Figure 7. Effect of pressure on the bond line thickness of the thermal
interface material for smooth surfaces for graphite platelets and carbon
black.
The effect of pressure on the bond line thickness of the thermal
interface material for rough surfaces is illustrated in Figure 8 for
graphite platelets and carbon black. Typically, information handling
systems include a plurality of thermal conducting members such as, for
example, processors, integrated heat spreaders, heat sinks, heat
transfer dies, and a variety of other thermal conducting materials. As
the heat production of thermal conducting members, for example,
processors, increases, the transfer of heat between thermal conducting
members, for example, the processor, an integrated heat spreader, a heat
transfer die, and a heat sink, raises a number of issues.
Conventionally, a thermal interface material such as, for example, a
thermal grease, a phase change thermal interface material, and a variety
of other thermal interface materials, is used between a plurality of
thermal conducting members such as, for example, a processor and a heat
sink, an integrated heat spreader and a heat sink, a heat transfer die
and a heat sink, and a pair of heat sinks, in order to fill air gaps in
the thermal conduction path between the two thermal conducting members.
It is optimum to apply an amount of thermal interface material to the
interface surfaces between the thermal conducting members such that the
thermal interface material engages approximately 100 percent of the
interfaces surfaces between the thermal conducting members and
completely occupies an interface volume between the thermal conducting
members. However, when pressure is applied to engage the thermal
conducting members the thermal interface material and then heat is
transferred between the thermal conducting members, the thermal
interface material thins and spreads across the interface surfaces
between the thermal conducting members. This can cause the thermal
interface material to flow out of the interface volume between the
thermal conducting members and migrate onto, for example, a silicon
substrate or a printed circuit board that the thermal conducting members
are coupled to. This phenomenon is known as pump out and is accelerated
by expansion and contraction of the thermal conducting members during
heating and cooling cycles, which results in the loss of the thermal
interface material from the interface volume between the thermal
conducting members. This can be particularly problematic in some
chipsets and processors that include power input pads located adjacent
the chipset or processor on the base substrate, as the thermal interface
material can migrate out of the interface volume between the thermal
conducting members and onto the power input pads, resulting in excessive
heating and part failure at the power interconnect. As the thermal
interface material spreads in the volume between the first thermal
conducting member and the second thermal conducting member, the excess
of thermal interface material becomes housed in the channel, preventing
the excess of thermal interface material from migrating off of the first
thermal transfer surface and onto the sensitive top surface and the
electrical contacts. The method then proceeds to the step where heat is
dissipated from the heat producing component. The heat producing
component is operated and produces heat, which is conducted through the
first thermal conducting member, the thermal interface material, and the
second thermal conducting member. The fins on the second thermal
conducting member allow the heat to be dissipated to the ambient.
Consequently, an apparatus and method are provided that allow excess
thermal interface material being used to help dissipate heat from a heat
producing component to be housed such that the excess thermal interface
material does not engage sensitive surfaces in the system that could
cause failure in the system.