Two-phase cooling
Thermal management is an ongoing concern for many designs. The process usually begins with a tactic for dissipating or removing heat from the primary sources (mostly but not exclusively “chips”), then progresses to keeping the circuit-board assembly cool, and finally getting the heat out of the box and “away” to where it becomes someone else’s problem. Passive and active approaches are employed, involving some combination of active or passive convection, conduction (in air or liquid), and radiation principles.
The search for an effective cooling and thermal transfer solution has inspired considerable research. One direct approach uses microchannels embedded within the chip itself. This allows coolant, usually water, to flow through, efficiently absorbing and transferring heat away.
The efficiency of this technique is constrained, however, by the sensible heat of water. (“Sensible heat” refers to the amount of heat needed to increase the temperature of a substance without inducing a phase change, such as from liquid to vapor.) In contrast, the latent heat of phase change of water—the thermal energy absorbed during boiling or evaporation—is around seven times greater than its sensible heat.
Two-phase cooling with water can be achieved by using the latent heat transition, resulting in a significant efficiency enhancement in terms of heat dissipation. Maximizing the efficiency of heat transfer depends on a variety of factors. These include the geometry of the microchannels, the two-phase flow regulation, and the flow resistance; adding to the task, there are challenges in managing the flow of vapor bubbles after heating.
Novel water-cooling system
Now, a team at the Institute of Industrial Science at the University of Tokyo has devised a novel water-cooling system comprising three-dimensional microfluidic channel structures, using a capillary structure and a manifold distribution layer. The researchers designed and fabricated various capillary geometries and studied their properties across a range of conditions to enhance thin-film evaporation.
Although this is not the first project to use microchannels, it presents an alternative physical arrangement that appears to offer superior results.
Not surprisingly, they found that both the geometry of the microchannels through which the coolant flows and the manifold channels that control the distribution of coolant influence the thermal and hydraulic performance of the system. Their design centered on using a microchannel heat sink with micropillars as the capillary structure to enhance thin-film evaporation, thus controlling the chaotic two-phase flow to some extent and mitigating local dry-out issues.
This was done in conjunction with three-dimensional manifold fluidic passages for efficient distribution of coolant into the microchannels, Figure 1.
Figure 1 Microfluidic device combining a microchannel layer and a manifold layer. (A) Schematic diagrams of a microfluidic device. Scale bar: 5 mm. (B) Exploded view of microchannel layer and manifold layer. The heater is located on the backside of the substrate with parallel microchannels. Both the microchannel layer and manifold layer are bonded with each other to constitute the flow path. (C) The coolant flows between the manifolds and microchannels to form an N-shaped flow path. The capillary structures separate the vapor flow from the liquid thin film along the sidewall. The inset schematic shows the ordered two-phase flow under ideal conditions. Scale bar: 50 mm. (D) Cross-sectional schematic view of bonded device showing the heat and fluid flow directions. (E) Clamped device is mechanically tightened using bolts and nuts. (F) Images of clamped device showing the isometric, top, and side views. Scale bar, 1 cm. Source: Institute of Industrial Science at the University of Tokyo
Testing this arrangement requires a complicated electrical, thermal, and fluid arrangement, with clamps to put just the right calibrated pressure on the assembly for a consistent thermal impedance, Figure 2. They also had to allow time for start-up thermal transients to reach steady-state and take other test subtleties into account.
Figure 2 The test setup involved a complicated arrangement of electrical, thermal, mechanical, and fluid inputs and sensors, all linked by a LabVIEW application; top: system diagram; bottom: the actual test bench. Source: Institute of Industrial Science at the University of Tokyo
Their test process included varying key physical dimensions of the micropillars, capillary microchannels, and manifolds to determine optimum performance points.
It’s difficult to characterize performance with a single metric, Figure 3.
Figure 3 Benchmark of experimentally demonstrated critical heat flux and COP of two-phase cooling in microchannel using water. Zone 1 indicates the results in this work achieving efficient cooling by using a mass flow rate of 2.0 g/min with an exit vapor quality of 0.54. The other designs using manifolds marked by solid symbols in zone 2 consume hundreds of times of water with an exit vapor quality of around 0.1. The results of microstructure-enhanced designs are marked by open symbols in zone 3. Zone 4 shows the performance of typical single-phase cooling techniques. Source: Institute of Industrial Science at the University of Tokyo
One such number, the measured ratio of useful cooling output to the required energy input (the dimensionless coefficient of performance, or COP) reached up to 105, representing a meaningful advance over other water-channel cooling techniques that are cited in the references.
Details including thermal modeling, physics analysis, device fabrication, test arrangement, full data, results, and data discussion are in their paper “Chip cooling with manifold-capillary structures enables 105 COP in two-phase systems” published in Cell Reports Physical Science.
As noted earlier, this is not the first attempt to use microchannels to cool chips; it represents another approach to implementing this tactic. Do you think this will be viable outside of a lab environment in the real world of mass-volume production and liquid interconnections? Or will it be limited to a very small subset, if any, of enhanced chip-cooling solutions?
Bill Schweber is an EE who has written three textbooks, hundreds of technical articles, opinion columns, and product features.
Related content
- Stop blaming the supply for your dissipation woes
- My long-running affection for heat sinks
- Put a diamond topping on your die to avoid heat stroke
- What’s a “thermal jumper” do, anyway?
- Micropillar surface yields lower-temperature boiling, better heat shedding
The post Can microchannels, manifolds, and two-phase cooling keep chips happy? appeared first on EDN.