MEMS Microphone

A commercial MEMS application with special packaging requirements is the MEMS microphone. The Knowles SiSonic MEMS microphone is targeted towards high volume consumer electronic products where cost is a key factor [58]. The advantage of MEMS microphones over low cost electret condenser micro­phones is the ability to withstand lead-free solder reflow cycles. This ability permits the MEMS microphone to be picked and placed like any other component and thus leads to significantly lower manufacturing costs.

The microphone itself is a fully clamped round polysilicon membrane, about 0.5 mm in diameter and 1 (im thick, micromachined on a standard silicon wafer. The silicon below the membrane has been removed using Deep Reactive Ion Etch (DRIE). The die size is 1.65x1.65 mm. The electronics for the microphone consisting of a series of different amplifiers is built on a separate CMOS chip.

Both die need to be packaged in a single package. This approach is extremely low cost, exposes the MEMS die to the environment, and still protects the MEMS and CMOS die both physically and from unwanted noise. A schematic of the package is shown in Fig. 12.5. It consists of a base, a wall, and a lid laminated together from FR4 PCB material. All inside surfaces are plated to provide a Faraday shield against EMI interference. The MEMS and CMOS die are adhesively bonded to the base and wire bonded with gold wire together as well as to the base. In order to protect the CMOS die from corrosion and incident light, the CMOS die is also encapsulated. A small port, offset from the MEMS die to prevent dirt accumulation or physical damage to the die, serves as acoustic interface. Packaging is performed on 4" by 4" panels containing hundreds of microphones and only standard semiconductor packaging equipment is used in the assembly of these microphones.

12.11.4 MEMS Shutters

Many MEMS devices are uniquely designed for special systems. Such an exam­ple is the shutters used for the ST-5 MEMS radiator [59, 60] and the James Webb Space Telescope [61]. In both of these systems, large MEMS arrays had to be built, assembled, packaged and tested for space.

For the MEMS package to fly on the New Millennium Program ST-5 Spacecraft by NASA Goddard Space Flight Center and the Johns Hopkins University Applied

Physics Laboratory, novel packaging techniques were needed to place MEMS based thermal control devices on the skin of a satellite for the Variable Emittance Coating Instrument.

Some unique packaging challenges included the need to protect the shutter ele­ments from particulate contamination during launch due to exterior mounting; to consider the spacecraft electrical surface grounding requirements; and to meet the requirements for thermal operation - high and low emissivities - for the different substrate/surface materials.

The micro-shutter-array (MSA) radiator is located on the bottom deck of the spacecraft. The gold-coated shutters open and close over the substrate and change the apparent emittance of the radiator. The shutter die, each 12.65 x 13.03 mm, consists of arrays of 150 (Jim long and 6 (Jim wide shutters driven by electrostatic comb drives, needed to cover as large an area as possible of a 9 cm x10 cm radiator, in order to expose their low emissivity top surface or, when open, the high emissivity substrate.

In order to manage the thermal expansion mismatch between Al and Si for the survival temperature range, - 45 to 65°C, an intermediate carrier made from alu­minum nitrite was used. Sets of six die, with wire bonds connecting all the common inputs, were attached to an aluminum nitride substrate with conductive epoxy. Six of these sets were then themselves attached to the aluminum radiator with epoxy. The thermal results associated with opening and closing the shutters are measured by thermistors that are co-located on the underside of the MSA radiator chassis with a heater to allow control of the radiator temperature. A top-view picture of the radi­ator is shown in Fig. 12.6 and an exploded view of the MSA radiator assembly is shown in Fig. 12.7. The package went through a full space qualification including thermal vacuum, vibration, shock, and acoustic testing at NASA Goddard and fin­ished its mission successfully in 2006, working for the full 3 months of the mission life.

The MEMS shutters for the James Webb Space Telescope (JWST) are of even higher complexity and demonstrate a system which had to be designed around the MEMS device. These shutters expose an infrared spectrometer to different sections

of the image to perform spectroscopy on up to 100 targets simultaneously. One of the requirements was an open fill-factor of more than 70%. The solution was a shut­ter array, made of silicon with a magnetic coating on the shutters, which is opened by a magnet moving across. The shutters are kept open by an electrostatic electrode, which can be addressed individually for each shutter to keep it open to view or block a portion of the sky. Each of the four micro-shutter arrays contains 171 x 365 shut­ters and each shutter is 200 micron-long and 100 micron-wide. Besides achieving space qualification (thermal vacuum, shock, vibration and acoustic testing), these shutters have to work at cryogenic temperatures (35 K) and have the ability to open and close more than 100,000 times. One of the challenges, unlike the ST-5 shut­ters, which are all on a common signal line, is that all shutters on the 38 mm x 35 mm die need to be individually addressed via rows and columns, which requires a large number of connections between the shutter die and the mother board. These connections also need to survive the transition to cryogenic temperatures. NASA's solution was the use of Indium bump bond pads to connect the shutter array to the substrate. A picture of the shutter array mounted into one of the quadrants is shown in Fig. 12.8.

Date: 2015-02-28; view: 2233