Chemical Engineering Science 62 (2007) 4751 – 4759

www.elsevier.com/locate/ces

Nanoparticle filtration by electrospun polymer fibers

Ki Myoung Yun a , Christopher J. Hogan Jr. b , Yasuko Matsubayashi c , Masaaki Kawabe c ,
Ferry Iskandar a , Kikuo Okuyama a,∗
a Department of Chemical Engineering, Graduate School of Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi-Hiroshima 739-8527, Japan
b Department of Energy, Environmental, and Chemical Engineering, One Brookings Drive, Box 1180, Washington University in St. Louis, St. Louis, MO
63130, USA
c Central Research Laboratory, Japan Vilene Company Ltd., 7 Kita-tone, Sowa-machi, Sashima-gun, Ibaraki 306-0213, Japan

Received 8 February 2007; received in revised form 6 June 2007; accepted 11 June 2007
Available online 16 June 2007

Abstract
Polyacrylonitrile (PAN) fibers with mean diameters in 270–400 nm range were prepared by electrospinning for use as a filter media.
Compared to commercial filters made of polyolefin and glass, the fibers of electrospun filters were more uniform in diameter. The performance
of electrospun filters was evaluated by measuring the penetration of monodisperse NaCl nanoparticles (below 80 nm in size) through the
filters. It was found that electrospun filters could be made which had nanoparticle penetration values comparable to commercial filters but
with substantially less filter mass. The penetration of nanoparticles through the electrospun filter media could be reduced by increasing the
filter thickness, which is controlled by the collection time during the electrospinning process. Nanoparticle collection by electrostatic forces
was found to be negligible for electrospun filters. Filter quality factors and single fiber collection efficiencies were found to be independent of
filter thickness for electrospun filters, and the penetration of nanoparticles through electrospun filters was in better agreement with theoretical
predictions than was the measured penetration through a commercial filter. This study shows that electrospinning is a promising technology
for the production of high performance nanoparticle filters.
䉷 2007 Elsevier Ltd. All rights reserved.

Keywords: Filtration; Aerosol; Electrospinning; Polymer processing; Nanoparticles; Nanofibers

1. Introduction a capillary where a voltage is applied will form a stable “Taylor

cone” because of a balance between surface tension and elec-
Aerosol filtration is the most widely utilized technique for trical forces within the liquid. A polymer solution jet is emitted
air cleaning and sampling, and is applied in a variety of in- from the cone, which, so long as the polymer solution is suffi-
dustries, such as respiratory protection, air cleaning of smelter ciently elastic (Yu et al., 2006), will result in the formation of
effluents, processing of nuclear and hazardous materials, and diameter controlled polymer fibers after solvent evaporation.
particle collection in clean rooms (Hinds, 1999). In particular, Many polymers have been successfully electrospun into
fibrous filters have been widely used to separate solid matter fibers from polymer solutions and polymer melts (Huang et al.,
from particulate laden air flow streams because of their simple 2003; Li and Xia, 2004). Several studies have investigated the
structure and low material cost. In recent years, a method to performance of electrospun fiber mats as filter media (Barhate
prepare ultrafine fibers, electrospinning, has received consid- et al., 2006; Gibson et al., 2001; Hajra et al., 2003; Shin et al.,
erable attention. Electrospinning occurs as the result of apply- 2005). However, electrospun filter media have not been ap-
ing an electrical field to a highly conducting polymer solution, plied previously to the collection and removal of nanoparticles.
which drives the formation of an electrically driven polymer Because of the deleterious effects of nanoparticles on health
solution jet (Taylor, 1969). A polymer solution flowing through and the environment (Biswas and Wu, 2005), it is necessary
to demonstrate that electrospun filter media can efficiently
∗ Corresponding author. Tel.: +81 82 424 7850. collect nanoparticles and compare the performance of electro-
E-mail address: okuyama@hiroshima-u.ac.jp (K. Okuyama). spun filters to commercial filters and theoretical filter models.
0009-2509/$ - see front matter 䉷 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ces.2007.06.007
4752 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759

The purpose of this study is to use electrospinning to produce using a field emission scanning electron microscope (FE-SEM,
polyacrylonitrile (PAN) fiber mats for use as filter media, and S-5000, Hitachi Ltd., Japan). To produce filters by electrospin-
evaluate the filtration performance of the electrospun fibers us- ning, multiple capillaries were used to increase the production
ing experimental measurements with nanoparticles. PAN was rate of fibers. Each filter took between 2 and 15 min to produce.
selected as the material for electrospun filters in this study be- The thicknesses of the electrospun filters were measured using
cause it is a versatile polymer that is widely used for making a micrometer screw gauge (Mitutoyo Ltd., Japan; Gopal et al.,
membranes, and is not only hydrophobic but is also insoluble in 2007) with ±1 m error.
a wide range of solvents (Musale and Kumar, 2000). The size
distribution of electrospun fibers, pressure drop across the filter 2.2. Aerosol filtration
media, hydraulic permeability, nanoparticle penetration, qual-
ity factor, and single fiber collection efficiency were calculated Fig. 2 shows the experimental system used to determine
from measurements of electrospun filters of varying thickness the penetration of nanoparticles through electrospun PAN
and compared to the measurements on commercial filters as fiber filters. N2 gas (3.9 l min−1 ) was sent into an NaCl
well as theoretical predictions. evaporation–condensation aerosol generator (Nakaso et al.,
2002; Panda and Pratsinis, 1995) which produced polydis-
2. Materials and methods perse NaCl nanoparticles. An Am-241 bipolar charger was
used to charge-neutralize NaCl particles, followed by subse-
2.1. Electrospinning quent particle size selection using a nanodifferential mobility
analyzer (DMA, TSI model 3085, Shoreview, MN, USA).
The experimental system used for polymer fiber production Singly charged, monodisperse particles exited the DMA and
by electrospinning was similar to that used previously for poly- were sent to the test filter. Charged, charge neutralized (equi-
mer particle production (Hogan et al., 2007) and is shown in librium distribution), and uncharged particles were used for
Fig. 1. Polymer fibers were prepared from a solution of PAN filter testing. To produce charge neutralized particles, particles
(Sigma–Aldrich, St. Louis, MO, USA) with a molecular weight were sent through an Am-241 bipolar charger and to produce
of 150 kDa dissolved in N, N -dimethylformamide. The solu- uncharged particles, particles were sent through an Am-241
tion was pumped through a capillary with an inner diameter of bipolar charger as well as a plate condenser to remove any re-
400 m using a syringe-pump (PHD 2000, Harvard Apparatus maining charged particles. The test filter was held in a 20 mm
Inc., Holliston, MA, USA). For stable system operation, a pos- diameter filter holder, and the flow rate at the DMA outlet
itive voltage of 3.0–10.0 kV was applied to the capillary with a was set at 1.0 l min−1 , which corresponded to a filter face
suitable liquid flow rate. A ground plate was placed 5 cm from velocity of 5.3 cm/s. The number concentration of particles
the capillary tip. All experiments were performed in a humid- at the filter outlet was measured using an ultrafine particle
ity controlled chamber in pure N2 gas and the relative humid- counter (CPC, TSI model 3025) operated at a flow rate of
ity was below 25% for all experiments. A high-speed video 0.3 l min−1 . The penetration of particles was determined by
camera (HPV-1, Shimadzu Ltd., Japan) was used to observe measuring the concentration of particles at the filter holder
electrospun fibers. After collection, PAN fibers were viewed outlet with and without a filter installed and taking the ratio of

High Voltage
Power Supply (DC)

M on itor
Monitor

Gas(N2)
CCD camera inlet
with Microscope
Hygrometer

Liquid Aluminum
Gas Supply Plate
Stainless
outlet
Capillary

Fig. 1. System used for the production of PAN fibers.

 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759 4753

Purge

N2 241Am
DMA
charger

N2

Filter MFC

NaCl Aerosol generator

Cooling
water
241Am
CPC
charger
(TSI 3025) Test Filter

Plate
Condenser

MFC
Pump

Fig. 2. System used for the measurement of nanoparticle penetration through filters.

the number concentration of particles with the filter installed the electric field (calculated approximately from the applied
to the number concentration without the filter (see Ichitsubo potential and distance between the capillary and ground plate)
et al., 1996; Heim et al., 2005). Prior to the start of each ex- and solution flow rate (Shin et al., 2001). Fig. 4 shows electric
periment, the system was flushed with clean N2 gas for a pe- field necessary for stable jet production as a function of flow rate
riod of at least 1 h to remove any particles which may have for a solution with 10% PAN mass fraction. For flow rate and
deposited on tubing walls in previous experiments. Six differ- electric field values below this curve, a jet was not formed and
ent types of electrospun PAN filters were produced and tested. solution dripping from the capillary was observed. Conversely,
For comparison, the penetration of nanoparticles through two for flow rate and electric field values above this curve, unstable
commercially produced medium grade polyolefin filters (PO whipping and jet breakage was observed (Fig. 3b). For uniform
filter, Japan Vilene Co. Ltd., Ibaraki, Japan), a high efficiency fiber production, the stable cone-jet (Fig. 3a) was used in all
particulate air filter (HEPA filter, Lydall Inc. Ltd., USA), and experiments.
an ultra low penetration air filter (ULPA filter, Lydall Inc. Ltd., SEM images of electrospun PAN fibers are shown in
USA) was also measured. Figs. 5a–d for fibers made from solutions with four different
PAN mass fractions. With lower PAN mass fraction solutions
3. Results and discussion (3% PAN), beads were visible in fibers (Fong et al., 1999).
At sufficiently low polymer concentrations, polymer solutions
3.1. Electrospun fibers do not electrospin; rather, the Taylor cone-jet breaks up into
droplets and solvent evaporation results in the formation of
Fig. 3 shows profiles of the electrically driven jets produced polymer particles (Hogan et al., 2007). Beaded fiber forma-
during electrospinning. Figs. 3a and b both show single jet tion occurs when solutions with intermediate polymer mass
modes, while Fig. 3c shows a multi-jet mode. The jet shown fractions are used such that droplet formation does not occur,
in Fig. 3a was stable, i.e., fibers were generated continuously but the solution is not sufficiently elastic for pure fiber for-
from these jets. The jet in Fig. 3b, however, resulted in the mation to occur. Images of beaded fiber formation were taken
production of beaded fibers and broken fibers (Dietzel et al., with the high-speed camera and are shown in Fig. 6a. Near-
2001). Jet breakage prior to fiber formation can be clearly seen spherical beads can be seen forming within the jet. Conversely,
in Fig. 3b. The jet in Fig. 3c was stable for a short period but electrospinning with higher mass fraction PAN solutions re-
eventually solution dripping from the capillary was observed. sulted in the formation of smooth, uniform fibers with the
The stability of the jet can be described by a correlation with fiber diameter increasing as a function of PAN mass fraction
4754 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759

a b c

Fig. 3. Characteristic electrically driven jets in the: (a) cone-jet, (b) cone-jet with jet breakup, and (c) multi-jet modes.

2.0 than did the conventional filters (Huang et al., 2003). Previous
work on droplets produced by electrospray has shown that the
1.8 initial droplets are relatively monodisperse (Chen et al., 1995;
de Juan and Fernandez de la Mora, 1997; Rosell-Llompart and
1.6 Fernandez de la Mora, 1994); thus it would be expected that
fibers produced from electrified jets which do not break into
1.4 droplets would also be uniform in size.
E [kV/cm]

1.2
3.2. Measured nanoparticle penetration
1.0
The penetration of charge neutralized nanoparticles in the
10–80 nm size range through all tested filters is shown in Fig. 8.
0.8
For nanoparticles, particle collection is typically diffusion dom-
inated and therefore penetration increases with increasing par-
0.6
ticle diameter (Emi et al., 1982; Kim et al., 2006). Nanoparticle
penetration was highest through the medium grade polyolefin
0.4
(PO-1 and PO-2) filters and was lowest through the ULPA filter
0 5 10 15 20 25 30 35
Q [μl/min]
and PAN filter ES-5. Nanoparticle penetration through electro-
spun filters (ES-1 to ES-5) was a decreasing function of filter
Fig. 4. Optimum electric fields as function of solution flow rates for stable thickness (see Table 1). Filters ES-1 through ES-5 all had the
jet operation of a 10% PAN mass fraction solution. same fiber diameter, similar fiber volume fractions, and dif-
fered in thickness, i.e., amount of electrospinning time used
to make each filter. This demonstrates that filters of all grades
(Dietzel et al., 2001). Fig. 6b shows high-speed camera images can be made by the same electrospinning process by simply
of an electrically driven jet of a polymer solution with suffi- increasing or decreasing the collection time for each filter. Fur-
ciently high polymer concentration that bead formation did not thermore, the pressure drop of the filter ES-5 is smaller than
occur. that of HEPA and ULPA filters (Table 1), but the penetration of
For comparison, Figs. 5e and f show SEM images of poly- nanoparticles through these filters is the same. This is primar-
olefin and glass (HEPA) fibers, which were less uniform in di- ily due to the smaller diameter fibers that can be produced by
ameter and larger than the electrospun fibers. The properties of electrospinning, which, based on filtration theories (Emi et al.,
all tested filters are shown in Table 1. Five electrospun PAN 1982; Kirsh et al., 1975; Lee et al., 2002; Podgorski et al., 2006)
fiber filters (ES-1 to ES-5) were produced with different thick- collect particles more efficiently. The major drawback to using
nesses but the same mean fiber diameter (0.27 m) and near such filters is that their low mass gives rise to low durability.
identical fiber volume fraction. A sixth electrospun filter was This could be overcome by placing an electrospun filter onto a
produced with a mean fiber diameter of 0.4 m for comparison support material, e.g. a low efficiency filter with low pressure
to the ULPA filter, which also had a mean fiber diameter of drop, which would limit damage to the electrospun filter over
0.4 m. SEM measured size distribution functions of the fibers time.
of the electrospun filter ES-6 and the ULPA filter are shown There is a tradeoff between nanoparticle penetration through
in Fig. 7. The size distribution functions were determined by the filter and the pressure drop across the filter. To decrease par-
measuring over 200 fibers from both of the filters. The coeffi- ticle penetration through an electrospun filter, it must be made
cient of variation of the fiber diameter for filter ES-6 was 0.21, thicker, and because flow through electrospun filters obeys
while for the ULPA filter it was 0.66, showing quantitatively Darcy’s Law, the pressure drop through the filters increases
that the electrospun filters had more uniform fiber diameters linearly with filter thickness (Miguel, 2003). The hydraulic
 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759 4755

a b

5 µm 5 µm

c d

5 µm 5 µm

e f

30 µm 10 µm

Fig. 5. SEM images of electrospun filter fibers with PAN mass fractions of 3% (a), 7% (b), 9% (c), and 11% (d), and SEM images of commercial filters with
polyolefin (e) and glass (HEPA, f) fibers.

a b

Fig. 6. High-speed camera observation of jets during electrospinning for PAN mass fractions of (a) 3% and (b) 7%.

permeability through electrospun filters was found to be 8.5 × ULPA filters which had hydraulic permeabilities of the order
10−14 m2 for filters ES-1 to ES-5, and 2.04×10−13 m2 for filter of 10−12 m2 ; thus, although low mass electrospun filters could
ES-6, as compared to the polyolefin filters which had hydraulic collect particles efficiently, they required a larger pressure drop
permeabilities of the order of 10−11 m2 and the HEPA and per unit filter thickness than conventional filters to do so.
4756 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759

Table 1
Properties of tested filters

Sample no. Material Mean fiber diameter (m) Filter thickness (mm) Pressure drop (Pa) at 5.3 cm/s Volume fraction (–)

ES-1 Polyacrylonitrile 0.27 0.004 59 0.152

ES-2 Polyacrylonitrile 0.27 0.007 76 0.124
ES-3 Polyacrylonitrile 0.27 0.010 112 0.113
ES-4 Polyacrylonitrile 0.27 0.016 161 0.112
ES-5 Polyacrylonitrile 0.27 0.020 220 0.113
ES-6 Polyacrylonitrile 0.4 0.018 85 0.19
P0-1 Polyolefin 10 1.050 3.3 0.100
P0-2 Polyolefin 3 1.100 32 0.100
HEPA Glass 0.6 0.500 310 0.065
ULPA Glass 0.4 0.500 420 0.060

6000 process. For a filter with penetration P and pressure drop p,
Electrospun Filter (ES-6) the quality factor, qF , is calculated as (Hinds, 1999):
dN/dlog(dp) [-]

4000
ln(1/P )
qF = . (1)
2000
p
The single fiber collection efficiency, , of a filter with pene-
0 tration P , fiber volume fraction , fiber diameter df , thickness
6000
ULPA Filter L, fiber diameter variance g , and volume fraction variance ,
dN/dlog(dp) [-]

is calculated as (Kirsh et al., 1975):

4000
 
1  (1 − ) df (1 + )
 = ln · · ·  (2)
2000 P 4  L

0 df , , and  in the above equation are given by

100 1000
Fiber Size [nm] df = dfg exp(0.5 ln2 g ), (3)

Fig. 7. Fiber diameter distributions of commercial and electrospun filter fibers.  = exp(ln2 g ) − 1, (4)
f
CD
= (5)
0
CD
10
ES1 dfg , g are the geometric mean diameter and the geometric
ES2 f
10–1 standard deviation of the fiber, respectively. CD is the resis-
ES3

–2
ES4 tance coefficient of a fan model filter and CD is the resistance
10
Penetration [-]

ES5
coefficient of measured filter, and are given by
ES6
–3 PO1
10 f 8
PO2 CD = , (6)
–4
HEPA Rek
10 ULPA
 
k = −0.5 ln − 0.52 + 0.64 + 1.43(1 − )Kn,
10 –5 1− 1−
(7)
10–6
10 20 30 40 50 607080 90100
Kn = 2/df , (8)
Particle Diameter [nm]  
 1 −  df (1 + ) f u20
CD = · · P , (9)
Fig. 8. Penetration of nanoparticles through all tested filter media. 4  L 2

where  and f are the mean free path (66.5 nm) and density
of carrier gas, respectively.
Calculation of filter quality factors and single fiber collection Figs. 9a and b show the quality factors and single fiber col-
efficiencies provided evidence of the high degree of uniformity lection efficiencies for PAN filters ES-1 to ES-5 as a function
of electrospun filters and consistency of the electrospinning of particle diameter. At a given diameter, all filters had near
 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759 4757

0.2 7

6
ES1
ES2 5
ES3

Single Fiber Efficiency [-]

ES4 4
Quality Factor [1/Pa]

ES5
0.1 3
0.09

0.08
0.07 2

0.06

0.05

0.04 1
20 30 40 50 60 70 80 20 30 40 50 60 70 80
Particle Diameter [nm] Particle Diameter [nm]

Fig. 9. Measured quality factors (a) and single fiber collection efficiencies (b) of electrospun filters.

identical quality factors and single fibers collection efficien- fiber collection efficiency based on interception. Pe, R, and Kn
cies, i.e., quality factor and single fiber collection efficiency in the above equation are given by
were primarily a function of the diameter of the filter fibers and
packing density. Therefore, the penetration of particles through Pe = u0 df /D, (11)
filters produced by a process such as electrospinning, which
R = dp /df , (12)
produces uniform fibers at a fixed size giving rise to filters with
thickness-independent quality factors, can be estimated based kB T C C
on the known quality factor, filter thickness, and hydraulic per- D= , (13)
3 dp
meability. Furthermore, because the pressure drop through the
  
tested filters (ES-1 to ES-5) increased linearly with filter thick- 2 dp
Cc = 1 + 1.257 + 0.4 exp −0.55 . (14)
ness (Table 1) and all filters had the same quality factor for a dp 
given particle diameter, it follows from Eqs. (1) and (2) that all
filters had the same single fiber collection efficiency curves. u0 , dp , kB , and T are the filter face velocity (5.3 cm/s), par-
ticle diameter, Boltzmann’s constant (1.38 × 10−23 J/K), and
temperature, respectively. Based the single fiber collection effi-
3.3. Comparison with theory
ciency, fiber diameter, fiber diameter variance, filter thickness,
and the filter fiber volume fraction, the penetration of parti-
Single fiber collection efficiency of a filter based on a fan
cles through a filter can be calculated using Eq. (2). Although
model of a filter (Kirsh et al., 1975) is calculated as
the electrospun filters had a higher fiber volume fraction than
f f conventional filters, it has been shown that the given equations
 = D + R + f (Pe, R) are valid for filter with fiber volume fractions in the 0.01–0.27
range (Emi and Kanaoka, 1984) and can therefore be used here.
= 2.7Pe−2/3 (1 + 0.39k −1/3 Pe1/3 Kn) + 0.624Pe−1 Fig. 10 shows the measured nanoparticle penetrations through
 filters ES-1 through ES-5 and the HEPA filter as well as the
1 1
+ − (1 + R) + 2(1 + R) ln(1 + R) theoretical nanoparticle penetration values. All electrospun fil-
2k 1 + R ters had measured nanoparticle penetration values which were
 slightly lower than, but overall in agreement with theoretical
2.86Kn(2 + R)R
+ + 1.24k −1/2 Pe−1/2 R 2/3 . (10) predictions. Conversely, the nanoparticle penetration through
1+R the HEPA filter was 1–2 orders of magnitude higher than was
f
predicted. This has previously been attributed to variations in
 is the single fiber collection efficiency, D is the single fiber the volume fraction of fibers along the cross section of the fil-
f
collection efficiency based on diffusion, and R is the single ter (Dhaniyala and Liu, 2001; Podgorski et al., 2006), which
4758 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759

10–1
10–1 ES-1 ES- 2E S-3
10–2
10–2

10–2
10–4
10–3
Penetration [-]

10–3 10–4 10–6

20 30 40 50 60 70 8090 20 30 40 50 60 70 8090 20 30 40 50 60 70 8090

10–2 10–3 10–3

ES-4 ES-5 HE PA

10–3 10–4 10–5

10–4 10–5 10–7

10–5 10–6 10–9

30 40 50 60 70 80 90 30 40 50 60 70 80 90 30 40 50 60 70 80 90
Particle Diameter [nm]

Fig. 10. Theoretically calculated (solid lines, broken lines: the range of error) and measured (black circles) nanoparticle penetration through electrospun filters
and a high efficiency particulate air filter.

10–1 uncharged, and charge neutralized particles through an electro-

Electrospun Filter (ES-2) spun filter as a function of particle diameter. Particle penetration
was independent of particle charge state, which implies that any
10–2 Coloumbic forces present between the particles and the fibers
did not lead to enhancement of particle collection (Kousaka
Penetration [-]

10–3
et al., 1990). Electrostatic collection may still be prevalent in
electrospun filters, as electrostatic collection and collection by
diffusion are not additive. Lee et al. (2002) numerically mod-
10–4 eled electret filters and showed that there is interaction between
Positively Charged Particles the diffusive and electrostatic collection mechanisms such that
Negatively Charged Particles
high collection by diffusion decreases electrostatic collection,
10–5 Neutralized Particles
Uncharged Particles
which may be the case for nanoparticle collection by electro-
spun filters.
10–6
20 30 40 50 60 70 80 90
Particle Diameter [nm] 4. Conclusions

Fig. 11. Penetration of nanoparticles of various charge states through elec- PAN fibrous filters were produced by electrospinning and the
trospun filters. penetration of nanoparticles through electrospun filters as well
as commercial polyolefin and glass fibrous filters was measured.
Fibers produced by electrospinning were much more uniform
implies that unlike the HEPA filter, electrospun filters not only in diameter than fibers in commercially produced filters. Par-
have uniform diameter fibers but also uniformly distributed ticle penetration through the electrospun filters was strongly a
fibers within the filter. Therefore, electrospun fiber filters could function of the filter thickness, while the filter quality factor
be potentially used as model systems for the testing of filtration and single fiber collection efficiency were found to be con-
models. stant for a given fiber diameter and particle diameter. To have
The slight decrease in measured penetration as compared the same nanoparticle penetration value as commercial filters,
to the predicted penetration for electrospun filters may have electrospun filters required significantly less filter mass but a
been due to particle collection by Coloumbic forces. Any higher pressure drop due to their high fiber volume fraction
residual charge remaining on the filter fibers after collection and low hydraulic permeability. The penetration of nanoparti-
would give rise to Coloumbic forces between charged or di- cles through electrospun fibrous filters was in better agreement
electric particles and the filter fiber, which may have enhanced with theoretical predictions than was the penetration through a
particle collection. Fig. 11 shows the penetration of charged, commercial filter, which was attributed to the high degree of
 K.M. Yun et al. / Chemical Engineering Science 62 (2007) 4751 – 4759 4759

uniformity in the volume fraction throughout the electrospun Heim, M., Mullins, B.J., Wild, M., Meyer, J., Kasper, G., 2005. Filtration
filter. Charged, uncharged, and charge neutralized particles all efficiency of aerosol particles below 20 nanometers. Aerosol Science and
had the same penetration through electrospun filters, which im- Technology 39, 782–789.
Hinds, W.C., 1999. Aerosol Technology. second ed. Wiley, New York.
plies that electrostatic collection was minimal. Overall, elec- Hogan, C.J., Yun, K.M., Chen, D.R., Lenggoro, I.W., Biswas, P., Okuyama, K.,
trospinning allows for advanced fibrous materials production 2007. Controlled size polymer particle production via electrohydrodynamic
(Li and Xia, 2004) with controlled, uniform fiber size which atomization. Colloids and Surfaces A-Physicochemical and Engineering
allows for the production of low nanoparticle penetration, low Aspects, in press.
mass filters. Huang, Z.M., Zhang, Y.Z., Kotaki, M., Ramakrishna, S., 2003. A review
on polymer nanofibers by electrospinning and their applications in
nanocomposites. Composite Science and Technology 63, 2223–2253.
Acknowledgements Ichitsubo, H., Hashimoto, T., Alonso, M., Kousaka, Y., 1996. Penetration of
ultrafine particles and ion clusters through wire screens. Aerosol Science
This work was supported by grants from the Japan Vilene and Technology 24, 119–127.
Company Ltd. and a Grant-in-Aid for Scientific Research (A) Kim, C.S., Bao, L., Okuyama, K., Shimada, M., Niinuma, H., 2006. Filtration
efficiency of a fibrous filter for nanoparticles. Journal of Nanoparticle
(No. 18206079) by Japan Society for the Promotion of Science Research 8, 215–221.
(JSPS). K.M.Y. acknowledges support from the Yoneyama- Kirsh, A.A., Stechkina, I.B., Fuchs, N.A., 1975. Efficiency of aerosol filters
rotari Scholarship Foundation, and C.J.H. acknowledges sup- made of ultrafine polydisperse fibres. Journal of Aerosol Science 6,
port from an NSF Graduate Research Fellowship. The authors 119–120.
thank Nobuyuki Tokuoka (Shimadzu Ltd., Japan) for assistance Kousaka, Y., Okuyama, K., Shimada, M., Takii, Y., 1990. Development of
a method for testing very high-efficiency membrane filters for ultrafine
in observation of spun fibers by a high-speed video camera. aerosol particles. Journal of Chemical Engineering of Japan 23, 568–574.
Lee, M.-H., Otani, Y., Namiki, N., Emi, H., 2002. Prediction of collection
efficiency of high-performance electret filters. Journal of Chemical
References Engineering of Japan 35, 57–62.
Li, D., Xia, Y.N., 2004. Electrospinning of nanofibers: reinventing the wheel?
Barhate, R.S., Loong, C.K., Ramakrishna, S., 2006. Preparation and Advanced Materials 16, 1151–1170.
characterization of nanofibrous filtering media. Journal of Membrane Miguel, A.F., 2003. Effect of air humidity on the evolution of permeability
Science 283, 209–218. and performance of a fibrous filter during loading with hygroscopic and
Biswas, P., Wu, C.Y., 2005. Critical review: nanoparticles and the environment. non-hygroscopic particles. Journal of Aerosol Science 34, 783–799.
Journal of the Air and Waste Management Association 55, 708–746. Musale, D.A., Kumar, A., 2000. Solvent and pH resistance of
Chen, D.R., Pui, D.Y.H., Kaufman, S.L., 1995. Electrospraying of conducting surface crosslinked chitosan/poly(acrylonitrile) composite nanofiltration
liquids for monodisperse aerosol generation in the 4 nm to 1.8 m diameter membranes. Journal of Applied Polymer Science 77, 1782–1793.
range. Journal of Aerosol Science 26, 963–977. Nakaso, K., Shimada, M., Okuyama, K., Deppert, K., 2002. Evaluation of the
de Juan, L., Fernandez de la Mora, J., 1997. Charge and size distributions of change in the morphology of gold nanoparticles during sintering. Journal
electrospray drops. Journal of Colloid and Interface Science 186, 280–293. of Aerosol Science 33, 1061–1074.
Dhaniyala, S., Liu, B.Y.H., 2001. Theoretical modeling of filtration by Panda, S., Pratsinis, S.E., 1995. Modeling the synthesis of aluminum particles
nonuniform fibrous filters. Aerosol Science and Technology 34, 170–178. by evaporation–condensation in an aerosol flow reactor. Nanostructured
Dietzel, J.M., Kleinmeyer, J., Harris, D., Beck Tan, N.C., 2001. The effect Materials 5, 755–767.
of processing variables on the morphology of electrospun nanofibers and Podgorski, A., Balazy, A., Gradon, A., 2006. Application of nanofibers to
textiles. Polymer 42, 261–272. improve the filtration efficiency of the most penetrating aerosol particles
Emi, H., Kanaoka, C., 1984. In: Takahashi, K. (Ed.), Ohyoearozorugaku. in fibrous filters. Chemical Engineering Science 61, 6804–6815.
Yokendo, Tokyo. (in Japanese) (Chapter 3). Rosell-Llompart, J., Fernandez de la Mora, J., 1994. Generation of
Emi, H., Kanaoka, C., Kuwabara, Y., 1982. The diffusion collection efficiency monodisperse droplets 0.3 to 4 m in diameter from electrified cone-jets
of fibers for aerosol over a wide range of Reynolds numbers. Journal of of highly conducting and viscous liquids. Journal of Aerosol Science 25,
Aerosol Science 13, 403–413. 1093–1119.
Fong, H., Chun, I., Reneker, D.H., 1999. Beaded nanofibers formed during Shin, Y.M., Hohman, M.M., Brenner, M.P., Rutledge, G.C., 2001.
electrospinning. Polymer 40, 4585–4592. Experimental characterization of electrospinning: the electrically forced
Gibson, P., Schreuder-Gibson, H., Rivin, D., 2001. Transport properties of jet and instabilities. Polymer 42, 9955–9967.
porous membranes based on electrospun nanofibers. Colloids and Surfaces Shin, C., Chase, G.G., Reneker, D.H., 2005. Recycled expanded
A-Physicochemical and Engineering Aspects 31, 469–481. polystyrene nanofibers applied in filter media. Colloids and Surfaces
Gopal, R., Kaur, S., Feng, C.Y., Chan, C., Ramakrishna, S., Tabe, S., Mtsuura, A-Physicochemical and Engineering Aspects 262, 211–215.
T., 2007. Electrospun nanofibrous polysulfone membranes as pre-filters: Taylor, G.I., 1969. Electrically driven jets. Proceedings of the Royal Society
particulate removal. Journal of Membrane Science 289, 210–219. of London A 313, 453–475.
Hajra, M.G., Mehta, K., Chase, G.G., 2003. Effects of humidity, temperature, Yu, J.H., Fridrikh, S.V., Rutledge, G.C., 2006. The role of elasticity in the
and nanofibers on drop coalescence in glass fiber media. Separation and formation of electrospun fibers. Polymer 47, 4789–4797.
Purification Technology 30, 79–88.