Behavior of liquid metal droplets in an aspirating nozzle

Measurements of particle size, velocity, and relative mass flux were made on a spray field produced by aspirating liquid tin into 350 °C argon flowing through a venturi nozzlevia a small orifice in the throat of the nozzle. Details of the aspiration and droplet formation process were observed through windows in the nozzle. The spatial distribution of droplet size, velocity, and relative number density was measured at a location 10 mm from the nozzle exit. Due to the presence of separated flow in the nozzle, changes in nozzle inlet pressure did not significantly effect resulting droplet size and velocity. This suggests that good aerodynamic nozzle design is required if spray characteristics are to be controlled by nozzle flow.

The driving force in all foul" modes is the velocity differential between the gas and the droplet. to 5 /_sec, produced sharp images of the droplets.
The system developed for simultaneous measurement of particle size and velocity integrates a crossed beam laser Doppler velocimeter (LDV) with a scattered light particle size measurement. A schematic of the measurement system optics appears in Figure 2. The beam launching optics consist of two Brewster angle dispersing prisms that separate the individual colors irl a multi-line 6 W argon ion laser beam. The blue 488 nm beam is routed, via a 1500 mmfocal length lens and aperture/beam stop, to the measurement volume by a series of mirrors.
The green 514 nrn beam is routed to the LDV optics. These optics consist of a polarization rotator, a beam splitter, a beam expander and a 600 mmfocal length focusing lens.
The system is aligned such that the L.DVmeasurement volume (0.173 x 4.2 mm) is located in the center of the large (2 mm) blue sizing beam.
The effective length of the LDV measurement volume is further shortened to approximately 0.75 mmby an aperture, PI, in the receiving optics. Localization of the LDV measu_,'ement volume in the center of the sizing beam and the requirement of coincidence between a particle size signal and a Doppler burst avoids the trajectory ambiguity problem of large particles clipping the periphery of the sizing beam and producing a signal representative of a smaller particle.
The receiving optics are also shown schematically in Figure 2.
Lens LI to an estimated one standard deviation in the measured diameter of 4.9 A_m.
In calibrating the laser Doppler velocimeter only the laser wavelength and the angle between the beams are required to establish the relationship between the velocity of the scattering particle and the frequency output of the photodetector. The accuracy of the counter processor is checked against a standard function generator. The estimated velocity uncertainty is less than 5%.

RESULTS
FLOWVISUALIZATION -Schlieren photography produced evidence of strong density gradients, i.e. shock waves, in the throat and divergent sections of the nozzle.
When the upstream pressure in the nozzle is increased to approximately twice atmospheric pressure, the velocity in the throat is sonic and a standing normal shock wave is formed at the exit of the throat. As the upstream pressure is  Although the velocity of the gas does not change the density of the gas does increaseby 34% over this range of inlet pressures. Since the aerodynamicforce on the droplets is proportionalto p_V2, this increase in density of the gas may explainthe variationin averagemeasuredparticle size towardssmallerparticles at higher nozzle inlet pressures shown in Figure 6. The data points shown in Figure 6 are an arithmeticmean of all the particle size measurementsmade at the 10 mm stand off distance with a tin melt temperatureof 600°C.

I0
The present Weber number of 9 is in good agreement with the data compiled by Pilch and Erdman [5] of critical Weber number data as a function of Ohnesorge number from a variety of authors. Figure 7 shows the vertical distribution of average particle size at a stand off distance of 10 mm from the nozzle exit. Again the nozzle inlet pressure was 168 kPa with the argon at a temperature of 350°C and the tin melt temperatLre at 600°C. The plot shows that the smaller particles tend to migrate to the outer edges of the flow and that the larger particles stay on the center line.
Thus, for a substrate passed vertically in front of this nozzle, the larger particles would be preferentially deposited in the middle of the sprayed layer.
The maximum particle mass flux also occurs near the centerline of the nozzle, a,." shown in Figure 8. This data was obtained for the same conditions as