Molten Aluminum Micro-Droplet Formation and Deposition for Advanced Manufacturing Applications

Accepted for publication in Aluminum Transactions Journal, 2000 Molten Aluminum Micro-Droplet Formation and Deposition for Advanced Manufacturing App...
Author: Sara Chambers
2 downloads 2 Views 574KB Size
Accepted for publication in Aluminum Transactions Journal, 2000

Molten Aluminum Micro-Droplet Formation and Deposition for Advanced Manufacturing Applications Melissa Orme*, Qingbin Liu and Robert Smith Department of Mechanical and Aerospace Engineering, University of California - Irvine Irvine, CA 92697-3975, USA

Abstract Molten aluminum droplets are generated from capillary stream break-up and are injected into an inert environment. The intrinsic fidelity of droplet formation from capillary stream break-up provides the allure for use in advanced manufacturing applications such as net-form manufacturing and electronics packaging for the two following reasons: first, because the droplets are generated with extremely regular diameters and inter-droplet spacings, and second, because of the high rates at which they are generated. Additionally, droplet formation from capillary stream break-up allows the customization of droplet streams for a particular application. The current status of the technology under development is presented, and issues affecting the microstructure and the mechanical properties of the manufactured components are studied in an effort to establish a relationship between processing parameters and properties.

Keywords: Net-Form Manufacturing, Solid Freeform Fabrication, Capillary Stream Break-Up, Droplet Deposition, Droplet Based Precision Manufacturing

1. Introduction Use of precisely controlled molten aluminum droplets for advanced manufacturing applications is gaining considerable academic and industrial interest due to the promise of improved component quality resulting from rapid solidification processing and the economic benefits associated with fabricating a structural component in one integrated operation.

The

manufacturing industrial sector seeks to gain a competitive advantage by developing new methods of droplet-based net-form manufacturing that allow the flexible data driven manufacture

Accepted for publication in Aluminum Transactions Journal, 2000

of complicated three-dimensional components in one integrated step with minimal posttreatment. University

Such a droplet based net-form manufacturing technique is under development at of

California-Irvine

manufacturing (PDM).

(UCI)

that

is

termed

precision

droplet-based

net-form

The crux of the technique lies in the ability to generate highly uniform

streams of molten metal droplets such as aluminum or aluminum alloys.

Though virtually any

Newtonian fluid that can be contained in a crucible is suitable for the technology, this work concentrates on the generation and deposition of molten aluminum alloy (AA2024) droplets that are generated and deposited in an inert environment.

Shown in Figure 1 is a conceptual schematic of the PDM process. Droplets are generated from capillary stream break-up in an inert environment and are deposited onto a substrate whose motion is controlled by a programmable x-y table. In this way, tubes with circular, square and triangular cross sections have been fabricated such as those illustrated in Figure 2. Tubes have been fabricated with heights as great as 11.0 cm. The surface morphology of the component is governed by the thermal conditions at the substrate. If we denote the solidified component and the substrate the "effective substrate," then the newly arriving droplets must have sufficient thermal energy to locally remelt a thin layer (with dimensions on the order of 10 microns or less) of the effective substrate.

Remelting action of the previously deposited and solidified material

will insure the removal of individual splat boundaries and result in a more homogeneous component.

The thermal requirements for remelting have been studied analytically in reference

[1]. It was shown in that work that there exists a minimum substrate temperature for a given droplet impingement temperature that results in remelting.

The "bumpiness" apparent in the

circular cylinder shown in Figure 2 is because the initial substrate temperature was insufficient to initiate the onset of remelting. In particular to the circular tube, it can be seen that the surface morphology is not uniform.

This is because the temperature of the substrate was not uniform,

but was lower at the positions where the component is rougher. Deposition on a low temperature surface caused the incoming droplets to land on a previously solidified and curved top surface, with insufficient thermal energy for the action of remelting. Hence, the droplets tend to roll from the top of the curved surface and roll from one side or the other prior to solidification, as is evident in the photograph.

Generally, as the component grows in height by successive droplet

deliveries, the effective substrate temperature increases due to the fact that droplets are delivered

Accepted for publication in Aluminum Transactions Journal, 2000

at high rates so that the cooling time between successive droplet deliveries is insufficient to allow cooling before the arrival of the next layer of droplets. Therefore, within the constraints of the current embodiment of the technology, there exists a certain height of the component for which remelting will occur. This height is demarcated at the location where the "bumpiness" is eliminated and relative "smoothness" prevails, as can be seen in the circular cylinder.

As the

component grows beyond this height, the remelting depth will continue to increase due to increased heating to the effective substrate.

Hence, the component walls will thicken due to

slower solidification rates. There is an ongoing work at UCI to identify the heat flux required for the minimum remelting of the effective substrate and to develop processing conditions for which this heat flux seen by the substrate remains constant for each geometry desired. In this manner, the fidelity of the microstructure, mechanical properties and geometry will remain intact.

At this stage in the development of the PDM technology, electrostatic charging and deflection is not employed.

However, in the final realization of the technology, charging and deflection will

be utilized in order to control the droplet density as a function of the component geometry, and to print fine details at high speed and at high precision.

The charging and deflection of droplets

bears many similarities to the technology of ink-jet printing, except that in the current application of PDM, large lateral areas are printed, thereby requiring significantly higher droplet charges than in ink-jet printing. The high charges on the closely spaced droplets result in mutual interdroplet interactions that are not apparent in the application of ink-jet printing.

Recent

experimental and numerical results on the subject of droplet interactions due the application of high electrostatic charges are presented elsewhere [2]. Droplet charging and deflection has been successfully applied to the "printing" of electronic components such as ball grid arrays (BGA's) [3] with a high degree of success. Hence, we aim in future work to bridge the technology of fine pitch printing of electronic components with that of net-form manufacturing.

A vast body of research on controlled droplet formation from capillary stream break-up over the past decade forms the basis of the PDM technology and has enabled ultra-precise charged droplet formation, deflection and deposition that makes feasible many emerging applications in net-form manufacturing and electronic component fabrication [4-9].

Unlike the Drop-on-Demand mode

of droplet formation, droplets can be generated at rates typically on the order of 10,000 to 20,000

Accepted for publication in Aluminum Transactions Journal, 2000

droplets per second, from capillary stream break-up and can be electrostatically charged and deflected onto a substrate with a measured accuracy of ±12.5 µm.

The technologies of 3D printing (3DP) [10-12] and shape deposition manufacturing (SDM) [1315] bear similarities to PDM in as much as they rely on the deposition of precisely controlled molten droplets. In 3DP, parts are manufactured by generating droplets of a binder material with the Drop-on-Demand mode of generation and depositing them onto selected areas of a layer of metal or ceramic powder. After the binder dries, the print bed is lowered and another layer of powder is spread in order to repeat the process. The process is repeated until the 3-D component is fabricated.

Similarly, the process of SDM relies on uniform generation of molten metal droplets. However, the droplet generation technique is markedly different than droplet generation from capillary stream formation, which is the method employed in this work. In SDM, the tip of a feedstock wire is melted with the aid of a plasma-welding torch to form droplets that are typically on the order of 1-10 mm in diameter.

The droplets impinge upon a substrate where they fuse with

previously deposited and solidified material.

With SDM, high quality metallic, ceramic and

polymer components have manufactured.

2. Experimental Procedures Figure 3 illustrates a conceptual schematic of the molten aluminum droplet generation and deflection apparatus. The main component is a graphite lined titanium cartridge that contains the molten aluminum or aluminum alloy.

The alumina orifice is contained at the lower end of the

cartridge and a vibrating plunger rod is inserted through an adapting plate at the upper end and is extended through the entire length of the cartridge. piezoelectric crystal from extreme temperatures.

An active cooling flange shields the

The molten aluminum is pressurized with a

driving pressure that is typically on the order of 20 psi. The pressure perturbation established by the vibrating rod is on the order of 5% of the driving pressure.

The pressure perturbation

translates to a radial perturbation on the capillary stream, which will grow until droplets are pinched off from the stream. The droplets are injected into an inert environment with an oxygen content of less than 25 ppm.

Accepted for publication in Aluminum Transactions Journal, 2000

3. Droplet Generation and Manipulation This paper focuses on the generation and deposition of highly uniform molten aluminum droplets generated from capillary stream break-up and their application to net-form manufacturing. For purposes of brevity, the basic physics of conventional droplet generation from capillary stream break-up is omitted and the reader is referred to the informative review papers [16, 17], or other numerous references on the subject in the literature base.

Our experiments have revealed the critical importance of maintaining an environment virtually devoid of oxygen. It was found that the aluminum capillary stream is very efficient in capturing oxygen.

Small trace amounts of oxygen in the jetting environment cause the droplet stream to

lose its angular stability. Larger amounts of oxygen prohibit entirely the capillary disturbance to grow on the stream, eliminating the possibility of generating droplets.

Additionally, it has been found in this and in previous work [9] that oxides within the melt are the leading cause of angular instabilities of the droplet stream. Hence, meticulous attention must be paid to filtering the molten aluminum and insuring that it is maintained in an inert environment at all times.

With the above precautions, we have been successful in generating uniform and angularly stable streams of droplets from capillary stream break-up. Care must also be taken with respect to the selection of the characteristics of the forcing disturbance in order to achieve a stream of drops with highly uniform diameters and inter-droplet spacings.

The nondimensional wavenumber

( k o* ) is defined as the ratio of the stream circumference to the wavelength of the applied disturbance and is likely the most important parameter in capillary stream break-up. It is well known that the capillary wave growth on an inviscid jet subject to conventional forcing grows the fastest for k o* = 0.697. It has also been shown previously [5] that droplet generation at the wavenumber corresponding to fastest wave growth results in the most uniform droplet stream and generation at wavenumbers far from that of maximum growth lead to droplet streams with highly irregular characteristics.

Therefore, droplet generation with conventional forcing

(sinusoidal disturbance resulting in a droplet stream such as that shown in Figure 3) is restricted

Accepted for publication in Aluminum Transactions Journal, 2000

to a small range of wavenumbers that will result in uniform droplets.

Use of amplitude-modulated (a-m) disturbances, however, will bypass the limitation imposed by conventional forcing. Figure 4 illustrates the ability to manipulate and control the droplet stream characteristics through variations in the applied force.

The images were obtained by

backlighting the droplet stream and capturing the droplet configuration on videotape.

Individual

frames were processed and reproduced as shown in Figure 4. In Figure 4a the droplets were generated with an a-m disturbance with a frequency ratio of N = fc / fm = 3, where the carrier frequency (f c) is selected to be in the region of Rayleigh growth, and is 12,000 Hz. Figure 4b illustrates the a-m disturbance that is applied to the piezoelectric crystal that initiated the stream's radial instability.

Examination of Figure 4 and other similar experimental realizations reveal that

when an a-m disturbance with an integer frequency ratio (N) is employed for droplet generation, groups of N droplets systematically merge in flight to form the final stream of droplets.

The

droplet configurations and the time required to achieve the droplet configurations are predictable with knowledge of the forcing disturbance characteristics.

Excellent comparison between

experiment and simulation are provided elsewhere [4, 6].

The final droplet configuration

consists of droplets that are separated by a distance of Nλ and have diameters equal to N1/3 times the diameter of the original carrier droplet, where 2 is the wavelength of the applied carrier disturbance, i.e., the wavelength of the disturbance employed in conventional Rayleigh mode droplet formation. The droplets shown form a highly-regularly spaced and sized droplet stream. In fact, the speed dispersion, which is a measure of the uniformity of the droplet stream, decreases as 1/N, so that the a-m droplet stream is more uniform than the conventionally generated stream. Additionally, the effective wavenumber of the a-m generated stream is k o* /N, so that the droplet streams shown in Figure 4 has an effective wavenumber of 0.232. Generation of these droplet streams with a conventional disturbance would be impossible due to the severe degradation in uniformity that occurs at generation at such low wavenumbers.

In addition to generating highly uniform streams of droplets at extended wavenumbers, droplet streams can be customized for a given application, where the droplet sizes and separations can be selected a priori.

Figure 5 illustrates one example of the evolution of a droplet pattern that

consists of repeated sequences of large-small-large droplets. This droplet stream was generated

Accepted for publication in Aluminum Transactions Journal, 2000

with a frequency ratio N = 3.5.

The microstructural characteristics of several tubes similar to those shown in Figure 2 that were fabricated with AA2024 aluminum alloy droplets were examined.

The processing parameters

are included below in Table I. In the table, To is the initial capillary stream temperature, Ts is the initial substrate temperature and m/l is the mass delivered per unit length to the substrate, which is directly related to the droplet speed and substrate speed. We chose not to characterize the results in terms of the mass flow rate

(m& )

since it contains no information regarding the amount

of material overlap on the substrate (a function of substrate speed), which is of critical importance for the geometric and mechanical quality of the component.

The cylinders were

sectioned, and the sections were polished and subsequently photographed with an optical microscope.

The samples were etched with a modified version of Keller's Reagent, which

consists of 31% H2 O, 31% HCl, 31% HNO3 and 7% HF.

Table I. Processing properties of cylindrical components. Microphotographs shown in Figure number

8

9

10

To (o C)

787

787

830

Ts (o C)

176

176

149

m/l (kg/m)

1.02

1.02

1.19

Micro-hardness tests were conducted on the above samples directly after fabrication (with no post-working) at OCM Test Laboratory in Anaheim California.

The results were provided as

Rockwell "B" Hardness values, which were converted to BNH values (500 kg load and a 10 mm diameter ball) in order to compare to the unworked AA2024 reference stock that has a Rockwell "B" Hardness value of less than zero and a BNH value of 47. The measurements were conducted by averaging five readings over a vertical distance of 1.0 cm in the sample. The procedure is identified as ASTM-E-384. The specific gravity was also measured by OCM Test Laboratories for the specimens.

4. Results

Accepted for publication in Aluminum Transactions Journal, 2000

Aluminum Ball Fabrication By selecting processing parameters that allow the droplets to solidify in flight, uniform aluminum balls can be fabricated with capillary stream break-up.

Shown in Figure 6 are

solidified aluminum (AA2024) balls that are 190 µm in diameter and were generated at a rate of 17,000 balls per second, this corresponds to 1.85 kg/hour. The droplet size can be estimated from conservation of mass by assuming that mass in one wavelength (λ) of the capillary stream of initial radius (ro ) forms a carrier droplet of radius rd :

4  rd =  ro2 λ 3 

1

3

And relating the wavelength of the applied disturbance to the nondimensional wavenumber

(k ) , * o

which is held constant in experiment, the above equation for rd can be rewritten as:

 8π rd = ro  *  3k o

  

1

3

The size of the droplets generated with a-m disturbances with integer frequency ratios is just rd N1/3 . The droplet production rate can be written as:

fc =

k o*V 2πro

or f c/N for droplets generated with integer frequency ratios. In the above, the stream speed (V) is estimated experimentally from conservation of mass measurements obtained by photographing the image of the droplet stream. The disturbance wavelength (λ) was measured by averaging up to 20 inter-droplet spacings off of several video images. Knowledge of the driving frequency (f c) and wavelength (λ) enables the estimate of stream speed V = fcλ. Measurements reveal that the balls are uniform to within 10.60% of the mean diameter without any sorting or sieving. Hence, droplet generation from capillary stream break-up provides a natural mechanism for highly

Accepted for publication in Aluminum Transactions Journal, 2000

uniform ball formation.

Figure 7 illustrates the microstructure in an individual solidified aluminum ball.

The scale

shown is 10 microns and hence, the small dimension of the grain size is on the order of 10 microns.

The use of such spheres is beneficial for a host of emerging applications and the

production rate can be scaled-up by using arrays of orifices. Microstructural Characteristics of Net-Formed Components Components were analyzed for their chemical composition and compared to the corresponding pre-jetted characteristics. It was found in all cases that the composition of the droplet-deposited components was consistent with AA2024 aluminum alloy, and therefore the process of droplet deposition does not significantly alter the alloy composition.

Figure 8 illustrates the microstructure of a tube component fabricated in a manner as illustrated in Figure 1. The microstructure is sampled from a location at 5.0 mm from the substrate. It can be seen that the grain size is on the order of 50 µm.

Figure 9 shows the microstructure of a sample that was taken from a distance of 1l0 mm from the substrate fabricated with the same processing conditions. It can be seen that the grain size is significantly larger than that of the corresponding sample closer to the substrate. The reason for the larger grain size is that the component temperature increases with increasing component height. This is because the droplets are delivered at a rate that is too high to allow for cooling prior to the arrival of the subsequent droplets.

Hence, as the component grows, the depth of

molten aluminum increases, thereby increasing the solidification time.

It is well known that an

increase in solidification time can be correlated to an increase in grain size.

As mentioned

earlier, current work is focused on modeling the heat flux required to locally remelt a thin layer of the previously solidified material in efforts to maintain a constant heat flux for any fabrication geometry.

Figure 10 is a microphotograph of a component illustrating coarse precipitation zones.

The

reason for the coarse precipitation zones is that both the droplet temperature and the mass

Accepted for publication in Aluminum Transactions Journal, 2000

delivered per unit length of substrate was higher than those for the components fabricated in Figures 8 and 9, thereby transferring more heat from the molten droplet stream to the component. Hardness It was found that the samples without the coarse precipitation zones had a hardness of over twice that of the raw stock, with typical BNH values in the range of 95-101. No significant difference in hardness was measured between samples obtained near the substrate or far from the substrate. The value of BNH hardness for sample shown in Figure 10 is 77, which is significantly lower than that of the other samples and it is concluded to be a result of the coarse precipitates that are observable in the micrograph and are due to elevated heating. Specific Gravity The specific gravity of specimens was found to vary from 2.62 to 2.71 g/cc, where the raw stock value of 2.70 g/cc is used as a reference. Hence, cylinders have been fabricated that are fully dense and devoid of measurable porosity. Even the specimen shown in Figure 10 is only 1.5% less dense than our reference value, yet the hardness is significantly lower than the other samples in this work.

Hence, the generation of widespread coarse precipitates appears to be the

influencing factor for the generation of softer components, and not the presence of porosity.

5. Conclusions This paper describes the generation and deposition of molten aluminum alloy droplets for advanced manufacturing applications.

Several important findings have been reported in this

paper, which chronicles the development of the PDM process. The most important findings are summarized below.

We have found that molten aluminum alloy droplets can be consistently generated from capillary stream break-up, provided that great care has been taken to remove traces of oxygen from the jetting environment, and to filter residual oxides from the melt, which both have the tendency to cause angular jitter in the droplet stream.

Large concentrations of oxygen in the jetting

environment will eliminate the possibility of generating droplets altogether.

Measurements have

revealed that aluminum alloy AA2024 droplets can be generated from capillary stream break-up

Accepted for publication in Aluminum Transactions Journal, 2000

and deposited in an inert environment without deviation from specified alloy composition limits.

Capillary stream break-up provides an attractive method for generating highly uniform metallic spheres at relatively high rates. Aluminum alloy AA2024 spheres have been fabricated at a rate of 18,000 spheres per second, which corresponds to 1.85 kg/hour for one orifice. Typical sample measurements reveal that the uniformity in sphere diameter is ±0.60% of the average diameter.

Molten aluminum-droplet deposition has been utilized for the fabrication of net-shape structures. Though the process is still immature, several important findings can be reported. 1) It has been found that structures as high as 11.0 cm can be formed (11.0 cm is the greatest height that our current apparatus will permit) with little measurable changes in component characteristics; 2) BNH of as-deposited material is roughly twice that of (annealed) raw stock; 3) an increase in hardness of 100% over the raw stock can be achieved under certain processing conditions and 4) fully dense components can be fabricated with this process.

With respect to microstructural and mechanical properties of the manufactured components, it was found that the processing parameters that govern the geometric, mechanical and microstructural integrity of the net-formed component are coupled with each other. It was found that the most important factor is the heat flux received by the cylinder during droplet deposition. The heat flux can be affected by varying the mass delivered per unit length, the droplet temperature or the substrate temperature.

Extreme heating of the cylinder through any of the

mechanisms discussed above results in coarsely precipitated microstructure, which can reduce the hardness of the component.

Additionally, extreme heat fluxes also thickened cylinder walls

due to slow solidification times. It is also important, however, that the droplets not be too cool prior to impingement, else the droplet boundaries will remain intact thereby reducing the structural integrity of the component.

Hence, there is a trade-off between high heat fluxes to

reduce porosity and increase hardness and low heat fluxes for higher control over component geometry and finer microstructures.

It is believed that the results presented here are an underestimate of the microstructural refinement and mechanical quality enhancements of the final realization of PDM technology.

Accepted for publication in Aluminum Transactions Journal, 2000

This is because the droplets did not solidify as rapidly as they will in the final PDM technology since electrostatic charging and deflection was not employed in the present work. In the final realization of PDM, the droplets will be deflected to different locations on the substrate, allowing the splats to cool prior to the next splats arrival.

In this manner, the temperature of the

previously deposited and solidified material will not increase with time, allowing the droplets to rapidly solidify which leads to refined microstructures and enhanced mechanical properties.

Acknowledgment: This work has received the generous support from Boeing Commercial Airplane Company (BCA-23483), Lawrence Livermore National Laboratories (B345710) and the National Science Foundation (DMI-9622400).

References 1.

M. Orme and C. Huang, "Phase Change Manipulation for Droplet-Based Solid Freeform

Fabrication of Aluminum Components," ASME J. Heat Transfer, vol. 119, 1997, pp 818 – 832

2. Q. Liu, C. Huang and M. Orme, "Mutual Electrostatic Interactions Between Closely Spaced Charged Solder Droplets" Journal of Atomization and Sprays, (in print, September 2000)

3. E. P. Muntz, M. Orme, G. Pham-Van-Diep and R. Godin, "An Analysis of Precision, FlyThrough Solder Jet Printing for DCA Components" presented at the 30th International Symposium on Microelectronics, Pennsylvania, published by IMAPS -Int. Microelectronics Packaging Soc., Reston, VA, USA 1997, pp. 671-680

4. M. Orme and E. P. Muntz, "The Manipulation of Capillary Stream Breakup Using Amplitude Modulated Disturbances: A Pictorial and Quantitative Representation," Phys. of Fluids, vol. 2, no. 7, 1990, pp. 1124 - 1140

5. M. Orme, "On the Genesis of Droplet Stream Microspeed Dispersions," Physics of Fluids, vol. 3, no. 12, 1991, pp 2936 - 2947 6. M. Orme, K. Willis and V. Nguyen, "Droplet Patterns from Capillary Streams," Physics of

Accepted for publication in Aluminum Transactions Journal, 2000

Fluids, vol. 5, 1993, pp 80 - 90

7.

M. Orme, "A Novel Technique of Rapid Solidification Net-Form Materials Synthesis,"

Journal of Materials Engineering and Performance, vol. 2, no. 3, 1993, pp 399 - 405

8.

M Orme, C. Huang and J. Courter, "Precision Droplet Based Manufacturing and Material

Synthesis: Fluid Dynamic and Thermal Control Issues", ILASS Journal of Atomization and Sprays vol. 6, 1996, pp 305 - 329

9.

M. Orme and R. Smith, "Enhanced Aluminum Properties with Precise Droplet Deposition"

ASME Journal of Manufacturing Science and Engineering, 2000 (in print, expected v. 122)

10.

E. M. Sachs, M. J. Cima, P. Williams, D. Brancazio and J. Cornie, "Three Dimensional

Printing: Rapid Tooling and Prototypes Directly from a CAD Model" J. Eng. Ind., vol. 114, 481488, 1992,

11. E. Sachs, M. Cima, J. Bredt and A. Curodeau, "CAD-Casting: The Direct Fabrication of Ceramic Shells and Cores by Three Dimensional Printing, "Man. Rev., vol. 5, no. 2, pp 118-126 1992

12. T. Jackson, N. Patrikalakis, E. Sachs, and M. Cima, "Modeling and Designing Components with Locally Controlled Composition," Solid Freeform Fabrication Symposium, Austin, Texas, 1998, pp. 259 - 266.

13. F. B. Prinz, L. E. Weiss, C. H. Amon and J. L. Beuth, "Processing, Thermal and Mechanical Issues in Shape Deposition Manufacturing," Solid Freeform Fabrication Symposium, Austin, Texas, 1995, pp. 118-129.

14. C. H. Amon, J. L. Beuth, R. Merz, F. B. Prinz, and L. E. Weiss, "Shape Deposition Manufacturing with Microcasting: Processing, Thermal and Mechanical Issues, ASME J. Manufacturing Science and Engineering, vol. 120, 1998, pp. 656-667.

Accepted for publication in Aluminum Transactions Journal, 2000

15.

J. Fessler, A. Nickel, G. Link, and F. Prinz, "Functional Gradient Metallic Prototypes

through Shape Deposition Manufacturing" Solid Freeform Fabrication Symposium, Austin, Texas, USA, 1997, pp. 521 - 528.

16. D. B. Bogy, "Drop Formation in a Circular Liquid Jet" Ann. Rev. Fluid Mech., vol. 11, 1979, pp. 207-228.

17. M. J. McCarthy and N.A. Molloy, "Review of Stability of Liquid Jets and the Influence of Nozzle Design" Chem. Engineering, vol. 7, 1974, pp. 1-20.

Accepted for publication in Aluminum Transactions Journal, 2000

Molten droplet stream

Inert environme

Fabricate d

Substrate motion

Figure 1: Conceptual schematic of cylinder fabrication on a flat-plate substrate with controlled droplet deposition.

Accepted for publication in Aluminum Transactions Journal, 2000

Figure 2: Examples of preliminary components fabricated with PDM. The tall square tube shown horizontally is 11.0 cm.

Accepted for publication in Aluminum Transactions Journal, 2000

Active cooling flange Stagnation pressure regulation

Periodic signal of wavelength λ to piezoelectric crystal

Molten aluminum Cartridge

orifice Droplet formation with an average inter-droplet spacing equal to λ

Figure 3: Schematic of experimental apparatus

Accepted for publication in Aluminum Transactions Journal, 2000

(a)

(b)

Figure 4: (a) evolution of droplets into modulation drops with an amplitude modulation ratio N=3, (b) amplitude modulated disturbance to the piezoelectric crystal

Accepted for publication in Aluminum Transactions Journal, 2000

(a)

(b)

Figure 5: (a) evolution of droplets into a stable configuration of large-small-large with an amplitude modulation ratio N=3.5, (b) amplitude modulated disturbance to the piezoelectric crystal

Accepted for publication in Aluminum Transactions Journal, 2000

Figure 6: SEM photograph of 2024 aluminum balls fabricated with capillary stream break-up

Accepted for publication in Aluminum Transactions Journal, 2000

Figure 7: Optical micrograph of cross-sections of several droplets illustrating grain structure (top view of a cut)

Accepted for publication in Aluminum Transactions Journal, 2000

Figure 8: Microstructure of aluminum cylinder sampled at 5.0mm from substrate

Accepted for publication in Aluminum Transactions Journal, 2000

Figure 9: Microstructure of cylinder sampled at 65mm from the substrate

Accepted for publication in Aluminum Transactions Journal, 2000

Figure 10: microstructure of a component that illustrates coarse precipitates

Suggest Documents