Liquid dibutyl phthalate (DBP) particles were generated by homogeneous nucleation as a heated gas stream laden with DBP vapour exited a nozzle and mixed with room temperature particle-free air. The nozzle diameters were 0.235 and 0.375 cm, the nozzle temperature was 413 K, and DBP vapour concentrations were 100-500 ppm. Conditions were chosen to make the jet turbulent (Reynolds number 4000-7000) and to keep nucleation rates low so that the particle growth could be studied in the absence of coagulation and in the early stages of the transition to coagulation. Aerosol size distributions and number concentrations were measured at different positions and jet velocities. For low values of DBP vapour mole fraction, the size distributions were unimodal and the aerosol concentration followed jet dilution profiles, indicating that the particle formation was confined to the initial region of the jet. The final aerosol concentrations were correlated by simple scaling laws. For higher vapour mole fractions, nucleation continued to occur downstream from the shear layer, resulting in bimodal size distributions. Data for nozzles of different diameters indicate that the overall particle formation rate can be decreased by splitting a large nozzle flow into multiple smaller streams.

The effect of turbulent mixing on the homogeneous nucleation rate was studied for a jet stream at 400 K injected into dry air a t 298 K and 1 atm. To simplify the calculations, it was assumed that the Lewis number was everywhere unity. Probability density functions for the temperature fluctuations in an axisymmetric shear layer were used to calculate the local particle formation rate, averaged over the fluctuations, and the resulting particle concentration profile. Taking turbulent fluctuations into account significantly reduces local maxima in the average nucleation rate and broadens the region of particle formation in the shear layer compared with calculations based on the mean saturation ratio profile. However, the overall rate of particle formation in the mixing layer, taking fluctuations into account, is of the same order of magnitude as the rates based on the mean concentration and temperature for the conditions of the calculation.

Experiments were conducted to test the theory of nucleation-controlled growth of spherical particles in a turbulent jet. Dibutyl phthalate (DBP) particles were formed in a bench-scale jet apparatus with nozzle diameters 0.235 cm and 0.375 cm. Size distributions and number concentrations of the particles were measured at different DBP vapor concentrations, jet velocities, and positions.There is evidence that the DBP particle nucleation was confined to the shear layer of the jet and that the particles grew by condensation as they moved away ffom the nozzle. Trends in the data suggest that for low rates of particle formation in jets, the final aerosol concentration can be predicted fiom simple scaling laws. IntroductionHomogeneous nucleation is a temperature-and concentration-dependent process that frequently occurs in turbulent flow. Examples of particle formation by nucleation are emissions from industrial, vehicular and aircraft sources, and process gas mixing. The complexity of turbulent flow fields has made it necessary to make simplifying assumptions, often oversimplifications, in predicting nucleation rates. Turbulent fluctuations have frequently been neglected and calculations based on mean velocity, temperature, and concentration profiles. As methods have been developed to incorporate turbulent fluctuations into models, the effects of these fluctuations have been found to be significant.A condensable vapor at high temperature may become supersaturated as it undergoes turbulent mixing with a gas at lower temperature. When the vapor attains only moderate saturation ratios such that rates of nucleation are in the range lo4 -10' particles/cm3 s, the resulting particle number concentrations are low. Particle-particle collisions are infrequent; growth is by heterogeneous condensation and the particle concentration changes only by dilution. In such cases, the growth is said to be nucleation-controlled because the final particle size and concentration are determined by the rates and duration of homogeneous nucleation in the initial mixing stage.This study focuses on a specific flow system of great practical interest, the turbulent jet. As shown in Figure 1, the flow pattern of a turbulent jet has two distinct regions, the initial and similarity regions. The initial region consists of the potential core, a cone of undisturbed nozzle fluid, surrounded by the shear layer. For aerosol formation studies, the shear layer is very important since it is where incipient turbulent mixing, supersaturation, and homogeneous nucleation occur. Hidy and Friedlander (1964) showed qualitatively that processes occurring in the shear layer of a jet strongly affect particle formation by homogeneous nucleation. Lesniewski and Friedlander (1995) hypothesized that particle formation occurs in the shear layer of the jet, but is quenched as the particles move down the axis. Quenching may occur by three mechanisms: (1) the condensable vapor is diluted by ambient fluid entrained into the jet; (2) the number concentration of newly formed partic...

No abstract