An alternative to pure thermals
BY IOAN SAUCIUC
Keeping die junction temperature below acceptable limits to ensure device performance and package reliability at an acceptable noise level has been the primary focus of thermal management in microelectronics applications. Most of the fundamental theories, such as forced convection, 2-phase flow, and radiation are dated, and most cooling technology issues and concerns are well-known in the thermal community. Because of the high cost and questionable reliability of future technical solutions, they are not feasible for many market segments. A viable alternative for thermal engineers is to focus on mechanical/material and reliability aspects of future cooling technologies, rather than pure thermals. Because performance limits for some important building blocks are already determined, it is important to investigate innovations which will bring the best return on investment (ROI).
Current air cooling, heatsink technologies used in large form factors, such as desktops and servers, improved their heat-handling capacity over the years. This started to present diminishing returns for research and development.
To facilitate the investigation, future CPU cooling technologies were categorized into two major building blocks: “spreading building block” and the “forced air convection building block.” These building blocks address the performance tradeoffs for large form factors and small form factors (i.e. handheld devices, cell phones, and laptops).
Spreading Building Block
The spreading building block can be divided into three major areas: (A) spreading through better, “effective” thermal conductivity; (B) spreading, which facilitates the use of large remote heat exchangers; and (C) spreading through refrigeration. Using high conductivity materials, better interface materials, and increasing the effective thermal conductivity by creating thin-film evaporation in heat pipes, thermosiphons, and vapor chambers achieves A. Figure 1 shows the possible reduction in thermal resistance when high conductivity (k) anisotropic materials are used.
 Figure 1. Particular study – anisotropic materials’ effect on sink-to-ambient resistance. |
These materials – graphite, for example – have a high in-plane thermal conductivity (X and Z directions), with a low thermal conductivity in the normal direction (Y). For this study, the normal thermal conductivity (kY) was varied between 50 to 360 W/mK. One anisotropic heatsink (80 × 80 × 60 mm) was compared with a heatsink made of copper for a typical CPU test-vehicle using an integral heat spreader. Both heatsinks have the same geometry and are using the same fan. For significant improvements, (kY) should be larger than copper’s conductivity, even for high in-plane conductivities. Current anisotropic materials have normal conductivities much less than 50 W/mK, and looks to find cost-effective anisotropic materials to improve the spreading resistance of the heatsink base. This is a key challenge for this building block and requires material engineers’ help.
 Figure 2. TEC-based concept thermal solution. |
Generating, or enhancing thin-film evaporation in heat pipes or vapor chambers also improves effective thermal conductivity. These devices are limited already and no significant improvements were observed in most of the empirical work. Academia recognized the need to determine the theoretical limit for thin-film evaporation, which will determine the gap compared to current technologies. Until that is done, enhancing of thin-film evaporation should focus on using active devices, such as 2-phase forced convection.
For the thermal interfaces used in CPU cooling, performance was improved to the limits where investing in future research presents diminishing returns.
Single-phase liquid cooling and 2-phase (liquid/vapor) forced convection cooling are main candidates for achieving sub-building block B. The performance and challenges for these technologies include bearing reliability. Innovation in materials and pump technologies will be necessary to help the CPU cooling.
Sub-building block C deals with spreading through refrigeration. Thermoelectric (TEC) modules or vapor compression refrigeration devices provide the most effective methods for CPU cooling. Figure 2 shows an innovative TEC concept which further improves the performance of current devices without using pumps or compressors. The spreading device conducts the heat to the cold side of a conventional thermoelectric device. Several TEC modules introduce a negative (effective) resistance into the chain of thermal resistances, thus decreasing the operating temperature of the spreading device. This TEC technology integration is the only passive extension of today’s conventional air cooling.
Air-convection Heat-transfer Coefficient Building Block
Recent work shows that future innovations of “forced air convection” for large form factors present diminishing returns. However, for small form factors, the conclusion may be different. In recent years, the convergence of communications with computing raises the need for devices to generate air flow in small form factors such as PDAs, cell phones, and laptops. One device which can accommodate this at very low power consumption is the Piezo actuator – a cantilever made out of metal or plastic with a piezoelectric material bonded to it. Under an alternating electrical current, the Piezo actuator oscillates back and forth, generating airflow. Compared to conventional fans, these actuators have much lower power consumption, lower noise (<100 Hz), and smaller overall volume.
Comparison of Technologies
Typically, thermal designers have limited time to perform design sensitivity studies and determine the technology capability envelope, making complex modeling for various configurations and scenarios impossible.
TEC-based, large form factor devices may be the only passive technology with significant advantages. Beyond conventional remote heat-pipe cooling, technology costs will increase by at least a factor of 2 to 4.
 Figure 3. Small-form-factor technology comparison. |
The introduction of different platforms allows for the fabrication of smaller-size devices with more multiple, simultaneous functions. This generates significant advancements into communications platforms for multiple wired and/or wireless technologies. Extra functionality within a smaller package makes thermal management a greater challenge. The cooling issues are more complex and significantly different from the large form factor. Figure 3 approximately quantifies the cost/size/performance tradeoff for the small-form-factor technologies using the test setup presented in Figure 4. The first data point (Piezo Off) shows the performance when no spreading device is used. Other technologies combined with the Piezo actuator are: heat pipe (HP) attached to a plate and small heatsinks (HS). The heat pipe and plate (Piezo Off) technology is used as a baseline for the comparison of the cooling solutions. The combination of a spreader plate and Piezo actuator results in significant improvements in thermal resistance.
 Figure 4. Piezo actuator in an enclosure. |
null
Future Innovation Focus
The innovation focus areas (for both small and large form factors) are presented in Table 1. Most of the innovation focus should be on materials and fabrication of reliable devices. Complex teams comprised of materials, thermal, and mechanical engineers are required to address these innovations for better future electronics.
 Table 1. Future proposed innovation focus. |
null
References
Contact the author for a complete list of references.
IOAN SAUCIUC, senior engineer, may be contacted at Intel Corp., Assembly Technology Development, 5000 W. Chandler Blvd., Chandler, AZ 85226; 408/552-0450; E-mail: [email protected].