Producing Molten Tin Droplets Smaller than the Nozzle Diameter by using a Pneumatic Drop-on-Demand Generator

Transcription

1 ILASS Americas, 2 th Annual Conference on Liquid Atomization and Spray Systems, Chicago, IL, May 27 Producing Molten Tin Droplets Smaller than the Nozzle Diameter by using a Pneumatic Drop-on-Demand Generator A. Amirzadeh Goghari *, S. Chandra Department of Mechanical and Industrial Engineering University of Toronto Toronto, ON M5S 3G8 Canada Abstract A pneumatic droplet generator to produce molten tin droplets smaller than the nozzle diameter is described. The generator consists of a heated aluminum cylinder in which a cavity is machined. A small nozzle is press-fit into a stainless steel nozzle holder attached to the bottom plate of the generator. The system is connected to a gas cylinder through a solenoid valve. Opening the valve for a preset time sends a pressure pulse to the molten metal and ejects a single droplet. Then, gas escapes through the exit vent and no further droplets emerge. To produce these droplets, the required pressure variation was created within the droplet generator by connecting a 12 cm long tube to the exit vent. The effect of various experimental parameters, such as exit vent tube length, ejection rate (frequency), and number of nozzles on droplet formation was investigated. Depending on the length of the vent tube, droplets were generated during the fourth or fifth peak pressure. Droplets were produced from a 12 µm diameter nozzle at different ejection rates. At an ejection rate of.5 Hz, 99% of droplets were smaller than the nozzle diameter and when the frequency was increased to 5 and 1 Hz, more than 9% of droplets were smaller than the nozzle diameter. The results showed that higher ejection rates, not only increases the production capacity, but produces small droplets in a relatively narrow size range (6-1 µm). * Corresponding author

2 ILASS Americas, 2 th Annual Conference on Liquid Atomization and Spray Systems, Chicago, IL, May 27 Introduction Producing metal droplets of desired diameter on demand is of great interest in various industrial and research areas, such as IC manufacturing, inkjet soldering, and rapid prototyping to fabricate complex shaped, three-dimensional parts. One of the most common used methods to produce liquid droplets on demand is the piezoelectric droplet generator [1]. The shape of the voltage pulse used to excite the piezo element, has various effects on the drop ejection process. By varying the pulse width, fluid jets with different diameters are generated. Applying more sophisticated manipulations of drive parameters, including sequential excitation of negative-positive-going pressure pulses, can similarly modulate the diameter of microdrops [1]. By using a piezoelectric inkjet, Sakai [2, 3] applied sequential negative-positive-going pressure pulses to alter the droplet diameter. A negative impulse was initiated to draw the fluid inward into the nozzle. While the fluid was in motion, a properly timed positive pressure was applied causing the central region to be accelerated outward at a higher rate than the liquid in the vicinity of the nozzle wall. By applying the above technique, ink droplets about 2% the nozzle diameter were produced. Chen and Basaran [4] investigated the formation of small water/glycerin droplets using a piezoelectric ejector with a squeeze mode nozzle consisting of a piezoelectric transducer bonded outside of a glass capillary tube with a 7 µm diameter. They applied a sequence of negative-positive- and again negative going voltage pulses. During the first negative voltage, the piezo expanded and liquid was drawn back into the nozzle; during the positive voltage pulse, a tongue emerged from the tip of a primary drop (the about to form droplet) with a high velocity relative to the liquid nearby; and finally, during the second negative pulse, the primary drop was suppressed and the tongue broke up and formed a small droplet. Using this method, water/glycerin droplets with diameters less than 5% the nozzle diameter were produced. It was observed that for intermediate values of glycerin concentration there exist a region near the interior wall of the nozzle where the viscous drag is important, and an inviscid core region near the centerline. Thus, in the vicinity of the centerline, the meniscus protrudes out of the nozzle with a deformation in phase with the pressure pulse and near the wall it moves inward and is out of phase with the pressure (Fig. 1). Ulmke et al. [5] studied the production of single droplets with different diameters using double-distilled water and a mixture with 5 wt% glycerin in a piezoelectric droplet ejector. Droplets were generated from a capillary system filled with liquid and enveloped by a piezoceramic tube. Switzer [6] produced uniform-size liquid droplets by operating a piezoelectric droplet generator at different operating modes. When applying the drop-on-demand-mode, droplets with diameters the same size as the nozzle diameter were produced. Considering the direct relationship between the droplet diameter, piezoelectric voltage, and liquid pressure, the droplet diameter could be varied. Molten metal droplets have also been produced using piezoelectric droplet generators. Orme and Bright [7] conducted experimental studies to produce aluminum droplets for net-form manufacturing., Yim et al. [8] applied a piezo electrically actuated droplet ejector to investigate the effect of oxygen concentration on the break-up behavior of a laminar molten tin jet. And furthermore, Haferl and Paulikakos [9] studied the transport and solidification phenomena in molten microdroplet pile up by employing a piezoelectric dispenser. However, none of these studies investigated the formation of droplets smaller than the nozzle diameter. Cheng et al. [1] and Foutsis [11] used a pneumatic drop-on-demand generator, originally developed by Chandra and Jivraj [12], to produce tin droplets. During his experiments, Cheng [13] observed irregularly sputtering, and small droplets emerged 1-15 ms after a large droplet was ejected. In further investigations, Foutsis [11] produced single and multiple small droplets ranging from µm and µm in diameter, respectively, from a 12 µm diameter nozzle. He also studied the effect of various operating parameters on small droplet formation, including the gas supply pressure, pressure pulse width, and the amount of molten tin in the droplet generator. In this paper a pneumatic droplet generator is described in which rapidly opening and closing a solenoid valve applies gas pressure pulses to the molten metal inside the droplet generator forcing single and multiple droplets out of a 12 µm diameter nozzle. The generator was originally developed and patented by Chandra and Jivraj [12] and unlike the piezoelectric ejector, it is robust and has no moving parts in contact with the liquid, is simple to build, and suitable for high temperature molten metals. Experimental Method The droplet generator consists of a 38.1 mm diameter, 35.5 mm long aluminum cylinder (main body) in which a 19.5 mm diameter cavity is machined. A sapphire nozzle of.12 mm in diameter (Model 12T47, Swiss Jewel Company, Philadelphia, PA) is press-fit into a stainless steel nozzle holder which is attached to the bottom surface of the droplet generator through a stainless steel supporting plate. The top of the droplet generator body is connected to a cross junction with one of its outlets connected to a nitrogen gas supply line (Oxygen free type nitrogen, O 2 less than 5 ppm, BOC, Mississauga, ON). The second outlet is attached to a pressure measurement system to record pressure varia-

3 tions inside the droplet generator. Finally, the last outlet is used as an exit vent to which a 12 cm long stainless steel tube was connected. Figures 2 and 3 show a photograph and a schematic diagram of the experimental apparatus, respectively. Initially, the nozzle holder was filled with tin shot (~3 mm in diameter). The tin shot was heated above their melting point (232 C) using a 275 watts/12 volts band heater (Model HBA-31427, Omega Engineering Inc., Stamford, CT), which was clamped around the droplet generator body. To ensure stable droplet ejection, the molten tin temperature was kept at 32 C using a temperature controller (CN9112A, Omega Engineering Inc., Stamford, CT). Heat loss from the droplet generator was minimized by placing heat insulators between different parts of the generator (Figure 4). The temperature of the generator was measured using a K- type (Chromel-Alumel) thermocouple placed next to the body of the droplet generator near the nozzle holder. To prevent oxidation effects, droplets were produced in a 15mm long Pyrex glass tube with an inner diameter of 44mm in which the oxygen content was kept below 15 ppm. The glass tube was clamped between two supporting plates 1 and 2 as shown in figure 5. The oxygen content inside the glass tube was monitored using an oxygen analyzer (Model 911, Illinois Instruments, Illinois) and could be measured over a range of.1 ppm to 3% with an accuracy of ±2% of the reading. The pressure inside the test chamber must be kept at atmospheric pressure; otherwise, it affected the droplet formation. For pressures much higher than atmospheric, either no droplets emerged, or a large pressure pulse was required to eject a droplet. If the pressure was much lower than atmospheric, uncontrollable liquid flow would occur. Therefore, a water manometer was connected to the chamber to precisely measure the chamber pressure during experiments (Figure 3). To photograph droplets emerging from the nozzle, a CCD camera, Sensi Cam High Speed (Type 37 KF, OPTIKON Corporation, Kitchener, ON), was used. The CCD chip was capable of recording 3 frames per second with a resolution of 128x124 pixels and an exposure time as low as 1 ns. Controlling software, CamWare, was also used to adjust the camera s parameters and start the system. The camera was equipped with a Tamron lens (Tamron AF SP 9/2.8 Macro, Japan) with a minimum object distance of.29 m. The photographing system also included a light source, a 15 watts/12 volts light bulb, placed close to the droplet generator. A Lab VIEW Control System controlled timing of signals for the camera, solenoid valve, and pressure transducer. To monitor the rapidly changing pressure inside the droplet generator, the data acquisition was chosen to be 5 samples/sec for a duration of 5 ms for each pressure pulse. Hence, 25 samples were obtained from the pressure transducer in each pulse. To produce droplets, the nozzle holder was initially filled with tin shot. Before starting the process, the oxygen content in the test chamber was reduced and then the tin shot was heated above its melting point. To decrease the oxygen concentration in the chamber, the glass tube was filled with nitrogen gas to a pressure of 5 kpa and then the gas was released through a nonreturn valve to the environment. By repeating this process for several times, the oxygen content was reduced to 3 ppm. After the pressure inside the chamber was adjusted and monitored using a non-return valve and water manometer, respectively, the heater was switched on until the tin temperature reached 32 C. When the solenoid valve was opened for a pre-set time (pressure pulse width), the gas flow into the droplet generator increased the pressure inside the generator and forced liquid out of the nozzle. Once the liquid jet emerged from the nozzle, it broke up into droplets due to fluid instability. By adjusting the pulse width, the opening time of the solenoid valve and hence, the pressure build-up duration in the droplet generator could be varied. Results and Discussion Molten Tin Droplets According to the required pressure variation described by Chen and Basaran [4], to produce molten tin droplets smaller than the nozzle diameter, a similar pressure variation was produced within the droplet generator. For this purpose, different tube lengths were used at the exit vent. In all experiments the supply pressure and the ejection rate were adjusted to 276 kpa and.5 Hz, respectively. At lower supply pressures, it was difficult to maintain stable and repeatable droplet ejection. Experiments showed that 12 and 24.5 cm long tubes produced the desired pressure variation (Figure 6). However, small droplets started emerging during the first peak pressure, which was contrary to our predictions. Because of poor repeatability of droplet ejection under this condition, a closed needle valve was connected to the other end of the 12 cm long vent tube which increased the peak pressures (Figure 7). In the new configuration, small droplets were ejected over the first five peak pressures. To obtain single droplets during one peak pressure only, the pressure pulse was reduced to 4.1 ms and single droplets were produced during the fifth peak pressure. Figures 8 and 9 show a single droplet emerging from the nozzle and the corresponding pressure variation within the droplet generator, respectively. As can be observed, there is no negative pressure and only positive pressure fluctuations exist; i.e., small droplets can be generated even without the presence of any

4 negative pressure. By increasing the pulse width to 4.5 ms, multiple small droplets were produced. These droplets and the corresponding pressure variation are shown in figures 1 and 11, respectively. In this case, the formation of multiple small droplets occurred during the fourth peak pressure. Effect of the Exit Vent Tube Length on Droplet Formation In addition to the 12 cm long tube, a 24.5 cm long tube was also used. Similar to the previous case, the droplet ejection process was stable and repeatable, and a comparable pressure variation was developed within the droplet generator. By keeping all the operating parameters (supply pressure, pulse width, and nozzle diameter) the same, single droplets were obtained during the fourth peak pressure. Figure 12 compares the pressure variations for two different tube lengths. Size of Droplets Droplet Generation at.5 Hz Ejection Rate- Figure 13 shows photographs of single droplets when the pulse width was set to 4.1 ms and a closed needle valve, connected to a 12 cm long tube was used at the exit vent. As can be observed, droplets are spherical and almost uniform-size. The corresponding size distribution is shown in figure 14. To produce multiple droplets, the pulse width was increased to 4.5 ms. These droplets are not as uniform as single droplets (Figure 15). The corresponding size distribution is demonstrated in figure 16. To produce droplets at higher capacities, instead of a single nozzle, four nozzles were press-fit into the nozzle holder next to each other in a row. The objective was to generate single droplets from each nozzle using the same experimental settings. However, because the nozzles were not exactly the same shape and size, droplets produced a much larger variation in size. When the pulse width was set to 4.9 ms, some of the nozzles produced one or two droplets, while others generated multiple droplets. Therefore, the uniformity of the produced droplets was even less than the case where multiple droplets were generated from one nozzle. These droplets and the corresponding size distribution are shown in figures 17 and 18, respectively. Droplet Generation at 5 and 1 Hz Ejection Rates- Another method used for mass production of tin droplets was increasing the droplet ejection rate (frequency) which required some modifications in the system. As mentioned before, the solenoid valve used to create pressure pulses was triggered by the Lab VIEW system. Although this system worked well at low rates (.5 Hz), it did not operate properly at higher frequencies. When the ejection rate was increased, the time delay increased as well. Therefore, instead of the Lab VIEW system, a function generator was used to activate the solenoid valve. To obtain stable droplet generation at a frequency of 5 Hz, the needle valve was kept fully open. Then, the pulse width was gradually increased until small droplets emerged from the nozzle. Multiple droplets were produced at a pulse width of 6 ms (Figure 19). The corresponding size distribution is demonstrated in figure 2. When the frequency was set to 1 Hz and the needle valve was kept fully open, multiple small droplets were ejected at a pulse width of 4.9 ms. These droplets and the size distribution are shown in figures 21 and 22, respectively. Table 1 demonstrates the percentage of droplet size distribution for different diameter ranges in each case. As can be observed, for a frequency of.5 Hz, when single droplets are produced, 99.5% of the droplets range from µm in diameter, and for multiple droplets, only 53% of the droplets lay in this range. For the case where four nozzles were used, the diameter of multiple droplets is almost evenly distributed from 8 to µm, which is 82% of the droplets. At an ejection rate of 5 Hz, 96.6% of the droplets range from µm and at a 1 Hz frequency, 93.7% of droplets lay in the range of µm. The mean diameter and standard deviation for each case is presented in Table 2. Summary and Conclusions This paper presented an experimental method to produce small molten tin droplets by using a pneumatic drop-on-demand generator. The required pressure variation was produced within the droplet generator using 12 and 24.5 cm long vent tubes. Interestingly, it was observed that droplets smaller than the nozzle diameter could be produced without the presence of any negative pressure and the important factor is the pressure oscillation. The results showed that single droplets emerged during the fourth and fifth peak pressures for 24.5 and 12 cm long tubes, respectively. Droplets were obtained using single and multiple nozzles (12 µm in diameter) at different ejection rates. For single droplets produced at a low rate (.5 Hz) from one nozzle, more than 99% of the droplets were ranging from µm in diameter. To produce droplets at higher capacities, four nozzles were used. This resulted in droplets with diameters evenly distributed from µm, but only 33% of droplets had diameters less than 1 µm. Droplet production at higher operation frequencies (5 and 1 Hz) using a single nozzle was also tested. When the ejection rate was set to 5 Hz, 9.5% of droplets ranged from µm and at a frequency of 1 Hz, more than 93% of droplets were in the range of µm in diameter. The results showed that using higher ejection rates, not only increases the production capacity, but produces small droplets in a comparably narrow size range. References 1. Lee, E.R., Microdrop Generation, Stanford Linear Accelerator Center, Stanford University, Sakai, S., U.S. Patent No. 5,933,168 (1999).

5 3. Sakai, S., IS & T s NIP 16, International Conference on Digital Printing Technologies, IS & T: The Society for Imaging Science and Technology, 2, pp Chen, A.U. and Basaran, O.A., Physics of Fluids, 14:1, L1-L4 (22). 5. Ulmke, H., Mietschke, M. and Bauckhage, K., Chem. Eng. Technol. 24, pp (21). 6. Switzer, G.L, Rev. Sci. Instrum. 62(11), pp (1991). 7. M. Orme and A. Bright, Recent advances in highly controlled molten metal droplet formation from capillary stream break-up with applications to advanced manufacturing, TMS Annual Meeting, P. Yim, J. Chun, N. Saka and J.C. Rocha, The Minerals, Metals & Materials Society, pp , S. Harfel and D. Paulikakos, J. Applied Physics, 92(3), Cheng, S.X., Li, T. and Chandra, S., J. Materials Processing Technology, 159(3): (25). 11. Foutsis, V., Producing small molten metal drops with a pneumatic generator, University of Toronto, M.A.Sc. Thesis, Chandra, S. and Jivraj, R., U.S. Patent No. 6,446,878(22). 13. Cheng, S.X., Development of a molten metal droplet generator for rapid prototyping, University of Toronto, M.A.Sc. Thesis (22).

6 ILASS Americas, 2 th Annual Conference on Liquid Atomization and Spray Systems, Chicago, IL, May 27 Figure 1. Formation of a small droplet Figure 2. Photograph of the experimental system Figure 3. Schematic diagram of the experimental apparatus

7 8 Pressure (kpa) Time (ms) Figure 4. Insulation in the droplet generator Figure 7. Pressure variation inside the droplet generator when a closed needle valve was attached to a 12 cm long vent tube, Nozzle diameter: 12 µm, Supply pressure: 276 kpa, Pulse width: 4.1 ms Figure 8. Single droplet emerging from a 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.1 ms, Exit vent: closed needle valve connected to a 12 cm long tube Figure 5. Test chamber assembly of the generator 12 Pressure (kpa) pulse width droplet starts emerging Pressure (kpa) pulse width 9 droplet starts emerging Time (ms) Time (ms) Figure 6. Single droplet emerging from a 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.9 ms, Vent tube length: 12 cm Figure 9. Pressure variation within the droplet generator when a single tin droplet is ejected from a 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.1 ms, Exit vent: closed needle valve connected to a 12 cm long tube

8 Pressure (kpa) 15 pulse width 12 multiple tin droplets start emerging Time (ms) Figure 1. Multiple droplets emerging from a 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.5 ms, Exit vent: closed needle valve connected to a 12 cm long tube Figure 11. Pressure variation inside the droplet generator when multiple droplets are ejected from a 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.5 ms, Exit vent: closed needle valve connected to a 12 cm long tube Pressure (kpa) fourth peak presssure fifth peak pressure Time (ms) 12 cm long tube 24.5 cm long tube Figure 12. Pressure variations inside the droplet generator for different tube lengths, Nozzle diameter: 12 µm, Supply pressure: 276 kpa, Pulse width: 4.1 ms Figure 13. Single droplets produced from a 12 m diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.1 ms, Exit vent: closed needle valve connected to a 12 cm long tube

9 Number of Droplets Droplet diameter (microns) Figure 14. Droplet size distribution of single droplets produced from 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.1 ms, Exit vent: closed needle valve connected to a 12 cm long tube Figure 15. Multiple droplets produced from a 12 m diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.5 ms, Exit vent: closed needle valve connected to a 12 cm long tube Number of droplets Droplet diameter (microns) Figure 16. Droplet size distribution of multiple droplets produced from a 12 µm diameter nozzle, Supply pressure: 276 kpa, Pulse width: 4.5 ms, Exit vent: closed needle valve connected to a 12 cm long tube

10 Figure 17. Multiple droplets produced from four 12 m diameter nozzles, Supply pressure: 276 kpa, Pulse width: 4.9 ms, Exit vent: closed needle valve connected to a 12 cm long tube 6 5 Number of droplets Droplet diameter (microns) Figure 18. Droplet size distribution of multiple droplets produced from four 12 µm diameter nozzles, Supply pressure: 276 kpa, Pulse width: 4.9 ms, Exit vent: closed needle valve connected to a 12 cm long tube Figure 19. Multiple droplets produced from a 12 m diameter nozzle at a 5 Hz ejection rate, Supply pres re: 276 su kpa, Pulse width: 6 ms, Exit vent: fully-open needle valve connected to a 12 cm long tube

11 4 Number of droplets Droplet diameter (microns) Figure 2. Droplet size distribution of multiple droplets produced from a 12 m diameter nozzle at a 5 Hz ejection rate, Supply pressure: 276 kpa, Pulse width: 6 ms, Exit vent: fully-open needle valve connected to a 12 cm long tube Figure 21. Multiple droplets produced from a 12 m diameter nozzle at a 1 Hz ejection rate, Supply pressure: 276 kpa, Pulse width: 4.9 ms, Exit vent: fully-open needle valve connected to a 12 cm long tube 35 ets Number of dropl Droplet diameter (microns) Figure 22. Droplet size distribution of multiple droplets produced from a 12 m nozzle at a 1 Hz ejection rate, Supply pressure: 276 kpa, Pulse width: 4.9 ms, Exit vent: fully-open needle valve connected to a 12 cm long tube

12 Table 1. Percentage of droplet size distribution Case Diameter Range (µm) % Single droplets (one nozzle) Rate=.5 Hz % Multiple droplets (one nozzle) Rate=.5 Hz % Multiple droplets (four nozzles) Rate=.5 Hz % Multiple droplets (one nozzle) Rate=5 Hz % Multiple droplets (one nozzle) Rate=1 Hz Table 2. Statistics on the diameter of tin droplets Case Ejection Rate (Hz) Mean Diameter (µm) Standard Deviation (µm) Single droplets (one nozzle) Multiple droplets (one nozzle) Multiple droplets (four nozzles) Multiple droplets (one nozzle) Multiple droplets (one nozzle)