Conference PaperPDF Available

A New Methodology for Measuring Filtration Efficiency as a Function of Particle Aerodynamic Diameter Using a Monodisperse Aerosol Source

Abstract

*** Note: the accompanying slideshow is accessed via the "Linked Data" tab below *** The Aerodynamic Aerosol Classifier (AAC) uses a centrifugal force and sheath flow between two concentric rotating cylinders to produce a truly monodisperse aerosol classified by aerodynamic diameter with a range from 25 nm to 6.8 µm. The instrument is unique in being able to produce a stream of particles of known finite aerodynamic diameter; other technologies such as impactors and virtual impactors provide either an upper or lower cut. Since the classification technique is independent of particle charge state, there are no losses in throughput associated with charging efficiency, and no multiple charging peaks such as those commonly associated with the differential mobility analyser (DMA). Samples from two respiratory filtering half masks conforming to international standards for personal protective equipment were mounted in an in-line filter holder. These were placed downstream of the AAC which was run at a series of rotational speeds and sheath flows to select aerodynamic diameters from two atomised aerosol sources: sodium chloride (NaCl) and dioctyl sebacate (DOS). The upstream and downstream particle number concentrations were measured using a condensation particle counter (CPC). Tests at each particle size were run back-to-back replacing the AAC with a DMA long column, which was configured to select particles of the same aerodynamic diameter when converted from electrical mobility size space. This enabled direct comparison of transmission efficiency and filter penetration behaviour between the two methods of classification. It was found that the classifier throughput of the AAC was 1.5 to 5.5 times higher than the DMA across the particle sizes tested from both aerosols. Filter penetration curves were generally in close agreement; in specific cases where there was a disparity, multiple charging from the DMA was a possible explanation. To investigate this effect, a particle diameter was selected from the left side of the DOS size distribution so that multiply-charged particles were drawn over the peak. The DMA output was then scanned with the AAC either side of a face mask sample which showed the presence of doubly and triply charged particles from the DMA downstream of the filter, leading to error in the penetration calculation.
A NEW METHODOLOGY FOR MEASURING FILTRATION
EFFICIENCY AS A FUNCTION OF PARTICLE AERODYNAMIC
DIAMETER USING A MONODISPERSE AEROSOL SOURCE
Simon Payne
1
, Martin Irwin
1
, Tyler Johnson
2
, Jonathan Symonds
1
1
Cambustion Ltd., Cambridge, UK
2
Department of Engineering, University of Cambridge, UK
ABSTRACT
The Aerodynamic Aerosol Classifier (AAC) uses a centrifugal force and sheath flow
between two concentric rotating cylinders to produce a truly monodisperse aerosol
classified by aerodynamic diameter with a range from 25 nm to 6.8 µm. The instrument
is unique in being able to produce a stream of particles of known finite aerodynamic
diameter; other technologies such as impactors and virtual impactors provide either an
upper or lower cut. Since the classification technique is independent of particle charge
state, there are no losses in throughput associated with charging efficiency, and no
multiple charging peaks such as those commonly associated with the differential
mobility analyser (DMA).
Samples from two respiratory filtering half masks conforming to international standards
for personal protective equipment were mounted in an in-line filter holder. These were
placed downstream of the AAC which was run at a series of rotational speeds and
sheath flows to select aerodynamic diameters from two atomised aerosol sources:
sodium chloride (NaCl) and dioctyl sebacate (DOS). The upstream and downstream
particle number concentrations were measured using a condensation particle counter
(CPC). Tests at each particle size were run back-to-back replacing the AAC with a
DMA long column, which was configured to select particles of the same aerodynamic
diameter when converted from electrical mobility size space. This enabled direct
comparison of transmission efficiency and filter penetration behaviour between the two
methods of classification. It was found that the classifier throughput of the AAC was
1.5 to 5.5 times higher than the DMA across the particle sizes tested from both aerosols.
Filter penetration curves were generally in close agreement; in specific cases where
there was a disparity, multiple charging from the DMA was a possible explanation. To
investigate this effect, a particle diameter was selected from the left side of the DOS
size distribution so that multiply-charged particles were drawn over the peak. The DMA
output was then scanned with the AAC either side of a face mask sample which
showed the presence of doubly and triply charged particles from the DMA downstream
of the filter, leading to error in the penetration calculation.
KEYWORDS
Aerodynamic Diameter, Aerosol, Classifier, Filter Test Equipment,
Filtration Efficiency, Particle Size
INTRODUCTION
The Aerodynamic Aerosol Classifier (AAC) was developed by Tavakoli and Olfert (2013) as a means of
classifying particles by their aerodynamic equivalent diameter, which is equal to the diameter of a sphere
with unit density that has the same settling velocity as the particle under test. The AAC selects particle
size by using two rotating cylinders to balance opposing centrifugal and drag forces for the desired
aerodynamic diameter, so that particles move across a sheath flow to the outlet. This new principle of
aerosol selection is independent of the aerosol charge state and only limited by diffusion and impaction
losses. The benefits to filtration experiments are:
1. Aerodynamic diameter is the most relevant equivalent particle size for characterising aerosol
flow through filter media, since it relates to the relaxation time of particles (defined below) as
flow conditions change.
2. The aerosol flow is truly monodisperse.
3. The classification range extends from 25 nm up to 6.8 microns.
4. The higher transmission efficiency reduces uncertainty in filter penetration measurements.
The particle relaxation time,
τ
, is related to aerodynamic diameter, d
a
, electrical mobility diameter, d
m
,
and volume equivalent diameter, d
ve
, as follows:
 =
m
∙
B
=
C
c
d
a
∙
0
∙d
a
2
18
=
C
c
d
m
∙
eff
∙d
m
2
18
=
C
c
d
ve
∙
p
∙d
ve
2
18∙
(1)
where m is the particle mass, B is the particle mobility, C
c
is the Cunningham slip correction, µ is the
dynamic viscosity of the carrier gas,
ρ
0
is unit density,
ρ
eff
is the particle effective density,
ρ
p
is the
material density and
χ
is the particle shape factor.
In order to investigate filtration efficiency as a function of particle size, impactors based on aerodynamic
diameter or electrical mobility classifiers such as the Differential Mobility Analyser (DMA) have generally
been used. Impactor devices provide a 50% size cut at specified particle diameters, meaning
determination of size-dependent filtration efficiency is a resource-heavy task with large uncertainties.
The DMA classifies particles based on their electrical mobility diameter by applying an electric field to
induce a known electrostatic force (Knutson & Whitby, 1975); it enables selection of sub-micron particle
sizes but relies on the accurate charge-conditioning of input aerosol by a neutraliser dependent on
radioactive or X-ray charging methods. Uncharged particles do not pass through the DMA which
significantly limits the instrument’s transmission efficiency (particularly at particle sizes below 50 nm),
and while the output aerosol contains particles of the same electrical mobility, it is not truly monodisperse
due to the presence of multiple charges (particularly at particle sizes above 100 nm).
In this study, particles from two aerosol sources were classified by the AAC and DMA back-to-back and
their penetration through filter media from commercially available face masks was measured.
METHOD
Filter media
Test samples were cut from two disposable filtering half masks conforming to international standards
for personal protective equipment (PPE):
The 3M 8710E face mask conforms to EN 149 (2009) and is assigned the FFP1 protection level,
meaning <20% by mass particle penetration at 95 lpm and in most test cases <22% total inward
leakage (including through the face seal). The face mask consists of an outer filtration layer and
a ribbed inner shell.
The Honeywell D7030V face mask is a single-layer device with an exhalation valve that
conforms to the US NIOSH N95 rating, meaning <5% penetration of all airborne oil-free particles
(this is closest to the FFP2 protection level in EN 149).
In this study a Pall stainless steel 47 mm in-line filter holder was used as the filter test housing; 47 mm
diameter disc samples were punched out from the central region of each face mask, secured in the filter
holder and tested at a face velocity of 10 cm/s correct at 0°C and 101.3 kPa, i.e. standard temperature
and pressure (STP). This is approximately equivalent to the 95 lpm test flow rate specified in EN 149
through an entire filtering half mask with 150 cm
2
surface area. The test sample is supported underneath
by a 47 mm diameter stainless steel screen which comprises a 35 mm diameter grid mesh surrounded
by a 6 mm wide solid annulus, meaning the effective filtration area is 9.6 cm
2
. Therefore the 10 cm/s at
STP face velocity condition is achieved with a volume flow rate of 6.2 lpm correct at a room temperature
of 22°C. Additional tests were conducted with samples from the 3M 8710E face mask at 5 cm/s at STP
(equivalent to 3.1 lpm at 22°C). Since these tests involve initial particle penetration of flat filter media
sealed by an o-ring in the housing, it should be noted that filtration efficiency results may differ from a
set-up involving an entire filtering half mask with leakage around the face seal.
Aerosol generation
Polydisperse aerosols of sodium chloride (NaCl) and dioctyl sebacate (DOS) were generated by
atomising solutions using a BGI Collison nebuliser. NaCl is the solid aerosol specified for filter testing in
numerous international standards and a solution was prepared by dissolving 1% NaCl by weight in
deionised water. Excess atomised fluid was removed by a liquid trap. The NaCl aerosol was then dried
in a desiccating column filled with silica gel beads before entering the classifier. DOS was selected as
a second aerosol because it provided a distribution with a significantly larger count median diameter
(CMD) and forms particles of known spherical morphology and density (914 kg/m
3
), which facilitates an
accurate conversion between aerodynamic and mobility size distributions.
AAC scans of both aerosols shown in Figure 1 were recorded in conjunction with a TSI 3775
Condensation Particle Counter (CPC) which was set to 0.3 lpm flow. Dilution upstream of the classifier
was necessary to ensure the maximum number concentration throughout the scan remained below 5 x
10^4 p/cc, above which the 3775 CPC operates in photometric mode instead of single-count mode. For
NaCl, a HEPA filter dilution bridge with a dilution ratio (DR) of 5 was sufficient. The number
concentration of the DOS aerosol was substantially higher and a DR of approximately 150 was required;
this was provided by a modified version of the rotary disk diluter used on the Cambustion DMS500
particle analyser by Symonds et al. (2007).
Scans in mobility space were subsequently recorded using the DMA & CPC, together commonly referred
to as the Scanning Mobility Particle Sizer (SMPS). The data were corrected for multiple charges (He &
Dhaniyala, 2013) and a Gumbel distribution was fitted to estimate the cut-off for particles beyond the
range of the SMPS and obtain particle size statistics, which are listed in Table 1 along with the values
obtained from the AAC scans. The normalised size distributions from both instruments are plotted in
Figure 1 along with the conversion of each AAC scan to mobility space by taking the expressions
containing d
a
and d
m
in Equation 1 and substituting
ρ
p
/
χ
for ρ
eff
, where
ρ
p
is the material density and
χ
is
the particle shape factor (there is an inherent assumption that the particle density is equal to the bulk
material density, which is justified for the aerosols used in this study because the particles are not porous
or agglomerated). The values used for NaCl were ρ
p
= 2160 kg/m
3
and
χ
= 1.08 (Kelly & McMurry, 1992).
For DOS, ρ
p
= 914 kg/m
3
and
χ
= 1 (spherical particles). Since the charge-corrected SMPS data and
AAC-converted data are in good agreement, the same density and shape factor values were used to
convert the aerodynamic diameter setpoints to mobility size space for filter penetration measurements.
Figure 1: Normalised AAC & SMPS scans of NaCl (top) and DOS (bottom) particle size distributions
0.E+00
1.E+06
2.E+06
3.E+06
4.E+06
5.E+06
6.E+06
7.E+06
8.E+06
9.E+06
1.E+07
10 100 1000 10000
Normalised size spectral density (dN/dLog(Da)/cc)
Particle equivalent diameter (nm)
NaCl - AAC scan of aerodynamic diameter
NaCl - AAC scan converted to mobility space
NaCl - SMPS scan of mobility diameter
0.0E+00
5.0E+07
1.0E+08
1.5E+08
2.0E+08
2.5E+08
3.0E+08
10 100 1000 10000
Normalised size spectral density (dN/dLog(Da)/cc)
Particle equivalent diameter (nm)
DOS - AAC scan of
aerodynamic diameter
DOS - AAC scan converted
to mobility space
DOS - SMPS scan of
mobility diameter
Count median
diameter (nm)
Geometric
standard deviation
NaCl, aerodynamic 168 1.82
NaCl, electrical mobility 67 2.05
DOS, aerodynamic 287 1.85
DOS, electrical mobility 289 1.80
Table 1: Particle size statistics from AAC and SMPS scans
Size classification
In common with usual flow setpoints for size distribution scanning, a constant 10:1 sheath:sample flow
ratio was maintained for particle size selection on both classifiers.
At reference classifier conditions (295 K and 1.013 bar(a)), the AAC can classify aerodynamic particle
diameters from 32 nm to 3
µ
m at low flow (0.3/3.0 lpm sample/sheath flow) and from 202 nm to 6.8
µ
m
at high flow (1.5/15 lpm sample/sheath flow). The lower size limit of the AAC is sensitive to particle
diffusion and the maximum obtainable rotational speed (note that this can be brought down to 25 nm if
the sheath flow is reduced by approximately 25% at the maximum rotational speed, which will broaden
the transfer function).
The DMA used in these experiments was a TSI model 3080, consisting of a Krypton-85 radioactive
source with a 3081 long column. At 295 K and 1.013 bar(a), this configuration can classify mobility
particle diameters (assuming single charges) from 13 nm to 833 nm at low flow (0.3/3.0 lpm
sample/sheath flow) and 6 nm to 239 nm at high flow (1.5/15 lpm sample/sheath flow). The operation of
the DMA to classify particles and measure particle size efficiency of flat sheet filter media is covered by
ISO 29463-3 (2011), which stipulates for a minimum of six approximately logarithmically equidistant
sizes in a single test. Seven aerodynamic diameters were selected for each aerosol in these
experiments, which were converted from aerodynamic to mobility size space using the same procedure
given for size distribution data in the previous section; the particle sizes are listed in Table 2.
Particle equivalent diameter (nm)
NaCl,
d
a
50 75 100 150 200 300 500
NaCl,
d
m
27 41 56 88 122 191 333
DOS,
d
a
75 100 150 200 300 500 750
DOS,
d
m
81 107 160 212 317 526 787
Table 2: Aerodynamic and electrical mobility diameters for filter penetration
Additional filter penetration tests were conducted with DOS aerosol using aerodynamic diameters
exceeding 1
µ
m; the corresponding mobility diameters are beyond the range of the DMA. The maximum
selected aerodynamic diameter was 3
µ
m, which is the upper size limit of the 3775 CPC. The full ranges
of operating setpoints used on the AAC and DMA for both aerosols are shown in Figure 2. Note the
opposite directions in which sample flow through each classifier relates to setpoint particle diameter: the
sample and sheath flows in the AAC must be reduced where the cylinder rotational speed approaches
its maximum (700 rad/s) in order to select
small
particle sizes, whereas in the DMA the sample and
sheath flows must be reduced where the column reaches its maximum potential (10 kV) in order to
select
large
particle sizes.
Figure 2: AAC and DMA operating setpoints used in filter testing experiments (a constant 10:1
sheath:sample flow ratio was maintained on each classifier)
Since the DMA output comprises positively-charged particles only, ISO 29463-3 requires that a
neutraliser is installed downstream of the column so that an equilibrium charge distribution is prescribed
on the aerosol incident on the filter sample. Accordingly, the classifier output (both AAC and DMA) was
passed through a second Krypton-85 neutraliser before mixing with make-up air.
Measurement of particle penetration
The short-term particle penetration test outlined in Clause 8.2 in ISO 16900-3 (2012) begins with sealing
of the filter housing, then the filter is exposed to (polydisperse) aerosol and the specified flow is passed
0
100
200
300
400
500
600
700
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
10 100 1000 10000
Cylinder rotational speed (rad/s)
Sample flow rate (lpm)
Aerodynamic diameter (nm)
AAC sample
flow
AAC cylinder
rotational speed
0
1
2
3
4
5
6
7
8
9
10
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
10 100 1000 10000
Column voltage (kV)
Sample flow rate (lpm)
Electric mobility diameter (nm)
DMA sample
flow
DMA column
voltage
through the filter. After a stabilisation time of 3 minutes, logging of filter penetration is initiated (at 1
concentration reading per second) and the short-term penetration value is the average taken over the
following 30 seconds. The same procedure is followed in these experiments for size-selected NaCl and
DOS particles. As shown in Figure 3, the classified & neutralised aerosol mixes with HEPA-filtered
make-up air (to establish the correct filter face velocity), which is supplied by a compressed zero air
cylinder containing < 3 ppm H
2
O and passed through a mass flow controller. This meant relative
humidity measurements made downstream of the filter sample remained below 20%, which is
recommended by Li et al. (2012) to prevent NaCl particles deliquescing from the filter. A tee-piece with
a concentric tube was installed so that the make-up air enters around the tube, mixing into the classified
aerosol sample from an outer annular volume (in order to prevent impaction of particles at the junction).
Figure 3: Experimental schematic for classified aerosol filter testing
Downstream of the filter, 1.5 lpm sample flow was drawn into the 3775 CPC while the remaining flow
was drawn by a vacuum pump through a second mass flow controller. Particle number upstream of the
filter was recorded by bypassing the filter holder and directing flow through an equivalent length tube.
The percent penetration at each particle size was calculated as the ratio of down- to upstream
concentrations measured alternately with the same 3775 CPC. For each particle size the AAC was used
first and CPC readings up- and downstream of the filter were logged, then particle classification was
immediately switched over to the DMA (with diameter converted to mobility size space) and up- and
downstream readings were logged again. These steps were repeated for all other particle sizes. DOS
aerosol was passed through the rotary disk diluter (DR ~ 50) before entering either classifier to bring
the classified particle concentration upstream of the filter to a level within the CPC single particle
counting mode of operation; for the dried NaCl aerosol no dilution was required before the classifier
since post-classifier dilution by the make-up air was sufficient.
For a final experiment with DOS aerosol, a particle size halfway up the left-hand side of the size
distribution was selected with the DMA (set equivalent to the AAC in mobility space) to represent the
worst case for multiple charging, since the double and triple charge peaks would occur either side of the
CMD. Then the AAC was scanned downstream of the DMA to quantify the multiple charging effect, both
up- and downstream of a filter sample.
RESULTS
Classifier transmission efficiency
The average particle concentrations upstream and downstream of the face mask filter sample are
denoted by
n
1
and
n
2
respectively and were measured directly by the CPC. The average particle number
concentration in the classifier output is denoted by
n
0
and is equal to
n
1
multiplied by the make-up air
DR. As shown in Figure 4, the AAC transmission efficiency was 1.5 to 5.5 times higher than the DMA
across the particle sizes tested from both aerosols. Only at the smallest NaCl particle size tested
(
d
a
= 50 nm and
d
m
= 27 nm) did
n
1
for the DMA exceed
n
1
for the AAC, because the lower AAC sample
flow at this operating setpoint (see Figure 2) meant the make-up air DR was approximately 4 times
higher. For all particles with aerodynamic diameters
75 nm however,
n
2
for the AAC exceeds
n
2
for
the DMA, which reduces the uncertainty in the penetration calculation.
Figure 4: NaCl (top) and DOS (bottom) particle number concentrations transmitted through each
classifier and measured upstream and downstream of the 3M 8710E filtering half mask sample at a
face velocity of 10 cm/s at STP
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
10 100 1000
Particle number concentration (p/cc)
Aerodynamic diameter (nm)
NaCl, AAC n0
NaCl, DMA n0
NaCl, AAC n1
NaCl, DMA n1
NaCl, AAC n2
NaCl, DMA n2
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1.E+06
10 100 1000
Particle number concentration (p/cc)
Aerodynamic diameter (nm)
DOS, AAC n0
DOS, DMA n0
DOS, AAC n1
DOS, DMA n1
DOS, AAC n2
DOS, DMA n2
Penetration curves
The filter penetration is equal to
n
2
/
n
1
. For each penetration value the error bars in Figures 5 to 7
represent a two-sided 95% confidence interval (O’Shaughnessy & Schmoll, 2013), which was calculated
in accordance with the specifications for particle counting statistics in Clause 7 of ISO 29463-2 (2011)
based on the Poisson distribution.
The results from the classified NaCl tests on samples cut from both filtering half masks are shown in
Figure 5. The confidence intervals for penetration values obtained using the AAC and DMA overlap in
most cases, except for particles with aerodynamic diameters of 50 nm and 75 nm incident on the 3M
8710E mask, through which a significantly higher penetration of DMA-classified particles was observed.
Since these size cuts are drawn from the left side of the NaCl size distribution (see Figure 1) and appear
to fall below the most penetrating particle size – which is nearest
d
a
= 100 nm for the 3M 8710E mask –
a possible explanation is the presence of larger, multiply-charged particles in the DMA output that are
more highly penetrating through the filter sample.
The Honeywell D7030V face mask adheres to a higher filtration rating and the most penetrating particle
size shifts down, falling closest to
d
a
= 75 nm. This would mitigate the influence of multiply-charged
particles from the left side of the NaCl size distribution on the penetration curve. At
d
a
= 150 nm, the
penetration of particles classified by the AAC was significantly higher than with the DMA. This size is
above the most penetrating particle size of the filter sample and just below the CMD of the aerosol,
which means transmission of larger, multiply-charged particles through the DMA would lower the
penetration result.
The results from the classified DOS tests on the 3M 8710E mask are shown in Figure 6. The smallest
aerodynamic diameter selected, 75 nm, is on the left tail of the DOS size distribution (therefore the
relatively low particle concentration leads to a large confidence interval) and it is inconclusive whether
this lies at or above the most penetrating DOS particle size. The average penetration values for DOS
particles classified by the DMA are lower than those for the AAC except for the largest size selected
(
d
a
= 750 nm and
d
m
= 787 nm), where the lower transmission efficiency through the DMA and greater
dilution by make-up air meant
n
2
was only 3 p/cc and bound by a particularly large confidence interval.
In order to measure penetration of micro-scale particles from the right tail of the DOS size distribution
selected by the AAC, the rotary disk diluter upstream of the classifier was bypassed. This ensured
n
1
never dipped below 2000 p/cc. At
d
a
= 3
µ
m, the upper limit of the 3775 CPC,
n
2
< 1 p/cc.
Figure 5: Penetration of NaCl particles classified by AAC and DMA back-to-back through disc
samples cut from 3M 8710E and Honeywell D7030V filtering half masks
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
10 100 1000
Number penetration
Aerodynamic diameter (nm)
3M 8710E face mask at 10 cm/s
NaCl, AAC n2/n1
NaCl, DMA n2/n1
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
10 100 1000
Number penetration
Aerodynamic diameter (nm)
Honeywell D7030V face mask at 10 cm/s
NaCl, AAC n2/n1
NaCl, DMA n2/n1
Figure 6: Penetration of DOS particles classified by AAC and DMA up to d
a
= 750 nm through disc
sample cut from 3M 8710E filtering half mask (top) then re-plotted on logarithmic y-axis with
additional results up to d
a
= 3
µ
m using the AAC only (bottom)
A comparison of the penetration of AAC-classified particles through the 3M 8710E mask is shown in
Figure 7. Higher penetration is seen with NaCl than with DOS particles for
d
a
> 100 nm, where
interception starts to become a more efficient filtration mechanism than diffusion. Interception occurs if
a particle passes within one particle radius of a collecting element in the filter; the interception efficiency
is usually expressed as a function of the ratio of particle size to the size of the collecting element. In this
context, the most appropriate particle size metric is the volume equivalent diameter,
d
ve
. For
d
a
= 100 nm,
use of Equation 1 shows that
d
ve
= 56 nm for NaCl and
d
ve
= 107 nm for DOS, therefore the interception
length of DOS particles is greater at the same aerodynamic diameter.
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
10 100 1000
Number penetration
Aerodynamic diameter (nm)
3M 8710E face mask at 10 cm/s
DOS, AAC n2/n1
DOS, DMA n2/n1
0.0001
0.001
0.01
0.1
10 100 1000 10000
Number penetration
Aerodynamic diameter (nm)
3M 8710E face mask at 10 cm/s
DOS 75-750 nm, AAC n2/n1
DOS 300-3000 nm, AAC n2/n1
DOS 75-750 nm, DMA n2/n1
A new set of data was acquired at half face velocity; penetration of NaCl particles throughout the size
range tested is substantially lower at 5 cm/s compared with 10 cm/s at STP since the timescale for
diffusive transport of particles within the filter has been doubled.
Figure 7: Penetration of AAC-classified particles from two aerosol sources at two face velocities
through 3M 8710E face mask sample
AAC scans of multiple charging peaks
The effect of multiply-charged particles in the DMA output on the face mask penetration measurements
has been considered – to investigate this in more detail, a particle diameter on the left side of the DOS
size distribution was selected with the DMA and the output was scanned with the AAC to measure the
multiple charging peaks.
The relationship between
d
m
and the equivalent diameter,
d
n
, of a multiply charged particle with the
same electrical mobility and carrying an integer number of charges,
n
e
, is:
d
m
C
c
d
m
=
n
e
d
n
C
c
d
n
(2)
where
C
c
is the Cunningham slip correction.
The selected singly-charged particle diameter for this experiment was
d
a
= 150 nm (where there was a
significant disparity in the penetration through the Honeywell D7030V face mask between particles
classified by the AAC and DMA, as shown in Figure 5). In mobility space this diameter is 160 nm. The
locations of multiply-charged particles were calculated using Equation 2; as shown in the SMPS scan in
Figure 8, the doubly charged and triply charged particles are located either side of the peak of the size
distribution.
The neutralised DMA output was then passed through a Honeywell D7030V mask sample and AAC
scans were recorded upstream and downstream. It was estimated from the plot of the AAC scans in
Figure 8 that approximately 47% of particles emerging from the DMA were singly charged (and therefore
of the desired mobility diameter), while 30% were doubly charged, 18% were triply charged and the
remainder carried at least 4 charges. Triply charged particles are clearly distinguishable in the scan
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
10 100 1000
Number penetration
Aerodynamic diameter (nm)
AAC-classified particles
and 3M 8710E face mask
NaCl at 5 cm/s
NaCl at 10 cm/s
DOS at 10 cm/s
downstream of the filter sample. The estimated effect on the penetration measurement is a 1/8
th
reduction from 4% for entirely singly charged particles to approximately 3.5% for the multiply charged
mix. This error would be larger in cases where the most penetrating particle size shifted upwards (i.e. if
the flow rate were lower and the interstitial distances within the filter medium were greater).
Figure 8: Sizes of larger, multiply-charged particles with same electrical mobility as singly-charged
d
m
= 160 nm particle (top) which are present in AAC scans of the DMA output recorded upstream
and downstream of a face mask sample (bottom)
0.0E+00
5.0E+07
1.0E+08
1.5E+08
2.0E+08
2.5E+08
3.0E+08
10 100 1000
Size spectral density (dN/dLog(Da)/cc)
Electrical mobility diameter (nm)
DOS - SMPS scan
dm = 160 nm
dm = 249 nm (2 charges)
dm = 338 nm (3 charges)
dm = 422 nm (4 charges)
dm = 500 nm (5 charges)
0.0E+00
1.0E+04
2.0E+04
3.0E+04
4.0E+04
5.0E+04
6.0E+04
7.0E+04
8.0E+04
9.0E+04
1.0E+05
10 100 1000
Size spectral density (dN/dLog(Da)/cc)
Aerodynamic diameter (nm)
Honeywell D7030V face mask at 10 cm/s
Upstream AAC scan
Downstream AAC scan
CONCLUSIONS
In this study, an AAC and a DMA were used alternately to classify NaCl and DOS particles for face mask
penetration measurements. While NaCl is the most common solid aerosol source specified in
international standards, the known spherical morphology and density of DOS facilitates accurate
conversion between aerodynamic and mobility size space. The lower penetration of DOS particles
compared with NaCl particles at the same aerodynamic diameter is attributed to their lower density and
greater volume equivalent diameter, which means higher collection efficiency by the interception
mechanism.
The classifier transmission efficiency of the AAC was found to be 1.5 to 5.5 times higher than the DMA
across the particle sizes tested from both aerosols up to 1
µ
m. Additional tests were conducted with the
AAC on particles larger than 1
µ
m (beyond the range of the DMA), limited by the 3
µ
m limit of the CPC.
The 95% confidence intervals on penetration curves from each classifier generally overlapped; in
specific cases where they did not, it is suggested that multiple charging from the DMA was the cause.
AAC scans of the DMA output at
d
m
= 160 nm on the left side of the DOS size distribution revealed
substantial number concentrations of doubly and triply charged particles that were still visible
downstream of the face mask sample, distorting the penetration calculation.
The use of an AAC in filtration efficiency testing provides new insights into particle deposition as a
function of aerodynamic diameter and enables classification of sizes from 25 nm to 6.8
µ
m. Experimental
uncertainties are reduced due to the AAC’s generation of a truly monodisperse aerosol at high
transmission efficiency.
REFERENCES
CEN. (2009).
EN 149:2001+A1:2009 Respiratory protective devices - Filtering half masks to protect
against particles - Requirements, testing, marking.
He, M., & Dhaniyala, S. (2013). A Multiple Charging Correction Algorithm for Scanning Electrical Mobility
Spectrometer Data.
Journal of Aerosol Science
, 13-26.
ISO. (2011). Part 2: Aerosol production, measuring equipment and particle-counting statistics. In
ISO
29463:2011 High-efficiency filters and filter media for removing particles in air.
ISO. (2011). Part 3: Testing flat sheet filter media. In
ISO 29463:2011 - High-efficiency filters and filter
media for removing particles in air.
ISO. (2012). Part 3: Determination of particle filter penetration. In
ISO 16900:2012 Respiratory protective
devices - Methods of test and test equipment.
Kelly, W. P., & McMurry, P. H. (1992). Measurement of Particle Density by Inertial Classification of
Differential Mobility Analyzer–Generated Monodisperse Aerosols.
Aerosol Science and Technology
,
199-212.
Knutson, E. O., & Whitby, K. T. (1975). Aerosol Classification by Electric Mobility: Apparatus, Theory
and Applications.
Journal of Aerosol Science
, 443-451.
Li, L., Zuo, Z., Japuntich, D. A., & Pui, D. Y. (2012). Evaluation of Filter Media for Particle Number,
Surface Area and Mass Penetrations.
Annals of Occupational Hygiene
, 581-594.
O’Shaughnessy, P. T., & Schmoll, L. H. (2013). Particle Count Statistics Applied to the Penetration of a
Filter Challenged with Nanoparticles.
Aerosol Science and Technology
, 616-625.
Symonds, J. P., Reavell, K. S., Olfert, J. S., Campbell, B. W., & Swift, S. J. (2007). Diesel Soot Mass
Calculation in Real-Time with a Differential Mobility Spectrometer.
Journal of Aerosol Science
, 52-68.
Tavakoli, F., & Olfert, J. S. (2013). An Instrument for the Classification of Aerosols by Particle Relaxation
Time: Theoretical Models of the Aerodynamic Aerosol Classifier.
Aerosol Science and Technology
, 916-
926.
... Unlike the differential mobility analyser (DMA), size classification by the AAC is independent of particle charge state; this means there are no losses in throughput associated with charging efficiency and no multiple charging peaks. Applications where the AAC has already found use as a DMA alternative include aerosol charger efficiency testing , calibration of the condensation particle counter (CPC) (Symonds, 2018) and filter penetration measurements (Payne, Irwin, Johnson, & Symonds, 2018). ...
... Its application is demonstrated by measuring the evolution of the ambient particle size distribution in a room before, during and after operation of an air purifier. To support analysis of the scan data, measurements of particle penetration are made on a filter sample from the air purifier using the AAC as a monodisperse particle selector, following the same methodology presented previously at FILTECH (Payne, Irwin, Johnson, & Symonds, 2018). ...
... The area of the pleated "HEPA" filter within the DeLonghi AC150 is 0.83 m 2 when unfolded, meaning at 160 m 3 /hr the filter face velocity is 5.4 cm/s (correct at ambient temperature & pressure) or 5.0 cm/s at standard temperature and pressure (STP) i.e. 0°C and 1.013 bar(a). This same face velocity was set for the size-selected NaCl and DOS experiments using filter samples cut and sealed in a 47mm filter holder, following the exact same procedure for AAC filter testing reported at the previous year's conference (Payne, et al., 2018). ...
Conference Paper
Full-text available
*** Note: the accompanying slideshow is accessed via the "Linked Data" tab below *** The Aerodynamic Aerosol Classifier (AAC) is a novel instrument that classifies particles based on their aerodynamic diameter. This is accomplished by passing the aerosol between spinning concentric cylinders with a coaxial sheath flow. Particles that have aerodynamic diameters larger than the AAC setpoint are dominated by the centrifugal force and impact the outer cylinder surface, while smaller particles remain entrained in the sheath flow; only particles with a narrow range of aerodynamic diameters pass through the AAC classifier. Particle sizes between 25 nm and 6.7 microns can be selected independent of the aerosol charge state, with a high transmission efficiency limited only by diffusion and impaction losses. This technique can be used to measure an aerosol's aerodynamic size distribution by stepping the AAC setpoint while recording the classified particle number concentration. This stepping procedure requires long measurement times to achieve reasonable spectral resolutions. To overcome this limitation, a scanning AAC inversion was developed. Following a similar methodology as the Scanning Mobility Particle Sizer (SMPS) inversion, this new approach allows the aerodynamic size distribution to be continuously measured by varying the classifier speed while measuring the classified particle concentration. By exponentially ramping the classifier speed, the change in the centrifugal force a particle experiences during its classifier residence time is independent of its inlet time. Furthermore, when the scan time is sufficiently longer than the scan speed ramping constant, the scanning AAC transfer function converges to the same shape as the steady state transfer function. This new methodology is called the "Scanning Aerodynamic Size Spectrometer" (SASS). Application of the SASS is demonstrated by collecting accelerated measurements of the ambient air particle size distributions in a room before, during and after operation of a commercial room air purifier. Additionally, a filter sample was taken from the air purifier and measurements of the size-dependent filtration efficiency were made, using the AAC to select monodisperse test particles by aerodynamic diameter. The performance of the air purifier was assessed by comparing size-dependent penetration measurements with the full real-world ambient spectra and accounting for the air exchange rate between the inside and outside of the room.
... The experimental method was originally presented at the 2018 FILTECH conference (Payne, Irwin, Johnson, & Symonds, 2018). As shown in Figure 3, the filter holder was placed downstream of the AAC which was run at a constant 10:1 sheath:sample flow ratio to classify particles with aerodynamic diameters between 50 nm and 5 m. ...
Conference Paper
Full-text available
*** Note: the accompanying slideshow is accessed via the "Linked Data" tab below *** During the COVID-19 pandemic, face coverings have been widely worn in public spaces to capture respiratory particles produced by the wearer and reduce spread of infection. Transmission generally results from virus-laden particles produced by infected people while coughing, speaking and breathing; the particle size distribution can be very broad with multiple modes. In this study the filtration properties of various face mask materials were investigated across a wide range of particle sizes. The interception mechanism is important for large particle capture, for which aerodynamic diameter is the most useful measure of size. The Aerodynamic Aerosol Classifier (AAC) can select particle sizes between 25 nm and >5 micrometres by using rotating cylinders to classify particles with the desired aerodynamic diameter (i.e. a set ratio of centrifugal and drag forces) so that particles move across the AAC's sheath flow to its outlet. This principle of aerosol selection is independent of particle charge state and produces a truly monodisperse aerosol with a high transmission efficiency limited only by diffusion and impaction losses. For measurements of particle penetration through various types of face coverings, samples were cut and sealed in a filter holder. These coverings included washable cloth face masks consisting of woven fabrics (with and without filter inserts), disposable surgical masks consisting of non-woven fabrics and disposable filtering half masks (with examples conforming to the FFP1 and FFP2 classes specified in EN 149). The AAC was run at a series of rotational speeds to select aerodynamic diameters in the range 50 nm-5 micrometres from a dioctyl sebacate (DOS) aerosol. The filter face velocities correspond to volumetric flows through the entire mask specified in EN 149 for steady-state breathing resistance tests. Subsequently, the pressure drop versus flow rate characteristic of each material sample was measured and the trade-off between filtration efficiency and breathing resistance was evaluated.
Conference Paper
Full-text available
The Aerodynamic Aerosol Classifier (AAC) is a new instrument which classifies aerosol by aerodynamic diameter. Like the established Differential Mobility Analyzer (DMA), nanoparticles are added to a sheath of clean air. In the DMA, a transverse electrical force is applied to pre-charged particles, such that particles of the desired charge:drag ratio (i.e. electrical mobility size) emerge through a slot. In the AAC, a centrifugal force is applied by rotating the whole classifier, and particles of the desired mass:drag ratio (i.e. aerodynamic diameter) emerge. The classification size is varied by varying the rotational speed. Unlike an impactor, or virtual impactor, the particles selected are over a narrow finite range of diameters. The particles' charge state does not affect the measurement, so no charger or neutralizer is required and no correction is needed for multiple charges. Thus truly monodisperse aerosol is generated. Recently a data inversion has been developed to allow the AAC to measure full size distributions, when step-scanned and coupled with a downstream Condensation Particle Counter (CPC). In this paper we use this technique to measure ambient urban aerosols with the AAC from 25 nm up to 2.5 μm with high resolution. In addition, we use the AAC to measure ambient aerosols before and after filtration with commercial air purification devices. The AAC provides a convenient way to assess filter media by selecting monodisperse test aerosol by aerodynamic diameter and using the aerosol to assess size dependent filtration efficiency. Filter media from the air purification devices is assessed using this technique, and compared with the results obtained on the full real-world ambient spectra.
Article
A new aerosol particle classifier, the aerodynamic aerosol classifier (AAC), is presented and its classifying characteristics are determined theoretically. The AAC consists of two rotating coaxial cylinders rotating at the same angular velocity. The aerosol to be classified enters through a gap in the inner cylinder and is carried axially by particle-free sheath flow. The centrifugal force causes the particles between the rotating cylinders to move in the radial direction and particles of a narrow range of particle relaxation times exit the classifier through a gap in the outer cylinder with the sample flow. Particles with larger relaxation times impact and adhere to the outer cylinder and particles with smaller relaxation times exit the classifier with the exhaust flow. Thus, the aerosol is classified by particle relaxation time from which the aerodynamic equivalent diameter can easily be found. Four theoretical models of the instrument transfer function are developed. Analytical particle streamline models (with and without the effects of particle diffusion), like those often used for mobility classifiers, are developed for the case when the centrifugal acceleration field is assumed to be uniform in the radial direction. More accurate models are developed when this assumption is not made. These models are the analytical limiting trajectory model which neglects the effects of diffusion and a numerical convective diffusion model that does not. It is shown that these models agree quite well when the gap between the cylinders is small compared to the radii of the cylinders. The models show that, theoretically, the AAC has a relatively wide classification range and high resolution.Copyright 2013 American Association for Aerosol Research
Article
Statistical confidence in a single measure of filter penetration (P) is dependent on the low number of particle counts made downstream of the filter. This paper discusses methods for determining an upper confidence limit (UCL) for a single measure of penetration. The magnitude of the UCL was then compared to the P value, UCL ≤ 2P, as a penetration acceptance criterion (PAC). This statistical method was applied to penetration trials involving an N95 filtering facepiece respirator challenged with sodium chloride and four engineered nanoparticles: titanium dioxide, iron oxide, silicon dioxide and single-walled carbon nanotubes. Ten trials were performed for each particle type with the aim of determining the most penetrating particle size (MPPS) and the maximum penetration, Pmax. The PAC was applied to the size channel containing the MPPS. With those P values that met the PAC for a given set of trials, an average Pmax and MPPS was computed together with corresponding standard deviations. Because the size distribution of the silicon dioxide aerosol was shifted towards larger particles relative to the MPPS, none of the ten trials satisfied the PAC for that aerosol. The remaining four particle types resulted in at least 4 trials meeting the criterion. MPPS values ranged from 35 - 53 nm with average Pmax values varying from 4.0% for titanium dioxide to 7.0% for iron oxide. The use of the penetration acceptance criterion is suggested for determining the reliability of penetration measurements obtained to determine filter Pmax and MPPS.
Article
A density measurement technique based on the selection of a monodisperse aerosol with a differential mobility analyzer followed by classification according to aerodynamic diameter with an impactor has been designed and tested. Experimental results were obtained for several laboratory aerosols (dioctyl phthalate, (NH4)2SO4, NaCl, and H2SO4 at a range of humidities) by using four different microorifice uniform deposit impactor stages with aerodynamic diameter cut-offs of 0.12–0.56 Jim. The average error in measured particle densities is 4% and a maximum error of 8% is observed for all of the materials tested except NaCl, for which the measured effective density is 14% smaller than the true density. The discrepancy for NaCl is attributed to nonspherical particle shape. The system will be applied in the future to measure the densities of submicrometer atmospheric particles. This is Particle Technology Laboratory Report No. 817.
Article
An improved version of the Hewitt (differential) electric mobility analyzer was developed and its classifying characteristics were determined theoretically. The central mobility of the classified aerosol was found to be (qc + qm)/4πΛV, where qc and qm are the clean air and main outlet flows, respectively, Λ is a geometric factor, and Λ is the center rod voltage. The half-width of the mobility band was found to be (qa + qs)/4πΛV, where qa and qs are the aerosol and sampling outlet flows, respectively. These expressions were verified by the tests with a monodisperse aerosol of known size and low charge.
Article
The National Institute for Occupational Safety and Health (NIOSH) developed a standard for respirator certification under 42 CFR Part 84, using a TSI 8130 automated filter tester with photometers. A recent study showed that photometric detection methods may not be sensitive for measuring engineered nanoparticles. Present NIOSH standards for penetration measurement are mass-based; however, the threshold limit value/permissible exposure limit for an engineered nanoparticle worker exposure is not yet clear. There is lack of standardized filter test development for engineered nanoparticles, and development of a simple nanoparticle filter test is indicated. To better understand the filter performance against engineered nanoparticles and correlations among different tests, initial penetration levels of one fiberglass and two electret filter media were measured using a series of polydisperse and monodisperse aerosol test methods at two different laboratories (University of Minnesota Particle Technology Laboratory and 3M Company). Monodisperse aerosol penetrations were measured by a TSI 8160 using NaCl particles from 20 to 300 nm. Particle penetration curves and overall penetrations were measured by scanning mobility particle sizer (SMPS), condensation particle counter (CPC), nanoparticle surface area monitor (NSAM), and TSI 8130 at two face velocities and three layer thicknesses. Results showed that reproducible, comparable filtration data were achieved between two laboratories, with proper control of test conditions and calibration procedures. For particle penetration curves, the experimental results of monodisperse testing agreed well with polydisperse SMPS measurements. The most penetrating particle sizes (MPPSs) of electret and fiberglass filter media were ~50 and 160 nm, respectively. For overall penetrations, the CPC and NSAM results of polydisperse aerosols were close to the penetration at the corresponding median particle sizes. For each filter type, power-law correlations between the penetrations measured by different instruments show that the NIOSH TSI 8130 test may be used to predict penetrations at the MPPS as well as the CPC and NSAM results with polydisperse aerosols. It is recommended to use dry air (<20% RH) as makeup air in the test system to prevent sodium chloride particle deliquescing and minimizing the challenge particle dielectric constant and to use an adequate neutralizer to fully neutralize the polydisperse challenge aerosol. For a simple nanoparticle penetration test, it is recommended to use a polydisperse aerosol challenge with a geometric mean of ~50 nm with the CPC or the NSAM as detectors.
Article
This paper presents a methodology to allow a real-time particle size spectrometer to produce a mass concentration output by calculation from its electrical mobility response. As part of this, a Bayesian statistical algorithm for parametrising spectral data from the Cambustion DMS500 in terms of a number of lognormal functions is outlined, allowing the nucleation and accumulation modes of a Diesel aerosol to be treated separately and also to reduce mass calculation noise and improve spectral resolution. Previous literature is combined with new experimental results to develop a size:mass power-law relationship for this instrument. The effective density as a function of size for this instrument is found to be closer to that for water droplets than equivalent relationships for DMA/SMPS measurements in the literature, therefore making DMS500 mass calculation less susceptible to error from liquid adsorbed on agglomerates. The technique is validated with two Diesel engines against the gravimetric methods of filter paper and Diesel particulate filter (DPF) weighings. Good agreement is achieved over a variety of engine conditions, with the mean and standard deviation of the percentage difference of the calculated mass concentration from DPF weighings being -12.1±11.4% and from filter paper weighings being -15.1±18.0%. The calculated mass concentrations are systematically below those of the gravimetric methods.
Respiratory protective devices -Filtering half masks to protect against particles -Requirements, testing, marking
  • Cen
CEN. (2009). EN 149:2001+A1:2009 Respiratory protective devices -Filtering half masks to protect against particles -Requirements, testing, marking.
Part 2: Aerosol production, measuring equipment and particle-counting statistics
  • Iso
ISO. (2011). Part 2: Aerosol production, measuring equipment and particle-counting statistics. In ISO 29463:2011 High-efficiency filters and filter media for removing particles in air.
Part 3: Testing flat sheet filter media
  • Iso
ISO. (2011). Part 3: Testing flat sheet filter media. In ISO 29463:2011 -High-efficiency filters and filter media for removing particles in air.