Texas Instruments has been a leading contributor in the field of semiconductor reliability for many years. Our staff is noted for being among the best in the world. It's no wonder that maintaining and improving product reliability is an essential part of what we do at TI.

Reliability engineering begins as soon as test structures can be built with a new process, and it continues throughout the lifetime of the process. We start with an assessment of the fundamental physics of the materials and components used in a process. We then design reliability models that are among the most comprehensive in the industry. These models help us ensure the reliability of our designs by taking into account all the process complexities and operating environments necessary to meet our customer requirements.

Among the critical factors that TI reliability engineers have to examine: the migration of atoms in interconnect metal, the time-dependent dielectric breakdown of gates and low-K films, the injection of electrons into gates that create channel hot carriers, and the negative bias temperature instability that breaks down conductivity because of heat.

Modeling the fundamental device physics alerts us to process modifications we must make to minimize these effects and extend the useful life of the product. With the smaller design rules and greater complexity of new CMOS processes at 65-nm and below, the challenges of creating valid reliability models continue to increase—even as the models become more important to the success of our products.

The effectiveness of reliability modeling and process control is measured by TI's sophisticated quality methodology, which is tied closely to market and business requirements. Statistical quality control enables us to minimize defects and helps us keep process distributions tight by analyzing margins to specification. In the end, quality and reliability controls are reflected in the figure of merit (FOM), a composite measurement that reflects all aspects of transistor performance improvement over time. The FOM, which is continually increasing, shows us how well we are improving our processes from generation to generation.