Environ. Sci. Technol. 1987, 21, 178-182 Isoaxial Aerosol Sampling: Nondimensional Representation of Overall Sampling Efficiency Ken Okazakl,+ Russell W. Wiener, and Klaus Willeke" Aerosol Research Laboratory, Department of Envlronmental Health, University of Cincinnati, Cincinnati, Ohio 45267-0056 Accurate aerosol measurements of ambient and in- dustrial air environments are essential for the protection of human health. In this study, the overall aerosol sam- pling efficiency of a tubular thin-walled inlet, sampling isoaxially from environmental airflows, has been deter- mined by a dynamic evaluation technique for various values of particle size, wind and inlet velocities, and inlet size. With the assumption of laminar flow, models de- veloped from the results discriminate the dominant mechanisms that modify the sample properties during sampling and evaluate quantitatively the possible sampling errors for relatively wide ranges of sampling conditions. For Stokes numbers less than 0.03, the sampling efficien- cies are 100 f 20%, while for Stokes numbers larger than about 0.3 particle losses in the inlet become significant. On the basis of a large number of experimental data, the overall sampling efficiency of the inlet has been described accurately by a new nondimensional parameter consisting of Stokes number, gravitational deposition parameter, and Reynolds number. Introduction When aerosols are sampled from ambient or industrial environments, air with particles suspended in it is drawn through an inlet to a filter or direct-reading instrument. It is essential that the sampled aerosol be representative of the aerosol upstream of the sampler; i.e., the aerosol concentration, size distribution, chemistry, and other properties should be unchanged by the sampling process. If changes do occur, they should be known or predicted quantitatively to the greatest extent possible. Particles above a few micrometers in aerodynamic di- ameter may deviate from the air streamlines as they bend toward the inlet face. This results in a change in particle concentration and size distribution of the sampled aerosol. Most of the theoretical and experimental studies have focused on this aspiration process (1-13). In addition to aspiration, particle losses by gravitational deposition and impaction on the inner wall of the inlet may also change the sample properties. Thus, the overall sampling effi- ciency is generally lower than the sampling efficiency based on aspiration alone. Previous studies considering the entire inlet have primarily evaluated specific inlet designs (14-22). The purpose of this study was to develop a com- prehensive and generalized model for evaluating the aer- osol sampling accuracy for a wide range of conditions. Laminar flow conditions are assumed throughout the study. The effect of turbulence is currently under inves- tigation and will be reported on at a later time. Our initial studies (23, 24) dealt with results and in- terpretations for an inlet of a specific size. We have now performed a large amount of systematic experiments utilizing different sizes of tubular thin-walled and sharp- edged inlets in our horizontal wind tunnel. Various factors were found to affect the overall sampling efficiency. The results have been modeled to discriminate the dominant t Present address: Department of Energy Engineering, Toyohashi University of Technology, Tempaku-cho, Toyohashi, 440 Japan. mechanisms that modify the properties of the sample depending on the Stokes number range of the sampled particle. In this paper, we discuss our findings for isoaxial sampling where the inlet is aligned parallel to the wind direction. Our findings for non-isoaxial sampling will be described in a separate paper. Experimental Procedures Our wind tunnel setup (23) has been computerized for fast data acquisition and manipulation of our sampling efficiency data. Test aerosols of oleic acid are generated by a specially designed vibrating-orifice aerosol generator (25,26) and are homogeneously dispersed into the air flow (23). The sampling inlet in the test section of the wind tunnel is integrated into the sensor of an optical single- particle counter that has been further modified so as to accomodate a high flow rate for the larger inlet sizes. The sampling inlet is surrounded by clean sheath air so that all aerosol particles that have passed through the inlet tube are registered by the sensor. All inlets used have been designed to be thin walled and sharp edged so as to meet the inlet design criteria of Belyaev and Levin (6, 7), which require the ratio of the exterior diameter to the interior diameter of the inlet face to be 51.05 and the angle of taper to be 515'. This assures that the rebound and reen- trainment effects at the edge of the inlet face are negligible. The sampling efficiency of the entire inlet system is defined as E, = c*/co (1) and may be characterized by the product of three distinct efficiencies (6, 7) Es = EaEBt (2) where C, is the particle concentration after passage through the inlet and Co is the true particle concentration in the original air environment. The aspiration efficiency, Ea, is the ratio of the particle concentration at the face of the inlet to that in the undisturbed environment. The entry efficiency, E,, is the ratio of the particle concentration passing the inlet face to that incident to the face. The transmission efficiency, E,, is the ratio of the particle concentration exiting from the inlet tube to that just past the inlet face, which accounts for the particle losses to the inside wall of the inlet by gravitational settling, impaction, and turbulent or laminar diffusion, depending on the sampling conditions. Because of the thin-walled and sharp-edged design of our inlet, the entry efficiency, E,, is assumed to be unity, so that E,, = (3) Here, a further subscript is used to show the velocity ratio, R, which is defined as R = U,/Ui (4) where U, and Ui are the wind velocity outside the inlet and the average sampling velocity in the inlet, respectively. In conventional wind tunnel studies the sampling is determined by collecting the sampled aerosol on to a filter 178 Environ. Sci. Technol., Vol. 21, No. 2, 1987 0013-936X/87/0921-0178$01.50/0 0 1987 American Chemical SocietyTable I. Experimental Conditions ' parameter d,, w uw, cmls Vi, cmjs Di(i.d.) cm