wall Superheat Excursions in the Boiling incipience of Dielectric Fluids
TL;DR: In this paper, a brief review of the mechanisms that may be responsible for delayed nucleation and examines the limited literature on incipience superheat excursions is presented. But the authors do not consider the effect of temperature variations on the nucleation of microelectronic components.
Abstract: Many of the candidate fluids for immersion cooling of microelectronic components possess both low surface tension and high gas solubility. As a consequence, ebullient heat transfer with such fluids is accompanied by nucleation anomalies and a frequently observed wall temperature overshoot. The difficulty in preventing this thermal excursion and in predicting its magnitude constrains the development of immersion cooling systems. This paper begins with a brief review of the mechanisms that may be responsible for delayed nucleation and examines the limited literature on incipience superheat excursions.
Citations
More filters
TL;DR: In this paper, a comprehensive review of published articles addressing passive enhancement of pool boiling using surface modification techniques is provided, including macroscale, microscale, and nanoscale surfaces, as well as multiscale (hybrid-scale), and hybrid-wettability techniques.
343 citations
TL;DR: In this paper, a correlation for predicting the wall superheat and wall heat flux at ONB has been developed from the data obtained in this study and that reported in the literature.
Abstract: The partitioning of the heat flux supplied at the wall is one of the key issues that needs to be resolved if one is to model subcooled flow boiling accurately. The first step in studying wall heat flux partitioning is to account for the various heat transfer mechanisms involved and to know the location at which the onset of nucleate boiling (ONB) occurs. Active nucleation site density data is required to account for the energy carried away by the bubbles departing from the wall. Subcooled flow boiling experiments were conducted using a flat plate copper surface and a nine-rod (zircalloy-4) bundle. The location of ONB during the experiments was determined from visual observations as well as from the thermocouple output. From the data obtained it is found that the heat flux and wall superheat required for inception are dependent on flow rate, liquid subcooling, and contact angle. The existing correlations for ONB underpredict the wall superheat at ONB in most cases. A correlation for predicting the wall superheat and wall heat flux at ONB has been developed from the data obtained in this study and that reported in the literature. Experimental data are within630 percent of that predicted from the correlation. Active nucleation site density was determined by manually counting the individual sites in pictures obtained using a CCD camera. Correlations for nucleation site density, which are independent of flow rate and liquid subcooling, but dependent on contact angle have been developed for two ranges of wall superheat—one below 15°C and another above 15°C. @DOI: 10.1115/1.1471522#
300 citations
TL;DR: The chapter summarizes analytical, numerical, and experimental work in literature, in order to facilitate the improvement of existing schemes and provide a basis for the development of new ones on the thermal control of semiconductor devices, modules, and total systems.
Abstract: Publisher Summary Thermal control of electronic components has one principal objective, to maintain relatively constant component temperature equal to or below the manufacturer's maximum specified service temperature, typically between 85 and 100°C. It is noted that even a single component operating 10°C beyond this temperature can reduce the reliability of certain systems by as much as 50%. Therefore, it is important for the new thermal control schemes to be capable of eliminating hot spots within the electronic devices, removing heat from these devices and dissipating this heat to the surrounding environment. Several strategies have developed over the years for controlling and removing the heat generated in multichip modules, which include advanced air-cooling schemes, direct cooling, and miniature thermosyphons or free-falling liquid films. The chapter summarizes analytical, numerical, and experimental work in literature, in order to facilitate the improvement of existing schemes and provide a basis for the development of new ones. The chapter focuses on investigations performed over the past decade and includes information on the thermal control of semiconductor devices, modules, and total systems.
285 citations
TL;DR: In this paper, an experimental study of boiling heat transfer from a simulated microelectronic component immersed in a stagnant pool of dielectric Fluorinert (FC-72) is presented.
Abstract: An experimental study of boiling heat transfer from a simulated microelectronic component immersed in a stagnant pool of the dielectric Fluorinert (FC-72) is presented. Various enhancement surfaces were attached to an electrically heated copper calorimeter bar having a vertically oriented heat transfer surface area of 12.7 {times} 12.7 mm{sup 2}. A number of enhancement schemes aimed at a reduction of the incipience temperature overshoot were tested, employing various arrangement of fins, studs, grooves, and vapor-trapping cavities. Atmospheric pressure testing revealed a variation in the magnitude of boiling curve incipience temperature excursion as a function of both macro- and microcharacterization of the surface geometry and initial conditions (pressure and temperature history) prior to boiling. Increased incipience temperatures accompanied prolonged periods of nonboiling. It is assumed that this is due to vapor embryos within surface cavities collapsing to smaller radii. Large artificially created cavities (0.3 mm diameter) were found incapable of maintaining a stable vapor embryo for time periods greater than 10 min. In comparison to flat surfaces, low-profile surface geometries having a structure scale of the order of one bubble departure diameter resulted in significant enhancement of nucleate boiling while drilled surfaces had minimal effectiveness. Surface finish and artificial cavities hadmore » no effect on CHF, but levels of critical heat flux computed on base area were strongly dependent on macrogeometry, due in part to increased surface area.« less
237 citations
TL;DR: In this article, the performance of micro-porous enhanced surfaces using five different sizes of diamond particles was studied using one-dimensional transient thermal conduction (TTC) to calculate the superheated liquid layer thickness.
226 citations
References
More filters
186 citations
TL;DR: In this article, the authors used a 15.8 mm o.d. plain copper tube and three copper enhanced surfaces: a Union Carbide High Flux surface, a Hitachi Thermoexcell-E surface and a Wieland Gewa-T surface.
Abstract: Pool boiling heat-transfer measurements were made using a 15.8 mm o.d. plain copper tube and three copper enhanced surfaces: a Union Carbide High Flux surface, a Hitachi Thermoexcell-E surface and a Wieland Gewa-T surface. The dielectric fluids were Freon-113 and Fluorinert FC-72, a perfluorinated organic compound manufactured to cool electronic equipment. Data were taken at atmospheric pressure, and at heat fluxes from 100 W/m/sup 2/ to 200,000 W/m/sup 2/. Prior to operation, each test surface was subjected to one of three aging procedures to observe the effect of surface past history upon boiling incipience. For Freon-113 the enhanced surfaces showed a two to tenfold increase in the heat-transfer coefficient when compared to a plain tube, whereas for FC-72 an increase of two to five was measured. The High Flux surface gave the best performance over the range of heat fluxes. The Gewa-T surface did not show as much of an enhancement at low fluxes as the other two surfaces, but at high fluxes its performance improved. In fact, it was the only surface tested which delayed the onset of film boiling with FC-72. The degree of superheat required to activate the enhanced surfaces was sensitive to both past history ofmore » the surface and to fluid properties.« less
183 citations
TL;DR: In this article, the limit of homogeneous nucleation in a liquid is shown to lie very close to its liquid spinodal line and it is also argued that the homogeneous prediction should be based on a comparison of the critical work of nucleation with the potential well energy instead of the kinetic molecular energy.
Abstract: The limit of homogeneous nucleation in a liquid is shown to lie very close to its liquid spinodal line. It is also argued that the homogeneous nucleation prediction should be based on a comparison of the critical work of nucleation with the “potential well” energy instead of the kinetic molecular energy. The result is a new prediction of the liquid spinodal line for water that is valid to large negative pressures. This prediction compares well with spinodal points obtained by extrapolating liquid and vapor water data with the Himpan equation.
92 citations
TL;DR: In this article, the heat transfer coefficients were obtained with two heater heights (5 mm, 10 mm) and varying width (2 mm, approximately 70 mm) in water and R-113.
Abstract: Microelectronic circuits were simulated with thin foil heaters supplied with d-c power. The heaters were arranged in two configurations: flush mounted on a circuit board substrate or protruding from the substrate about 1 mm. Heat transfer coefficients (midpoint) were obtained with two heater heights (5 mm, 10 mm) and varying width (2 mm {approximately} 70 mm), in water and R-113. The height effect for single flush heaters agrees qualitatively with convectional theory; however, even the widest heaters agrees qualitatively with convectional theory; however, even the widest heaters have coefficients higher than predicted due to leading edge effects. The heat transfer coefficient increases with decreasing width, with the coefficient for 2 mm being about 150% above that for 20 mm {approximately} 70mm. This is attributed to three-dimensional boundary layer effects. The protruding heaters have a coefficient about 15 percent higher. Data were obtained for in-line and staggered arrays of flush heaters with varying distance between heaters. Coefficients for the upper heaters are below those for lower heaters, with the differences diminishing as the vertical or horizontal spacing increases. For the protruding heaters, the upper heaters have higher coefficients than the lower heaters.
77 citations
TL;DR: In this article, the effect of component size on convective heat transfer from small devices (with surface areas between 2·00 and 0·01 sq cm) was investigated analytically and experimentally.
62 citations