The heat generated by Integrated Circuit and discrete semiconductor chips has a damaging and limiting effect on performance. Thermal spreader technology allows a concentrated area of high heat to be distributed over a larger area, thereby reducing the average temperature of the chip.

It is cheaper and more efficient to cool a component the closer the cooling solution is to the heat source. For example, you can produce a chilled glass of lemonade faster by putting ice in it than by air-conditioning the surrounding room. Similarly, a computer chip can be cooled more rapidly and at less cost by using a heat spreader directly adjacent to the chip than by increasing the performance of a heat sink, or increasing airflow with larger or faster fans. This efficiency is responsible for the benefits of using a heat spreader rather than merely a heat sink with a simple metal base plate in lieu of a heat spreader.

Seen in Figures 1 and 2 (FCBGA and QFP packages, respectively), a heat spreader is typically bonded to the die with a thermal adhesive, thereby facilitating heat transfer to the ambient air and PCB (via the package leads/balls) where it can be dissipated.

Heat spreaders are most efficient when heat is uniformly applied over the entire plate. Heat spreaders with large contact areas are attached to heat sources with much smaller contact areas due to the further densification of electronic device packages. The result, a "spreading resistance," creates a higher local temperature at the location where the heat source is placed.

Figure 3 illustrates how the surface temperature of a heat sink base-plate would respond as the size of the heat source is progressively reduced from left to right with all other conditions unchanged: the smaller the heat source, the more spreading has to take place.

On a chip (die) level, the heat generated on a specific region also exhibits spreading resistance and a corresponding temperature gradient across the die. Therefore, while one area of the die may have a temperature well below the design point, another area of the die may exceed the maximum temperature at which the design will function reliably. Figure 4 is an example of a simulated temperature plot of the Pentium 4 processor.

There are several critical criteria in the selection of a heat spreader material:

• Performance
• Coefficient of Thermal Expansion (CTE) which defines the rate at which the material changes size as it changes temperature
• Cost
• Density of the material, and hence its weight
• Anisotropy—the ability of a material to conduct heat better in a preferred direction

The challenge in the selection of a heat spreader material is in balancing these criteria. Optimizing one criterion can generally only be done at the expense of one or more of the others.

Heat spreaders have seen significant evolution in their materials. Originally simple and homogeneous spreaders of solid aluminum or copper were used. Copper in particular has excellent thermal spreading performance, with a bulk conductivity around 380 W/mK. However it does not match the thermal expansion of the silicon chips it needs to cool, necessitating the use of interface materials of relatively high thermal resistance and poor bulk conductivity.

Materials with matching CTE include the metal matrix composites AlSiC, Copper-Tungsten, Copper-Molybdenum and others. These materials have a CTE match to silicon, but are far inferior to copper in thermal spreading performance. Some are also quite dense, and hence introduce weight problems as well.

Much current research and interest has centered on various diamond and carbon composites, as well as anisotropic materials such as carbon nanotubes and Highly-Oriented Pyrolitic Graphite (HOPG). These materials offer excellent performance, and some are very light and/or can be tailored to match the CTE of chips. However the sometimes breathtaking cost of these materials make them unsuitable for many applications.

Find out how Polara compares to other heat spreading materials