Environmentally safe molding compound
11/01/2001
Evaluation of new flame retardants
BY RODEL MANALAC, BEN CHEAH AND JOEL ALIMAGNO
A recent development in the campaign for environmental and health awareness is the introduction of environmentally safe flame retardant systems for plastic encapsulants in the electronics industry. Flame retardant agents, like antimony (Sb) and halogens that are used in conventional molding compounds, are considered to be environmental hazards. Such elements are eliminated in new "green" molding compounds and are replaced by an environmentally friendly flame retardant system, such as magnesium hydroxide.
This article discusses an evaluation of several antimony- and halogen-free molding compounds. A number of industry standard tests were adapted to evaluate the reliability performance, and due consideration was given to the properties and characteristics of these green molding compounds.
Molding Compound Evaluation
A study on these new molding compounds was performed to understand compound behavior with alternate flame retardant components. Two types of Br/Sb-free compounds with slightly differing flame retardant systems were chosen (referred to as compound A and compound B).
 Table 1. Storage test conditions for evaluating new molding compounds. |
The reliability test for this project involves compound adhesion, electrical test response and moisture sensitivity level, which can all be affected by the following factors: package size; leadframe design and coating; die attach materials; die size; die coating and passivation; and die pad size, shape, and locking features.
To qualify a molding compound that can pass environmental tests and conditions is a major challenge because the material must meet requirements for package integrity, processing and performance con-sistency before it can be recom-mended for pro-duction use. The assessment includes verification that the material withstands a reflow peak tem-perature of 260°C. This project aims to attain the same or better moisture sen-sitivity level with a reflow temperature of 260°C.
Test Methodology
The following process steps were performed during the entire evaluation: material preparation; die attach; wire bond; encapsulation; post-mold cure (PMC); dambar cut; plating; marking; forming and singulation; and reliability testing (preconditioning at level 3 - reflow at 260°C).
 Figure 1. Typical moisture absorption trends in mold compound A and mold compound B for the 14x14x1.4 LQFP. Results were similar for other package sizes. |
Samples of lead quad flat pack (LQFP) packages (leadframe material C7025) were molded with compounds A and B and then submitted to the storage conditions as shown in Table 1. The samples were subjected to reliability testing. Scanning acoustic microscopy was conducted before and after JEDEC level 3 preconditioning to check for package delamination. A total of 420 units per package type were subjected to a pressure cooker test, temperature cycling, high temperature storage and liquid thermal shock. Electrical tests were done after preconditioning and at each read-out point of the reliability tests.
 Table 2. Molding and reliability results. |
Processability and moldability: The results in Table 2 show that the wires are able to withstand the flow of the molding compounds with a maximum wiresweep significantly below 5 percent. In addition, no internal or external voids were detected for either compound A or compound B.
 Figure 2. The desorption trend for mold compound A and mold compound B for the 14 x14x1.4 LQFP. The results were similar for other package sizes. |
Reliability results: No delamination was observed on the die top after JEDEC level 3 preconditioning at a reflow temperature of 260°C. However, the pad bottom and pad top areas were found to have delamination before and after preconditioning for all package types molded on the two compounds.
Electrical test results: Electrical tests were done after preconditioning all units to the level 3 condition. No failures were seen on any units, and all reliability read-out points passed electrical test. All packages were able to pass electrical testing after undergoing 168 hours of pressure cooker test, up to 1,000 cycles of temperature cycling, up to 1,000 cycles of high temperature storage and up to 1,000 cycles of liquid thermal shock.
Failure Analysis
Failure analysis was conducted on those units that had delamination. Units were decapped and found to have wire bonds still intact - no broken or cracked wires were found in the delaminated area. Thruscan (which differs from normal CSAM in that it can detact delamination on all interfaces at one time) confirmed delamination at the pad interface. Cross-sectional analysis showed microscopic gaps between the mold compound and the pad interface.
Moisture Absorption and Desorption
Moisture absorption data was obtained by baking 25 units of each compound for each package at 24 hours, and then soaking them at 85°C/85%RH for 192 hours. They were weighed at 24-hour intervals, and a plot was constructed to show the moisture absorption trend. The trends for both of the compounds are comparable, with maximum moisture levels of 0.15 to 0.16 percent for LQ7x7x1.4; 0.13 to 0.15 percent for LQ14x14x1.4 and 0.17 percent for LQ28x28x1.4 packages. All three packages show similar absorption trends, as shown in Figure 1.
 Figure 3. Adhesion strength comparison. |
After the saturation period, units were subjected to a moisture desorption chamber. Weight loss was then measured by placing all of the packages inside the baking oven at 125°C for 48 hours. Results show that within four hours of desorption, the moisture content in all of the packages with both compounds was reduced to the standard requirement of 0.04 percent moisture level, as shown in Figure 2. The moisture content after 24 hours was negligible.
Mold Releasability Test
Mold releasibility tests determined the adhesion properties of each compound, which affects the ability of the mold compound to avoid sticking the mold cavity. Results showed that compound A required an average of 1.12 kg compared to 1.88 kg for compound B to be separated from the mold cavity. Both of these readings are relatively low in that normal acceptable range is between 2.0 and 4.0 kg.
Adhesion Strength
 Figure 4. DSC plot showing percentage of cure vs. cure time. |
Adhesion strength tests were conducted to determine the adhesion of both compounds to different materials, including copper, silicon and silver plating. The results are summarized in Figure 3. Compound A shows higher shear strengths than compound B.
Compound Cure Time Response
Differential scanning calorimetry was done to determine the percentage of cure of the compounds. Figure 4 shows the consistency of both compounds in achieving 98 percent cure even at two hours of post mold cure.
Glass Transition Temperature
Measurement of the compounds' Tg (glass transition temperature) was performed using thermo-mechanical analysis (TMA). Figure 5 shows a Tg of 120°C or higher for both compounds with 0 to 6 hours of PMC, except for compound A which has a Tg of 108°C at 0 hours PMC. The graph also shows an indirect correlation of crosslink density versus PMC hours, with an increase in PMC time resulting in an increase in crosslink density.
Discussion
The two molding compounds that were evaluated show no process issues. Both compounds passed electrical test on all reliability read-out points. However, based on CSAM results, both compounds exhibit delamination after JEDEC level 3 preconditioning at 260°C reflow. The units were confirmed to have microscopic gaps propagating from the pad and compound interface. The delamination occurring on the pad interface might affect devices having ground bonds and downbonding.
 Figure 5. Glass transition temperature evaluated by TMA for different cure times. |
Both green compounds meet the standard requirement of 0.04 percent moisture level for moisture absorption/desorp-tion tests. Compound A has better adhesion properties, requir-ing less force for release from the mold cavity while retaining high adhesion strength with the relevant packaging materials.
Conclusion
Based on these results, the two types of compound passed the material qualification test at L3 preconditioning at 260°C. However, continuous improvement must be focused to address the presence of delamination above and below the pad. Molding compound suppliers were informed of the results of the green compound evaluation and encouraged to improve the current compound properties. Evaluations on the performance of these mold compounds are continuing. AP
Rodel Manalac, engineering manager, Ben Cheah, mold process engineer, and Joel Alimagno, senior process engineer, can be contacted at ST Assembly Test Services (STATS), 5 Yishun Street 23, Singapore 768442; 65-824-1278; E-mail: [email protected] or [email protected].