Pumped Two-Phase Cooling for Thermal Management of High Heat Flux Electronics

February 1, 2014

Pumped two-phase cooling provides a compact, low pumping power option for thermal management in high heat flux applications (300-500W/cm²). Compared to single-phase convection, significantly higher heat transfer coefficients can be achieved at substantially lower flowrates. Two-phase cooling also provides a high degree of isothermality, which is important in many applications such as lasers whose emission wavelengths are temperature-dependent.

Figure 1. Schematic of a pumped two-phase cooling system showing key components.

The phase change (latent heat) of the coolant enables two-phase systems to handle high heat fluxes with low pumping power compared to single-phase systems for a given heat load. Two-phase cooling systems are prone to flow/ thermal instabilities, yet engineering solutions are available to address such issues. Specifically, two-phase systems are not new and techniques to effectively manage instabilities have been studied and are available, including the application of engineered microporous coatings on the heated surface(s), evaluated in this study. The porous coatings considered here enhance the boiling performance by increasing the nucleation site density and provide a capillary-driven mechanism for resupplying the liquid/ coolant to the heat transfer surface that postpones dry-out.

Figure 2. a) Experimental pumped two-phase cooling system; b) test section containing a minichannel heat sink/evaporator, housing and heater block used to simulate the heat load.

A schematic of a typical pumped two-phase cooling system is shown in Figure 1. Key components include a pump, preheater, surge tank, evaporator/ heat sink, condenser and accumulator. The surge tank and the preheater differentiate this system from a conventional liquid cooling loop. The surge tank consists of vapor and liquid at saturation; by controlling the pressure in the tank, the saturation (boiling) temperature of the working fluid can be controlled. The preheater heats the subcooled liquid exiting the condenser to a temperature close to the saturation temperature before it enters the evaporator/ heat sink. This is important as boiling heat transfer is most efficient at saturation (minimal subcooling).

For the sake of illustration, representative experimental results on the cooling performance of a pumped two-phase cooling loop are presented here, along with some practical considerations concerning the design and operation of two-phase cooling systems.

Experimental Hardware

The laboratory test setup shown in Figure 2a is outfitted with a copper minichannel heat sink having overall dimensions 20.4mm × 12.3mm × 6.0mm with 8 rectangular channels each having a hydraulic diameter 1.8 mm and a channel aspect ratio (H/W) of 2.8 (housed within the Test Section Assembly). The heat load applied to the heat sink was simulated using a copper heater block containing cartridge heater inserts. As shown in Figure 2b, the heater block narrows down to a pedestal through which heat is transferred to the heat sink. The heat flux applied to the minichannel heat sink is approximated by the product of the thermal gradient (ΔT/Δx) measured with thermocouples positioned along the pedestal length using the 1-D heat conduction formula:

where kc is the thermal conductivity of the copper heater block.

Boiling Enhancement Coatings

Micro- and nano-textured porous surface coatings can help stabilize boiling and suppress undesirable flow oscillations ^{1, 2}. In addition, the use of coatings in pool boiling experiments have been shown to lower wall superheat ΔT (difference between the wall temperature and the saturation temperature of the coolant) for a given input heat flux as well as increase the Critical Heat Flux (CHF)³. In the current study, the performance of porous coatings was evaluated in flow boiling conditions where both nucleate boiling and forced convection play important roles. Images of select uncoated and coated heat sinks/ evaporators considered in this study are shown in Figure 3.

Results

The thermal performance of coated and uncoated minichannel heat sinks was evaluated in the pumped two-phase cooling system shown in Figure 2. The heat transfer coefficient (HTC) [W/m²K] and the Incipient Wall Superheat [K] were evaluated as a function of the input heat flux and coolant mass flux [kg/m²s] using refrigerant R134a. Here, the HTC was calculated using a one-dimensional resistance model that yields:

Here, T_s is the average temperature of the heat sink base, T_f is the average fluid temperature (average of the inlet and outlet temperatures) and R_b and R_c are the thermal resistances of the copper base and channel. Notably, Rb = thickness of the base/thermal conductivity of copper and:

where W_w is the fin width, W_c is the channel width, ζ is the fin efficiency (function of HTC) and H_c is the channel height. Substituting the experimentally measured values, the equation was then solved for the HTC.

Figure 4. (a) HTC as a function of input heat flux for a coated and uncoated (bare) minichannel heat sink and (b) HTC as a function of coolant mass flux for a coated minichannel heat sink with two different applied heat fluxes (40W/cm2 and 160W/cm2)

Figure 4a shows the HTC versus heat flux for a fixed coolant flowrate (0.6 gpm equivalent to a mass flux of 1400 kg/m²s). Results of a bare copper heat sink and a coated (sintered copper powder – mesh 200-230) are compared. As shown, HTC depends on the input heat flux. Regarding the effect of the microporous coating, it can be seen that the HTC increases by as much as a factor of 2 with the application of the coating. Also note that the coating has a negligible effect on the pressure drop across the heat sink since the coating is very thin (order 100 microns) relative to the channel dimensions. Additionally, the HTC as a function of mass flux is plotted in Figure 4b. At low heat fluxes (i.e., 40W/cm²), HTC does not vary much as a function of mass flux. This is characteristic of nucleate boiling. However, at high heat fluxes (160W/cm²), the HTC increases with increasing mass flux, characteristic of single-phase forced convection.

Figure 5. a) Incipient wall superheat is lower for coated heat sinks; b) CHF increases with application of coatings.

The incipient wall superheat is also plotted in Figure 5a as a function of input heat flux on coated and uncoated (bare) heat sinks. As shown, the coating enhances the thermal performance of the heat sinks as evident by lower values of incipient wall superheat; this means that an electronic device mounted on a two phase heat sink with coating can be maintained at a lower temperature compared to one that is mounted on an identical uncoated heat sink. In Figure 5b, the Critical Heat Flux (CHF) is reported for three different surface coatings (lower mesh numbers correspond to coarser coatings having larger particle sizes) and compared to the baseline (no coating). This means that the coating enables the heat sink to handle higher heat fluxes before drying out, at which point the heat sink can no longer efficiently dissipate the high heat fluxes. While all coatings increase the CHF above the baseline, there is an optimum coating that provides the best performance for the heat sink and operating conditions tested. In other words, coating thickness, particle and pore size of the coating need to be optimized for a given application.

In short, coated heat sinks have higher heat transfer coefficients, lower wall superheat, and higher CHF than uncoated heat sinks. Moreover, temperature and pressure fluctuations (not presented) in the heat sink/evaporator are suppressed with the coated heat sinks, thus enabling accurate thermal control and isothermalization of high-power electronics and laser diodes.

Summary

Pumped two-phase cooling systems can handle very high heat fluxes, operate with small pumps (low pumping power), and can be designed to be compact and reliable. At ACT, efforts are underway to develop an ultra-compact two-phase cooling system for ground and air-based platforms. In this study, the heat sink temperature for all cases considered (heat fluxes as high as 320 W/cm² or ~105 W/cm² on the wetted surface) was maintained below 90°C (with a condenser temperature of 30°C). Among other benefits, a two-phase system allows for precise control of the boiling temperature, which in turn enables accurate control of the device’s operating temperature.

Flow boiling heat transfer on minichannel copper heat sinks was evaluated as a function of coolant mass flux, input heat flux, and boiling enhancement coatings. The microporous coatings substantially improve heat transfer by promoting nucleate boiling on the heat transfer surfaces as demonstrated in the substantial reduction in the incipient wall superheat required to dissipate a given heat flux. For the best performance, for a given application, special attention should however be given to optimizing the thickness of the porous coating and its particle and pore size.

This article was written by Ehsan Yakhshi-Tafti, Xudong Tang, Pete Ritt, and Howard Pearlman, Advanced Cooling Technologies, Inc. (ACT) (Lancaster, PA). For more information, Click Here .

Acknowledgements

Special thanks are given to Dr. Tadej Semenic who developed and tested some of the initial prototypes and Dr. Seung You at the University of Texas Dallas for coating select heat sinks and helping characterize their thermal performance under pool boiling conditions. Funding for this work was provided by the National Science Foundation under contract #1127293.

References

Forster, H. K., & Zuber, N. (1955). Dynamics of vapor bubbles and boiling heat transfer. AIChE Journal, 1(4), 531-535.
Semenic, T., & You, S. M. (2013). Two-Phase Heat Sinks with Microporous Coating. Heat Transfer Engineering, 34(2-3), 246-257.
Chang, J. Y., & You, S. M. (1997). Boiling heat transfer phenomena from microporous and porous surfaces in saturated FC-72. International Journal of Heat and Mass Transfer, 40(18), 4437-4447.