0
Review Article

A Review of Two-Phase Forced Cooling in Three-Dimensional Stacked Electronics: Technology Integration OPEN ACCESS

[+] Author and Article Information
Craig Green, Peter Kottke, Xuefei Han, Casey Woodrum, Pouya Asrar, Yogendra Joshi, Andrei Fedorov, Suresh Sitaraman

G. W. Woodruff School of Mechanical Engineering,
Georgia Institute of Technology,
771 Ferst Drive,
Atlanta, GA 30332

Thomas Sarvey, Xuchen Zhang, Muhannad Bakir

School of Electrical and Computer Engineering,
Georgia Institute of Technology,
777 Atlantic Drive NW,
Atlanta, GA 30332

Contributed by the Electronic and Photonic Packaging Division of ASME for publication in the JOURNAL OF ELECTRONIC PACKAGING. Manuscript received July 22, 2015; final manuscript received August 30, 2015; published online September 25, 2015. Assoc. Editor: Mehdi Asheghi.

J. Electron. Packag 137(4), 040802 (Sep 25, 2015) (9 pages) Paper No: EP-15-1068; doi: 10.1115/1.4031481 History: Received July 22, 2015; Revised August 30, 2015

Three-dimensional (3D) stacked electronics present significant advantages from an electrical design perspective, ranging from shorter interconnect lengths to enabling heterogeneous integration. However, multitier stacking exacerbates an already difficult thermal problem. Localized hotspots within individual tiers can provide an additional challenge when the high heat flux region is buried within the stack. Numerous investigations have been launched in the previous decade seeking to develop cooling solutions that can be integrated within the 3D stack, allowing the cooling to scale with the number of tiers in the system. Two-phase cooling is of particular interest, because the associated reduced flow rates may allow reduction in pumping power, and the saturated temperature condition of the coolant may offer enhanced device temperature uniformity. This paper presents a review of the advances in two-phase forced cooling in the past decade, with a focus on the challenges of integrating the technology in high heat flux 3D systems. A holistic approach is applied, considering not only the thermal performance of standalone cooling strategies but also coolant selection, fluidic routing, packaging, and system reliability. Finally, a cohesive approach to thermal design of an evaporative cooling based heat sink developed by the authors is presented, taking into account all of the integration considerations discussed previously. The thermal design seeks to achieve the dissipation of very large (in excess of 500 W/cm2) background heat fluxes over a large 1 cm × 1 cm chip area, as well as extreme (in excess of 2 kW/cm2) hotspot heat fluxes over small 200 μm × 200 μm areas, employing a hybrid design strategy that combines a micropin–fin heat sink for background cooling as well as localized, ultrathin microgaps for hotspot cooling.

Three-dimensional interconnect technology promises significant performance advantages for the next generation of computing and communications systems, ranging from shorter interconnect lengths to the ability to integrate heterogeneous technologies into a single 3D stack [1,2]. However, the vertical stacking of multiple functional dice comes with the challenging thermal problem of significantly increased power densities per chip footprint area. These power densities will be difficult to dissipate using traditional top of the die heat sinking approaches. Interlayer fluidic cooling offers an attractive solution to this difficult thermal problem, as it gives each tier an embedded thermal solution, allowing the cooling to scale directly with the buildup of the 3D stack [3].

While the dissipation of very large heat fluxes with embedded single-phase heat sinks was demonstrated over 30 years ago [4], significant technical challenges remain when designing systems to address the largest expected heat fluxes, which may be in excess of 1 kW/cm2 on average, with local hotspots exceeding 5 kW/cm2 [5]. At these very large heat fluxes, the coolant flow rates required to maintain the temperature gradients across the chips within reasonable constraints are large, as are the associated pressure drops needed to pump the fluid. As a result, two-phase cooling has received significant attention in the past decade [6,7]. It has the potential to allow significantly reduced flow rates and pressure drops for similar power dissipation when compared to single-phase cooling [8], while offering greater temperature uniformity. In typical forced two-phase cooling in high-power 3D systems, saturated or subcooled fluid is delivered to the inlet of a multitier stack and undergoes convective boiling as it traverses the device. This process is shown schematically in Fig. 1 for pin fin style heat sinks integrated into the backside silicon dice in a two-tier stack.

Given the increasing attention two-phase convective cooling has received in recent years, numerous studies and reviews have been published covering the fundamental physics and performance of standalone microcoolers [6,7,911]. To shed light on a newly emerging technology focus of thermal management research, this work will focus on the less covered questions of system integration for the two-phase heat sinks.

In emerging chip architectures, system power consumption and dissipation maps are not uniform, both in space and time, which brings about a significant complication to thermal design unlike a simpler case of thermally uniform heat loads, which has been addressed in numerous studies. The implications of the localized hotspots that arise from these nonuniformities will be reviewed. Next, consideration will be given to the need for and approaches to addressing stabilization of the fluid flow field in two-phase microcoolers. This will be followed by an examination of the coolants being considered for use in two-phase systems, as well as available strategies for choosing optimal coolants for a target application. After considering the available coolants, an overview of the approaches to routing and delivery of those coolants to the 3D stacks will be presented. Finally, an overview of a cohesive thermal design that takes these factors into account will be presented to demonstrate the degree of cross-functional codesign that will be needed to realize high-performance information technology systems that utilize two-phase convective cooling.

Hotspot Cooling.

A key challenge to implementation of two-phase cooling in 3D systems is the inevitable nonuniformity of power consumption within individual tiers, which gives rise to localized hotspots. Hotspots are especially problematic in two-phase cooling systems, as the local variation in heat flux can cause increases in local fluid thermodynamic quality that results in flow oscillations and flow by-pass with reduction in the local heat transfer coefficient. For example, Bogojevic et al. [12] evaluated the impact of local hotspots in various locations in a microchannel heat sink with water as a coolant. They found that the activation of hotspots resulted in pressure fluctuations that were accompanied by significant variations in transverse temperature profile. In a study by Hetsroni et al. [13], it was found, using parallel triangular microchannels, that a relatively mild variation in heat flux (22 W/cm2 average 26 W/cm2 peak) resulted in a large temperature variation exceeding 60 K. As a comparison, this was more than twice the temperature variation found for the case of uniform heating at 22 W/cm2.

Hotspot-induced instabilities are especially problematic in heat-sink designs that implement parallel channels because the fluid flow cannot communicate with adjacent channels to relieve pressure buildup due to vapor generation. As a result, alternative layouts have been investigated to improve flow stability in the presence of local hotspots. Alam et al. [14] proposed microgap heat sinks as an alternative to the microchannel design to improve lateral fluid communication and temperature uniformity. They found that when subjected to similar mass fluxes, a microgap cooler provided a uniform temperature profile in the presence of a 25 W/cm2 hotspot (15 W/cm2 background), while the microchannel design showed a significant temperature increase in the presence of the hotspot.

Microgap coolers provide advantages in terms of flow stability, but they trade surface area enhancement for that benefit. Sahu et al. [15] evaluated recondensation of the vapor generated by a hotspot by a superlattice cooler inserted in the channel. While numerical simulation found that this method could reduce the pumping power needed to move fluid through the heat sink, only a single microchannel was evaluated. As a result, no conclusions regarding flow stability can be reached. Thermoelectric cooling was also investigated as a means of addressing hotspots in a two-phase microchannel heat sink by Marcinichen et al. [16]. They found the technique to be potentially effective at dissipating hotspot heat fluxes approaching 5 kW/cm2. However, the numerical routine manually adjusted mass flux distribution to balance pressure in the outlet plenum, effectively suppressing oscillations.

Green et al. evaluated a hotspot cooling approach that implemented two-phase microgap coolers on the hotspot scale [17]. As shown schematically in Fig. 2, they developed a “fluid to fluid, spot to spreader (F2/S2) hybrid heat sink” that relied on two key ideas. First, the background and hotspot coolers utilized two separate and distinct coolants for global and hotspot cooling with separate inlets for on-chip delivery that allowed the lengths scales, flow rates, and fluid delivery method to be tailored to the demands of the local or global heat fluxes present on the die. Fluidically separating the hotspot coolant from the background coolant helps to mitigate the potential for flow instabilities in the background coolant due to a power nonuniformity. The second driving idea of the F2/S2 approach was utilization of the relatively larger thermal capacity of the background coolant for on-chip regeneration of the hotspot coolant via either direct mixing downstream or heat exchange across a common heat transfer interface—thus, simplifying the packaging, routing, and operation of the overall system. Computational analysis of the F2/S2 design showed the potential to dissipate hotspots of up to 1 kW/cm2 using convective boiling techniques while operating within the limits of commercially available micropump technologies.

Flow Stabilization.

Hotspots may exacerbate fluid flow oscillations; however, two-phase flow has shown the potential for instability even in the presence of uniform heat fluxes. Numerous researchers have observed significant oscillations in parallel microchannels of various cross sections, including rectangular [18,19], trapezoidal [20,21], and triangular [22] channels even in the presence of uniform heating. These flow oscillations give rise to large swings in temperature and premature critical heat flux conditions. As a result, several strategies have been suggested to improve the flow stability.

Flow restrictors at the inlet of the microchannel array have been successfully demonstrated by several researchers, including Refs. [20], [23], and [24]. The pressure drop created by the flow restrictor changes the slope of the pump curve, such that the system cannot operate beyond the system’s fundamental static instability point [25]. As a result, perturbations may not result in the oscillations that would occur without the restrictors in place. In order to be effective, the pressure restriction must be large enough to disallow operation beyond the instability point, thus some knowledge of the microchannels’ pressure drop versus flow rate characteristics is needed for proper design. In addition, these flow restrictors represent an additional system pressure drop, reducing overall system coefficient of performance.

An alternative means of stabilizing the flow in parallel microchannels involves the use of diverging channel cross sections. Lu and Pan found that an array of microchannels with a 0.5 deg diverging angle delivered improved stability over straight channels of similar dimensions [26]. They suggested that the stability improvement might be explained through examination of the nondimensional number K1 first identified by Kandlikar [27] Display Formula

(1)K1=(qcGifg)2ρlρv
Here, qc is the heat flux in the channel, G is the mass flux, ifg is the latent heat, and ρl and ρv are the saturated liquid and vapor densities, respectively. As a stability criterion, this represents the tradeoff between the evaporation momentum and the inertia forces at the liquid–vapor interface.

While the diverging channel design of Lu and Pan reduced the magnitude of pressure oscillations, the instabilities were still present. Through further investigation, Lu and Pan found that the addition of laser-etched artificial nucleation sites could further improve the flow stability [28]. When the wall superheat required for onset of nucleate boiling (ONB) is high, the vapor pressure at ONB may be sufficient to force flow backward into the inlet plenum. Thus, they hypothesize that the nucleation sites allow nucleation at lower superheat, reducing the tendency for backflow.

Fundamentally, instabilities in two-phase heat sink arise as a result of vapor generation in the channels. David et al. developed a technique for removing the vapor phase from the coolant in the channels through a hydrophobic semipermeable membrane [29]. The membrane allows the vapor to escape while trapping the liquid phase in the microchannels. They found that this approach not only improved the hydrodynamic stability of the test device but also reduced two-phase pressure drop.

Many techniques have been developed to address the parallel channel instability associated with microchannel arrays, to varying degrees of success. An alternative to engineering controls to reduce the parallel channel instabilities is to implement a surface area enhancement scheme that does not rely on parallel channels. Micropin–fin arrays, for example, allow the flow to communicate in both the spanwise and streamwise directions. In theory, this should minimize the parallel channel type instabilities.

Qu and Siu-Ho studied a staggered array of square micropins and found that the pressure fluctuations were significantly reduced when compared to a microchannel array subjected to similar conditions [30].The micropin array dissipated 230 W/cm2 with water at a mass flux of 389 kg/m2 s with only very mild fluctuations, while the microchannel array dissipated 200 W/cm2 at 400 kg/m2 s with much larger oscillations in inlet pressure. A throttle valve was available at the inlet of the test section to suppress some of the flow instabilities; however, the study was conducted with the valve wide open to fully highlight the difference in performance between the designs. In general, however, upstream compressible volume flow instabilities may still be present in pin fin style heat sinks, even when the parallel channel instability is minimized.

David et al. investigated a micropin–fin heat sink cooled with R134a under steady and time-varying heat fluxes ranging from 30 W/cm2 to 170 W/cm2 [31]. Under the conditions of their study, the authors did not find the types of temporal temperature fluctuations commonly associated with flow instability. Instead they found that even at 170 W/cm2, the temperature deviation was ±2 °C from the average value.

Coolant Selection.

A wide variety of coolants have been evaluated for potential use in two-phase cooling systems. De-ionized (DI) water, the most abundant and cost friendly of the available liquids, is an excellent heat transfer fluid in its own right. However, its high saturation temperature at atmospheric pressure requires reliance on either subcooled boiling [32] or operation at reduced pressure [33,34]. Furthermore, its relatively low dielectric strength makes it a less conservative choice from a reliability standpoint [35]. Furthermore, the large difference in specific volume between the liquid and vapor phases of water places a limit on the achievable outlet quality for small gap heights before reaching the critical heat flux conditions [36].

A second class of coolants commonly investigated for two-phase cooling are fluorocarbons. These include perfluorocarbons (e.g., FC-72 and FC-77 [37,38]), hydrofluoroethers (e.g., HFE-7000 and HFE-7001 [39,40]), and refrigerants (e.g., R245fa [41,42], R134a [43], R236fa [44], and R141b [45]) among others. Fluorocarbons generally have the high dielectric strength and dielectric constant needed for direct contact liquid cooling, but also have significantly lower thermal conductivity and latent heat of vaporization than water. While many modern fluorocarbons have been engineered to have limited ozone depletion, many still have an undesirably high global warming potential (GWP) [46]. Finally, care must be taken to ensure the long-term compatibility of refrigerants with the elastomers and polymers that are present in the fluidic system packaging, pumps, and seals, as the coolants can have a degrading effect on certain materials [47].

Finally, organic liquids, such as CO2 [48], methanol [49], pentane [50], or acetone [51], have received significant interest as coolants due to their environmentally friendly nature, specifically their low GWP. CO2 is considered especially attractive, as it has a net global warming impact of zero (it is a byproduct of pre-existing industrial processes) and unlike the hydrocarbon-based organic liquids, it is both nontoxic and nonflammable [46].

Kottke et al. conducted a thermodynamic analysis for the design of microgaps for the removal of high heat fluxes, i.e., 1 kW/cm2, at low wall temperature (∼85 °C) via a strategy of very high mass flux (>1000 kg/m2s), high quality (outlet vapor mass quality >90%), two-phase forced convection to obtain performance trends across a wide range of candidate coolants [36]. A first-principles thermodynamic model for coolant selection was adopted because it enabled rapid assessment of tradeoffs with no empirical correlations. The thermodynamic model imposed mass, energy, and momentum conservation on a control volume that consists of all fluid within the microgap. The pressure drop from inlet to outlet is due only to acceleration. A broad portfolio of coolants was considered with properties available in the engineering equation solver [52] to identify the impact of coolant selection on predicted performance. The algebraic conservation equations used in the model are then Display Formula

(2)G=constant
Display Formula
(3)G(iout+ek,outiinek,in)=QH
Display Formula
(4)G2(1ρout1ρin)=PinPout

Here, G is the mass flux of the coolant, i is the fluid enthalpy, ek is the kinetic energy of the fluid, Q is the heat input rate per unit depth of a microgap cooler, H is the height of the microgap, and ρ is the fluid density.

For each coolant, there is a maximum attainable mass flux for a given inlet temperature under the constraint that the coolant starts as saturated liquid and exits as saturated vapor. Allowable microgap height is minimum at the conditions for which the maximum mass flux can be attained, and microgap height is expected to strongly influence heat transfer coefficient, with higher heat transfer coefficients occurring at smaller heights [53,54], due to the inverse scaling of heat transfer coefficients with gap height to some power exponent, n. Because heat transfer coefficients also scale directly with the thermal conductivity of the coolant (k), the results of the thermodynamic analysis are presented in terms of k/Hn in Fig. 3 (with n chosen to be unity for simplicity and lack of a better choice without diverting to empiricism).

As seen in Fig. 3, the highest heat transfer coefficients should occur at the highest system pressures. Water and methanol have anomalously superior performance at low pressure due to combined effects of their larger latent heats and thermal conductivities as compared to the dielectric fluids. However, they are not suitable for many electronics cooling applications involving direct contact with electronic components due to their appreciable electrical conductivities unless they are extremely pure or continuously de-ionized. The trend of improved performance at higher system pressure is a direct result of the small achievable gaps at higher pressure. Smaller gaps can be utilized in higher pressure systems because of larger vapor densities at higher outlet pressures, which result in smaller accelerational pressure drops. Thus, when high system pressures (Pout in Fig. 3) can be tolerated, a coolant such as R134a may be selected. However, structural or reliability concerns may drive selection toward lower pressure refrigerants, such as R245fa.

Green et al. performed a first-principles analysis to determine the scaling relationships for selection of optimal coolants in two-phase heat sinks according to design parameters, such as pumping power or total package volume minimization [8]. From this analysis, a figure of merit parameter (FOM) was developed that could be used to choose coolants that minimized pumping power (P˙) requirements for a given device power dissipation Display Formula

(5)P˙vf2+Cvfvfg/ifg3
Minimizing the property relationship vf2+Cvfvfg/ifg3 in the coolant of choice scales with a reduction in pumping power for the overall two-phase heat sink. Here, vf is the saturated fluid-specific volume, vfg is the difference between saturated fluid and vapor specific volume, ifg is the latent heat of vaporization, and C is a constant related to the two-phase friction factor. Similar to the finding of Kottke et al., this FOM suggests that high-performance (low pumping power requirement) coolants may exist at high pressures where the difference between saturated fluid and vapor specific volume is minimized; however, this is balanced by the typical reduction in ifg near the critical point of fluids.

Warrier et al. used a computer-aided molecular design (CAMD) process to design novel heat transfer fluids for both convective and pool boiling [55]. In CAMD, new materials are developed by combining atoms, molecules, or functional groups in a computer algorithm that compares the properties of the resulting materials to FOM for pool or flow boiling. The FOM that was considered for flow boiling was based on either correlations from Lazarek and Black (vertical channels) [56] or Tran (horizontal channels) [57]. The resulting FOM that was implemented was

LararekandBlack:103kμ0.857ifg0.714
Tran:106k(μifg)0.62(ρvρl)0.607
Here, μ is the fluid viscosity and the remainder of the parameters is as previously defined. Based on the results of their algorithmic design, a new coolant, C6H11F3, was designed and tested experimentally in a 7 wt.% mixture combined with HFE-7200 (93%). The experimental validation was done in a pool boiling experiment where an increase of 7% critical heat flux was observed with the new fluid.

Fluid Delivery.

In order to deliver coolant to microfluidic heat sinks in a 3D stack, complementary metal-oxide semiconductor (CMOS) compatible processes for fabricating and sealing fluidic input/output (I/Os) and fluidic vias must be developed. Dang et al. first demonstrated a low-temperature CMOS compatible process for connecting polymer-based fluidic I/Os from the chips to a FR-4 like substrate following back end of the line processing [58]. King et al. successfully demonstrated fluid delivery to multiple electrically connected stacked tiers, using Avatrel-based fluidic I/Os sealed with underfill epoxy and C4-based microbumps for electrical connectivity [59].

An all solder-based I/O (both fluidic and electrical) technique was developed by King et al. that eliminated the need for polymer piping [60]. All solder-based processing allows simultaneous processing with the electrical I/Os, greater flexibility in fluidic via height, reworkability, as well as other potential benefits.

Oh et al. fabricated and tested annular-shaped fluidic microbumps that were cofabricated with fine pitch electrical microbumps to provide both electrical and fluidic connectivity in a 3D stack. The respective diameter and pitch of the electrical microbumps were 25 μm and 50 μm; the inner and outer diameters of the fluidic microbumps were 150 μm and 210 μm, respectively; the fluidic vias were 100 μm in diameter. Figure 4 shows the fabricated silicon die with fluidic microbumps, fluidic vias, micropin–fin heat sink, and electrical microbumps [61]. The micropin–fin heat sink was hydraulically tested with single-phase water, successfully sealing against about 100 psi (gauge) system pressure. Electrical testing showed low resistances at the interconnects (13.8 Ω average). Oh et al. flip-chip bonded a device with the characteristics of the silicon die in Ref. [61] onto a silicon interposer, demonstrating a 2.5D implementation of fluidic routing and delivery [62]. Upon fluidic and electrical testing of the flip-chip bonded device, the system maintained the performance of the one-dimensional device, validating the efficacy of the bonding and sealing process.

Inspired by the F2/S2 hybrid heat-sink design [17], the three-dimensional stackable evaporative cooler (3D STAECOOL) heat sink has been recently proposed for imbedded cooling. It utilizes thermally coupled background and hotspot coolers to dissipate the substantial heat fluxes expected in 3D-stacked electronics (>500 W/cm2 background and >2 kW/cm2 hotspot). The 3D STAECOOL concept, shown in Fig. 5, uses an integrated micropin–fin enhanced microgap cooler with convective boiling to dissipate the large average heat fluxes generated by 3D-stacked electronics. The design also includes flow redistribution fins in the manifolds (Fig. 6) that work in concert with the micropin design to maximize flow stability. Both R134a (high pressure and high performance) and R245fa (lower pressure and lower mechanical stress) are evaluated as refrigerants. In addition, water at reduced pressure is evaluated. Thermally coupled to the background cooler are dedicated hotspot coolers that pump the dielectric coolant through ultrathin microgaps to deliver maximum heat transfer coefficients. In this section, the thermal-fluidic design of the STAECOOL approach is presented considering the thermal performance of both the background and hotspot coolers. Finally, a thermomechanical analysis of the device’s performance is conducted to evaluate its potential reliability.

Background Cooling With Flow Boiling of Dielectric Liquids.

In the STAECOOL design, the background heat fluxes generated by the devices will be dissipated using micropin–fin heat sinks integrated into the back side of the silicon in each tier of a 3D-stacked device. To efficiently dissipate the large design heat fluxes, two-phase convective boiling has been selected as a cooling technique. To characterize the background cooling approach, test vehicles with 200 μm tall cylindrical pins (shown in Fig. 6) with a range of diameters from 30 μm to 150 μm covering an area of 1 cm2 were fabricated and tested [34]. Transverse pitch to diameter ratios ranged from 1.2 to 4, and lateral pitch to diameter ratios ranged from 1 to 4. In addition to fluid inlet and outlet ports, pressure ports were included on either side of the micropin–fin array in order to accurately measure pressure. A single line of pins also exists on either side of the pin array to promote an even flow distribution and suppress two-phase flow instabilities. Finally, oval support structures were added near the inlet and outlet for mechanical support for operation with high-pressure dielectric coolants. Platinum heaters are deposited on the back side of micropin–fin heat sink, covering an area of 1 cm2. In addition to heating, the platinum heaters function as resistance temperature detectors (RTDs).

Fabrication.

The process used to fabricate micropin–fin samples is shown in Fig. 7. First, the micropin–fins and manifolds were etched into a 500 μm thick double side polished wafer using the Bosch process. The cavities formed during etching were then sealed using an anodically bonded Pyrex cap. Platinum heaters were then deposited on the backside of the wafer with a 2 μm thick insulating silicon dioxide layer. Finally, inlet, outlet, and pressure measurement ports were etched from the backside of the wafer.

Experimental Characterization.

The micropin–fin heat sink was tested in a closed flow loop system with a gear pump and liquid cooled condenser (Fig. 8). Both subcooled, subatmospheric pressure DI water and hydrofluorocarbon refrigerants were evaluated experimentally as coolants. The system was instrumented to measure flow rate, pressure drop, and fluid temperature at both the inlet and outlet of the heat sink. Fluid was delivered to the chips via an external polycarbonate package that seals against the chip using o-rings. The DI water flow loop ran at subatmospheric pressure of 25 kPa. The pressure drop across the water cooled heat sinks varied from 20 kPa to 350 kPa, depending on the pin diameters and pitches. The micropin–fin heat sink dissipated heat fluxes approaching 500 W/cm2, with DI water exit quality of 0.07 at the highest flux, as is shown in Fig. 9. To achieve the high fluxes using water, the DI water entered the heat sink subcooled, resulting in low-outlet quality. High-outlet qualities are more readily achievable using high-pressure refrigerants as compared to water due to the large difference in liquid and vapor specific volume of water at subambient pressure.

Hotspot Cooler.

At the hotspot level, the target fluxes are extremely high (5 kW/cm2), so dedicated hotspot coolers are essential. The ability to route and deliver dedicated cooling streams to the background and hotspot coolers separately means that the dimensions of the hotspot cooler can be tailored to address the local hotspot fluxes, without the excessive pressure drop penalty that would be present if the flow rates present on the global scale were delivered to the hotspot cooler. As a result, ultrathin (less than 10 μm) microgaps can be implemented to achieve the heat transfer coefficients needed to maintain the chips within temperature allowances. Furthermore, the dedicated cooling stream allows the conditions of the coolant to be matched to the desired inlet conditions of the microgap cooler (e.g., saturated, subcooled, etc.), instead of relying on the local flow conditions in the background heat sink.

Sub 10 μm microgaps are not covered by the existing data sets used to develop traditional two-phase convective boiling correlations, so an experimental investigation was undertaken to better understand the heat transfer characteristics of short, ultrathin microgaps [63]. The geometry and layout of the test devices are shown in Fig. 10. A 5 μm deep microgap was etched into a 300 μm silicon substrate that was capped via an anodically bonded borosilicate glass wafer. The microgap is 200 μm wide in the spanwise direction and 300 μm long in the streamwise direction. Fluid is delivered to the microgap via inlet and outlet plena that are 50 μm deep and 200 μm wide. Aligned to the backside of the microgap is a 200 μm × 200 μm thin film platinum heater that is calibrated to act as a RTD and heater simultaneously.

R134a was used as the coolant and was delivered slightly subcooled (∼22 °C) at a volumetric flow rate of 0.4 ml/min. On the back side of the Si substrate, the 200 μm hotspot heater delivers fluxes ranging from 1.6 kW/cm2 up to 2.17 kW/cm2, boiling the fluid as it traverses the microgap. Notably, even at hotspot heat fluxes of 2.17 kW/cm2, the device junction maintains temperatures at or near design thresholds for modern semiconductors. The trends seen in Fig. 11, with the temperature initially rising in the streamwise direction, followed by a subsequent cooling, is characteristic of two-phase flow in microgaps. First, the device temperature rises in the streamwise direction as the fluid goes from subcooled to saturated state. Once the fluid reaches saturated conditions, temperature decreases in the streamwise direction as the saturation temperature decreases with decreasing pressure along the gap. Given the small size of the hotspot (200 μm in the streamwise direction), the overall temperature gradient across the hotspot remains small in absolute terms.

Thermomechanical Reliability Modeling.

In analyzing thermomechanical reliability of microfluidic cooling systems, the pin–fins stand out as a unique feature which arouses concern. These pin–fins should have the capability to reliably encase high aspect ratio through-silicon vias (TSVs) to ensure optimal routing of electrical signals across the microfluidic channel. A 3D finite-element model was developed to predict resultant stresses in a strip of the microfluidic channel with inclusion of copper TSVs. Figures 12 and 13 illustrate the arrangement of copper vias and the surrounding oxide liner inside a pin–fin in the model. The model includes a single 3D strip of 150 μm diameter, 225 μm pitch, 100 μm tall pin fins inside of a 150 μm thick Si die that is heated from the bottom. Convective thermal boundary conditions and pressure loading are applied at the wetted surfaces in the model. Figure 14 shows the resultant first principle stress at the top of the vias due to copper pumping. Next, a parametric study was performed to explore the effects of assumed copper stress-free temperature, copper via diameter and spacing, and oxide thickness on stress in the pin–fins near the TSVs. Figure 15 shows the trend of three stress states with regard to assumed stress-free condition in copper. Oxide stresses and copper pumping both tend to decrease with increasing anneal temperature. This indicates that fabrication and processing guidelines should be developed to include an anneal step for the copper vias, in order to reduce stress in the pin–fins near the TSVs.

Given the high pressures needed to support refrigerants, such as R134a, an analysis of the test vehicle described in Ref. [34] was undertaken to better understand the maximum pressure rating of the device as well as the potential failure modes, so that the design could be improved [64]. It was found that in order to achieve the highest expected system pressures, large unsupported areas, such as those found in the inlet manifold, are structural weak points that must be buttressed with support structures to maintain system integrity at high internal pressures.

Successful integration of two-phase cooling systems into next generation 3D chip stacks will require careful consideration of design parameters crossing multiple disciplines. The focus must not rest only on the global cooling problem, but must pay equal attention to addressing the localized variations in device power density, which give rise to hotspots. The fluid routing and delivery must successfully traverse the path from off-chip interconnects, through multiple tiers, orifices, manifolds, and other flow distribution and support structures to deliver the necessary flow rates to support the cooling. Furthermore, the approach for fluidic connection and routing may not have a deleterious effect on the electrical performance and integration processes needed to deliver a functional system. Equal care must be given to selecting an optimal fluid among the wide variety of available coolants, balancing concerns of not only thermal performance but also the safety and environmental compatibility of the fluids, as well as the long-term reliability concerns that may arise from the necessary operating pressures and material compatibility of the selected coolant. However, as the STAECOOL design has begun to demonstrate, all of these goals may be met with cross-disciplinary collaboration.

DARPA ICECool Fundamentals Program (Award No. HR0011-13-2-0008) as well as ICECool Applications Program (Award No. HR0011-14-1-0002) provided financial support for this work.

Souri, S. J. , Banerjee, K. , Mehrotra, A. , and Saraswat, K. C. , 2000, “ Multiple Si Layer ICs: Motivation, Performance Analysis, and Design Implications,” 37th Annual Design Automation Conference (DAC), Los Angeles, CA, June 5–9, pp. 213–220.
Lau, J. H. , 2014, “ Overview and Outlook of Three-Dimensional Integrated Circuit Packaging, Three-Dimensional Si Integration, and Three-Dimensional Integrated Circuit Integration,” ASME J. Electron. Packag., 136(4), p. 040801. [CrossRef]
Brunschwiler, T. , Michel, B. , Rothuizen, H. , Kloter, U. , Wunderle, B. , Oppermann, H. , and Reichl, H. , 2009, “ Interlayer Cooling Potential in Vertically Integrated Packages,” Microsyst. Technol., 15(1), pp. 57–74. [CrossRef]
Tuckerman, D. B. , and Pease, R. , 1981, “ High-Performance Heat Sinking for VLSI,” IEEE Electron. Device Lett., 2(5), pp. 126–129. [CrossRef]
Bar-Cohen, A. , Maurer, J. J. , and Felbinger, J. G. , 2013, “ DARPA’s Intra/Interchip Enhanced Cooling (ICECool) Program,” Compound Semiconductor Manufacturing Technology Conference (CSMANTECH), New Orleans, LA, May 13–16, pp. 171–174.
Kandlikar, S. G. , 2012, “ History, Advances, and Challenges in Liquid Flow and Flow Boiling Heat Transfer in Microchannels: A Critical Review,” ASME J. Heat Transfer, 134(3), p. 034001. [CrossRef]
Szczukiewicz, S. , Magnini, M. , and Thome, J. R. , 2014, “ Proposed Models, Ongoing Experiments, and Latest Numerical Simulations of Microchannel Two-Phase Flow Boiling,” Int. J. Multiphase Flow, 59, pp. 84–101. [CrossRef]
Green, C. E. , Fedorov, A. G. , and Joshi, Y. K. , 2009, “ Scaling Analysis of Performance Tradeoffs in Electronics Cooling,” IEEE Trans. Compon. Packag. Technol., 32(4), pp. 868–875. [CrossRef]
Harirchian, T. , and Garimella, S. V. , 2011, “ Boiling Heat Transfer and Flow Regimes in Microchannels—A Comprehensive Understanding,” ASME J. Electron. Packag., 133(1), p. 011001. [CrossRef]
Mudawar, I. , 2011, “ Two-Phase Microchannel Heat Sinks: Theory, Applications, and Limitations,” ASME J. Electron. Packag., 133(4), p. 041002. [CrossRef]
Tibirica, C. B. , and Ribatski, G. , 2013, “ Flow Boiling in Micro-Scale Channels–Synthesized Literature Review,” Int. J. Refrig., 36(2), pp. 301–324. [CrossRef]
Bogojevic, D. , Sefiane, K. , Walton, A. , Lin, H. , Cummins, G. , Kenning, D. B. R. , and Karayiannis, T. G. , 2011, “ Experimental Investigation of Non-Uniform Heating Effect on Flow Boiling Instabilities in a Microchannel-Based Heat Sink,” Int. J. Therm. Sci., 50(3), pp. 309–324. [CrossRef]
Hetsroni, G. , Mosyak, A. , and Segal, Z. , 2001, “ Nonuniform Temperature Distribution in Electronic Devices Cooled by Flow in Parallel Microchannels,” IEEE Trans. Compon. Packag. Technol., 24(1), pp. 16–23. [CrossRef]
Alam, T. , Lee, P. S. , Yap, C. R. , and Jin, L. , 2013, “ A Comparative Study of Flow Boiling Heat Transfer and Pressure Drop Characteristics in Microgap and Microchannel Heat Sink and an Evaluation of Microgap Heat Sink for Hotspot Mitigation,” Int. J. Heat Mass Transfer, 58(1–2), pp. 335–347. [CrossRef]
Sahu, V. , Joshi, Y. K. , and Fedorov, A. G. , 2009, “ Hybrid Solid State/Fluidic Cooling for Hot Spot Removal,” Nanoscale Microscale Thermophys. Eng., 13(3), pp. 135–150. [CrossRef]
Marcinichen, J. B. , d’Entremont, B. P. , Thome, J. R. , Bulman, G. , Lewis, J. , and Venkatasubramanian, R. , 2013, “ Thermal Management of Ultra Intense Hot Spots With Two-Phase Multi-Microchannels and Embedded Thermoelectric Cooling,” ASME Paper No. IPACK2013-73276.
Green, C. , Fedorov, A. G. , and Joshi, Y. K. , 2009, “ Fluid-To-Fluid Spot-To-Spreader (F2/S2) Hybrid Heat Sink for Integrated Chip-Level and Hot Spot-Level Thermal Management,” ASME J. Electron. Packag., 131(2), p. 025002. [CrossRef]
Bogojevic, D. , Sefiane, K. , Walton, A. J. , Lin, H. , and Cummins, G. , 2009, “ Two-Phase Flow Instabilities in a Silicon Microchannels Heat Sink,” Int. J. Heat Fluid Flow, 30(5), pp. 854–867. [CrossRef]
Chang, K. , and Pan, C. , 2007, “ Two-Phase Flow Instability for Boiling in a Microchannel Heat Sink,” Int. J. Heat Mass Transfer, 50(11–12), pp. 2078–2088. [CrossRef]
Wang, G. , Cheng, P. , and Bergles, A. , 2008, “ Effects of Inlet/Outlet Configurations on Flow Boiling Instability in Parallel Microchannels,” Int. J. Heat Mass Transfer, 51(9–10), pp. 2267–2281. [CrossRef]
Wu, H. , and Cheng, P. , 2003, “ Visualization and Measurements of Periodic Boiling in Silicon Microchannels,” Int. J. Heat Mass Transfer, 46(14), pp. 2603–2614. [CrossRef]
Hetsroni, G. , Mosyak, A. , Pogrebnyak, E. , and Segal, Z. , 2006, “ Periodic Boiling in Parallel Micro-Channels at Low Vapor Quality,” Int. J. Multiphase Flow, 32(10–11), pp. 1141–1159. [CrossRef]
Szczukiewicz, S. , Borhani, N. , and Thome, J. R. , 2013, “ Two-Phase Heat Transfer and High-Speed Visualization of Refrigerant Flows in 100 × 100 μm2 Silicon Multi-Microchannels,” Int. J. Refrig., 36(2), pp. 402–413. [CrossRef]
Park, J. E. , Thome, J. R. , and Michel, B. , 2009, “ Effect of Inlet Orifice on Saturated CHF and Flow Visualization in Multi-Microchannel Heat Sinks,” 25th Annual IEEE Semiconductor Thermal Measurement and Management Symposium (SEMI-THERM 2009), San Jose, CA, Mar. 15–19.
Bergles, A. E. , and Kandlikar, S. G. , 2005, “ On the Nature of Critical Heat Flux in Microchannels,” ASME J. Heat Transfer, 127(1), pp. 101–107. [CrossRef]
Lu, C. T. , and Pan, C. , 2008, “ Stabilization of Flow Boiling in Microchannel Heat Sinks With a Diverging Cross-Section Design,” J. Micromech. Microeng., 18(7), p. 075035. [CrossRef]
Kandlikar, S. G. , 2003, “ Heat Transfer Mechanisms During Flow Boiling in Microchannels,” ASME Paper No. ICMM2003-1005.
Lu, C. T. , and Pan, C. , 2009, “ A Highly Stable Microchannel Heat Sink for Convective Boiling,” J. Micromech. Microeng., 19(5), p. 055013. [CrossRef]
David, M. P. , Miler, J. , Steinbrenner, J. E. , Yang, Y. , Touzelbaev, M. , and Goodson, K. E. , 2011, “ Hydraulic and Thermal Characteristics of a Vapor Venting Two-Phase Microchannel Heat Exchanger,” Int. J. Heat Mass Transfer, 54(25–26), pp. 5504–5516. [CrossRef]
Qu, W. , and Siu-Ho, A. , 2009, “ Measurement and Prediction of Pressure Drop in a Two-Phase Micro-Pin-Fin Heat Sink,” Int. J. Heat Mass Transfer, 52(21–22), pp. 5173–5184. [CrossRef]
David, T. , Mendler, D. , Mosyak, A. , Bar-Cohen, A. , and Hetsroni, G. , 2014, “ Thermal Management of Time-Varying High Heat Flux Electronic Devices,” ASME J. Electron. Packag., 136(2), p. 021003. [CrossRef]
Kureta, M. , and Akimoto, H. , 2002, “ Critical Heat Flux Correlation for Subcooled Boiling Flow in Narrow Channels,” Int. J. Heat Mass Transfer, 45(20), pp. 4107–4115. [CrossRef]
Koşar, A. , Kuo, C.-J. , and Peles, Y. , 2005, “ Reduced Pressure Boiling Heat Transfer in Rectangular Microchannels With Interconnected Reentrant Cavities,” ASME J. Heat Transfer, 127(10), pp. 1106–1114. [CrossRef]
Zhang, X. , Han, X. , Sarvey, T. E. , Green, C. E. , Kottke, P. A. , Fedorov, A. G. , Joshi, Y. , and Bakir, M. , 2015, “ 3D IC With Embedded Microfluidic Cooling: Technology, Thermal Performance, and Electrical Implications,” ASME 13th International Conference on Nanochannels, Microchannels, and Minichannels (InterPACK/ICNMM2015), San Francisco, CA, July 6–9.
Tong, X. , 2011, “ Liquid Cooling Devices and Their Materials Selection,” Advanced Materials for Thermal Management of Electronic Packaging, Vol. 30, Springer, New York, pp. 421–475.
Kottke, P. , Yun, T. M. , Green, C. E. , Joshi, Y. K. , and Fedorov, A. G. , 2016, “ Two Phase Convective Cooling for Ultra-High Power Dissipation in Microprocessors,” ASME J. Heat Transfer, 138, p. 011501. [CrossRef]
Megahed, A. , and Hassan, I. , 2009, “ Two-Phase Pressure Drop and Flow Visualization of FC-72 in a Silicon Microchannel Heat Sink,” Int. J. Heat Fluid Flow, 30(6), pp. 1171–1182. [CrossRef]
Chen, T. , and Garimella, S. V. , 2006, “ Measurements and High-Speed Visualizations of Flow Boiling of a Dielectric Fluid in a Silicon Microchannel Heat Sink,” Int. J. Multiphase Flow, 32(8), pp. 957–971. [CrossRef]
Kuo, C.-J. , and Peles, Y. , 2009, “ Flow Boiling of Coolant (HFE-7000) Inside Structured and Plain Wall Microchannels,” ASME J. Heat Transfer, 131(12), p. 121011. [CrossRef]
Lee, J. , and Mudawar, I. , 2009, “ Experimental Investigation and Theoretical Model for Subcooled Flow Boiling Pressure Drop in Microchannel Heat Sinks,” ASME J. Electron. Packag., 131(3), p. 031008. [CrossRef]
Agostini, B. , Thome, J. R. , Fabbri, M. , Michel, B. , Calmi, D. , and Kloter, U. , 2008, “ High Heat Flux Flow Boiling in Silicon Multi-Microchannels—Part II: Heat Transfer Characteristics of Refrigerant R245fa,” Int. J. Heat Mass Transfer, 51(21–22), pp. 5415–5425. [CrossRef]
Isaacs, S. , Kim, Y. J. , McNamara, A. J. , Joshi, Y. , Zhang, Y. , and Bakir, M. S. , 2012, “ Two-Phase Flow and Heat Transfer in Pin-Fin Enhanced Micro-Gaps,” 13th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), San Diego, CA, May 30–June 1, pp. 1084–1089.
Madhour, Y. , Olivier, J. , Costa-Patry, E. , Paredes, S. , Michel, B. , and Thome, J. R. , 2011, “ Flow Boiling of R134a in a Multi-Microchannel Heat Sink With Hotspot Heaters for Energy-Efficient Microelectronic CPU Cooling Applications,” IEEE Trans. Compon., Packag. Manuf. Technol., 1(6), pp. 873–883. [CrossRef]
Agostini, B. , Thome, J. R. , Fabbri, M. , Michel, B. , Calmi, D. , and Kloter, U. , 2008, “ High Heat Flux Flow Boiling in Silicon Multi-Microchannels—Part I: Heat Transfer Characteristics of Refrigerant R236fa,” Int. J. Heat Mass Transfer, 51, pp. 5400–5414. [CrossRef]
Dong, T. , Yang, Z. , Bi, Q. , and Zhang, Y. , 2008, “ Freon R141b Flow Boiling in Silicon Microchannel Heat Sinks: Experimental Investigation,” Heat Mass Transfer, 44(3), pp. 315–324. [CrossRef]
Bolaji, B. O. , and Huan, Z. , 2013, “ Ozone Depletion and Global Warming: Case for the Use of Natural Refrigerant—A Review,” Renewable Sustainable Energy Rev., 18, pp. 49–54. [CrossRef]
Szymurski, S. R. , Hourahan, G. C. , and Godwin, D. S. , 1993, “ Materials Compatibility and Lubricants Research on CFC-Refrigerant Substitutes,” Quarterly MCLR Program Technical Progress Report, U.S. Department of Energy, Washington, DC, Report No. DOE/CE/23810-8, available at: http://www.osti.gov/scitech/servlets/purl/7076439-wca8qF/
Cheng, L. , and Thome, J. R. , 2009, “ Cooling of Microprocessors Using Flow Boiling of CO2 in a Micro-Evaporator: Preliminary Analysis and Performance Comparison,” Appl. Therm. Eng., 29(11–12), pp. 2426–2432. [CrossRef]
Lin, P. , Fu, B. , and Pan, C. , 2011, “ Critical Heat Flux on Flow Boiling of Methanol–Water Mixtures in a Diverging Microchannel With Artificial Cavities,” Int. J. Heat Mass Transfer, 54(15–16), pp. 3156–3166. [CrossRef]
Hanks, D. F. , Zhengmao, L. , Narayanan, S. , Bagnall, K. R. , Raj, R. , Rong, X. , Enright, R. , and Wang, E. N. , 2014, “ Nanoporous Evaporative Device for Advanced Electronics Thermal Management,” IEEE Intersociety Conference Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), Orlando, FL, May 27–30, pp. 290–295.
Ribatski, G. , Zhang, W. , Consolini, L. , Xu, J. , and Thome, J. R. , 2007, “ On the Prediction of Heat Transfer in Micro-Scale Flow Boiling,” Heat Transfer Eng., 28(10), pp. 842–851. [CrossRef]
Klein, S. , and Alvarado, F. , 2002, “ Engineering Equation Solver,” F-Chart Software, Madison, WI.
Alam, T. , Lee, P. S. , Yap, C. R. , and Jin, L. , 2012, “ Experimental Investigation of Local Flow Boiling Heat Transfer and Pressure Drop Characteristics in Microgap Channel,” Int. J. Multiphase Flow, 42, pp. 164–174. [CrossRef]
Kim, D. W. , Rahim, E. , Bar-Cohen, A. , and Han, B. , 2008, “ Thermofluid Characteristics of Two-Phase Flow in Micro-Gap Channels,” 11th Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITHERM 2008), Orlando, FL, May 28–31, pp. 979–992.
Warrier, P. , Sathyanarayana, A. , Patil, D. V. , France, S. , Joshi, Y. , and Teja, A. S. , 2012, “ Novel Heat Transfer Fluids for Direct Immersion Phase Change Cooling of Electronic Systems,” Int. J. Heat Mass Transfer, 55(13–14), pp. 3379–3385. [CrossRef]
Lazarek, G. M. , and Black, S. H. , 1982, “ Evaporative Heat Transfer, Pressure Drop and Critical Heat Flux in a Small Vertical Tube With R-113,” Int. J. Heat Mass Transfer, 25(7), pp. 945–960. [CrossRef]
Tran, T. , Wambsganss, M. , and France, D. , 1996, “ Small Circular- and Rectangular-Channel Boiling With Two Refrigerants,” Int. J. Multiphase Flow, 22(3), pp. 485–498. [CrossRef]
Dang, B. , Bakir, M. S. , and Meindl, J. D. , 2006, “ Integrated Thermal-Fluidic I/O Interconnects for an On-Chip Microchannel Heat Sink,” IEEE Electron. Device Lett., 27(2), pp. 117–119. [CrossRef]
King, C. R. , Sekar, D. , Bakir, M. S. , Dang, B. , Pikarsky, J. , and Meindl, J. D. , 2008, “ 3D Stacking of Chips With Electrical and Microfluidic I/O Interconnects,” 58th Electronic Components and Technology Conference (ECTC 2008), Lake Buena Vista, FL, May 27–30.
King, C. R., Jr. , Zaveri, J. , Bakir, M. S. , and Meindl, J. D. , 2010, “ Electrical and Fluidic C4 Interconnections for Inter-Layer Liquid Cooling of 3D ICs,” 60th Electronic Components and Technology Conference (ECTC), Las Vegas, NV, June 1–4, pp. 1674–1681.
Oh, H. , Zhang, Y. , Zheng, L. , and Bakir, M. S. , 2014, “ Electrical Interconnect and Microfluidic Cooling Within 3D ICs and Silicon Interposer,” ASME Paper No. ICNMM2014-21813.
Zheng, L. , Zhang, Y. , Huang, G. , and Bakir, M. S. , 2014, “ Novel Electrical and Fluidic Microbumps for Silicon Interposer and 3-D ICs,” IEEE Trans. Compon., Packag. Manuf. Technol., 4(5), pp. 777–785. [CrossRef]
Green, C. E. , Kottke, P. A. , Sarvey, T. E. , Fedorov, A. G. , Joshi, Y. , and Bakir, M. S. , 2015, “ Performance and Integration Implications of Addressing Localized Hotspots Through Two Approaches: Clustering of Micro Pin Fins and Dedicated Microgap Coolers,” ASME 13th International Conference on Nanochannels, Microchannels, and Minichannels (InterPACK/ICNMM2015), San Francisco, CA, July 6–9.
Woodrum, D. C. , Sarvey, T. , Bakir, M. S. , and Sitaraman, S. K. , 2015, “ Reliability Study of Micro-Pin Fin Array for On-Chip Cooling,” IEEE 65th Electronic Components and Technology Conference (ECTC), San Diego, CA, May 26–29, pp. 2283–2287.
Copyright © 2015 by ASME
View article in PDF format.

References

Souri, S. J. , Banerjee, K. , Mehrotra, A. , and Saraswat, K. C. , 2000, “ Multiple Si Layer ICs: Motivation, Performance Analysis, and Design Implications,” 37th Annual Design Automation Conference (DAC), Los Angeles, CA, June 5–9, pp. 213–220.
Lau, J. H. , 2014, “ Overview and Outlook of Three-Dimensional Integrated Circuit Packaging, Three-Dimensional Si Integration, and Three-Dimensional Integrated Circuit Integration,” ASME J. Electron. Packag., 136(4), p. 040801. [CrossRef]
Brunschwiler, T. , Michel, B. , Rothuizen, H. , Kloter, U. , Wunderle, B. , Oppermann, H. , and Reichl, H. , 2009, “ Interlayer Cooling Potential in Vertically Integrated Packages,” Microsyst. Technol., 15(1), pp. 57–74. [CrossRef]
Tuckerman, D. B. , and Pease, R. , 1981, “ High-Performance Heat Sinking for VLSI,” IEEE Electron. Device Lett., 2(5), pp. 126–129. [CrossRef]
Bar-Cohen, A. , Maurer, J. J. , and Felbinger, J. G. , 2013, “ DARPA’s Intra/Interchip Enhanced Cooling (ICECool) Program,” Compound Semiconductor Manufacturing Technology Conference (CSMANTECH), New Orleans, LA, May 13–16, pp. 171–174.
Kandlikar, S. G. , 2012, “ History, Advances, and Challenges in Liquid Flow and Flow Boiling Heat Transfer in Microchannels: A Critical Review,” ASME J. Heat Transfer, 134(3), p. 034001. [CrossRef]
Szczukiewicz, S. , Magnini, M. , and Thome, J. R. , 2014, “ Proposed Models, Ongoing Experiments, and Latest Numerical Simulations of Microchannel Two-Phase Flow Boiling,” Int. J. Multiphase Flow, 59, pp. 84–101. [CrossRef]
Green, C. E. , Fedorov, A. G. , and Joshi, Y. K. , 2009, “ Scaling Analysis of Performance Tradeoffs in Electronics Cooling,” IEEE Trans. Compon. Packag. Technol., 32(4), pp. 868–875. [CrossRef]
Harirchian, T. , and Garimella, S. V. , 2011, “ Boiling Heat Transfer and Flow Regimes in Microchannels—A Comprehensive Understanding,” ASME J. Electron. Packag., 133(1), p. 011001. [CrossRef]
Mudawar, I. , 2011, “ Two-Phase Microchannel Heat Sinks: Theory, Applications, and Limitations,” ASME J. Electron. Packag., 133(4), p. 041002. [CrossRef]
Tibirica, C. B. , and Ribatski, G. , 2013, “ Flow Boiling in Micro-Scale Channels–Synthesized Literature Review,” Int. J. Refrig., 36(2), pp. 301–324. [CrossRef]
Bogojevic, D. , Sefiane, K. , Walton, A. , Lin, H. , Cummins, G. , Kenning, D. B. R. , and Karayiannis, T. G. , 2011, “ Experimental Investigation of Non-Uniform Heating Effect on Flow Boiling Instabilities in a Microchannel-Based Heat Sink,” Int. J. Therm. Sci., 50(3), pp. 309–324. [CrossRef]
Hetsroni, G. , Mosyak, A. , and Segal, Z. , 2001, “ Nonuniform Temperature Distribution in Electronic Devices Cooled by Flow in Parallel Microchannels,” IEEE Trans. Compon. Packag. Technol., 24(1), pp. 16–23. [CrossRef]
Alam, T. , Lee, P. S. , Yap, C. R. , and Jin, L. , 2013, “ A Comparative Study of Flow Boiling Heat Transfer and Pressure Drop Characteristics in Microgap and Microchannel Heat Sink and an Evaluation of Microgap Heat Sink for Hotspot Mitigation,” Int. J. Heat Mass Transfer, 58(1–2), pp. 335–347. [CrossRef]
Sahu, V. , Joshi, Y. K. , and Fedorov, A. G. , 2009, “ Hybrid Solid State/Fluidic Cooling for Hot Spot Removal,” Nanoscale Microscale Thermophys. Eng., 13(3), pp. 135–150. [CrossRef]
Marcinichen, J. B. , d’Entremont, B. P. , Thome, J. R. , Bulman, G. , Lewis, J. , and Venkatasubramanian, R. , 2013, “ Thermal Management of Ultra Intense Hot Spots With Two-Phase Multi-Microchannels and Embedded Thermoelectric Cooling,” ASME Paper No. IPACK2013-73276.
Green, C. , Fedorov, A. G. , and Joshi, Y. K. , 2009, “ Fluid-To-Fluid Spot-To-Spreader (F2/S2) Hybrid Heat Sink for Integrated Chip-Level and Hot Spot-Level Thermal Management,” ASME J. Electron. Packag., 131(2), p. 025002. [CrossRef]
Bogojevic, D. , Sefiane, K. , Walton, A. J. , Lin, H. , and Cummins, G. , 2009, “ Two-Phase Flow Instabilities in a Silicon Microchannels Heat Sink,” Int. J. Heat Fluid Flow, 30(5), pp. 854–867. [CrossRef]
Chang, K. , and Pan, C. , 2007, “ Two-Phase Flow Instability for Boiling in a Microchannel Heat Sink,” Int. J. Heat Mass Transfer, 50(11–12), pp. 2078–2088. [CrossRef]
Wang, G. , Cheng, P. , and Bergles, A. , 2008, “ Effects of Inlet/Outlet Configurations on Flow Boiling Instability in Parallel Microchannels,” Int. J. Heat Mass Transfer, 51(9–10), pp. 2267–2281. [CrossRef]
Wu, H. , and Cheng, P. , 2003, “ Visualization and Measurements of Periodic Boiling in Silicon Microchannels,” Int. J. Heat Mass Transfer, 46(14), pp. 2603–2614. [CrossRef]
Hetsroni, G. , Mosyak, A. , Pogrebnyak, E. , and Segal, Z. , 2006, “ Periodic Boiling in Parallel Micro-Channels at Low Vapor Quality,” Int. J. Multiphase Flow, 32(10–11), pp. 1141–1159. [CrossRef]
Szczukiewicz, S. , Borhani, N. , and Thome, J. R. , 2013, “ Two-Phase Heat Transfer and High-Speed Visualization of Refrigerant Flows in 100 × 100 μm2 Silicon Multi-Microchannels,” Int. J. Refrig., 36(2), pp. 402–413. [CrossRef]
Park, J. E. , Thome, J. R. , and Michel, B. , 2009, “ Effect of Inlet Orifice on Saturated CHF and Flow Visualization in Multi-Microchannel Heat Sinks,” 25th Annual IEEE Semiconductor Thermal Measurement and Management Symposium (SEMI-THERM 2009), San Jose, CA, Mar. 15–19.
Bergles, A. E. , and Kandlikar, S. G. , 2005, “ On the Nature of Critical Heat Flux in Microchannels,” ASME J. Heat Transfer, 127(1), pp. 101–107. [CrossRef]
Lu, C. T. , and Pan, C. , 2008, “ Stabilization of Flow Boiling in Microchannel Heat Sinks With a Diverging Cross-Section Design,” J. Micromech. Microeng., 18(7), p. 075035. [CrossRef]
Kandlikar, S. G. , 2003, “ Heat Transfer Mechanisms During Flow Boiling in Microchannels,” ASME Paper No. ICMM2003-1005.
Lu, C. T. , and Pan, C. , 2009, “ A Highly Stable Microchannel Heat Sink for Convective Boiling,” J. Micromech. Microeng., 19(5), p. 055013. [CrossRef]
David, M. P. , Miler, J. , Steinbrenner, J. E. , Yang, Y. , Touzelbaev, M. , and Goodson, K. E. , 2011, “ Hydraulic and Thermal Characteristics of a Vapor Venting Two-Phase Microchannel Heat Exchanger,” Int. J. Heat Mass Transfer, 54(25–26), pp. 5504–5516. [CrossRef]
Qu, W. , and Siu-Ho, A. , 2009, “ Measurement and Prediction of Pressure Drop in a Two-Phase Micro-Pin-Fin Heat Sink,” Int. J. Heat Mass Transfer, 52(21–22), pp. 5173–5184. [CrossRef]
David, T. , Mendler, D. , Mosyak, A. , Bar-Cohen, A. , and Hetsroni, G. , 2014, “ Thermal Management of Time-Varying High Heat Flux Electronic Devices,” ASME J. Electron. Packag., 136(2), p. 021003. [CrossRef]
Kureta, M. , and Akimoto, H. , 2002, “ Critical Heat Flux Correlation for Subcooled Boiling Flow in Narrow Channels,” Int. J. Heat Mass Transfer, 45(20), pp. 4107–4115. [CrossRef]
Koşar, A. , Kuo, C.-J. , and Peles, Y. , 2005, “ Reduced Pressure Boiling Heat Transfer in Rectangular Microchannels With Interconnected Reentrant Cavities,” ASME J. Heat Transfer, 127(10), pp. 1106–1114. [CrossRef]
Zhang, X. , Han, X. , Sarvey, T. E. , Green, C. E. , Kottke, P. A. , Fedorov, A. G. , Joshi, Y. , and Bakir, M. , 2015, “ 3D IC With Embedded Microfluidic Cooling: Technology, Thermal Performance, and Electrical Implications,” ASME 13th International Conference on Nanochannels, Microchannels, and Minichannels (InterPACK/ICNMM2015), San Francisco, CA, July 6–9.
Tong, X. , 2011, “ Liquid Cooling Devices and Their Materials Selection,” Advanced Materials for Thermal Management of Electronic Packaging, Vol. 30, Springer, New York, pp. 421–475.
Kottke, P. , Yun, T. M. , Green, C. E. , Joshi, Y. K. , and Fedorov, A. G. , 2016, “ Two Phase Convective Cooling for Ultra-High Power Dissipation in Microprocessors,” ASME J. Heat Transfer, 138, p. 011501. [CrossRef]
Megahed, A. , and Hassan, I. , 2009, “ Two-Phase Pressure Drop and Flow Visualization of FC-72 in a Silicon Microchannel Heat Sink,” Int. J. Heat Fluid Flow, 30(6), pp. 1171–1182. [CrossRef]
Chen, T. , and Garimella, S. V. , 2006, “ Measurements and High-Speed Visualizations of Flow Boiling of a Dielectric Fluid in a Silicon Microchannel Heat Sink,” Int. J. Multiphase Flow, 32(8), pp. 957–971. [CrossRef]
Kuo, C.-J. , and Peles, Y. , 2009, “ Flow Boiling of Coolant (HFE-7000) Inside Structured and Plain Wall Microchannels,” ASME J. Heat Transfer, 131(12), p. 121011. [CrossRef]
Lee, J. , and Mudawar, I. , 2009, “ Experimental Investigation and Theoretical Model for Subcooled Flow Boiling Pressure Drop in Microchannel Heat Sinks,” ASME J. Electron. Packag., 131(3), p. 031008. [CrossRef]
Agostini, B. , Thome, J. R. , Fabbri, M. , Michel, B. , Calmi, D. , and Kloter, U. , 2008, “ High Heat Flux Flow Boiling in Silicon Multi-Microchannels—Part II: Heat Transfer Characteristics of Refrigerant R245fa,” Int. J. Heat Mass Transfer, 51(21–22), pp. 5415–5425. [CrossRef]
Isaacs, S. , Kim, Y. J. , McNamara, A. J. , Joshi, Y. , Zhang, Y. , and Bakir, M. S. , 2012, “ Two-Phase Flow and Heat Transfer in Pin-Fin Enhanced Micro-Gaps,” 13th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), San Diego, CA, May 30–June 1, pp. 1084–1089.
Madhour, Y. , Olivier, J. , Costa-Patry, E. , Paredes, S. , Michel, B. , and Thome, J. R. , 2011, “ Flow Boiling of R134a in a Multi-Microchannel Heat Sink With Hotspot Heaters for Energy-Efficient Microelectronic CPU Cooling Applications,” IEEE Trans. Compon., Packag. Manuf. Technol., 1(6), pp. 873–883. [CrossRef]
Agostini, B. , Thome, J. R. , Fabbri, M. , Michel, B. , Calmi, D. , and Kloter, U. , 2008, “ High Heat Flux Flow Boiling in Silicon Multi-Microchannels—Part I: Heat Transfer Characteristics of Refrigerant R236fa,” Int. J. Heat Mass Transfer, 51, pp. 5400–5414. [CrossRef]
Dong, T. , Yang, Z. , Bi, Q. , and Zhang, Y. , 2008, “ Freon R141b Flow Boiling in Silicon Microchannel Heat Sinks: Experimental Investigation,” Heat Mass Transfer, 44(3), pp. 315–324. [CrossRef]
Bolaji, B. O. , and Huan, Z. , 2013, “ Ozone Depletion and Global Warming: Case for the Use of Natural Refrigerant—A Review,” Renewable Sustainable Energy Rev., 18, pp. 49–54. [CrossRef]
Szymurski, S. R. , Hourahan, G. C. , and Godwin, D. S. , 1993, “ Materials Compatibility and Lubricants Research on CFC-Refrigerant Substitutes,” Quarterly MCLR Program Technical Progress Report, U.S. Department of Energy, Washington, DC, Report No. DOE/CE/23810-8, available at: http://www.osti.gov/scitech/servlets/purl/7076439-wca8qF/
Cheng, L. , and Thome, J. R. , 2009, “ Cooling of Microprocessors Using Flow Boiling of CO2 in a Micro-Evaporator: Preliminary Analysis and Performance Comparison,” Appl. Therm. Eng., 29(11–12), pp. 2426–2432. [CrossRef]
Lin, P. , Fu, B. , and Pan, C. , 2011, “ Critical Heat Flux on Flow Boiling of Methanol–Water Mixtures in a Diverging Microchannel With Artificial Cavities,” Int. J. Heat Mass Transfer, 54(15–16), pp. 3156–3166. [CrossRef]
Hanks, D. F. , Zhengmao, L. , Narayanan, S. , Bagnall, K. R. , Raj, R. , Rong, X. , Enright, R. , and Wang, E. N. , 2014, “ Nanoporous Evaporative Device for Advanced Electronics Thermal Management,” IEEE Intersociety Conference Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), Orlando, FL, May 27–30, pp. 290–295.
Ribatski, G. , Zhang, W. , Consolini, L. , Xu, J. , and Thome, J. R. , 2007, “ On the Prediction of Heat Transfer in Micro-Scale Flow Boiling,” Heat Transfer Eng., 28(10), pp. 842–851. [CrossRef]
Klein, S. , and Alvarado, F. , 2002, “ Engineering Equation Solver,” F-Chart Software, Madison, WI.
Alam, T. , Lee, P. S. , Yap, C. R. , and Jin, L. , 2012, “ Experimental Investigation of Local Flow Boiling Heat Transfer and Pressure Drop Characteristics in Microgap Channel,” Int. J. Multiphase Flow, 42, pp. 164–174. [CrossRef]
Kim, D. W. , Rahim, E. , Bar-Cohen, A. , and Han, B. , 2008, “ Thermofluid Characteristics of Two-Phase Flow in Micro-Gap Channels,” 11th Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITHERM 2008), Orlando, FL, May 28–31, pp. 979–992.
Warrier, P. , Sathyanarayana, A. , Patil, D. V. , France, S. , Joshi, Y. , and Teja, A. S. , 2012, “ Novel Heat Transfer Fluids for Direct Immersion Phase Change Cooling of Electronic Systems,” Int. J. Heat Mass Transfer, 55(13–14), pp. 3379–3385. [CrossRef]
Lazarek, G. M. , and Black, S. H. , 1982, “ Evaporative Heat Transfer, Pressure Drop and Critical Heat Flux in a Small Vertical Tube With R-113,” Int. J. Heat Mass Transfer, 25(7), pp. 945–960. [CrossRef]
Tran, T. , Wambsganss, M. , and France, D. , 1996, “ Small Circular- and Rectangular-Channel Boiling With Two Refrigerants,” Int. J. Multiphase Flow, 22(3), pp. 485–498. [CrossRef]
Dang, B. , Bakir, M. S. , and Meindl, J. D. , 2006, “ Integrated Thermal-Fluidic I/O Interconnects for an On-Chip Microchannel Heat Sink,” IEEE Electron. Device Lett., 27(2), pp. 117–119. [CrossRef]
King, C. R. , Sekar, D. , Bakir, M. S. , Dang, B. , Pikarsky, J. , and Meindl, J. D. , 2008, “ 3D Stacking of Chips With Electrical and Microfluidic I/O Interconnects,” 58th Electronic Components and Technology Conference (ECTC 2008), Lake Buena Vista, FL, May 27–30.
King, C. R., Jr. , Zaveri, J. , Bakir, M. S. , and Meindl, J. D. , 2010, “ Electrical and Fluidic C4 Interconnections for Inter-Layer Liquid Cooling of 3D ICs,” 60th Electronic Components and Technology Conference (ECTC), Las Vegas, NV, June 1–4, pp. 1674–1681.
Oh, H. , Zhang, Y. , Zheng, L. , and Bakir, M. S. , 2014, “ Electrical Interconnect and Microfluidic Cooling Within 3D ICs and Silicon Interposer,” ASME Paper No. ICNMM2014-21813.
Zheng, L. , Zhang, Y. , Huang, G. , and Bakir, M. S. , 2014, “ Novel Electrical and Fluidic Microbumps for Silicon Interposer and 3-D ICs,” IEEE Trans. Compon., Packag. Manuf. Technol., 4(5), pp. 777–785. [CrossRef]
Green, C. E. , Kottke, P. A. , Sarvey, T. E. , Fedorov, A. G. , Joshi, Y. , and Bakir, M. S. , 2015, “ Performance and Integration Implications of Addressing Localized Hotspots Through Two Approaches: Clustering of Micro Pin Fins and Dedicated Microgap Coolers,” ASME 13th International Conference on Nanochannels, Microchannels, and Minichannels (InterPACK/ICNMM2015), San Francisco, CA, July 6–9.
Woodrum, D. C. , Sarvey, T. , Bakir, M. S. , and Sitaraman, S. K. , 2015, “ Reliability Study of Micro-Pin Fin Array for On-Chip Cooling,” IEEE 65th Electronic Components and Technology Conference (ECTC), San Diego, CA, May 26–29, pp. 2283–2287.

Figures

Grahic Jump Location
Fig. 1

Schematic of a typical two-phase forced flow heat sink embedded in 3D system

Grahic Jump Location
Fig. 2

Schematic of the F2/S2 hybrid heat-sink concept

Grahic Jump Location
Fig. 3

Ratio of thermal conductivity to microgap height (k/H), which serves as a proxy for heat transfer coefficient, as a function of outlet pressure for a variety of coolants [36]

Grahic Jump Location
Fig. 4

Scanning electron microscopy images of the fabricated electrical microbumps (25 μm diameter), fluidic microbumps (210 μm OD), fluidic via (100 μm), and micropin–fins (150 μm) [61,62]

Grahic Jump Location
Fig. 5

Schematic of the 3D STAECOOL design

Grahic Jump Location
Fig. 6

Micropin–fin heat sink [34]

Grahic Jump Location
Fig. 7

Micropin–fin sample fabrication process

Grahic Jump Location
Fig. 8

Flow loop for characterizing micropin–fin heat sinks

Grahic Jump Location
Fig. 9

Max wall temperature versus heat flux: 30 μm pin diameter, 60 μm pitch DI water at 1784 kg/m2 s [34]

Grahic Jump Location
Fig. 10

Schematic of 5 μm × 200 μm × 300 μm microgap with aligned Pt heaters [63]

Grahic Jump Location
Fig. 11

Temperature variation in the streamwise direction along the hotspot under varying thermal loads [63]

Grahic Jump Location
Fig. 12

Schematic of computational domain modeled for mechanical stress analysis

Grahic Jump Location
Fig. 13

Close-up view up computational domain for thermomechanical modeling showing TSVs, oxide layer, and Si pins

Grahic Jump Location
Fig. 14

Shear stress near the top of TSVs showing concentration in oxide liner (stress in MPa)

Grahic Jump Location
Fig. 15

Stress results (MPa) versus assumed copper stress-free temperature (°C)

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In