Abstract Preview

Here is the abstract you requested from the Thermal_2011 technical program page. This is the original abstract submitted by the author. Any changes to the technical content of the final manuscript published by IMAPS or the presentation that is given during the event is done by the author, not IMAPS.

Mitigating Transient Hotspots Using Dynamic Thermal Management and Two-Phase Microfluidic Heat Exchangers
Keywords: Dynamic Thermal Management, Two-Phase Microfluidic Cooling, Modeling
To prevent electromigration and mitigate thermal stresses in microprocessors, increasing attention is being paid to managing chip hotspots, localized regions of the chip where heat fluxes are as much as 20 times greater than neighboring regions. Two general approaches can be taken, either independently or in combination, to reduce the peak operating temperature of a hotspot. First, dynamic thermal management (DTM) can be used to reduce the power output in the hotspot region, resulting in reduced hotspot temperatures. Second, advanced cooling techniques can be used to decrease the effective thermal resistance of the heat sink. For this purpose, two-phase microfluidic heat exchangers are an attractive candidate solution capable of providing very high heat flux cooling at relatively low pumping power, as compared to single phase alternatives. In this work, we present two complementary modeling efforts that inform both approaches - DTM and advanced microfluidic cooling - for thermal management of transient hotspots. The first model focuses on quantifying the inherent uncertainty in DTM schemes commonly used to mitigate hotspots in multi-core processors. Much attention has focused on various response methods for throttling power in DTM schemes; however relatively little research has considered the propagation of uncertainty in signals from the chip's embedded temperature sensors that are used to trigger DTM responses. The present work introduces a model to determine the accuracy and resolution at which hotspots can be measured using distributed temperature sensors. The model uses a novel, computationally-efficient, spatial frequency domain inverse heat transfer solution. The uncertainty in the hotspot location and magnitude is computed for numerous randomized chip heat flux profiles over a range of thermal sensor spatial frequencies that represent nominal sensor sizes from 150 um to 2 mm. The results show the minimum uncertainty achievable in the calculated heat flux profile using distributed temperature sensors, and a comparison is made to direct interpretation of the temperature. The simulation tools presented in this study can be used to determine the optimal spacing of distributed temperature sensor arrays for DTM in chips. The second portion of the study focuses on cooling transient hotspots using two-phase microfluidic heat exchangers. In these systems, it remains a challenge to maintain stable two-phase flow without introducing detrimental pressure drop elements. In this study, we develop a model for the multi-scale, highly transient physics involved in flow instabilities, and the resulting effects on the heat exchanger performance. We employ a reduced-order modeling approach that captures transient thermal and fluidic transport using circuit analysis for a non-uniform, time-variant heat source applied to a parallel microchannel, two-phase heat exchanger. The model is compared to previously published models and empirical results for validation. We report criteria for stable two-phase heat exchanger operation considering factors such as pumping pressure, channel geometry, and applied heat flux. The results provide needed insight into the robust design of two-phase microfluidic heat exchangers. By integrating two-phase microfluidic heat exchangers with improved DTM schemes, transient hotspots can be effectively suppressed.
Josef L. Miler, Ph.D. Candidate
Stanford University
Stanford, CA

  • Amkor
  • ASE
  • Canon
  • EMD Performance Materials
  • Honeywell
  • Indium
  • Kester
  • Kyocera America
  • Master Bond
  • Micro Systems Technologies
  • MRSI
  • Palomar
  • Plexus
  • Promex
  • Qualcomm
  • Quik-Pak
  • Raytheon
  • Specialty Coating Systems