium-ion batteries and it
allows adequate time for charge adsorption and desorption on the
LiFePO4 particle surfaces to reach equilibrium. Only the zeta potential
results after 24 h stabilization are used in this work, and each data point
was averaged from 10 measurements.
The influence of dispersant concentration on viscosity and
agglomerate size was investigated for four different suspensions:
LiFePO4 suspensions; C45 suspensions; LiFePO4 and C45 suspensions;
and LiFePO4, C45, and XG suspensions. Viscosity of the suspensions
was measured at 25 °C by a controlled stress Rheometer (AR-G2, TA
Instruments). Agglomerate size distribution of the suspensions was
measured in water media at 25 °C by laser diffraction (Partica LA-950
V2, Horiba Scientific). The first step in preparing the suspensions was to
dissolve 10 wt % PEI in deionized (DI) water. The desired amount of
PEI solution was mixed with DI water for 2 min followed by mixing with
LiFePO4 or C45 by a high-shear mixer (model 50, Netzsch) for 30 min
to obtain a LiFePO4 or C45 suspension. For LiFePO4−C45
suspensions, LiFePO4 was first added to the PEI solution and mixed
for 15 min. Then, C45 was added to the suspensions and mixed for
another 15 min. For LiFePO4−C45−XG suspensions, a similar process
was used to obtain LiFePO4−C45 suspensions except the mixing time
was reduced to 10 min for each step. Finally, the desired amount of
1 wt % XG solution in DI water was added to the suspension and mixed
for 10 min. The ratio of components in all suspensions was maintained
at LiFePO4/C45/XG/H2O = 100/10/2.5/250 wt fraction.
The suspensions were coated by a custom slot-die coater (Frontier Figure 1. Zeta potential of each component of LiFePO4 aqueous
Industrial Technology) on pretreated Al foil and dried in a vacuum cathode suspensions with and without PEI. (a) C−LiFePO4. The
oven at 90 °C for 2 h. The Al foil was pretreated by corona discharge presence of PEI changes the charge on C−LiFePO4 surface to positive;
(Compak 2000, Enrocon) at 1 kW and 10 ft min−1 to improve its (b) C45, the presence of PEI changes the charge on C45 surface from
surface energy. The distribution of each component in the composite predominantly negative to positive.

3784 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790

Langmuir Article

(IEP) of LiFePO4 was pH = 4.3, the point where the charge on

the particle surfaces is zero. It is here that the suspension is in its
least stable state with respect to particle agglomeration. The IEP
also provides guidance on selection of the type of dispersant
(surfactant) required to obtain a stable dispersion. A dispersant
is an additive used to control particle separation, which allows
for increasing solids loading in a powder suspension and for
controlling particle packing during processing.12 It has been
reported that anionic dispersants such as poly(acrylic acid)
(PAA)7a and poly(4-styrene sulfonic acid) (PSSA)8g were used
in the LiFePO4 aqueous suspensions. According to Figure 1a,
the charge on the surface of the LiFePO4 particles is negative at
pH > 4.3. The appropriate pH for LiFePO4 suspensions should
be close to neutral to avoid corrosion on aluminum current
collectors in the subsequent coating processes.
Addition of PEI has been shown to yield a positive charge in
aqueous ceramic suspensions at pH < 1013 while PAA has been
shown to yield a negative charge at pH > 3 due to the
protonation. These findings can be extended to our suspensions
in that a cationic dispersant, such as PEI, with positive charge
should adsorb more easily on LiFePO4 particles than an anionic
dispersant with negative charge. The zeta potentials of LiFePO4
with PEI dispersant is also shown in Figure 1a, and it is seen that
the values changed to positive across the entire pH range
indicating PEI successfully adsorbed on the surface of LiFePO4
particles. Note that in typical zeta potential measurements,
solutions are aged 24 h to achieve an equilibrium charge state.
The scatter seen in the zeta potentials measured in this study
are attributed to nonequilibrium conditions existing at the time
of measurement. An experiment to study the effect of longer
equilibrium times on the resulting zeta potential values is planned.
In LiFePO4 aqueous cathode suspensions for lithium-ion
battery electrodes, there are two other components required,
the C45 conductive carbon black additive and the XG binder
(water-soluble). To have maximum efficacy as a dispersant for
the LiFePO4 aqueous cathode suspension, PEI also needs to be
efficient in dispersing the C45. Therefore, the zeta potential of
C45 was also investigated with and without PEI and is shown in
Figure 1b.
According to Figure 1b, the IEP of C45 was found to be pH =
3.4, and the charge on the surface of the C45 particles was
negative at pH > 3.4. This data suggests that PEI is an appropriate Figure 2. Rheological properties of LiFePO4 nanoparticle suspensions
dispersant for C45 similar to the findings for LiFePO4. However, with various amounts of PEI (a) viscosity vs shear rate and (b) shear
this is confirmed by the zeta potential of C45 when PEI stress vs shear rate.
dispersant is added (see Figure 1b). The charge on the C45
particle surfaces changed from predominantly negative without properties were further verified by the shear stress results shown
PEI to highly positive with PEI. Therefore, it can be concluded in Figure 2b, which were fitted with the Herschel−Bulkley
that PEI is an appropriate dispersant for both solid components (H−B) model. The H−B equation is one of the most employed
in LiFePO4 aqueous cathode suspensions. models for situations where a nonlinear dependence exists of
Suspension Rheological Properties. In order to under- shear stress on shear rate. It is described by the power law
stand the effect of PEI on aqueous processing of LiFePO4 equation and is stated as the following:
cathode suspensions, rheological properties of the suspensions
with individual and multiple components were investigated with ⎧
⎪ τ = τ0 + K γ̇
n
if τ > τ0
different PEI concentrations. All PEI concentrations investigated ⎨
⎪
in this work are referenced to the LiFePO4 weight fraction. ⎩ γ̇ = 0 if τ ≤ τ0 (1)
Figure 2a, b shows the apparent viscosity and shear stress,
respectively, as a function of shear rate for LiFePO4 suspensions. where τ, τ0, K, γ̇, and n are the shear stress, yield stress (stress
The viscosity significantly decreased with addition of PEI, and needed to initiate the flow), consistency index, shear rate, and
the suspensions also showed shear thinning behavior without power-law index, respectively. If n = 1, this function reduces to
PEI (i.e., when viscosity decreases with increasing shear rate). the classical Bingham plastic equation. If τ0 = 0 and n = 1, this
Newtonian or slightly shear thickening (i.e., when viscosity is function describes Newtonian behavior. The LiFePO4 suspen-
constant or slightly increases with increasing shear rate) behavior sion without PEI showed shear thinning behavior with n = 0.86
was observed when PEI was added to the suspensions. These and τ0 = 18.3 Pa, as shown in Table 1, which indicates relatively
3785 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790
Langmuir Article

Table 1. Parameters from the H−B Model for the LiFePO4

Suspensions
PEI conc τ0 (Pa) n shear rate (s−1)
0 wt % 18.3 0.86 whole range
0.5 wt % 13.1 1.00 <1310
−12.4 1.14 >1310
1.0 wt % 0.01 0.91 <500
−2.2 1.15 >500
1.5 wt % 0.04 0.98 <500
−1.72 1.22 >500
2.0 wt % 0.14 0.98 <500
−1.434 1.22 >500

strong interaction between agglomerates. In contrast, the addi-

tion of 0.5 wt % PEI changed the LiFePO4 suspension from
shear thinning to Bingham plastic with n = 1.00 and τ0 = 13.1 Pa
for a shear rate below 1310 s−1 and slightly shear thickening
with n = 1.14 and τ0 = −12.4 Pa for shear rate above 1310 s−1,
respectively. For higher PEI concentrations, the LiFePO4
suspensions showed Newtonian behavior with n ≈ 1 and
τ0 ≈ 0 Pa at γ̇ < 500 s−1, and slightly shear thickening
at γ̇ > 500 s−1, respectively. The decrease in yield stress with
increasing PEI concentration indicates that the interaction
between the agglomerates was dominated by van der Waals
forces for low or zero PEI concentrations. For higher PEI
concentrations, such forces were offset and eventually dominated
by Coulomb forces, induced by the adsorption of PEI on the
LiFePO4 particles. Thus, the suspensions were stabilized and
formed an ordered, layered structure. For further increases in
shear rate, the suspensions showed slightly shear thickening
behavior, which is ascribed to the order−disorder transition of
the ordered, layered structure above a critical shear rate. At this
point, the suspensions became unstable and more viscous.14 The
critical shear rate provides important information for suspension
processing, and it is demonstrated that 1.0 wt % PEI is enough to
stabilize LiFePO4 in water.
Figure 3 shows the rheological results of the C45 sus-
pensions, and their fitting parameters from the H−B model
are listed in Table 2. All of the suspensions showed shear
thickening behavior (n > 1) at high shear rate. At low shear
rate, however, the suspensions with PEI showed Newtonian
behavior (n = 1) as compared to shear thinning behavior Figure 3. Rheological properties of C45 suspensions with various amounts
(n < 1) for the suspension without PEI. The suspensions with of PEI (a) viscosity vs shear rate and (b) shear stress vs shear rate.
PEI were well-dispersed and stable with the yield stress equal or
close to 0 Pa. The shift from Newtonian to shear thickening at Table 2. Parameters from the H−B Model for the C45
high shear rate is also due to the order−disorder transition of Suspensions
the layered structure of C45 particles. Thus, 0.5 wt % PEI is
enough to well disperse C45 in water. PEI conc τ0 (Pa) n shear rate (s−1)
Rheological properties of suspensions composed of LiFe- 0 wt % −6.57 0.35 <820
PO4−C45 were also measured (Figure 4), and the fitting 77.9 2.36 >820
parameters from the H−B model are listed in Table 3. All of 0.5 wt % −0.03 0.98 <600
the suspensions except the one with 1.5 wt % PEI showed shear 2.23 2.42 >600
thinning behavior, and the suspensions with 1.0 wt % and 1.0 wt % −0.01 1.01 <800
2.0 wt % PEI exhibited quasi-Newtonian behavior with the 0.10 1.76 >800
power-law index close to unity, that is, n ≈ 0.90. In contrast, the 1.5 wt % 0.00 1.01 <750
suspension with 1.5 wt % PEI showed shear thickening 0.10 1.76 >750
behavior. From the weight fraction of materials, it could be 2.0 wt % 0.01 1.03 <750
assumed that 1.5 wt % PEI is enough to disperse the LiFePO4− 0.1 1.75 >750
C45 suspensions, since the LiFePO4 and C45 suspensions
need 1.0 wt % and 0.5 wt % PEI, respectively, to achieve including weight fraction of materials and mixing method,
Newtonian behavior. However, the rheological properties are sequence, time, and so forth. This phenomenon of shear
more complicated and affected by several additional factors, thickening is under investigation and will be discussed in future
3786 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790
Langmuir Article

Figure 4. Rheological properties of LiFePO4−C45 suspensions with Figure 5. Rheological properties of LiFePO4−C45−XG suspensions
various amounts of PEI (a) viscosity vs shear rate and (b) shear stress with various amounts of PEI (a) viscosity vs shear rate and (b) shear
vs shear rate. stress vs shear rate.

Table 3. Parameters from the H−B Model for the LiFePO4− Table 4. Parameters from the H−B Model for the LiFePO4−
C45 Suspensions C45−XG Suspensions
PEI conc τ0 (Pa) n shear rate (s−1) PEI conc τ0 (Pa) n shear rate (s−1)
0 wt % 452.1 0.41 whole range 0 wt % 111.2 0.36 whole range
0.5 wt % 249.7 0.64 whole range 0.5 wt % −3244 0.03 whole range
1.0 wt % 71.1 1.68 <100 1.0 wt % 314.9 0.22 whole range
214 0.89 >100 1.5 wt % −22.3 0.24 <150
1.5 wt % 17.4 1.52 whole range 216.3 0.92 >150
2.0 wt % 1.57 0.90 whole range 2.0 wt % 28.5 0.33 <100
260 0.87 >100
publications. Nevertheless, the decrease in the yield stress
confirmed that the interaction between agglomerates decreased LiFePO4−C45 suspensions. The addition of XG also increased
with increasing PEI concentration. the yield stress significantly, which is ascribed to the interlinking
Finally, the rheological properties of the LiFePO4 aqueous of particles by XG molecules. The suspensions with 1.5 wt %
cathode suspensions with all components were investigated and and 2.0 wt % PEI exhibited the lowest yield stresses and closest
are shown in Figure 5. The fitting parameters from the H−B power-law indices to unity resulting in the highest quality
model are listed in Table 4. All of the suspensions with XG suspensions.
binder showed shear thinning behavior, and the addition of PEI Effect on Suspension Agglomerate Size. Figure 6
XG further reduced their power-law indices compared to the shows the d50 values of the agglomerates for all suspensions.
3787 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790
Langmuir Article

Figure 6. d50 of the suspensions as a function of PEI. Note: The PEI

concentration is based on LiFePO4. It will be 10 times if based on C45,
since LiFPO4/C45 = 10/1 wt.

The d50 decreased with increasing PEI concentration for the

LiFePO4 suspensions demonstrating that better dispersion was
obtained at higher PEI concentrations. Based on the rheological
results in the Suspension Rheological Properties section, the
suspensions exhibited Newtonian behavior with 1.0 wt % or
more PEI. However, the distance between agglomerates might
not have been large enough at 1.0 wt % PEI, and multiple
agglomerates might have acted as one large agglomerate during
the size distribution measurement. Smaller d50 with higher PEI
concentration translates to a higher quality dispersion, either
smaller agglomerate size or greater distance between agglomer-
ates. In contrast, the C45 suspensions exhibited Newtonian
behavior with 0.5 wt % to 2.0 wt % PEI; however, the cor-
responding d50 increased with increasing PEI concentration. This
finding indicates PEI reached saturated adsorption on C45
particles at 0.5 wt %, and at higher PEI concentrations, additional
PEI molecules in the suspensions bridged with other PEI
molecules adsorbed on C45 particles forming larger agglomer-
ates. For the LiFePO4−C45 suspensions, the d50 values
decreased with increasing PEI and were relatively constant
with 1.5 wt % and 2.0 wt % PEI indicating the optimum PEI
concentration was 1.5 wt %. A minimum d50 was observed in the
LiFePO4−C45−XG suspensions with 1.5 wt % PEI demonstrat-
ing saturated adsorption of PEI on the agglomerate surfaces was
reached.
Morphology of Coated LiFePO4 Cathodes. Homoge-
neity of each component distributed in coated LiFePO4 Figure 7. SEM obervation of LiFePO4 composite cathodes. The
cathodes (coatings made with the suspensions containing intense red, blue, and yellow represent C, P, and Fe, respectively. (a, b)
all components) was qualitatively characterized by energy without PEI; (c, d) 0.5 wt % PEI; (e, f) 1.0 wt % PEI; (g, h) 1.5 wt %
dispersive spectroscopy (EDS) mapping (Figure 7). The PEI; and (i, j) 2.0 wt % PEI.
cathodes without PEI showed a rough surface with lots of
voids (Figure 7a), and the components were not distributed Performance of LiFePO4 Cathodes. It is common that
uniformly (i.e., carbon rich spots). In contrast, the cathodes with the capacity of LiFePO4 cathodes increases in the first couple
PEI exhibited a smoother surface and were much denser. There cycles, which is attributed to the reintercalation of the
were still some carbon rich spots in the cathode with 0.5 wt % deintercalated lithium via air and water exposure.15 Figure 8
(Figure 7b). However, it was observed that the components shows the voltage profiles of LiFePO4 cathodes at the fifth
were well-dispersed for the cathodes with 1.0 wt % PEI or cycle at which the deintercalated lithium was reintercalated. The
above. This finding verifies the addition of PEI improves addition of PEI significantly improved the LiFePO4 cathode
the suspension quality and the resulting dispersion of each performance. The LiFePO4 with 2.0 wt % PEI exhibited the best
component in the coated LiFePO4 cathodes. capacity (∼165 mAh/g). Furthermore, the plateaus were much
3788 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790
Langmuir Article

authors further wish to acknowledge John Henry, Kevin Cooley,

Fred Montgomery, Roberta Ann Meisner, Sergiy Kalnaus, Ed
Kenik, and Larry Walker for the assistance with the experiments.
The authors also thank Nancy Dudney, Joanne Meisner,
Phostech Lithium Inc., and Cabot Corporation for supplying
materials.

■ REFERENCES
(1) Tarascon, J.-M.; Armand, M. Issures and challenges facing
rechargeable lithium batteries. Nature 2001, 414, 359−367.
(2) Gainese, J.; Cuenza, R. Costs of Lithium-Ion-Batteries for Vehicles;
Argonne National Laboratory: Argonne, IL, 2000.
(3) Li, J.; Daniel, C.; Wood, D. Materials processing for lithium-ion
batteries. J. Power Sources 2011, 196 (5), 2452−2460.
(4) Striebel, K.; Shim, J.; Sierra, A.; Yang, H.; Song, X.; Kostecki, R.;
McCarthy, K. The development of low cost LiFePO4-based high
power lithium-ion batteries. J. Power Sources 2005, 146 (1−2), 33−38.
(5) (a) Arnold, G.; Garche, J.; Hemmer, R.; Ströbele, S.; Vogler, C.;
Wohlfahrt-Mehrens, M. Fine-particle lithium iron phosphate LiFePO4
synthesized by a new low-cost aqueous precipitation technique.
Figure 8. Voltage profile versus capacity curves at 0.2 C/−0.2 C and J. Power Sources 2003, 119−121, 247−251. (b) Padhi, A. K.;
25 °C confirming the addition of PEI improved the LiFePO4 cathode Nanjundaswamy, K. S.; Goodenough, J. B. Phospho-olivines as
performance. positive-electrode materials for rechargeable lithium batteries.
J. Electrochem. Soc. 1997, 144 (4), 1188−1194.
broader with PEI, and the polarization was much smaller. This (6) Courtel, F. M.; Niketic, S.; Duguay, D.; Abu-Lebdeh, Y.;
could be attributed to the smaller agglomerate size and more Davidson, I. J. Water-soluble binders for MCMB carbon anodes for
homogeneous distrubtion of cathode components, which favors lithium-ion batteries. J. Power Sources 2011, 196 (4), 2128−2134.
lithium diffusion inside the LiFePO4 agglomerates and the (7) (a) Lee, J.-H.; Kim, J.-S.; Kim, Y. C.; Zang, D. S.; Paik, U.
interfacial charge transfer. Dispersion properties of aqueous-based LiFePO4 pastes and their

■
electrochemical performance for lithium batteries. Ultramicroscopy
CONCLUSIONS 2008, 108 (10), 1256−1259. (b) Lee, J.-H.; Kim, J.-S.; Kim, C.; Zang,
D. S.; Choi, Y.-M.; Park, W. I.; Paik, U. Effect of carboxymethyl
It has been demonstrated that PEI is highly effective in cellulose on aqueous processing of LiFePO4 cathodes and their
dispersing the components of LiFePO4 cathodes for lithium-ion electrochemical performance. Electrochem. Solid-State Lett. 2008, 11
batteries. The effect of PEI concentration on suspensions with (10), A175−A178.
individual or multiple cathode components were studied. (8) (a) Guerfi, A.; Kaneko, M.; Petitclerc, M.; Mori, M.; Zaghib, K.
Concentrations of 1.0 wt % and 0.5 wt % PEI are required to LiFePO4 water-soluble binder electrode for Li-ion batteries. J. Power
stabilize LiFePO4 and C45, respectively, in water. For the Sources 2007, 163 (2), 1047−1052. (b) Li, C.-C.; Lee, J.-T.; Lo, C.-Y.;
suspensions with all components, 1.0 wt % PEI showed shear Wu, M.-S. Effects of PAA-NH4 addition on the dispersion property of
thinning behavior, whereas 2.0 wt % PEI exhibited close to aqueous LiCoO2 slurries and the cell performance of as-prepared
Newtonian behavior. The addition of PEI increased the LiCoO2 cathodes. Electrochem. Solid-State Lett. 2005, 8 (10), A509−
dispersion homogeneity of LiFePO4 cathode components and A512. (c) Li, C. C.; Lee, J. T.; Peng, X. W. Improvements of dispersion
homogeneity and cell performance of aqueous-processed LiCoO2
reduced the agglomerate size. These results provide guidance in
cathodes by using dispersant of PAA-NH4. J. Electrochem. Soc. 2006,
suspension processing and electrode coating of LiFePO4 153 (5), A809−A815. (d) Li, C.-C.; Lee, J.-T.; Tung, Y.-L.; Yang, C.-R.
composite cathodes for lithium-ion batteries through aqueous Effects of pH on the dispersion and cell performance of LiCoO2
processing. The addition of PEI dramatically improved the cathodes based on the aqueous process. J. Mater. Sci. 2007, 42 (14),
LiFePO4 cathode performance.

■
5773−5777. (e) Cai, Z. P.; Liang, Y.; Li, W. S.; Xing, L. D.; Liao, Y. H.
Preparation and performances of LiFePO4 cathode in aqueous solvent
AUTHOR INFORMATION with polyacrylic acid as a binder. J. Power Sources 2009, 189 (1), 547−
Corresponding Author 551. (f) Porcher, W.; Lestriez, B.; Jouanneau, S.; Guyomard, D. Design
*E-mail: [email protected]. Phone: +1-865-574-5158. Fax: of aqeous processed thick LiFePO4 composite electrodes for high-
energy lithium battery. J. Electrochem. Soc. 2009, 156 (3), A133−A144.
+1-865-574-4357.
(g) Li, C. C.; Peng, X. W.; Lee, J. T.; Wang, F. M. Using poly(4-
Notes styrene sulfonic acid) to improve the dispersion homogeneity of
The authors declare no competing financial interest. aqueous-processed LiFePO4 cathodes. J. Electrochem. Soc. 2010, 157

■ ACKNOWLEDGMENTS
This research was performed at Oak Ridge National Laboratory
(4), A517−A520. (h) Porcher, W.; Lestriez, B.; Jouanneau, S.;
Guyomard, D. Optimizing the surfactant for the aqueous processing of
LiFePO4 composite electrodes. J. Power Sources 2010, 195 (9), 2835−
2843. (i) Lee, J.-T.; Chu, Y.-J.; Wang, F.-M.; Yang, C.-R.; Li, C.-C.
(ORNL), managed by UT Battelle, LLC, for the U.S. Department Aqueous processing of lithium-ion battery cathodes using hydrogen
of Energy (DOE) under contract DE-AC05-00OR22725 and was peroxide-treated vapor-grown carbon fibers for improvement of
sponsored by the Office of Energy Efficiency and Renewable electrochemical properties. J. Mater. Sci. 2007, 42 (24), 10118−
Energy (EERE) Advanced Manufacturing Office and Vehicle 10123. (j) Lee, J.-T.; Chu, Y.-J.; Peng, X.-W.; Wang, F.-M.; Yang, C.-R.;
Technologies Program (VTP). Part of the work was conducted at Li, C.-C. A novel and efficient water-based composite binder for
the Shared Research Equipment (SHaRE) user facility, which is LiCoO2 cathodes in lithium-ion batteries. J. Power Sources 2007, 173
sponsored by the ORNL Scientific User Facilities Division. The (2), 985−989.

3789 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790

Langmuir Article

(9) Nahass, P.; Rhine, W. E.; Pober, R. L.; Bowen, H. K.; Robbins,
W. L. Ceramic Transactions, Materials and Processes in Microelectronic
Systems; Nair, K. M., Pohanka, R., Buchanan, R. C., Eds.; American
Ceramic Society: Westerville, OH, 1990; p 355.
(10) Hunter, R. J. Foundations of Colloid Science; Oxford University
Press: New York, 2001.
(11) Lee, J.-H.; Paik, U.; Hackley, V. A.; Choi, Y.-M. Effect of
poly(acrylic acid) on adhesion strength and electrochemical perform-
ance of natural graphite negative electrode for lithium-ion batteries.
J. Power Sources 2006, 161 (1), 612−616.
(12) Mistler, R. E.; Twiname, E. R. Tape Casting Theory and Practice;
The American Ceramic Society: Westerville, OH, 2000.
(13) Kirby, G. H.; Cooley, K. M.; Armstrong, B. L. Tailored
rheological behavior of mullite and bsas suspensions using a cationic
polyelectrolyte. Proc. ASME Turbo Expo: Power Land, Sea, Air 2005, 1,
307−313.
(14) (a) Boersma, W. H.; Laven, J.; Stein, H. N. Shear Thickening
(Dilatancy) in Concentrated Dispersions. AIChE J. 1990, 36 (3), 321−
332. (b) Derjaguin, B. V.; Landau, L. D. Theory of stability of highly
charged lyophobic sols and adhesion of highly charged particles in
solutions of electrolytes. Acta Physicochim. URSS 1941, 14, 633−652.
(15) Martin, J. F.; Yamada, A.; Kobayashi, G.; Nishimura, S. I.;
Kanno, R.; Guyomard, D.; Dupre, N. Air exposure effect on LiFePO4.
Electrochem. Solid-State Lett. 2008, 11 (1), A12−A16.

3790 dx.doi.org/10.1021/la205157d | Langmuir 2012, 28, 3783−3790