High mechanical strength, low electrical resistance, long pot life and stability through temperature are vital features of any conductive epoxy used in aerospace quality hybrid microcircuits.
Microcircuit Engineers at Sperry Flight Systems assessed a number of conductive epoxies to choose one that best satisfied these requirements. Epo-Tek's H20E easily matched all of the electrical and mechanical performance criteria and was designated as the primary die attachment material for hybrid microcircuits.
Epo-Tek H20E epoxy is a silver-filled, two component system which cures at 150 °C within five minutes. It is a thixotropic paste which can be stamped, dispensed or screen printed. The manufacturer claims that the two components may be blended either by equal volumes or by equal weights.
Sperry's first production shipment of H20E was received in pre-measured bi-packs. Incoming material testing revealed that the lap shear strength of this early production shipment was considerably lower than the evaluation samples had demonstrated. Additional investigation revealed that, while the evaluation samples had been mixed by equal weights, the bi-packs contained equal volumes of the two components. Subsequent testing proved that epoxy mixed by equal weights had a considerably higher lap shear strength than that mixed by equal volumes.
Concern developed over the reliance of other properties, such as electrical resistance, on mix ratio. A study was conducted to relate mechanical and electrical performance to mix ratio so that suitable operational and material controls could be imposed in the manufacturing area. Furthermore, the effects of pot age (the time expired between mixing and using) were analyzed.
This article determined that:
- Mix ratio greatly affects both electrical and mechanical performance.
- Pot age of up to one week has negligible effects on electrical performance and mechanical strength, but does have an effect on handling features.
This article discusses the specific relationship of both mix ratio and pot age to electrical and mechanical performance.
Mix Ratio Effects
Seven mix ratios of resin (Part "A") to hardener (Part "B") ranging from 1:2 by weight to 2:1 were tested. This range includes both equal volume and equal weight mix ratios. The seven mix ratios shown in the table below were analyzed electrically and mechanically.
(Weight "A" to Weight "B")
|Weight % Part "A"
||Volume % Part "A"
For a number of reasons, the collector-emitter saturation voltage (VCEsat) on 2N2222 transistors was used as a measure of the electrical performance of the epoxy. This saturation voltage is a strong function of the quality of the chip attach medium. The VCE sat method measures all the effects of a bond, including the bulk resistance and interfacial effects. Besides being simple to measure, VCEsat is highly sensitive to minor changes in bond resistance. Over the years, Sperry Flight Systems has collected a large amount of VCEsat data associated to chip attachment techniques.
Twenty five 2N2222 transistors were linked to thick-film gold substrates with each of the seven mix ratios of epoxy. All were wire bonded thermosonically with 0.001" diameter gold wire and hermetically sealed in 16-pin DIPs, five transistors per package. A photograph of this construction can be seen in Figure 1. All testing was conducted with a collector current, Ic, of 150 ma and a base current Ib of 15 ma.
Figure 1. Dip circuit used for VCEsat tests. Five 2N2222 transistors were bonded in each package.
VCEsat measurements were done at the following process points:
- After environmental screens (screening included temperature cycle, stabilization bake, centrifuge, fine leak and gross leak testing in compliance with MIL-STD-883A)
- After 168 hours burn-in at 125 °C with power on the chips.
- After 500 hours burn-in at 125 °C with power on
- After 1000 hours burn-in at 125 °C with power on
Electrical Test Results
All VCEsat data is shown in terms of resistance. Preliminary measurements, captured at pre-seal, were all approximately 1 ohm. However, as each group of mix ratios was exposed to screening and temperature, those samples containing over 50 % resin, by weight, considerably increased in bond resistance. Those samples with 50 % or less resin were very stable throughout the entire 1000 hours at temperature. Figure 2 illustrates the highest average bond resistance that occurred throughout exposure to temperature for each mix ratio. The numbers in parenthesis indicates the time and temperature at which the average resistance of any particular mix ratio was the highest. Figure 3 illustrates the bond resistance history through temperature for each mix ratio sample.
Figure 2. Bond resistance as a function of epoxy mix ratio. Mix ratio has a pronounced effect on bond resistance.
Figure 3. Bond resistance as a function of time at temperature. Samples containing more than 50% resin exhibited resistance increases with time at temperature.
The uniformity of bond resistance from transistor to transistor within a mix ratio group was also a robust function of mix ratio. Figure 4 shows the reliance of resistance uniformity on mix ratio. Those samples having 50 % resin or less are extremely tightly grouped throughout temperature exposure, while those having over 50 % resin display a great deal of non-uniformity.
Figure 4. Bond resistance uniformity as a function of time at temperature.
Tests for screenability, lap shear strength and dispensability were conducted.
Four lap shear test samples of each mix ratio were examined after curing at 150 °C for 30 minutes in a dry nitrogen atmosphere.
Screenability of each mix ratio was analyzed by screen printing through a 200 mesh patterned screen onto thick-film printed alumina substrates. Printing operators noticed features such as ease of printer set-up, print thickness, print definition and voids in the printed pattern.
Dispensability samples of each mix ratio were dispensed onto thick-film printed substrates using a commercial epoxy die bonder. Operators detected features such as the repeatability of dot size and the control of the size of the epoxy dot.
Mechanical Test Results
Figure 5 illustrates the results of the lap shear tests. In the lower mix ratio ranges, lap shear strength is sharply reliant on mix ratio. Shear strength is comparatively stable in the upper mix ratio range. Note that the "equal volume" mix ratio falls within the steep portion of the curve. This describes the low lap shear strength issue with the early production shipment referenced earlier.
Figure 5. Lap shear strength as a function of mix ratio. Lap shear strength is highly dependent on mix ratio.
Mix ratio also considerably affected dispensability and screenability. Samples having less than 45% resin displayed very poor screening qualities, including stringiness, poor print definition and printer set-up problems. All the other samples tested showed satisfactory printing features. Dispensability was less impacted by mix ratio. All samples were dispensable, though the lower ratios were somewhat more difficult to control.
Irrespective of the application technique, the lower ratios displayed substantial vehicle bleedout. This bleedout can cause wire bonding issues, mainly on thin-film gold since the texture of the gold tends to "wick" the vehicle out by capillary action. On thick-film gold, the pockets and pores in the gold die bond pad are inclined to collect the vehicle and halt its journey to wire bonding pads.
Summary of Mix Ratio Results
Mix ratio has a certain impact on the vital performance properties of H20E, specifically
- Lower mix ratios display superior electrical performance in terms of resistance, stability through uniformity and temperature
- Ratios above 1:1 by weight display higher average bond resistance, are more suitable to change with time at temperature and are less unvarying from bond to bond
- Lower mix ratios are inclined to have a lower shear strength and tend to present more handling problems. In the lower mix ratio range, lap shear strength is sensitive to minor variations in mix ratio.
- The higher mix ratios display a higher, more stable shear strength and display acceptable dispensability, screenability and bleedout features.
Pot Age Effects
The pot age of the epoxy is defined as the amount of time expired mixing the two components and using the mixture. Eight samples, with pot ages ranging from freshly mixed to one week old were tested electrically and mechanically. All pot age samples were mixed 1:1 by weight. Two extra "special case" samples were made to mimic unusual production circumstances. One sample was aged after placing the dice in the epoxy and the other sample was aged before putting the dice in the epoxy.
As in the mix ratio tests, VCEsat was used as a measure of electrical performance. Eight samples of epoxy were mixed and allowed to age at room temperature in a dry nitrogen atmosphere for the definite period of time, ranging from one hour to 168 hours, after which 25 2N2222 transistors were bonded with each sample and cured. Also, two samples were produced to mimic unusual production conditions. The first of these, identified here as S-l, was assembled by positioning the transistors in fresh epoxy, then allowing 80 hours to expire (corresponding to a long weekend) before curing the parts. For the second of these special samples, S-2, the epoxy was placed on the substrate fresh, then 80 hours later the transistors were positioned in the epoxy and the assembly was cured. All storage was conducted in dry nitrogen.
Substrates were brought together in 16-pin Dual-in-Line hermetic packages and subjected to the same screening and testing processes as the mix-ratio samples.
Electrical Test Results
Pot age had slight effect on the electrical performance of the epoxy samples. Figure 6, a plot of the maximum average resistance through temperature as a function of pot age, reveals bond resistances steadily below 1 ohm. Figure 7, a plot of bond resistances of all samples as a function of time at higher temperature, specifies that pot age has little impact on resistance or stability of the epoxy bond.
Figure 6. Bond resistance as a function of pot age. Bond resistance is independent of pot age.
Figure 7. Resistance as a function of time at temperature. All pot age samples exhibited stable bond resistance through time at temperature.
Lap shear strength tests were carried out using the same samples as those used in the electrical tests. Screenability and dispensibility were also analyzed.
Mechanical Test Results
Figure 8 illustrates that there does appear to be some reliance of lap shear strength on pot age, but at all ages tested it stayed well over 1000 psi and even reached 2000 psi for a good portion of the samples. The special samples S-l and S-2 showed sufficient strength. The 120 hour sample and the 168 hour sample were adhesive failures with the remainder being cohesive failures.
Figure 8. Lap shear strength as a function of pot age. Lap shear strength is relatively independent of pot age.
Samples aged less than 72 hours were considered screenable and dispensable. The 72 hour sample was dispensable but offered weak screenability. The 96 hour and older samples had room temperature cured excessively to be screenable or dispensable in a production environment. Screenability and dispensability are subjective properties, thus each user must make his own decisions based on his specific production needs.
The mix ratio of H20E has a great effect on its electrical performance. Mixtures having 50 % or less resin by weight display very low, very stable bond resistances while those having greater than 50 % resin act rather poorly, both primarily and after exposure to temperature. The mechanical strength of H20E increases with mixtures having 50 % or greater resin. Mixtures having less than 50 % resin display reduced strength. Handling features such as screenability and dispensability are sensitive to mix ratio, but are poor only in the lower mix ratios.
It is seen that for the best total performance, H20E should be mixed 1:1 by weight. An equal volume mixture provides less strength than an equal weight mixture, though electrical performances are the same.
The electrical performance is essentially independent of pot age. Though pot age does impact the handling features, with the older epoxy being more hard to handle, it does not significantly impact the mechanical strength.
The results of this article have enabled Sperry Flight Systems to write important specifications to control the incoming material. Epoxy Technology has also been able to execute controls on manufacturing procedures that enhance the use of the epoxy on the production line without conceding reliability.
This information has been sourced, reviewed and adapted from materials provided by Epoxy Technology, Inc.
For more information on this source, please visit Epoxy Technology, Inc.