Study of the rheological behaviour of Sn-Ag-Cu solder pastes and their correlation with printing performance

Solder paste plays an important role in the electronic assembly process by providing electrical, mechanical and thermal bonding between the components and the substrate. The rheological characterisation of pastes is an important step in the design and development of new paste formulations. With the ever increasing trend of miniaturisation of electronic products, the study of the rheological properties of solder pastes is becoming an integral part in the R&D of new paste formulations and in the quality monitoring and control during paste manufacture and electronic assembly process. This research work outlines some of the novel techniques which can be successfully used to investigate the rheology of leadfree solder pastes. The report also presents the results of the correlation of rheological properties with solder paste printing performance. Four different solder paste samples (namely paste P1, P2, P3 and P4) with different flux vehicle systems and particle size distributions were investigated in the study. As expected, all the paste samples showed shear thinning behaviour. Although the samples displayed similar flow behaviour at high shear rates, differences were observed at low shear rates. In the stencil printing trials, round deposits showed better results than rectangular deposits in terms of paste heights and aperture filling. Our results demonstrate a good correlation between higher paste viscosity and good printing performance. The results of the oscillatory and thixotropy tests were also successfully correlated to the printing behaviour of solder paste.

Microwave curing of thermosetting polymer materials in microelectronics applications

Thermosetting polymer materials are widely utilised in modern microelectronics packaging technology. These materials are used for a number of functions, such as for device bonding, for structural support applications and for physical protection of semiconductor dies. Typically, convection heating systems are used to raise the temperature of the materials to expedite the polymerisation process. The convection cure process has a number of drawbacks including process durations generally in excess of 1 hour and the requirement to heat the entire printed circuit board assembly, inducing thermomechanical stresses which effect device reliability. Microwave energy is able to raise the temperature of materials in a rapid, controlled manner. As the microwave energy penetrates into the polymer materials, the heating can be considered volumetric – i.e. the rate of heating is approximately constant throughout the material. This enables a maximal heating rate far greater than is available with convection oven systems which only raise the surface temperature of the polymer material and rely on thermal conductivity to transfer heat energy into the bulk. The high heating rate, combined with the ability to vary the operating power of the microwave system, enables the extremely rapid cure processes. Microwave curing of a commercially available encapsulation material has been studied experimentally and through use of numerical modelling techniques. The material assessed is Henkel EO-1080, a single component thermosetting epoxy. The producer has suggested three typical convection oven cure options for EO1080: 20 min at 150C or 90 min at 140C or 120 min at 110C. Rapid curing of materials of this type using advanced microwave systems, such as the FAMOBS system [1], is of great interest to microelectronics system manufacturers as it has the potential to reduce manufacturing costs, increase device reliability and enables new device designs. Experimental analysis has demonstrated that, in a realistic chip-on-board encapsulation scenario, the polymer material can be fully cured in approximately one minute. This corresponds to a reduction in cure time of approximately 95 percent relative to the convection oven process. Numerical assessment of the process [2] also suggests that cure times of approximately 70 seconds are feasible whilst indicating that the decrease in process duration comes at the expense of variation in degree of cure within the polymer.