John G. Watson, Judith C. Chow, and Jitendra J. Shah. Environmental Research .... are termed "fine"), and 3 to 30 /xm (particles in this mode are termed "coarse"). ..... J. B. Wedding, M. Weigand, W. John, S. Wall "Sampling effec- tiveness of the ...
The Effect of Sampling Inlets on the PM-10 and PM-15 to TSP Concentration Ratios
John G. Watson, Judith C. Chow, and Jitendra J. Shah Environmental Research & Technology, Inc. (ERT) Concord, Massachusetts
Thompson G. Pace Environmental Protection Agency Office of Air Quality Planning and Standards Research Triangle Park, North Carolina
The products of sampling effectiveness as measured in wind tunnel tests and typical size distributions were integrated over particle size to determine mass collection efficiencies for the HIVOL sampler, the HIVOL with a size selective inlet, and the Sierra 244 and the Beckman SAMPLAIR dichotomous samplers. Hypothetical sampling effectiveness curves were proposed for a 10 fim inlet and the calculated mass collection efficiencies of such inlets were compared to those of existing 15 fim inlets. Average ratios of IP to TSP from measurements in EPA's Inhalable Particulate (IP) matter sampling network were compared to the calculated ratios; the agreement of these ratios is close.
There is much current interest in mass concentrations of size-classified suspended particulate matter (PM) as a result of the proposed 10 ^m standard, PM-10. It is therefore important to evaluate the particle collection characteristics of different sampling devices. Several simultaneous sampling comparisons have been made1 but these tests cannot cover the full range of enironmental conditions and sampler constructions which might be of interest. Wedding et al.2 define the "sampling effectiveness" of an aerosol sampling device as the functional relationship between the aerodynamic particle diameter and the fractional transmission of particles through the sampler inlet to the sampling substrate. These functions for several samplers are illustrated in Figures 1 and 2. Particle generation methods and wind tunnel tests to measure these sampling effectiveness functions have been devised and applied to achieve these curves.3'4-5'6
The "sampling efficiency" of a sampler is defined as the integral, over all particle sizes, of the product of the sampling effectiveness and particle mass distribution functions. This corresponds to the fraction of the total aerosol mass which is collected by the sampler. The sampling effectiveness depends only on the sampler and possibly on wind conditions. The sampling efficiency depends on both sampling effectiveness and the nature of the particles being sampled. The objectives of this study were to: • Develop a method for calculating sampling efficiency. • Calculate sampling efficiencies for standard high-volume samplers (HIVOL), the HIVOL with a 15 yum Anderson size-selective inlet (HIVOL(SSI)), the Beckman SAMPLAIR dichotomous sampler (DICHOTB), the Sierra 244 dichotomous sampler (DICHOTS), and a hypothetical 10 yum size-selective inlet, for typical particle size distributions. • Compare ratios of these calculated sampling efficiencies with the average mass concentration ratios obtained from simultaneous measurements with the aforementioned samplers.
100
Acceptable performance range, RTI
90
|
70
I 60 .1 50
u
75 40 A: B: C: D: E: F:
5 30 Dr. Watson was formerly with ERT, Environmental Research & Technology, Inc.; his present address is Desert Research Institute, P. 0. Box 60200, Reno, NV 89506. Ms. Chow and Dr. Shah are associated with ERT, Environmental Research & Technology, Inc., 696 Virginia Road, Concord, MA 01742. Mr. Pace is with the Office of Air Quality Planning and Standards, U.S. EPA, Research Triangle Park, NC 27711. This paper was submitted for review on August 10, 1982; the revised manuscript was received on November 29, 1982.
Copyright 1983-Air Pollution Control Association
114
Hypothetical 10 pm Step function 10 jjm HIVOL(SSI) HIVOL 24 km/h HIVOL 8 km/h HIVOL 2 km/h 0.5
1
1.0
2
1.5
2.0 ln(D)
2.5
3.0
3.5
4.0
3 4 5 6 7 8 10 15 20 30 40 50 60 70 Aerodynamic particle diameter (D), ^m
Figure 1. Collection effectiveness of standard HIVOL and HIVOL(SSI) under various wind speeds.4 Logarithms of particle diameters are placed on the abscissa to facilitate reading of collection effectiveness as a function of In D. The acceptable performance range was proposed for 15 / i samplers.
Journal of the Air Pollution Control Association
100
Sampling Efficiencies Acceptable performance range, RTI
90 ^ 80
Given the results of these wind tunnel tests, there is no doubt that the fractions of total suspended particulate matter measured by different samples depend on both the sampler design and atmospheric conditions. The expected difference between simultaneous collocated measurements with different instruments can be examined as follows. If the ambient mass concentrations are distributed as a function F(D) of aerodynamic particle diameter, D, and the sampling effectiveness of the sampler as a function of particle size is E(D), then the mass concentration, C, measured by the sampler is
.£ 70
40 Sierra 40 km/h Sierra 15 km/h Beckman 2 km/h Sierra 5 km/h
30 20
C = §F(D) E(D) dlnD
10 0.5
1.0
1.5
2.0 2.5 ln(D)
3,0
3 4 5 6 7 8 10 15 20 Aerodynamic particle diameter (D),
3.5.
4.0
30 40 50 60 70
Figure 2. Collection effectiveness of Beckman5 and Sierra 2446 dichotomous samplers. The Beckman SAMPLAIR collection effectiveness also varies with wind speed but only the 2 km/h curve was available.
A more detailed description of the procedures and results presented here is reported by Watson et al.7
Whitby and Sverdrup9 formulated seven different categories into which their hundreds of measured particle size distributions fell. In all of these categories, more than 90% of the aerosol volume was concentrated in the superposition of two log-normally shaped curves. Lundgren and Paulus10 also found atmospheric aerosols to have bimodal mass distributions which could be fitted by two additive log-normal curves. The two modes have been observed to have geometric mean diameters in the range of 0.2 to 0.7 //m (particles in this mode are termed "fine"), and 3 to 30 /xm (particles in this mode are termed "coarse"). The geometric mean diameter and geometric standard deviation for selected categories are listed in Table I. The atmospheric aerosol mass distribution, assuming it to be additive bimodal log-normal, may be represented by
Sampling Effectiveness
The sampling effectiveness curves of the sampling devices studied are shown in Figures 1 and 2. Included in each diagram is the acceptable performance range suggested by EPA8 for sampling particles less than 15 jum aerodynamic diameter. Figure 1 also includes two sampling effectiveness curves for hypothetical samplers with a 50% cut-size of 10 urn. Curve A was obtained by shifting the 15 fim HIVOL(SSI) curve. The slope5 of the 10 /im curve is 1.35. For comparison, a perfect 10 j«m sampling effectiveness curve (Curve B) with a slope equal to 1 is also considered. The sampling effectiveness of the dichotomous sampler inlets and of the HI VOL sampler inlet are dependent on wind speed for aerodynamic particle diameters greater than 7 /xm. The wind tunnel tests of the circular HIVOL(SSI) inlet show its sampling effectiveness to be virtually independent of wind speed.
(1)
-((ln D - In D/)2/2 ln2 af)
F(D) =
a) ^ e - ( d n D - In (2) In ov P = the fractional mass in the fine mode and (1 — P) is the fractional mass in the coarse mode; D — particle aerodynamic diameter; Df&Dc = the geometric mean aerodynamic diameters for the fine and coarse modes, respectively; af&oc= the geometric standard deviations for the fine and coarse modes. The integral of Eq. 1 was calculated by dividing the logarithmic size range indicated in Figures 1 and 2 into intervals of 0.2. For the tth interval c
d = F(Di)E(Di) A In D = 0.2 F{Di)E{Di).
(3)
Table I. Geometric mean aerodynamic diameters and geometric standard deviations for selected atmospheric particle size distribution.8
Aerosol Classification 9
Average
Fine mode Geometric mean aerodynamic Geometric diameter standard deviation Df
Coarse mode Geometric mean Geometric aerodynamic standard diameter deviation Dc
(fim)
0.38
2.02
10.20
2.26
Urban average
0.50
5.0
20.00
2.00
Background and aged urban plume9
0.47
1.84
7.27
2.12
Marine9
0.39
2.0
19.30
2.70
10
a
Optical diameters, Do reported in reference 9 have been converted to aerodynamic diameters, Da, by Da — yj~p x Do where a density, p, of 1.7 gm/cm3 has been used for the fine mode and 2.6 gm/cm3 has been used for the coarse mode. While this approximation may not be true under all circumstances, it is legitimate for making comparisons between the collection efficiencies of different samplers under a variety of circumstances. February 1983
Volume 33, No. 2
115
Table II. Theoretical collection efficiencies for different aerosol sampler inlets for different wind speeds and size distributions. HIVOL Samplers
Sierra Samplers % fine
5 km/h
Wind speed 15 km/h
40 km/h
2 km/h
Average
33 50
0.85 0.88
0.73 0.80
0.63 0.72
0.95 0.96
0.89 0.92
0.85 0.89
Urban average
33 50
0.67 0.75
0.52 0.63
0.42 0.56
0.89 0.92
0.76 0.82
0.70 0.77
Background and aged urban plume
33 50
0.92 0.94
0.83 0.87
0.73 0.80
0.97 0.98
0.94 0.96
0.91 0.93
Marine
33 50
0.70 0.77
0.59 0.69
0.51 0.63
0.89 0.92
0.77 0.83
0.72 0.79
Aerosol classification8
Wind speed 24 8 km/h km/h
Table II, continued
Aerosol classification8
Beckman HIVOLb Hypothetical PM-10b Lowerc Wind (SSI) envelope speed 2 (slope = 1.35) (slope = 1.0) km/h
Upperc envelope
Average
0.74 0.80
0.76 0.82
0.66 0.74
0.57 0.67
0.69 0.77
0.89 0.91
Urban average
0.55 0.65
0.55 0.66
0.45 0.58
0.37 0.52
0.47 0.59
0.71 0.78
Background and aged urban plume
0.82 0.87
0.85 0.89
0.76 0.82
0.67 0.75
0.79 0.85
0.95 0.96
Marine
0.60 0.70
0.61 0.71
0.53 0.65
0.47 0.60
0.55 0.66
0.73 0.80
a b c
See Table I for aerosol classification. Collection efficiencies are independent of wind speed. Collection efficiencies which would be achieved by inlets with effectiveness curves corresponding
to the lower and upper limits of Reference 8.
The particle diameter at which F(Di) was evaluated via Eq. 2 was chosen as the center of the interval. E{D{) was read from the curves of Figures 1 and 2 at the same value of In D. The calculations were carried out using the VISICALC® software tool to an APPLE II® + microcomputer. The functions F{D) and E{D) have been normalized so that the summation of fractional mass concentrations, 2C,-, over all intervals between .01 and 100 /urn gives the fraction of the total mass collected by the sampler under consideration. The maximum error introduced by numerical approximation over the intervals was estimated to be less than 3%. Other sources of error in this treatment result from the simplification with respect to real aerosol size distributions inherent in Eq. 2, measurement uncertainties of the collection effectiveness curves, and the lack of consideration of mass measurement interferences. For the present analysis, these errors are noted but not quantified. Limitations imposed on the conclusions because of these shortcomings will be noted where appropriate. The sampling efficiency of the sampler (i.e. mass collected by inlet/total suspended mass) for the size distributions specified by the parameters in Table I and the sampling effectiveness curves in Figures 1 and 2 were calculated and are presented in Table II. The fine particle fraction, P, of the total mass was set equal to 0.33 and 0.50 for these calculations. This has been shown to represent a reasonable ambient range.7 A close examination of the results in Table II provides some important insights into the variability of the different inlets 116
with respect to particle size distributions, wind speeds, and the proposed acceptable performance range for 15 /*m samplers. The sampling efficiencies of all inlets under all wind speed conditions vary with particle size distribution. The efficiencies increase as the total aerosol mass is shifted toward the smaller particles because the collection effectiveness increases as particle size decreases. The lowest efficiencies are obtained for the urban average and marine distributions, which Table I shows to have the largest geometric mean aerodynamic particle diameters (approximately 20 fim) in the coarse mode. At average wind speeds of 8 to 15 km/h DICHOTS efficiencies range from 0.52 to 0.87, HIVOL(SSI) efficiencies range from 0.55 to 0.89, and HIVOL efficiencies range from 0.76 to 0.96, for the size distributions considered. The size distribution which is representative of the most common sampling sites is probably the urban average10 with P = 0.33. For this distribution, the DICHOTS, HIVOL(SSI) and HIVOL sampling efficiencies at typical wind speeds are 0.52,0.55 and 0.76, respectively. The variation of sampling efficiency with wind speed is most pronounced for the DICHOTS and HIVOL samplers. For the urban average size distribution (P = 0.33) the sampling efficiency of the DICHOTS decreases from 0.67 to 0.42 as wind speeds rise from 5 km/h to 40 km/h. These extremes represent variations of +29% and -19% with respect to the efficiency under typical (approximately 15 km/h) wind speed conditions. For the HIVOL sampling the same size distributions, efJournal of the Air Pollution Control Association
Table III. Ratios of theoretical PM-15 sampler collection efficiencies to HIVOL collection efficiencies as a function of wind speed and particle size distribution.
Aerosol classification11
% fine
Relative3 PM-15 Collection Efficiencies with respect to HIVOL DICHOTS/HIVOL HIVOL(SSI)/HIVOL DICHOTB/HIVOL Wind speed Wind speed Wind speed c 5 km/h 40 km/h d 8km/h 24 km/h 2km/h 2 km/h 15 km/hc
Average
33 50
0.80 0.89
0.85 0.91
0.89 0.92
0.78 0.83
0.96 0.96
0.82 0.87
0.74 0.81
Urban average
33 50
0.62 0.72
0.72 0.80
0.79 0.86
0.62 0.71
0.88 0.91
0.68 0.77
0.60 0.73
Background
33 50
0.88 0.91
0.90 0.93
0.93 0.96
0.85 0.89
0.98 0.98
0.88 0.91
0.80 0.86
0.67 0.76
0.91 0.93
0.77 0.83
0.71 0.80
0.79 0.85 0.69 0.90 0.86 0.77 a Sampling efficiencies at the same or similar wind speeds are compared. b See Table I for aerosol classification. c Compared to 8 km/hr HIVOL efficiency. d Compared to 24 km/hr HIVOL efficiency. Marine (Whitby)
33 50
ficiencies decrease from 0.89 to 0.70 as wind speeds increase from 2 km/h to 24 km/h. These extremes constitute +17% and - 8 % deviations with respect to typical wind speed conditions. The upper and lower sampling efficiencies of the proposed acceptable performance range for PM-15 samplers offer a considerable range for acceptable sampling devices. The acceptable performance range exhibits +29% and -15% deviations with respect to the HIVOL(SSI) sampling efficiency, for which the sampling effectiveness curve falls in the center of the range (see Figure 1). The acceptable performance range collects 47 to 71% of a typical urban size distribution (see Table II). A sampler with a sampling effectiveness curve close to the lower envelope will collect 66% of the mass collected by a sampler with a sampling effectiveness curve close to the upper envelope. The DICHOTS, DICHOTB, and the HIVOL(SSI) sampling efficiencies fall within the limits of the acceptable performance range regardless of the size distribution sampled and the wind speed with the exception of the DICHOTS sampler under 40 km/h wind speeds. This is in contrast to their sampling effectiveness curves illustrated in Figures 1 and 2, for which only the HIVOL (SSI) sampler meets the criteria. HIVOL sampling efficiencies are marginally within the upper limit of the acceptable performance range under 24 km/h wind speed conditions, but they exceed this limit for all size distributions in 2 km/h and 8 km/h wind speeds. Both of the hypothetical PM-10 samplers exhibit sampling efficiencies less than that of the lower boundary of the acceptable performance range for PM-15 samplers. Each of the sampling effectiveness curves in Figures 1 and 2 is distinct from every other, but the sampling efficiencies listed in Table II are much more similar than would be initially suspected by comparing the effectiveness curves. This occurs
because one-third to half of the mass is in the size range for which the sampling effectiveness is close to 100%. Relative Sampling Efficiencies of Different Sampler Inlets
Three inter-sampler sampling efficiency comparisons are relevant to the interpretation of size-classified particulate matter data: 1. DICHOTS, DICHOTB, and HIVOL(SSI) measurements with respect to HIVOL measurements. 2. DICHOTS and DICHOTB measurements with respect to HIVOL (SSI) measurements. 3. Hypothetical PM-10 measurements with respect to DICHOTS, DICHOTB, and HIVOL(SSI) measurements. The first comparison is important because of the plethora of HIVOL measurements available and the desire to develop a predictive model allowing size-classified concentrations to be predicted from TSP data. The fraction of measured TSP which would be measured as PM-15 by collocated samplers under various wind speed and particle size distributions can be estimated by this comparison and is given in Table III. For the HIVOL(SSI) measurements, which show no wind speed dependence, this ratio increases with increasing wind speeds. In contrast, for the DICHOTS measurements, the ratios decrease with increasing wind speeds. The second comparison demonstrates the extent to which samplers which are intended to measure the same fraction of the aerosol mass, the PM-15 fraction, can be expected to do so. The ratios of collection efficiencies of these samplers tabulated in Table IV show the degree of equivalency which can be expected under ambient measurement conditions. The DICHOTS/HIVOL(SSI) ratio is close to, but generally less
Table IV. Ratios of theoretical DICHOT sampling efficiencies to HIVOL(SSI) sampling efficiencies as a function of wind speed and particle size distribution.
Aerosol classification8
February 1983
% fine
DICHOTS HIVOL(SSI) Wind speed 5 km/h 15 km/h 40 km/h
DICHOTB HIVOL(SSI) 2 km/h
Average
33 50
1.12 1.07
0.96 0.98
0.83 0.88
0.97 0.98
Urban average
33 50
1.22 1.14
0.95 0.95
0.76 0.85
1.00 0.98
Background
33 50
1.08 1.06
0.98 0.98
0.86 . 0.90
0.96 0.98
Marine (Whitby)
33 50
1.14 1.08
0.97 0.97
0.84 0.89
0.98 0.99
Volume 33, No. 2
117
Table V. Ratios of hypothetical PM-10 sampling efficiencies to PM-15 sampling efficiencies as a function of wind speed and particle size distribution.
2 km/h
PM-10/ HIVOL(SSI) All wind speeds
Average
33 50
0.78 0.84
0.90 0.93
1.05 1.03
0.89 0.93
0.87 0.90
Urban average
33 50
0.67 0.77
0.87 0.92
1.07 1.04
0.82 0.88
0.82 0.88
Background and urban plume
33 50
0.83 0.87
0.92 0.94
1.04 1.03
0.93 0.94
0.89 0.92
Marine
33 50
0.76 0.84
0.90 0.94
1.04 1.03
0.88 0.93
0.87 0.92
Aerosol classification3
a
% fine
PM-10/ DICHOTB
PM-10/DICHOTS Wind speed 5 km/h 15 km/h 40 km/h
See Table I for aerosol classification.
than 1.0. For the typical 15 km/h wind speed and urban average size distribution, the samplers should measure the same mass concentrations within 5% of each other. There is a discrepancy between the DICHOTB and DICHOTS samplers at low wind speeds; the DICHOTS samples up to 22% more aerosol mass than the DICHOTB or HIVOL(SSI) according to these tests. At high wind speeds, a substantial difference, approaching 25% between DICHOTS and HIVOL(SSI), is apparent. The final comparison is important to relate conclusions drawn from a PM-15 measurement system to a standard which would possibly address the mass fraction in the 0 to 10 yum size range (PM-10). The ratios of the hypothetical PM-10
sampler (slope = 1.35) sampling efficiencies to the DICHOTS, DICHOTB, and HIVOL(SSI) efficiencies appear in Table V. For the urban average aerosol (P = 0.33) and 15 km/h wind speeds, the calculations show that the PM-10 sampler would collect 87% of the mass collected by a DICHOTS sampler and 82% of the mass collected by a HIVOL(SSI). These factors, rounded to reflect their uncertainties, 0.9 for DICHOTS measurements and 0.8 for HIVOL(SSI) measurements, might be used to estimate mass concentrations in the 0 to 10 fxm size range from existing PM-15 measurements. These factors must be used with great care, however, with full recognition of the range of values they can take under different wind speed conditions and aerosol size distributions.
Table VI. Observed relationship among DICHOT, HIVOL(SSI), and HIVOL mass concentration measurements. HIVOL(SSI)/HIVOL
DICHOT/HIVOL
Average Study and samplers ratioa
Std.a
Watson et al.1 Sierra 244, Beckman SAMPLAIR, and HIVOL
0.73 0.74 0.68 0.74 0.81
0.18 0.14 0.21 0.16 0.28
1135 259 320 412 88
0.72 0.73 0.71 0.77 0.82
0.74 0.71 0.72 0.75
0.66 0.66 0.60 0.69 0.72
0.20 0.13 0.13 0.28 0.27
62 18 22 13 9
0.68 0.68 0.60 0.68 0.71
0.70 0.69 0.69 0.72 0.70
n
Suggs et al. Sierra 244, Beckman SAMPLAIR, and HIVOL
No. of Samples
Most similar theoretical Average ratiob ratio3
Std.a
No. of Samples
Most similar theoretical ratiob
0.13 0.15 0.15 0.19
683 318 217 129
0.72 0.72 0.72 0.72
35 12 15 5 3
0.72 0.72 0.72 0.72 0.72
c
0.10 0.08 0.07 0.09 0.12 -
DICHOT/HIVOL(SSI) Most similar theoretical ratjob
Average ratioa
Std.a
No. of samples
Watson et al. Sierra 244, Beckman SAMPLAIR, and HIVOL
0.97 1.00 0.93 0.93
0.24 0.21 0.27 0.21
285 102 99 64
0.97 1.06 0.97
Suggs et al.11 Sierra 244, Beckman SAMPLAIR, and HIVOL
0.81 0.91 0.76 0.83 0.72
0.01 0.17 0.25 0i39 0.07
27 8 13 4 2
0.83 0.90 0.76 0.83 0.76
Study and samplers 7
1.0
c
Site type All site types Industrial Commercial Residential Rural All site types Industrial Commercial Residential Rural
a
Arithmetic average and standard deviation of individual ratios. Value taken from Table III. c Insufficient HIVOL(SSI) data. b
118
Journal of the Air Pollution Control Association
Comparison of Calculated and Measured Relative Sampling Efficiencies
The sampling effectiveness curves and the sampling efficiencies derived from them in the previous sections are based on laboratory measurements and assumptions about size distributions which may not be valid under ambient sampling conditions. However, these results can be compared with those obtained from simultaneous sampling using DICHOTS, DICHOTB, HIVOL(SSI), and HIVOL samplers to evaluate their validity. Researchers have used three methods to establish the relationships between simultaneous measurements of two samplers, the average ratio, the ratio of averages, and the slope of a linear regression between one measurement and the other. Results of each of these methods from several studies comparing dichotomous sampler measurements are reported in Watson et al.7 The average ratio should be compared with the calculated values in Tables III and IV. This comparison is made in Table VI. The average DICHOT/HIVOL, HIVOL(SSI)/HIVOL, and DICHOT/HIVOL(SSI) ratios reported by Suggs11 are generally lower than those found by Watson et al.1 even though the basic data set is the same. (This may be due to the different methods of calculation and the different data validation procedures used.) Substantial variability in the individual measured ratios .exists, as evidenced by the large standard deviations (which range from 0.07 to 0.39 for different sitetypes). This is reasonable since the size distributions and the wind speeds vary significantly for the periods of time and geographical distances over which the samples were taken. The difference between the calculated and measured average ratios is small when the wind speeds and size distributions chosen for the calculated ratios are close to those expected at the type of site where the measurements were made. The calculated ratios come from size distributions and wind speed conditions that are plausible for the site-types. The DICHOT/HIVOL(SSI) ratio is an important one, since both of these samplers are used to estimate PM-15. The observed ratio varied from 0.72 to 1.05. Thus HIVOL(SSI) measured higher concentration of PM-15 than DICHOT. Part of the discrepancy is probably due to the filter artifact associated with alkaline filters employed by HIVOL(SSI) as described further by Watson et al.7 This comparison of measurements from different researchers using the same samplers highlights the difficulty of determining an experimental relationship between the collection efficiencies of different samplers. The data chosen and methods of summarizing those data must be carefully considered. Conclusions
The following conclusions can be drawn from this evaluation of sampling efficiencies: • Though the dichotomous samplers studied here do not meet the stated sampling effectiveness requirements of Reference 8, they do meet the sampling efficiency requirements under typical wind speeds and particle size distributions. HIVOL (SSI) samplers attain both sampling effectiveness and sampling efficiency requirements. A criterion for particle size sampling efficiency for acceptable sampling devices might be specified in addition to or in place of a criterion for sampling effectiveness. The efficiencies of samplers which would fall within the window of Reference 8 range from 47 to 71% for a typical urban size distribution. A sampler with a sampling effectiveness curve designed to meet the lower limit would sample 66% of the mass collected by a sampler designed to correspond to the upper limit. • PM-10 inlets can be expected to sample between 80 and 90% of the mass sampled with the present dichotomous February 1983
Volume 33, No. 2
•
•
and HIVOL(SSI) samplers. HIVOL(SSI) and dichotomous samplers collect equivalent mass concentrations within 5% of each other under typical situations. However, under low or high wind speeds, in particle size distributions with much coarse material, or when interferences are present in one or both of the samplers, this equivalency degrades. Results of the sampling efficiency calculations agree with average ratios determined from ambient measurements. These measured average ratios show substantial variability, however, and a knowledge of the ambient particle size distribution and wind speed during ambient sampling is required to truly verify the calculations.
Acknowledgment
This work was performed under EPA contract 68-02-2542, Task 6 for the EPA Office of Air Quality Planning and Standards. The authors would like to thank Steven Heisler and Benjamin Green from ERT for participating in several useful discussions and review of this manuscript. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. The statements made here are those of the authors and are not the official policy of the EPA. References
1. D. C. Camp, A. L. Van Lehn, B. W. Loo, Intercomparison of Samplers Used in the Determinations of Aerosol Composition, EPA-600/7-78-118, Research Triangle Park, NC, 1978. 2. J. B. Wedding, "Ambient Aerosol Sampling: History, Present Thinking, and a Proposed Inlet for Inhalable Particulate Matter," in Proceedings: The Technical Basis for a Size Specific Particulate Standard Parts I&II Specialty Conference, Air Pollution Control Association, Pittsburgh, PA, 1980. 3. J. B. Wedding, A. R. McFarland, J. E. Cermak, "Large particle collection characteristics of ambient aerosol samplers," Environ. Sci.Technol. 11:387(1977). 4. A. R. McFarland, C. A. Ortiz, C. E. Rodes, "Characteristics of Aerosol Samplers Used in Ambient Air Monitoring," presented at 86th National Meeting of the American Institute of Chemical Engineers, Houston, TX, 1979. 5. A. R. McFarland, G. A. Rost, "Design and Evaluation of an Automated Aerosol Sampler," Texas A&M Air Quality Laboratory Report 3397/01/79/ARM, College Station, TX, 1979. 6. J. B. Wedding, M. Weigand, W. John, S. Wall "Sampling effectiveness of the inlet to dichotomous sampler," Environ. Sci. Technol. 14:1367 (1980). 7. J. G. Watson, J. C. Chow, J. J. Shah, "Analysis of Measurements from the Inhalable Prticulate Matter Sampling Network," EPA-450/4-81-035, Research Triangle Park, NC, Dec. 1981. 8. M. B. Ranade, E. R. Kashdan, "Critical Parameters for the Federal Reference Method for the Inhalable Particulate Standard," Final Report Technical Directive 222, EPA Contract 68-02-2720, Research Triangle Park, NC, 1979. 9. K. T. Whitby, G. M. Sverdrup, "California Aerosols: Their Physical and Chemical Characteristics" in The Character and Origins of Smog Aerosols, edited by G. M. Hidy, P. R. Mueller, D. Grosjean, B. R. Appel, J. J. Wesolowski, John Wiley & Sons, New York, 1980. 10. D. A. Lundgren, H. J. Paulus, "The mass distribution of large atmospheric particles," JAPCA 25:1227 (1975): 11. J. C. Suggs, C. E. Rodes, G. E. Evans, R. E. Baumgardner, "Inhalable Particulate Network Annual Report: Operations and Data Summary (Mass Concentrations Only) April 1979-June 1980," U.S. EPA Environmental Monitoring Systems Laboratory Report, Research Triangle Park, NC, 1981. 12. S. Miller, "Middle scale distribution of inhalable particulates and TSP within a metropolitan area," JAPCA 30:1320 (1980). 13. J. C. Wendt, K. J. Torre, "Field Test of Four Size-Segregated Samplers," presented at 74th annual APCA Meeting, Philadelphia, PA, 1981. 14. G. E. Pashel, D. R. Egner, J. C. Winkley, R. E. Sistek, "Preliminary Results of AISI Particulate Matter Monitoring Study" in Proceedings: The Technical Basis for a Size Specific Particulate Standard Parts I & II Specialty Conference, Air Pollution Control Association, Pittsburgh, PA, 1980, p. 252. 15. J. A. Grantz, "Inhalable Particulate Matter in the Vicinity of an Integrated Iron and Steelmaking Complex", presented at 74th annual APCA Meeting, Philadelphia, PA, 1981. 119
