Search

Impinging jet heat transfer is a promising method to cool electronic components. In this paper, a numerical study has been carried out to examine the conjugate heat transfer under a confined impinging jet using a plate-fin heat sink as the target plate. Effects of geometric parameters such as fin number, fin height, and fin-to-spacing ratio are examined over a range of jet Reynolds numbers using dielectric fluid FC-72 as the fluid medium. Thermal resistance, pressure drop, and Nusselt number are the main criteria used to evaluate the thermal and fluid dynamic performance of this flow system. Furthermore, the effects of fin height, fin-to-spacing ratio, and jet Reynolds number on impinging jet heat transfer are obtained. The concept of an effective Nusselt number is introduced for computing the heat transfer effectiveness of heat sinks with different fin numbers. [DOI: 10.1115/1.2809447] Keywords: thermal resistance, fin height, fin-to-spacing ratio, conjugate heat transfer

An entropy generation minimization method is applied to study the thermodynamic losses caused by heat transfer and pressure drop for the fluid in a cylindrical pin-fin heat sink and bypass flow regions. A general expression for the entropy generation rate is obtained by considering control volumes around the heat sink and bypass regions. The conservation equations for mass and energy with the entropy balance are applied in both regions. Inside the heat sink, analytical/empirical correlations are used for heat transfer coefficients and friction factors, where the reference velocity used in the Reynolds number and the pressure drop is based on the minimum free area available for the fluid flow. In bypass regions theoretical models, based on laws of conservation of mass, momentum, and energy, are used to predict flow velocity and pressure drop. Both in-line and staggered arrangements are studied and their relative performance is compared to the same thermal and hydraulic conditions. A parametric study is also performed to show the effects of bypass on the overall performance of heat sinks. [DOI: 10.1115/1.2965209]

An experimental study was conducted to investigate the heat transfer from a parallel flat plate heat sink under a turbulent impinging air jet. A horizontal nozzle plate confined the target surface. The jet was discharged from a sharp-edged nozzle in the nozzle plate. Average Nusselt numbers are reported for Pr=0.7, 5000 [less than or equal to] Re [less than or equal to] 30,000, L/d=2.5, and 0.833 at H/d=3 where L, H, and d define the length of the square heat source, nozzle-to-target spacing, and nozzle diameter, respectively. Tests were also conducted for an impinging flow over a flat plate, flush with the top surface of the target plate. The average Nusselt numbers from the heat sink were compared to those for a flat plate to determine the overall performance of the heat sink in a confined impingement arrangement. The experimental results were compared with the numerical predictions obtained in an earlier study. Although the average Nusselt numbers obtained from numerical simulations differed from the experimental measurements by 18%, the disagreement is much less significant when related to the junction temperature. Under typical conditions, it was shown that such discrepancy in the Nusselt number lead to an error of 6% in the prediction of the junction temperature of the device.

The design and thermal performance of a synthetic-air-jet-based heat sink for high-power dissipation electronics is discussed. Each fin of a plate-fin heat sink is straddled by a pair of two-dimensional synthetic jets, thereby creating a jet ejector system that entrains cool ambient air upstream of the heat sink and discharges it into the channels between the fins. The jets are created by periodic pressure variations induced in a plenum by electromagnetic actuators. The performance of the heat sink is assessed using a thermal test die encased in a heat spreader that is instrumented with a thermocouple. The case-to-ambient thermal resistance under natural convection with the heat sink is 3.15 [degrees]C/W. Forced convection with the synthetic jets enables a power dissipation of 59.2 W at a case temperature of 70[degrees]C, resulting in a case-to-ambient thermal resistance of 0.76[degrees]C/W. The synthetic-jet heat sink dissipates ~40% more heat compared to steady flow from a ducted fan blowing air through the heat sink. The synthetic jets generate a flow rate of 4.48 CFM through the heat sink, resulting in 27.8 W/CFM and thermal effectiveness of 0.62. The effect of fin length on the thermal resistance of the heat sink is discussed. Detailed measurements on an instrumented heat sink estimate that the average heat transfer coefficients in the channel flow between the fins is 2.5 times that of a steady flow in the ducts at the same Reynolds Number. [DOI: 10.1115/1.1869509]

Knowledge of flow pattern and flow pattern transitions is essential to the development of reliable predictive tools for pressure drop and heat transfer in two-phase micro-channel heat sinks. In the present study, experiments were conducted with adiabatic nitrogen-water two-phase flow in a rectangular micro-channel having a 0.406 x 2.032 [mm.sup.2] cross-section. Superficial velocities of nitrogen and water ranged from 0.08 to 81.92 m/s and 0.04 to 10.24 m/s, respectively. Flow patterns were first identified using high-speed video imaging, and still photos were then taken for representative patterns. Results reveal the dominant flow patterns are slug and annular, with bubbly flow occurring only occasionally; stratified and churn flow were never observed. A flow pattern map was constructed and compared with previous maps and predictions of flow pattern transition models. Features unique to two-phase micro-channel flow were identified and employed to validate key assumptions of an annular flow boiling model that was previously developed to predict pressure drop and heat transfer in two-phase micro-channel heat sinks. This earlier model was modified based on new findings from the adiabatic two-phase flow study. The modified model shows good agreement with experimental data for water-cooled heat sinks.

Heat sink designers have to balance a number of conflicting parameters to maximize the performance of heat sinks. This multi-parameter problem lends itself naturally to mathematical optimization techniques. The paper illustrates how mathematical optimization techniques combined with a semi-empirical thermal simulation program can be used to construct a trade-off curve (Pareto-optimal set) between the heat sink mass and thermal resistance for a given heat sink configuration. This trade-off curve can be used by the engineer to decide on the optimal heat sink design that is the best compromise between heat sink mass and thermal resistance for the specific application under consideration.

Dedicated fan-duct-heatsink combinations have become a standard means of cooling computer processors. Most previous studies have considered optimization of fin geometry (pitch and thickness) with overall heatsink dimensions (width, height, length)fixed. The present study considers size requirements for the constraints of fixed air volume flow rate and pressure drop, fixed fan/blower power, and fixed thermal conductance. First, an ideal heatsink with infinite fin thermal conductivity is considered, providing simple power-law prediction of performance. Next, fins of finite thermal conductivity and thickness, as well as effects of developing flow are included in the analysis, permitting prediction and minimization of weight. Models of both levels of complexity can be used, previous to more detailed numerical and/or experimental studies, to design optimized heatsinks. [DOI: 10.1115/1.1569262]

The effort described herein extends the use of least-material single rectangular plate-fin analysis to multiple fin arrays, using a composite Nusselt number correlation. The optimally spaced least-material array was also found to be the globally best thermal design. Comparisons of the thermal capability of these optimum arrays, on the basis of total heat dissipation, heat dissipation per unit mass, and space claim specific heat dissipation, are provided for several potential heat sink materials. The impact of manufacturability constraints on the design and performance of these heat sinks is briefly discussed. [DOI: 10.1115/1.1568361] Keywords: Heat Sink, Rectangular Plate-Fin, Natural Convection, Least-Material, Optimization

3D finite element modeling of thermal emulator cube and its composition consisting of composite stack of multi-layer chip are developed. Thermal analysis of the Multi-Chip Module consisting of 16 alternate layers Si processor and heat sink layers with Si spacers and AlN ceramic cap is undertaken. The various alternatives for design of the emulator cubes such as thermal cube floating in free-space, thermal cube-on-substrate, thermal cube-on-flex cable with a continuous joint of solder and thermal cube embedded in rectangular Si-spacer are investigated for their heat extraction capability. Thermal modeling of a composite structural unit stack of chip offers first hand information as to the operating performance of the entire thermal emulator cube to be used in the construction of buffer cube. The scientific understanding of the mode of heat transfer of the emulator cube, heat extraction of the various heat sink materials, ceramic and the metallic substrates are investigated. A thin sheet of ceramic (AlN) substrate is at least three times more effective in extraction of heat than thick block of steel under similar conditions. The homogeneous and heterogeneous nature of the composite structure of thermal emulator in heat transfer is analyzed. The primary and secondary hotspots in thermal cubes with AlN heat sink are found in thermal simulations. The mode of heat transfer advances normal to lateral and transverse directions of stacking from the central core of the cube towards the outward face. The sharp corners of the cube typically exhibit edge convection due to chilling effect. Buffer-on-flex cable is modeled with a continuous solder joint and its further improvement with alternate hinges of solder joints and micro-channels is proposed for enhanced heat transfer analysis. The embedded emulator cubes are developed for thermal analysis of the optical layers on top of buffer. The optical layers with an interconnection of solder joints on top of the embedded emulator cubes coupled with microchannels and hinged solder joints will be used for further enhanced heat transfer and higher dissipation of heat (1 ~ 3 W/layer) resulting in efficient and cost effective thermal management technique. [DOI: 10.1115/1.1481894]

An extensive numerical analysis of the temperature distribution and fluid flow in a heat sink currently being used for cooling desktop computers was conducted, and demonstrated that if the base of a heat sink was fabricated as a heat pipe instead of a solid material, the heat transfer performance could be significantly increased. It was shown that as the heat sink length increases, the effect of the thermal conductivity of the base on the heat transfer performance increases to be a predictable limit. As the thermal conductivity is increased, the heat transfer performance of heat sinks is enhanced, but cannot exceed this limit. When the thermal conductivity increases to 2,370 W/m-K, the heat transfer performance of the heat sinks will be very close to the heat transfer performance obtained assuming a base with infinite thermal conductivity. Further increases in the thermal conductivity would not significantly improve the heat transfer performance of the heat sinks. [DOI: 10.1115/1.1478058]

In this paper, we present recent experimental results on forced convective heat transfer in novel finned metal foam heat sinks. Experiments were conducted on aluminum foams of 90 percent porosity and pore size corresponding to 5 PPI (200 PPM) and 20 PPI (800 PPM) with one, two, four and six fins, where PPI (PPM) stands for pores per inch (pores per meted and is a measure of the pore density of the porous medium. All of these heat sinks were fabricated in-house. The forced convection results show that heat transfer is significantly enhanced when fins are incorporated in metal foam. The heat transfer coefficient increases with increase in the number of fins until adding more fins retards heat transfer due to interference of thermal boundary layers. For the 20 PPI samples, this maximum was reached for four fins. For the 5 PPI heat sinks, the trends were found to be similar to those for the 20 PPI heat sinks. However, due to larger pore sizes, the pressure drop encountered is much lower at a particular air velocity. As a result, for a given pressure drop, the heat transfer coefficient is higher compared to the 20 PPI heat sink. For example, at a [DELTA] p of 105 Pa, the heat transfer coefficients were found to be 1169 W/[m.sup.2]-K and 995 W/[m.sup.2]-K for the 5 PPI and 20 PPI 4-finned heat sinks, respectively. The finned metal foam heat sinks outperform the longitudinal finned and normal metal foam heat sinks by a factor between 1.5 and 2, respectively. Finally, an analytical expression is formulated based on flow through an open channel and incorporating the effects of thermal dispersion and interfacial heat transfer between the solid and fluid phases of the porous medium. The agreement of the proposed relation with the experimental results is promising. [DOI: 10.1115/1.1464877]

Dynamic phasors models capture transients in (main) harmonic coefficients of periodically dominated systems, and their utility in state approximations is supported by machine and power systems case studies. This article explores analytical plausibility arguments, and inherent restrictions of such approximations in dissipative systems with quadratically nonlinear lossless components. Index Terms--Dissipative system, dynamic phasors.

Heat transfer by conduction, natural convection, and radiation together in discretely heated side-vented compact enclosures has been studied with focus on the enhancement of overall transfer through the opening. The roles of the various modes were of interest as well. Natural convection was found to be weakened by radiation. Overall heat transfer is enhanced when radiation is put into the computations. Radiation plays a significant role in the overall heat transfer. Radiative transport is more pronounced for low Ra.

A study of combined radiation, conduction, and natural convection from discrete heat sources in a horizontal narrow-aspect-ratio enclosed space has been carried out in three dimensions. Excellent agreement between temperature predictions on the heat sources and substrate and experimental measurements was found for modified Rayleigh numbers in the 9.7 x (10 to the fifth) - 1.6 x (10 to the seventh) range.

Forced convection two-phase flow and its complete condensation in a miniature tube are discussed. A physical and mathematical model of annular film condensation in such a tube has been developed. The liquid flow in the model is coupled with the vapor flow along the liquid-vapor interface through the interfacial temperature, shear stress, heat flux, and pressure sump conditions because of surface tension effects. The model predicts the shape of the liquid-vapor interface along the condenser and the length of the two-phase flow region. Findings show complete condensation of the incoming vapor is possible at comparatively low heat loads.

An impingement air flow heat sink performance study which considers the effects of geometry, nozzle-to-sink vertical placement, material and Reynolds number has been performed. The experiments consisted of varying the dimensionless nozzle-to-sink vertical distance from 2 to 12 and the Reynolds number from 1 x [10.sup.4] to 5 x [10.sup.4]. A new carbon composite material manufactured by AlliedSignal was tested and shown to have superior heat transfer characteristics when compared to aluminum. The study also showed that optimal performance is achieved when the nozzle is placed within the dimensionless vertical distance of eight to twelve from the heat sink, and when the Reynolds number lies between 4 x [10.sup.4] and 5 x [10.sup.4]. Both experimental and analytical estimates of the spreading resistance within the heat sinks tested were obtained.

Use of chimneys to enhance heat transfer from free-convection heat sinks is discussed. Experimental results for thermal systems made up of a vertical, parallel-plate, isothermal heat sink and a chimney are given. A ridge of maximum total heat transfer is seen relative to the plate spacing and heat-sink height. Cold inflow at the chimney exit decreases overall heat transfer by about 4%.

A study is presented for laminar natural convection heat transfer from isothermal, vertical disks with horizontal support cylinders, as found in annular-fin heat sinks. A distinction is made between the external and internal surfaces of the heat sink, and, hence, the heat transfer mechanisms which dominate in each of these surfaces can be more easily applied. A boundary layer solution for laminar natural convection, a fully developed flow solution, and a diffusive limit are successfully applied to this complex geometry to form the present model. The model shows good agreement with present experimental data and previous correlations.

An experimental investigation aims to obtain the critical heat flux that can be used uniformly to the two wide evaporator walls for a range of heated lengths, total mass flow rates and outlet saturation temperature. Four copper-water flat miniature heat sinks are considered with relevant geometrical parameters. Results illustrate very similar trends for critical heat flux for equivalent test parameters for all four flat miniature heat sinks.

A re-investigation of the fundamental heat transfer problem of connecting a concentrated heat sink to a finite-size heat-generating volume was conducted. The objective is to show that there is a design approach wherein the volume is fixed, in addition to the constructal growth method of minimizing the thermal resistance between a volume and one point. Results demonstrate that the optimized shapes produced analytically are robust from a design standpoint.

The heat-sink/chimney concept has been extended to systems with pin-fin heat sinks. The results have been applied to problems where the size of the overall system is constrained. Optimal values of the pin-fin diameter and heat-sink porosity were seen, and the optimization was used to show that minimum thermal resistance from a pin-fin heat sink is twice as large as that of an inviscid flow based model.

A study was conducted on previous solutions for the influence of single layers on the thermal spreading resistance of heat flux tubes, compound disks and other circular heat sources undergoing different heat flux distributions. A compound disk solution was found to effectively compute the spreading resistance for all heat sources. Proposed approximations offer upper and lower bounds on the total thermal resistance for a spreader-heat sink problem that can be used on diamond spreader-copper heat sink devices.

Research was conducted to examine two-phase heat dissipation in high-conductivity porous channel heat sinks for microelectronics. Subcooled water and sintered copper particles were utilized to fabricate porous channels while numerical computations were carried out by using measured permeability and test conditions. Results indicated the direct correlation between fluid pressure drop and saturation temperature. They also showed that porous channels were characterized by smaller diameter spheres with higher heat transfer coefficients.

A thermal evaluation of the passive heat sink for a PowerPC 620 microprocessor in a multiprocessor computer shows that the microprocessor junction temperature is 85 degrees celsius for a 25 W heat dissipation. The temperature is maintained at 101 degrees celsius for a 30 W heat dissipation. Fluid dynamic computer simulations model the airflow at the system- and board-level. The results obtained using a compact heat exchanger are discussed.

Thermal performance of a pin fin heat sink of a height up to 15 mm, base size of 25 X 25 mm and 6 X 6 pin fin configuration is optimal in low velocity, open flow and unshrouded type conditions. The fin cross section is 1.5 X 1.5 mm. The performance of the sink is independent of flow direction. A study on different heat sinks shows that the effect of anodizing on the thermal performance is negligible and chromated sinks are suitable for most applications.

The problem of estimating the geometrical parameters of a heat sink for maximum heat dissipation of power semiconductor devices (PSD) is considered. To this end, an experiment to investigate heat exchange 'in cooling channels under forced movement of dielectric coolant.' The experiment involves 'liquid boiling on finned surfaces' for removing high-intensity heat fluxes from cooled PSD components.

A particulate heat sink paste, consisting of submicron diamond dispersed in n-decane, provides a short thermal path between the working face of a semiconductor device and its casing. Its thermal conductance increases with increasing temperature without mass flow, contrary to the linear decrease in thermal conductance of its constituents, namely the diamond and n-decane. An increase in packing fraction results in a greater gas-like behaviour and a consequent increase in thermal conductance. Micropore geometry also affects the heat transport in a direction normal to the particle surface.

Gold-tin solders in 80%-20% ratio are used to mount commercial 1 watt laser diodes soldered on chemical vapor deposition diamond heat sinks. An additional metallization of 2 micrometers gold reduces the mechanical stress in the laser diode and forms reproducible bonds. An analytical model calculates the maximal shear, tensile, and peel stress. High resolution X-ray microscopic results show the absence of voids in the bond interface. The results of the burn-in and the first-accelerated aging tests are presented.

Vapor-deposited micro heat pipe arrays (VDMHP) were fabricated as an integral part of semiconductor devices to act as efficient heat spreaders by reducing the thermal path between the heat sources and heat sink. Fabrication of the VDMHP was accomplished by first establishing a series of grooves in a silicon wafer. Orientation dependent etching (ODE) using a KOH-1-propanol-[H.sub.2]O solution on a (100) wafer with a (111) fiat covered with an oxide mask, resulted in grooves 25 [[micro]meter] wide and 25 [[micro]meter] deep with sharp, perpendicular edges. The wafers were predeposited with a layer of chromium followed by a layer of gold to improve the adhesion characteristics. Dual electron beam vapor deposition, followed by planetary process using molybdenum crucibles, were used to deposit copper 31.5-33.0 [[micro]meter] thick, and provide complete closure of the grooves. A glass cover slip was bonded on the top of the deposited layer. The grooves were finally charged and sealed. A computer model Simulation and Modeling of Evaporated Deposition Profiles (SAMPLE) was used to optimize the metal step coverage and successfully predict the cross-sectional profile of the VDMHP.

A two-phase thermosyphon system was developed for cooling power devices having high heat fluxes, such as thyristas. The system utilizes an inert dielectric fluorocarbon as the heat-transfer medium, effecting heat transfer via boiling and condensation. The thermosyphon system demonstrated a maximum cooling capacity of 3 kW per power device.

A comparison of the cooling effect of a microconfigured heat sink of differing channel widths, based on pool boiling, indicates that decreasing relative channel widths result in less cooling power, until a point is reached where the thermosyphon effect begins to manifest itself. The thermosyphon effect helps to draw cooling fluid through the heat sink thereby improving heat transfer performance.

The cooling behavior of a heat sink with plate fins for a multichip module was determined through experiment and numerical analysis, and nodal network analysis. The cooling behavior, calculated using nodal network analysis, was found to be in good agreement with the results obtained experimentally.

Cooling systems for electronic equipment can be designed using computational fluid dynamics (CFD) with as few as eight cells in the heat sink. A comparison was made between the coarse and detailed CFD modeling of a finned heat sink, both with identical dimensions and boundary conditions but differing in the number of cells and in the representation of the heat sink. Performance was compared of the two CFD models with experimental results indicating good agreement between them.

The conjugated heat transfer problem caused by a thin flush heat source mounted on a conductive substrate is solved by a new formulation. The formulation involves a geometry which consists of a paradigm for direct air cooling of components on the conducting boards. The thermal field in the air cooling board are estimated by the application of printed in circuit board algorithms.

A new methodology and analysis is proposed for the development of the heatsink adhesive attachment strategy into the manufacturing process of semiconductor devices. Semiconductor devices exhibit a high risk of cracking during the manufacturing process. This high risk is caused by the stress produced by the coefficient of Thermal Expansion (CTE) mismatch occurring in the semiconductor packages. The new methodology involves the use of Low Modulus Adhesive (LMA).

The thermal characterization of the Tape Carrier Package (TCP) involves the consideration of the unenhanced performance of the TCPs. The performance of a 20 millimeter TCP on 8 layer boards is found to be 19 degree C/W. The package thermal resistance is 2 degree C/W. The addition of thermal bias with the use of low profile heat sinks enhances the thermal performance in accordance with the requirements for mobile computing systems.

A junction-to-liquid resistance can be achieved by a low profile 3-d multichip, module is less than 2.0 C/W for liquid flow rates which are less than 0.10 gpm. The use of miniature liquid-cooled heat sinks provide a junction-to-liquid resistances of 0.6 c/w and lesser values. The heat capacity rate of the liquid in both cases is increased significantly by the MicroPCM used in the process. This facilitates the production of several multi-chip modules.

Single-drift GaAs TUNNETT diodes were mounted on diamond heat sinks for improved thermal resistance and evaluated around 100 GHz in a radial line full height waveguide cavity. The diodes were fabricated from MBE-grown material originally designed for diodes that operate in CW mode around 100 GHz on integral heat sinks. An RF output power of more than 70 mW with a corresponding dc to RF conversion efficiency of 4.9% was obtained at 105.4 GHz. This is the first successful demonstration of GaAs TUNNETT diodes mounted on diamond heat sinks. To the authors' knowledge, these dc to RF conversion efficiencies and RF power levels are the highest reported to date from TUNNETT diodes and exceed those of any single discrete device made of group III-V materials (GaAs, InP, etc.) at this frequency. Free-running TUNNETT diode oscillators exhibit clean spectra with an excellent phase noise of less than -94 dBc/Hz, measured at a frequency off-carrier of 500 kHz and an RF output power of 40 mW.

Single-drift GaAs TUNNETT diodes on integral heat sinks with RF power levels up to 40 mW at 102.33 GHz exhibit no sign of saturation in RF output power or efficiency and are thought to he thermally limited. Therefore, diodes from the same MBE grown material were fabricated using a different selective etching technology and mounted on diamond heat sinks for improved heat dissipation. RF output power levels of up to 95 mW with corresponding dc-to-RF conversion efficiencies of 5.85% were measured at 104.16 GHz. To further increase the available RF output power, two TUNNETT diodes at a time were power combined using two radial line full-height waveguide cavities. The combined RF output power was 46 mW at 102.3 GHz from two diodes on integral heat sinks and 142 mW at 103.87 GHz from two diodes on diamond heat sinks with corresponding combining efficiencies of 80% and more. This is the first successful demonstration of power combining with TUNNETT diodes and the power levels and efficiencies from either single or combined diodes are the highest reported to date from any devices made of III-V materials (eg., GaAs and InP). Single diodes in second-harmonic mode yielded RF power levels of 0.57 mW at 219.7 GHz and 0.19 mW at 235.7 GHz.

Heat transfer experiments are reported on the thermal performance of longitudinal fin heat sinks attached to an electronic package which is part of a regular array of packages undergoing forced convection air cooling. The effect of coolant bypass on the performance of the heat sink is assessed and performance correlations for reduced heat transfer due to this effect are developed. These correlations are used to develop design guidelines for optimal performance.

Flat rectangular plate heat sinks are often used to cool large electronic components by the combined effects of natural convection and thermal radiation. There is, however, a paucity of rational design techniques for these devices. Thus, a systematic program to investigate the use of flat, rectangular plate heat sinks with surface coatings to enhance the net radiative exchange with the surroundings has been undertaken. The preliminary results of this program are presented in this work. A two-dimensional numerical model of a single electronic component mounted on a vertically oriented, flat rectangular plate heat sink that is located immediately above an upward-facing, horizontal component board was developed for this investigation. This model, which is solved using a control volume method based on the SIMPLER algorithm, accounts for the fully-coupled natural convection, conduction and radiative heat transfer processes that occur in the two-dimensional heat sink configuration described above. The results of a parametric study performed with the numerical model confirm the necessity of employing a heat sink, since for the ranges of values investigated, from 64 to 88 percent of the energy dissipated in the component is transferred to the surroundings from the heat sink. The parametric study examines the effects of component power, heat sink size (height), the thickness and emissivity of the heat sink, the vertical location of the component on the heat sink, and the temperature of the horizontal component board on the temperature of the component mounted on the heat sink.

Enormous increase in the power density of microelectronic devices and packages is obtained from the miniaturization of integrated circuits. High speed performance of these circuits yields an increase in power density which depends on the temperature distribution in the heat dissipation system. Temperature distribution is determined by deriving integrals based on Dini series and Hankel transform.

Liquid immersion cooling is rapidly becoming the mechanism of choice for the newest generation of supercomputers. Miniaturization at both the chip and module level places a severe constraint on the size of the heat sink employed to dissipate the high heat fluxes generated. A study was conducted to develop a surface that could augment boiling heat transfer from silicon surfaces under these constraints. The surface created consists of reversed pyramidal features etched directly on to the silicon surface. Experiments were conducted in a saturated pool of refrigerant-113 at atmospheric pressure. The inexpensive crystallographic etching techniques used to create the enhanced features are described in the paper. The main characteristics of interest in the present study were the incipient boiling superheat and the magnitude of the temperature overshoot at boiling incipience. Results were obtained for test sections with various cavity densities, and compared with data for the smooth untreated surface. It was found that incipient boiling superheat was reduced from a range of 27.0-53.0 degrees C for the untreated silicon surface, to a range of 2.5-15.0 degrees C for the enhanced surfaces. The overshoot also decreased considerably; from about 12.0-18.0 degrees C for two classes of untreated surfaces, to a range of 1.5-5.3 degrees C for the enhanced surfaces. The values of the incipient boiling superheat, and those of the overshoot decreased with a decrease in cavity mouth size. Two ratios of heat source surface area to the area of the enhanced surface were studied. The overshoot values obtained for these surfaces were compared with those observed for some commonly used enhanced surfaces. An elementary numerical study was conducted to estimate the magnitude of heat spreading.

A silicon chip heat loss measurement model is tested for 'thermal via' and 'heat sink conditions' to gauge its effectiveness for use in the design of a high-speed, high-density module. The boundary conditions for the heat transfer coefficient and the thermal conductivity coefficients in the calculations improve the model's precision. Multi-chip module detailing is different because of compound materials in the module.