Comparison of hermetic sealing using SAC and SnPb solder for a MEMS pressure sensor

Abstract

In order to minimize the influence of packaging stress on the signal of MEMS pressure sensors, the pressure inlet can be reduced in footprint and mechanically decoupled from the mechanically moving parts. Moreover, the formation of a hermetic seal between the sensor inlet and an external inlet becomes more challenging as the footprint is reduced. Soldering is a preferred solution as a hermetic seal is achievable despite some surface roughness at the surfaces to be joined. However, the metallization on the MEMS, the solder and the metallization on the external inlet must be carefully matched to assure long term stability; the solder will react quickly with the metal layers deposited on the surfaces during the reflow process and later at a reduced rate during storage and application. The formation of intermetallic compounds (IMC) can catastrophically degrade the integrity of a joint if large amounts of voids are formed or the mechanical compliance significantly reduces as a result of the IMC formation. The metallization alternatives for the MEMS in this case were sputtered TiW/Au and NiCr/Au. The TiW and NiCr are adhesion layers whereas the Au is applied as a wetting layer which is normally fully consumed during the soldering process. A thick layer of plated Au, or a thick layer of plated Ni with a thin surface finish layer of Au, were possible metallization alternatives for the external inlet. Dewetting of solder from TiW is frequently mentioned in literature, but less conclusive work is published about soldering to NiCr/Au [1–3]. In particular, limited work has been published on long term effects of soldering to NiCr/Au surfaces using a SAC solder. In this work TiW and NiCr were compared as adhesion layers. In addition, SAC and SnPb were compared as solder, and Au and Ni/Au were compared as metallization on the external inlet. A total of 10–20 assemblies were prepared for each of 12 tested combinations. Half of the assemblies were exposed to high temperature storage (HTS) for ~300 hours at 130–150 °C. Shear testing and inspection of fracture surfaces and cross sections using light microscopy, scanning electron microscopy, and energy-dispersive X-ray spectroscopy were performed for samples