Jetting Small Dots of High Viscosity Fluids for Packaging Applications

Applications for various fluids in electronics assembly have increased at a rapid pace over the years. In the 1970s, the packaging of die was completed using a leadframe, die bonding, wire bonding, and plastic molding process. Printed circuit board (PCB) assembly used wave soldering equipment and IC inserters. The two fluid dispensing applications were die attach and temporary solder masks. With the growth of hybrid assembly and surface-mount technology (SMT), the dispensing applications expanded to solder paste, surface-mount adhesives, die-attach materials, glob top, temporary solder mask, and die coating. In the late 1980s, the demand for higher speed and performance drove semiconductor packaging advances to area-array packages, which brought applications for dam and fill, encapsulation and the first large production volume applications for flip-chip underfill. As the market demands more performance in terms of smaller size, lower power, lighter weight, greener components, higher functionality and higher speed, the applications for more types of fluids, adhesives, coatings, encapsulants, pastes and fluxes increase.

In the past, the application of fluids was looked on as a process to eliminate because of the tough nature of applying fluids in the assembly process. Today, fluids are becoming integral parts of the electronics package as evidenced by liquid crystal displays (LCDs), biotechnology labs on a chip, lenses, and MEMS devices. The transition of wire harnesses to PCB to SMT to system-in-package (SiP) modules drive more new fluids and applications. Similarly, there is an evolution in the technology for applying fluids from needle/contact methods to jetting/non-contact methods. This transition is as important in our niche market as the transition from contact/pin printing to inkjet printing for computer printers. Change is inevitable, and new customer requirements drive the automation market to innovate and develop new technologies, such as jetting fluids used in assembly.

Types of jets

Thermal inkjets are the most common jets available today. The device is very appropriate for the consumer market because it can be manufactured in volume with semiconductor and thin-film techniques at very low production prices. The technology jets small dots of low-viscosity inks, particularly acceptable in the printing industry.

The history of thermal inkjets goes back to using thin-film resistors for thermal printing on special thermal paper. At one point, it was recognized that the hot resistor could be used to boil an ink and create dots. The thermal inkjet works by rapidly heating a small resistor to create a bubble at the resistor by initiating boiling, then turning off power to the resistor and allowing the surrounding ink to cool and the bubble to collapse. The action of bubble formation and subsequent collapse imparts momentum to the fluid directly above the bubble. Having a hole above the resistor or nearby provides an easy escape path for the fluid. Consequently, a dot of ink is expelled from the orifice. The advantages of this technology are the size of dots, speed of operation, insensitivity to entrapped air, low-cost manufacturing of heads, and the ability to package many nozzles closely together. However, the thermal technology is difficult to extend into other areas beyond inks. The technology works best with low-viscosity fluids below 30 cps. Also, the process of heating fluids to boiling creates many opportunities for unwanted chemical reactions in the fluid and at the resistor.

Underfilling a flip-chip in package, the Asymtek DispenseJet DJ-9000 jet shows how jetting provides a new way to apply adhesives and package electronics in smaller spaces.

This style of jet is typically limited to low-viscosity fluids below 30 cps. Also, since the deflections are small, any entrapped air in the chamber being squeezed by the piezoelectric significantly dampens the response and prevents material from being ejected from the nozzle. Piezoelectric technology is more chemically inert than thermal inkjets, and this author believes it has a significant advantage over thermal inkjets for applications beside ink.

Another practical way to create a jet is to rapidly open and close an orifice. In this method, a fluid is put under relatively high pressure (>0.2 mPa for fluids in the 30 cps range and significantly higher for more viscous materials), then an orifice is opened. A stream of fluid starts to flow out the orifice, which then closes. The rapid closure cuts off the flow, and the stream momentum carries the fluid away from the nozzle. Piezoelectric actuators attached to a lever system allow a rapid means of valve actuation. Rapid and repeatable actuation is required to accurately control the amount of material streamed out of the orifice. Also, to form small dots, small nozzles are required with higher pressures and faster actuations. This process is completely dependent on the fluid viscosity, and is similar to an air-over-valve actuation. This jet technology has found good success in applying ultraviolet-curable adhesives for electronics encapsulation. Jets of this type are available from Picodostec, Delo and Vermes.

Another new jet technology was recently introduced by Mydata. This technology uses a piezoelectric rod as an excitation device in a semi-closed chamber. The chamber is open to a supply line that feeds in solder paste from a rotary positive displacement pump (RPDP). As material is pushed into the chamber, the pumping action of the piezoelectric standing wave propels consistent-sized dots of material from the nozzle at rates up to 500 dots/sec. This type of piezoelectric jet is new to the industry, but seems promising in providing a new means of applying solder paste as an alternative to stencil printing in prototype or high-mix production lines. This technology has the advantage of having a positive displacement component because of the addition of the RPDP.

The mechanical jet works in another unique manner (Fig. 1). In this case, fluids are fed into a chamber at a relatively low pressure. Typically, underfill adhesives are pressurized at <0.1 mPa and lower-viscosity materials, such as liquid crystal, at 0.01 mPa. The mechanical jet is designed to mitigate flow out the nozzle under fluid pressure alone and eject a dot based on the nozzle size, ball size and seat geometry. The advantage of this technology is that it creates very high local pressures at the nozzle and is able to jet very high-viscosity fluids. The disadvantage is that the dot sizes are much larger than piezoelectric or thermal inkjets. However, the mechanical jet has found many applications in jetting adhesives and fluids typically found in electronics assembly, such as underfill, epoxy, flux, surface-mount adhesive, and liquid crystal. Almost every type of fluid used in electronic assembly has been jetted with this technology.

1. The mechanical jet is designed to mitigate flow out of the nozzle under fluid pressure alone and eject a dot based on the nozzle size, ball size and seat geometry.

Theory of operation

The supply pressure is used to refill the ball/seat area. The jet works by moving the ball away from the seat to allow fluid to fill the seat area. As the shaft moves up, it makes the first chamber larger; consequently, fluid flows into the chamber from the fluid supply. The jet nozzle is small enough and the supply pressure is great enough so that air is not drawn into the nozzle. The ball is then moved down rapidly with a known velocity to impact the seat. As the shaft moves down, fluid is displaced. Fluid in contact with the shaft moves with the shaft, but fluid in the center of the space between the shaft and wall of the chamber moves back toward the supply. This process continues until the ball approaches the seat. At the point just prior to contact with the seat, a volume of fluid is trapped in the seat and finds its only exit path out the nozzle orifice. The fluid pressures become extremely high, and fluid is jetted out of the orifice in a stream. However, the source of additional fluid is already cut off, and the final impact of the ball on the seat snaps the fluid stream.

Jet vs. needle

The various jetting technologies all provide advantages over needle dispensing by solving the needle's inherent weaknesses. The process of depositing material with a needle requires that the needle, substrate and fluid are all in contact with each other at the same time. Deposition occurs after the flow from the needle is stopped and the needle is extracted away for the surface. As the needle moves up, the break off and amount of material that remains on the surface and needle is uncontrolled. In the case of the jet, the fluid leaves the jet nozzle and impacts the surface. Therefore, one uncontrollable variable is eliminated; consequently, the reliability, repeatability and process window gets better.

Another advantage is that the small geometries of the jet orifice allow for small streams and relative high steam velocity (1.5 m/sec). The physics of flow in a nozzle tube and needle are the same (Equation).

However, the 100 µm jet nozzle has a length of 0.5 mm and can be positioned 2 mm away from a surface. An equivalent needle would be 32 gage and need to be >2.5 mm in length. Given the same fluid pressures, the jet delivers 5× more fluid. The jetting stream is unconstrained. The needle provides a casing all around the fluid to the point of delivery. Consequently, the position of the fluid is dependent on the position of the needle, which may be bent or out of the closed-loop position of the positioning robot. Furthermore, the needle may not wet the substrate consistently and cause the fluid to be biased to one side of the needle or another, contributing to additional positional errors on deposition. At this point, one must ask why the needle is bent, which alludes to the point that it hit something. This attribute to needle dispensing gave rise to "die clipping," which means the edge of a flip-chip was damaged by an errant needle. Jetting can never chip a die (Fig. 2).

2. A 100 µm jet nozzle has a length of 0.5 mm and can be positioned 2 mm above the surface of the die. An equivalent needle would need to be positioned down below the die edge, increasing the chances of clipping and chipping the die, shown in the above top view.

Jetting dots to make lines for seals provides advantages that are not initially obvious. A series of dots at the correct spacing will make an almost perfectly straight and non-scalloped line once compressed between two parts. As the dots are compressed, they initially make bigger circles as the fluid flows out equally. However, once the dots touch one other along the line axis, there is an equalization of flow symmetrically at the point of contact (Fig. 3). As the dots are further squashed, the fluid flows toward the unrestricted boundaries. At first, the boundary is scalloped. As the dots are compressed further, the fluid will take the shortest path, which is to the deepest scallop. Consequently, the flow equalizes along the free surface to a straight line. This has real advantages in making seals for OLED, LCOS and flat panel display (FPD) assemblies. Since the beginning and end of a rectangle or continuous shape made with equally spaced dots is unrecognizable, there is no beginning or ending blob in the dispensed line. This condition is next to impossible to achieve with needle dispensing.

3. Jetting dots merge on compression between two substrates to form seals (patent pending).

Enabling technology

All jetting applications have enabled new packaging technology. The ability to place fluids in smaller spaces provides new ways to use adhesives and package electronics in smaller spaces. Flip-chip on flex (FCOF) is used in many hand-held devices, cameras and hard disk drive assemblies. Flexible circuit substrates have been growing faster than rigid PCBs for the past few years, and will continue to find more applications in packaging.

In the case of small flip-chip die, the ability to apply underfill quickly enables the end product introduction. In order to meet cost goals, the manufacturer must meet high units per hour (uph) throughput on the production line. If a single 1 mm die required four dots of underfill, a jet could apply the dots at 200/sec, which equates to 50 pps. If a needle process was used at the high rating of 14 dots/sec, the throughput would be 3.5 pps. When considering the other parts of the process, jetting provides >3× the advantage in throughput — therefore, at least 3× the machine utilization.

The FPD assembly market has unique applications that use the inherent advantages of non-contact dispensing. FPDs are made on large panels of glass. The latest generation of glass called Gen 7 is a substrate 2 × 3 m. Even in much smaller panels, needles are traditionally used to dispense a bead of sealant around each display. Good needle dispensing requires the distance between the needle and substrate to be tightly maintained to make a good deposition. Complicated mechanics and servo controls are required to meet these demands. Fortunately, jetting does not require tight height controls. The jet applies fluid over substrates that vary 1 mm, therefore higher dispense speeds with less costly mechanics provide lower-cost manufacturing of displays.

In the application of underfill chip-scale packages, new challenges occur as the wireless applications become more prevalent. The new, sophisticated, high-end personal data assistant/cell phones usually require electromagnetic interference shielding. If a shield needs to be placed over a component that requires secondary underfill, the shield may not be applied until the underfill operation is complete. This would require an additional solder reflow of the shield onto the board, which is highly undesirable, expensive, and subjects the almost-finished product to undue stress. Fortunately, a small hole in the shield placed above a corner of the component requiring underfill will solve the problem. Since the jet can deliver high flow rates in the range of 20 mg/sec with 100 µm stream size, jetting the underfill material into the hole provides a viable production process. Depositing underfill in one location along the edge or corner of the part will provide adequate underfilling.

Another enabling technology is the process called one-drop filling (ODF). The name is somewhat of a misnomer because more than one drop is used. However, the usual way to fill the space between the glasses of an FPD was to apply a vacuum to one end of the display and draw the liquid crystal in from a supply tray. As screens got larger, this process could take a day. The new method is to jet the liquid crystal in an array and then laminate the display in a vacuum. This method increases the throughput of the process on large displays from 24 hours to 3 hours (Fig. 4).

4. Jetting is replacing time- consuming vacuum techniques for filling a flat panel display with liquid crystal. (Source: Shin-Etsu).

Conclusion

Jetting is rapidly becoming the standard method for applying the various fluids used in electronics assembly, semiconductor packaging and FPD assembly. It may be easy to predict that automated needle dispensing will decrease to niche processes in the future because of the inherent advantages of jetting and its enabling features. The small size, tighter tolerances, and high speeds of jetting allow newer and smaller products because the manufacturers can produce the higher-performance components economically.