Find a downloadable version of this story in pdf format at the end of the story.
As the heat density increases at the die level of more powerful semiconductor devices, there becomes a greater need for new and innovative cooling solutions. Standard Aluminum extrusions and bonded fin heat sinks no longer have the capacity to sufficiently cool the semiconductor dies. These heat sinks are limited by the amount of thermal spreading in the base of the heat sink. As shown in Fig. 1, with increased flux, hot spots are apparent in the base of a standard aluminum heat sink. The fins on the extremities of the part are cool and are not providing any additional cooling. The location above the heat source has the concentrated heat load, causing localized high temperature areas in the sink and ultimately increased junction temperatures.
By inserting a vapor chamber into the base of the heat sink, fins now heat up evenly ultimately making the heat sink more efficient (Fig. 2). Also, the hot spot over the die is now evenly spread out, significantly decreasing the junction temperature.
In implementing a vapor chamber, the gains from adding additional surface area by increasing the footprint of the heat sink are much more realized. In a standard heat sink as shown in Fig. 1, if the size of the heat sink were to increase, there would be no real improvements in overall resistance of the heat sink since the fins that were added would be ineffective (shown in blue in Fig. 1). The vapor chamber allows you to expand the heat sink in width or length nearly without limit and still greatly improved thermal resistance.
Simulating Thermal Performance
In order to show this spreading phenomenon, a CFD (computation fluid dynamics) simulation was created for a typical power semiconductor IGBT and its cooling sink (Fig. 3). For this example, a 9 X 4 X 2 in. low-profile aluminum heat sink with a standard extruded pitch (7 FPI @ .070 in. thick fin) was used to cool an IGBT component (Standard Die locations and thermal spreading in package, Total Power = 2400 watts). As shown in Fig. 1, the high temperature gradients are evident in the cross-section view. The fins toward the outside edges are quite cool compared to the hot spot directly over the electrical component. The thermal resistance of the heat sink with 50 CFM of airflow is 0.054 °C/Watt, which is the temperature rise from ambient to the hottest spot on the surface of the heat sink over the total power dissipated.
The challenge now becomes the ability to spread that heat efficiently through the base of the heat sink without changing its existing geometry. Meeting this challenge allows the designer to stay within the same form factor or original package size without a long and costly redesign of the component enclosure. In order to reduce this high thermal resistance, the heat sink metallic base needs to be replaced with a “super” conducting material. In this example, a vapor chamber can be used as the medium to spread heat in the base more efficiently.
Vapor Chamber Benefits
Vapor chambers are essentially flat or planar heat pipes that use the principles of evaporation and condensation to produce a very high conductivity thermal plane. Vapor chambers, like traditional cylindrical heat pipes, are evacuated vessels with an internal wick and a working fluid. The wick helps transport the working fluid back to the evaporator surface without the use of any moving parts. Once the fluid evaporates, it travels to the cooler section of the chamber, condenses in the wick and the cycle continues.
Vapor chambers can have a number of different shell materials and working fluid combinations. The selection of these materials depends mainly on the operating temperature of the cooling system. The most common combination in the electronics cooling field is copper and water due to its operating temperature of about 10°C to 250°C, but other liquids and materials can be used for extreme temperature ranges.
Bulk conductivities for vapor chambers have been measured at over 30 times the conductivity of copper, and over 10 times the conductivity of pyrolytic graphite and diamond in the same flat plane configuration. In addition, vapor chambers can be bonded into an existing extrusion or used as the base itself, in which case fins can be soldered directly to it. Vapor chamber sizes can range from as small as 1 X 1 in. to as large as 13 X 30 in. Standard thicknesses range from 3 to 9 mm so they can be easily inserted into existing bases.
In today’s electronics cooling market, vapor chambers are used in various applications. The military uses these high-conductivity heat sinks in cooling radar TWTs (traveling wave tubes), IGBTs (Insulated-gate bipolar transistors), and other high-flux electronics. The medical industry uses them to warm blood uniformly. A multitude of heat sinks in mid- to high-range computer servers use vapor chamber technology to manage the heat from high-flux, high-computing-power CPUs that define the speed and performance of the system.
Continue to next page
To illustrate the thermal performance improvement that a vapor chamber can provide to an all-metallic heat sink, let’s examine the same heat sink described earlier, but with a vapor chamber integrated into the heat sink base. As shown in Fig. 4, the heat is spread much more evenly across the entire heat sink, causing a drop in overall thermal resistance of 19 percent. The heat sink shown here with an embedded vapor chamber and all of the parameters held constant exhibits a resistance of only 0.044 °C/watt. Vapor chambers also allow easier implementation of a folded fin or stamped fin. These fins can be directly soldered or glued onto the top of the chamber. The fin densities can be dramatically improved by using these types of fins due to the ability to increase pitch. Since the heat flux is so low exiting the vapor chamber, the additional resistance going through this fin-attach interface is inconsequential.
Higher Metal Heat Sink Performance
As illustrated, the enhanced performance of the vapor chamber improves the thermal performance of an all-metallic heat sink significantly. The improved thermal performance allows the electronic component designer to easily manage component frequency speed and power increases within the existing architecture, and at the same time, allow for much more computing/transmitting power for new designs in a more compact space. For example, if an IGBT component in a given system is reaching its maximum junction temperature at 1500 Watts, the vapor chamber can potentially increase the dissipated power to 1800 Watts without changing junction temperature. This is a great advantage for mechanical designers where changing the form factor of a given heat sink would prompt a costly overhaul of the enclosure layout.
Fig. 5 shows a Therma-Base vapor chamber, which enables lower device temperature and greater component reliability. Like conventional heat sinks, vapor chambers are versatile enough to be freeze/thaw resilient and able to withstand military shock and vibration standards, but it’s important to consider those variables in the specification. Operating temperature, gravity orientation, and length of power transport are all factors to account for when tailoring a vapor chamber’s internal structure for a given application.
While conventional heat sinks may be suitable for use in low power, large heat source applications, it is applications where the performance of the system is limited by thermal/mechanical constraints (heat flux, overall power, space, mass) that the vapor chamber offers the ability to obtain the next level of speed and power in the same space.
The Therma-Base vapor chamber has an enhanced capability to accept higher heat fluxes than a traditional aluminum or copper surface. Its smaller size improves system packaging and provides quieter operation through less air flow. Able to operate in any orientation, the Therma-Base passes shock and vibration testing, and thermal cycling (freeze/thaw).”
Download the story in pdf format here.