Influence of probe nozzle diameter on aerosol particle size distribution for isokinetic sampling in gas streams

Isokinetic sampling from stacks and ducts is a widely used method for air pollution control in many industrial and environmental applications.
During the last 40 years, U.S. EPA developed reference methods for air pollutant emission sampling of stationary sources, which included a compendium of technical guidelines for the monitoring of aerosol particles. One of the main parts of the sampling system is the sampling probe nozzle for which it was suggested in the reference methods that the lower bound diameter should be greater than di = 6 mm (Cohen, 2001). The influences of sampling nozzle tube diameter, the velocity ratio between free stream and the sampling tube as well as the tube inclination have been considered widely (Vincent, 2007). On the other hand, the influence of tube diameter has not been quantified by means of aerosol particle size spectrometry yet.
McFarland et al. (1997) as well as Peters and Leith (2004) experimentally investigated the deposition losses of aerosol particles in pipe bends. Both found a strong increase of deposition losses in the sampling tube for particle Stokes number of Stk > 0.3.
In this work, the influence of the nozzle diameter of the sampling probe on the aerosol particle size distribution was investigated by means of TSI APS 3321 and TOPAS LAP 321 measurements.
The considered gas/aerosol test facility is a small scale wind tunnel, which consists of a HEPA at the inlet and a square duct section in diameter of dhyd = 10 cm and length of l = 20dhyd = 2 m. The test section, where sampling takes place, is followed by a diffuser stage, an electrical precipitator and a radial fan in order to generate the desired pressure head. The average air velocity can be regulated over the range u = 1..7 m/s, corresponding to Reynolds numbers between Red = 8.9k..42k. In order to achieve isokinetic sampling, the flow rate of the sampling device has been accurately adjusted using Particle Image Velocimetry data.
AC Fine Test Dust (ISO 12103-1, dnom = 0.1..80 µm, bimodal distribution) was injected 17 diameter lengths upstream of the sampling point by means of a dust disperser TOPAS SAG 410. Four probe nozzle diameters di = [3,5,7,9] mm were chosen.
Figure 1 displays the volume weighted aerodynamic particle size distribution of AC Fine under variation of nozzle tube diameter as measured by the APS 3321. The modal dP,aero = 5 µm is roughly independent of nozzle size, whereas the values of the distribution for particle diameters above dP,aero > 12 µm increase continuously with increasing nozzle diameter.
A similar upward tendency in size distribution with increasing nozzle diameter is observed in the LAP measurement results (not shown here). This leads to the conclusion that deposition losses due to inertia, which consequently affect the larger particles, can be considerably reduced by accordingly increasing tube nozzle diameter. In addition, a comparison of aerodynamic particle size (APS 3321) and scattering particle size (LAP 321) will be presented.