Science of the Total Environment 409 (2011) 48684874

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v

Removal of airborne nanoparticles by membrane coated lters

Jingxian Liu a, David Y.H. Pui b, Jing Wang c, d,
a

Filter Test Center, Northeastern University, Shenyang, Liaoning, China Particle Technology Laboratory, Mechanical Engineering, University of Minnesota, Minneapolis, Minnesota, USA c ETH Zurich, Institute of Environmental Engineering, 8093 Zurich, Switzerland d Empa, Analytical Chemistry, 8600 Dbendorf, Switzerland
b

a r t i c l e

i n f o

a b s t r a c t
The increasing amount of nanoparticles with the development of nanotechnology gives rise to concerns about potential negative impact on the environment and health hazards posed to humans. Membrane lter is an effective media to control nanoparticles. Three lters coated with polytetrauoroethylene (PTFE) membrane were investigated in this study. A series of experiments on the lter efciency and relevant parameters such as the particle size and face velocity were carried out. The data show that the efciency curves for the membrane lters demonstrate the typical shape of v for particle sizes from 10 to 300 nm at face velocities from 0.3 to 15 cm/s. Membrane lters with larger pore sizes have larger Most Penetrating Particles Sizes (MPPS), and the MPPS decreases with increasing face velocity. The efciencies decrease with increasing face velocity for particle sizes from 10 to 300 nm. We present the ltration efciency data as a novel three-dimensional graph to illustrate its dependence on the particle size and face velocity. The membrane coated lter can be considered as two combined layers, one brous layer and one membrane layer. We develop a new ltration efciency model which is a combination of the models for the two layers. Results from the model calculation agree with experimental data well. The study can help to optimize the lter product and to determine the operational parameters of lters, thus contributing to reduction of air pollution by rapidly emerging nanoparticles. 2011 Elsevier B.V. All rights reserved.

Article history: Received 21 June 2011 Received in revised form 20 July 2011 Accepted 8 August 2011 Available online 1 September 2011 Keywords: Nanoparticle Membrane lter Filtration efciency Face velocity Model

1. Introduction In recent years, nanotechnology is becoming a revolutionary eld due to its unique applications across a wide range of industries, from the eld of medicine to alternative energy technology. Nanoparticles often possess unique electrical, optical, chemical, and biological properties. On the other hand, the special properties of nanoparticles can also potentially lead to new hazards or increased risks to the environment (McMurry et al., 2004; Oberdrster et al., 2005; Maynard and Pui, 2007; Wang et al., 2011a). Filtration is the simplest and most common method for particle control and air cleaning (Maynard and Pui, 2007). Aerosol ltration is used in diverse applications, such as respiratory protection, air cleaning of smelter efuent, processing of nuclear and hazardous materials (Hinds, 1999), removal of asbestos bers (Spurny, 1986) and diesel particles (Kittelson et al., 1984). However, the process of ltration is complicated, and although the general principles are well known there is still a gap between theory and experiments (Hinds, 1999).

Corresponding author at: ETH Zurich, Institute of Environmental Engineering, 8093 Zurich, Switzerland. E-mail address: jing.wang@ifu.baug.ethz.ch (J. Wang). 0048-9697/$ see front matter 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2011.08.011

Two major classes of lters exist, the brous lters and the membrane lters. Both types of lters have been widely used in different industrial elds. The membrane lters generally possess relatively high solid fractions and may provide high efciency and excellent pressure drop features (Sutherland, 2008; Galka and Saxena, 2009; Kuo et al., 2010). The membrane lters rely more on the surface ltration than on the depth ltration for particles larger than the rated pore sizes in the membrane (Rubow and Liu, 1986; Ling et al., 2010). The ltration performance of membrane lters depends on lter structures, particle properties and operation parameters (Marre and Palmeri, 2001; Cyrs et al., 2010). The research on the relationship between efciency and particle size may help to optimize the lter product, and the research on the relationship of efciency and face velocity may help to determine the operational parameters of lters. Our study is limited to clean membrane lters. Loading of particles on membranes is important in practical applications; however, it is out of the scope of this study. Fibrous ltration has been extensively studied experimentally and theoretically; models based on the single-ber efciency are well developed and systematically documented by Brown (1993), Hinds (1999) and Lee and Mukund (2001). Studies on membrane lters are less compared to those on brous lters. While the capillary tube model has been shown to accurately predict the particle collection characteristics of Nuclepore membrane lters (Spurny et al., 1969; Manton, 1978, 1979;

J. Liu et al. / Science of the Total Environment 409 (2011) 48684874

4869

Marre and Palmeri, 2001; Cyrs et al., 2010), the brous lter model gives more accurate prediction for the conventional solvent-cast membranes (Rubow, 1981; Rubow and Liu, 1986). Good agreement can be found between the effective ber diameter used in the model and the diameter of the ber-like structures in the conventional membrane (Rubow and Liu, 1986). The samples in our study are coated with PTFE membrane, for which a model based on the effective ber diameter is suitable. In this study, a series of experiments on three commercial membrane lter samples were performed to study the effects of nanoparticle size and face velocity on ltration efciency. The ltration efciency data are shown as a novel three-dimensional graph with the particle size and face velocity as the axes. The graph gives an intuitive summary of the efciency data and can better illustrate the dependence of the Most Penetrating Particle Size (MPPS) on the face velocity. Meanwhile a new theoretical model is developed combining the efciencies of the brous layer and PTFE membrane layer. It is used to calculate the efciency of the tested lters, and the results are compared to experimental ones.

Table 1 Parameters of three membrane coated lters. Filter media Media Filter A Non-woven nylon/ PTFE membrane 0.43 137.63 3.24 1.5 Filter B Non-woven polyester/PTFE membrane 0.38 63.00 2.43 1.3 Filter C Non-woven polyester/PTFE membrane 0.15 29.03 2.20 1.5

Thickness [mm@3.45 kPa] Basis weight [g/m2] Permeability [cm/s @127 Pa] Effective pore size [m]

2.2. Experimental scheme In the process of ltration, the aerosol ow goes through the lter media at a given face velocity, meanwhile the particles are captured by the lter. The parameters of ltration efciency and pressure drop are two important factors for lters. In this study, the pressure drop of membrane lter samples at different face velocities were rst measured, then a series of ltration efciencies for different particle sizes at different face velocities was obtained. The parameters of the three membrane coated lter samples are listed in Table 1. Fig. 2 shows SEM images of the PTFE membrane lter A, which consists of a series of interconnected ber links between the adjacent void spaces. For particle generation, 0.1% NaCl solution was used. The mean particle size was about 50 nm. In the experiments, the particles of 10,

2. Experimental method 2.1. Experimental setup The experimental setup is shown in Fig. 1. An atomizer (TSI 3078) was used for generating aerosol particles with sodium chloride solution. A differential mobility analyzer (DMA, TSI 3080) was applied for particle classication according to the particle mobility sizes. A Po210 neutralizer was used to give Boltzmann equilibrium charging status to the classied particles. The lter holder was the supporter for lter samples. Two condensation particle counters (CPCs) were used for measuring the number concentrations of the particles upstream and downstream of the lter. A pressure gauge was used for measuring the pressure drop of the lter media. The ltration face velocity was controlled by the gas ow rate. Our experimental method involved measuring ltration efciency for monodisperse particles with the same electrical mobility. We have used this approach to study ltration of nanoparticles down to 3 nm (Kim et al., 2007; Wang et al., 2007), penetration through nanober composite lters (Wang et al., 2008a, 2008b), and ltration of nanoparticle agglomerates (Kim et al., 2009) and carbon nanotubes (Wang et al., 2011b, in press). This approach led to consistent results and easier data analysis compared to ltration tests using polydisperse aerosols.
Neutralizer

DMA

CPC P CPC Atomizer Filter

Fig. 1. The experimental setup.

Fig. 2. SEM images of the PTFE membrane lter A.

4870

J. Liu et al. / Science of the Total Environment 409 (2011) 48684874

1000 900 800

Filter A Filter B Filter C

100

99.995

Pressure drop (Pa)

700 600 500 400 300 200 100 0 0 2 4 6 8 10 12 14 16 99.975 99.98

Efficiency (%)

99.99
V=0.3cm/s V=1cm/s

99.985

V=5.3cm/s V=10cm/s V=15cm/s

Face velocity (cm/s)

Fig. 3. Pressure drop as a function of the face velocity for the three lter samples.

10

100

1000

Particle size (nm)

Fig. 5. Efciencies of Filter B at different face velocities.

20, 50, 100, 200, and 300 nm sizes were selected using the DMA and used to challenge the lter with face velocities of 0.3, 1, 5.3, 10, and 15 cm/s. 3. Result analysis 3.1. Pressure drop features Pressure drop is an important parameter to evaluate the performance of lters. High pressure drop means high energy consumption, which usually leads to high efciency. Pressure drop is related to the face velocity, and increases with increasing face velocity linearly. The pressure drop values of the three lter samples are shown in Fig. 3. Filter A had the lowest pressure drop. In contrast, Filter C had the highest one. Filter B was in the middle and close to Filter C. At the face velocity of 1 cm/s, the pressure drop values of Filters A, B and C were 39.2 Pa, 52.3 Pa and 57.8 Pa, respectively. Permeability is related to the pressure drop; higher permeability leads to lower pressure drop. The data in Fig. 3 are consistent with the data of lter permeability. 3.2. Efciency at different face velocities As air penetrates a lter, the trajectories of particles deviate from the air streamlines due to several mechanisms. As a result, particles may collide with the lter surface and become deposited on them. The

important mechanical mechanisms causing particle deposition include diffusion, inertial impaction, interception, and gravitational settling. Because Brownian motion generally increases with decreasing particle size, the diffusive deposition of particles is stronger when the particle size is reduced. Inertial impaction mechanism becomes stronger with increasing particle size and increasing air velocity. Interception and gravitational settling are also related to the particle size. The curve for the total efciency for all capture mechanisms vs. the particle size takes a typical v shape as shown in Brown (1993), Hinds (1999), Lee and Mukund (2001). The ltration efciencies for different particle sizes at different face velocities were measured in the experiments. The efciency data of Filters A, B and C are shown in Figs. 4, 5, and 6, respectively. As the particle size increases from 10 nm to 300 nm, the efciency curves demonstrate the typical shape of v for all samples. As the face velocity increases from 0.3 to 15 cm/s, the efciency decreases and the bottom of the v-shaped curve drops. The lowest point of the v-shaped curve is the minimum efciency and corresponds to the MPPS. At 5.3 cm/s, for Filters A, B and C, the minimum efciencies are 99.800%, 99.997% and 99.993%, respectively and the MPPS values are 100 nm, 70 nm and 100 nm, respectively. Membrane lters with larger pore sizes allow more penetration of large particles. The data of MPPS are consistent with the pore sizes of the three lter samples. The

100 100 99.9 99.99 99.995

Efficiency (%)

99.8

Efficiency (%)

99.985
V=0.3cm/s

99.7
V=0.3cm/s

99.98 99.975 99.97 99.965 10

V=1cm/s V=5.3cm/s V=10cm/s V=15cm/s

99.6 99.5 99.4 99.3

V=1cm/s V=5.3cm/s V=10cm/s V=15cm/s

10

100

1000

100

1000

Particle size (nm)

Fig. 4. Efciencies of Filter A at different face velocities.

Particle size (nm)

Fig. 6. Efciencies of Filter C at different face velocities.

J. Liu et al. / Science of the Total Environment 409 (2011) 48684874

4871

theoretical model of Lee and Liu (1980) shows that the MPPS decreases with increasing face velocity. It can be seen from the efciency curves that the bottom point is moving to the left as the face velocity increases. For Filter A, the MPPS values are 100 nm, 90 nm and 75 nm for the face velocities of 5.3, 10 and 15 cm/s, respectively. If we compare the efciencies of the three lter samples, Filter A has the lowest one, Filter B and Filter C have almost the same values, with Filter B slightly higher. This can be seen from Fig. 7, in which the efciencies of the three lter samples at the face velocity of 5.3 cm/s are compared. The quality factor (QF) is a parameter to evaluate the lter performance, which is dened as: ln1=P QF = P where P is the penetration and P is the pressure drop. The quality factor changes with the particle size and face velocity. The quality factor curves of the three lter samples at the face velocity of 5.3 cm/s are shown in Fig. 8. Generally, Filter B has higher QF compared to the others. For 50 nm particles, Filter B and Filter C have the same QF; for 100 nm particles, lter B has the highest QF and Filter A has the lowest one; for 300 nm particles, Filter C has the lowest QF.

0.06

0.05

Quality factor (-)

0.04

0.03
Filter A Filter B Filter C

0.02

0.01

0 10

100

1000

Particle size (nm)

Fig. 8. The quality factor of the three lter samples for different particle sizes at 5.3 cm/s.

3.4. Efciency surface analysis In order to better describe the ltration efciency as a function of both the particle size and face velocity, a three-dimensional graph of efciency surface with the particle size and face velocity is generated, as shown in Figs. 9, 10 and 11 for Filters A, B and C, respectively. The graphs for the three samples are similar, which indicates that all the three membrane lter samples possess similar characteristics. The shape of the efciency graph is like a half funnel. The efciency is relatively high for small and large particles at all face velocities; it is also relatively high at very low face velocities for the particle size range in our study. As the face velocity increases, the efciency for intermediate particle sizes (50100 nm) is becoming signicantly smaller than those for smaller and larger particles. Thus a trough region is formed on the efciency surface and it becomes deeper as the face velocity increases. The trough region represents the MPPS for different face velocities. The three-dimensional efciency surface gives a summary of the data and shows the MPPS intuitively.

3.3. Efciency for different particle size The collection efciency is related to various ltration mechanisms, which depend on the particle size. The data show that for small particles (such as 10 or 30 nm) and large particles (such as 200 or 300 nm), the ltration efciencies are higher than for the intermediate sizes (such as 50 and 100 nm). An increase in the particle size causes increased ltration by interception and inertial impaction mechanisms, whereas a decrease in the particle size enhances collection by Brownian diffusion. As a consequence, there is an intermediate particle size region where two or more mechanisms are simultaneously operating, yet none is dominant. This is the region where the particle penetration through the lter is a maximum and the lter efciency a minimum. All lters have specic particle sizes for which the efciencies are the lowest and the efciency decreases sharply. The MPPS values for the three samples are between 50 and 100 nm. At low velocities, the efciency decreases slowly with increasing velocity; but at high velocities, the efciency decreases sharply.
100

99.95

Efficiency (%)

99.9

99.85

Filter A Filter B Filter C

99.8

99.75

10

100

1000

Particle size (nm)

Fig. 7. Comparison of the efciencies of samples at 5.3 cm/s. Fig. 9. The efciency surface for Filter A.

4872

J. Liu et al. / Science of the Total Environment 409 (2011) 48684874

composite lters with multiple layers has been used by Wang et al. (2008a, 2008b) to study composite nanober lters. The efciency of the non-woven brous layer s is related to the single ber efciency sf (Hinds, 1999): s = 1 exp ! 4 Lsf ; Df 2

where is the solidity of the lter, L is the thickness, sf is the single ber efciency, Df is the ber diameter. The efciency of the membrane layer m is related to the pore structures, and the so-called effective single ber efciency mf. The relation is the same as Eq. (2), but with sf replaced by mf, ! 4 Lmf : m = 1 exp Df 3

Fig. 10. The efciency surface for Filter B.

4.1.1. Single ber efciency of the non-woven brous layer Particles in the ow are captured by bers in non-woven brous lter. For a single ber, the four major mechanical capture mechanisms are Brownian diffusion, interception, inertial impaction and gravitational setting. The total ltration efciency of a single ber is the sum of the efciencies by the individual ltration mechanisms. The overall single ber efciency sf can be written as: sf sfDiff + sfInter + sfimp + sfGrav 4

4. Model calculation 4.1. Filtration efciency model The membrane lter samples used in the experiment were made by surface coating technology. The lters included two layers: a non-woven brous layer and a PTFE membrane layer. Thus the function of ltration was due to the two layers. The total efciency of the membrane coated lter is: E = 1Pm Ps = 11 m 1 s ; 1

where sfdiff, sfInter, smp, sfGrav, are the single ber efciencies due to Brownian diffusion, interception, inertial impaction and gravitational setting mechanisms, respectively. These individual efciencies can be computed by (Lee and Mukund, 2001; Wang and Pui, 2009): sfDiff = 2:58   1 1=3 2=3 Pe ; K 5

sf inter =

1 R2 ; K 1 + R  i 1 h 0:62 2 2:8 R 27:5R Stk; 29:628 2 2K Vg =U ; 1 + Vg =U

where Pm and Ps are the penetrations for the membrane layer and non-woven layer, respectively; m and s are the efciencies of the membrane layer and non-woven layer, respectively. Such a model for

sfimp =

sfGrav =

where U is the face velocity; K is the hydrodynamics factor, K = df U 0.5 ln 0.75 0.25 2 + ; Pe is the Peclet number, Pe = D with D the particle diffusion coefcient; R is the ratio of particle diameter to ber diameter R = dp/df; Stk is the Stokes number, and Vg is the settling velocity. Calculation results show that the efciencies due to impaction and gravitational settling are negligible compared to those due to diffusion and interception for the particle size range in our study. 4.1.2. Effective single ber efciency of the membrane layer The PTFE membrane consists of a series of interconnected ber links between the adjacent void spaces. The ltration mechanisms are similar to those of the non-woven brous lters, also with the same four mechanical mechanisms. In application of the brous lter model to membrane lters, the actual lter thickness and solidity values are used in the model. The effective ber diameter is dependent of the micro-structure in the membrane lter. Rubow and Liu (1986) found that the effective ber diameter agreed well with the average diameter of the ber-like structures in membrane lters. We found that in our samples the ber-like structures have similar dimension as the pores. Thus the effective diameter is set to be 1.5 m, the same as

Fig. 11. The efciency surface for Filter C.

J. Liu et al. / Science of the Total Environment 409 (2011) 48684874 Table 2 Parameters used in model calculation. Parameters (Effective) ber size [m] Solidity Filter thickness [mm] Particle density [g/cm3] Mean free path [m] Viscosity [g/cm s] Fibrous layer 20 0.2 0.3 2.165 0.066 0.000181 Membrane layer 1.5 0.3 0.14

4873

100 99.8 99.6

Efficiency (%)

99.4 99.2 99 98.8 98.6 10

the average pore size in Filter A. The Knudsen number based on this effective ber diameter is about 0.09, which is considered to be in the slip ow regime. The Kuwabara ow eld in the slip ow regime is used (Pich, 1965). In the particle size range of our experiments, diffusion and interception are the dominant capture mechanisms. Thus we only include these two in the calculation of the overall single ber efciency. The overall single ber efciency can be written as (Rubow, 1981):       0 1 1=3 2=3 1 Pe 1=3 mf = 2:86 Pe 1 + 0:389 K K 0! 2 1 R 2 1+ + K 1+R R

100

1000

Particle size (nm)

Fig. 12. Comparison of the lter efciency curves from the model calculation and experimental data.

5. Conclusion Based on the experiments, analysis and model calculation, the following conclusions are obtained: (1) The efciency curves for all membrane coated lter samples show a typical shape of v for face velocities from 0.3 to 15 cm/ s and the particle sizes from 10 to 300 nm. (2) The MPPS values of lter samples are related to the pore sizes; larger MPPS corresponds to larger pore size. The MPPS decreases with increasing face velocity. (3) Under our experimental conditions, the efciency decreases with increasing face velocity for all particle sizes and decreases sharply at high velocities. (4) The efciency surface graph presents a typical half funnel shape with high efciency plateau for small and large particles and low face velocities. The trough region represents the MPPS at different face velocities.

where ' is a dimensionless coefcient used to present the degree of ow slippage at the solid surface boundary. The other terms are dened same as above.

4.2. Model calculation According to the theoretical models listed above, the total efciency of Filter A is calculated. Table 2 gives the parameters used in the calculation. We list the values of the ltration efciency from the experiment and model in Table 3 for the face velocity of 5.3 cm/s. The total efciency from the model has contributions from both the membrane and brous layers (Eq. 1). The model indicates that the efciency by the membrane layer is dominant for Filter A. Fig. 12 is the comparison of the efciencies between model calculation and experimental results. Fig. 13 gives the efciency surface graph from model calculation compared with experimental points. The gures show that the data from model calculation and experiments agree very well, and the theoretical model is suitable for predicting efciency of membrane coated lters. Pressure drop is inversely proportional to the square of the ber size if the slip effect is not considered (Davies, 1973; Hinds, 1999). Since the effective ber size in the membrane layer is signicantly smaller than the ber size in the brous layer (Table 2), the pressure drop of the membrane layer is substantially higher than that of the brous layer. Therefore, the membrane makes the major contribution in terms of both ltration efciency and pressure drop for Filter A.

Table 3 Filtration efciency from the experiment and model for Filter A at 5.3 cm/s. Particle size (nm) 10 20 50 100 200

Experiment efciency (%) 100 99.9980 Model efciency (%) 100 100 Efciency of the brous layer 64.4964 34.1619 s (%) 100 Efciency of the membrane 100 layer m (%)

99.9135 99.800 99.9664 99.9910 99.7817 99.9744 12.2050 5.8535 5.3602 99.9898 99.7681 99.9730 Fig. 13. Comparison of the efciency surface graph. The results from model calculation are represented by the surface and the experimental data are represented by the points.

4874

J. Liu et al. / Science of the Total Environment 409 (2011) 48684874 Ling TY, Wang J, Pui DYH. Measurement of retention efciency of lters against nanoparticles in liquids using an aerosolization technique. Environ Sci Technol 2010;44:7749. Manton MJ. The impaction of aerosols on a nuclepore lter. Atmos Environ 1978;12: 166975. Manton MJ. Brownian diffusion of aerosols to the face of a nuclepore lter. Atmos Environ 1979;13:52531. Marre S, Palmeri J. Theoretical study of aerosol ltration by nucleopore lters: the intermediate crossover regime of Brownian diffusion and direct interception. J Colloid Interface Sci 2001;237:2308. Maynard AD, Pui DYH. Nanoparticles and occupational health. .. ISBN-10-1-4020-58586Springer; 2007. p. 186. McMurry PH, Shepherd M, Vickery JS. Particulate matter science for policy makers: a NARSTO assessment. Cambridge, UK: Cambridge University Press; 2004. 660. Oberdrster G, Oberdrster E, Oberdrster J. Nanotoxicology: an emerging discipline evolving from studies of ultrane particles. Environ Health Perspect 2005;113: 82339. Pich J. The ltration theory of highly dispersed aerosols. Staub Reinhalt Luft 1965;5: 1623. Rubow, K. L., Submicrometer Aerosol Filtration Characteristics of Membrane Filters. Ph. D. thesis, University of Minnesota, 1981. Rubow KL, Liu BYH. Characteristics of membrane lters for particle collection. In: Raber RR, editor. Fluid Filtration: Gas, Volume I, ASTM STP 975. Philadelphia: American Society for Testing and Materials; 1986. Spurny KR. On the ltration of brous aerosols. Sci Total Environ 1986;52:18999. Spurny KR, Lodge JP, Frank ER, Sheesley DC. Aerosol ltration by means of nuclepore lters: structural and ltration properties. Environ Sci Technol 1969;3:45364. Sutherland K. Filtration overview: a closer look at depth ltration. Filtr Sep 2008;45(8): 258. Wang J, Pui DYH. Numerical investigation of ltration by bers with elliptical crosssections. J Nanopart Res 2009;11(1):18596. Wang J, Asbach C, Fissan H, Hlser T, Kuhlbusch TAJ, Thompson D, et al. How can nanobiotechnology oversight advance science and industry: examples from environmental, health and safety studies of nanoparticles (nano-EHS). J Nanopart Res 2011a;13:137387. Wang J, Chen DR, Pui DYH. Modeling of ltration efciency of nanoparticles in standard lter media. J Nanopart Res 2007;9:11725. Wang J, Kim SC, Pui DYH. Carbon nanotube penetration through a screen lter: numerical modeling and comparison with experiments. Aerosol Sci Technol 2011b;45:44352. Wang J, Kim SC, Pui DYH. Measurement of Multi-wall Carbon Nanotube Penetration through a Screen Filter. J Nanopart Res in press (Published online, 19 May 2011). doi:10.1007/s11051-011-0415-y. Wang J, Kim SC, Pui DYH. Investigation of the gure of merit for lters with a single nanober layer on a substrate. J Aerosol Sci 2008a;39:32334. Wang J, Kim SC, Pui DYH. Figure of merit of composite lters with micrometer and nanometer bers. Aerosol Sci Technol 2008b;42:7228.

(5) The membrane coated lter can be modeled with a brous layer and a membrane layer. The efciency results of model calculation agree well with experimental data. Acknowledgement The work was partially supported by the National Institute of Environmental Health Sciences grant # 1RC2ES018741-01 (sub-grant 100029-D) on Hazard Assessment and Risk Estimation of Inhaled Nanomaterials Exposure. The authors thank the support of members of the Center for Filtration Research: 3M Corporation, Boeing Company, Cummins Filtration Inc., Donaldson Company, Inc., Entegris Inc, Hollingsworth & Vose Company, Samsung Semiconductor Inc., Shigematsu Works CO., LTD, TSI Inc., and W. L. Gore & Associates and the afliate member National Institute for Occupational Safety and Health (NIOSH). References
Brown RC. Air ltration. London: Pergamon Press; 1993. Cyrs WD, Boysen DA, Casuccio G, Lersch T, Peters TM. Nanoparticle collection efciency of capillary pore membrane lters. J Aerosol Sci 2010;41(7):65564. Air ltration. In: Davies CN, editor. London: Academic Press; 1973. Galka N, Saxena A. High efciency air ltration: the growing impact of membranes. Filtr Sep 2009;46(4):225. Hinds WC. Aerosol technology: properties, behavior, and measurement of airborne particles. Second ed. New York, USA: Wiley-Interscience; 1999. Kim SC, Harrington MS, Pui DYH. Experimental study of nanoparticles penetration through commercial lter media. J Nanopart Res 2007;9:11725. Kim SC, Wang J, Emery M, Shin W-G, Mullholand G, Pui DYH. Structural property effect of nanoparticle agglomerates on particle penetration through brous lter. Aerosol Sci Technol 2009;43(4):34455. Kittelson DB, Moon KC, Liu BYH. Filtration of diesel particles. Sci Total Environ 1984;36: 1538. Kuo Y-M, Huang S-H, Lin W-Y, Hsiao M-F, Chen C-C. Filtration and loading characteristics of granular bed lters. J Aerosol Sci 2010;41(2):2239. Lee KW, Mukund R. Filtration collection. In: Baron PA, Willeke K, editors. Aerosol Measurement Principles, Techniques, and Application. 2nd Ed. John Wiley and Sons; 2001. Lee KW, Liu BYH. On the minimum efciency and most penetrating particle size for brous lters. J Air Pollut Control Assoc 1980;30:37781.