Experimental study of nanoparticle penetration through commercial filter media

Abstract

In this study, nanoparticle penetration was measured with a wide range of filter media using silver nanoparticles from 3 nm
to 20 nm at three different face velocities in order to define nanoparticle filtration characteristics of commercial fibrous
filter media. The silver particles were generated by heating a pure silver powder source via an electric furnace with a temperature
of 870°C, which was found to be the optimal temperature for generating an adequate amount of silver nanoparticles for the
size range specified above. After size classification using a nano-DMA, the particle counts were measured by an Ultrafine
Condensation Particle Counter (UCPC) both upstream and downstream of the test filter to determine the nanoparticle penetration
for each specific particle size. Particle sampling time continued long enough to detect more than 105 counts at the upstream and 10 counts at the downstream sampling point so that 99.99% efficiency can be detected with the
high efficiency filter. The results show a very high uniformity with small error bars for all filter media tested in this
study. The particle penetration decreases continuously down to 3 nm as expected from the classical filtration theory, and
together with a companion modeling paper by Wang et al. in this same issue, we found no significant evidence of nanoparticle
thermal rebound down to 3 nm.

Full text

Special issue: Nanoparticles and Occupational Health
Experimental study of nanoparticles penetration through commercial filter media
Seong Chan Kim, Matthew S. Harrington and David Y. H. Pui*
Particle Technology Laboratory, Department of Mechanical Engineering, University of Minnesota, 111
Church St. S.E., Minneapolis, MN, 55455, USA; *Author for correspondence (Tel.: +64-3-364-2507;
E-mail: dyhpui@tc.umn.edu)
Received 29 August 2006; accepted in revised form 1 September 2006
Key words: nanoparticles, penetration, filtration, thermal rebound, diffusion, occupational health
Abstract
In this study, nanoparticle penetration was measured with a wide range of filter media using silver nano-
particles from 3 nm to 20 nm at three different face velocities in order to define nanoparticle filtration
characteristics of commercial fibrous filter media. The silver particles were generated by heating a pure
silver powder source via an electric furnace with a temperature of 870°C, which was found to be the optimal
temperature for generating an adequate amount of silver nanoparticles for the size range specified above.
After size classification using a nano-DMA, the particle counts were measured by an Ultrafine Conden-
sation Particle Counter (UCPC) both upstream and downstream of the test filter to determine the nano-
particle penetration for each specific particle size. Particle sampling time continued long enough to detect
more than 10
5
counts at the upstream and 10 counts at the downstream sampling point so that 99.99%
efficiency can be detected with the high efficiency filter. The results show a very high uniformity with small
error bars for all filter media tested in this study. The particle penetration decreases continuously down to
3 nm as expected from the classical filtration theory, and together with a companion modeling paper
by Wang et al. in this same issue, we found no significant evidence of nanoparticle thermal rebound down
to 3 nm.
Introduction
Nanotechnology, which involves the manipulation
of matter at nanometer length scales to produce
new materials, structures and devices, has the
potential to start the new industrial revolution.
The potential for new products leading to
improvements in our lives is astounding. Nano-
particles often behave much differently than bulk
samples of the same materials, resulting in unique
electrical, optical, chemical, and biological prop-
erties. The special properties of nanoparticles give
rise to recent concerns about the potential health
hazards posed to workers or users that are exposed
to them (Oberdo
¨rster, 2000; Maynard, 2003).
Therefore, nanoparticle research has received
considerable attention in many laboratories and
industrial fields, especially for studying the health
effects of nanoparticles and their control.
Filtration is the simplest and most common
method for air cleaning, and aerosol filtration is
used in diverse applications, such as respiratory
protection, air cleaning of smelter effluent, pro-
cessing of nuclear and hazardous materials, and
clean rooms. However, the process of filtration is
complicated, and although the general principles
are well known there is still a gap between theory
and experiment (Hinds, 1999). In particular, recent
Journal of Nanoparticle Research (2007) 9:117–125 ÓSpringer 2006
DOI 10.1007/s11051-006-9176-4

modeling and experiments pointed to the potential
penetration of nanoparticles through the filters
due to thermal rebound. Further, nanoparticle
penetration has not been shown clearly due to the
difficulties of system set-up and penetration mea-
surement. Wang and Kasper (1991) suggested a
numerical model for nanoparticle penetration
showing that the thermal impact velocity of a
particle will exceed the critical sticking velocity in
the size range between 1 and 10 nm depending
sensitively on elastic and surface adhesion
parameters. Ichitsubo et al. (1996) conducted an
experimental work of nanoparticle penetration
using wire screens, and showed the nanoparticle
penetration below two nm in size was higher than
the theoretical results due to the thermal rebound.
Following this, Alonso et al. (1997) used a tandem
DMA technique, and detected no particle bounce
effects in the same size range as Ichitsubo et al. As
of now, the thermal rebound effect on nanoparticle
filtration is not well proven, and it is very impor-
tant to study the air filtration properties of nano-
particles to determine the filtration requirements
of personal protective equipment or HVAC filter.
In this study, the nanoparticle penetration test
system has been established and the nanoparticle
penetration was tested with a wide range of filter
media (four fiberglass filter media, four electret
filter media and one nanofiber filter media) using
silver nanoparticles from 3 nm to 20 nm at face
velocities of 5.3, 10 and 15 cm/s.
Experiments
Figure 1 shows a schematic diagram of a nano-
particle filtration test system, and it consists of a
nanoparticle generation system, a size classifica-
tion system and a penetration measurement
system. An electric furnace is used to generate
silver nanoparticles from a pure silver powder
source (99.999%, Johnson Mattney Electronics),
and clean compressed air is used as a carrier gas
with flow rate of 3.0 lpm. The silver powder source
located in the center of a heating tube is vaporized
and condensed into silver nanoparticles with a
wide size distribution at stainless steel tube parts.
The particle size distribution can be controlled by
changing the furnace temperature as shown in
Figure 2. The average size and the particle number
concentration of silver nanoparticles generated by
the furnace increases with the furnace temperature,
because at a higher temperature, the evaporation
rate increases, giving rise to a larger amount of
condensable vapor which allows the particles to
grow to larger sizes by agglomeration and con-
densation (Ku & Maynard, 2006).
The silver nanoparticles are given a Boltzman
charge distribution by Po-210 and classified by a
differential mobility analyzer (nano-DMA, Model
3085, TSI). The neutralized silver nanoparticles
(by another Po-210) are then introduced to the test
filter and the number counts upstream and
downstream of the filter are measured by an
Figure 1. Schematic diagram of the nanoparticle penetration test.
118

ultrafine CPC (Model 3025A, TSI) for the nano-
particle penetration calculation during a certain
sampling time that is long enough to show high
reliability. Each test is repeated more than five
times with different test conditions to show the
variability caused by the set-up and the measure-
ment itself.
Chen et al. (1998) studied nanoparticle trans-
portation in the nano-DMA in order to reduce
nanoparticle loss and suggested a new inlet design
to reduce the recirculation problem. Here, the slit
width is reduced to improve the matching of the
flow velocity in the classifying region and to avoid
electric field penetration into the upstream side of
the entrance slit. As a result, the nano-DMA has
the potential for high resolution in sizing and
classifying nanoparticles. Figure 3 shows the cali-
bration results of nano-DMA measured by a
Scanning Mobility Particle Sizer (SMPS, Model
3080, TSI) and CPC (Model 3022A, TSI). These
size distributions were measured next to the nano-
DMA for the particle size classification and the
results show that nanoparticle size distribution
classified by nano-DMA is monodisperse and
Figure 2. Silver nanoparticle size distribution as a function of temperature.
Figure 3. Nano-DMA calibration.
119

acceptable for a discrete nanoparticle penetration
test.
Table 1 shows the specifications of the four
different fiberglass filter media tested in this study.
The filter media were made by the manufacturer,
Hollingsworth and Vose of East Walpole,
MA 02032, U.S.A, and were originally donated for
use in the establishment of the precision and
accuracy statement for the ASTM F1215-89
Standard ‘‘Standard Method for Determining the
Initial Efficiency of a Flatsheet Filter Media in an
Airflow Using Latex Spheres’’. The filter media are
of very low variability, with coefficients of varia-
tion for thickness, mass per area, initial pressure
drop and initial DOP penetration of less than 4, 1,
2 and 3%, respectively (Japuntich, 1991). The HE
series approaches HEPA regime for small particle
size and the HF series is more common to
standard HVAC systems. These filter media have
different combinations of supporting fibers to keep
filter shape and main fibers to capture particles,
and the filtration efficiency is proportional to
the amount of the main fibers. Figure 4 shows the
SEM images of H&V fiberglass filter media mag-
nified 500 times. The pore size of the HE filter
media is much smaller than that of the HF filter
media, and the main fiber ratio of the HE
filter media is much higher than that of the HF
filter media.
Table 2 shows a list of commercial filter media
that were tested in this study. Four different
electret filter media (media A, B, C, and D) are
made by the 3M Company and Lydall, Inc. and
applied on commercial respirators that are widely
used in the working field. Media E is a nanosized
e-PTFE (expanded polytetrafluoroethylene)
membrane filter medium made by W.L. Gore, and
is used for ultra high efficiency filtration industrial
applications. Figure 5 shows the SEM images of
the commercial filter media tested in this study.
The uniformity of the porosity is not as good when
compared to the H&V fiberglass filter media as
shown in the SEM images, but media E can be
expected to show high repeatability in test results
due to its uniform porosity all over the filter
medium.
Each filter sample was placed in a portable filter
holder with a filtration area of 17.34 cm
2
, and the
face velocity through the test filter medium was
controlled by a regulated vacuum pump located at
Table 2. Specifications of the specialized filter media
Name Type Manufacturing method
Media A Corona charged blown fiber (mid-size fiber) Melt blowing process
Media B Highly charged blown Fiber (mid-size fiber) Melt blowing process
Media C Split film fiber Film extrusion process
Media D Highly charged blown fiber (fine-size fiber) Melt blowing process
Media E e-PTFE Membrane Filter –
Table 1. Specifications of H&V fiberglass filter media
Filter Parameters Media
HE1073 HE1021 HF0031 HF0012
Thickness (cm) Ave. 0.053 0.069 0.074 0.074
%COV 2.3 4.3 2.3 2.3
Basis Weight (g/m
2
) Ave. 63.9 80.3 82.6 69.2
%COV 0.53 0.67 0.86 0.92
Pressure Drop at 5.3 cm/s (mmH
2
O) Ave. 8.4 4.7 3.5 1.3
%COV 1.48 1.35 1.94 1.47
DOP % Penetration 0.3 lm at 5.3 cm/s Ave. 12.8 39 45.8 79.9
%COV 2.2 1.7 0.92 1.24
Fiber Density(g/m
3
) – 2.4 2.4 2.4 2.4
Solidity – 0.050 0.049 0.047 0.039
Effective Fiber Diameter (lm) – 1.9 2.9 3.3 4.9
Effective Pore Diameter (lm) – 8.8 13.4 16.1 26.2
120

the end of the system. The tests were conducted
with face velocities of 5.3, 10.0, and 15.0 cm/s.
Prior to each measurement, a particle count mea-
surement at the downstream end of the test filter
with applying zero Volts to the nano-DMA was
conducted in order to check the leakage of the
filter holder. Because no particle can pass through
the DMA without any electric field inside of the
DMA, we can check a system leakage by checking
the zero particle count in case of zero Volts DMA
test. Furthermore, after switching the sampling
point from upstream to downstream, another zero
Volts DMA test was conducted to make sure that
there are no residual particles inside the sampling
tube.
Results and discussion
Nanoparticle penetration efficiencies were mea-
sured for four different fiberglass filter media, four
different electret filter media and one nanofiber
filter medium using silver nanoparticles. All
experimental results are shown in terms of the
percent penetration with respect to the electrical
mobility diameter that is classified by the
nano-DMA. Each data point is an average of
at least five replicates with the maximum and
minimum values as error bars. Figure 6 shows the
nanoparticle penetration of the H&V fiberglass
filter media at the face velocity of 5.3 cm/s, which
is a standard test velocity for a respirator filter
medium. The furnace setting temperature was
870°C, which can generate an adequate amount of
silver nanoparticles for the size range of 3 to
20 nm. The particle sampling time was 600 s for
particle sizes smaller than 5 nm and 60 s for the
rest of particle size in order to get more than 10
5
counts at the upstream and 10 counts at the
downstream sampling point so that 99.99%
efficiency can be detected with the high efficiency
filter. The results show very high uniformity with
small error bars. In the case of the HF 0012, which
has the lowest filtration efficiency among the H&V
fiberglass filter media, data were obtainable for all
particle sizes down to 3 nm, while the particle
Figure 4. SEM images of the H&V filter papers (500). (a) HF0012 (b) HF0031 (c) HE1021 (d) HE1073.
121

penetration less than 9 nm for the HE 1073 could
not be measured due to its high filtration effi-
ciency. In these cases, particles could not be
detected at the downstream sampling point, even
with an extended sampling time of 30 min. The
results show that particle penetration decreases
continuously down to 3 nm as expected from the
classical filtration theory, and there is no signifi-
cant evidence of the nanoparticle thermal rebound
down to 3 nm.
Figures 7 and 8 show the nanoparticle penetra-
tion of the H&V fiberglass filter media at the face
velocity of 10 and 15 cm/s, respectively. The
higher face velocities show a higher penetration
percentage due to a shorter residence time through
the filtration region. These results show the same
trend as the case of the 5.3 cm/s face velocity.
Figure 9 shows the nanoparticle penetration of the
commercial filter media at the face velocity of
5.3 cm/s. Nanoparticle penetration decreases
Figure 5. SEM image of the specialized filter media. (a) Media A (100) (b) Media B (100) (c) Media C (50) (d) Media D
(100) (e) Media E (3,000) (f) Media E (30,000).
122

continuously with decreasing particle size down to
3 nm with no evidence of thermal rebound. These
results show larger error bars than those of
the H&V fiberglass filter media due to the
non-uniformity of the fiber diameter, porosity and
fiber charging condition except media E as
mentioned previously.
Figure 10 shows the combination of the nano-
particle penetration measured in this study with
the submicron particle (from 20 to 200 nm) pene-
tration measured by Japuntich et al. (2007) for
H&V fiberglass filter media at the face velocity of
5.3 cm/s. They used a TSI 8160 automated filter
tester for the particle penetration test with sodium
chloride (NaCl) particles generated by an atom-
izer. As shown in the graph, the results agree well
with each other at the particle size of 20 nm, even
though different test particles were used in the
Figure 6. Nanoparticle penetration of the H&V filters at the face velocity of 5.3 cm/s.
Figure 7. Nanoparticle penetration of the H&V filters at the face velocity of 10.0 cm/s.
123

two studies. This is because the most dominant
filtration mechanism for nanoparticles is Brownian
diffusion, which is not affected by the particle
density.
Conclusion
In this study, the nanoparticle penetration was
tested with a wide range of filter media (four
glassfiber filter media, four electret filter media and
one nanofiber filter medium) using silver nano-
particles from 3 nm to 20 nm at face velocities of
5.3, 10 and 15 cm/s. Nano-DMA calibration and
adequate leakage tests show that the test system
can produce repeatable and reliable data. The
furnace setting temperature and particle sampling
time were determined experimentally in order to
generate enough amounts of silver nanoparticles in
the size range of 3–20 nm, so that 99.99% of effi-
Figure 8. Nanoparticle penetration of the H&V filters at the face velocity of 15.0 cm/s.
Figure 9. Nanoparticle penetration of the specialized filter media.
124

ciency can be measured for the high efficiency filter
media. The results show a very small variability
with small error bars for all filter media tested in
this study, even though each test was repeated
many times with a variety of test conditions
(operator, test date and sample). The particle
penetration decreases continuously down to 3 nm
as expected from the classical filtration theory, and
there is no significant evidence of nanoparticle
thermal rebound down to 3 nm for nine different
filter media with three different face velocities.
Further, the result shows a good agreement in the
overlapping size range of previous test results
using submicron particles (from 20 to 200 nm).
Acknowledgement
The authors thank the support of members of
the Center for Filtration Research (CFR): 3M,
Donaldson, Fleetguard, Samsung Digital Appliance,
Samsung Semiconductor, TSI, and W. L. Gore &
Associates.
References
AlonsoM.,Y.Kousaka,T.Hashimoto&N.Hashimoto,1997.
Penetration of nanometer-sized aerosol particle through wire
screen and laminar flow tube. Aerosol Sci. Technol. 27, 471–480.
Chen D.-R., D.Y.H. Pui, D. Hummes, H. Fissan, F.R. Quant &
G.J. Sem, 1998. Design and evaluation of a nanometer
aerosol differential mobility analyzer (Nano-DMA). J. Aer-
osol Sci. 29(5/6), 497–509.
Hinds W.C., 1999 Aerosol Technology: Properties, Behavior,
and Measurement of Airborne Particles. New York: John
Wiley & Sons.
Ichitsubo H., T. Hashimoto, M. Alonso & Y. Kousaka, 1996.
Panetration of ultrafine particles and ion clusters through
wire screen. Aerosol Sci. Technol. 24, 119–127.
Japuntich D.A., 1991 Particle Clogging of Fibrous Filters,
Ph.D. Thesis. Loughborough, U.K: Loughborough Univer-
sity of Technology.
Japuntich, D.A., L. Franklin, D.Y.H. Pui, T. Kuehn & S.C.
Kim, 2007. A comparison of two nano-sized particle air
filtration tests in the diameter range of 10 to 400 nm,
J. Nanoparticle Res.
Ku B.K. & A. Maynard, 2006. Generation and investigation of
airborne silver nanoparticle with specific size and morphol-
ogy by homogeneous nucleation, coagulation and sintering.
J. Aerosol Sci. 37(4), 452–470.
Maynard A.D., 2003. Estimating aerosol surface area from
number and mass concentration measurement. Ann. Occu-
pational Hygiene 47, 123–144.
Oberdo
¨rster G., 2000. Toxicology of ultrafine particles: in vivo
studies. Phil. Trans. Roy. Soc. London Ser. A 358, 2719–
2740.
Wang H.-C. & G. Kasper, 1991. Filtration efficiency of
nanometer-size aerosol particles. J. Aerosol Sci. 22, 31–41.
Figure 10. Comparison of test results with other study for H&V fiberglass filter media.
125