Why Standard Materials Metrics Aren’t Enough
BY PATRICK TRIPPEL, Henkel
For decades, materials developers and formulators have used common, industry-standard measurement tools for assessing a given material’s properties and characteristics to provide some indication of performance. When examining die-attach paste properties, for example, metrics such as die-shear, viscosity, volume resistivity, and moisture absorption are often part of the analysis. While these have been the main metrics used for evaluation - in part because to-date they’ve been the only thing available - these materials-characteristic indicators aren’t enough to predict in-process and in-field performance.
Examples of this disconnect between material properties and performance are countless, but this discussion will focus specifically on die-attach materials. When packaging specialists are trying to determine the ideal die-attach material for a particular application, a viscosity analysis is often used as a measure for the material’s dispensability. The thixotropic index measure is commonly implemented as the indicator of viscosity and, therefore, dispensability characteristics. But, in reality, the performance correlation between viscosity and dispensability is very weak. The viscosity of a die-attach material only predicts the dispense rate for a consistent PSI (pounds-per-square-inch pressure) using a specific-sized dispensing needle. So, what happens if the dispensing tool changes? Is the dispense rate the same? Absolutely not, and that’s why these dated metrics should be closely examined. When considering the variety of different dispensing tools on the market and, therefore, the different rates with which material can pass through these systems, it becomes painfully obvious how misguided it is to use these indices as performance prognosticators. Plus, it’s not just about dispense speed: other critical performance characteristics including material tailing, clean breakaway, and nozzle clogging cannot be predicted by viscosity alone. So, a lower viscosity may not always mean a faster and better dispensing result. In fact, recent research has shown that when optimized chemistries and fillers are used, even materials with higher viscosities may dispense more quickly than those at lower viscosity - and with superior results.
Die shear is another common measure used in the packaging industry - a metric that is often implemented to predict the die-attach material’s bond strength. Again, this is a flawed approach to understanding the die-attach performance in process. Although the die-shear test will indicate how strong a material bond is for various-sized die, keep in mind this test generally only evaluates die-to-substrate bond strength. What happens when the mold compound is added and the package moves through reflow? In many cases, die-attach materials that show good die shear have failed JEDEC testing once they are evaluated in-package. If the bond strength is good, but the package comes apart in reflow, a good die-shear metric is meaningless. Package assemblers may try to predict overall package performance by compiling the results of die shear, volume resistivity, modulus, and moisture absorption among other common metrics and often make the judgment that good results in all of these categories equate to good performance. Unfortunately, that’s usually not the case.
I’m certainly not suggesting that all of these metrics should be thrown out the window. These various tests and material-properties measures do provide useful analysis for materials production and are excellent tools to use for consistency validation during the materials manufacturing process. However, results from these industry-standard tests should absolutely not be interpreted as performance indicators. There is only one way to evaluate how a given material or a set of materials will perform in-package and in the field: build the package. Analysis of these materials in the context of their ultimate environment is the single best way to get meaningful results. Because materials interactions and various process conditions can impact a set of materials in many different ways, only a thorough evaluation of performance within a manufacturing environment will provide useful data.
If your materials supplier can’t provide you with proven, tested in-process performance results for their materials, you should be shopping for another materials partner. Only those materials companies that make the investment in the tools necessary to build devices and analyze materials performance within a production environment should be worthy of your business. After all, your customers are counting on products that work; shouldn’t you?
PATRICK TRIPPEL, president, electronics group, may be contacted at Henkel Technologies, 15350 Barranca Parkway, Irvine CA 92618; 949/789-2556; E-mail: firstname.lastname@example.org.