Journal of Membrane Science 526 (2017) 315–322

Contents lists available at ScienceDirect

Journal of Membrane Science

journal homepage: www.elsevier.com/locate/memsci

Effect of polymer and ceramic morphology on the material and MARK

electrochemical properties of electrospun PAN/polymer derived ceramic
composite nanofiber membranes for lithium ion battery separators
⁎
Soshana A. Smitha, Brian P. Williamsb, Yong Lak Joob,
a
Department of Fiber Science and Apparel Design, Cornell University, Ithaca, NY 14853, USA
b
School of Chemical & Biomolecular Engineering, Cornell University, Ithaca, NY 14853, USA

A BS T RAC T

Electrospun polyacrylonitrile (PAN)/ceramic fibers were created by combining the PAN polymer and various
room temperature curable organopolysilazane (OPSZ) precursors in a one step electrospinning process. The
resultant PAN/ceramic (70:30 wt%) hybrid fiber mats consisted of well-formed, uniform fibers with evenly
distributed ceramic, both in the longitudinal and cross-sections of the fibers. The presence of the tetraethyl
orthosilicate (TEOS) pendant chain attached to the polysilazane (PSZ) backbone impacted the ability of the
ceramic to form a continuous ceramic network within the polymer network. Results show that the membranes
with an interconnected ceramic domain showed superior mechanical and electrochemical properties. The
interconnected ceramic domains affect the polymer's ability to crystallize, leading to an increase in amorphous
regions within the fiber. The increase in amorphous regions leads to increased electrolyte uptake, superior ionic
conductivity and excellent cycling performance. Batteries containing 40 wt% TEOS:PSZ and 20 wt% TEOS:PSZ
separators showed greater than 90% capacity retention at a C/5 discharge rates after 100 cycles. C-rate tests
show a similar trend with 40 wt% TEOS:PSZ and 20 wt% TEOS:PSZ membranes maintaining greater capacity
retention even at high charge rates.

1. Introduction
Lithium ion batteries (LIB) play a crucial role in the expanding
popularity of portable electronic devices (e.g. digital camera, laptop,
smartphone) and the electronic vehicles (EV) market, due to their high
energy density, light-weight, long lifespan, and environmental friendli-
ness [1–4]. As our use of these devices continues to increase, the need
for more efficient lithium ion batteries becomes more apparent. The
separator, a membrane which lies between the cathode and anode in
electrochemical cell, serves as a physical barrier to prevent short
circuiting. Though the separator must be present within a small
compact battery, it should be as thin as possible so as to reduce cell
resistance and also highly porous to aid in ion transport. In addition,
separators should be mechanically, thermally, and electrochemically
stable over the operative temperature range of the battery [5].
Commercially, the most widely used separators are polyolefin mem-
branes, which fulfill some of the previously mentioned criteria however
they also possess many disadvantages. Due to the hydrophobic nature
of polyolefin polymers used to make the membranes, there are issues
with low electrolyte uptake and wettability [6]. These combined factors
of poor wettability and low porosity lead to poor ionic transport over
the lifetime of the early batteries containing polyolefin separators.
In order to improve the ionic transport, researchers first increased
the porosity of separator membranes by producing non-woven mem-
branes made by the electrospinning technique [7–9]. Electrospun
membranes have been made from a variety of polymers including
polyethylene terephthalate (PET) [10], polyacrylonitrile (PAN) [11,12],
polyvinylidene fluoride (PVDF) [13] and polyimides (PI) [14–16].
Electrospun membranes result in a highly porous fiber network with
an average porosity of 70–80%, with Jung et al. [11] reporting
porosities as high as 90%. This increase in porosity is correlated with
an increase in ionic conductivity due to a higher effective diffusivity of
the electrospun membranes. In addition, the use of less hydrophobic
materials leads to increases in electrolyte uptake and retention. All
these factors lead to batteries with higher discharge capacity and
superior capacity retention compared to Celgard.
Both commercially and in academic institutions, researchers are
incorporating inorganic ceramic fillers and coating to their polymer
separators. The most commonly used ceramic are aluminum oxide
(Al2O3) [17–21] and silicon dioxide (SiO2) [22–25]. These polymer/

⁎
Corresponding author.
E-mail address: ylj2@cornell.edu (Y.L. Joo).

http://dx.doi.org/10.1016/j.memsci.2016.12.052
Received 26 September 2016; Received in revised form 22 December 2016; Accepted 24 December 2016
Available online 25 December 2016
0376-7388/ © 2016 Elsevier B.V. All rights reserved.
S.A. Smith et al.
ceramic composite membranes are made by two methods: 1) incorpor-
ating the ceramic directly into the polymer matrix [11,26,27], 2)
coating the membrane [28]. Membranes with ceramic have shown
superior electrochemical performance compared to the neat polymer
membranes and polyolefin separators. The improved electrochemical
performance is attributed to either ceramics deterring polymer crystal-
lization or contributing to a highly conductive interfaces between the
ceramic and polymer [22,29]. Ceramics can also increase the surface
area of the membrane by introducing mesopores (~2–50 nm) which
dramatically increases the ionic conductivity [11,27]. However, the
dip-coating method can lead to issues of ceramic delamination which
could lessen the impact of the ceramic in the battery [30]. Moreover,
the use of particles within the polymer matrix requires extensive
sonication to ensure good ceramic particle dispersion. In addition,
when ceramic is incorporated into the polymer matrix, the amount of
ceramic that can be incorporated into the polymer is limited. After a
certain percentage, ceramic particles will not disperse well leading to
aggregation and a reduction in ionic conductivity and cycling perfor-
mance [11].
For our current research, we aim to study the role of internal fiber
morphology on the material and electrochemical properties of hybrid
polymer/ceramic membranes. Past works have found that changing the
ceramic nanoparticle loading, in a sample, changes the internal
morphology of the polymer in the fiber. As mentioned previously,
researchers typically observe that the ceramic interferes with the
polymer's ability to crystallize [11,31–33]. This reduction in fiber
crystallinity leads to large changes in ionic conductivity and overall
electrochemical performance. Work done by Smith et al. shows a
similar trend when using ceramic precursors [34]. Previous works
using similar liquid ceramic precursors allowed researchers to make
membranes with a much higher percentage of ceramic compared to
using ceramic nanoparticles [34,35]. The use of ceramic precursors
leads to a range of ceramic morphology ranging from discrete domains
to interconnected networks with various loadings of ceramic precursor
in the fiber. By fixing the ceramic loading within the fiber while
changing the ceramic morphology within the PAN fiber, we aim to
investigate how different ceramic morphologies will affect the mor-
phology of the PAN polymer. We will them aim to study how these
changes in morphology affect the material and electrochemical proper-
ties of our membrane.
In this research, polymer/ceramic non-wovens were created via a
one-step electrospinning process using three different
organopolysilazane(OPSZ). The three precursors have a similar che-
mical backbone but differ on the ratios of the tetraethyl orthosilicate
(TEOS) pendant chain attached to the polysilazane (PSZ) backbone. All
OPSZ solutions are miscible in the PAN/DMF polymer solution
allowing the OPSZ to be mixed easily in the solution. This creates a
solution that is easy to electrospin into polymer/ceramic nanofibers
without the need for sonication for further post treatment. The
presence of TEOS chains enables the ceramic to form an independently
connected ceramic network within the polymer differentiating it from
the discrete ceramic materials that have been used in the past. TEM
images show that the resultant fibers contain both ceramic inside the
fiber as well as a sheath layer on the outside. In this work, even though
the ceramic loading is the same, there are two distinctly different
polymer/ceramic morphologies. We aim to show that the internal
morphology of hybrid fiber has a major impact on the electrochemical
properties of membranes.
2. Experimental
2.1. Material preparation
The PAN/ceramic non-woven separator was prepared using the
electrospinning method. The electrospinning solution was prepared by
combining polyacrylonitrile (PAN, MW =200,000, Polysciences) with
316
Journal of Membrane Science 526 (2017) 315–322
N,N-Dimethylformamide (DMF) to form a 8 wt% solution. The solution
was placed in a 95 °C oven for 12 h to thoroughly dissolve the polymer.
The PAN/DMF solution was then left to cool to room temperature.
Ceramic precursors, organopolysilazanes (EMD Performance
Materials), were added to obtain 30 wt% ceramic in relation to the
PAN. The mixture was stirred on a vortex mixer for two minutes to
ensure thorough mixing of the components. The solution was left
sitting for 10 min to allow any bubbles within the solution to disappear
before spinning the solution. The PAN/OPSZ solution was electrospun
using a 20-gauge needle at a flow rate of 0.6 mL/h for both the 20 wt%
and 40 wt% TEOS:OPSZ solutions and 0.72 mL/h for the 0 wt%
TEOS:PSZ solution. A 17.5 kV voltage and 16 cm distance between
the needle and the collector plate was used for all samples. The
collected samples were then annealed at 120 °C for 12 h to ensure
the removal of all residual solvent.
2.2. Characterization
The fiber morphology and diameter were examined using scanning
electron microscopy (Tecsa Mira3 Field Emission SEM) with an
accelerating voltage of 5 kV. The average fiber diameters were deter-
mined by choosing 50 randomly selected fibers using Image J. In order
to investigate the structure of the ceramic within the fibers, membranes
were soaked in DMF for 24 h in order ensure the total removal of PAN.
Samples were then dried and imaged using the Tescan SEM. The
morphology of the cross section of the fibers was investigated by
transmission electron microscopy (FEI Tecnai G2 T12 Spirit TEM
STEM). Thermal stability was characterized using differential scanning
calorimetry (DSC, TA Instrument Q200) under nitrogen from 40 °C to
350 °C at a 10 °C/min. Wide angle X-ray diffraction (XRD) patterns
were recorded from 10° to 70° for all prepared electrospun membranes
using Theta-Theta X-ray Diffractometer (Scintag) with Cu radiation
(λ=1.5405 Å; 40 mA and 40 kV). XRD was also used to estimate fiber
crystallinity. Elemental composition of ceramic component was done
using X-ray photoelectron spectroscopy (XPS, Surface Science
Instruments SSX-100). Electrolyte uptake was calculated using the
following relation:
Ma − Mi
Uptake (%) = ×100
Mi (1)
where Ma and Mi are the masses of the wet and dry membrane
respectively. The porosity (P) of the fabricated membranes was
calculated by using the measured data and the density of material
with the equation below [36]:
W
P =1 − ( )
ρ·V (2)
where W is the dried membrane weight, ρ the membrane density, and
V the apparent volume of the membrane. The membrane thickness was
measured by a digital thickness gauge.
Electrochemical tests are carried out in a 2302 coin cell. Cell
impedance was measured by sandwiching the separator membrane
between two lithium disks inside a cell. The cells were tested on a
PARSTAT 4000 (Princeton Applied Research), with the impedance
measurements at an amplitude of 5 mV over a frequency range of
20,000 kHz to 1Hz.
Full cells consisted of a LiCoO2 cathode (MTI Corporation), a
graphite anode (MTI Corporation), and LiPF6-EC/DMC/DEC electro-
lyte; cycling was carried out at a voltage range between 2.5 and 4.2 V at
0.2 C and 0.5 C charge and discharge rates. Discharge rate capabilities
were tested by charging cells to 4.2 V at 0.1 C rate for the first and last
cycle and 0.2 C rate for all other cycles. Cells were discharged at 0.1 C,
0.2 C, 0.5 C, 1 C, 2 C, 5 C and 0.1 C respectively.
S.A. Smith et al. Journal of Membrane Science 526 (2017) 315–322

Table 1 within the 40 wt% TEOS:PSZ and 20 wt% TEOS:PSZ fibers form a
Weight change in polymer/ceramic hybrid fibers after annealing. connected network independent of the polymer matrix. Further
magnification of the ceramic fiber surface (Fig. 2d-e) reveals the
Sample membrane Initial weight Final weight Weight change (%)
(mg) (mg) ceramic fibers appear to be porous. The porosity of the ceramic surface
indicates the PAN polymer and the ceramic intertwine along the fiber
0 wt% TEOS:PSZ 13.20 13.07 1 axis forming a co-continuous network. In contrast to the other two
20 wt% TEOS:PSZ 18.04 17.62 2
fibers, the 0 wt% TEOS:PSZ ceramic does not form a continuous
40 wt% TEOS:PSZ 10.87 10.73 1
network within the fiber. Instead the 0 wt% TEOS:PSZ ceramic forms
small discrete domains. The inability of 0 wt% TEOS:PSZ to form a
3. Results and discussion connected ceramic network could be due to the lack of TEOS branches
on the original precursor backbone. The TEOS branches dramatically
It is important to keep the amount of precursor in each sample the increase the amount of available reactive sites along the precursor
same to make sure that there is a fair comparison between the backbone to react with other OPSZ chains. The increased reaction sites
membranes because previous researchers have shown that the amount enabled the connected ceramic network seen in the 40 wt% TEOS:PSZ
of precursor used can affect membrane properties [11,25,26]. After and 20 wt% TEOS:PSZ samples. In addition, the 0 wt% precursors have
electrospinning, each membrane was weighed before and after being the fastest curing rate among the precursors; the fast curing rate of the
placed in the 120 °C oven for 12 h. There was minimal weight change 0 wt% precursors allows it to gel and solidify before it can create a
in the membranes (Table 1), showing there was no thermal degradation continuous structure.
of the ceramic or the polymer at that temperature. The minimal weight DSC measurements (Fig. 3) were done to investigate how the
loss observed is due to the evaporation of water from the membrane. presence of the ceramics affected the thermal properties of the polymer
During the electrospinning process, all three precursors undergo a network. Unlike most polymer, PAN undergoes a degradation before
moisture crosslinking converting the three OPSZ precursors to poly- melting leading to the cyclization of the PAN polymer backbone [39].
mer-derived ceramics [37]. OPSZ precursors can undergo room Between 200 and 300 °C, a free radical reaction is initiated in the
temperature curing in two ways: 1) hydrolysis reaction via the TEOS nitrile groups of the polymer leading to a network of hexagonal carbon-
pendant chains, 2) reaction along to Si-N backbone similar to the sol- nitrogen rings. The cyclization temperature (Tc) of the PAN polymer
gel reaction seen with perhydropolysilazane (PHPS) [38]. After a DMF without the presence of ceramic is 290.1 °C. The cyclization tempera-
rinse of the membranes to remove PAN, XPS analysis was done to ture of the PAN within the 0 wt% TEOS:PSZ fiber is 295 °C indicating
determine the chemical composition of the resulting ceramic inside inclusion of 0 wt% TEOS:PSZ has a positive impact on the thermal
each fiber mat. The chemical composition of the three OPSZ precursors stability of the fibers. Discrete ceramic particles within the fiber, within
and resulting polymer-derived ceramics are shown in Table 2. As the the fiber act as an inhibiting agent, preventing the free radicals from
precursors cure, there is a reduction in the percentage of nitrogen and getting to other PAN polymer chains thus increasing the cyclization
an increase in the percentage of oxygen due to the precursors reactions temperature [40,41]. On the other hand, there is a 13 °C reduction in
with moisture in the atmosphere. Fig. 1 shows external and internal the Tc temperature for the 20 wt% TEOS:PSZ fiber, reducing Tc to
morphology of the various nonwoven fiber PAN/PDC mats. SEM 277 °C. Cyclization first occurs within the amorphous region of the
images (Fig. 1a,d,g) show well-formed fibers, free of beads and PAN fiber due to the loose orientation of the polymer chains before
droplets. Fiber characteristics including fiber diameter and porosity moving to the crystalline regions [42,43]. A decrease in the cyclization
are presented in Table 3. There is a decrease in fiber diameter with temperature in comparison to the neat PAN fiber is a strong indication
TEOS ratio along the precursor backbone. This suggests that the gelling of a larger disorganization of the polymer allowing for free radicals to
of the TEOS chains during the spinning process helps reduce fiber better degrade the polymer chains. In addition, previous researchers
entanglement leading to smaller fibers. This is similar to previous have shown that the decrease in cyclization temperature and reduction
works, which show that the increase in SiO2 results in a decrease in in heat of enthalpy can also be caused by an outside initiator to nitrile
fiber diameter due to repulsive force minimizing polymer entanglement cyclization. The reduction in the onset temperature is most likely
[11,25]. caused by thermal degradation of the 20 wt% membrane, lending free
TEM images reveal the ceramic is evenly dispersed within the fibers radicals to the system to help propagate the thermal degradation of the
along the fiber axis for all samples. TEM images also show the ceramic polymer.
forms a continuous sheath layer (approximately 5 nm thick) around all In addition to good thermal property retention, separators must
fibers followed by a space void of ceramic towards the edge of the fiber. also be sufficient mechanical properties to withstand any punctures or
Towards the fiber center, there is also dispersed ceramic present within tension the membranes might experience. Fig. 4 shows the typical
the polymer matrix. Domain sizes of the ceramic within the fiber stress/strain curve of the three membranes. 20 wt% has the best
decrease with increasing TEOS percentage; 0 wt% TEOS:PSZ fibers highest strength at 14.8 MPa. This is followed by the 40 wt% mem-
have domain size of 40 ± 3 nm, 20 wt% and 40 wt% TEOS:PSZ fiber brane which has a ultimate strength of 12.2 MPa. This membrane
have domain sizes of 25 ± 2 nm and 38 ± 3 nm respectively. Removal of begins to exhibit brittle behavior due to the high ceramic loading; it's
the PAN polymer with a DMF rinse reveals the structure of the ceramic breaks at the yield point before entering the ductile region. The poorest
within the polymer network (Fig. 2). Fig. 2a. and b., show the ceramic performing membrane is 0 wt% TEOS:PSZ membrane strength at
11.2 MPa. These results suggest the interconnected ceramic network
Table 2
within the 40 wt% TEOS:PSZ and 20 wt% TEOS:PSZ offers added
Elemental composition of OPSZ precursors and resultant ceramic inside the fiber after mechanical stability to the membranes.
electrospinning. Fig. 5 shows the XRD of various PAN/PDC membranes.
Electrospun PAN membranes typically have peaks at 17° and 28°,
Sample Si (%) O (%) C (%) N (%)
indicating hexagonal packing of PAN chains [44,45]. 40 wt%
40 wt% TEOS:PSZ Before 32 14 33 12 TEOS:PSZ and 20 wt% TEOS:PSZ show a broadening of the peak at
40 wt% TEOS:PSZ After 15 25 50 9 17° which indicates a disruption in the crystalline structure and hence
20 wt% TEOS:PSZ Before 38 7 29 17 an increase in the amorphous regions. On the other hand, there is an
20 wt% TEOS:PSZ After 10 19 57 13
increase in the crystalline regions and crystal size for the 0 wt%
0 wt% TEOS:PSZ Before 44 0 25 22
0 wt% TEOS:PSZ After 21 29 43 6 TEOS:PSZ fibers compared to neat PAN polymer. This indicates that
the ceramic domains in 0 wt% TEOS:PSZ fibers act as nucleating

317
S.A. Smith et al. Journal of Membrane Science 526 (2017) 315–322

Fig. 1. SEM images and TEM images cross sectional and longitudinal cross section of various separator membranes. (a–c) 40 wt% (d–f) 20 wt% (g–i) 0w%.

Table 3 sample has the highest electrolyte uptake with an uptake of 1189 ± 32%
Fiber diameter, porosity, pore size distribution and electrolyte uptake of various followed by 20 wt% with an uptake of 1066 ± 24%. This difference in
separator membranes.
uptake is due to the difference in amorphous regions in the two
Sample Fiber Porosity (%) Pore Size Electrolyte membranes. The 40 wt% TEOS:PSZ sample has a 21% crystallinity
Diameter Distribution Uptake (%) compared to 24% of the 20 wt% TEOS:PSZ sample. However there is a
(nm) (nm) large decrease in uptake for the 0 wt% TEOS:PSZ membranes with an
0 wt% 660 ± 81 82 ± 4 619 ± 76 801 ± 75
uptake value of 801 ± 75%, this is due to the increased crystalline
TEOS:PSZ regions of 0 wt% TEOS:PSZ fibers compared to the other two mem-
20 wt% 600 ± 130 82 ± 3 490 ± 107 1066 ± 24 branes.
TEOS:PSZ
40 wt% 490 ± 69 82 ± 5 463 ± 72 1109 ± 32
4. Electrochemical performance
TEOS:PS-
Z
The ionic conductivity of a li-ion separator represent the transport
of ions during the charge-discharge process of the batteries. The ionic
agents allowing for slight increase in crystallinity. However, the conductivity of the three separators is presented in Fig. 6. Ionic
interconnected ceramic network disrupts the crystallinity during the conductivity tests show 40 wt% TEOS:PSZ membranes have the high-
spinning process, leading to a reduction in crystallinity for the other est calculated ionic conductivity at 1.04 ± 0.05 mS/cm followed by
fibers. 20 wt% TEOS:PSZ with a ionic conductivity of 0.59 ± 0.04 mS/cm and
Membranes are soaked in 1 M LiPF6/DC/DMC/DEC to investigate 0 wt% TEOS:PSZ at 0.44 ± 0.07 mS/cm. Increase in ionic conductivity
the ability to uptake and retain electrolyte. The ability of the membrane is usually attributed to increase in amorphous regions of the polymer
separator to uptake electrolyte is vital in reducing the internal ionic due to inorganic inclusion, and ceramics acting as the Lewis acid–base
resistance of the cell. Results in Table 3 show 40 wt% TEOS:PSZ center that promote salt disassociation which could free more cations

318
S.A. Smith et al. Journal of Membrane Science 526 (2017) 315–322

Fig. 2. SEM images of various membranes after removal of PAN using DMF rinse (a,d) 0 wt%, (b,e) 20 wt%, (c,f) 40 wt%.

Fig. 3. DSC thermograms of various separator membranes.

Fig. 4. Typical stress-strain curves of various separator membranes.

[46]. In Fig. 6, there is an indirect correlation between ionic con-

initial capacity is due to the high ionic conductivity of the PAN/PDC
ductivity and PAN polymer matrix crystallinity and crystal size. This
membranes. After 100 cycles, the membrane with the highest ionic
confirms that the ceramic inclusion disrupts the polymer matrix to
conductivity, 40 TEOS: PSZ has the highest capacity retention, 93%
varying degrees thus increasing the ionic conductivity. The decrease in
followed by 20 wt% TEOS: PSZ with a 91% capacity retention; 0 wt%
the degree of crystallization allows for the formation of more tunnels
TEOS: PSZ capacity has a 88% capacity retention after 100 cycles. A
allowing for greater Li+ migration. Electrolyte uptake results (Table 3)
pure PAN membrane has a capacity retention of 86%. All electrospun
also show 0 wt% has the lowest uptake compared to the other
membranes outperform the microporous membrane, Celgard, which
membranes due to have the larger pores sizes and fiber diameter.
showed 82% retention for a 0.2 C charge rate. The cells, operated at a
The reduced pore sizes of the other membranes allowed for a better
0.5 C charge rate, show a similar trend to the cells operated at 0.2 C
entrapment of the electrolyte leading to improved ionic conductivity.
rate, with 40 wt% TEOS:PSZ and 20 wt% TEOS:PSZ showing the best
Fig. 7 shows the cycling performance comparison of the three PAN/
capacity retention. Starting with an initial capacity of 121 mAh/g, the
PDC separators, at 0.2 C and 0.5 C charging rates in LiCoO2/graphite
40 wt% TEOS:PSZ is able to retain 90% of its initial discharge capacity,
full cells, in comparison to Celgard and PAN membranes. At the 0.2 C
ending with final capacity of 109 mAh/g. The 20 wt% TEOS:PSZ
charging rate, all PAN/ceramic separators start off at a similar initial
separator was able to retain 87% retention, ending at 105 mAh/g after
capacity, approximately 134 mAh/g. Both the PAN and Celgard
100 cycles. Analogous to 0.2 C case, 0 wt% TEOS:PSZ separator shows
membranes start off at a lower capacity at 130 mAh/g. The higher
the fastest decay ending with 84% retention. The superior performance

319
S.A. Smith et al. Journal of Membrane Science 526 (2017) 315–322

Fig. 5. XRD patterns of electrospun PAN/PDC composite membranes with various

ceramic contents.

Fig. 7. The cycle performances of LiCoO2/graphite full cells with various separator
Fig. 6. Ionic conductivity of various membranes in relation to polymer crystal size and membranes at a) 0.2 C b) 0.5 C.
crystallinity.

of the 20 wt% and 40 wt% membranes is due to their excellent

electrolyte uptake and ionic conductivity in relation to the 0 wt%
membrane and Celgard. The presence of Si-O-Si on the surface of the
fiber can also enable greater salt disassociation of LiPF6, producing a
higher concentration of Li+ ions in the cell for the polymer/ceramic
hybrid materials compared to the pristine PAN polymer and Celgard.
Rate capability testing for coin cells with the three separators can be
seen in Fig. 8. For the mild charge rates of 0.1 C, all three separators
show similar results. However as the cells are tested using more
strenuous charging conditions, there is a more noticeable difference
between the cells. At higher charge rates, the 40 wt% membrane shows
the best capacity retention due to its superior ionic conductivity. At the
most strenuous discharge capacity, 5 C, the differences between the
membranes becomes more apparent. Compared to the average C/10
discharge capacity, 132 mAh/g, there is a 55%, 44% and 33% retention
at 5 C for the 40 wt% TEOS:PSZ, 20 wt% TEOS:PSZ, and 0 wt%
TEOS:PSZ membranes respectively. The Celgard membrane maintains
28% retention, underperforming compared to the other membranes. As Fig. 8. Results of rate capability tests for the LiCoO2/graphite full cells with various
the cells are returned to a C/10 discharging rate, all PAN/PDC separator membranes.

membranes were able to recover a significant amount of their initial

capacity. The PAN and the Celgard membranes also show poor
recoverability when returned to a C/10 discharge rate. Celgard has a
capacity recovery of 70%, while PAN shows a capacity recovery of 82%.
However, all PAN/PDC membranes show capacity recovery of greater
than 95% due to the inclusion of ceramics into the polymer network.
5. Conclusions
In this work, we created three different types of electrospun
polycrylonitrile/polymer derived ceramic nanofibers using three dif-
ferent types of organopolysilazane precursors. The use of precursors
allows for the creation of a polymer-ceramic bi-component in a single
step process. The three precursors have a similar chemical backbone

320
S.A. Smith et al.
but differ on the ratio of the Si(OCH2CH3)3 pendant chain attached to
the polysilazane backbone. The presence of the Si(OCH2CH3)3 chain
enabled the ceramic to form an independently connected network
within the polymer matrix. For the precursor without the
Si(OCH2CH3)3 pendant chain, such a co-continuous polymer ceramic
network was not formed, leading to discrete ceramic domains in the
polymer matrix. These small domains acted as nucleating agents
increasing the crystallinity of the PAN regions. On the other hand,
the membranes with an interconnected ceramic portion showed
reduced crystallinity and crystal size. The curing of the ceramic
interrupts the polymers ability to form a well-ordered structure. The
connected polymer network results in fiber membranes with superior
mechanical and electrochemical performance compared to the mem-
brane that had discrete ceramic domains in the polymer network.
Acknowledgements
This work made use of the Cornell Center for Materials Research
(CCMR) Shared Facilities which are supported through the NSF
MRSEC Program (DMR-1120296). This work was partly supported
by AZ Electronic Materials (Cornell OSP No. 70729) and Axium
Nanofibers (Cornell OSP No. 70530). Soshana Smith acknowledges
the Cornell Provost Diversity Fellowship.
References
[1] J.-M. Tarascon, M. Armand, Issues and challenges facing rechargeable lithium
batteries, Nature 414 (2001) 359. http://dx.doi.org/10.1038/35104644.
[2] F.L. Guangmin Zhou, Progress in flexible lithium batteries and future prospects,
Energy Amp. Environ. Sci. 7 (2014) 1307. http://dx.doi.org/10.1039/c3ee43182g.
[3] B. Scrosati, J. Garche, Lithium batteries: status, prospects and future, J. Power
Sources 195 (2010) 2419–2430. http://dx.doi.org/10.1016/j.jpow-
sour.2009.11.048.
[4] P. Arora, Z. (John) Zhang, Battery separators, Chem. Rev. 104 (2004) 4419–4462.
http://dx.doi.org/10.1021/cr020738u.
[5] X. Huang, Separator technologies for lithium-ion batteries, J. Solid State
Electrochem. 15 (2010) 649–662. http://dx.doi.org/10.1007/s10008-010-1264-9.
[6] S.S. Zhang, A review on the separators of liquid electrolyte Li-ion batteries, J.
Power Sources 164 (2007) 351–364. http://dx.doi.org/10.1016/j.jpow-
sour.2006.10.065.
[7] K. Gao, X. Hu, C. Dai, T. Yi, Crystal structures of electrospun PVDF membranes and
its separator application for rechargeable lithium metal cells, Mater. Sci. Eng. B 131
(2006) 100–105. http://dx.doi.org/10.1016/j.mseb.2006.03.035.
[8] Y.-W. Ju, J.-H. Park, H.-R. Jung, W.-J. Lee, Electrochemical properties of
polypyrrole/sulfonted SEBS composite nanofibers prepared by electrospinning,
Electrochim. Acta 52 (2007) 4841–4847. http://dx.doi.org/10.1016/j.electac-
ta.2007.01.028.
[9] Y. Ding, W. Di, Y. Jiang, F. Xu, Z. Long, F. Ren, P. Zhang, The morphological
evolution, mechanical properties and ionic conductivities of electrospinning
P(VDF-HFP) membranes at various temperatures, Ionics 15 (2009) 731–734.
http://dx.doi.org/10.1007/s11581-009-0326-4.
[10] J. Hao, G. Lei, Z. Li, L. Wu, Q. Xiao, L. Wang, A novel polyethylene terephthalate
nonwoven separator based on electrospinning technique for lithium ion battery, J.
Membr. Sci. 428 (2013) 11–16. http://dx.doi.org/10.1016/j.memsci.2012.09.058.
[11] H.-R. Jung, D.-H. Ju, W.-J. Lee, X. Zhang, Electrospun hydrophilic fumed silica/
polyacrylonitrile nanofiber-based composite electrolyte membranes, Electrochim.
Acta 54 (2009) 3630–3637. http://dx.doi.org/10.1016/j.electacta.2009.01.039.
[12] T.-H. Cho, M. Tanaka, H. Onishi, Y. Kondo, T. Nakamura, H. Yamazaki, S. Tanase,
T. Sakai, Battery performances and thermal stability of polyacrylonitrile nano-fiber-
based nonwoven separators for Li-ion battery, J. Power Sources 181 (2008)
155–160. http://dx.doi.org/10.1016/j.jpowsour.2008.03.010.
[13] Y. Liang, S. Cheng, J. Zhao, C. Zhang, S. Sun, N. Zhou, Y. Qiu, X. Zhang, Heat
treatment of electrospun polyvinylidene fluoride fibrous membrane separators for
rechargeable lithium-ion batteries, J. Power Sources 240 (2013) 204–211. http://
dx.doi.org/10.1016/j.jpowsour.2013.04.019.
[14] X. Liang, Y. Yang, X. Jin, Z. Huang, F. Kang, The high performances of SiO2/
Al2O3-coated electrospun polyimide fibrous separator for lithium-ion battery, J.
Membr. Sci. 493 (2015) 1–7. http://dx.doi.org/10.1016/j.memsci.2015.06.016.
[15] Y.-E. Miao, G.-N. Zhu, H. Hou, Y.-Y. Xia, Electrospun polyimide nanofiber-based
nonwoven separators for lithium-ion batteries, J. Power Sources 226 (2013) 82–86.
http://dx.doi.org/10.1016/j.jpowsour.2012.10.027.
[16] Q. Wang, W.-L. Song, L. Wang, Y. Song, Q. Shi, L.-Z. Fan, Electrospun polyimide-
based fiber membranes as polymer electrolytes for lithium-ion batteries,
Electrochim. Acta 132 (2014) 538–544. http://dx.doi.org/10.1016/j.electac-
ta.2014.04.053.
[17] W.-K. Shin, Y.-S. Lee, D.-W. Kim, Hybrid composite membranes based on
polyethylene separator and Al2O3 nanoparticles for lithium-ion batteries, J.
Nanosci. Nanotechnol. 13 (2013) 3705–3710.
321
Journal of Membrane Science 526 (2017) 315–322
[18] H. Wang, H. Li, L. Yu, Y. Jiang, K. Wang, Synthesis of porous Al2O3-PVDF
composite separators and their application in lithium-ion batteries, J. Appl. Polym.
Sci. 130 (2013) 2886–2890. http://dx.doi.org/10.1002/app.39499.
[19] H.-S. Jeong, S.C. Hong, S.-Y. Lee, Effect of microporous structure on thermal
shrinkage and electrochemical performance of Al2O3/poly(vinylidene fluoride-
hexafluoropropylene) composite separators for lithium-ion batteries, J. Membr.
Sci. 364 (2010) 177–182. http://dx.doi.org/10.1016/j.memsci.2010.08.012.
[20] H. Liu, J. Xu, B. Guo, X. He, Effect of Al2O3/SiO2 composite ceramic layers on
performance of polypropylene separator for lithium-ion batteries, Ceram. Int. 40
(2014) 14105–14110. http://dx.doi.org/10.1016/j.ceramint.2014.05.142.
[21] T. Lee, W.-K. Kim, Y. Lee, M.-H. Ryou, Y.M. Lee, Effect of Al2O3 coatings prepared
by RF sputtering on polyethylene separators for high-power lithium ion batteries,
Macromol. Res. 22 (2014) 1190–1195. http://dx.doi.org/10.1007/s13233-014-
2163-1.
[22] H. Liu, J. Xu, B. Guo, X. He, Effect of SiO2 content on performance of
polypropylene separator for lithium-ion batteries, J. Appl. Polym. Sci. 131 (2014).
http://dx.doi.org/10.1002/app.41156.
[23] J.-H. Park, J.-H. Cho, W. Park, D. Ryoo, S.-J. Yoon, J.H. Kim, Y.U. Jeong, S.-Y. Lee,
Close-packed SiO2/poly(methyl methacrylate) binary nanoparticles-coated poly-
ethylene separators for lithium-ion batteries, J. Power Sources 195 (2010)
8306–8310. http://dx.doi.org/10.1016/j.jpowsour.2010.06.112.
[24] M. Yanilmaz, C. Chen, X. Zhang, Fabrication and characterization of SiO2/PVDF
composite nanofiber-coated PP nonwoven separators for lithium-ion batteries, J.
Polym. Sci. Part B Polym. Phys. 51 (2013) 1719–1726. http://dx.doi.org/10.1002/
polb.23387.
[25] M. Yanilmaz, M. Dirican, X. Zhang, Evaluation of electrospun SiO2/nylon 6,6
nanofiber membranes as a thermally-stable separator for lithium-ion batteries,
Electrochim. Acta 133 (2014) 501–508. http://dx.doi.org/10.1016/j.electac-
ta.2014.04.109.
[26] M. Yanilmaz, Y. Lu, Y. Li, X. Zhang, SiO2/polyacrylonitrile membranes via
centrifugal spinning as a separator for Li-ion batteries, J. Power Sources 273
(2015) 1114–1119. http://dx.doi.org/10.1016/j.jpowsour.2014.10.015.
[27] Y. Zhai, K. Xiao, J. Yu, B. Ding, Fabrication of hierarchical structured SiO2/
polyetherimide-polyurethane nanofibrous separators with high performance for
lithium ion batteries, Electrochim. Acta 154 (2015) 219–226. http://dx.doi.org/
10.1016/j.electacta.2014.12.102.
[28] J.-A. Choi, S.H. Kim, D.-W. Kim, Enhancement of thermal stability and cycling
performance in lithium-ion cells through the use of ceramic-coated separators, J.
Power Sources 195 (2010) 6192–6196. http://dx.doi.org/10.1016/j.jpow-
sour.2009.11.020.
[29] A. Manuel Stephan, K.S. Nahm, Review on composite polymer electrolytes for
lithium batteries, Polymer 47 (2006) 5952–5964. http://dx.doi.org/10.1016/
j.polymer.2006.05.069.
[30] M.-Y. An, H.-T. Kim, D.-R. Chang, Multilayered separator based on porous
polyethylene layer, Al2O3 layer, and electro-spun PVdF nanofiber layer for lithium
batteries, J. Solid State Electrochem. 18 (2014) 1807–1814. http://dx.doi.org/
10.1007/s10008-014-2412-4.
[31] M. Sethupathy, P. Pandey, P. Manisankar, Photovoltaic performance of dye-
sensitized solar cells fabricated with polyvinylidene fluoride–polyacrylonitrile–
silicondioxide hybrid composite membrane, Mater. Chem. Phys. 143 (2014)
1191–1198. http://dx.doi.org/10.1016/j.matchemphys.2013.11.020.
[32] Y.-J. Kim, C.H. Ahn, M.B. Lee, M.-S. Choi, Characteristics of electrospun PVDF/
SiO2 composite nanofiber membranes as polymer electrolyte, Mater. Chem. Phys.
127 (2011) 137–142. http://dx.doi.org/10.1016/j.matchemphys.2011.01.046.
[33] O. Padmaraj, B. Nageswara Rao, P. Jena, M. Venkateswarlu, N. Satyanarayana,
Electrochemical studies of electrospun organic/inorganic hybrid nanocomposite
fibrous polymer electrolyte for lithium battery, Polymer 55 (2014) 1136–1142.
http://dx.doi.org/10.1016/j.polymer.2014.01.015.
[34] S.A. Smith, J.H. Park, B.P. Williams, Y.L. Joo, Polymer/ceramic co-continuous
nanofiber membranes via room-curable organopolysilazane for improved lithium-
ion battery performance, J. Mater. Sci. (2016) 1–13. http://dx.doi.org/10.1007/
s10853-016-0574-4.
[35] M. Yanilmaz, Y. Lu, J. Zhu, X. Zhang, Silica/polyacrylonitrile hybrid nanofiber
membrane separators via sol-gel and electrospinning techniques for lithium-ion
batteries, J. Power Sources 313 (2016) 205–212. http://dx.doi.org/10.1016/
j.jpowsour.2016.02.089.
[36] X. Huang, Development and characterization of a bilayer separator for lithium ion
batteries, J. Power Sources 196 (2011) 8125–8128. http://dx.doi.org/10.1016/
j.jpowsour.2011.05.054.
[37] H.T. Chiu, T. Sukachonmakul, M.T. Kuo, Y.H. Wang, K. Wattanakul, Surface
modification of aluminum nitride by polysilazane and its polymer-derived amor-
phous silicon oxycarbide ceramic for the enhancement of thermal conductivity in
silicone rubber composite, Appl. Surf. Sci. 292 (2014) 928–936. http://dx.doi.org/
10.1016/j.apsusc.2013.12.081.
[38] R. Saito, Y. Fujii, T. Kumagai, Synthesis of epoxy resin–silica nanocomposites
provided from perhydropolysilazane as a curing reagent and the precursor of silica
domain, J. Appl. Polym. Sci. 127 (2013) 2074–2081. http://dx.doi.org/10.1002/
app.37804.
[39] S. Lee, J. Kim, B.-C. Ku, J. Kim, H.-I. Joh, Structural evolution of polyacrylonitrile
fibers in stabilization and carbonization, Adv. Chem. Eng. Sci. 2 (2012) 275. http://
dx.doi.org/10.4236/aces.2012.22032.
[40] L. Ji, X. Zhang, Ultrafine polyacrylonitrile/silica composite fibers via electrospin-
ning, Mater. Lett. 62 (2008) 2161–2164. http://dx.doi.org/10.1016/j.mat-
let.2007.11.051.
[41] L. Ji, K.-H. Jung, A.J. Medford, X. Zhang, Electrospun polyacrylonitrile fibers with
dispersed Si nanoparticles and their electrochemical behaviors after carbonization,
S.A. Smith et al.
J. Mater. Chem. 19 (2009) 4992–4997. http://dx.doi.org/10.1039/B903165K.
[42] A. Gupta, I.R. Harrison, New aspects in the oxidative stabilization of PAN-based
carbon fibers, Carbon 34 (1996) 1427–1445. http://dx.doi.org/10.1016/S0008-
6223(96)00094-2.
[43] Y. Hou, T. Sun, H. Wang, D. Wu, Thermal-shrinkage investigation of the chemical
reaction during the stabilization of polyacrylonitrile fibers, J. Appl. Polym. Sci. 114
(2009) 3668–3672. http://dx.doi.org/10.1002/app.30303.
[44] S.W. Choi, J.R. Kim, S.M. Jo, W.S. Lee, Electrochemical and spectroscopic
322
Journal of Membrane Science 526 (2017) 315–322
properties of electrospun PAN-based fibrous polymer electrolytes, J. Electrochem.
Soc. 152 (2005) A989–A995. http://dx.doi.org/10.1149/1.1887166.
[45] Z. Song, X. Hou, L. Zhang, S. Wu, Enhancing crystallinity and orientation by,
Materials 4 (2011) 621–632. http://dx.doi.org/10.3390/ma4040621.
[46] N. Salem, Y. Abu-Lebdeh, Effect of nanoparticles on electrolytes and electrode/
electrolyte interface, in: Y. Abu-Lebdeh, I. Davidson (Eds.), Nanotechnol. Lithium-
Ion Batter., Springer US, 2012: pp. 221–244. doi:10.1007/978-1-4614-4605-7_9