Abstract

Improvements in IC mold compounds and Die Attach materials have enabled improved Moisture Sensitivity Level (MSL) classifications at higher temperatures. As an example, a 48 pin HPQFP that had been MSL classified at 5a (24 hour floor life) for 235 Degree C reflow has now successfully passed MSL3 (168 hours floor life) for 260 Degree reflow. To achieve this level of improvement, both the Mold compound and Die attach needed to be changed. These materials are highly specialized compounds which are designed to perform a number of (often competing) tasks. As the properties needed for higher temperature and better MSL performance are improved, a number of trade-off's are inevitable. This paper deals with a number of these trade-offs, and the changes that were necessary to successfully implement these new materials. Specifically, the thickness of the Die Attach material and the "Resin Bleeding" performance of the leadframe surface were found to be key control items.

Article

The industry standard test method for moisture sensitivity is JEDEC J-STD- 020 "Moisture Induced Stress Sensitivity for Plastic Surface Mount Devices" (MSL test). This document outlines the stress conditions, failure modes, and suggested analysis techniques for detecting moisture induced damage in IC packages due to Soldering Printed Circuit Board (PCB) assembly. In addition to electrical testing, acoustical imaging is an extremely valuable, non-destructive tool to see inside the packaged device and understand if damage has occurred due to the stress condition. The standards for this procedure are contained in JEDEC --STD-035 "Acoustic Microscopy for Non hermetic Encapsulated Electronic Components". Delamination can be detected by a number of different techniques. Each technique has advantages & disadvantages, and it is up to the analyst to use the appropriate technique for the intended purpose. Especially when looking at new materials and processes, it is important to understand the basic tests and requirements, and the relationship of these to the finished product. The MSL test is designed to evaluate damage that occurs as a result of the soldering stress exposure - it specifically excludes initial delamination as a failure criteria. This is an important distinction, because it was discovered that the materials which achieve the best MSL performance are also more prone to have initial delamination. Although existence of initial delamination may not fail the letter of the J-STD-020, it should be considered as a failure for purposes of materials development. Thru-Scan acoustical imaging is especially useful in addition to more traditional Reflected imaging when the delamination plan is unknown. In this case not only was topside die surface of interest, but the backside surface as well.

Die Attach has typically been Polyimide or Epoxy based adhesive chemistry, often with Silver as a high conductivity filler material. Both chemistries are relatively "mature" and have properties which can be optimized to achieve a balance of purity, cure properties, adhesive strength, bulk strength, and thermal stability. Newer developments such as Acrylates and Liquid Bismaleimides (BMI) are available for IC assembly. These allow hybridized materials to achieve an even wider range of properties2. In the quest for shorter cure cycles, the "Free Radical" cure system has also been developed which can allow elimination of the traditional oven cure. The properties of a BMI/Acrylate based, free radical cure Die Attach gave the best MSL performance during preliminary tests (see table 1) and was chosen for production implementation. This material exhibited the second highest room temperature shear strength, and the highest, High Temperature strength after 85 Deg C / 85& RH exposure, and has very low moisture absorption.

Table 1 data:
MSL delamination @ 260 C
NiPdAu Silver Bare Cu
Epoxy Die Attach 50% 40% 94%
BMI/Acrylate 3% 23% 0%

The Lead(Pb)-Free product was targeted to use a NiPdAu plated leadframe, so the BMI/Acrylate showed a step-function improvement in delamination. The Die Attach change produced the most significant increase in MSL performance. The higher shear strength (~ 2x) of the new material in conjunction with a higher modulus generates a higher stress at the Die to Leadframe interface. The TCE difference between the L/F (CU) and the Si die is ~ 13 ppm per Deg C. Since the Die Attach cure temp is 220 Degrees C, this sets up a significant stress between the two when cooled to room temperature after D/A cure. Finite element modeling predicts that the peak shear stress (in the corner of the die) will be about 80 MPa at 1 mil Bond Line Thickness. Since actual testing showed that the epoxy could withstand only about 14 MPa, and the BMI/Acrylate only ~ 25 MPa before fracture at room temperature, it is clear that the bulk and surface adhesion stress on the Die Attach will be critical, and there is a risk of the die "popping off" as the die & Leadframe cool after cure. The higher strength or the BMI helps in this regard, but it also has a higher modulus, which results in a higher stress level. The cool down ramp rate after cure must be controlled to allow for some "creep" to occur in the Die Attach and relieve a portion of this stress. Bond Line Thickness is also a key control factor, and a minimum bond line thickness was specified to further reduce the potential die to Lead frame stress. Although these phenomenons are not new to this industry, the sensitivity occurs for smaller dies with the stronger adhesives or higher peak cure temperatures. A cool down ramp rate of 2-3 Degrees per minute to 80 Degrees, and a minimum Bond line Thickness of 1.5 mil (.0015") was specified for this application.