Deposition Removal of Monodisperse and Polydisperse Submicron Particles by a Negative Air Ionizer

Indoor particulate matter exposure is one of the main environmental risk factors influencing human health. Previous studies have indicated that the degree of hazards about ultrafine particles was associated with the particle surface area and number concentration. Polydisperse submicron particles (PSPs) contribute most of the indoor aerosol particle number and are usually used for removal testing. However, the behaviors of monodisperse submicron particles (MSPs) are easier to analyze. This study aims to investigate the difference between the deposition removal efficiency of MSP and PSP by the negative air ionizer (NAI) under different settings of freestream air velocity (FAV). The particle-removal experiments were conducted in a stainless steel test chamber under the condition of 50% relative humidity and FAV of 0.56, 1.2 and 2.0 m/s. The NaCl and oleic acid PSPs were generated from a constant output atomizer and the NaCl MSPs (30, 50, 100, 170 and 300 nm) were generated by passing the polydisperse NaCl aerosol through a Differential Mobility Analyzer. The concentration and size distribution of the particles in the chamber were monitored with a Scanning Mobility Particle Sizer (SMPS). The decay constant of particles number concentration, particle removal efficiency (PRE), enhancement factor of deposition rate (EF) and air cleaner effectiveness (ACE) were determined from the SMPS data. The decay constant of particle number concentration increased with the decrease of FAV and decreased with the increase of particle size. Both of the PREs of MSP and PSP were quite high, especially for the smaller particles and under low FAV. The EFs ranged from 4.4 to 115.5 and were proportional to the particle size. In addition, the ACE ranged from 0.57 to 0.96 and increased as the particle size increased. These results give a better understanding of the deposition of 30– 300-nm particles enhanced by the NAI and the influential factors.


INTRODUCTION
Up to 30 percent of the total burden of diseases was caused by particulate matter (PM) (Morawska et al., 2013), and most of the PM exposure happens in indoor environments where modern people spend most of their time (Klepeis et al., 2001).Thus, indoor PM exposure is one of the main environmental risk factors influencing human health.Ultrafine particles (UFP < 100 nm) have been shown to contribute 19% to 76% of indoor-generated particles (Morawska et al., 2013).The inhalation of UFP can cause numerous adverse health effects on the lungs, including pulmonary fibrosis, pulmonary inflammation, pleural effusion, granuloma, and increase of cancer risk (Song et al., 2009;Wang et al., 2009;Sze-To et al., 2012).The neuronal translocation through nasal tracheobronchial, olfactory and trigeminal nerves and translocation across the blood-brain barrier in certain regions of the brain might serve as pathways for central nerves system exposure to UFPs (Oberdörster et al., 2004).In view of the serious hazards caused by the inhalation of indoor aerosol particles, a focus on exposure control is severely needed.
Many researchers have shown that a negative air ionizer (NAI) can charge the airborne particles and force them to be deposited more rapidly than the uncharged ones due to the electrostatic force (Grabarczyk, 2001;Lee et al., 2004;Mayya et al., 2004;Grinshpun et al., 2005;Wu et al., 2006;Grinshpun et al., 2007;Shiue et al., 2011;Yu, 2012).Many factors have been shown to influence the effectiveness of air ionizer on the deposition of particles, such as particle size and composition, relative humidity, operation time of air ionizer and wall surface materials (Grabarczyk, 2001;Lee et al., 2004;Mayyaet al., 2004;Grinshpun et al., 2005;Wu et al., 2006;Grinshpun et al., 2007;Shiue et al., 2011;Yu, 2012).Other factors, including air movement and particle size distribution (monodisperse or polydisperse), may also affect the efficiency of NAI on the deposition removal of particles (Lai and Nazaroff, 2000;Zhao et al., 2004;Zhao and Wu, 2006).Generally, indoor aerosol particles are polydispersely distributed.However, the behavior of monodisperse particles is simpler and more easily proven by the theoretical model.There were rare researches on the deposition removal of monodisperse submicron particles (MSPs), with the enhancement of NAI.Understanding the difference between the deposition removal of MSPs and polydisperse submicron particles (PSPs), which has become an urgent work, can help us apply the theoretical model in the real situation.
The objective of this study is to investigate the effects of particle size distribution (MSPs and PSPs), the freestream air velocity (FAV) and particle composition (NaCl and oleic acid), on the particle deposition enhanced by the NAI.The particle sizes of 30-300 nm were targeted because most of pathogenic viruses, commonly used nano materials, and the most penetrating particle size of filters fall within these size ranges.Many previous studies have indicated that the degree of hazards about fine and ultrafine particles was associated with the particle number rather than particle mass (Donaldson et al., 1998;Oberdörster et al., 2005;Grass et al., 2010).Therefore, the particle removal focuses on the number concentration rather than the particle mass concentration.The result of this study can help establish a better understanding of the deposition of 30-300 nm particles enhanced by the NAI and the influential factors.

Experiment System
The schematic diagram of the experimental setup is shown in Fig. 1.In this setup, the zero air supply system which is composed with oil-free air compressor, diffusion dryer and active carbon and HEPA filter offered the clean air for the polished stainless steel cubic test chamber (60 cm × 60 cm × 60 cm).The relative humidity (RH) of the chamber air was maintained at 50% ± 3% by controlling the ratio of saturated air flow rate to dry air flow rate with a pair of mass flow controllers (MFCs).To investigate the effect of freestream air velocity on the deposition removal of particles, the experiments were conducted in the test chamber with a metal fan providing freestream air velocity of 0.56, 1.2 and 2.0 m/s (were defined as low, medium and high freestream air velocities).These air velocities approximately cover the range for mechanically ventilated indoor space and the ventilation duct (Lai and Nazaroff, 2000).The freestream air velocity, temperature, RH and CO 2 concentrations of the chamber air were recorded by the indoor air quality monitor (Q-Trak 7565-X, TSI Inc.) every one minute.The air velocity patterns in the chamber under various freestream air velocities are shown in Fig. 2. The air moved very fast to various directions and mixed well in the center of the chamber (mixing level > 95%).However, the air flow rate is much slow near the surface of the chamber wall.And the air flow patterns agree with the widely accepted description that there are three zones in a turbulent boundary layer (Lai and Nazaroff, 2000).
Both PSPs and MSPs were employed in the particleremoval experiments to investigate the effect of particle size distribution on the deposition removal of particles.The PSPs were generated from a Collison-type constant output atomizer (Model 3076, TSI Inc. USA).According to the electrical mobility of particles, the MSPs with size of 30, 50, 100, 170 and 300 nm were generated by passing the polydisperse aerosols through a Differential Mobility Analyzer (DMA, Model 3081, TSI Inc., USA).All the generated particles were neutralized by the Kr-85 radioactive Aerosol Neutralizer (Model 3077, TSI Inc., USA) before entering the chamber.To examine the effect of aerosol particles' relative permittivity on the deposition removal of particles enhanced by an operating NAI, we selected two kinds of aerosol particles: NaCl (relative permittivity = 6.1) and oleic acid (relative permittivity = 2.5) to conduct the particle-removal experiments.
In the particle-removal experiments, the aerosol particle number concentration and size distribution in the test chamber were monitored in real time by the Scanning Mobility Particle Sizer (SMPS, Model 3936L76, TSI Inc., USA).A commercial NAI (STR-11, Rayuan Co. Ltd.) with a carbon fiber emission source was used for testing.The concentration of negative air ion and strength of electrostatic field generated by the NAI were measured by an Air Ion Counter (AIC-20M, Alpha Lab, Inc. USA) and an Electrostatic Field Meter (EFM 231, Kleinwächter Ltd.), respectively.
Before each experiment got started, we would clean up the aerosol particles in the chamber by purging the chamber with zero air and operating the NAI for about half an hour, until the aerosol particle number concentration in the chamber reduced to less than 100 particles/cm 3 .After each experiment, we would clean up the chamber wall with 75% ethanol.
where k is the decay constant.From the experiment, two decay constants were obtained: the natural decay constant (k n ) which represents the decay rate constant of aerosol particle number concentration with no NAI operating, and the decay constant (k a ) which represents the decay rate constant of aerosol particle number concentration with an operating NAI.Enhancement factor, EF, is defined as the ratio of the effective decay rate constant of aerosol particle number concentration with an operating NAI to that without any NAI operating and this parameter represents the net effectiveness of NAI.EF is given by: where k is the decay constant of particles; λ is the air change per hour; the subscripts a and n denote the conditions with an NAI and with no NAI operating, correspondingly.Air cleaner effectiveness (ACE) of the NAI for the removal of aerosol particles in the test chamber is expressed as what follows (Waring et al., 2008): in which, β is the particle deposition loss-rate coefficient; V is the volume of test chamber; ECR is the effective clean rate

Statistical Method
The data size is small [replica for each data point (n = 3); particle sizes (n = 5), combined (n = 15)], and the type of data distribution is not a normal distribution.Thus, the correlations between various parameters and factors were tested with the nonparametric method (Spearman's rho) by using the SPSS Statistics 17 Software.

Natural Decay of Aerosol Particle Concentrations
In this study, the "natural decay" of aerosol particle represents the particle loss due to ventilation and deposition onto the chamber wall without any NAI operating.Because all the tested aerosol particles in this study were smaller than 0.5 μm, the inertial drift was assumed to be a minor factor for particle deposition.The Brownian diffusion and turbulent diffusion are the two mechanisms governing the particle transport through the boundary layer to the chamber wall.
Table 1 shows the particle removal efficiency (PRE) at various time points under different conditions.In general, smaller particles have higher removal efficiency.Besides, increasing freestream air velocity can enhance the removal efficiency of NaCl MSP.However, there seems to be no obvious impact of the freestream air velocity on the removal efficiency of PSP.
In the NaCl MSP deposition experiments, there was no observable coagulation between particles.On the other hand, the polydisperse coagulation coefficient is larger than the monodisperse ones and it increased rapidly when the ratio of the particle sizes increased as shown in Fig. 3. Additionally, the initial number concentrations of PSP were higher than 10 5 particles/cm 3 .Thus, particle coagulation occurred in the natural decay experiments of PSP as demonstrated in Fig. 4. The PSP coagulation led to a higher initial numberconcentration decay rate as compared to that of MSP (Fig. 5).However, when the particle number concentration reduced to < 10 5 particle/cm 3 , the natural number-concentration decay rate of PSP would slow down and the decay rate constant of PSP would become similar to the decay constants of MSP, as shown in Figs.4(a Although the coagulation of particles can reduce the particle number concentration, the coagulated particles were still suspended in the air and would still cause inhalation exposure.

Effect of Freestream Air Velocity and Particle Size on Natural Decay Rate
As demonstrated in Fig. 6, the natural decay constants of aerosol number concentration (k n ) increased with the increase of freestream air velocity.This result was associated with the friction velocity (u * ) which directly affected the particle transport from the region outside the boundary layer toward the chamber wall.For particles smaller than 0.5 μm, the deposition loss-rate coefficient (β) increased as the friction velocity increased.Friction velocity can be expressed as what follows (Lai and Nazaroff, 2000): in which ρ a is the density of air; τ w is the shear stress at thechamber wall and can be expressed as the following equation (Lai and Nazaroff, 2000): where, γ is the kinematic viscosity of air; U is the timeaveraged air velocity parallel to the chamber wall surface; y is the distance from the chamber wall surface.The average velocity gradient is a function of freestream air velocity (U ∞ ) (Lai and Nazaroff, 2000): in which L is a characteristic length of chamber surface.Substituting Eq. ( 7) into Eq.( 6), and then substituting the resulting equation into Eq.( 5), we have the friction velocity: According to Eq. ( 8), the friction velocity is proportional to the 9/10 power of freestream air velocity and in this study we employed the mean air speed (wind speed) to represent the freestream air velocity.We employed the Lai and Nazaroff'smodel (Lai and Nazaroff, 2000) to estimate the deposition loss-rate coefficient (β) of particle concentration: where A v , A u and A d are the areas of vertical, upward-facing and downward-facing surfaces, respectively; V is the volume of test chamber.The parameters v dv , v du and v dd are the deposition velocities of particles on vertical surface, upward horizontal surface and downward horizontal surface, respectively.The estimations of v dv , v du and v dd were described with details in Lai and Nazaroff's study (2000).In this model, the friction velocity and particle size are the two major factors that determine the deposition rate of particle toward the surface of chamber wall.
The theory of Lai and Nazaroff's deposition model is based on the understanding that there is a thin particleconcentration boundary layer within the turbulent boundary layer above a wall surface.Because the air velocity is low, no particle rebound or resuspension is considered (the surface is a perfect sink for particles).There is no source or sink of particle with the boundary layer.Thus, the particle flux, J, is constant through the concentration boundary layer and is described by a modified form of Fick's first law: in which D is the Brownian diffusivity of the particles; ε p is the particle eddy diffusivity; C is the particle number concentration in air; y is the distance from the chamber wall surface; v s is the gravitational settling velocity of particles.The terms on the right-hand side of Eq. ( 10 and Nazaroff, 2000).There is another assumption required for Eq. ( 10) to be valid: the air in the chamber is well mixed so that the particle-concentration gradient exists only very close to the deposition surface.In Lai and Nazaroff's model, the particle-concentration boundary layer is treated as three sublayers when defining the dependence of particle eddy diffusivity (ε p ) on air turbulent viscosity (ν t ), and the wall effect is also considered.The summation of β and air exchange per hour (λ) under different friction velocity is illustrated in Fig. 6.The trends of these curves (β + λ) are consistent with the natural decay constants of number concentration (k n ).The k n is negatively correlated with the particle size as shown in Fig. 7 and the correlation coefficients are shown in Table 2.This is owing to the larger particles possess higher inertia and does not repond to the turbulent fluctuations of air flow as well as smaller particles.Thus, the eddy diffusivity of larger particles in the boundary layer is lower than those of smaller ones.Consquently, larger particles deposit slower than smaller ones.

Enhancement Effect of NAI on the Deposition of Aerosol Particles
As shown in Table 1, there were still 0.5% to 76% of particles suspended in the chamber air after 24-minute "natural decay."However, the particles were totally removed within 6-24 minutes after the NAI started operating, indicating a powerful enhancement effect of NAI on the deposition of aerosol particles.When the NAI was operating, an electric field would be generated, and the aerosol particles would be charged by the negative air ion clusters.The number of charges obtained [n d (t)] by a particle by diffusion charging in a period of time t can be estimated by the following equation (Hinds, 1999): in which d p is the particle diameter; T is the absolute temperature; b is the Boltzmann constant (1.38 × 10 -23 J/K); K E is a constant (9.0 × 10 9 N m 2 /Coulomb 2 ); e is the electronic charge (1.6 × 10 -19 Coulomb); N i is the concentration of negative air ion generated by the NAI, and c i is the mean thermal speed of ions (240 m/s).The number of charges [n f (t)] acquired by a particle via field charging during a period of time t is (Hinds, 1999): where p is the relative permittivity of the particle; E is the electric field strength; Z i is the mobility of negative air ions (approximately 0.00015 m 2 /V•s).The terminal electrostatic velocity (v TE ) of a particle carrying [n d (t) + n f (t)] charges in an electric field is (Hinds, 1999): where Z p is the particle electric mobility; C c is the Cunningham slip correction factor; η is viscosity of the air.By measuring the N i and E generated by the NAI and using Eq. ( 12), we have the v TE , and the particle deposition lossrate coefficient (β NAI ) is: Eq. ( 13), which includes the effect of NAI on particle deposition, is analogous to Eq. ( 9).Fig. 6 demonstrates that the particle-size-dependent trend of β NAI + λ is similar to that of k a .Moreover, increasing freestream air velocity would reduce the concentration of negative air ions and electric field strength generated by the operating NAI, resulting in the decrease of the terminal electrostatic velocity of the particles.Thus, k a in "high" freestream air velocity was significantly lower than that in "low" and "medium" freestream air velocity.

Enhancement Factor of Particle Deposition Rate
The enhancement factor of particle deposition rate (EF) is the β NAI /β ratio, indicating that how many folds of particle deposition rate was increased by the operating NAI.As shown in Fig. 7, the EF is propotional to the particle size and ranges from 2.4 to 115.5, and the correlation coefficients are listed in Table 2. Increasing freestream air velocity would reduce the negative air ion concentration and electric field strength generated by the operating NAI, and would make the EF decrease.However, the effect of freestream air velocity on the EF of PSP was not as significant as EF of MSP.Kinematic coagulation of PSP driven by airflow might be one of the reasons to this phenomenon.Although increasing freestream air velocity would decrease the enhancement effect of NAI on the particle deposition.It enhanced the kinematic coagulation of PSP, which impaired the apparent impact of freestream air velocity on the EF.In   (Hinds, 1999): which is consistent with the positive correlation between EF and particle diameter as shown in Fig. 7.
The EF of oleic acid PSP deposition was lower than that of NaCl PSP and there were some reasons for this result.First, the relative permittivity of NaCl particle is higher than that of oleic acid particle.Therefore, NaCl particles can acquire more charges by field charging, which results in higher terminal electrostatic velocity.However, this effect is insignificant when particles are smaller than 200 nm.Additionally, only a minor fraction in the size distribution of oleic acid PSP was constituted by the particles smaller than 40 nm.In order to obtain an observable level of oleic acid particles smaller than 40 nm and facilitate the measurement of particle deposition, we employed a higher initial concentration of oleic acid PSP as compared to that of NaCl PSP.However, higher number concentration would result in a higher loss rate (higher k n ) due to coagulation and a lower ion/particle ratio, which might reduce the charging efficiency of an operating NAI.

Air Cleaner Effectiveness
The efficiency of NAI on the removal of aerosol particles in a given space may be evaluated by the "air cleaner effectiveness (ACE)" (Waring et al., 2008).The ACE assumes a well-mixed and steady-state space, is independent of indoor particle source, and is dependent on the volume of the spaces.Waring et al. (2008) has demonstrated that ACE can be obtained in an unsteady-state chamber by measuring the air exchange per hour of the test chamber (λ), size-resolved particle deposition loss-rate coefficient (β), and clean air delivery rate [CADR = (k a -k n )V] instead of measuring the steady-state particle concentration: where C AC and C No AC is the steady-state particle concentrations with and without an operating air cleaner, respectively.The air cleaner effectiveness ranges between 0% and 100%, which indicates the removal percentage of the particles.As shown in Fig. 8, the NAI's air cleaner effectiveness for particles larger than 100 nm was higher than 85%.However, the NAI was inefficient for the removal of oleic acid PSP of size smaller than 50 nm (ACE < 70%).The well-mixed space assumption and the dependence of the space volume are the two major considerations when the air cleaner effectiveness of this study is applied in real environments.
In general, our result of air cleaner effectiveness (ACE = 0.58-0.96) is comparable to previous studies as demonstrated in Table 3.Both Grinshpun's (Lee et al., 2004;Grinshpun et al., 2005) and Hu's groups (Shiue et al., 2011) found that ACE increased with the operation time of NAI.And we also found that the concentration of negative air ions would increase with the operation time of NAI.Our conclusion that in the testing range (30-300 nm particles) the ACE for larger particles is better than that for smaller particles is consistent with the result of Wu et al. (2006).

Implication in Indoor Environments
The concentration of negative air ion and electric field strength generated by the NAI decreased significantly with the distance away from the NAI (Yu, 2012).Thus, the effectiveness of NAI on the deposition removal of aerosol particles decreased considerably with the increase of room size and the distance between the NAI and wall.A poor performance could be expected, when the NAI is applied in a large room whose volume is much larger than that of the test chamber.However, a meticulous design, including an air cleaner equipped with an NAI and a deposition chamber, or application of NAI in the metal duct of a ventilation system, or using NAI in a confined indoor space such as a car or a cabin (Grinshpun et al., 2005), could help the NAI to achieve its better performance.Besides, there is a lower aerosol particle concentration in normal indoor environments than in our test chamber.Thus, a higher ion/particle ratio could be found in the indoor environments when applying the NAI, and this condition could be beneficial to the performance of NAI.The criterion for the efficient operation of NAI presents as what follows: v TE >> v dv , v du and v dd ; that is v TE ≥ 10 × v dv , v du and v dd .
In most cases in indoor environments, v du > v dv and v dd , and thus, the criterion becomes v TE ≥ 10 × v du .However, the relationships between ion concentration, v TE (Eq.( 12)) and v du are complicated: in which I is the Integral (Lai and Nazaroff, 2000).A practical method is to set a typical v du value (1 × 10 -5 -1.2 × 10 -4 m/s in this study) first, and then to calculate the corresponding concentration of ions (N i t > 3 × 10 14 ion sec/m 3 ; the value of t was assumed to be 1/k a = 70-250 seconds.And the concentration of ions was 1.2-4.28× 10 12 ion/m 3 ).Typical indoor particle concentration level ranged from 1 × 10 9 to 6 × 10 9 particles/m 3 , and thus the expected ion/particle ratio for rooms with the NAI operating is 200-4280.

Reference
This study Yu, 2012Shiue et al., 2011Waring et al., 2008Grinshpun et al., 2007Wu et al., 2006Grinshpun et al., 2005Lee et al., 2004Grabarczyk, 2001 Chamber The conductivity (or relative permittivity, p) of the particles could be another concern when applying the NAI in indoor environments.According to Eq. ( 11) the relative permittivity of particle materials could affect the number of charges acquired by a particle via field charging.However, the diffusion charging is the predominant mechanism for charging particle smaller than 200 nm (Hinds, 1999).Shin et al. (2009) employed three different materials (Ag, p = ∞; NaCl, p = 6.1; sucrose, p = 3.3) to exam the effect of relative permittivity of particle materials on unipolar diffusion charging of submicron particles and found that the material dependence of unipolar diffusion charging for particles ranging from 100 to 200 nm is insignificant.Therefore, even thought the indoor submicron particles cover various relative permittivity, the performance of NAI on the removal of these particles could be similar.

CONCLUSIONS
Because the coagulation coefficient of PSP was larger than that of MSP, the natural decay constant of PSP was higher than that of MSP.In the tested particle size region, the decay constant decreased with the increase of particle size.In contrast, the enhancement factor of particle deposition rate (EF)is positively proportional to the particle size.The EF of NaCl MSPs was higher than that of NaCl PSPs.By operating an NAI, the deposition rate of particles could be enhanced to as high as 100 times compared to that without any NAI operating.The air cleaner effectiveness (ACE) of NaCl MSP under the low freestream air velocity was the highest among all the test conditions and ranged from 94.3% to 96.0%, while the ACE of 30 nm oleic acid particles (PSPs) under high freestream air velocity was the lowest (57.6%) among all the test conditions.Both the EF and ACE decreased when the freestream air velocity increased.
Conclusively, the NAI performed effectively on enhancing the deposition removal of submicron particles.Although, the interaction between particles and negative air ions in the chamber was complicated and was involved with diffusion, field charging and coagulation, the trend of decay constant measured in the MSP and PSP experiments was corresponding to that predicted by the theoretical model.(Lai and Nazaroff, 2000) k: decay rate constant of aerosol particle number concentration; the subscripts a and n denote the conditions with an NAI and with no NAI operating, correspondingly K E : a constant (9.0 × 10 9 N m 2 /Coulomb 2 ) L: a characteristic length of chamber surface λ: air change per hour; the subscripts a and n denote the conditions with an NAI and with no NAI operating, correspondingly n d (t): number of charges obtained by a particle by diffusion charging in a period of time t n f (t): number of charges acquired by a particle via field charging during a period of time t N i : concentration of negative air ion generated by the NAI p: relative permittivity of the particle ρ a : the density of air γ: the kinematic viscosity of air u * : friction velocity T: absolute temperature τ w : the shear stress at the chamber wall U: the time-averaged air velocity parallel to the chamber wall surface v dv , v du and v dd : the deposition velocity of particle on vertical surface, upward horizontal surface and downward horizontal surface, respectively v s : gravitational settling velocity of particle v TE : terminal electrostatic velocity of a particle carrying [n d (t) + n f (t)] charges in an electric field V: volume of test chamber y: the distance from the chamber wall Z i : the mobility of negative air ions (approximately 0.00015 m 2 /V•s) Z p : particle electric mobility

Fig. 2 .
Fig. 2. Air velocity patterns in the chamber (a) high freestream air velocity (b) medium freestream air velocity (c) low freestream air velocity.

Fig. 5 .
Fig. 5.The time profile of (a)(d)NaCl MSP; (b)(e) NaCl PSP; (c)(f) Oleic acid PSP number concentrations under low freestream air velocity (a)(b)(c) NAI off; (d)(e)(f) NAI on; the short dash fitting curves in (b) and (c) have similar slopes to the corresponding fitting curves in (a), indicating that the decay rate is similar.

Fig. 6 .
Fig. 6.The relationship between decay constants and particle sizes under various friction velocities (a) NaCl MSP (b) NaCl PSP (c) Oleic acid PSP.FAV: freestream air velocity.

Fig. 7 .
Fig. 7.The relationship between EF and particle size (diameter) under various freestream air velocities (Low, Medium and High).

Fig. 8 .
Fig. 8.The relationship between ACE and particle size under different freestream air velocities (Low, Medium and High).

Table 1 .
Particle removal efficiency (PRE %) at various time points and under different conditions.Brownia coagulation coefficient K 1,2 for coagualtion of particles of diameters d p1 and d p2 .
p1 is the diffusion coefficient of particle with diameter of d p1 ; k is the Boltzmann constant (1.38 × 10 -23 J/K); T is absolute temperature; B is the mechanical mobility of particle with diameter of d p1 ; C c is Cunningham slip correction factor; η is the viscosity of air.

Table 2 .
The correlation coefficients (Spearman's rho) between particle size and various parameters under different experimental conditions.

Table 3 .
Brief comparison of ACE between this study and previous studies.