Managing Electronics Thermal Management

Are we fooling ourselves? As dutiful engineers and scientists, it’s our job to find solutions to “impossible” problems. But there are some indications that we are approaching a dead end in thermal management. We’ve all heard the story dozens of times before: as the miniaturization of semiconductor electronics continues with no slowdown in sight, the number of transistors and devices per unit area of the substrate is increasing ferociously, as is the associated power consumed and dissipated in these devices. More transistors are being packed into a single chip, more dice into multi-chip modules, and more devices into small confined spaces in systems. In addition, the heat generated by each device is increasing, due to higher operating frequencies, Moore’s Law, etc. This has become such a familiar, repetitive mantra for thermal engineers that we’re in danger of forgetting its true meaning. The truth is that, like it or not, physical systems have practical limits. Someday, maybe soon, we’re going to smack hard into the thermal ceiling. The heat densities in electronic devices are spiraling exponentially upward, but our cooling technologies are not keeping pace. Already, designers are limiting clock speeds and throttling performance for lack of adequate cooling.