United States Environmental Protection Agency Environmental Monitoring Systems Laboratory Research Triangle Park NC 27711 Research and Development EPA/600/S4-87/004 Apr. 1987 vvEPA Project Summary The PM10 Sampler Evaluation Program: January 1985 to July 1986 Mark Woods, Fu-Lin Chen, and M. B. (Arun) Ranade Test inlets designed to measure the atmospheric concentration of particu- late matter smaller than a nominal 10 micrometers (|im) aerodynamic diame- ter (PM10) were tested at the EPA wind tunnel test facility in Research Triangle Park, North Carolina. The Andersen Samplers Model 321A and the Wedding IP10 size selective inlets (SSI) were eval- uated following the procedures pro- posed in 40 CFR Part 53 (Ambient Air Monitoring Reference and Equivalent Methods). The tests consisted of meas- uring the inlet effectiveness over a range of particle sizes (3 to 20 |im nom- inal aerodynamic diameter) using both liquid and solid test particles at two windspeeds (2 and 8 km/h). Addition- ally two field-tested Wedding IP10 SSI's (EPA field study Phoenix, Arizona, May 1986) were compared with a clean "reference" Wedding IP10 SSI. Before conducting the above inlet tests, the procedures and requirements of 40 CFR Part 53 were revised to make the test procedures more practical. Several shortcomings in the sampling and analysis procedures as previously practiced were eliminated to ensure compliance with requirements of 40 CFR Part 53. At the time the tests were conducted, the wind tunnel and the wind tunnel test procedures fully met the requirements of 40 CFR Part 53. The Andersen Samplers Model 321A and the Wedding IP10 SSI's were both found to meet the 50-percent cutpoint requirement at 2 and 8 km/h for both liquid and solid particles. The "used" Wedding SSI's did not meet the 50- percent cutpoint criterion. However, all samplers met the expected mass ratio requirement. This Project Summary was devel- oped by EPA's Environmental Monitor- ing Systems Laboratory, Research Tri- angle Park, NC, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction On March 20, 1984, the U.S. Environ- mental Protection Agency (EPA) pro- posed revisions to the National Ambi- ent Air-Quality Standards (NAAQS) for paniculate matter. New primary stand- ards were proposed for particulate mat- ter measured as PM10 (particles with aerodynamic equivalent diameters less than a nominal 10 (im). A new Federal Reference Method (FRM) for the deter- mination of PM-ig in the atmosphere and provisions for the designation of refer- ence and equivalent methods for PM10 determination also were proposed. Under the provisions of these proposed regulations, candidate reference and equivalent methods for PM10 determi- nation would have to be tested in ac- cordance with the explicit procedures contained in 40 CFR Part 53 (Ambient Air Monitoring Reference and Equiva- lent Methods). Following satisfactory completion of all test requirements, candidate methods would be desig- nated formally as reference or equiva- lent methods. The proposed testing re- quirements of Part 53 include wind tunnel tests for sampling effectiveness and a 50-percent cutpoint. Sampling ef- fectiveness is the ratio of the mass col- lected by a test inlet compared to the mass collected by an isokinetic high- volume sampler. The 50-percent cut- ------- point is the particle size that is collected by the test inlet with 50 percent effi- ciency. Sampling effectiveness is determined in the following manner. A mono- disperse aerosol of known size is gener- ated using a Berglund-Liu vibrating ori- fice aerosol generator. The aerosol is introduced into the wind tunnel so that a uniform distribution of particles is achieved in the test section. The unifor- mity is evaluated by using an array of isokinetic samplers (referred to as an isokinetic rake) to measure the aerosol concentration profile in the sampling zone of the test section. To be accept- able, the concentration at each sample point on the rake must be within 10 per- cent of the mean concentration of all sample points on the rake. After sam- pling the concentration profile, a test inlet is placed in the wind tunnel and operated at a nominal flow rate of 40 cubic feet per minute (ft3/min) for 20 min. A high-volume isokinetic sampler (40 ft3/min) is then run for 20 min to determine the true wind tunnel aerosol concentration. The aerosol mass con- centration collected by both the test inlet and the isokinetic sampler is deter- mined using fluorometry. Liquid aero- sol particles contain uranine and solid particles are ammonium fluorescein, both of which are fluorescent materials. All inlet tests covered in this project summary included two inlets run se- quentially during a single test. The pro- cedure was to first run the isokinetic rake and then three replicates of the series consisting of test inlet one, the isokinetic high-volume inlet, and test inlet two. During most of calendar year 1984,13 tests were performed on various inlets at the EPA test facility. During most of calendar year 1985, shortcomings in the sampling and analysis procedures were identified and eliminated, and in 1986 inlet tests were resumed. The aerosol distribution system used during 1984 worked well at windspeeds of 2 and 8 kilometers per hour (km/h). However, at a windspeed of 24 km/h, the aerosol concentration was high at the center of the wind tunnel and low near the sides. The aerosol distribution system was modified (largely by trial and error) until acceptable performance was achieved at each of the three re- quired windspeeds. After the aerosol distribution system was redesigned, it was necessary to measure again the velocity and turbu- lence profiles in the test section. The ve- locity and turbulence profiles were measured using a TSI Model 1050-1 Anemometer and a ruggedized probe along with a TSI Model 1056 RMS, mean-square, direct current (dc) volt- meter. The probe was calibrated on a TSI Model 1125 calibrator, and the cali- bration curve was linearized using a fourth order least squares polynominal fit to the original data. The reproducibil- ity of the velocity profiles also was de- termined over the period of several weeks. The isokinetic high-volume sampler was used as the aerosol concentration reference for all sampling effectiveness calculations. Because a single isokinetic high-volume measurement was used as a reference for two test inlets (one run before the isokinetic reference and one run after the isokinetic reference), it was necessary to determine the repeatabil- ity of duplicate isokinetic concentration measurements. This was accomplished by making a series of four or five con- secutive isokinetic high-volume sam- pler concentration measurements. Ad- ditionally, it was necessary to stop the wind tunnel fan to replace inlets in the wind tunnel. Therefore, the effect on concentration of stopping the wind tun- nel fan between isokinetic sampler runs also was investigated. The rake was used to measure aero- sol uniformity, and the high-volume sampler was used to measure the wind tunnel aerosol concentration. Both are isokinetic sampling devices, and there- fore the two methods should yield iden- tical concentration values. In reality, a discrepancy exists, and a record has been kept on the quality of agreement obtained when both measurements are made. Fluorometry is the analytical tech- nique used to determine the mass con- centration of aerosol particles collected by a test inlet. An SLM Aminco Fluoro- Colorimeter II is used at the EPA wind tunnel. A calibration curve was devel- oped by preparing serial dilutions of a uranine solution (liquid particles) or of an ammonium fluorescein solution (solid particles). The precision of the in- strument was determined by measuring the fluorescence of five samples of the same solution concentration. After the test protocol was modified, the inlet tests were resumed. First, the Wedding IP10 SSI was run at a wind- speed of 8 km/h using liquid and solid aerosol particles. This was a prelimi- nary test designed chiefly to evalual the new test protocol. Following thi preliminary test, the Wedding IP10 SJ and the Andersen Samplers Mode 321A SSI were tested using both soli and liquid aerosol particles at wine speeds of 2 and 8 km/h. A "dirty" fiek tested Wedding IP10 SSI (Phoenix, Ar zona, May 1986) then was tested wit the "clean" Wedding SSI used in prev ous wind tunnel tests at a windspeed c 2 km/h with liquid particles. Liquid part cle tests then were conducted at 2 km/ on a cleaned, field-tested Wedding IP- SSI and the "clean" Wedding referenc inlet. Results The final aerosol distribution configi ration consisted of a six-point injectio system, a 4 ft x 4 ft baffle located jus upstream of the injection points, a 16-i diameter mixing fan located just dowr stream of the injection points (and d reeled into the bulk air flow), and flov straighteners located just upstream c the 4 ft x 4 ft baffle. Using this configu ration, the aerosol concentration a each rake nozzle was found to be withii ±10 percent of the mean concentratioi at 2, 8 and 24 km/h. The velocity and turbulence profile again were measured. Atypical velocit and turbulence profile obtained at windspeed of 8 km/h is shown in Fie ure 1. The air velocity is within 10 pei cent of the mean velocity, and the turbi lence intensity is less than 5 percent a all points measured in the test sectior The maximum deviation in day-to-da widespeed measurements at a give point was 1.7 percent, and the max mum deviation in turbulence intensit was 1.0 percent. Concentration stability was detei mined by making a series of isokineti high-volume sampler measurements £ each of the three required windspeed (2, 8, and 24 km/h). The coefficient c variation (the standard deviation d vided by the mean) was less than 3 pei cent at each of the three windspeeds The isokinetic measurements takei after stopping the wind tunnel fan be tween runs all fell within one standan deviation of their respective means ex cept in one case where the concentra tion was within two standard deviation of the mean. A record of the agreement betweei concentration values indicated by th isokinetic rake and the isokinetic high volume sampler has been kept for run ------- 10 9 I' Velocity Profile fb * = 6" above centerline 9 = Centerline O = 6" below centerline •no I j_ 18 30 42 Distance from East Wind Tunnel Wall (in) 54 5.0 a 4.0 3.0 Turbulence Intensity O + • O I j_ 6 18 30 42 54 Distance from East Wind Tunnel Wall (in) Figure 1. Velocity and turbulence profiles in the test section at 8 km/h. in which both measurements were made. The concentration indicated by the isokinetic rake has been on average 7 percent higher than the concentration indicated by the isokinetic high-volume sampler. The relative error of the fluorometer is calculated using the equation RE = (p x 100%)/(x) (1) where p = instrument precision (fxg/mL) x = measured concentration (n,g/mL). The precision of the fluorometer was determined over its range of operation using uranine solutions of 1, 0.1, 0.01, and 0.001 p,g/mL. Six samples were pre- pared at each concentration mimicking the procedures used to handle filter samples from an actual inlet test. Fluo- rometer readings were obtained for each solution concentration and con- verted to uranine concentrations using a current calibration curve. The stand- ard deviation of replicate uranine con- centration values was taken to be the precision of the fluorometer. The rela- tive error (required to be less than 5 per- cent in 40 CFR Part 53) then was calcu- lated using Equation 1. A plot of uranine concentration versus relative error shows that a uranine concentration of 0.014 (jig/mL corresponds to a relative error of 5 percent. Uranine concentra- tions encountered in the wind tunnel are typically much greater than 0.014 |j.g/mL except when small particles (3 or 5 n-m) are used at high windspeeds (24 km/h). A new rake flow system is being installed to provide for the collection of greater quantities of uranine in the same amount of sampling time. The two most important parameters obtained from the inlet tests are the 50- percent cutpoint and the expected mass ratio. The 50-percent cutpoint was de- fined previously. The expected mass ratio is the mass collected by the test inlet (found by integrating the inlet's sampling effectiveness curve against an assumed aerosol size distribution) com- pared to the mass predicted for an "ideal" sampler. The sampling effec- tiveness values for the ideal sampler and the aerosol size distribution are both found in 49 FR, 10461, Table D-3. The sampling effectiveness curve must be such that the test inlet has the 50- percent cutpoint between 9.0 and 11.0 (jim and an expected mass ratio be- tween 0.90 and 1.10. The results of the inlet tests are summarized in Table 1. Except where the test inlet is designated as a field test inlet, the same Wedding IP10 SSI and the same Andersen Model 321A SSI were used. The two field-tested Wedding IP10 inlets do not meet the 50-percent cut- point criterion; however, all inlets tested meet the expected mass ratio re- quirement. Both the Wedding IP10 SSI and the Andersen Samplers Model 321A SSI meet the 50-percent cutpoint requirement at windspeeds of 2 and 8 km/h using both solid and liquid aerosol particles. Conclusions and Recommendations The changes made in the test protocol at the EPA wind tunnel test facility have made it possible to meet all require- ments set forth in 40 CFR Part 53 regard- ing wind tunnel tests. The wind tunnel tests performed on the Wedding IP10 and Andersen Sam- plers Model 321A SSI's show that both inlets meet the proposed specification for 50-percent cutpoint and sampling ef- ------- Table 1. Summary of Inlet Test Results Test inlet Wedding IPW Wedding IPW Wedding IPW Wedding IPW Wedding IPW Wedding IPW Wedding IPW ("Dirty" field- tested inlet) Wedding IPW Wedding IPW ("Cleaned" field tested inlet) Model 321 A Model 321 A Model 321A Model 321 A Windspeed (km/h) a 8 8 2 2 2 2 2 2 8 8 2 2 Particle type3 L L S L S L L L L L S L S 50-percent outpoint (pm) 9.20 9.20 9.30 9.35 9.35 9.35 6.95 9.35 8.70 10.40 10.60 10.50 10.20 Slope (Vg)" 1.37 1.32 1.37 1.28 1.28 1.30 1.68 1.28 1.36 1.44 1.48 1.39 1.40 Expected mass ratio 0.976 0.985 0.990 0.989 0.987 0.989 0.936 0.989 0.965 1.021 1.025 1.030 1.026 aL = Liquid, S = Solid. bThe slope of the sampling effectiveness curve is estimated by the parameter cra = U.S. Environmental Protection Agency under Contract Number 68-02-3992 to Research Triangle Institute (RTI). It has been subjected to the Agency's peer and administrative review, and it has been approved for publication as an EPA document. fectiveness at windspeeds of 2 and 8 km/h using both solid and liquid aero- sol particles. The two field-tested Wed- ding IP10 SSI's met the sampling effec- tiveness criterion but not the 50-percent cutpoint criterion. The new rake flow system should be completed before tests are run at 24 km/ h, to ensure adequate quantities of ura- m'ne are collected for fluorescence measurements. The sampling effectiveness curves generated for the test inlets have not been corrected for the presence of mul- tiplets. Multiplets occur when two or three particles aggregate forming a sin- gle larger particle. At present, no single algorithm for the correction of multi- plets has been widely accepted. A single correction method should be developed and applied consistently to all sampling effectiveness measurements involving the PM10 test inlets. Th§ information in this document has beerrfiunded wholly or in part by the Mark Woods, Fu-Lin Chen, and M. B. (Arun) Ranade are with Research Triangle Institute (RTI), Research Triangle Park, NC 27709. Kenneth A. Rehme is the EPA Project Officer (see below). The complete report, entitled "The PMto Sampler Evaluation Program: January 1985 to July 1986." (OMer No. PB 87-145 801/AS; Cost: $18.95. subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Environmental Monitoring Systems Laboratory U.S. Environmental Protection Agency Research Triangle Park. NC 27711 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 '"••"- : :c-^~ 0,22 Official Business Penalty for Private Use S300 EPA/600/S4-87/OC4 ------- |