Curing stress solvent

The numerical simulations of the stress distributions are carried out on porous materials submitted to uniaxial loading. In order to check the validity of the numerical simulations, macroporous epoxies are prepared via the CIPS technique. Cyclohexane is selected as the solvent, thus resulting in the formation of a closed porosity, and the statistical distribution of the voids coincides with the random distribution of the model system. The structural characteristics of these materials prepared by curing at T=80 °C are summarized in Table 4. 

Since solvent evaporation and imidization in themselves are not destructive processes, the most crucial temperature regime lies between 150 °C and 250 °C. Here solvent removal and maximum imidization occurs simultaneously causing tremendous shrinkage and the creation of maximum stress in the polymer film. At this point it is not unusual to observe cracking problems in the polymer film, depending on the inherent mechanical properties of the partially cured poly-... 

In the formation of PI films, the material undergoes a solvent removal step plus several thermal treatments to cyclize the polymer and anneal it. Each of these processes contributes to the development of residual stresses. Goldsmith, et. al. ilA) have shown that the resulting stresses from curing PI films are independent of film thickness and the maximum room temperature stress developed for a fully cured film is 70 MPa. 

Many fectors must be taken into consideration in designing an adhesive. The requirements include low level of ionic impurities, no voids under the chip caused by evaporation of solvent or other volatiles, no resin bleed during cure, and thermal expansion properties that match those of the substrate and chip. A significant mismatch in the thermal expansion coeflScient can lead to development of thermal stresses that can result in cracking or distortion of the chip. This problem is becoming more and more important as die sizes continue to increase. 

When a dimensional change in a polymer coating does not match that of its constraint (such as the substrate to which it adheres), the resulting strain generates stress in both the polymer and the substrate. Strain arises in polymers by thermal expansion and contraction ( ), from solvent and by-product evaporation (2), from moisture absorption (2), from cure (2), and from physical aging (12). ... 

The conversion of strain mismatch into stress is a function of the stress relaxation modulus exhibited by the polymer. A predictive stress model must incorporate the complex dependencies of the modulus and stress relaxation behavior on temperature, glass transition temperature, degree of cure, crosslink density, solvent-plasticization, and reaction kinetics. 

Typical procedures of the bending beam experiments were as follows. Resin coatings were applied to one side of a clean quartz strip by spin-coating from an appropriate solvent at 2000 to 8000 rpm for about 30 seconds. Coatings were then evaluated for uniformity, and acceptable beams were then inserted into the bending beam apparatus. Stress measurements were taken during temperature ramps and holds. The specific cure schedule followed for each material depended on the manufacturers recommendations or the results of microdielectric cure studies (lfi.). Upon completion of all stress experiments, the beam was removed and the film thickness measured by an Alphastep profilometer. The quartz beams were 82 2 pm thick,... 

Both materials were subjected to the same processing conditions. The cure profile consisted of heating from room temperature to 2 60°C at 5°C/min, holding at 260°C for 2 hours, then cooling to room temperature at 5°C/min. As can be seen in Figure 3, their stress-temperature profiles are quite different. Both films left the spin-coater with approximately zero stress. Upon heating, the polyimide film developed substantial tensile stress due to film contraction from solvent evaporation while the BCB film exhibited only mild tensile stress buildup. The stress in the BCB film relaxed at 260°C while the stress in the polyimide did not. 

Upon cooling, after a 2 hour cure, both materials exhibit linear stress-temperature profiles. This indicates that the glass transition temperature is at or above the cure temperature, and that measurements have been made in the glassy elastic regime. The glassy-state Ea can be calculated from the slopes of these curves. For the polyimide it is 0.13 MPa/°C and for the BCB it is 0.16 MPa/°C. Note that the polyimide bears a higher cumulative stress at room temperature because of the stress induced by solvent evaporation, in spite of its lower Ea. 

The materials chosen for obstacles, polycarbonate (PC) and nylon-6, had elastic properties very similar to those of the epoxy matrix, so that stress concentrations would not be present and the analysis and interpretation of data would be simplified. All rods and spheres were solvent-cleaned and then thoroughly dried prior to their inclusion in the specimens to prevent plasticization of adjacent epoxy by diffusion of water from the rods during curing of the epoxy. Interparticle spacings (R/L, as defined in Figure 1) of 0.125, 0.187, and 0.250 were used, corresponding to equivalent volume fractions of approximately 0.06, 0.14, and 0.27. 

We have described a new type of conformal coating which is shelf-stable, solventless, one-component, and rapidly cured by a combination of UV and heat. Once cured, this coating has very good electrical properties, thermal stress resistance, and hydrolytic stability. With the proper choice of solvents, all or a portion of the cured coating can be removed from a circuit board to repair the electronic parts. 

---