ASME Conference Presenter Attendance Policy and Archival Proceedings

2015;():V003T00A001. doi:10.1115/IPACK2015-NS3.

This online compilation of papers from the ASME 2015 International Technical Conference and Exhibition on Packaging and Integration of Electronic and Photonic Microsystems (InterPACK2015) represents the archival version of the Conference Proceedings. According to ASME’s conference presenter attendance policy, if a paper is not presented at the Conference, the paper will not be published in the official archival Proceedings, which are registered with the Library of Congress and are submitted for abstracting and indexing. The paper also will not be published in The ASME Digital Collection and may not be cited as a published paper.

Commentary by Dr. Valentin Fuster

Advanced Fabrication and Manufacturing: 3D Additive Manufacturing

2015;():V003T03A001. doi:10.1115/IPACK2015-48027.

Topology optimization of an air-cooled heat sink considering heat conduction plus side-surface convection is presented. The optimization formulation is explained along with multiple design examples. A post-processing procedure is described to synthesize water-tight solid model computer-aided design (CAD) geometry from 3-D point-cloud data extracted from the optimization result. Using this process, a heat sink is optimized for confined jet impingement air cooling. A prototype structure is fabricated out of AlSi12 using additive layer manufacturing (ALM). The heat transfer and fluid flow performance of the optimized heat sink is experimentally evaluated, and the results are compared with benchmark plate and pin-fin heat sink geometries that are conventionally machined out of aluminum and copper. In two separate test cases, the experimental results indicate that the optimized ALM heat sink design has a higher coefficient of performance relative to the benchmark heat sink designs.

Commentary by Dr. Valentin Fuster
2015;():V003T03A002. doi:10.1115/IPACK2015-48187.

Sub-mm wavelength 3-D antennas are emerging as critical elements for ultrafast data transfer for various applications. The inherent 2-D nature of lithographic processes severely limits the available manufacturing routes to fabricate such antennas. In this work, we demonstrate a novel additive manufacturing method to fabricate 3-D metal-dielectric antenna structures at sub-mm length scale. A UV curable dielectric is dispensed from an Aerosol Jet system and instantaneously cured to form complex 3-D shapes. A metal nano particle ink is then dispensed over the 3-D dielectric, also by the Aerosol Jet technique, followed by thermal sintering. This novel method opens up the possibility of fabricating an entirely new class of 3-D antenna structures at sub-mm length scales.

Commentary by Dr. Valentin Fuster
2015;():V003T03A003. doi:10.1115/IPACK2015-48637.

Polymeric materials have several favorable properties for heat transfer systems, including low weight, low manufacturing cost, antifouling, and anticorrosion. Additionally, polymers are typically electrical insulators, making them favorable for applications in which electrical conductivity is a concern. Examples of utilizing these favorable properties are discussed. The drawbacks to raw polymer materials include low thermal conductivity, low structural strength, and poor stability at elevated temperatures. Methods of mitigating these unfavorable properties, including loading the polymer with other materials and developing new polymers, are discussed. Enhanced geometric designs enabled by additive manufacturing can also improve thermal performance of polymer heat exchangers. Results of a research study utilizing additive manufacturing toward developing high-performance and cost-effective polymer heat exchangers for an air-to-liquid application are reviewed and discussed. Finally, needs for further research on enhancing polymer thermal performance are discussed.

Commentary by Dr. Valentin Fuster

Advanced Fabrication and Manufacturing: Novel Fabrication Methods for Micro- and Nano- Scale Devices

2015;():V003T03A004. doi:10.1115/IPACK2015-48046.

This paper reports on a study and application of laser ablation for machining of micro-serrations on surgical blades. The proposed concept is inspired by nature and mimics a mosquito’s maxilla, which is characterized by a number of serrations along its edge in order to painlessly penetrate human skin and tissue. The focus of this study is to investigate the maxilla’s penetration mechanisms and its application to commercial surgical blades. The fundamental objective is to understand the friction and cutting behavior between a serrated hard surface and soft materials, as well as to identify serration patterns that would minimize the cutting force and the friction of the blade during tissue cutting. Micro-serrations characterized by different patterns and sizes ranging from 200 μm to 400 μm were designed and manufactured on surgical blades. As supported by finite element methods (FEM), a reduction of 20∼30% in the force during blade cutting has been achieved, which encourages further studies and their applications to biomedical devices.

Topics: Surgery , Blades
Commentary by Dr. Valentin Fuster
2015;():V003T03A005. doi:10.1115/IPACK2015-48381.

A singular electric flux density field may occur when mechanical compression or electrical load to piezoelectric joints causes a singular electric flux density field is applied in piezoelectric material corners. This means that a large electric displacement occurs arround a singular point. It can be thought that there are many possible applications of this local large electric displacement. Hence, electric singularity characteristics should be investigated to increase the performance of piezoelectric joints.

It is known that the intensity of stress singularity reduces with decreasing the thickness of piezoelectric joints; conversely, the intensity of electric singularity increases under a constant external voltage and load. This implies that there is a relationship between electric singularity and stress singularity. In the present paper, the electric singularity on piezoelectric joints is investigated numerically. Usefulness of piezoelectric joints will then be discussed from the results of analysis.

Commentary by Dr. Valentin Fuster
2015;():V003T03A006. doi:10.1115/IPACK2015-48449.

Advective molding in vapor-permeable templates is an evaporation-driven process for submicron molding of nanoparticles with high fidelity. In this process, nanoparticle ink is drawn through channels in a vapor permeable template. The ink solvent is sorbed into the channel walls and evaporated through the template. As the complexity (e.g., width variation and turns in a channel) of the desired features increases, so does the likelihood of incompletely patterned nanoparticles. Patterning difficulties arise from dry-out, a condition where the nanoparticle ink dries before reaching the end of the channel and blocks the flow of more ink. Predicting dry-out during the template development stage is a critical step in patterning complex features. In this work, we present a method for predicting dry-out by incorporating two layers of finite element analysis. First, models for ink fluid flow and solvent diffusion through the template are used to determine wall sorption rate correlations. Fluid flow through complex templates is then modeled in a fluid-only model, with the flux rate into the template walls determined by the sorption rate correlations. The fluid velocities and wall sorption rates are then used to determine the likelihood of dry-out. The linked simulations successfully predict points of improper nanoparticle patterning in real templates.

Commentary by Dr. Valentin Fuster

Emerging Technology Frontiers: 2D Materials, Graphene and CNT Thermal Characterization and Applications

2015;():V003T04A001. doi:10.1115/IPACK2015-48213.

Determination of transport coefficients in 3D topological insulators is one of the highlighted topics in the literature. The main difficulty of the calculation of transport coefficients comes from the contribution of bulk modes. In this work, electrical conductivity in 3D topologic insulators is considered under zero magnetic fields and it is written in a size dependent form. For this purpose, Weyl conjecture is used. The contribution of the bulk modes to electrical conductivity is analytically derived by depending on the material size, hybridization energy and degeneracy of the carriers.

Commentary by Dr. Valentin Fuster
2015;():V003T04A002. doi:10.1115/IPACK2015-48239.

Graphene has great potential for ultra-sensitive strain sensors applications due to its high mechanical strength and good compatibility with the traditional semiconductor process. In the current study, we investigated the effect of tensile and bending deformations on the electronic states of graphene nanoribbons (GNRs) using density functional theory (DFT) to clarify the underlying mechanism of the piezoresistive properties of graphene. It is found that the electronic structure of armchair graphene nanoribbons (AGNRs) is very sensitive to the tensile deformation. When a uniaxial tensile stress is applied to AGNRs with width Na = 10, the band structure is modified, leading to the change in band gap approximately from 0 eV to 1.0 eV. The band gap values of bent AGNRs decrease significantly when the maximum local dihedral angle exceeds a critical value due to the orbital hybridization. Based on these knowledge, we fabricated a strain sensor using the graphene film grown by thermal chemical vapor deposition (CVD) method on Cu foil. The strain sensor is fabricated directly on the graphene-coated Cu foils by using the standard photolithography process and reactive ion etching (RIE) and then transferred onto a stretchable and flexible polydimethysiloxane (PDMS) substrate. The one-dimensional tensile test and three-dimensional bending test are performed to investigate the piezoresistive properties. A gauge factor 3.4 was achieved under the tensile deformation. The fabricated strain sensor also exhibits good performance to detect bending deformation.

Topics: Graphene
Commentary by Dr. Valentin Fuster
2015;():V003T04A003. doi:10.1115/IPACK2015-48297.

The spectral components of the phonon transport in the locally thermally excited graphene samples were studied by molecular dynamics (MD) method. In order to be able to select and analyze separate phonon modes in the time of propagation, the transient Green-Kubo approach to the definitions of density of states (DOS) and thermal conductivity was tested in quasi-equilibrium regimes for limited region of the graphene sample studied. Propagation of single modes at the background of diffusional phonon distribution and energy decay of such modes are studied by calculation of the DOS and dispersion relations, their dependence on the heating condition and temperature is studied. Similar conditions can be generated at localized heating of small areas of graphene structures in electronic devices. In transient regime, many issues of thermal transport evaluation still remain not sufficiently tested, especially phonon dynamics. Thermal conductivity of graphene samples related to transport of separate phonon modes is still not completely investigated, however, recent result give indication on the difference in the contribution of phonon modes. In the study, we consider mostly high temperature transport modes that are generated at the heated spot in order to be able to define their velocities and lifetimes in the limit of transient MD sampling.

The single-layer graphene nanoribbon of 150 nm to 40 nm was relaxed and prepared in equilibrium in zigzag and armchair orientations. REBO potential for graphene was utilized. Our calculation has shown that at the heating to high temperatures of 1000K and higher, the G mode of graphene remains stationary and has a minimal contribution into thermal transport by coherent modes. The coherent phonon mode or modes that contribute the most into thermal transport were confined in the vicinity of 30 THz and can possibly be attributed to the D modes of graphene.

Commentary by Dr. Valentin Fuster
2015;():V003T04A004. doi:10.1115/IPACK2015-48444.

Thermal boundary conductance (TBC) at Ti-graphene-Ti and Au-graphene-Au interfaces was measured using time-domain thermoreflectance (TDTR). Graphene grown on Cu foil by chemical vapor deposition was transferred to metallized Si substrates and TDTR was performed following the deposition of a top metal contact to form the metal-graphene-metal sandwiched structure. The results show that TBC was reduced at Au-graphene-Au interface in comparison to Au-Au interface and the value was similar to TBC measured at Au-HOPG interface. TBC at Ti-graphene-Ti interface was measured to be in the same range as Ti-Ti interface. This result is believed to be due to formation of oxide on surface of Ti, but further investigation is required. To the best of our knowledge, this is the first experimental study of TBC in metal-graphene-metal architecture.

Commentary by Dr. Valentin Fuster
2015;():V003T04A005. doi:10.1115/IPACK2015-48587.

Forests comprised of nominally vertically aligned carbon nanotubes (CNTs) are excellent candidates for thermal interface materials (TIMs) due to their theoretically predicted outstanding thermal and mechanical properties. Unfortunately, due to challenges in the synthesis and characterization of these materials reports of the thermal conductivity and thermal contact resistance of CNT forests have varied widely and typically fallen far short of theoretical predictions. In particular, the micro- and nano-length scales characteristic of the heat transfer in CNT forests pose significant challenges and may lead to misreported results. Here we examine the ability of a popular and well-established thermal metrology technique, time-domain thermoreflectance (TDTR), to resolve the properties of CNT forest TIMs. The characteristic heating frequencies of TDTR (1–10 MHz) are used to probe heat transfer at length scales spanning ∼0.1–1 μm, applicable for measuring the contact resistance between the CNT forest free tips and an opposing substrate. We identify the range of CNT forest-opposing substrate interface resistances that can be resolved with TDTR, and simultaneously demonstrate the effectiveness of several processes developed to reduce the resistance of these interfaces. The limitations of characterizing CNT forests with TDTR are discussed in terms of uncertainty and sensitivity to parameters of interest.

Commentary by Dr. Valentin Fuster

Emerging Technology Frontiers: Near Junction Thermal Transport in Electronic Devices

2015;():V003T04A006. doi:10.1115/IPACK2015-48179.

GaN on Diamond has been demonstrated to enable notable increases in RF power density without impacting High Electron Mobility Transistor (HEMT) peak junction temperature. However, Monolithic Microwave Integrated Circuits (MMICs) fabricated using GaN on Diamond substrates are subject to the same packaging thermal limitations as their GaN on SiC counterparts. Therefore, efforts to exploit GaN on Diamond to achieve substantial increases in MMIC power are stymied by external packaging thermal resistances that characterize the current “remote cooling” paradigm. This paper explores an intra-chip cooling alternative to the “remote cooling” paradigm, eliminating various heat spreader, heat sink and thermal interface layers in favor of integral microfluidic cooling in close proximity to the device junction. We describe an intra-chip cooling structure comprised of GaN on Diamond with integral micro-channels fed using a Si fluid distribution manifold. This structure exploits GaN on Diamond substrate technology to support increased HEMT areal power density while employing diamond microfluidics to affect scalable, low thermal resistance die-level heat removal. Thermal-electrical-mechanical co-design of integrated circuit (IC) features is performed to optimize conjugate heat transfer performance and manage the electrical and mechanical impacts associated with the presence of fluidic cooling near the electrically active region of the device. Through this, MMICs with significantly greater RF output than typical of the current state-of-the-art (SoA), dissipating die and HEMT heat fluxes in excess of 1 kW/cm2 and 30 kW/cm2, respectively, can be operated with junction temperatures that support reliable operation. The modeling, simulation and micro-fabrication results presented here demonstrate the potential of diamond microfluidics-based intra-chip cooling as a means to alleviate thermal impediments to exploitation of the full electromagnetic potential of GaN.

Commentary by Dr. Valentin Fuster
2015;():V003T04A007. doi:10.1115/IPACK2015-48429.

Under the DARPA-sponsored ICECool Applications program, a microchannel cooling system using a 50-50 ethylene glycol-water mixture was optimized for cooling a high-power GaN-on-Diamond Monolithic Microwave Integrated Circuit (MMIC). Automated multi-objective optimization of the microchannel passages yielded an optimized design with a predicted thermal resistance of 22.4 K·cm2/kW at a pressure drop of only 121.4 kPa for an inlet temperature of 40°C. These values were corroborated by a coupled thermofluid analysis that included a detailed treatment of both the gate region and microchannel cooling geometry. Several versions of prototype coolers were fabricated, with one set consisting of pairs of coolers joined at their heated faces. These cooler pairs were used in heat exchange tests to characterize the average thermal resistance and the flow performance of the coolers. The performance testing results were consistent with the analytic predictions. Based on the analytical and experimental results, the system may be operated at inlet temperatures as high as 65°C without exceeding the transistor junction temperatures of 240 °C required for 106 hour mean-time-to -failure. The higher inlet temperature ameliorates system penalties associated with rejection of waste heat to ambient heat sinks.

Commentary by Dr. Valentin Fuster

Emerging Technology Frontiers: Power Electronics and Electric Machines

2015;():V003T04A008. doi:10.1115/IPACK2015-48029.

The present work is generally related to the design of a manifold microchannel heat sink with high modularity and performance for electronics cooling, utilizing two well established (i.e., jet impingement and channel flow) cooling technologies. The present cold plate design provides flexibility to assemble manifold sections in five different configurations to reach different flow structures, and thus different cooling performance, without redesign. The details of the modular manifold and possible configurations of a cold plate comprising three manifold sections are shown herein. A conjugate flow and heat transfer 3-D model is developed for each configuration of the cold plate to demonstrate the merits of each modular design. Parallel flow configurations are used to satisfy a uniform cooling requirement from each module, but a “U-shape” parallel flow “base” configuration cools the modules more uniformly than a “Z-shape” flow pattern due to intrinsic pressure distribution characteristics. A serial fluid flow configuration requires the minimum coolant flow rate with a gradually increasing device temperature along the flow direction. Two mixed (i.e., parallel + serial flow) configurations achieve either cooling performance similar to the “U-shape” configuration with slightly more than half of the coolant flow rate, or cooling of a specific module to a much lower temperature level. Generally speaking, the current cold plate design significantly extends its application to different situations with different cooling requirements.

Commentary by Dr. Valentin Fuster
2015;():V003T04A009. doi:10.1115/IPACK2015-48259.

The development of gallium nitride (GaN) on silicon (Si) substrates is a critical technology for potential low cost power electronics. These devices can accommodate faster switching speeds, hotter temperatures, and high voltages needed for power electronics applications. However, the lattice mismatch and difference in crystal structure between 111 Si and c-axis hexagonal GaN requires the use of buffer layers in order to grow device quality epitaxial layers. For lateral high electron mobility transistors, these interfacial layers act as a potential source of increased thermal boundary resistance (TBR) which impedes heat flow out of the GaN on Si devices. In addition, these interfacial layers impact the growth and residual stress in the GaN epitaxial layer which can play a role in device reliability. In this work we use optical methods to experimentally measure a relatively low TBR for GaN on Si with an intermediate buffer layer to be 3.8 ± 0.4 m2K/GW. The effective TBR of a material stack that encompasses GaN on Si with a superlattice (SL) buffer is also measured, and is found to be 107 ± 1 m2K/GW. In addition the residual state of strain in the GaN layer is measured for both samples, and is found to vary significantly between them. Thermal conductivity of a 0.8μm GaN layer on AlN buffer is determined to be 126 ± 25 W/m-K, while a 0.84 μm GaN layer with C-doping on a SL structure is determined to be 112 ± 29 W/m-K.

Commentary by Dr. Valentin Fuster
2015;():V003T04A010. doi:10.1115/IPACK2015-48382.

Thermal management for electric machines (motors/generators) is important as the automotive industry continues to transition to more electrically dominant vehicle propulsion systems. Cooling of the electric machine(s) in some electric vehicle traction drive applications is accomplished by impinging automatic transmission fluid (ATF) jets onto the machine’s copper windings. In this study, we provide the results of experiments characterizing the thermal performance of ATF jets on surfaces representative of windings, using Ford’s Mercon LV ATF. Experiments were carried out at various ATF temperatures and jet velocities to quantify the influence of these parameters on heat transfer coefficients. Fluid temperatures were varied from 50°C to 90°C to encompass potential operating temperatures within an automotive transaxle environment. The jet nozzle velocities were varied from 0.5 to 10 m/s. The experimental ATF heat transfer coefficient results provided in this report are a useful resource for understanding factors that influence the performance of ATF-based cooling systems for electric machines.

Commentary by Dr. Valentin Fuster
2015;():V003T04A011. doi:10.1115/IPACK2015-48516.

This paper presents a detailed approach to provide improved cooling and heat spreading in electric machine rotors using centrifugally-pumped revolving thermosiphons. Design concepts are discussed that offer the following advantages: (1) high thermal performance across a wide range of operating points; (2) low-impedance heat paths; (3) excellent opportunities for integration with electric machine design for improved electromagnetic performance and structural design, as well as practical, cost-effective manufacturing. It takes advantage of centrifugal force to provide effective inertial pumping over a wide range of operating conditions. In addition, the new thermosiphon design is compatible with existing standard electric machine manufacturing techniques and cooling needs. A condenser section fin and ramp structure provides consistently high condensation performance. Surface texture design to promote effective nucleate boiling at high speeds is discussed, and fluid fill factor is analyzed. Applications include induction and PM synchronous machines. Benefits of these thermosiphons include increased steady-state power and torque density, increased and more consistent efficiency, and reduced permanent magnet volume and cost in PM synchronous machines. Other applications may include centrifugal gas compression, chemical processes, and machine tools.

Topics: Cooling , Engines , Motors , Rotors
Commentary by Dr. Valentin Fuster
2015;():V003T04A012. doi:10.1115/IPACK2015-48602.

In this paper, we describe the system-level packaging of a 30 kW continuous, 55 kW peak, traction inverter to showcase the electro-thermal-mechanical performance enhancements of silicon carbide (SiC), a wide bandgap (WBG) semiconductor, over silicon. Higher efficiency, larger gravimetric and volumetric power densities, and smaller thermal management system requirements may be achieved through higher operating junction temperatures afforded by SiC versus silicon power devices. By applying advanced system-level packaging techniques, high-temperature control circuitry, utilizing 105°C-rated capacitors, and reducing the number of system interconnects and attaches to enable higher system reliability, a substantial cost reduction from the die level to the system level can be demonstrated by completely eliminating an electric vehicle’s secondary low-temperature cooling loop. The endgame is to reduce the traction inverter size (≥ 13.4 kW [peak]/L), weight (≥ 14.1 kW [peak]/kg), and cost (≤ $182/100,000) relative to output power while maintaining 15-year reliability metrics [1].

Commentary by Dr. Valentin Fuster
2015;():V003T04A013. doi:10.1115/IPACK2015-48625.

In this paper, we present a non-intrusive case temperature measurement method of the direct-water-cooled power module. It uses the structure function, which in this case comprises the cumulative thermal capacitance and the cumulative thermal resistance. Since the effective heat transfer rate of the pinfin heatsink varies with the water flow rate, in this study we assumed the inflection point of the structure function corresponding to the change in the flow rate was junction-case thermal resistance. We compared numerical simulation results with experimental results and present our findings. Finally, we show that the design area in which the heat spreading angle of 45 degrees, the well-known rule of thumb, is suitable.

Commentary by Dr. Valentin Fuster
2015;():V003T04A014. doi:10.1115/IPACK2015-48714.

In this paper an attempt has been made to demonstrate various package design considerations to accommodate series connection of high voltage Si-IGBT (6500V/25A die) and SiC-Diode (6500V/25A die). The effects of connecting the cathode of the series diode to the collector of the IGBT versus connecting the emitter of the IGBT to the anode of the series diode has been analyzed in regards to gate terminal operation and the parasitic line inductance of the structure. ANSYS Q3D/MAXWELL software have been used to analyze and extract parasitic inductance and capacitances in the package along with electromagnetic fields, electric potentials, and current density distributions throughout the package for variable parameters. SIMPLIS-SIMETRIX is used to simulate typical switch behavior for different parasitic parameters under hard switched conditions. Various simulation results have then been used to redesign and justify the optimized package structure for the final current switch design. The thermal behavior of such a package is also conducted in COMSOL in order to ensure that the thermal ratings of the power devices is not exceeded, and to understand where potentially harmful hotspots could arise and estimate the maximum attainable frequency of operation. The main motivation of this work is to enumerate detailed design considerations for packing a high voltage current switch package.

Topics: Design , Switches , Packaging
Commentary by Dr. Valentin Fuster

Emerging Technology Frontiers: Thermal Management: LED

2015;():V003T04A015. doi:10.1115/IPACK2015-48326.

The demand for high power LEDs for illumination applications is increasing. LED package encapsulation is one of most critical materials that affect the optical path of the generated light by LEDs, and may result in lumen degradation. A typical encapsulation material is a mixture of phosphor and a polymer based binder such as silicone. After LED chips are placed at the base of a cavity, phosphor particles are mixed with silicone and carefully placed into the cavity. One of the important technical challenges is to ensure a better thermal conductivity than 0.2 W/m-K of current materials for most of the traditional polymers in SSL applications.

In this study, we investigated an unconventional material of the silk fibroin proteins for LED applications, and showed that this biomaterial provides thermal advantages leading to an order of magnitude higher thermal performance than conventional silicones. Silk fibroin is a natural protein and directly extracted from silk cocoons produced by Bombyx mori silkworm. Therefore, it presents a “green” material for photonic applications with its superior properties of biocompatibility and high optical transparency with a minimal absorption. Combining these properties with high thermal performance makes this biomaterial promising for future LED applications.

An experimental and computational study to understand the optical and thermal performance is performed. A computational fluid dynamics study with a commercial CFD software was performed and an experimental set-up was developed to validate the computational findings to determine the thermal conductivity of the proposed material.

Topics: Cavities , Proteins
Commentary by Dr. Valentin Fuster
2015;():V003T04A016. doi:10.1115/IPACK2015-48445.

In 2012, 12% of US electricity consumption was due to residential and commercial lighting. By 2030, LED lightbulbs will be more plentiful, projected to drive a 40% reduction in energy [1]. Yet today’s LED lightbulbs have bulky heat sinks that do not leverage modern thermal management techniques. Reducing the size of the heat sink offers lower cost and increased energy efficiency.

To leverage modern techniques, it is valuable to understand how thermal performance degrades light quality as people utilize lights over weeks and months. However, using current data acquisition systems to obtain such massive data sets is expensive, time-consuming, and cable-ridden.

We have developed a robust ground-to-cloud system to acquire temperature and luminosity data for building lighting systems. The data acquisition units are small and may be positioned in existing lighting installations. During acquisition, data is backed up automatically. Usability studies demonstrate that novice users can configure an experiment in less than four minutes.

Commentary by Dr. Valentin Fuster
2015;():V003T04A017. doi:10.1115/IPACK2015-48635.

This paper describes the workflow and improved accuracy of a combined characterization and simulation of an LED luminaire. The achieved measurement results of the thermal and radiometric characterization, the process of implementing these results into a thermal simulation, and the benefit for the luminaire designer are presented.

Commentary by Dr. Valentin Fuster

Emerging Technology Frontiers: Thermal Management: Mobile Applications

2015;():V003T04A018. doi:10.1115/IPACK2015-48175.

Vapor chamber technologies offer an attractive approach for passive cooling in portable electronic devices. Due to the market trends in device power consumption and thickness, vapor chamber effectiveness must be compared with alternative heat spreading materials at ultra-thin form factors and low heat dissipation rates. A test facility is developed to experimentally characterize performance and analyze the behavior of ultra-thin vapor chambers that must reject heat to the ambient via natural convection. The evaporator-side and ambient temperatures are measured directly; the condenser-side surface temperature distribution, which has critical ergonomics implications, is measured using an infrared camera calibrated pixel-by-pixel over the field of view and operating temperature range. The high thermal resistance imposed by natural convection in the vapor chamber heat dissipation pathway requires accurate prediction of the parasitic heat losses from the test facility using a combined experimental and numerical calibration procedure. Solid Metal heat spreaders of known thermal conductivity are first tested, and the temperature distribution is reproduced using a numerical model for conduction in the heat spreader and thermal insulation by iteratively adjusting the external boundary conditions. A regression expression for the heat loss is developed as a function of measured operating conditions using the numerical model. A sample vapor chamber is tested for heat inputs below 2.5 W. Performance metrics are developed to characterize heat spreader performance in terms of the effective thermal resistance and the condenser-side temperature uniformity. The study offers a rigorous approach for testing and analysis of new vapor chamber designs, with accurate characterization of their performance relative to other heat spreaders.

Commentary by Dr. Valentin Fuster
2015;():V003T04A019. doi:10.1115/IPACK2015-48367.

Use of Thermal Interface Materials (TIM) is a common thermal solution approach in handheld devices to reduce junction temperature and control device skin temperature. This work summarizes the thermal benefits of using a TIM for enhancing the user experience and increasing System on Chip (SoC) performance. On the other hand, TIM induces a load on the package which in turn can impose stress on the package solder joints. This paper explains the impact of a variety of parameters such as TIM material and thickness and system boundary conditions on thermal performance of the SoC/system and the load distribution on solder joints. The complexity of mechanical load distribution is discussed through extensive data collection and simulation in phone and tablet form factors. Design guidelines for selection of appropriate TIM are proposed to improve the thermal performance without compromising the reliability of the SoC package.

Commentary by Dr. Valentin Fuster
2015;():V003T04A020. doi:10.1115/IPACK2015-48523.

Even as the use of flexible graphite heat spreaders becomes ubiquitous in mobile electronics, numerically quantifying the heat dissipation remains a challenge. The rapid pace of development of mobile devices has deterred the industry from establishing standards, and rules of thumb are few, as are closed-form solutions. Users have requested numerical methods and tools to simplify the selection of flexible graphite heat spreaders from among the standard thicknesses and grades, as well as to quantify the effect of changing heat transfer area and configuration.

In the presence of adjacent layers — adhesives, dielectrics, or still air gaps — the thin nature of the materials and the high, orthogonal thermal conductivity ratios of the graphite combine to create a complex conjugate heat transfer problem. Although the thinnest of these sheets constitute but a tiny fraction of the thickness of a cell phone or tablet, their dominant role in the heat transfer requires that they not be neglected in the calculations. Some CFD software guidelines advise using multiple meshing layers to capture fully the heat transfer in these spreaders, while others (primarily FEA based) provide a plate element that negates the need for discretization. In the former, a fully meshed spreader confounds the goal of a quick calculation, but the flexibility of 3D solution also demands meticulous attention to the details, provides “an answer” that is easy to misinterpret, and in the hands of an unskilled user, invites error.

The goal of this project is to establish the guidelines for computing heat spreading in graphite, including cell dimension ratio, mesh density, spreading radius, and transport capacity and to marry the orthogonal properties of the material with the row-column format of a spreadsheet or matrix software. It also reviews methods for addressing the non-orthotropic situations such as angled plates, and the curved surfaces seen in the case of graphite wraps and flexible hinges. There are cases in which a simple contact resistance value adequately represent a graphite thermal interface material, but others that require an accounting for the lateral conductivity that increases the efficacy of the TIM. Finally, the error of the calculation is assessed for a simple representative geometry.

Commentary by Dr. Valentin Fuster
2015;():V003T04A021. doi:10.1115/IPACK2015-48563.

As demands on performance for mobile electronics continue to increase, traditional packaging technology is facing its limit in number of input/outputs (I/Os) and thermal challenges. Glass interposers offer many advantages over previous packaging technology for mobile electronics, including ultra-high electrical resistivity, low loss, and lower cost at processed interposer levels. However, it has two fundamental limitations; brittleness and relatively low thermal conductivity (∼1 W/mK), compared to Si (∼150 W/mK). This paper presents a study on thermal performance enhancement of glass interposer based on thermal modeling, and compares it with silicon interposer. The model captures in-plane and out-of-plane thermal performance enhancement with copper structures incorporated in the interposer. To further study the effect of advanced cooling schemes on interposer technology, an integrated vapor chamber design is evaluated through computational modeling.

Topics: Glass
Commentary by Dr. Valentin Fuster
2015;():V003T04A022. doi:10.1115/IPACK2015-48769.

Cutting edge tablet and phone designs are increasingly utilizing OLED (Organic Light Emitting Diode) displays. Although these displays are very thin and have excellent picture quality, they come at the cost of higher power than traditional displays which presents new thermal challenges. For thermal purposes, there are essentially 2 basic types of OLED displays based on their power distribution. In this presentation, I show thermal modeling methods I developed for both types of OLED displays and show the successful correlation of these models to experimental results. In addition, this presentation discusses some of the unique structural design requirements with OLED displays and ways to integrate both the thermal and structural solutions into an elegant combined solution. A thermal engineer can use the simulation methods and design recommendations described in this presentation to enable integration of this technology into products at an early stage of the design process.

Topics: Design , Modeling
Commentary by Dr. Valentin Fuster
2015;():V003T04A023. doi:10.1115/IPACK2015-48787.

It has been reported that tablet computer surface temperatures can rise from room temperature up to 47°C. Holding a warm or hot computer surface might cause user’s thermal discomfort and possibly skin burns. The use of a tablet often requires holding the device for prolonged time with multiple fingers and palm areas in contact with the tablet lower surface. Previous research has not tested whole finger/palm thermal sensation at a specific surface temperature in a moderate environmental heat range. The current research investigates user’s thermal sensations on the palm and fingers, in response to warm/heat stimuli in a tablet size device with a longer contact duration than used in previous studies, to provide ergonomic design guidelines for electronic device designers and manufacturers. A tablet-size heating surface was developed comprising of nine rectangular aluminum heating pads connected with computer-controlled heaters and thermal sensors. Participants were asked to report their finger/palm thermal sensation and comfort every 45 seconds when they held the prototype for 90 seconds. Results showed a positive linear relationship between surface temperature and user’s thermal sensation and thermal discomfort. Duration of holding the prototype had no significant effect on user’s thermal comfort, but it did significantly affect thermal sensation ratings.

Topics: Computers
Commentary by Dr. Valentin Fuster

Energy, Health and Water- Applications of Nano-, Micro- and Mini-Scale Devices: Biomicrofluidics and Lab-on-a-Chip

2015;():V003T05A001. doi:10.1115/IPACK2015-48134.

This paper reports on a proof-of-concept study of applying a two-dimensional (2D) microfluidic-based tactile sensor for tissue palpation under the influence of misalignment. Two unavoidable misalignment issues, uncertainty in contact point and non-ideal normal contact, severely distort the genuine elasticity distribution of a tissue region, yielding false identification of abnormality. The core of the 2D tactile sensor is one whole microstructure embedded with an electrolyte-enabled 2D resistive transducer array underneath. This unique configuration allows the tactile sensor to interact with a tissue region in a continuous manner that mimics manual palpation: the whole microstructure (fingertip) presses a tissue region and the corresponding deflection distribution is captured concurrently by the embedded transducer array (distributed sensors under the skin). This continuous manner tackles the misalignment issues encountered by an individual sensor or a sensor array, in that any misalignment encountered by the 2D sensor is manifested as an increasing trend of the distributed deflection-depth relations along the tilt direction. Tissue phantoms with embedded nodules and extrusions are prepared and are measured using the 2D tactile sensor, validating the capability of the tactile sensor to identify abnormalities in soft tissue under the influence of misalignment.

Commentary by Dr. Valentin Fuster
2015;():V003T05A002. doi:10.1115/IPACK2015-48491.

Hydrogen sulfide (H2S) is rapidly emerging as a biologically significant signaling molecule. In recent studies, sulfide level in blood or plasma has been reported to be in the concentration between 10 and 300 μM suggesting it acts in various diseases. This work reports progress on a new Lab-on-a-Chip (LOC) device for these applications. The uniquely designed, hand-held device uses advanced liberation chemistry that releases H2S from liquid sample and an electrochemical approach to detect sulfide concentration from the aqueous solution.

The device itself consists of three distinct layers of Polydimethylsiloxane (PDMS) structures and a three electrode system for direct and rapid H2S concentration measurement.

In this work specifically, the oxidation of sulfide at the gold (Au) and platinum (Pt.) electrodes has been examined. This is the first known application of electrochemical H2S sensing in an LOC application. The analytical utility and performance of the device has been assessed through direct detection using chronoamperometry (CA) scan and cyclic voltammetry (CV). An electrocatalytic sulfide oxidation signal has been recorded for sulfide concentration range vs, Ag/AgCl at different pH buffers at the trapping chamber. The calibration curve in the range 1 μM to 1 M was obtained using this electrode setup. The detection limit was found to be 0.1 μM. This device shows promise for providing fast and inexpensive determination of H2S concentration in aqueous samples.

Topics: Hydrogen , Biomedicine
Commentary by Dr. Valentin Fuster

Energy, Health and Water- Applications of Nano-, Micro- and Mini-Scale Devices: Fuel Cells and Other Energy Devices

2015;():V003T05A003. doi:10.1115/IPACK2015-48816.

In this study, development of a novel system for combined water heating, dehumidification, and space cooling is discussed. The system absorbs water vapor from an air stream into an absorbent. The latent heat of absorption, released into the absorbent, is transferred into the process water that cools the absorbent. The solution is regenerated in the desorber, where it is heated by a heating fluid. The water vapor generated in the desorber is condensed and its heat of phase change is also transferred to the process water. The condensed water is then used in an evaporative cooling process to cool the dehumidified air exiting the absorber. Essentially, this open-absorption cycle collects space sensible heat and transfers it to hot water. Another novel feature of the cycle is recovery of the heat energy from the solution exiting the desorber by heat exchange with process water rather than with the solution exiting the absorber. This approach has enabled heating the process water from an inlet temperature of 15°C to 57°C (conforming to the required DOE building hot water standard) and compact fabrication of the absorber, solution heat exchanger, and desorber in plate and frame configuration. The system under development currently has a water heating capacity of 1.5 kW and a thermal coefficient of performance (COP) of 1.45.

Commentary by Dr. Valentin Fuster

Energy, Health and Water- Applications of Nano-, Micro- and Mini-Scale Devices: Mixing, Mass Transfer, and Chemical Reactions

2015;():V003T05A004. doi:10.1115/IPACK2015-48767.

We aimed to effectively use unutilized rice straw by producing Bio-coke, which is a new briquette (in the rest of this document referred to as the BIC) with high density and hardness, from rice straw with various conditions based on initial water content and processing temperature and evaluated characteristics of rice straw BIC. First of all, the apparent density of BIC was calculated from its weight and volume, and the cold compressive strength for each BIC was measured. From the results, it showed that the relationship between apparent density and maximum compressive strength derived from the compression test had a positive correlation. Furthermore, the hot compressive strength of the BIC produced with 5% initial water content and 453K processing temperature was measured. The rice straw BIC had a maximum compressive strength of 4.8MPa at a high temperature of 973K. This hot maximum compressive strength is equal to about one third of the hot maximum compressive strength of coal coke, which is 12MPa. Also, it was determined that the maximum compressive strength of rice straw BIC is highest on both cold and hot compression tests, and BIC produced from agricultural biomass like rice straw and rice husk had higher maximum compressive strength at room and high temperatures than BIC produced from other materials. Thus, it seemed that fiber and silica contained in agricultural biomass helps maintain of structure of BIC.

Topics: Coke , Biomass
Commentary by Dr. Valentin Fuster

Energy, Health and Water- Applications of Nano-, Micro- and Mini-Scale Devices: Transport in Membranes

2015;():V003T05A005. doi:10.1115/IPACK2015-48298.

Development of the 1D and 2D IR spectroscopy of small organic molecules and clusters opens yet another way of possible identification of small organic molecules in the state of motion in the graphene nanopore scanning device. With the advantage of obtaining qualitative and at least semi-quantitative information of specimens real-time and non-invasively, vibrational spectroscopy techniques, infrared (IR) and Raman have become more and more important in the analysis of biomolecular samples. At present, the sensitivity and spatial resolution of these techniques stands at the challenge of the detection and analysis of biosamples at very low concentration (single molecule) and high spatial resolution (nanometer/sub-nanometer scale). Spectral analysis requires theoretical assignment of vibrational modes to each biomolecule.

We considered vibrational spectra of DNA nucleobase at the time when they are translocated through the graphene nanopore. The Fourier transform of the density of states (DOS) of each base was calculated and the spectra of the base molecules and C atoms of graphene pore edge were obtained. Translocation rate was fixed to have maximum interaction of the base with 1.5 nm pore and single orientation of nucleobases was evaluated relative to molecular plane. Whether interaction of nucleobase and nanopore is able to enhance the signal is still remains unanswered. But we have shown that the spectra of each nucleobase are different and can be considered the fingerprint of the particular molecule.

The interaction forces between pore and base are structure dependent and time-limited by translocation time. In such case, transient correlation functions were utilized for the DOSes of the individual bases and forces on each atom of the particular base were sorted by intensity. The spectra of individual atoms in the bases as well as of whole molecule were compared and frequencies of most intense peaks were related to particular atoms. Molecular dynamics method is used for the DNA base and graphene nanopore calculations with the MM2/MM3 potentials for the base and REBO graphene potential. Interaction potential between the bases can simultaneously give additional information for the electronic transport calculations with possible tra and graphene are of the MM2/MM3 part of the Van der Waals interaction only has been considered. Possibility of base identification by spectral signature is confirmed. Calculated spectra are compared with results of the existing IR measurements for nucleobases.

Commentary by Dr. Valentin Fuster

MEMS and NEMS: N/MEMS: Emerging Technology

2015;():V003T07A001. doi:10.1115/IPACK2015-48286.

Recently two-dimensional layered semiconductors with promising electronic and optical properties, have opened up a new way for applications in atomically thin electronics and optoelectronics. Here we demonstrate large area synthesis of monolayer MoS2 and WSe2 using a chemical vapor deposition method at ambient pressure. The atomic analysis of the as-grown monolayer was conducted by spherical-aberration-corrected high resolution scanning transmission electron microscopy and atomic force microscopy. Raman spectroscopy was utilized to identify the monolayer configuration of the as-grown samples. Strong photoluminescence peaks at a visible wavelength were observed at room temperature in the as-grown monolayer samples. The mobility and carrier concentrations were calculated in as-grown monolayer-based transistor devices. The emergency of these two dimensional materials provides grand possibilities for future semiconductor device applications.

Commentary by Dr. Valentin Fuster
2015;():V003T07A002. doi:10.1115/IPACK2015-48585.

Two types of on-chip RFIC transformers based on CMOS compatible strain-induced self-rolled-up membrane (S-RuM) nanotechnology, with extremely small footprint, are demonstrated. The rolled-up transformers, with their 3D tubular form factors, dramatically reduce the substrate parasitic effects and push the maximum working frequency into millimeter wave bands with a coupling coefficient, k, as high as 0.92. The 3D stand-up nature also allows the tube transformers to be less susceptible to residue stress in the substrate and thus compatible with flexible platforms for wearable RF applications. The demonstrated samples with a turn ratio, n, of 5.5:1 only occupies 805 μm2 on-chip area (s) which is 12x smaller than that of the best planar transformer with the same turn ratio, and its figure of merit n·k/s, is therefore ∼ 6046/mm2, enhanced by 15x.

Commentary by Dr. Valentin Fuster
2015;():V003T07A003. doi:10.1115/IPACK2015-48590.

We report on experimental demonstration of multilayer molybdenum disulfide (MoS2) nanomechanical resonators integrated on microchannels, with the potential for resonant operation and sensing applications in microfluidics. Microchannels with width of ∼5μm and depth of ∼1–2μm are fabricated on polydimethylsiloxane (PDMS) substrate. We transfer MoS2 flakes with both singly-clamped cantilever and doubly-clamped membrane structure onto the PDMS channels, and measure both undriven thermomechanical resonances and optically driven responses from the devices. The devices show up to 6 thermomechanical resonances, with highest resonance frequency (fres) of 16.3MHz and quality (Q) factor ∼200. This type of MoS2 nanomechanical resonators, when integrated with microfluidics in microchannels, would make new interesting candidates for biosensing and chemical sensing in fluids.

Commentary by Dr. Valentin Fuster

MEMS and NEMS: N/MEMS: Fabrication, Integration, Packaging, Test and Reliability

2015;():V003T07A004. doi:10.1115/IPACK2015-48010.

Microelectromechanical system (MEMS) packages are vulnerable to stresses due to its functional structure. During the assembly process of the package, stresses stemming out of CTE mismatches of the structural elements and curing of the die attach material can cause warpage of the MEMS die [1]. Even though die attach material takes relatively small volumetric portion of the package, it plays a critical role in warpage of the die due to its location and sensitivity of a MEMS sensor.

Most of virgin die attach adhesives are in a state of viscous liquid and, as it is cured the material properties such as modulus and CTE change. Accordingly, residual strain is cumulated on MEMS die after curing process and signal trim process is required. Therefore, the material properties changes depending on the curing profile is valuable information for assembly process of the MEMS package.

To monitor the material properties changes and shrinkage during curing process, strain and modulus of a die attach material are measured in each curing step. Also, to investigate the material property change depending on the curing profile, two different curing profiles are used.

Experimental data show that die attach materials are gradually cured after each thermal cycling, which cause the increment of the modulus and glass transition temperature (Tg) with shrinkage at elevated temperature. Using the measurement data, FEA model is built to predict the warpage of the MEMS die. In the FEA model, residual strain on MEMS die is calculated by inputting material properties of die attach in each curing step. Also, die warpage of the package during the curing process is monitored using an optical profiler for the validation of the simulation results.

Commentary by Dr. Valentin Fuster
2015;():V003T07A005. doi:10.1115/IPACK2015-48086.

Polymeric die-attach adhesives have been widely utilized for attaching the sensor die to the lead-frame substrate in the packaging of micro-electro-mechanical systems (MEMS). Although much study has been performed regarding viscoelastic behaviors of packaging materials, less literature has addressed the issue of post-cure shrinkage, which occurs in isothermal storage and in accelerated tests. However, the influence of post-cure shrinkage cannot be ignored in characterizing long-term viscoelastic properties or evaluating long-term reliability for highly sensitive automotive sensors.

In this study, a non-contact digital image correlation (DIC) method was used to characterize the post-cure aging shrinkage on a commercial epoxy-based die attach adhesive. The shrinkage strain of bulk specimens was monitored over a period of isothermal aging. Analysis indicates that the process of postcure shrinkage can be expressed in terms of exponential time functions. Using hygroscopic shrinking as an analogous process, this aging was introduced into a time-dependent finite element (FE) simulation in ANSYS®15.0. FE case studies combining both stress and dimensional relaxation indicate that certain differences can occur in the long-term finite element analysis if the shrinkage strain is ignored.

Commentary by Dr. Valentin Fuster
2015;():V003T07A006. doi:10.1115/IPACK2015-48409.

High aspect ratio microchannels using high thermal conductivity materials such as silicon carbide (SiC) have recently been explored to locally cool micro-scale power electronics that are prone to on-chip hot spot generation. Analytical and finite element modeling shows that SiC-based microchannels used for localized cooling should have high aspect ratio features (above 8:1) to obtain heat transfer coefficients (300 to 600 kW/m2·K) required to obtain gallium nitride (GaN) device channel temperatures below 100°C. This work presents experimental results of microfabricating high aspect ratio microchannels in a 4H-SiC substrate using inductively coupled plasma (ICP) etching. Depths of 90 μm and 80 μm were achieved with a 5:1 and 12:1 aspect ratio, respectively. This microfabrication process will enable the integration of microchannels (backside features) with high-power density devices such as GaN-on-SiC based electronics, as well as other SiC-based microfluidic applications.

Commentary by Dr. Valentin Fuster
2015;():V003T07A007. doi:10.1115/IPACK2015-48457.

MEMS accelerometers have found applications in harsh environments with pressure, temperatures above ambient conditions, high g shock and vibrations. The complex structure of these MEMS devices has made it difficult to understand the failure modes and failure mechanisms of present day MEMS accelerometers. Little work has been done by the researchers in investigating the high g reliability of the MEMS accelerometers by continuous high g drops and quantifying the failure modes. There is little literature addressing the multiphysics finite element modelling of MEMS accelerometers subjected to high g shocks. In defense applications, where these devices are integrated with several other compactly assembled subsystems, lack of knowledge on the physics of failure for the MEMS sensor in harsh environment operation, can be detrimental to the success of the system on the whole. Being able to successfully model inside of an accelerometer, enables the user to better understand the change in parameters like time delay induced in response of successive drops, change in pulse width that indicate failure, reduction in sensed g levels. Some researchers have subjected various accelerometers to repeated drops at their maximum sensing g(not high g) level, and used optical microscopy to detect damaged sensing elements [Beliveau, 1999]. Few researchers have modeled the internal structure of the MEMS device, along with the device packaging under the stresses of operation [Fang 2004, Ghisi 2008, Xiong 2008]. In this paper, a multiphysics model of capacitive and the moving elements of the accelerometer has been developed to model the change in capacitance with respect to stroke and understand the correlation with g-levels, in addition to the transient dynamic response of the accelerometer under high-g shock. This has not been much explored in the past. The accelerometer studied in the paper is the ADXL193, and subjected to repeated drops of 3000g in each 3 axes as per 2002.4 of MIL-STD-883 without preconditioning. A characteristic graph of capacitance vs accelerometer stroke has been obtained from a series of electrostatic simulations and is then used to relate g levels, capacitance, stroke deflection and voltage change using electromechanical transducer elements. The drift in the performance characteristics of the accelerometer have been measured versus the number of shock events. In addition, an attempt has been made to investigate the failure mode in the accelerometer.

Commentary by Dr. Valentin Fuster
2015;():V003T07A008. doi:10.1115/IPACK2015-48611.

Ultrasound is increasingly in demand as a medical imaging tool and can be particularly beneficial in the field of intracardiac echocardiography (ICE). However, many challenges remain in the development of a 3D ultrasound imaging system.

We have designed and fabricated a quad-ring capacitive micromachined ultrasound transducer (CMUT) for real-time, volumetric medical imaging. Each CMUT array is composed of four concentric, independent ring arrays, each operating at a different frequency, with 128 elements per ring. In this project, one ring will be used for imaging. A large (5mm diameter) lumen is available for delivering other devices, including high intensity focused ultrasound transducers for therapeutic applications or optical fibers for photoacoustic imaging.

We address several challenges in developing a 3D imaging system. Through wafer vias are incorporated in the fabrication process for producing 2D CMUT arrays. Device integration with electronics is achieved through solder bumping the arrays, designing a flexible PCB, and flip chip bonding CMUT and ASICs to the flexible substrate. Finally, we describe a method for integrating the flex assembly into a catheter shaft. The package, once assembled, will be used for in-vivo open chest experiments.

Commentary by Dr. Valentin Fuster
2015;():V003T07A009. doi:10.1115/IPACK2015-48814.

A key component of the MEMS cryogenic cooling systems we are developing is a MEMS compressor. Its function is to compress the refrigerants used in the cooling cycle. Layers of polyimide are stacked and patterned on a silicon wafer to create micro check-valves and a compression chamber over which a diaphragm is suspended. To achieve the high (4:1) pressure ratio needed for the refrigeration cycle, the polyimide diaphragm needs to be fabricated with minimal dead volume beneath it, hence the need for a sacrificial layer with thickness of 100–300 nm. The topography created by the check-valves and valve-seats makes atomic layer deposition (ALD) ideal due to its conformality. Furthermore, following sacrificial layer release, the inside of the compression chamber will also need to be coated with a hermetic moisture barrier layer to enable the device to operate at 4 atmospheres without leaking. ALD is therefore also ideal for the final internal coating because it does not require line-of-sight. Towards this end, we demonstrate here the concept of using ALD TiO2 as a sacrificial layer to create a 5 mm × 5 mm × 20 m-thick polyimide membrane, suspended by ∼ 100–350 nm above a silicon wafer, followed by a second thinner ALD coating of the interior surfaces bounded by wafer and membrane. The air gap under the membrane, defined by the released sacrificial layer, was measured at about 130–370 nm using two independent methods: reflectometry, and FIB cross sectioning followed by SEM imaging of the air gap’s cross section. The membrane was removed from one chip and the thickness of the internal coating on the underlying silicon was measured with the reflectometer to be about 40 nm. We thus demonstrate the use of ALD TiO2 as both a sacrificial layer for fabricating nanoscale gaps, as well as for coating nanoscale internal cavities and channels.

Commentary by Dr. Valentin Fuster

MEMS and NEMS: N/MEMS: Sensors, Actuators, and Resonators

2015;():V003T07A010. doi:10.1115/IPACK2015-48133.

This paper reports on a microfluidic-based tactile sensor capable of detecting forces along two directions and torque about one direction. The 3-Degree-Of-Freedom (3-DOF) force/torque sensor encompasses a symmetric three-dimensional (3D) microstructure embedded with two sets of electrolyte-enabled distributed resistive transducers underneath. The 3D microstructure is built into a rectangular block with a loading-bump on its top and two microchannels at its bottom. Together with electrode pairs distributed along the microchannel length, electrolyte in each microchannel functions as a set of three resistive transducers. While a normal force results in a resistance increase in the two sets of transducers, a shear force causes opposite resistance changes in the two sets of transducers. Conversely, a torque leads to the opposite resistance changes in the two side transducers in each set. Soft lithography and CNC molding are combined to fabricate a prototype tactile sensor. The experimental results validate the feasibility of using this microfluidic-based tactile sensor for 3-DOF force/torque detection.

Commentary by Dr. Valentin Fuster
2015;():V003T07A011. doi:10.1115/IPACK2015-48567.

We demonstrate the first contact resistance measurements of graphene–galinstan (g-g) ohmic contacts in an effort to improve the performance of graphene photonic devices. The nobility of carbon materials provide an interesting graphene sensor application to explore an oxidation free liquid metal - semimetal interface that can be used to lower contact resistance at source/drain terminals of a standard graphene phototransistor. Our methods utilize photopolymerization of the reactive monomer Trimethlylolpropane Triacrylate (TMPTA) in order to fabricate micro structures necessary to overlay liquid metal contacts on graphene. With the use of an industry standard transfer length method (TLM), a contact resistance of −124±28Ω was measured at both standard temperature and pressure. The results from our study suggest that liquid metals such as galinstan are comparable alternatives to rigid semiconductor interfaces and demonstrates interesting boundary characteristics that may lead to heavy chemical doping and associated low resistance contacts that are required to increase sensitivity in graphene photonic devices.

Commentary by Dr. Valentin Fuster
2015;():V003T07A012. doi:10.1115/IPACK2015-48733.

We present the operation of capacitive micromachined ultrasonic transducers (CMUTs) in permanent contact mode as an efficient transducer. The gap height of our transducers is chosen to be slightly smaller than the static deflection of the plate due to the pressure difference between the ambient and the vacuum cavity. Thus, the plates are in contact with the bottom of the cavities even with no dc bias applied. The devices were fabricated based on the thick box process. High-temperature assisted direct wafer bonding technique was used to fabricate devices with such large cell size (radii ∼ 2000 μm) featuring low frequencies ∼100–150 kHz. Extensive acoustic characterization was performed to demonstrate the behavior of such CMUTs in terms of displacement profile, output pressure and acoustic pitch-catch response. A maximum sound pressure of ∼145 dB (SPL) at the transducer surface is measured at 240 V dc and 10 V ac with 100 cycles of burst signal. This is a great improvement from conventional CMUTs (with deeper gap height, operating at 55 kHz), which requires 350 V dc and 200 V ac in order to achieve an output pressure of 129 dB (SPL) at the transducer surface. The results presented in this paper demonstrate that operating CMUTs in permanent contact mode indeed enhances the device output pressure, and provides a good candidate for efficient ultrasonic transducers.

Topics: Pressure
Commentary by Dr. Valentin Fuster
2015;():V003T07A013. doi:10.1115/IPACK2015-48790.

Reliability data on MEMS accelerometers operating in harsh environments is scarce. Micro-electro-mechanical systems (MEMS) are used in a variety of military and automotive applications for sensing acceleration, translation, rotation, pressure and sound. This research work focuses on dual axis MEMS accelerometer reliability in harsh environments. Structurally an accelerometer behaves like a damped mass on a spring. Commercially there are three types of accelerometers namely piezoelectric, piezoresistive and capacitive depending on the components that go into the fabrication of the MEMS device. Previously, majority of concentration was focused on an effective internal design, performance enhancement of CMOS-MEMS accelerometers and packaging techniques Cheng [2002], Qiao [2009], Lou [2005], and Weigold [2001]. Studies have also been conducted to obtain an enhanced inertial mass SOI MEMS process using a high sensitivity accelerometer Jianbing [2013], Chen [2005]. There have been prior test(s) conducted on MEMS accelerometers, Jiang [2004], Cao [2011], Chun-Sun [2009], Lou [2009], Tanner [2000] and Yang [2010] but the availability of data on reliability degradation of such devices in harsh environments Brown [2003] is almost little to none which thereby generates the importance of this work and also makes way for a whole new path involving the reliability assessment techniques for MEMS devices. Concentration of our work is primarily on the reliability of this accelerometer upon sequential exposure to harsh environment(s) and drop-shock. Reliability of accelerometers in high G environments is unknown. The effects of these pre-conditions along with the drop test condition has been studied and analyzed. In this piece of research work, a test vehicle with a MEMS accelerometer, ADXL278 dual axis capacitive accelerometer, has been tested under high/low temperature exposure followed by subjection to high-g and low-g shock loading environments. The test boards have been subjected to mechanical shocks using the method 2002.5, condition G, under the standard MIL-STD-883H test. The stress environment and the test condition used for this paper are 1500g and 70g respectively where 70g is the full scale range output of ADXL278 in the drop direction with pulse duration set to 0.5millisecond. The deterioration of the accelerometer output has been characterized using the techniques of Mahalanobis distance and Confidence intervals. Scanning Electron Microscopy (SEM) has been used to study the different failure modes inside of the accelerometer, which were potted and polished and later de-capped. Furthermore, the non-destructive evaluations of the MEMS accelerometer have been demonstrated through X-rays and micro-CT scans.

Commentary by Dr. Valentin Fuster

Technology Update Talks

2015;():V003T08A001. doi:10.1115/IPACK2015-48145.

Microwave and power electronics based on GaN enables the performance of systems and their safe operating area to be driven to ‘extremes’. One of the major issues that then arises is thermal management. This includes heat transfer limitations across interfaces, however also the need of incorporating novel high thermal conductivity materials such as diamond. Thermal parameters of these novel device systems and their implications on the near junction temperature in the devices are not well known. The role of interfaces between the GaN transistor and the diamond substrate, and of the diamond thermal properties themselves near this interface are discussed, and novel thermal characterization approaches, such as enabling fast determination of the thermal resistance on the wafer level, as well as of lateral diamond thermal conductivity, are presented.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Advanced and Oscillating Heat Pipes

2015;():V003T10A001. doi:10.1115/IPACK2015-48665.

Self-rewetting fluids (SRWFs) are non-azeotropic solutions enjoy a particular surface tension behavior — an increase in the surface tension with increasing temperature. Due to the unique property, the SRWF can spontaneously wet hotter region and enhance heat transfer. The interesting behavior makes the SRWF become the research hotspot in phase change heat transfer research field. To clarify the heat transfer characteristics of SRWF, a series of boiling experiments have been carried out by employing dilute heptanol aqueous solution as SRWF. It is found out that, the bubble size of the SRWF is much smaller than that of pure water, and the critical heat flux of SRWF is much higher than that of water, which is beneficial for application in heat pipes. To find out the heat transfer performance of SRWF in heat pipes, experimental studies are performed on oscillating heat pipe (OHP) consisting of 4 meandering turns, with heat transfer length (L) of 150 mm and inner diameter (Di) of 1.3 mm. Compared with the water, the SRWF exhibits much better thermal performance, which indicates that SRWF is a promising and useful working liquid for the application in high efficient cooling devices with micro structure.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Cold Plate Integration

2015;():V003T10A002. doi:10.1115/IPACK2015-48154.

Computer processor speeds have increased in recent years to the extent that water cooling is becoming an attractive and sometimes a necessary way of cooling the processors and their associated electronics. Water can conduct heat much faster than air, allowing processors to run at higher speeds at lower acoustic levels. The ability of water to cool computer electronics can be diminished, or even lost, if there is excessive loss of water by permeation through, or leaks past, the various materials and/or connections between materials. Knowing the rate of loss of coolant from cooling systems can help designers determine the maintenance procedures and schedules for their cooling systems. The paper describes a novel, accurate and convenient method for measuring the moisture leakage rates out of water-carrying hardware. The water-filled hardware under test is placed in a chamber that is purged dry with flowing nitrogen gas and the chamber is then sealed. The rate of rise of relative humidity in the chamber is used to determine the rate of moisture leakage out of the water-filled hardware. The errors arising from the hose terminations and the adsorption of moisture by the metal chamber walls and the plastic fittings can be accounted for and corrected. The test duration is typically less than 10 days. The paper presents examples of water leakage out of hoses and tubes and their terminations and out of quick connects.

Topics: Cooling , Hardware , Water , Leakage
Commentary by Dr. Valentin Fuster
2015;():V003T10A003. doi:10.1115/IPACK2015-48324.

When contemplating processor module cooling, the notion of maximum cooling capability is not simple or straight forward to estimate. There are a multitude of variables and constraints to consider; some more rigid or fixed than others. This paper proposes a theoretical maximum cooling capability predicated on the treatment of the module heat sink or cold plate as a heat exchanger with infinite conductive and convective behavior. The resulting theoretical minimum heat sink thermal resistance is a function of the bulk thermal transport of the fluid dependent only on the fluid’s density, specific heat (at constant pressure) and volumetric flow rate. An ideal module internal thermal resistance will also be defined. The sum of the two resistances constitutes the theoretical minimum total module thermal resistance and defines the ideal thermal performance of the module. Finally, a module cooling effectiveness relating the actual module thermal performance to the ideal thermal performance will defined. Examples of both air and water cooled modules will be given with discussion on the relevance and utility of this methodology.

Topics: Cooling
Commentary by Dr. Valentin Fuster
2015;():V003T10A004. doi:10.1115/IPACK2015-48427.

We demonstrate the Lid-Integral Silicon Coldplate topology as a way to bring liquid cooling closer to the heat source IC. It allows to eliminate one thermal interface material (TIM2), to establish and improve TIM1 during packaging, to use wafer-level processes, and to ease integration in 1st level packaging. We describe the integration, and analyze reliability aspects of this package using modeling and test vehicle builts. To compare the impact of geometry, materials and mechanical coupling on warpage, strains and stresses, we simulate finite element models of five different topologies on an organic LGA carrier. We measure the thermal performance in terms of thermal resistance from coldplate base to inlet liquid and obtain 15mm2K/W at 30 kPa pressure drop across the package. We build two different topologies using silicon coldplates and injection molded lids. Gasket-attached coldplates pass an 800 kPa pressure test, direct-attached coldplates fracture in the coldplate. The results advise to use a compliant layer between coldplate and the manifold lid and promise a uniformly thick TIM1 layer in the Si-Si matched topology. The work shows the feasibility of composite lids with integrated silicon coldplates in high heat flux applications.

Topics: Reliability , Topology
Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Embedded Cooling for 3D ICs

2015;():V003T10A005. doi:10.1115/IPACK2015-48103.

In this paper we reported an advanced structure, the Piranha Pin Fin (PPF), for microchannel flow boiling. Fluid flow and heat transfer performance were evaluated in detail with HFE7000 as working fluid. Surface temperature, pressure drop, heat transfer coefficient and critical heat flux (CHF) were experimentally obtained and discussed. Furthermore, microchannels with different PPF geometrical configurations were investigated. At the same time, tests for different flow conditions were conducted and analyzed. It turned out that microchannel with PPF can realize high-heat flux dissipation with reasonable pressure drop. Both flow conditions and PPF configuration played important roles for both fluid flow and heat transfer performance. This study provided useful reference for further PPF design in microchannel for flow boiling.

Commentary by Dr. Valentin Fuster
2015;():V003T10A006. doi:10.1115/IPACK2015-48341.

Hot spots and temperature non-uniformities are critical thermal characteristics of current high power electronics and future three dimensional (3D) integrated circuits (ICs). Experimental investigation to understand flow boiling heat transfer on hot spots is required for any two-phase cooling configuration targeting these applications. This work investigates hot spot cooling utilizing novel radial microchannels with embedded pin arrays representing through-silicon-via (TSV) interconnects. Inlet orifices were designed to distribute flow in radial channels in a manner that supplies appropriate amounts of coolant to high-power-density cores. Specially designed test vehicles and systems were used to produce non-uniform heat flux profiles with nominally 20 W/cm2 background heating, 200 W/cm2 core heating and up to 21 W/mm2 hot spot (0.2 mm × 0.2 mm) heating to mimic a stackable eight core processor die (20 mm × 20 mm) with two hot spots on each core. The temperatures associated with flow boiling heat transfer at the hot spots were locally measured by resistance temperature detectors (RTDs) integrated between the heat source and sink. At nominal pressure and flow conditions, use of R1234ze in these devices resulted in a maximum hot spot temperature (Ths) of under 63 °C and average Ths of 57 °C at a hot spot power density of 21 W/mm2. A semi-empirical model was used to calculate the equivalent heat transfer rate around the hot spots which can provide a baseline for future studies on local thermal management of hot spots.

Commentary by Dr. Valentin Fuster
2015;():V003T10A007. doi:10.1115/IPACK2015-48348.

Thermal performance for embedded two phase cooling using dielectric coolant (R1234ze) is evaluated on a ∼20 mm × 20 mm large die. The test vehicles incorporate radial expanding channels with embedded pin fields suitable for through-silicon-via (TSV) interconnects of multi-die stacks. Power generating features mimicking those anticipated in future generations of processor chips with 8 cores are included. Initial results show that for the types of power maps anticipated, critical heat fluxes in “core” areas of at least 350 W/cm2 with at least 20 W/cm2 “background” heating in rest of the chip area can be achieved with less than 30 °C temperature rise over the inlet coolant temperature. These heat fluxes are significantly higher than those seen for relatively long parallel channel devices of similar base channel dimensions. Experimental results of flow rate, pressure drop, “device,” and coolant temperature are also provided for these test vehicles along with details of the test facility developed to properly characterize the test vehicles.

Topics: Cooling
Commentary by Dr. Valentin Fuster
2015;():V003T10A008. doi:10.1115/IPACK2015-48584.

In this paper, a novel thermal testbed with an embedded micropin-fin heat sink is designed and fabricated. The micropin-fin array has a nominal height of 200 μm and a diameter of 90 μm. Single phase and two phase thermal testing of the micropin-fin array heat sink are performed using deionized (D.I.) water as the coolant below atmospheric pressure. The measured pressure drop is as high as 100 kPa with a mass flux of 1637 kg/m2s at a heat flux of 400 W/cm2 in a two-phase regime. The heat transfer coefficient and the vapor quality are calculated and reported. The impact of microfluidic cooling on the electrical performance of the 3D interconnects is also analyzed. The high aspect ratio through silicon vias (TSVs) used in the electrical analysis have a nominal diameter of 10 μm.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Flow-Boiling Experimental Investigations

2015;():V003T10A009. doi:10.1115/IPACK2015-48288.

The main aim of the current paper is to demonstrate the capability of a two-phase closed thermosyphon loop system to cool down a contemporary datacenter rack, passively cooling the entire rack including its numerous servers. The effects on the performance of the entire cooling loop with respect to the server orientation, micro-evaporator design, riser and downcomer diameters, working fluid, and approach temperature difference at the condenser have been modeled and simulated. The influence of the thermosyphon height (here from 5 to 20 cm with a horizontally or vertically oriented server) on the driving force that guarantees the system operation whilst simultaneously fulfilling the critical heat flux (CHF) criterion also has been examined. In summary, the thermosyphon height was found to be the most significant design parameter. For the conditions simulated, in terms of CHF, the 10 cm-high thermosyphon was the most advantageous system design with a minimum safety factor of 1.6 relative to the imposed heat flux of 80 W cm−2. Additionally, a case study including an overhead water-cooled heat exchanger to extract heat from the thermosyphon loop has been developed and then the entire rack cooling system evaluated in terms of cost savings, payback period, and net benefit per year. This approximate study provides a general understanding of how the datacenter cooling infrastructure directly impacts the operating budget as well as influencing the thermal/hydraulic operation, performance, and reliability of the datacenter. Finally, the study shows that the passive two-phase closed loop thermosyphon cooling system is a potentially economically sound technology to cool high heat flux servers of datacenters.

Commentary by Dr. Valentin Fuster
2015;():V003T10A010. doi:10.1115/IPACK2015-48734.

As the proper cooling of the electronic devices leads to significant increase in the performance, two-phase heat transfer to dielectric liquids can be of an interest especially for thermal management solutions for high power density devices with extremely high heat fluxes. In this paper, the pressure drop and critical heat flux (CHF) for subcooled flow boiling of methanol at high heat fluxes exceeding 1 kW/cm2 is investigated. Methanol was propelled into microtubes (ID = 265 and 150 μm) at flow rates up to 40 ml/min (mass fluxes approaching 10000 kg/m2-s), boiled in a portion of the microtube by passing DC current through the walls, and the two-phase pressure drop and CHF were measured for a range of operating parameters. The two-phase pressure drop for subcooled flow boiling was found to be significantly lower than the saturated flow boiling case, which can lead to lower pumping powers and more stability in the cooling systems. CHF was found to be increasing almost linearly with Re and inverse of inner diameter (1/ID), while for a given inner diameter, it decreases with increasing heated length.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: High Heat Flux Hot Spot Cooling

2015;():V003T10A011. doi:10.1115/IPACK2015-48242.

The ability of various arrays of micro pin-fins to reduce maximum temperature of an integrated circuit with a 4 × 3 mm footprint and a 0.5 × 0.5 mm hot spot was investigated numerically. Micro pin-fins having circular, symmetric airfoil and symmetric convex lens cross sections were optimized to handle a background uniform heat flux of 500 W cm−2 and a hot spot uniform heat flux of 2000 W cm−2. A fully three-dimensional conjugate heat transfer analysis was performed and a multi-objective, constrained optimization was carried out to find a design for each pin-fin shape capable of cooling such high heat fluxes. The two simultaneous objectives were to minimize maximum temperature and minimize pumping power, while keeping the maximum temperature below 85 °C. The design variables were the inlet average velocity and shape, size and height of the pin-fins. A response surface was generated for each of the objectives and coupled with a genetic algorithm to arrive at a Pareto frontier of the best trade-off solutions. Stress–deformation analysis incorporating hydrodynamic and thermal loads was performed on the three Pareto optimized configurations. Von-Mises stress for each configuration was found to be significantly below the yield strength of silicon.

Commentary by Dr. Valentin Fuster
2015;():V003T10A012. doi:10.1115/IPACK2015-48496.

This work presents the experimental design and testing of a two-phase, embedded manifold-microchannel cooler for cooling of high flux electronics. The ultimate goal of this work is to achieve 0.025 cm2-K/W thermal resistance at 1 kW/cm2 heat flux and evaporator exit vapor qualities at or exceeding 90% at less than 10% absolute pressure drop. While the ultimate goal is to obtain a working two-phase embedded cooler, the system was first tested in single-phase mode to validate system performance via comparison of experimentally measured heat transfer coefficient and pressure drop to the values predicted by CFD simulations. Upon validation, the system was tested in two phase mode using R245fa at 30°C saturation temperature and achieved in excess of 1 kW/cm2 heat flux at 45% vapor quality. Future work will focus on increasing the exit vapor quality as well as use of SiC for the heat transfer surface upon completion of current experiments with Si.

Topics: Cooling , Fluids , Electronics
Commentary by Dr. Valentin Fuster
2015;():V003T10A013. doi:10.1115/IPACK2015-48632.

In this study, a new two-phase heat sink architecture is introduced that operates in two different phase change modes. At low wall superheat temperatures, the heat sink operates at the thin film evaporator mode and transitions to boiling when the wall superheat temperature is increased. This unique function is enabled through constraining the liquid and vapor phases into separate domains using capillary-controlled meniscus formed within a hierarchical 3D structure. The structure is designed to form thin layers of vertically oriented liquid films that directly evaporate into their neighboring vapor space. The dominant mode of heat transfer in this design is thin film evaporation, a very effective boiling sub-process. As the surface superheat temperature is increased and boiling starts, the capillary-controlled meniscus breaks down. A heat transfer coefficient of greater than 200 kW/m2K is achieved at less than 1 °C wall superheat temperature.

Commentary by Dr. Valentin Fuster
2015;():V003T10A014. doi:10.1115/IPACK2015-48689.

Embedded cooling techniques, utilizing microfluidic channels directly within a die substrate or in a miniature heat sink attached to the base of the die, have been known for decades [1] [2] [3]. Despite their demonstrated thermal benefits, such techniques and devices have not been successfully transitioned from the laboratory to either the commercial or military arena. Some of the hurdles preventing implementation to date have been the inability to miniaturize the supporting hardware, the high unit cost of fabricating individual microfluidic coolers, and the unknown reliability of the new technologies introduced. Recent advances in micro manufacturing and co-design/simulation capabilities have enabled significant progress to be made, aided by a significant new focus on embedded cooling technologies was recently initiated by DARPA [4]. This paper will present a recently fabricated embedded cooling system consisting of a state of the art, micro-miniature, 3D microfluidic manifold suitable for near term integration into existing systems. Computational fluid dynamics (CFD) and conjugate heat transfer (CHT) simulations demonstrate the ground breaking thermal performance achieved by the device, and coupled-field flow, thermal, structural and erosion simulations are also presented to address some of the reliability concerns. Finally, measured thermal performance is presented, validating the predicted thermal performance.

Commentary by Dr. Valentin Fuster
2015;():V003T10A015. doi:10.1115/IPACK2015-48757.

Power electronic modules are exhibiting ever increasing power density as a result of compound semiconductor devices being placed in packages of decreasing size. This has led, in turn, to higher volumetric heat generation, which is driving the development of advanced thermal management approaches, including integration of single and two-phase microchannel coolers into the power electronics package. Reliable integration and operation of these coolers is essential for maintaining the performance and reliability of the power electronic system as a whole. This paper will present models for the critical failure mechanisms in microchannel coolers, including erosion/corrosion and cooler fracture.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Liquid and Synthetic Jets

2015;():V003T10A016. doi:10.1115/IPACK2015-48094.

Impinging synthetic jets have been considered as a possible solution for cooling miniature structures. It has been shown that synthetic jet performance is sensitive to the distance between the jet nozzle and the target surface where enhancement of heat transfer decreases with a reduction in nozzle-to-target plate distance. At low nozzle-to-target spacing, no detailed information about the momentum and temperature fields have been shown in prior literature, therefore further investigation is needed. In this study, a 3-D computational fluid dynamics model was constructed to determine the flow and temperature fields of a meso-scale synthetic jet at a nozzle-to-target surface spacing of H/D = 2, ReD,j= 1400 and f = 500 Hz. Unlike the majority of previous computational studies, rather than specifying the boundary conditions at the nozzle, the flow inside the synthetic jet device was solved by specifying the time dependent boundary conditions on the vibrating diaphragm and utilizing the moving mesh technique. Local surface pressure and heat transfer coefficient distributions were determined and discussed. It was found that the pulsating flow at the nozzle exit for a round jet generates vortex rings and these rings seem to have some considerable effects on the target surface profiles.

Commentary by Dr. Valentin Fuster
2015;():V003T10A017. doi:10.1115/IPACK2015-48393.

The local surface temperature, heat flux, heat transfer coefficient, and Nusselt number were measured for an inline array of circular, normal jets of single-phase, liquid water impinging on a copper block with a common outlet for spent flow, and an experimental 2-D surface map was obtained by translating the jet array relative to the sensors. The effects of variation in jet height, jet pitch, confining wall angle, and average jet Reynolds number were investigated. A strong interaction between the effects of the geometric parameters was observed, and a 5° confining wall was seen to be an effective method of managing the spent flow for jet impingement cooling of power electronics. The maximum average heat transfer coefficient of 13,100W=m2K and average Nusselt number of 67.7 were measured at an average jet Reynolds number of 14,000.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Single Phase Cooling Fundamentals

2015;():V003T10A018. doi:10.1115/IPACK2015-48123.

This study considers the optimization of a complex micro-scale cooling geometry that represents a unit-cell of a full heat sink microstructure. The configuration consists of a channel with a rectangular cross section and a hydraulic diameter of 100 μm, where the fluid flows between two cooling fins connected by rectangular crossbars (50 × 50 μm). A previous investigation showed that adding these crossbars at certain locations in the flow can increase the heat transfer in the microchannel, and in the present work we perform an optimization to determine the optimal location and number of crossbars. The optimization problem is defined using 12 discrete design parameters, which represent 12 crossbars at different locations in the channel that can either be turned off and become part of the fluid domain, or turned on and become part of the solid domain. The optimization was done using conjugate heat transfer computational fluid dynamics (CFD) simulations using Fluent 15.0. All possible 4096 configurations were simulated for one set of boundary conditions. The domain was discretized using about 1 million nodes combined for the fluid and solid domains and the computational time was around 1 CPU hour per case. The results show that further improvements in heat transfer are feasible at an optimized pressure drop. The results cover a range of pressure drops from 25 kPa to almost 90 kPa and the heat transfer coefficient varies from 60 to 120 kW/m2K. The configurations on the Pareto front show the trend that crossbars closer to the maximal fluid-solid interface result in a more optimal performance than crossbars positioned farther away. In addition to performing simulations for all possible configurations, the potential of using a genetic algorithm to identify the configurations that define the Pareto front was explored, demonstrating that a 80% reduction in computational time can be achieved. The results of this study demonstrate the significant increase in performance that can be obtained through the use of computational tools and optimization algorithms for the design of single phase cooling devices.

Commentary by Dr. Valentin Fuster
2015;():V003T10A019. doi:10.1115/IPACK2015-48346.

Most methods for designing electronics cooling schemes do not offer the information on what levels of heat fluxes are maximally possible to achieve with the given material, boundary and operating conditions. Here, we offer an answer to this inverse problem posed by the question below. Given a micro pin-fin array cooling with these constraints:

- given maximum allowable temperature of the material,

- given inlet cooling fluid temperature,

- given total pressure loss (pumping power affordable), and

- given overall thickness of the entire electronic component,

find out the maximum possible average heat flux on the hot surface and find the maximum possible heat flux at the hot spot under the condition that the entire amount of the inputted heat is completely removed by the cooling fluid. This problem was solved using multi-objective constrained optimization and metamodeling for an array of micro pin-fins with circular, airfoil and symmetric convex cross sections that is removing all the heat inputted via uniform background heat flux and by a hot spot. The goal of this effort was to identify a cooling pin-fin shape and scheme that is able to push the maximum allowable heat flux as high as possible without the maximum temperature exceeding the specified limit for the given material. Conjugate heat transfer analysis was performed on each of the randomly created candidate configurations. Response surfaces based on Radial Basis Functions were coupled with a genetic algorithm to arrive at a Pareto frontier of best trade-off solutions. The Pareto optimized configuration indicates the maximum physically possible heat fluxes for specified material and constraints.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Two Phase Boiling Computational Modeling Challenges

2015;():V003T10A020. doi:10.1115/IPACK2015-48033.

This work presents a new boiling heat transfer prediction method for slug flow within microchannels, which is developed and benchmarked against the results of two-phase CFD simulations. The proposed method adopts a two-zone decomposition of the flow for the sequential passage of a liquid slug and an evaporating elongated bubble. The heat transfer is modeled by assuming transient heat conduction across the liquid film surrounding an elongated bubble and sequential conduction/convection within the liquid slug. Embedded submodels for estimating important flow parameters, e.g. bubble velocity and liquid film thickness, are implemented as “building blocks”, thus making the entire modeling framework totally stand-alone. The CFD simulations are performed by utilizing ANSYS Fluent v. 14.5 and the interface between the vapor and liquid phases is captured by the built-in Volume Of Fluid algorithm. Improved schemes to compute the surface tension force and the phase change due to evaporation are implemented by means of self-developed functions. The comparison with the CFD results shows that the proposed method emulates well the bubble dynamics during evaporation, and predicts accurately the time-averaged heat transfer coefficients during the initial transient regime and the terminal steady-periodic stages of the flow.

Commentary by Dr. Valentin Fuster
2015;():V003T10A021. doi:10.1115/IPACK2015-48066.

A novel fully dynamic model of a microchannel evaporator is presented. The aim of the model is to study the highly dynamic parallel channel instabilities that occur in these evaporators in more detail. The numerical solver for the model is custom-built and the majority of the paper is focused on detailing the various aspects of this solver. The one-dimensional homogeneous two-phase flow conservation equations are solved to simulate the flow. The full three-dimensional conduction domain of the evaporator is also dynamically resolved. This allows for the correct simulation of the complex hydraulic and thermal interactions between the microchannels that give rise to the parallel channel instabilities. The model uses state-of-the-art correlations to calculate the frictional pressure losses and heat transfer in the microchannels. In addition, a model for inlet restrictions is also included to simulate the stabilizing effect of these components. In the final part of the paper, initial validation results of the model are presented, in which stability results of the model are compared to existing experimental data from literature. Finally, some representative dynamic results are also given to demonstrate some of the unique capabilities of the model.

Topics: Microchannels
Commentary by Dr. Valentin Fuster
2015;():V003T10A022. doi:10.1115/IPACK2015-48129.

This paper presents a comparison of Volume-of-Fluid simulation results with experiments [1] for two-phase flow and heat transfer in a micro channel. Mass transfer between the phases is modeled using a reduced-order model, requiring the definition of a time relaxation constant, r. A two-step solution procedure is used, where first a fixed temperature boundary condition is imposed at the heater to avoid overheating of the device during the initial development of the two-phase flow. After obtaining a quasi-steady-state solution this is changed to a heat flux boundary condition to determine the final solution. Results using three different values for r indicate that the value of the constant should vary throughout the domain. A final simulation where r is defined as a function of the streamwise location results in a prediction of the base temperature within 1K of the experimental result, a pressure drop within 30%, and a prediction of the location of transition from subcooled to saturated flow within 2mm.

Commentary by Dr. Valentin Fuster
2015;():V003T10A023. doi:10.1115/IPACK2015-48178.

High-fidelity simulation of flow boiling in microchannels remains a challenging problem, but the increasing interest in applications of microscale two-phase transport highlight its importance. In this paper, a volume of fluid (VOF)-based flow boiling model is proposed with computational expense-saving features that enable cost-effective simulation of two-phase flow and heat transfer in realistic geometries. The vapor and liquid phases are distinguished using a color function which represents the local volume fraction of the tracked phase. Mass conservation is satisfied by solving the transport equations for both phases with a finite-volume approach. In order to predict phase change at the liquid-vapor interface, evaporative heat and mass source terms are calculated using a novel, saturated-interface-volume phase change model that fixes the interface at the saturation temperature at each time step to achieve stability. Numerical oscillation of the evaporation source terms is thus eliminated and a non-iterative time advancement scheme can be adopted to reduce computational cost. The reference frame is set to move with the vapor slug to artificially increase the local velocity magnitude in the thin liquid film region in the relative frame, which reduces the influence of numerical errors resulting from calculation of the surface tension force, and thus suppresses the development of spurious currents. This allows use of non-uniform meshes that can efficiently resolve high-aspect-ratio geometries and flow features and significantly reduces the overall numerical expense. The proposed model is used to simulate the growth of a vapor bubble in a heated 2D axisymmetric microchannel. The bubble motion, bubble growth rate, liquid film thickness, and local heat transfer coefficient along the wall are compared against previous numerical studies.

Commentary by Dr. Valentin Fuster
2015;():V003T10A024. doi:10.1115/IPACK2015-48334.

Modeling and simulation of two-phase phenomena, as well as their impact on electrical performance and physical integrity are critical to the success of embedded cooling strategies. In DARPA’s Intrachip/Interchip Embedded Cooling (ICECool) program, thermal/electrical/mechanical co-simulation and modeling tools are being applied to the analysis and design of RF GaN MMIC (Monolithic Microwave Integrated Circuit) Power Amplifiers (PA) and digital ICs, with the ultimate goal of achieving greater than 3X electronic performance improvement. This paper addresses various simulation strategies and numerical techniques adopted by the DARPA ICECool performers, with attention devoted to co-simulation through coupled iterations of thermal, mechanical and electrical behavior for capturing device characteristics and predicting reliability and “best in class” simulations that can provide an understanding of device behavior during rugged operating conditions impacted by multi-physics environments. The effect of CTE (Coefficient of Thermal Expansion) mismatch on bond and structural integrity, the impact of cooling fluid choice on performance, the factors affecting erosion/corrosion in the microchannels, as well as electro-migration limits and joule heating effects, will also be addressed. A separate discussion of various two-phase issues, including interface tracking, system pressure drops, conjugate heat transfer, estimating near wall heat transfer coefficients, and predicting CHF (Critical Heat Flux) and dryout is also provided.

Commentary by Dr. Valentin Fuster
2015;():V003T10A025. doi:10.1115/IPACK2015-48350.

A pulsating heat pipe (PHP), also known as an oscillating heat pipe (OHP), is a passive thermal transport device which consists of a single meandering microchannel making multiple passes each through an evaporator and condenser. With a sufficient number of such passes, intermittent boiling of liquid slugs within each evaporator pass perturbs flow in adjacent channels leaving the device in a perpetually unstable state of oscillation. A PHP is thus distinguished operationally from a loop thermosyphon by having a motive force other than buoyancy and the ability to operate in all gravitational orientations.

The most successful PHP models to date track liquid slug motion, sensible heating of the slugs, and mass transfer between liquid slugs and vapor plugs due to evaporation and condensation. However, the predictive capabilities of PHP models remain poor and the numbers assigned to evaporation and condensation heat transfer coefficients are generally not well justified by any realistic physical process. The current study applies methods consistent with state of the art prediction methods in microchannel boiling, to obtain results which predict the PHP’s heat transfer performance and the effect of gravitational orientation on performance.

Topics: Heat pipes
Commentary by Dr. Valentin Fuster
2015;():V003T10A026. doi:10.1115/IPACK2015-48441.

A numerical method is described to study two-phase flows for single and multiple bubbles with phase change. The fluid flow equations are based on the Arbitrary Lagrangian-Eulerian formulation (ALE) and the Finite Element Method (FEM), creating a new two-phase method with an improved model for the liquid-gas interface in microchannels. A successful adaptive mesh update procedure is also described for effective management of the mesh at the two-phase interface to remove, add and repair surface elements, since the computational mesh nodes move according to the flow. The Lagrangian description explicitly defines the two-phase interface position by a set of interconnected nodes which ensures a sharp representation of the boundary, including the role of the surface tension. The methodology proposed for computing the curvature leads to accurate results with moderate programming effort and computational cost and it can also be applied to different configurations with an explicit description of the interface. Such a methodology can be employed to study accurately many problems such as oil extraction and refinement in the petroleum area, design of refrigeration systems, modelling of biological systems and efficient cooling of electronics for computational purposes, being the latter the aim of this research. The obtained numerical results will be described, therefore proving the capability of the proposed new methodology.

Commentary by Dr. Valentin Fuster
2015;():V003T10A027. doi:10.1115/IPACK2015-48489.

The 3D (three dimensional) integration of microelectronic chips into chip stacks is an enabling technology to provide a possible path for increasing computational performance. However, 3D chip stacks require a solution to significant new thermal challenges. As a feasible solution, two-phase cooling utilizing a chip-to-chip interconnect-compatible dielectric fluid can be used. This chip-integrated micrometer scale two-phase cooling technology can be essential to fully optimize the benefits of improved integration density and modularity of 3D stacking of high performance integrated circuits (ICs) for future computing systems; but is faced with significant developmental challenges including high fidelity modeling.

In the present work, an Eulerian multiphase model has been developed for simulating two-phase evaporative cooling through chip embedded microscale cavities. First, the model was used to predict the flow and heat transfer characteristics for coolant R245fa flowing through a single straight micro channel with cross section 100 × 100 um and length 10 mm. The flow is sub-cooled in the initial section of the channel and saturated in the remaining. The results were compared to experimental data available from literature, focusing on the model capability to predict the correct flow pattern, temperature profile and pressure drop. Next, the validated model was extended to the simulation of complex flow geometries expected in microprocessor chip-stacks with chip-to-chip interconnects.

Commentary by Dr. Valentin Fuster

Thermal Management Using Micro Channels, Jets, Sprays: Vapor Chambers and Condensation

2015;():V003T10A028. doi:10.1115/IPACK2015-48308.

A capillary-wick heat pipe having the dimensions of 5.0 mm × 5.0 mm × 100 mm (length) is fabricated on a surface of a plastic board, and the experimental investigations are conducted on the operational characteristics of the heat pipe. Plastics are easy to manufacturing, lightweight, low cost, flexible, and besides, the present study aims at the phase-change heat transfer inside the plastic board. A sintered copper powder and water are used as the wick structure and the working fluid of the heat pipe, respectively. In experiments, an evaporator section of the heat pipe is heated by a heater while a condenser section is water-cooled by a heat sink. A heat input and a liquid volume inside the heat pipe are changed, and the temperature distribution of the heat pipe is measured by thermocouples. Moreover, a one-dimensional thermal circuit model is made to evaluate the effective thermal conductivity of the heat pipe. From the experimental results, the continuous phase-change heat transfer inside the plastic board and its effectiveness are confirmed. It is also revealed that the effective thermal conductivity of the heat pipe is 854 W/(m·K) in maximum under the present experimental conditions.

Topics: Heat , Heat pipes
Commentary by Dr. Valentin Fuster
2015;():V003T10A029. doi:10.1115/IPACK2015-48554.

A vapor chamber is a flat-plate heat pipe, where a cooled (condenser) section is much larger than a heated (evaporator) section, and has been used as a heat spreader to enhance the cooling of electronic devices. An objective of this study is to integrate the vapor chamber into a polycarbonate board. Plastic materials are easy to manufacturing, light weight, low cost, flexible, and then the present study aims at performing a phase-change heat transfer and a heat spreading inside the polycarbonate board. A sintered copper powder and water are used as a wick structure and a working fluid, respectively. In experiments, the heat is applied by a heater while the cooling water is circulated between a thermostatic bath and a cooling jacket. The experiments are conducted changing a liquid volume and a heat input, and the transient temperature distribution of the vapor chamber is measured by thermocouples. For comparison, the experiment is also conducted where the working fluid is not charged into the vapor chamber. Moreover, based on a thermal resistance network, an analytical model of the vapor chamber is made and the analysis is performed on the phase-change heat transfer inside the vapor chamber. From the experimental and analytical results, the heat transfer characteristics of the polymer-based vapor chamber and the effectiveness of the phase-change heat transfer are confirmed.

Topics: Vapors , Manufacturing
Commentary by Dr. Valentin Fuster

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