* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download MUTUAL ELECTROSTATIC INTERACTIONS BETWEEN
Survey
Document related concepts
Transcript
MUTUAL ELECTROSTATIC INTERACTIONS BETWEEN CLOSELY SPACED CHARGED SOLDER DROPLETS By Qingbin Liu, Changzheng Huang and Melissa Orme Department of Mechanical and Aerospace Engineering University of California, Irvine, CA 92697-3975 ACCEPTED FOR PUBLICATION IN JOURNAL OF ATOMIZATION AND SPRAYS ABSTRACT Emerging technologies of net-form manufacturing and electronic packaging rely on the use of precisely deposited molten metal droplets with sizes on the order of 100 µm in diameter. In many technological realizations, closely spaced droplets are electrostatically charged and deflected onto a substrate in a manner similar to ink-jet printing in order to “print” fine features onto a board for electronics applications or onto a substrate for net-form manufacturing. Unlike ink-jet printing, the aforementioned emerging technologies require the printing of large lateral dimensions onto the substrate by means of electrostatic charging and deflection (on the order of several centimeters), and hence these applications require the droplets to have significantly higher charges than in the ink-jet printing technology. The high charges of the closely spaced droplets can lead to inter-droplet electrostatic interactions that will cause significant deviations in the droplets’ trajectories. Hence, the understanding of the physics of inter-droplet electrostatic interactions is of primary importance in order to assure the fidelity of the net-formed component or the printed electronic package. In this work, we present a model that predicts the trajectories of the droplets when charged and deflected and corresponding experimental validations. Conditions for which electrostatic interactions are negligible are sought. The authors acknowledge the generous grant from the National Science Foundation Grant #DMI-9622400 which supported this work. 1 NOMENCLATURE A a b Cd D E Fi,i+1 f g I i k o∗ k o∗ max m cross sectional area of the droplet distance between two deflection plates radius of charge tube drag coefficient aerodynamic drag force electric field strength electrostatic repulsive force between droplets i and i+1 frequency of forcing disturbance gravitational constant current drawn by droplet stream droplet index nondimensional wavenumber =2πro /λ nondimensional wavenumber at maximum growth rate Q Re r rd ro Vo Vx Vy Vz Vc droplet mass mass flow rate of droplet stream droplet charge Reynolds number instantaneous distance between two droplets droplet radius initial stream radius initial stream speed component of velocity in x direction component of velocity in y direction component of velocity in z direction charge potential α β εo λ ν ρ ρa σ direction of repulsive force growth rate of the applied disturbance permittivity of free space wavelength of the applied disturbance kinematic viscosity of the fluid density of molten metal density of ambient fluid surface tension of molten metal m & 2 1. INTRODUCTION There is recent interest surrounding the use of highly controlled molten metal droplet streams for net-form manufacturing of structural components [1-9] and printing of electronic packages [1012]. In a subset of these developing technologies, the droplet streams are electrostatically charged and deflected for rapid high precision printing onto a substrate [1,3,7, 9-12]. Figure 1 is a conceptual schematic of the droplet printing technology, which is implemented in this work. Droplets that are generated from capillary stream breakup acquire an electrostatic charge by passing through the charge tube at the point of droplet formation. The initial direction of the droplet stream is along the negative z-axis. The charge signal is carefully synchronized with the droplet formation disturbance waveform so that droplets can be charged on a drop-to-drop basis. The charged droplets pass through an electric field, which is established by two vertical deflection electrodes as indicated in the figure. Passing through the electric field causes the droplets to acquire a component of velocity in the x direction, the magnitude of which is directly related to the charge on the drop. Continuous substrate motion in the y direction allows printing in two dimensions as shown for the example of Ball Grid Array (BGA) fabrication for electronics packaging. More details of the specific conditions employed to print the shown BGA are provided in reference [12]. For the similar application of net-form manufacturing which entails the fabrication of three-dimensional cohesive structural components without any postworking, droplets are deposited in an overlapping regime such that fluid and thermal interactions occur, and successive layers are deposited in order to build the component in the z dimension. Since many applications require droplet placement over large lateral areas, it is of interest to determine the maximum charge that can be applied to the droplets before the onset of significant inter-droplet interactions that prevent accurate droplet targeting. Inter-droplet interactions occur 3 since the droplets are closely spaced and carry like charges. Hence, the droplets act to repel each other but their motion is impeded by the existence of their droplet neighbors on the opposite side, which also possess charges of the same sign but different magnitude. Figure 2 illustrates the phenomenon of mutual electrostatic interactions. Shown are photographs of two molten solder droplet streams issuing vertically into an inert environment from a 100 µm diameter orifice. In each experimental realization the droplet generation frequency is 12 kHz and the initial vertical droplet speed is 5.5 m/s. In the droplet trace shown to the left, the droplets were charged by passing the capillary stream through a charge electrode at the time of droplet formation. A constant voltage of 250 Volts DC is applied to the charge electrode, giving a charge of -20.24×10-13 Coulombs to every droplet. The negative charge on the droplets is a result of using a positive charge voltage. This causes electrons to be drawn onto the stream by the attraction to the oppositely charged electrode. Since the droplets each posses the same charge, they act to repel each other, however their motion along the vertical axis is constrained by the existence of their charged neighbors on the opposite side. Therefore, the droplets move laterally out of the main stream creating a helix spiraling down in x-y-and z, that appears to look like the zigzag when viewed in one plane as shown. The current experimental apparatus is only equipped with observation ports for photographing the x-z plane. Our interpretation of the helical behavior of the charged droplet stream is based on our observations of the droplet trajectory after rotating the droplet generator 90o in order to observe y-z plane. These observations also revealed that the zigzag path was not limited to the x-z plane, but also occurred in the y-z plane. While the experimental results of this paper are limited to observations in the x-z plane, the phenomenon is three-dimensional, and therefore a three-dimensional model is provided here. The droplet stream shown in the plate to the right is included as a reference of an uncharged solder droplet 4 stream and was generated with the same conditions (excluding charge) as the droplet stream shown to the left. Since the charging waveform is unique for each pattern to be printed and is likely to be more complicated than the DC example shown above, general methods of compensating for the interactions are non-existent at this time and are not within the purview of this work. This work is aimed at presenting a model which is corroborated by new experimental results that predicts the droplet trajectory subject to highly charged, closely spaced droplets. 2. BACKGROUND 2.1. Droplet Formation from Capillary Stream Break-up Capillary stream break-up has been used to generate streams of uniformly sized droplets for well over 100 years, and is typically achieved by applying the well-known Rayleigh instability [13]. In this work, the periodic motion of a piezoelectric crystal is used to establish a disturbance on the surface of a molten metal capillary stream in order to initiate droplet formation as illustrated in Figure 3. According to the linear theory of Rayleigh, the radial disturbance on the surface of the jet will be unstable and grow resulting in droplet formation for conditions when the nondimensional wavenumber, k o * , which is defined as the ratio of the initial stream circumference to the wavelength of the imposed disturbance, is less than unity. Lord Rayleigh developed the first linear stability analysis where he considered an infinitely long, circular, inviscid jet subject to a temporal disturbance growth. He found that disturbances of the radius grow in time t as eβt, where β is the growth rate of the disturbance. A more detailed description of the droplet formation process from capillary stream break-up can be found in the 5 review articles by Bogy [14] and McCarthy and Molloy [15]. Droplet formation from capillary streams has been the subject of numerous experimental and theoretical works. A complete review of the subject is not within the purview of this paper, which is focused on droplet charge interactions. The droplet generation technique defined by the Rayleigh instability is termed “continuous” droplet generation since a capillary stream is broken into a continuous stream of droplets. Sheild et al. [16] and Bousfield et al. [17] have developed techniques of generating droplets at given pulses and thus are termed as drop-on-demand. The drop-on-demand mode of droplet generation is limited by lower droplet production rates than the continuous mode. 2.2. Droplet Charging A great deal of background work on droplet charging has been performed in the 1960’s and 70’s for the well-known application of ink-jet printing. A few of the more important experimental works are given by Sweet [18, 19], Schneider et al. [20], Kamphoefner [21], and Fillmore et al. [22]. Analogous to the applications described in this work, an ink-jet jet printer produces characters on paper by deflecting charged droplets on one axis while the print head moves along the perpendicular axis. The droplet charge and the strength of the electric field through which the droplet moves determine the amount of deflection achieved by a droplet. The net charge on the droplet is acquired at the time of droplet formation. In contrast to the ink-jet droplets, the molten metal droplets in this work attain significantly higher charges so that large lateral areas can be printed in order to fabricate a structural component of meaningful dimensions. The high charges cause significant inter-droplet mutual electrostatic interactions to occur which are not evident in the ink-jet printing technology as was 6 illustrated in Figure 2. A conceptual schematic of the droplet charging and deflection configuration is shown in Figure 3. As previously discussed, droplets are charged by passing the capillary stream through a charge electrode at the point of droplet break-off from the stream. The molten jet is grounded, and a positive potential, Vc, is applied to the charge electrode. The amount of charge given to the droplet is determined by the level of voltage applied to the electrode at the time of droplet formation. It is assumed that the conductivity of the liquid is high enough to allow a steady charge to quickly form on the surface of the jet prior to break-up, and that the charge acquired by each droplet is the surface charge on a length λ of the capillary stream. With these assumptions, the charge to mass ratio can be predicted by the relation given by Schneider et al. [20]. Q 2πε V = 2 o c m ρro ln( b / ro ) (1) Here, m is the droplet mass, εo is the permittivity of free space, Vc is the charge potential, b is the radius of the charge tube, ρ is the molten metal density, and ro is the radius of the unperturbed capillary stream. In order to attain 100% charge efficiency, the fluid attached to the capillary stream that will form the leading droplet must reside in the charge tube for a time sufficient for the charge tube to acquire the desired charge value. After droplet formation, the voltage is changed in order to apply a different charge on the next droplet to be formed. Since the conductivity of the liquid is high, it is assumed that charge relaxation occurs faster than the transit time for an element of fluid to move through the charge electrode. Since the droplets acquire a charge at the time the droplet pinches from the capillary stream, the charge electrode should be positioned so that any 7 excursions in droplet formation length are still contained in the charge tube. The synchronization of the charge pulse with break-up time is adjusted manually by tuning the phase of the voltage pulse to be synchronous with the break-up time. Excursions in break-up distance can occur due to fluctuations in the driving pressure, fluctuations in amplitude of the disturbance due to the resonance of unwanted frequency components that are apparatus specific, and variations in fluid properties with temperature such as viscosity and surface tension. It is of critical importance to eliminate any excursions in break-up distance since this will lead to a loss of synchronization between droplet formation and droplet charge. Hence, in the undesirable occurrence of fluctuations in break-up distance, droplet charges will be “assigned” to the wrong droplets, causing gross errors in droplet targeting. 3. MODEL DEVELOPMENT Consider a sequence of similarly charged droplets emanating from a capillary stream that is oriented along the z-axis. Suppose that at time t=t1, droplet i experiences a disturbance that causes it to drift a modest distance from the main stream. Then the nearest neighbors on both sides of the drop, i+1 and i-1, immediately feel the repulsion forces from the droplet i, which in return feels the anti-reaction force from its neighbors. Similarly, at a later time t=t2 , the droplets i+1 and i-1 have been pushed out to the opposite side of the main stream. Then the droplets i+2 and i-2 feel the repulsion and they will move to the same side of droplet i. The process continues until only a very few (if any) droplets will remain on their intended trajectory, deeming a sequence of similarly charged droplets fragile to transverse disturbances. The transverse disturbance could be as simple as the droplet passing through a set of deflection electrodes for the purpose of droplet targeting. 8 A similar situation occurs when a stream of droplets carry different charges of high magnitude (e.g., to print a horizontal line or other feature). A droplet stream in which each droplet has a charge that is different in magnitude from its nearest neighbor, will also induce repulsive reactions thereby causing them to adjust their positions. The adjustment of one droplet position is not an isolated event, since all of the droplets will adjust their positions to account for the adjustment of their neighbors’ position. Hence, in the situation of non-uniform droplet charges (or AC charging), the droplet excursions will not be the uniform helical behavior as shown in the two-dimensional cross-section of Figure 2 for charges of equal magnitude and sign, but will be a complicated sequence of droplet positions wholly dictated by the characteristics of the charging waveform. Governing Equations The equations of motion used to simulate the trajectory of the ith charged droplet traveling through an electric field of intensity E in the x direction are given as: m jmax dVx,i = − D cosθ x + ∑ Fij cosαx + Qi E dt j =1 (2) j ≠i m m dV y,i dt = − D cosθ y + dVz ,i = − D cosθ z + dt jmax ∑ Fij cos αy (3) j =1 j ≠i jmax ∑ Fij cos αz + mg (4) j =1 j ≠i 9 Where we assume that the undeflected droplet stream is aligned with the z coordinate which is parallel to the gravity vector, and the deflection plates establish a field in the x-z plane so that the magnitude of the deflection due to passage through the deflection plates is measured in the perpendicular x coordinate. Ideally, droplet trajectories will deviate from the z-axis along the xaxis due to the established electric field. However, mutual electrostatic repulsion will cause the droplets to deviate from their ideal trajectory in both the x and y plane. In the above, Vx, Vy and Vz are the components of velocity in the x, y, and z coordinates defined by: cosθ x = Vx V (5) cosθ y = Vz V (6) cosθ z = Vz V (7) V 2 = Vx2 + Vy2 + Vz2 (8) and the electrostatic repulsive force between the ith and j th droplet is given by: Fij = 1 Qi Q j 4πεo r 2 (9) Where Qi and Qj are the charges of the ith and j th droplets given by Equation (1). The angle α is the direction of the repulsive motion given by: 10 cosαx = cosαy = cosαz = xi − x j ( xi − x j ) + ( yi − y j ) + ( zi − z j ) 2 2 (10) 2 yi − y j ( xi − x j ) + ( y i − y j ) + ( zi − z j ) 2 2 (11) 2 zi − z j ( xi − x j ) + ( yi − y j ) + ( zi − z j ) 2 2 (12) 2 Where x i, x j, yi, yj, zi, and zj are the x, y and z positions of the ith and j th droplets with the z-axis pointing vertically downward and the origin of the x and y-axes are placed at the top of the deflection plates. In the above, g is the gravitational acceleration, and D is the aerodynamic drag force defined by: 1 D = Cd ⋅ ρaV 2 ⋅ A 2 (13) where ρa is the density of the atmosphere, A is the cross sectional area of the droplet, and Cd is the drag coefficient of a sphere given by White [23] Cd = 24 6 + + 0.4 Re 1 + Re and Re is the Reynolds number based on the diameter of the droplet. (14) The drag coefficient described by Eq (14) is for an isolated sphere traveling in a quiescent environment. Since the droplets in this work are deflected by various amounts out of the main stream, the above model 11 may overestimate the drag force since many of the droplets are protected by the wake provided by their nearest neighbor. The system of ordinary differential equations described by Equations (2), (3) and (4) above is solved iteratively for each successive time step. 4. EXPERIMENTAL APPARATUS The experimental setup consists of a droplet generator that injects droplets into an environmental chamber, capable of being evacuated to a few times 10-5 torr with the aid of a diffusion pump. Although the experiments are performed in an inert environment at a pressure of one atmosphere, we have found that it is necessary to evacuate traces of oxygen from the chamber in order to avoid the disruptive effects of oxidation which work to impede molten metal droplet formation. Immediately below the droplet generator are the charging electrode and the parallel deflection plates. The separation between the orifice, which is contained at the lower end of the droplet generator, and the substrate, is 295 mm. The substrate is mounted on an x-y table that is capable of moving at speeds up to 13 cm/s. allow optical observation in the x-z plane. Two windows on opposite sides of the chamber A schematic of the facility is shown in Fig 4. A microscope with long working distance optics and equipped with a camera is mounted outside the chamber. Values used in the experiment are given in Table 1 below unless otherwise noted in the text. The droplet deflection is highly sensitive to the initial unperturbed stream radius ro . The initial stream radius may not necessarily be identical to the radius of the orifice depending on the orifice shape. An orifice which has a smoothly contoured entrance region followed by a short tube will have a relatively flat velocity profile at the orifice exit resulting in a stream radius 12 droplet fluid specific gravity surface tension Viscosity orifice diameter ambient and stagnation gas driving pressure solder reservoir temperature disturbance frequency charge electrode inner diameter deflection plate separation deflection plate length Eutectic Solder, 63% Sn, 37% Pb 8.420 0.49 kg/s2 1.58x10-7 m2 /s 100 microns Nitrogen 138 kPa 200o C 12,000 Hz 0.318 cm 1.27 cm 5.08 cm Table 1: Properties used in experiment commensurate with the orifice radius. Streamlines through a flat plate orifice, on the other hand, will be curved at the exit, causing a non-uniform pressure distribution across the radius of the stream. The velocity distribution in the jet becomes uniform a short distance away from the orifice where the jet area is smaller than the orifice area. Stream radius measurements were made by measuring the overall mass flow rate of the droplet stream, m& and relating it to measured stream quantities. In each experimental realization, the number of droplets collected was large enough to insure negligible errors due to turn-on/off transients. 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, and wavelength λ, enables the estimate of stream speed V= fλ. From conservation of mass, the undisturbed stream radius, ro , and the droplet radius, rd , are given by: ro = & m , ρVπ & 3 m rd = 4 ρπf (15) 13 5. RESULTS 5.1. Droplet Charge Measurements As a benchmark to our deflection experiments, we have measured the droplet charge as a function of charge electrode voltage. In order to measure the droplet charge the droplets are collected in a conductive container that is set on an electrically insulated surface. The container is connected to ground through a picoammeter and the current is measured. The picoammeter is connected between the droplet stream collector and the ground instead of between the voltage source and charge electrode to avoid erroneously high current readings due to the leakage of current through the thin insulator between the charge electrode and the grounded droplet generator. The current measured corresponds to that drawn by the stream due to charging. The following relationship is used to determine the droplet charge to mass ratio from our measurements: Q I = m m& (16) In the above, I is the current drawn by the droplet stream, and m & is the mass flow rate of the solder through the orifice. Operating conditions for this measurement are as listed in Table 1 except for the fact that the charge electrode voltage is varied from zero to 200 volts DC. At the largest charge potential used for this measurement of 200 volts, a current of -16.3 pA is drawn by the droplet stream. This corresponds to an average droplet charge of -1.36×10-12 Coulombs. Our experimental results are shown as circular symbols in Figure 5, which also shows the theoretical prediction due to Schneider (Eq. 1) as a solid line. Note that agreement between experiment and 14 theory is excellent, and we can therefore be assured that any deviations between predicted and experimental deflection do not arise due to errors in charge efficiency. 5.2. Charged Droplet Interactions Figures 6 and 7 provide illustrative examples of the detrimental effects of mutual droplet charge interactions. In the figures, the characters “UCI” were printed by electrostatically deflecting the charged droplets along the x-axis as shown while the substrate was in continuous motion along the y-axis. The fluid dynamic properties of the droplets are provided in Table 1. waveforms used to print the characters are provided in Figure 8. The charging The maximum charge used to generate the characters in Figure 6 is -6×10-13 Coulombs, which is relatively low, and hence, this charging waveform does not create observable errors in droplet targeting. The character height is 4.7 mm. The characters shown in Figure 7, on the other hand, were obtained with relatively high charges as is evident from the height of the characters (illustrating the extent of the deflection) which is 9.4 mm. The maximum charge used to print the characters in Figure 7 is –1.5×10-12 Coulombs. Also apparent in the figure is the undesired waviness of the lines formed by droplet splats that create the characters. This waviness is due to electrostatic interactions, which are more severe when the droplets are highly charged as evidenced in the example shown here. A few comments are required at this point to describe the nature of the deviations from the ideal target for the high charging case. First, it can be seen that there is not a systematic displacement in droplet position with applied voltage that one may expect from the theoretical expressions. To understand this finding, we consider the waveform that was used to print the character “U” as shown in Figure 8. The first five peaks illustrate the relative charges applied to the five droplets that created the first of the two vertical lines along the x-axis of the “U”. One 15 may expect the interactions between drops i=4 and i=5 to be greater than the interactions between drops i=2 and i=3 due to the magnitude of their charges. It should be pointed out, however, that drop i=5, which is a highly charged end drop, feels the repulsive force from only one neighboring drop, i=4, whereas drops i=2 and i=3 feel the repulsive forces from neighboring drops on both sides. Hence, there is a competition between the magnitude of the droplet charge and the magnitude of the charges of the neighboring drops. Also, it appears that the y-axis error is greater than the x-axis error. This is not necessarily the case, and the x-axis error and the y-axis error are on a par. The waviness (y-axis error) shown on the vertical segments made from the five closely spaced charged drops is due to the helical behavior of the droplet charges as they travel along z. However, there is also an error in the placement along the x-axis, which is the y-axis error that is on a par in magnitude with the x-axis error that can be observed by noting the relative differences in inter-droplet spacing along the x-axis. Additionally, the base of the “U” character was printed with the two isolated droplets as shown in Figure 8. These droplets are sufficiently isolated to exhibit negligible inter-droplet interactions, and the apparent error in droplet placement (which appears to be an x-axis error) is somewhat misleading. To illustrate this point, consider all of the droplets in the “UCI” except those that were printed with closely spaced drops (i.e., except the four lines of five drops along the x-axis). Examination of the remaining drops illustrates that they exhibit very little x-axis errors, and these positions are very near to their “ideal” positions. The gray horizontal line has been included as a reference to illustrate the point that the only drops exhibiting errors from this reference are those associated with the vertical lines of five drops. Hence, it is the lines that were printed with closely spaced charged drops that exhibit mutual repulsions that cause the end drops to “kick-out” along the x-axis farther from the “ideal” position. 16 The model described in this work was developed as a tool to study the effect of electrostatic mutual repulsion, and was used to simulate the trajectory of five droplets under the influence of gravity, aerodynamic drag, and mutual electrostatic repulsive forces. plane were also conducted to verify our model results. Experiments in the x-z Figure 9 illustrates the charging waveform that was used in both the simulation and experiment. In this case, the maximum charge is –8.2×10-13 Coulombs which is considered to be relatively low. Figure 10 illustrates the model simulation of the trajectories of five droplets subject to the charging waveform shown in Figure 9. excluded. Here, the effects of gravity and drag are included, while the repulsive forces are The circular symbols are our experimental measurements. In both experiment and simulation, the droplets traveled a downward vertical distance of 295 mm until impacting with the substrate which is indicted by the horizontal dashed line. In the absence of repulsive interactions, the simulations predict that the droplets should be evenly spaced as shown where the solid lines intersect the horizontal dashed line. However, our experimental measurements indicate that there is a small amount of repulsive interactions occurring as can be seen by the increased separation in experimental data between drops i=3 and i=4. This is due to the fact that the charges on the drops i=1 and i=2 are too small to cause measurable inter-droplet repulsions, and the charges on drops i=3 and i=4 are larger than on any other bounded set of droplets (droplet i=5 is unbounded and hence only feels the repulsive force from drop i=4). Including the effects of mutual repulsion in the simulation leads to the results shown in Figure 11. In this case, the model simulations are in excellent agreement with experiment, which are reproduced on Figure 11 for comparative purposes. Figure 12 illustrates the charging waveform used in the experiment and simulations for the high charging case shown in Figures 13 and 14. In Figure 13, the simulations included the effect of 17 gravity and drag, but not electrostatic repulsion. As can be seen by the solid lines, which depict the trajectories of the droplets, the simulation predicts that the droplets will be uniformly separated when they encounter with the substrate. Our experimental measurements show that the droplet positions are highly non-uniform which is a consequence of the electrostatic repulsive forces. Unlike the data of the low charging case as shown in Figures 10 and 11, the magnitude of the charge on all of the droplets (even drops i=1 and i=2) is large enough to cause mutual interactions, and therefore, the deviation between experimental and the ideal trajectories is significant. It can be seen that droplet i=3, the central droplet is the only droplet that hits its target. This is because, unlike the case for low charging, the interactions from both drops i=2 and i=4 are significant, and the deviations become more symmetric (though not completely symmetric due to the variations in absolute droplet charge). The discrepancy between experiment and simulation here illustrates the magnitude of the repulsive forces. Figure 14 illustrates the simulated droplet trajectories with the effect of gravity, drag, and electrostatic repulsion included. As is evident from the figure, there is excellent agreement between experiment and simulation, and electrostatic interactions cause significant deviations from the ideal droplet trajectories. It is worth noting that without considering the effects of repulsive forces in the simulation, one may be led to believe that there is scatter due to experimental error in the measurement. However, the consideration of the effects of the repulsive forces allows the correct conclusion to be drawn that there is indeed little experimental error and the droplet placement can be predicted to a high degree of accuracy. There are two distinct issues that define upper limit for the charge that can be applied to a drop. The first, which is the subject of this paper, defines the maximum charge that can be applied to the stream before mutual charge interactions cause unacceptable printing resolution. In this 18 work, we assume that droplet placement deviations greater than ±12.5 µm at the substrate resulting from charge interactions are unacceptable as suggested by individuals representing certain sectors of the electronics assembly industry [12]. The second issue is that of charge induced droplet disintegration. If the electrostatic forces are strong enough to overcome the surface tension force that holds the drop together, the drop will catastrophically disintegrate. Lord Rayleigh [24] estimated the maximum charge before disintegration with the following expression: Qmax = 64π 2 ε o rd3 σ (17) For the molten solder droplets generated in this work, Qmax is equal to 48.0 ×10-12 Coulombs, which is well above the values used in this work. We have used the model presented in this work to estimate the maximum charge that can be applied to a droplet stream before mutual electrostatic interactions cause deviations in placement on the substrate greater than ±12.5 µm. The deviations in trajectories depend on flight distance from the exit of the deflection plates and the magnitude of the deflection field. In order to make the analysis independent of our apparatus dimensions, we have defined the origin of the coordinate system to be coincident with the exit face of the deflection electrodes, and plotted the maximum charge as a function of flight distance for several values of the deflection plate voltage. The results are shown in Figure 15, where it can be seen that the mechanism governing the limiting droplet charge is due to mutual droplet interactions and not to droplet disruption, as the maximum charge for droplet disruption is over an order of magnitude higher than that for unacceptable interactions. Since the targeting errors due to interactions amplify over flight distance, the maximum charge that may be applied to the drop decreases with flight distance. It is 19 also seen that droplets traveling through a weaker electric field can tolerate higher charges before significant interactions are apparent. It is shown in previous work [25], that the droplet deflection is a function of the product of the charge voltage and the deflection voltage. Hence it can be reasoned that the errors in droplet deflection also depend on the product of droplet charge and charge potential, and therefore, an increase in droplet charge can be compensated for by decreasing the deflection potential as evidenced by the simulation results in Figure 15. For conditions employed in this work, the simulations indicate that the charge must not exceed –6.1×10-13 Coulombs for a deflection voltage of ±2,500V if droplet placement accuracy less than ±12.5µm is to be maintained, which is consistent with all of the results presented in this work. The model provided in section 3 could be used to predict the maximum charge for any required printing accuracy. 6. CONCLUSIONS Through model simulations and experimental measurements, it is shown that electrostatic repulsive forces between highly charged droplets are significant and can cause considerable deviations from the ideal droplet trajectory, thereby leading to printing errors if the repulsive forces are ignored. Such errors are not apparent in the technology of ink-jet printing since smaller deflections are required, deeming the use of high charges unnecessary. Hence, it is the emerging applications of net-form manufacturing and electronic package fabrication from metal droplet printing, which rely on the deposition of highly charged metallic droplets in order to print large dimensions onto a substrate that will suffer from the existence of electrostatic repulsive forces. In those applications, methods to compensate for the repulsive forces must be developed. Since the series of repulsive interactions for a droplet stream is unique for each 20 charging waveform, optimization of the charge waveform in an effort to compensate for the interactions can be a tedious endeavor since general optimization methods are currently not available. Also presented was a prediction of the maximum charge that can be applied to the stream before electrostatic interactions that cause deviations greater than ±12.5µm in droplet placement at the substrate occur, which is a requirement imposed by certain sectors of the electronic package industry. The model presented in this work can also be used to ascertain the maximum charge for any required printing accuracy. It is shown that the maximum charge that can be applied decreases with flight distance to the substrate, which can be reasoned to be due to the fact that the errors in droplet trajectory are amplified with flight distance. Also, it is shown that increasing the magnitude of the electric field through which the charged drop travels results in a decrease in the maximum charge that can be applied to the drop. For the conditions employed in this work, the limiting charge is due to repulsive interactions, and the maximum charge before droplet disintegration is nearly two orders of magnitude higher. 21 1. Orme M., Muntz E.P. 1992, United States patent Number 5,171,360, December 15 2. Sachs E., Cima M., Bredt, J., and Curodeau, A., “CAD-Casting: The Direct Fabrication of Ceramic Shells and Cores by Three Dimensional Printing,” Manufacturing Review, Vol. 5, No. 2, pp 118-126, 1992 3. Orme M., 1993, “A Novel Technique of Rapid Solidification Net-Form Materials Synthesis,” Journal of Materials Engineering and Performance, 2, (3) 4. Prinz, F.B., and Weiss, L.E., 1994, “Method and Apparatus for Fabrication of ThreeDimensional Metal Articles by Weld Deposition,” U.S. Patent No. 5,207,371 5. Chin, R.K., Beuth, J.L., and Amon, C.H., 1995, “Control of Residual Thermal Stresses in Shape Deposition Manufacturing, Solid Freeform Fabrication Symposium, Austin, TX, 221-228 6. Argarwala M.K., Van Werren R, Jamalabad V., Langrana N., Whalen P., Danforth S.C. and Ballard C. “Quality of Parts Processed by Fused Deposition” Proceedings to the Solid Freeform Fabrication Symposium, University of Texas at Austin, 1995 7. Orme M, Huang C and Courter J, 1996, “Precision Droplet Based Manufacturing and Material Synthesis: Fluid Dynamic and Thermal Control Issues”, ILASS Journal of Atomization and Sprays vol. 6 8. Amon, C.H., Schmaltz, K.S., Merz, R., Prinz, F.B., 1996 “Numerical and Experimental Investigation of Interface Bonding Via Substrate Remelting of an Impinging Molten Metal Droplet” ASME J. of Heat Transfer, 118, 164-172 9. Orme, M.E., Huang, C. 1997, “Phase Change Manipulation for Droplet-Based Solid Freeform Fabrication”, ASME Journal of Heat Transfer, 119, 818 – 823 10. Gao F. and Sonin, A.A., 1994 “Precise deposition of molten microdrops: the physics of digital microfabrication,” Proc. R. Soc. London, Ser. A 444, 533 11. Prinz, F.B., Weiss, L.E., and Siewiorek, D.P., 1994 “Electronic Packages and Smart Structures Formed by Thermal Spray Depositions,’ U.S. Patent No., 5,278,442 22 12. Muntz E.P., Orme, M.E., Pham-Van-Diep, G., and Godin, R. 1997 “An Analysis of Precision, Fly-Through Solder Jet Printing for DCA Components’ Proceedings of the 30th International Symposium on Microelectronics, Pennsylvania , October 1997 13. Rayleigh, Lord, 1879, On the instability of jets. Proc. London Math. Soc. 10, 4-13 14. Bogy, D.B., “Drop Formation in a Circular Liquid Jet” Ann. Rev. Fluid Mech., 1979, 11: 207-228 15. McCarthy and Molloy, “Review of Stability of Liquid Jets and the Influence of Nozzle Design’ Chem. Engineering, 7, 1-20, 1974 16. Shield T.W., Bogy D.B., and Talke F.E. “Drop Formation by DOD Ink-jet Nozzles: A Comparison of Experimental and Numerical Simulation,” IBM J. Res. Dev. 31, 96, 1987 17. Bousfield D.W. and Denn M.M. “Jet Breakup Enhanced by an Initial Pulse,” Chem. Eng. Comm. 53, 61, 1987 18. Sweet R. G. “High-Frequency Oscillography with Electrostatically Deflected Ink Jets,” Stanford Electronics Laboratories Technical Report No. 1722-1, Stanford University, CA, 1964 19. Sweet R. G. “High Frequency Recording with Electrostatically Deflected Ink Jets”, Rev. Sci. Instrum. 36, 2, 131, 1965 20. Schneider J.M., N.R.Lindblad, and Hendricks C.D, “Stability of an Electrified Liquid Jet”, J Applied Physics. 38, 6, 2599, 1967 21. Kamphoefner F.J. “Ink Jet Printing”, IEEE Trans. Electron Devices ED-19, 584, 1972 22. Fillmore G.L., Buehner, W.L., West, D.L., “Drop Charging and Deflection in an Electrostatic Ink Jet Printer”, IBM J. Res. Develop. Jan, 1977 23. White F. M. Viscous Fluid Flow, 2nd Edition, McGraw-Hill Inc. 1991 24. Lord Rayleigh, Phil Mag. 14, 184 (1882) 25. M. Orme, Q. Liu, J. Courter, and R. Smith, “Electrostatic Charging and Deflection of Arbitrary Molten Solder Droplet Formations” Physics of Fluids, A, (in review) 1998. 23 FIGURES Figure 1: Schematic of droplet printing with electrostatically charged and deflected molten metal droplets. The Ball Grid Array shown was printed in 0.14 seconds from a stream of droplets generated from capillary stream breakup. Ball diameters are approximately 190 microns. Figure 2: Photograph of two molten solder droplet streams. The top droplet in each plate has traveled 90 mm from the orifice exit. The droplet stream on the left is composed of droplets which charges of 20.24×10-13 C, and the droplets on the right are uncharged. Figure 3: Photograph of molten solder droplet formation from capillary stream break-up and illustrative sketch of generation and charging phenomenon. Figure 4: Schematic of experimental apparatus Figure 5: Measured charge on a solder droplet stream as a function of charge electrode voltage compared to the theory of Schneider. Figure 6: Molten solder splats composing the characters “UCI”. Droplets that were 189 microns in diameter were charged and deflected onto a black paper substrate where they underwent the splatting action as shown. Droplets were charged with fairly low charges as is evident from minimal charging interactions. Figure 7: Molten solder splats composing the characters “UCI”. Droplets that were 189 microns in diameter were charged and deflected onto a black paper substrate where they underwent the splatting action as shown. Droplets were highly charged as is 24 evident from the size of the characters and the charging interactions, which cause observable non-uniform character spacing. Figure 8: Charging waveforms used to print the characters U, C, and I as shown. The maximum charge used in the low and high charging (figures 6 and 7 above respectively) case is –6×10-13 and –1.5×10-12 coulombs respectively. Figure 9: Charge waveform used in the experiments and simulations shown in Figures 10 and 11. Figure 10: Model simulations shown as solid lines of the trajectories of five droplets for low charging with gravity and drag included. Circular symbols illustrate experimental results. Figure 11: Model simulations shown as solid lines of the trajectories of five droplets for low charging with gravity, drag and electrostatic repulsion included. Circular symbols illustrate experimental results. Figure 12: Charge waveform used to generate the experimental results shown in Figures 13 and 14. Figure 13: Model simulations shown as solid lines of the trajectories of five droplets for high charging with gravity and drag included. Circular symbols illustrate experimental results. 25 Figure 14: Model simulations shown as solid lines of the trajectories of five droplets for high charging with gravity, drag and electrostatic repulsion included. Circular symbols illustrate experimental results. Figure 15: Maximum charge that can be applied to the solder droplet stream before interactions cause deviations greater in value than ±12.5µm at the substrate. The three curves correspond to values of the deflection plate voltage of ± 1500, 2500, and 3500 volts with a 1.27cm separation between deflection plates. 26 Inert environment Molten droplet stream Charge electrode Deflection plates Substrate motion 27 28 orifice plate charge tube vibrating rod 2rd b λ λ 2ro λ droplet formation signal charge s ignal 29 30 31 x, deflection axis y, substrate motion 32 x, deflection axis y, substrate motion 33 34 35 36 37 38 39 40 41