Full paper to be presented at APEX 2003, Mar. 31 – Apr 2, 2003.

Dr. Ken Gilleo
Cookson Electronics

One of the most critical issues facing MEMS, MOEMS, and optoelectronic devices is packaging. While electronic packaging and assembly have undergone a steady evolution to deliver low cost and high-level automation, hermetic packaging remains expensive, with a low degree of standardization and poor adaptability to automated SMT assembly. The work reported here is fabrication and preliminary testing of a possible molded plastic replacement for ceramic/metal hermetic packages.

The next decision is which type of plastic, thermoset or thermoplastic? Thermoplastics have the best barrier properties of any polymers, with the added advantage of remeltability. Remelting provides two key benefits: thermoplastics can be molded into a precise shape including a cap or a cavity, and thermoplastics can be selectively remelted to seal a lid, glass window, or fiber.

Which thermoplastic is optimum? The first criterion must be a melting point well above the processing temperature of solder, including lead-free types. The next criterion is a high gas barrier, especially to moisture. Other desirable properties are low outgassing, low dielectric constant, platability, and reasonable economics. We conclude, as have several others, that the liquid crystal polymer (LCP) class best meets all these criteria.

LCP is a popular choice for OE connectors and many precision parts that require high dimensional stability and low moisture absorption. This plastic can be molded into caps suitable for testing like the one shown in Figure 1 from Foster-Miller, Inc. (Waltham, MA). The cap was made into a hermetic enclosure by bonding a "lid", a piece of glass in our initial work.

Figure 1. Molded liquid crystal cap with flange.

A variety of energy sources can be used to bring an LCP package to the melting point of slightly over 300^oC. Initial testing showed that Near-Infrared (NIR) was a suitable energy source for sealing LCP to glass. Both IR and NIR readily travel through glass without appreciable loss. LCP becomes a strong IR/NIR absorber when a small amount (>1% by weight) of carbon black is added.

A laser beam was directed through glass placed on top of the molded LCP cap, onto the plastic surface, as shown in Figure 2. LCP surface melting occurred at 5 watts, although 15 – 30 watts gave the most satisfactory results. The X-Y table was programmed so that the laser spot traveled along the perimeter of the cap, producing a sealed enclosure.

Figure 2. Process diagram, laser sealing glass lid.

Bond strength tests gave values of about 100 PSI, with an excellent bond to glass. Samples of LCP caps with sealed glass lids were tested for hermeticity using the well-accepted helium leak method. The standard for hermeticity is a value < 1 x 10^-8 cc/@ 1 atm. He. An independent test lab reported that the package simulator passed full hermeticity. Values for "good" samples were in the 10^-9 range, the limit of the equipment.

Preliminary testing on package simulators shows that plastic packaging might provide acceptable hermeticity for some MEMS and MOEMS packaging. Plastic packaging of real devices followed by extensive testing appears warranted, based on the preliminary data.