Mold Compound
10/01/2003
High Performance Requirements
By Philip Procter
Mold compounds are the plastics used to encapsulate many types of electronic packages, from capacitors and transistors to central processing units (CPUs) and memory devices. The modern mold compound has evolved into a complex formulation containing as many as 20 raw materials and multiple processing steps, each statistically controlled to yield a uniform and predictable end product.
At the most basic level, mold compounds contain five classes of raw materials. Organic resins are typically meltable, such as epoxy resin. Fillers are non-melting inorganic materials. Catalysts accelerate the cure reaction. The mold release material allows the naturally adhesive epoxy resin to come out of the mold. The final raw material is a pigment or colorant. Other materials, such as flame retardants, adhesion promoters, ion traps and stress relievers are added to the mold compound as appropriate.
Mold compounds have evolved over the years to keep pace with industry needs. Each innovation in chip or package design required a similar change in the design of the encapsulant. In 1969, encapsulants typically were filled with fused silica at about a 68 percent w/w loading. Currently, materials are filled with 90 percent fused silica. This shift was not made for reasons of cost, but driven by end-user performance requirements. Over the last decade, every raw material and process in mold compounds has been re-examined — and almost all have seen major changes.
Balanced Properties
Mold compounds must meet 20 or more clearly defined performance requirements (and many others that are less clear). Change in any one formulation property affects overall performance and requires a compensatory shift elsewhere in the formulation. One example of such a property would be filler content.
As chip sizes increase, reducing the thermal expansion (CTE) of the mold compound will minimize internal stresses. Since the CTE of fused silica is 1/10 of epoxy resin, a simple increase in the filler loading will reduce CTE (Figure 1).
 Figure 1. A simple increase in the filler loading will reduce CTE. |
null
However, increasing filler content also degrades the flow and modulus of the material (Figure 2). In response, formulators have adjusted other ingredients to maintain flow and modulus, for example, by lowering polymer melt viscosity. Unfortunately, lowering polymer melt viscosity reduces process leeway.
 Figure 2. Increasing filler loading reduces CTE, but at the same time, this also degrades the flow and modulus of the material. |
null
With each change in chip design, package design, mold process or test requirement, formulators must meet the new performance requirement, evaluate the product and then begin the rebalancing process.
The Latest Transition
In response to environmental concerns, the industry is moving rapidly to eliminate lead solders from electronic equipment. One characteristic common to tin/lead replacement alloys is a significantly higher melt eutectic. Board soldering temperatures are rising to 245° or 260°C. Certain device types, such as stacked package configurations, are requiring more than 300°C solder resistance. The stresses placed on the package during these brief thermal shocks are significant, leading to weakened packages and immediate, catastrophic failure.
Failure occurs after molding, when the encapsulant absorbs a small amount of moisture to reach equilibrium with its environment. During thermal shock, any moisture present in the package — especially at the interface between the plastic and the substrate — will turn to steam and exert considerable internal pressure. If the pressure exceeds the strength at any point in the package, permanent delamination occurs.
While package design improvements and toughened encapsulant formulations minimize gross package cracking, less visible, internal damage still occurs due to temperature shocks. The damage can occur between any dissimilar interfaces, the most critical of which are 1) mold compound to lead frame or laminate surface, 2) mold compound to die surface, 3) die-attach to lead frame, or 4) plastic resin to filler particle.
To research and eliminate these failures, mold compound formulators depend on sophisticated techniques and instrumentation, from high-frequency C-SAM to specially designed lead frames and high-powered modeling programs.
Rebalancing the Encapsulant
Several factors can improve the package's resistance to failure. Since Joint Electronic Devices Engineering Council (JEDEC) moisture preconditioning (per J-STD-20B) is done to near saturation, limiting the total amount of moisture absorbed by the plastic is the most obvious solution. Likewise, controlling the diffusion rate of moisture into the encapsulant reduces the moisture content at JEDEC test point, but not at true saturation, which will remain unchanged. This may also reduce the effect of moisture on other properties of the plastic.
Increasing the adhesion of the encapsulant beyond the power of the steam pressure at the same interface will eliminate delamination. Reducing thermal expansion mismatches generated across the package at reflow temperature will minimize internal package stress.
Reducing the package's moisture absorption limits the volume of water present and affects the steam pressure generated. As the encapsulants absorb moisture, physical properties such as strength and glass transition temperature (Tg) degrade.
The majority of encapsulants are based on the reaction of epoxy cresol novolac with phenol novolac. Together, these two polymers comprise the majority of the organics in an encapsulant composite. Changing the moisture content of the encapsulant means changing the reaction product by 1) changing the catalyst, by 2) changing the epoxy resin or novolac hardener to a more hydrophobic resin such as dicyclopentadiene (DCPD) or MAR, or 3) reducing the volume percent of resin by increasing filler content. Filler is predominantly fused silica (amorphous SiO2), which absorbs almost no water.
Bond strength at any interface is based on the surface properties of the substrate and encapsulant. If conducted under controlled conditions, direct adhesion measurements can provide a valuable comparative design tool during development. Valid measurement methods include lap strength, die shear, puck shear, contact angle and lead frame tab pull. While C-SAM often is called an adhesion test, it actually looks for delamination.
Encapsulants form both physical and chemical bonds to the surface of the substrate. In a typical package, several different substrates coexist, including the lead frame, the die or passivation, the die-attach fillet, and the die pad, which is often coated with splatter or outgassing traces from the die-attach.
Epoxies are capable of wetting many high-tension synthetic surfaces and forming strong, stable bonds to metallic oxides. An epoxy's bond to copper oxide often is stronger than the bond between the oxide and the base metal, resulting in oxide delamination during stress. The delamination is visible by C-SAM, but the determination of exact failure mode requires additional surface chemical analysis — single in-line memory modules (SIMMS), energy-dispersive X-ray analysis (EDX) or similar. Physically, as with most adhesives, mildly rougher surfaces form stronger bonds.
While using an encapsulant formulated for maximum adhesion to copper only, a change to the lead frame plating can alter adhesion by a factor of 50, creating process problems.
Adhesion technology can be tailored to the specific requirements of a device by adjusting highly reactive additives, polymer viscosity and reaction rates capable of significant increases in adhesion to specific substrates. Figure 3 compares an otherwise identical mold compound that has been modified for improved adhesion to nickel.
null
A mere 45°C increase in solder temperature can change the integrity of the package. Compounds capable of withstanding 215°C temperatures will fail at 260°. In some cases, even the increase from 245°C to 260°C can cause catastrophic failure. The glass transition temperature (Tg) is the temperature at which the polymer changes from a glassy solid to a rubbery solid. Due to the variation in an encapsulant's molecular structure and blended composition, the Tg actually is a range of 15°C to 30°C. Though real reductions in modulus occur at Tg, encapsulants are never rubbery.
By itself, a reduction in modulus would aid thermal shock resistance. Unfortunately, other changes occur in the polymer as well. The coefficient of thermal conductivity (CTE) can increase by a factor of five and the physical strength can decrease by a factor of 10. While effective at modulus modification, low Tg stress relieving additives can exacerbate CTE and physical strength changes.
Changes that occur when passing through Tg are unavoidable. The magnitude of change can be minimized by using the right encapsulant formulation. Is lower Tg or higher Tg the better formulation? The answer is that either can be, in the right circumstance.
Typical mold compounds have Tgs of about 150°C, while newer, low-stress chemistries can dip below 100°C. Lowering the Tg extends the range between the Tg and the reflow temperature, effectively forcing more of the temperature shock to occur above Tg, where the modulus is much lower but the CTE is higher.
Increasing the Tg to a temperature higher than the reflow temperature minimizes thermal expansion mismatches and internal strains, but can lead to a more brittle polymer and higher moisture absorption, due to the increase in free volume. Once limited to 200°C, encapsulants are now appearing with Tgs exceeding 230°C. When combined with the intrinsically stable RF frequency dielectric properties of an anhydride-cured epoxy, this approach may offer a unique solution in GHz power and hybrid application.
Conclusion
No approach is ideal for all applications. Mold compound formulators have become experts at adapting the performance of their product to suit the unique demands of each application. Changes in design or process, such as a small increase in reflow temperature from 215°C to 260°C, should be discussed early in the process to ensure that the delicate balance in the end product is maintained.
PHILIP PROCTER, technical service laboratory manager, may be contacted at Henkel Loctite Corp., 15051 E. Don Julian Rd., Industry, CA 91746; e-mail: [email protected].