SynopsisStress relaxation has been studied in networksof dihydroxy-terminated polybutadiene (mostly cis:trans:uinyl = 34:4026) crosslinked by triphenyl methane-4,4',4"-triisocyanate and containing about 9.5% by weight of unattached linear random styrene-hutadiene copolymer with various molecular weights (from 1.4 to 3.3 X lo5) and with styrene content and butadiene microstructure chosen to match the average solubility parameter of the end-linked network. Stress relaxation measurements were made also on networks containing no unattached species and containing 9.3% hydrocarbon oil, and on the various uncrosslinked linear polymers. The stretch ratio was 1.25 and the Young's relaxation modulus was calculated from the neo-Hookean stress-strain relation. For the uncrosslinked linear polymers, the relaxation modulus E l l ( t ) corresponds to a rather narrow distribution of relaxation times whose magnitudes were approximately proportional to the 3.4 power of viscosity-average or weight-average molecular weight; for one polymer, the time dependence agreed closely with the prediction of the Doi-Edwards theory modified for a small degree of molecular weight distribution. The disengagement times calculated from the Doi-Edwards theory as modified by Graessley appeared to be of the correct order of magnitude. The contribution of the unattached species in the networks E l ( t ) was calculated by difference; after multiplication by (1 -u%)-', where uz is the volume fraction of network, and correction for the difference in monomeric friction coefficient associated with the difference in fractional free volume in the two environments, E l ( t ) was compared with E l l ( t ) for each linear polymer. The relaxation was slower in the network than in the uncrosslinked polymer by about an order of magnitude, but the form of the relaxation modulus was similar in both environments except for two linear polymers for which the relaxation in the network became very much slower at long times. This behavior appeared to he correlated with a broader molecular weight distribution.
Eight commercial semiconductor grade epoxy compounds that are used to encapsulate 1C (integrated circuit) devices have been evaluated for their ability to minimize the development of thermal stresses which can cause failure during device temperature cycling. Thermal expansion, dynamic modulus and adhesion studies are used to describe the mechanical interaction between the plastic package and the silicon device it surrounds. A “figure of merit” is defined for the development of stress on the 1C device as it is cooled after the packaging process. The stress is shown to be proportional to the product of three terms: (αp‐αs) Ep (Tanch‐T) where αp and αs are the expansion coefficients for the plastic and silicon, respectively, Ep is the modulus of the epoxy and Tanch is the temperature at which the epoxy becomes anchored to the silicon device during transfer molding. In addition, the importance of good adhesion between the epoxy encapsulant and the silicon device to prevent package cracking has been demonstrated by finite element analysis and a novel adhesion test.
Further stress relaxation experiments, mostly at 50°C, are reported on mixtures of crosslinkable ethylene–propylene terpolymer with saturated ethylene–propylene copolymer (molecular weights 3.6 and 45 × 104) containing up to 50% by weight of copolymer, crosslinked by sulfur to leave the saturated copolymer unattached and free to reptate in the copolymer network. Stress relaxation was measured in small simple elongations (stretch ratio about 1.15) on samples which had been extracted to remove a large part of the unattached copolymer and dried. The relative increase in modulus at long times (104 sec) increased with the proportion extracted; at short times (1 sec), extraction of the lower molecular weight copolymer increased the modulus to about the same extent but extraction of the higher molecular weight copolymer affected it very little. The relaxation modulus of the copolymer extracted from sample 50H (50% copolymer of high molecular weight), obtained by difference, agreed with that for the total copolymer except for a small difference probably attributable to molecular weight selectivity in the extraction. Stress relaxation was measured on sample 50H at six higher elongations up to a stretch ratio of 3. The dependence of stress on time and strain was consistent with an analysis based on the following assumptions: (a) linear additivity of the network and unattached copolymer contributions, (b) strain–time factorization of the stress contributions from the individual components, (c) a strain dependence for the unattached component corresponding to the presence of a Mooney–Rivlin C2 term only, (d) a strain dependence for the network component which does not follow the Mooney–Rivlin equation but is dominated by a simple neo‐Hookean term.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.