J Solid State Electrochem (2011) 15:649–662

DOI 10.1007/s10008-010-1264-9

REVIEW

Separator technologies for lithium-ion batteries

Xiaosong Huang

Received: 11 October 2010 / Revised: 20 November 2010 / Accepted: 28 November 2010 / Published online: 30 December 2010
# Springer-Verlag 2010

Abstract Although separators do not participate in the energy (about 150 Wh/kg), high-energy density (about
electrochemical reactions in a lithium-ion (Li-ion) battery, 400 Wh/L), long lifetime cycle (>1,000 cycles), low self-
they perform the critical functions of physically separating the discharge rate (2–8%/month), and high-operational voltage
positive and negative electrodes while permitting the free flow (2.5–4.2 V) [1]. They have been widely used in consumer
of lithium ions through the liquid electrolyte that fill in their electronics, such as portable computers, camcorders, and cell
open porous structure. Separators for liquid electrolyte Li-ion phones [2]. Recently, their applications have been expanding
batteries can be classified into porous polymeric membranes, into aerospace technology, electric vehicles (EV), and hybrid-
nonwoven mats, and composite separators. Porous mem- electric vehicles (HEV) [2].
branes are most commonly used due to their relatively low
processing cost and good mechanical properties. Although Lithium-ion battery design Typically produced in either
not widely used in Li-ion batteries, nonwoven mats have the spiral wound or prismatic design, the four fundamental
potential for low cost and thermally stable separators. Recent components of a Li-ion battery are the negative electrode
composite separators have attracted much attention, however, (anode), the positive electrode (cathode), the electrolyte, and
as they offer excellent thermal stability and wettability by the the separator, as shown schematically in Fig. 1. The first three
nonaqueous electrolyte. The present paper (1) presents an components participate in the electrochemical reactions in a
overview of separator characterization techniques, (2) reviews battery cell, while the separator is an inactive component.
existing technologies for producing different types of separa- Most of Li-ion batteries use carbon as the active anode
tors, and (3) discusses directions for future investigation. material, although silicon, lithium titanate, tin dioxide, and
Research into separator fabrication techniques and chemical some intermetallics have also been intensely researched [3–
modifications, coupled with the numerical modeling, should 6]. This active material is adhered to a copper current
lead to further improvements in the performance and abuse collector with a binder, typically poly(vinylidene fluoride)
tolerance as well as cost reduction of Li-ion batteries. (PVDF), and a high-surface-area carbon (e.g., acetylene
black) is used in the electrode formulation in order to
Keywords Lithium-ion battery . Separator . Porous provide an electron flow path to the current collector.
membrane . Battery abuse tolerance . Thermal runaway Lithiated transition metal oxides have been widely used as
the active materials for cathodes. Similarly to the negative
electrode, the active material particles are adhered to an
Introduction aluminum current collector with a binder. A high-surface-area
carbon is also incorporated to provide a conductive path for
Secondary lithium-ion (Li-ion) batteries provide an attractive the electrons to reach the active material particles.
landscape for energy storage systems due to their high specific The last active component in a Li-ion battery is the
electrolyte. An electrolyte composition requires good chemi-
X. Huang (*) cal, thermal, and electrochemical stability, as well as high-ionic
Chemical Sciences and Materials Systems Lab, conductivity. Despite its known disadvantages (problematic
General Motors Global R&D,
thermal stability and susceptibility to hydrolysis with the
30500 Mound Rd.,
Warren, MI 48090, USA formation of hydrofluoric acid), lithium hexafluorophosphate
e-mail: [email protected] (LiPF6) dissolved in a mixture of organic solvents represents
650 J Solid State Electrochem (2011) 15:649–662

Separator
S Separator requirements An ideal separator should have an
A d
Anode
C th d
Cathode Current
C t collector
ll infinite electronic but a zero ionic resistance. In practice, the
electrical resistivity of the polymers used for separators is in
Current collector
the order of 1012–1014 Ω cm, i.e., they are electrical
insulators. In the meantime, a low internal ionic resistance
is especially important for HEV/EV applications where a
battery needs to be able to offer high power. However, the
existence of the separator always increases the ionic
resistance of the inter-electrode medium (consisting of the
separator and the liquid electrolyte) because (1) the finite
porosity of a separator implies a restricted contact area
Charge between the electrolyte and the electrodes, and (2) the
P i i electrode
Positive l d LiMO2 Li1-xMO2+ xLi++ xe- tortuosity of the open porous structure results in a longer
Discharge
mean path for the ionic current compared to when the liquid
Charge
Negative
g electrode C + xLi++ xe- (Li+)xC-x electrolyte is used alone. Generally, a thin membrane with a
Discharge
high porosity and a large mean pore size can minimize the
Fig. 1 Schematic of a liquid electrolyte Li-ion cell ionic resistance, enabling high-specific battery power. How-
ever, too great of a porosity and a small membrane thickness
can reduce the mechanical strength of the membrane and
the most widely used liquid electrolyte due to its relatively increase the risk of inner battery electrical shorting. In
low-cost, high-ionic conductivity, and the ability of passivat- practice, most separators for liquid electrolyte batteries in use
ing the aluminum current collector for the positive electrode. today are 20 to 30 μm thick, have submicron-sized pores,
Generally, the organic solvents used for dissolving LiPF6 are and possess porosity ranging from 40% to 70%.
carbonates, such as ethylene carbonate (EC), propylene In addition, the separator should be mechanically strong,
carbonate (PC), dimethyl carbonate (DMC), ethyl methyl with no skew or yield, to keep the anode and cathode from
carbonate (EMC), and diethyl carbonate (DEC). During the contacting each other during the whole battery lifetime.
battery operation, lithium ions (Li+) are shuttled between the Separators also must possess dimensional stability at
anode and cathode via the electrolyte—from the anode to the elevated temperatures, especially for applications in high-
cathode during discharging and in the reverse direction power batteries. Fully charged batteries have highly
during the charging of the battery. oxidizing and reducing environments at electrode–electro-
The only inactive fundamental component of a Li-ion battery lyte interfaces, the stability of a separator in these environ-
cell is the separator, necessarily sandwiched between the anode ments is therefore critical. The specifications for a liquid
and cathode. It plays two main roles during the successful electrolyte battery separator provided by the US Advanced
operation of a cell: (1) resting between the anode and cathode to Battery Consortium (USABC) are listed in Table 2.
prevent internal short circuiting and (2) providing a path for
ionic conduction in the liquid electrolyte throughout the Significance of separator development Despite their advan-
interconnected porous structure. Separators for liquid electrolyte tages over the alternative battery technologies, the application
batteries are currently engineered as porous membranes, of large-capacity Li-ion batteries is hindered mainly due to the
nonwoven mats, or multilayers consisting of porous membranes cost, the abuse tolerance concerns, and the loss of perfor-
and/or nonwoven mats [7, 8]. Table 1 lists the major separator mance, especially at extreme operating temperatures. The
manufacturers, showing that most of the commercial separators further development in separator technologies will help
are made of porous polyolefin membranes. facilitate the expansion of the application of Li-ion batteries.

Table 1 Major separator

manufacturers Manufacturer Material Separator design

Asahi Kasei chemicals Polyolefin and ceramic-filled polyolefin Biaxially orientated

Celgard LLC PE, PP, and PP/PE/PP Uniaxially orientated
Entek membranes Ceramic-filled UHMWPE Biaxially orientated
PE polyethylene, PP polypro- ExxonMobil/Tonen PE and PE/PP mixtures Biaxially orientated
pylene, PET poly(ethylene tere- SK energy PE Biaxially orientated
phthalate), UHMEPE ultrahigh Ube industries PP/PE/PP Uniaxially orientated
molecular weight polyethylene
J Solid State Electrochem (2011) 15:649–662 651

Table 2 USABC requirements for Li-ion battery separators

Parameter Goal Test method

Sales price ($/m2) ≤1.00

Thickness (μm) ≤25 ASTM D5947-96, ASTM D2103
MacMullin number ≤11 N/A
Gurley (s/10 cm3) <35 ASTM D726
Wettability Complete wet-out in typical battery electrolyte N/A
Chemical stability Stable in battery for 10 years N/A
Pore size (μm) <1 ASTM E128-99
Puncture strength >300 g/25.4 μm ASTM F1306-90
ASTM D3763
Thermal stability <5% shrinkage at 200 °C ASTM D1204
Purity <50 ppm H2O N/A
Tensile strength <2% offset at 1,000 psi ASTM D882-00
Skew (mm/m) <2 N/A
Pin removal Easy removal from all major brands of winding machines N/A
Melt integrity ≥200 °C Thermal-mechanical analysis

One of the major concerns for Li-ion batteries centers Separator characterization
on the thermal events triggered by hard internal shorts.
The cause can be traced back to the manufacturing Electrolyte resistance derating factor The incorporation of
process when defects (metal fines) are introduced into the a porous separator increases the ionic resistance of the inter-
inter-electrode space. Separators are considered to be a electrode medium. The so-called MacMullin number (NM),
crucial battery component for avoiding thermal runaway shown below in Eq. 1, is a relative value used to describe
conditions. Although the separators with thermal shut- the increased resistance of the inter-electrode medium,
down ability have been commercially available since the caused by the presence of a separator, relative to the
1990s, they have been ineffective for preventing the hard resistance of the liquid electrolyte by itself, assuming
internal shorts that originate from the manufacturing defects. equivalent dimensions between the electrodes for each [9–
Two (not necessarily incompatible) technological solutions 11]:
for mitigating the internal shorting have therefore been
Rs
proposed during the past several years: (1) separators NM ¼ ð1Þ
R0
exhibiting a high melting point, a low high-temperature
shrinkage, and improved mechanical properties (in where Rs is the resistance of the inter-electrode medium
particular, puncture resistance) and (2) ceramic-enhanced once a separator is sandwiched between the electrodes, and
separators, either containing ceramic layers on their surfaces R0 is the resistance of the liquid electrolyte.
or having ceramic powders dispersed in the polymeric Figure 2 is a simplified alternating current impedance
material, where the major role of the ceramic is to spectrum (Nyquist plot), commonly used to determine the
prevent the collapse of the inter-electrode space and resistance. In Fig. 2, the semicircle (or overlapped semi-
thus avoid an internal short in the event of thermal circles) is due to different physical electrochemical pro-
runaway conditions. cesses and the observed Warburg line in some cases can be
In addition, separator accounts for a large portion of the related to the diffusion of Li+. Ideally, NM should be close
cost of a battery cell, which can exceed 20% for a high-power to one, while the typical values of the NM for Li-ion battery
battery. The USABC cost target for a separator is $1/m2 as separators range from 5 to 15.
indicated in Table 2. Research into separator fabrication Besides the development of solid electrolyte interphase
techniques to develop low-cost separators is therefore critical (SEI) on the electrode particle surface, the compatibility
to reduce the overall cost of the battery system. between the separator and the electrodes can also change
The present paper presents an overview of separator the cell resistance. As indicated in Fig. 2, the total
characterization techniques, reviews existing technologies interfacial resistance (Ri or RSEI) can be estimated by
for producing different types of separators, touches upon the subtracting the bulk resistance (Rb) in the complex
numerical modeling, and discusses directions for future impedance spectrum. Jeong and Kim [12] observed that
investigation. after increasing the thickness of a porous gellable poly
652 J Solid State Electrochem (2011) 15:649–662

Z"
The tortuosity factor indicates the ratio of the mean actual
Frequency increases path in comparison with the direct distance. Tortuosity τ [3,
16] can be estimated from Eq. 4 as shown below:
rffiffiffiffiffiffiffiffiffi
Rs
t¼ " ð4Þ
R0
Z' where ε is the porosity ratio. A high tortuosity hinders the
growth of Li dendrites but increases the ionic resistance of
Rb RSEI
the inter-electrode medium.
Fig. 2 Schematic of a simplified AC impedance spectrum The permeability is typically characterized using the
Gurley number according to ASTM D726. It measures the
time for a given amount of gas to flow through the membrane
(acrylonitrile-methyl methacrylate), layered on a polyolefin at a constant pressure. Clearly, the Gurley number will depend
separator, from 4 to 14 μm, the Ri dropped from 50.4 to on the porosity, the fraction of open pores, and the tortuosity
40.1 Ω cm2. This observation in reduction was contributed of a separator. The permeability can also be evaluated from
to better interfacial adhesion. the permeability coefficient B according to Darcy’s law, from
Eq. 5 shown below. Here, l is the sample thickness, μ is the
Porosity, mean pore size, tortuosity, and permeability The fluid viscosity, v is the fluid velocity, and ΔP is the pressure
morphology of a separator affects not only the battery drop across the separator [17].
performance and abuse tolerance but also the efficiency of
lmv
assembling and handling of a battery through its mechan- B¼ ð5Þ
ical and physical properties. ΔP
Porosity can be determined by a simple liquid absorption
test according to ASTM D2873 [13], from Eq. 2, Wettability The separator should quickly absorb and retain
the electrolyte during operation. Poor wettability limits the
wT  ws
P¼ ð2Þ performance of a cell by increasing the internal ionic
rl  Vs
resistance. Good wettability shortens the electrolyte filling
where ws is the weight of the dry separator sample, wT is time during the cell assembly and extends the battery’s life
the total weight of the separator once the liquid is fully cycle under normal operating conditions. There is no
absorbed, ρl is the density of the liquid, and Vs is the standard method for the wettability test although measuring
apparent volume of the separator sample. the contact angle between the liquid electrolyte and the
Separators should have small mean pore sizes and separator is a simple indicator. An alternative method to
uniform pore size distributions. Small mean pore sizes are determine wettability is by hanging a separator vertically in
preferred as they can effectively prevent shorts caused by a liquid electrolyte and calculating the wicking (capillary
dendrite growth or the migration of the active material rise) rate or height.
particles between battery electrodes. Uniform pore size
distributions can avoid the nonuniform current densities Thermal behavior Most commercial separators provide
across the electrode-separator interfaces. A nonuniform thermal shutdown (fuse) capability at various temper-
current accelerates the formation of Li dendrites in a cell. atures. Upon reaching a certain temperature, the separa-
Pore dimension and pore size distribution can be estimated tor can flow and fuse into a film by pore collapsing,
from porosimetry techniques based on the Laplace equation thus, minimizing the ionic conduction between the
[14] as shown below in Eq. 3: electrodes and protecting batteries from overheating or
electrical shorting. Thermal shutdown must start before
4g cos q
d¼ ð3Þ the occurrence of the uncontrollable thermal runaway
ΔP
and can be evaluated by measuring the resistance as the
where d is the pore diameter, ΔP is the pressure difference temperature increases. Venugopal et al. [13] tested
across the pore, γ is the surface tension of the liquid, and θ different membranes, as shown in Fig. 3, where a
is the contact angle. Based on this equation, different types polypropylene (PP) membrane exhibited shutdown at
of porometers have been developed to characterize complex around 165 °C and the resistance increased approximately
porous structures as discussed in the review paper by Jena by two orders of magnitude; a polyethylene (PE) separator
and Gupta [15]. For a more qualitative approach, cross showed shutdown at around 135 °C and the electrical
section and surface pore morphologies can be examined resistance increased by about three orders of magnitude. It
simply from scanning electron microscopy images. was reported that at least a three order of magnitude
J Solid State Electrochem (2011) 15:649–662 653

employ stretching in one or two directions to impart/

increase the porosity and improve the tensile strength. Both
processes use low-cost polyolefin materials, and thus the
majority of the separator cost is due to manufacturing
methodologies.
In the dry process, the melt-extruded polyolefin films are
annealed at an elevated temperature immediately below the
polymer melting temperature to induce crystallite formation
and/or to increase the size and amount of crystallites. The
well-aligned crystallites have lamellae arranged in rows,
perpendicular to the machine direction. Uniaxial stretching
at a low temperature and a subsequent high temperature is
followed. The stretching of about 150–250% in the
machine direction can be achieved by utilizing nip rolls
Fig. 3 Separator thermal shutdown [13] with different circumferential rates. The porous structure is
formed in this step due to the fracturing of the amours
increase in the electrical resistance was required to phase or tearing apart of the lamellar crystallites. Heat
effectively shut down the reactions. treatment is then followed to fix the pores and relax the
A separator must also prevent the electrodes from residual stress in the membrane. Porous membranes
contacting one another at high temperatures. Therefore, its produced through this dry process usually show character-
shrinkage at an elevated temperature needs to be mini- istic slit-like pore structures, as shown in Fig. 4a [7]. It is
mized. Thermal-mechanical analysis can be used to obvious that nanosized fibrils connect adjacent crystalline
characterize the high-temperature shrinkage of a separator. regions. The stretching ratio, speed, temperature, and other
The separator is held under a constant load while processing parameters dictate the membrane morphology.
simultaneously recording the elongation and temperature US patents 4994335, 4138459, 3801404, and 3843761
change. Porous separators that are either uniaxially or have described this process in detail. Generally speaking, a
biaxially stretched usually have a large shrinkage in the uniaxially stretched membrane exhibits good mechanical
stretched direction, while ceramic separators can offer high- properties (>150-MPa tensile strength) but a high thermal
dimensional stability at elevated temperatures. shrinkage in the machine direction, while a low tensile
strength (<15 MPa) but a negligible thermal shrinkage in
Mechanical properties A certain mechanical strength is the transverse direction. Dry process requires no use of
required for separator handling, battery assembly, and abuse solvents. However, this type of process is only applicable to
tolerance consideration. The most important mechanical polymers that can form semicrystalline structures.
properties are the tensile strength and puncture strength. In the wet process, plasticizers (or low molecular weight
Tensile properties are measured by the conventional substances, e.g., paraffin oil and mineral oil) are added into
stretching process according to ASTM D882. The strength the polymer before it is extruded or blown into a thin film
is strongly influenced by the post-extrusion stretching. at an elevated temperature. The extruded film is then
Porous membranes usually show a high tensile strength in calendared until a desired thickness is achieved. After
the stretched direction. The puncture strength is the load solidification, the plasticizer is extracted from the film with
required for a needle to puncture a separator and can be a volatile solvent (e.g., methylene chloride and trichloro-
tested according to ASTM F1306. The USABC puncture ethylene) to leave submicron-sized pores. The porous
strength target is 300 g/25.4 μm, as indicated in Table 2. membrane then passes through a solvent extractor to
High puncture strength helps to deter dendrites or particles remove the solvent in it after it leaves the oil extraction
from penetrating the battery separators. tank. The separators fabricated through this wet process are
generally stretched biaxially to enlarge the pore size and
increase the porosity. Therefore, the pores are more round-
like as indicated in Fig. 4b [7]. Tensile strengths in the
Polymeric porous membranes machine and transverse directions are comparable and can
both exceed 100 MPa. Compared with the dry process, this
Manufacturing processes The most widely used processes wet process can be applied on a wider range of polymers
for producing porous polymeric membranes for liquid since there is no requirement on the formation of a
electrolyte Li-ion batteries are the dry and wet processes. semicrystalline structure before stretching. In addition, the
Both processes are conducted through an extruder and application of plasticizers reduces the viscosity, thus,
654 J Solid State Electrochem (2011) 15:649–662

Fig. 4 Separators fabricated by the dry (a) and the wet (b) processes [7]

improving the polymer processability. However, the plasti- parameters that can change the membrane morphology
cizer extraction in this wet process adds cost to the include but are not limited to solvent, nonsolvent, compo-
membrane production. This step needs to be designed to sition of the polymer solution, and coagulation temperature.
prevent an undesired release of solvent from the system. Zhang et al. [29] used acetonitrile as a solvent and water as
Parameters such as dope composition, quenching tem- a nonsolvent to fabricate a porous membrane of poly
perature, stretching rate, and film thickness can all (vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP)
influence the membrane morphology. Kim and Kim [18] by this phase-inversion process. They observed different
investigated the pore size control by generating a phase morphologies on the two sides of the porous membrane as
diagram. Membranes were fabricated with PE/dioctyl one side of the cast solution film had been in contact with a
phthalate (DOP), PE/isoparaffin, and PE/DOP/isoparaffin, solid substrate while the other side had been exposed to the
respectively. The average pore size varied from 0.07 to nonsolvent. Andrieu and Laurence [30] cast a PVDF
0.5 μm depending mainly on the temperature gap between solution directly onto a negative electrode and, by immers-
the phase separation and PE crystallization. Ma et al. [19] ing the negative electrode in water, a porous separator
have shown that stretching before plasticizer extraction can membrane with a porosity of 75% was formed.
produce membranes with a smaller mean pore size and a The evaporation-induced phase separation is very similar
narrower pore size distribution compared with stretching to the abovementioned phase-inversion process. It takes the
after the extraction. Fillers can also be used to change the advantage of the differential evaporation rates of the solvent
morphology and improve the mechanical and thermal and the nonsolvent to form a porous structure. A compo-
performance of a membrane [20, 21]. Yoneda et al. [20] sition consisting of a polymer, a volatile polymer solvent,
investigated the phase separation behavior of a three- and a less volatile nonsolvent can be cast into a thin film.
component system consisting of PE, nanosilica powder, The solvent and the nonsolvent ideally should be com-
and a solvent (or plasticizer). They reported that the pletely miscible. With the evaporation of the solvent, the
addition of silica resulted in a greater mean pore size. film solidifies and the nonsolvent is homogeneously
Besides the dry and wet processes, there are a couple of dispersed in the matrix. The subsequent removal of the
other techniques that have been studied to produce porous nonsolvent yields a porous structure. Similarly, solvent,
membranes for Li-ion battery applications. Phase inversion nonsolvent, polymer concentration, polymer solubility, etc.
is a well-known technique to prepare porous membranes can all change the membrane morphology. It has also been
[22–28]. A polymer solution is cast into a thin film and then reported that fine fillers can be added to the polymer
soaked in a nonsolvent coagulation bath. Since the solvent solution to maintain the porous structure and provide
and nonsolvent are miscible, a rapid diffusion rate of the structural strength to the produced membrane [31–33].
nonsolvent and solvent causes phase separation to take
place, forming a solid-like polymer rich phase and a liquid- Polymers for the fabrication of porous separators Most of
like polymer lean phase. In the case when the polymer rich the commercially available Li-ion battery separators are
phase is a continuous phase, a porous membrane can be made of PE, PP, other polyolefins, or their mixtures, or
formed when the polymer lean phase is replaced by the copolymers through either a dry or wet process. Polyolefin
nonsolvent after a further solvent/nonsolvent exchange. The usually provides good mechanical properties and chemical
J Solid State Electrochem (2011) 15:649–662 655

stability. Most of the polyolefin separators can also provide a dry process. Compared with PP and PE, PMP has a higher
thermal shutdown capability at various temperatures by melting temperature of about 235 °C. The PMP precursor
closing the pores and turning the membrane into a film produced by a tubular extrusion process (dry process)
nonporous film. The thermal shutdown temperature of a showed a semicrystalline structure with planar-stacked
PP membrane is around 160 °C, while the thermal lamellae. The stacked lamellar morphology and the orien-
shutdown temperature of PE is between 120 and 150 °C, tation state of the membranes heavily relied on the
depending upon its morphology. Ultrahigh molecular extrusion stress and the weight-average molecular weight
weight PE (UHMWPE) has good melt integrity and (Mw) of PMP. The evaluation of this type of membranes as
sufficient polymer flow to cause pore collapse, thereby battery separators was not presented.
providing better thermal shutdown through the formation of Porous power technologies fabricated PVDF separators
a nonporous film. using the evaporation-induced phase separation process
Although polyolefin materials can provide thermal [41]. Acetone and water were used as the solvent and
shutdown at an appropriate temperature range, the temper- nonsolvent, respectively. The separator showed a high
ature in a cell may continue to increase after the pores are porosity of about 80%, and the batteries with this porous
closed. It is possible that the separator would shrink, melt, PVDF separator exhibited reduced internal resistance.
and finally lead to the shorting of the electrodes. Therefore, However, investigations on the abuse tolerance (i.e.,
to ensure good abuse tolerance, the difference between the tolerance of short) of this type of separators need to be
shutdown temperature and the melting temperature should conducted as PVDF losses its mechanical strength in liquid
be as large as possible. Coextrusion or lamination of PP and electrolytes, especially when the temperature is high.
PE to produce multilayer membranes has been used for this Attempts have been made to improve the battery abuse
purpose. PP/PE bilayer separators and PP/PE/PP trilayer tolerance through increasing the thermal stability of the
separators have been developed [34, 35]. The PE layer separator by using high-temperature polymers (such as
transforms into a nonporous film at a temperature lower polyamide and polyimide). These activities mostly remain
than the thermal runway temperature, which increases the in the research stage and the membranes are mainly being
electrical resistance and provides thermal shutdown. In the prepared using the phase-inversion approach. A paraor-
meantime, the PP layers still maintain mechanical integrity iented polyamide porous film comprising of fibrils having a
and keep the electrodes separated. diameter less than 1 μm has been fabricated [42]. The
Instead of lamination, Higuchi et al. [36] patented a polyamide solution was first cast into a film and then
battery separator made of a blend of PE and PP. The cooled to deposit the polyamide. Next, the film was
separator was produced by a dry process. Since PP and PE immersed in an aqueous or alcoholic solution to elute the
are not completely miscible, the separator showed porous solvent. The dry porous film exhibited good mechanical
regions of PE and PP. The separator provided the thermal properties and a porosity ranging from 38% to 87% with
shutdown ability at a temperature close to the melting point good uniformity. Due to the thermal stability of the
of PE. PP maintained the mechanical integrity of the aromatic polyamide, its linear thermal expansion coefficient
separator up to its melting temperature. between 200 and 300 °C was about 50×10−6/°C. However,
Ihm et al. [37] made separators using the blends of a its performance in a liquid electrolyte was not discussed.
high-density polyethylene (HDPE) and an UHMWPE by a UBE Industries, Ltd. has developed polyimide separators
wet process. They concluded that the mechanical strength [43]. The polyimide precursor was a polyamic acid
of the membranes increased with the molecular weight of prepared by polymerizing a tetracarboxylic acid and a
UHMWPE and its content in the blends. The film diamine. A porous precursor film was fabricated through
consisting of 6 wt.% UHMWPE at a draw ratio of 5 the phase-inversion method by casting the polyamic acid
showed a tensile strength of about 100 MPa. Although the solution, forming a liquid film comprising of a mixture of a
tensile strength increased with the draw ratio, the puncture solvent and a nonsolvent on a glass substrate, and
strength decreased. The pore size was very uniform and immersing the cast film into a coagulating solution
most of the pores arranged in size from 0.10 to 0.12 μm. composed of the solvent and the nonsolvent. The precursor
Nitto Denko Co. developed a UHMWPE separator film was then dried and imidized into a polyimide porous
having a crosslinked structure. UHMWPE resin blended film. The resulting polyimide separator showed relatively
with a crosslinkable rubber was made into a porous uniform pore size distribution. Although the membrane had
membrane by a wet process [38, 39]. The porous membrane no thermal shutdown ability, it offered excellent thermal
was then crosslinked by air oxidation. The final structure stability.
showed high puncture strength and heat rupture resistance. Overcharge protection has also been investigated
Johnson and Wilkes [40] prepared porous membranes through the separator design [44–47]. Xiao et al. [44]
made of an isotactic poly(4-methyl-1-pentene) (PMP) using prepared a composite membrane comprised of p-poly-
656 J Solid State Electrochem (2011) 15:649–662

phenyl (PPP) and polyaniline (PAn) layers. The separator celly [50] modified a PP membrane with acrylic acid (AA)
was p-doped at high-oxidation potentials into a conductive and diethyleneglycol-dimethacrylate by electron beam irra-
phase to shunt the current and returns to the de-doped diation; Yao and Ranby [51] developed a continuous process
isolating state at normal operating voltages. The PPP/PAn for the modification of polyolefin separators with AA
composite separator was evaluated in a Li/LiMn2O4 cell. through an ultraviolet irradiation for alkaline batteries; and
When cycled between 3.6 and 4.3 V at 25 mA/g, the Urairi et al. [52] treated polyolefin separators with plasma to
composite separator showed behaviors similar to a com- improve their hydrophilicity for alkaline battery application.
mercial PP porous membrane. However, upon overcharg- Porous polyolefin membranes have also been laminated
ing, the voltage of the cell with the PP separator rose to with gellable polymer films to improve their compatibility
4.8 V in 8 min, while the voltage of the cell containing the with the electrolyte and electrodes. A separator comprising of
composite separator showed a steady plateau at 4.31 V. a porous polyolefin base and a porous PVDF layer on at least
Chen and Richardson [45] impregnated a PP porous one surface of the base material was proposed [19, 53]. The
separator with an electrochemically active poly(3-butylth- presence of PVDF provided improved electrolyte retention
iophene) for overcharge protection. TiS2, Li, and 1 M LiPF6 and bondability to the electrodes. Polyolefin separators have
in PC/EC (1:1 by volume) were used as the cathode, the also been laminated with a porous gellable poly(acryloni-
anode, and the electrolyte, respectively. When charged at a trile-methyl methacrylate) layer by dipping the separator in
C/10 rate to a limit of 4 V, the discharge capacity of the cell the copolymer solution [12]. The base PE membrane was
with a regular PP separator degraded rapidly due to the 25 μm thick, and the coated separator had a total thickness
degradation of the cathode. The cell with the poly(3- of 30–45 μm. The authors attributed the good capacity
butylthiophene) impregnated separator was undamaged by retention and rate performance to the strong bonding
overcharging up to ten times the normal capacity, since between the separator and the electrodes. Kim et al. [54]
poly(3-butylthiophene) shunted the excess charge and further incorporated ceramic particles into a gellable polymer
limited the potential to about 3.2 V. However, at 1 °C rate coating by pasting a 2–3-μm thick layer of gellable polymer
overcharging, the overcharge protection was not achieved. containing SiO2 particles on both sides of a base separator
Feng et al. [46] used an electroactive polytriphenylamine as membrane. The addition of silica particles showed improved
the separator material. Upon testing Li/LiFePO4 cells, the wettability and enabled increased ionic conduction.
experimental results demonstrated that the electroactive
separator produced a resistive internal short circuit to
maintain the cell voltage near 3.75 V at overcharge.
However, for this type of separator, the major barriers to Other types of separators
practical applications are the battery self-discharge and the
effectiveness of the overcharge protection, especially at a Nonwoven mats/webs Although nonwoven mats/webs have
charge rate. long been used as separators for different batteries, they have
very limited applications in Li-ion secondary batteries
Separator modification Pristine polyolefin membranes lack nowadays. Nonwoven mat separators have certain advantages
sufficient compatibility with the electrodes and carbonate such as low processing cost, high porosity, and lightweight. In
solvents. However, good surface contact between separators addition, mats with different organic and inorganic fibers can
and electrodes is critical since a poor interface can cause an be conveniently fabricated to provide high heat resistance and
uneven current distribution resulting in the formation of Li good mechanical properties. However, nonwoven mat sepa-
dendrites and increase the battery internal resistance rators typically have a large pore size, so to prevent dendrites
deteriorating the battery performance. from growing through the pores, their thickness must be
The modification of membranes has then been conducted increased. The recent technology has made the fabrication of
to optimize the performance of the separator and the overall nanofiber nonwoven mats feasible. Thin nonwoven separators
battery system. Grafting polymerization can attach functional can thus be produced with nanofibers, which provide the
groups on the membrane surface through covalent bonding so potential for Li-ion battery applications. A nonwoven mat is
as to change the separator surface polarity permanently. Ko et usually produced through a dry laid process, a wet-laid
al. [48] used an electron beam to attach glycidyl methacry- process, a melt-blown process, or an electrospinning process.
late on the surface of a porous PE separator. They attributed These processes allow fiber adhesion by sprayed resin
the improved life cycle of the mesocarbon microbeads/ bonding, thermoplastic fiber bonding, fiber self bonding, or
LiCoO2 cell to an enhanced electrolyte holding capability mechanical bonding.
and a reduced interfacial resistance. Similarly, Gao et al. [49] The air-laid process is a dry process where individualized
grafted methylmethacrylate on polyolefin separators by an staple fibers are air conveyed onto a moving belt to form a
electron-beam-induced polymerization; Gineste and Pour- mat. A latex binder is applied at the same time to incorporate
J Solid State Electrochem (2011) 15:649–662 657

the mechanical integrity. The formed mat can also be ogy is affected by the solvent evaporation rate, the polymer
thermally bonded through a calendaring process. concentration, the distance from the nozzle tip to the target,
The wet-laid process can be conducted on a papermak- the accelerating voltage, etc. The fibers fabricated through
ing machine. Fibers are first suspended in water. Paper this process can be submicron or nanometer sized. Thus,
machines are used to separate water to form a uniform web, electrospun mats can have much smaller thickness com-
which is then bonded and dried to achieve mechanical pared with those fabricated by other mat-forming processes.
integrity. Kritzer [55] fabricated a polyester mat using this Due to the challenges of designing high-throughput electro-
wet-laid process. Polyester fibers together with binder jetting processes, investigation continues to be exploratory
fibers were suspended in water and randomly laid on a and no commercial battery separators are currently avail-
screen belt. The nonwoven mat, formed by drying the fiber able. Porous membranes composed of submicron fibers
suspension, was thermally bonded at 230 °C. The electro- made of polyimide [61], polyacrylonitrile (PAN) [62], and
chemically and thermally stable nonwoven mat showed an PVDF [63, 64] have been investigated. Bansal et al. [61]
average pore size of 20–30 μm and a porosity of 55–65%. prepared a polyether imide nanofiber web through electro-
Pinholes were not observed due to the labyrinth-like spinning. Polyimide and the blends of polyimide and PAN/
structure of the nonwoven material, yet, because of the PVDF-HFP were spun into mats with 65–85% porosity.
large mean pore size, this nonwoven mat was not used as a However, the effect of liquid electrolytes on the properties
separator directly, but as a support layer for battery of these mats was not discussed and their performance in a
separators. Ashida and Tsukuda [56] disclosed the fabrica- real battery was not systematically investigated. Cho et al.
tion of a nonwoven mat made by blending organic and up [62] prepared porous PAN separators using the electro-
to 80 wt.% ceramic fibers through this wet-laid process spinning process. The fibers had a homogeneous diameter
showing improved heat resistance. Organic fibers such as of 250–380 nm. Compared with the conventional porous
polyolefin and polyamide fibers were used to improve separators, PAN nonwoven mats showed better cycling
formability and increase strength. The separator showed performance and rate capability at room temperature. Kim
good adhesion to the electrodes, displaying no signs of et al. [63] was able to control the pore size and porosity
slippage or gaps between the separator and the electrodes, PVDF-HFP porous membranes by electrospinning the
during the fabrication of the battery. Wang et al. [57] copolymer solutions of distinct polymer concentration.
prepared a wet-laid nonwoven separator using fibrillated The electrospun polymer separators enabled good ionic
aromatic polyamide fibers observing, thereafter, an im- conduction at room temperature and a wide electrochemical
proved permeability and electrolyte absorption rate com- stability window of up to 4.5 V vs. Li+/Li when 1 M LiPF6
pared with a conventional polyolefin separator. Capillary in EC/DMC/DEC (1/1/1 by weight) was used as the
porometer tests showed the average pore sizes of the two electrolyte.
mats prepared were as small as 0.180 and 0.372 μm, In another effort, electrospinning was modified into an
respectively. It should be noted that a capillary porometer electroblowing process by simultaneously applying electrical
measures the smallest pore constraints. force and an air shearing force [65–69]. An air-blowing
Melt-blown mats are formed by extruding molten system, with flow and temperature control, delivered a high-
polymers through a die and then attenuating the resulting velocity air around the spinneret along the fiber spinning
filaments with a high-velocity air to fine diameters. This is direction. This process resulted in higher efficiency and
a promising low-cost process to fabricate thermally stable yield-combining characteristics of electrospinning and melt-
thin membranes suitable for battery separator applications. blowing processes [65].
The preparation of thin but strong separators made of Currently, nonwoven mats are mainly used as a support
polyester, polyamide, and polymethylpentene and its layer for separators. Patent application WO06123811
copolymers through this process has been suggested [58– disclosed a separator for Li-ion batteries produced by
60]. Mats consisting of a fibrous sheet of polymethylpen- coating a nonwoven mat with porous aromatic polyamide
tene showed a high-temperature shutdown characteristic, layers on both sides [70]. The poly(ethylene terephthalate)
and those consisting of high temperature polymers such as (PET) nonwoven mat with a thickness less than 30 μm was
polyester and polyamide exhibited excellent dimensional fabricated through the wet-laid process and the polyamide
stability at elevated temperatures. coating was applied through the phase-inversion method.
Electrospinning/electrospraying is another attractive The separators showed high heat resistance, and no
method for preparing nonwoven mats, especially those shrinkage was observed when tested at 150 °C for 2 h.
composed of nanofibers. The electrospinning/electrospray- Lee et al. [71, 72] coated a porous PVDF layer on the
ing process is conducted by applying a strong electric field surface of a PE nonwoven mat of 20 μm thick. The PVDF
to a polymer solution and collecting a fine-charged jet from coating was applied by casting a solution of PVDF in N-
a small orifice on a grounded collector. The mat morphol- methylpyrrolidone on the PE mat and then, by the phase-
658 J Solid State Electrochem (2011) 15:649–662

inversion method, converted to a 33-μm thick porous authors evaluated the membrane with a CaCO3/PTFE
structure. The authors reported that these separators weight ratio of 92/8 by soaking it in a liquid electrolyte of
displayed sufficient mechanical properties and improved 1 M LiPF6 in a mixture of EC and EMC (3:7 by weight).
capacity retention in Li-ion cells. Pekala and Khavari [73] The “effective ionic conductivity” of the electrolyte wetted
passed a ∼18-μm thick porous UHMWPE nonwoven mat membrane was 2.4 mS/cm at 20 °C as compared with
through a hot PVDF solution bath impregnating the 8.0 mS/cm for the virgin liquid electrolyte. However, the
polymer into the mat. The time required to achieve a compatibility of the filler with the battery components,
uniform electrolyte distribution throughout the battery was especially the liquid electrolyte, needs to be investigated.
reduced with this PVDF-coated mat due to increased Cho et al. [75] fabricated a ceramic separator by casting
compatibility with the battery system. PE–PP sheath-core composite fibers (a fiber diameter of
Ceramic coatings have also been applied on nonwoven 2 μm) with nanosized silica particles using an air-laid
mats for improved thermal performance. This topic is process. The formed webs were thermally bonded in an
covered in the following section. oven. The cell was assembled using LiCoO2 as the cathode,
graphite as the anode, and 1 M LiPF6 in EC/DEC (1/1 by
Composite/ceramic-enhanced separators As mentioned volume) as the electrolyte. The composite separator showed
earlier, to improve the battery abuse tolerance, Li-ion improved wettability and better performance than neat
battery separators with high thermal stability are needed, polyolefin-based membranes when cycled in the voltage
especially for applications in EVs or HEVs. This can be range of 3.0–4.2 V at a C/2 rate. The composite membrane
achieved either by using high melting temperature poly- was thermally stable, and its shrinkage at 150 °C was only
mers, such as polyamide and polyimide (as described in the 3%, as compared with the 37% shrinkage for a conven-
previous sections), or by incorporating inorganic fillers into tional PP separator. The authors also claimed that in order
the porous or nonwoven membranes. The latter are called to suppress the microshorting caused by the Li dendrite
composite or ceramic-enhanced separators. In general, growth, the Gurley number for this type of separators must
composite separators have excellent wettability by the be higher than 200 s/100 cm3.
nonaqueous electrolyte solvents, even when the tempera- Lamination by coating a supporting porous polyolefin or
ture is low. These separators also show excellent thermal nonwoven structure with a layer of inorganic particles has
stability, which is critical to ensure satisfactory battery also been used to fabricate freestanding ceramic-enhanced
abuse tolerance at high temperatures. In addition, inorganic separators. Shinohara et al. [76] fabricated a heat-resistant
ceramic particles have relatively high thermal conductivity, membrane by coating a mat or porous film with a
thus, dissipating heat at a fast rate. Due to the above- dispersion of ceramic particles in a heat-resistant aromatic
mentioned advantages, composite separators have attracted polymer solution of either a polyamide or a polyimide. The
increasing attention recently. However, although composite prepared laminate membranes typically showed a porosity
separators generally offer high punch strength, their tensile of about 45% and a tensile strength of above 45 MPa.
strength is usually lower compared with the plain polymeric Evonik–Degussa commercialized a ceramic separator by
separators. Other potential issues related to ceramic- coating an ultrathin PET nonwoven support layer with
enhanced separators are increased separator weight due to oxides including alumina, zirconia, and silica, as shown in
the incorporation of heavy ceramic particles, unknown Fig. 5 [8, 77–79]. Oxide particles were first suspended in an
electrochemical stability of the ceramic and the ceramic inorganic binder sol that was prepared by hydrolyzing a
binder in the highly oxidizing and reducing environments mixture of tetraethoxysilane, methyltriethoxysilane, and (3-
encountered in Li-ion batteries, and particle shedding due to glycidyloxypropyl) trimethoxysilane in the presence of a
the insufficient binding. HCl aqueous solution. The suspension was then coated on a
Composite separators can be produced by binding nonwoven PET mat. A composite separator was obtained
micron or submicron ceramic particles with polymeric by drying the coated PET at 200 °C. The obtained separator
binders. Zhang et al. [74] prepared a freestanding and had a small average pore size of 0.08 μm and a thickness of
flexible porous membrane by using CaCO3 particles as the about 24 μm. For this type of separators, the PET
main component and 2–8 wt.% of polytetrafluoroethylene nonwoven of about 20 μm thick provides the tensile
(PTFE) as the binder. The process for producing the strength and flexibility, while the fine ceramic particle
membrane involved dispersing 10 μm CaCO3 particles in coating helps avoid pinholes while providing dendrite
a PTFE dispersion, casting and hot rolling the dispersion penetration resistance and thermal stability.
into a flexible membrane with a thickness of 175–190 μm, LG Chem also developed ceramic-enhanced separators
and drying the membrane at 120 °C under vacuum. [80–83]. The separator was produced by coating a porous
However, to reduce the membrane thickness for realistic substrate (typically a polyolefin porous film) with a porous
applications, smaller particles will need to be used. The layer formed by casting a slurry of inorganic particles and a
J Solid State Electrochem (2011) 15:649–662 659

Fig. 5 Schematic of an Evonik–

Degussa ceramic separator [8,
77]

polymeric binder. The membrane with a PE support and a [92] applied separator coating on an electrode by spraying a
BaTiO3/butylacrylate-acrylic acid copolymer-coating layer dispersion of fumed silica in a PVDF solution. When the
showed less than 10% shrinkage at 150 °C for 1 h. silica/PVDF weight ratio was 65/35, the separator coating
Hitachi Maxell Ltd. has reported the development of a showed a NM similar to a commercial polyolefin separator.
thermally stable separator made by coating a polyolefin Kim et al. [93] formed composite separators by a dip-
porous film with metal oxide flakes [84]. The coated separator coating approach. The nanoalumina loading in the ceramic-
exhibited a shrinkage of less than 5% even at 180 °C. coating layer was up to 94 wt.% with the rest being a PVDF
In addition, composite separators have been fabricated binder. The cell with this separator coating showed
simply by adding ceramic particles into the material enhanced rate performance when compared with a cell
formulations for making separators using the wet process. with a conventional separator.
Sony Corporation developed a separator composed of
HDPE and fine alumina particles [85]. A mixture of PE
resin, alumina particles and/or inorganic fibers, and mineral
oil was kneaded into a sheet at the melting point of the Numerical modeling of Li-ion cells/separators
polyolefin resin. A porous membrane was formed through
the extraction of the mineral oils. Since the porous separators have very complex pore
Besides the applications in porous separators for liquid structures, a simplified analysis has been employed to
electrolyte batteries, ceramic powders have also been used evaluate the effect of separators in practice [94, 95]. The
in a gel polymer electrolyte (gel polymer separator). The following empirical equation (Eq. 6) has been used
effects of SiO2 [86], Al2O3 [87], TiO2 [88], and MgO [89] extensively in mathematical modeling.
on the electrochemical properties of gellable-polymer-based
Rs ¼ "a  R0 ð6Þ
electrolytes have been studied. It has been shown that the
incorporation of ceramic powder enhances the ionic where Rs is the resistance of the separators filled with the
conductivity, improves the mechanical and thermal stability, liquid electrolyte, R0 is the resistance of the native liquid
and provides high chemical integrity (no liquid leakage) electrolyte, ε is the void volume fraction in a separator, and
even at elevated temperatures. The improvement has been α is the Bruggeman exponent. However, Patel et al. [11]
attributed to the change of the molecular morphologies and found that only porous membranes with idealized morphol-
physical properties of the polymers due to the addition of ogies (spherical or slightly prolate ellipsoids) followed the
the inorganic particles. Abraham et al. [90] prepared a Bruggeman law with an α of about 1.3. Porous networks
ceramic-PVDF electrolyte separator and Hikmet [91] based on other particle morphologies resulted in either a
obtained a porous Li0.35La0.55TiO3-UHMWPE separator. significant increase in α, or a complete deviation from the
However, they stated that addition of the ceramic particles power law, mainly due to the increased tortuosity. The
did not significantly change the conductance. models indicated that spherical or slightly prolate ellipsoi-
Composite separator layers have also been coated on an dal pores should be preferred for batteries where a high-rate
electrode directly to reduce the production cost. Less et al. performance is required.
660 J Solid State Electrochem (2011) 15:649–662

Thorat et al. [96] developed an empirical relationship abuse tolerance and performance. Effort in this area is still
between porosity and the tortuosity of the porous structures. needed in order to further optimize the separator perfor-
Their results demonstrated that the tortuosity-dependent mance in future battery systems.
mass transport resistance in porous separators and electrodes
was significantly higher than predicted by the frequently
used Bruggeman relationship. Tye [16] investigated the Conclusions
effect of tortuosity on the electrolyte diffusion ingress
response of a separator. He showed that separators with Semicrystalline polyolefin-based microporous separators
identical average tortuosities and porosities could have have dominated the Li-ion separators market. They have
different unsteady-state behaviors due to the difference in small thickness, excellent chemical resistance, and good
tortuosity distributions. Newman and Srinivasan [97] studied mechanical properties. However, polyolefin separators
the performance of a natural graphite/LiPF6 in a mixture of generally are not thermally stable. They shrink significantly
EC and DEC/iron phosphate lithium-ion cell. They built a at 90–120 °C and melt at 150–200 °C depending on the
model to optimize the porosity and thickness of the positive type of the polymers. Ceramic-enhanced composite separa-
electrode, the thickness and porosity of the separator, the tors exhibit the advantages of good liquid electrolyte
electrolyte concentration, and the porosity of the negative wettability and the ability to prevent thermal runaway
electrode. Xiao et al. [98] attempted to evaluate the stress conditions in the event of an internal short, which is critical
field in a separator during normal battery cycling. The in high-power applications. Although traditional nonwoven
current industrial requirement on the separator tensile separators have a low production cost and exhibit good
property comes from the battery assembly demands. The thermal stability, they have a large thickness, which reduces
stress distribution in a separator when the battery is under the energy and power densities of a battery, and a large
normal cycling conditions is not well understood. This work mean pore size, which increases the risk of internal shorting
has indicated that the stress is affected by the active material through dendrite growth. Recent techniques for fabricating
properties, electrode geometries, separator wrapping patterns, nonwoven mats comprised of nanosized fibers augment the
charging–discharging protocols, etc. feasibility for producing high-performance separators with
high tortuosity, high porosity, and good thermal stability.
The modification (including ceramic coating, crosslinking,
Future work and functionalization) of polyolefin microporous separators
and the development of alternative polymers for porous
In the light of the above discussion, the goals of the future membranes are also the potential directions for improving
research will be to (1) develop low-cost fabrication the performance of separators.
processes, (2) discover new materials with high thermal
stability, (3) further understand the behavior of separators Acknowledgments The author would like to thank Ion C. Halalay,
Ingrid A. Rousseau, Hamid G. Kia, and Mark W. Verbrugge at
during the operation of the battery, and (4) extend the
General Motors and Jonathon Hitt at Optimal Resources LLC for the
functions of the separator in a battery in order to improve valuable discussions.
the battery performance under all operating conditions
expected in the automotive usage. Abuse tolerance and
cost are especially important for large-scale applications in References
HEV/EV and should drive all these research directions.
The modification of commonly used polyolefin micro- 1. Ehrlich GE (2002) In: Linden D, Reddy TB (eds) Handbook of
porous separators can lead to improved separator perfor- batteries, 3rd edn. McGraw-Hill, New York
mance. The modification can be achieved by coating and 2. Johnson BA, White RE (1998) J Power Sources 70:48–54
3. Djian D, Alloin F, Martinet S, Lignier H, Sanchez JY (2007) J
chemical crosslinking. Ceramic separators have shown
Power Sources 172:416–421
excellent thermal stability and electrolyte wettability. 4. Jansen AN, Kahaian AJ, Kepler KD, Nelson PA, Amine K, Dees
Increasing numbers of ceramic separators have been DW, Vissers DR, Thackeray MM (1999) J Power Sources 81–
commercialized to meet the abuse tolerance and perfor- 82:902–905
5. Orsini F, du Pasquier A, Beaudouin B, Tarascon JM, Trentin M,
mance requirements for HEV/EV applications. Develop-
Langenhuizen N, de Beer E, Notten P (1999) J Power Sources 81–
ment of modern processes, from an interdisciplinary 82:918–921
approach, may lower the production costs. Nonwoven mats 6. Liu HK, Wang GX, Guo Z, Wang J, Konstantinov K (2006) J
of nanosized fibers have the potential to be used as low-cost Nanosci Nanotechno 6:1–15
7. Arora P, Zhang Z (2004) Chem Rev 104:4419–4462
separators with good thermal stability and should be
8. Zhang SS (2007) J Power Sources 164:351–364
developed for battery applications. Numerical modeling 9. Caldwell DL, Poush KA (1984) US Patent 4464238
should assist the design of separators for improved battery 10. Abraham KM (1993) Electrochim Acta 38:1233–1248
J Solid State Electrochem (2011) 15:649–662 661

11. Patel KK, Paulsen JM, Desilvestro J (2003) J Power Sources 53. Sugiyama M, Totsuka H, Mitani S, Takahata M (2007) US Patent
122:144–152 7311994
12. Jeong YB, Kim DW (2004) J Power Sources 128:256–262 54. Kim JY, Kim SK, Lee SJ, Lee SY, Lee HM, Ahn S (2004)
13. Venugopal G, Moore J, Howard J, Pendalwar S (1999) J Power Electrochim Acta 50:363–366
Sources 77:34–41 55. Kritzer P (2006) J Power Sources 161:1335–1340
14. Piatkiewicz W, Rosinski S, Lewinska D, Bukowski J, Judycki W 56. Ashida T, Tsukuda T (2001) US Patent 6200706
(1999) J Membr Sci 153:1–102 57. Wang Y, Zhan H, Hu J, Liang Y, Zeng S (2009) J Power Sources
15. Jena A, Gupta K (2002) Fluid Part Sep J 4:227–241 189:616–619
16. Tye FL (1983) J Power Sources 9:89–100 58. Kamei T, Yamazaki M (2004) US Patent 6730439
17. Liu SJ, Maslivah JH (1996) In: Schramm LL (ed) Suspensions: 59. Suzuki H, Sudo Y (2003) JP-A-2003-142064
fundamentals and applications in the petroleum industry. Amer- 60. Sudou Y, Suzuki H, Nagami S, Ikuta K, Yamamoto T, Okijima S,
ican Chemical Society, Washington Suzuki S, Ueshima H (2007) US Patent 7183020
18. Kim LU, Kim CK (2006) J Polym Sci Polym Phys 44:2025–2034 61. Bansal D, Meyer B, Salomon M (2008) J Power Sources
19. Ma JC, Megahed ES, Stachoviak TJ, Craanen SA, Schneider DA, 178:848–851
Nestler JP (2002) US Patent 6444356 62. Cho TH, Sakai T, Tanase S, Kimura K, Kondo Y, Tarao T, Tanaki
20. Yoneda H, Nishimura Y, Doi Y, Fukuda M, Kohno M (2010) M (2007) Electrochem Solid-State 10:A159–A162
Polym J 42:425–437 63. Kim JR, Choi SW, Jo SM, Lee WS, Kim BC (2005) J
21. Weighall MJ (1991) J Power Sources 34:257–268 Electrochem Soc 152:A295–A300
22. Strathmann H, Kock K (1977) Desalination 21:241–255 64. Lee SW, Choi SW, Jo SM, Chin BD, Kim DY, Lee KY (2006) J
23. Yilmaz L, McHugh AJ (1986) J Membr Sci 28:287–310 Power Sources 163:41–46
24. Reuvers AJ, van der Berg JWA, Smolders CA (1987) J Membr 65. Chu B, Hsaio BS, Fang D (2006) US Patent application
Sci 34:45–65 2006049542
25. Reuvers AJ, Smolders CA (1987) J Membr Sci 34:67–86 66. Um IC, Fang D, Hsiao BS, Okamoto A, Chu B (2004)
26. Radovanovic P, Thiel SW, Hwang ST (1992) J Membr Sci Biomacromolecules 5:1428–1436
65:213–229 67. Armantrout JE, Bryner MA, Davis MC, Kim YM (2006) US
27. Wienk IM, Boom RM, Beerlage MAM, Bulte AMW, Smolders Patent application 2006012084
CA, Strathmann H (1996) J Membr Sci 113:361–371 68. Kim YM, Sung YB, Jang RS, Ahn KR (2009) US Patent 7618579
28. Bottino A, Camera-Roda G, Capannelli G, Munari S (1991) J 69. Peng M, Sun Q, Ma Q, Li P (2008) Microporous Mesoporous
Membr Sci 57:1–20 Mater 115:562–567
29. Zhang SS, Xu K, Foster DL, Ervin MH, Jow TR (2004) J Power 70. Hiroki S, Satoshi N, Hiroyuki H, Takahiro D (2006) Patent
Sources 125:114–118 application WO 06123811
30. Andrieu X, Laurence J (1998) US Patent 5811205 71. Lee YM, Kim JW, Choi NS, Lee JA, Seol WH, Park JK (2005) J
31. Dupasquier A, Tarascon JM (2003) US Patent 6537334 Power Sources 139:235–241
32. Dupasquier A, Warren PC, Culver D, Gozdz AS, Amatucci GG, 72. Lee YM, Choi NS, Lee JA, Seol WH, Cho KY, Jung HY, Kim JW,
Tarascon JM (2000) Solid State Ionics 135:249–257 Park JK (2005) J Power Sources 146:431–435
33. Li ZH, Zhang P, Zhang HP, Wu YP, Zhou XD (2008) Electrochem 73. Pekala RW, Khavari M (2003) US Patent 6586138
Commun 10:791–794 74. Zhang SS, Xu K, Jow TR (2005) J Power Sources 140:361–364
34. Lundquist JT, Lundsager B, Palmer NI, Troffkin HJ, Howard J 75. Cho TH, Tanaka M, Onishi H, Kondo Y, Nakamura T,
(1987) US Patent 4650730 Yamazaki H, Tanase S, Sakai T (2008) J Electrochem Soc
35. Yu WC (1997) US Patent 5691077 155:A699–A703
36. Higuchi H, Matsushita K, Ezoe M, Shinomura T (1995) US Patent 76. Shinohara Y, Tsujimoto Y, Nakano T (1999) US Patent 6447958
5385777 77. Augustin S, Hennige VD, Horpel G, Hying C (2002) Desalination
37. Ihm DW, Noh JG, Kim JY (2002) J Power Sources 109:388–393 146:23–28
38. Noumi S, Yamamura Y, Nakayama S (2010) US Patent 7704597 78. Henninge V, Hying C, Horpel G (2010) US Patent 7807286
39. Yamamoto K, Fujita S, Uetani Y, Noumi S, Emori H, Yamamura 79. Hennige V, Hying C, Horpel G, Novak P, Vetter J (2010) US
Y (2003) US Patent 6559195 Patent 7709140
40. Johnson MB, Wilkes GL (2002) J Appl Polym Sci 83:2095–2113 80. Kim SK, Sohn JY, Park JH, Jang HM, Shin BJ, Lee SY, Hong JH
41. Beard KW (2006) US patent application 20060081530 (2010) US Patent 7709152
42. Takahashi T, Tateno T (1999) US Patent 5856426 81. Park JH, Lee SY, Hong JH, Nam MJ, Yoo JA, Kim SS, Han CH
43. Ohya S, Fujii Y, Yao S, Asano Y, Nakayama K, Fukunaga K (2010) US Patent 7695870
(2003) US Patent 6565962 82. Seo DJ, Kim K, Hong JH, Sohn JY, Lee SY, Ahn SH (2007) US
44. Xiao LF, Ai XP, Cao Y, Wang YD, Yang HX (2005) Electrochem Patent application 20070122716
Commun 7:589–592 83. Lee SY, Seo DJ, Sohn JY, Kim SK, Hong JH, Kim YS Jang HM
45. Chen G, Richardson TS (2004) Electrochem Solid-State 7:A23–A26 (2009) US Patent 7638241
46. Feng JK, Ai XP, Cao YL, Yang HX (2006) J Power Sources 84. Katayama H, Kojima E, Aoyama S, Sato Y (2008) Patent
161:545–549 application WO08029922
47. Thomas-Alyea KE, Newman J, Chen G, Richardson TJ (2004) J 85. Tomi H, Sato E, Murayama S, Itou F, Imoto H (2006) Japan
Electrochem Soc 151:A509–A521 Patent JP3831017
48. Ko JM, Min BG, Kim DW, Ryu KS, Kim KM, Lee YG, Chang 86. Lee KH, Lee YG, Park JK, Seung DY (2000) Solid State Ionics
SH (2004) Electrochim Acta 50:367–370 133:257–263
49. Gao K, Hu GX, Yi TF, Dai CS (2006) Electrochim Acta 87. Appetecchi GB, Romagnoli P, Scrosati B (2001) Electrochem
52:443–449 Commun 3:281–284
50. Gineste JL, Pourcelly G (1995) J Membr Sci 107:155–164 88. Chung SH, Wang Y, Persi L, Croce F, Greenbaum SG, Scrosati B,
51. Yao ZP, Ranby B (1990) J Appl Polym Sci 41:1469–1478 Plichta E (2001) J Power Sources 97–98:644–648
52. Urairi M, Tachibana T, Matsumoto K, Shinomura T, Iida H, 89. Kumar B, Scanlon L, Marsh R, Mason R, Higgins R, Baldwin R
Kawamura K, Yano S, Ishida O (1996) US Patent 5558682 (2001) Electrochim Acta 46:1515–1521
662 J Solid State Electrochem (2011) 15:649–662

90. Abraham KM, Koch VR, Blakley TJ (2000) J Electrochem Soc 94. Doyle M, Newman J, Gozdz AS, Schmutz CN, Tarascon JM
147:1251–1256 (1996) J Electrochem Soc 143:1890–1903
91. Hikmet RAM (2001) J Power Sources 92:212–220 95. MacMullin RB, Muccini GA (1956) AICHE J 2:393–403
92. Less GB, Knapp A, Babinec SJ (2009) US Patent application 96. Thorat IV, Stephenson DE, Zacharias NA, Zaghib K, Harb JN,
20090155678 Wheeler DR (2009) J Power Sources 188:592–600
93. Kim M, Nho YC, Park JH (2010) J Solid State Electrochem 97. Srinivasan V, Newman J (2004) J Electrochem Soc 151:A1530–A1538
14:769–773 98. Xiao X, Wu W, Huang X (2010) J Power Sources 195:7649–7660