Wafer bonding is the microfabrication process of joining two wafers (usually silicon) together. Wafer bonding is not currently used in integrated circuit processing, but it is a useful technique for MEMS researchers. Two wafers with patterned features on them can be bonded to eachother (with patterned-faces touching), creating new structures. For example, two wafers with trenches etched into their surfaces could have completely-enclosed channels after the wafer bonding adjoins their trenches. These channels are useful to researchers working on microfluidics. Wafer bonding is also one way to create silicon on insulator wafers, which could be used in future integrated circuits.

There are three main ways to bond wafers: direct bonding, field-assisted bonding, and intermediate-layer bonding.

In direct bonding, the wafers are first cleaned thoroughly and hydrated. Then they are pressed together, and the hydration results in gentle hydrogen bonding between the wafers. The wafers are heated to about 1000 °C, and they crystallize into one continuous wafer.

In field-assisted, or anodic bonding, one wafer is coated with a glass layer such as Pyrex 7740. This glass contains highly mobile sodium ions. The two wafers are brought together and a voltage (300-700V) is applied between them. This voltage causes the sodium ions in the glass to move away from the interface, leaving it negatively charged. This attracts the silicon to the glass. At a temperature of about 500 °C, the wafers fuse.

Intermediate layer bonding is the simplest bonding method. One wafer is coated with an adhesive substance or some material that will become adhesive at an elevated temperature. The two wafers are brought together, and heated if necessary. The problem with this bonding method is that the adhesives are likely to contaminate the wafers or make them non-functional.

Log in or registerto write something here or to contact authors.