RELIABILITY is defined as the ability of a product to

Transcription

1 658 IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY, VOL. 4, NO. 4, DECEMBER 2004 In Situ Temperature Measurement of a Notebook Computer A Case Study in Health and Usage Monitoring of Electronics Nikhil Vichare, Student Member, IEEE, Peter Rodgers, Valérie Eveloy, and Michael G. Pecht, Fellow, IEEE Abstract Reliability prediction methods do not generally account for the actual life cycle environment of electronic products, which covers their environmental, operating and usage conditions. Considering thermal loads, thermal management strategies still focus on a design for continuous operation that is often determined based on an accumulation of worst-case assumptions. Health monitoring is a method of assessing the reliability of a product in its actual application conditions. A case study in health and usage monitoring of electronic products is presented for a commercial notebook computer. Internal temperatures were dynamically monitored in situ and statistically analyzed during all phases of the life cycle, including usage, storage, and transportation. The effects of power cycles, usage history, CPU computing resources usage, and external thermal environment on peak transient thermal loads were characterized. Such monitored life cycle temperature data could be applied in a life consumption monitoring methodology, to provide damage estimation and remaining life prediction due to specific failure mechanisms influenced by temperature. These findings could contribute to the design of more sustainable, least-energy consumption thermal management solutions. Index Terms Electronics cooling, health monitoring, in situ monitoring, notebook computer, reliability prediction, temperature measurement, thermal characterization. I. INTRODUCTION RELIABILITY is defined as the ability of a product to perform as intended (i.e., without failure and within specified performance limits) for a specified time, in its life cycle application environment. The accuracy of any reliability prediction depends upon both the prediction methodology used, and accurate knowledge of the product, generally including the structural architecture, material properties, fabrication process, and product life cycle conditions [1]. The life cycle conditions consist of the assembly, storage, handling, and use of the product. The life cycle loads include environmental conditions (e.g., temperature, humidity, vibration, shock) and operational parameters (e.g., voltage, current, power dissipation) [2]. The severity and duration of such loads can be significantly influenced by the product usage profile (e.g., utilization duration and frequency, transportation and storage) [3]. Commonly-used electronics reliability prediction methods generally do not accurately account for the life cycle en- Manuscript received September 6, 2004; revised September 17, The authors are with the CALCE Electronic Products and Systems Center, University of Maryland, College Park, MD USA ( pecht@calce.umd.edu). Digital Object Identifier /TDMR vironment of electronic equipment. This arises from either fundamental flaws in the reliability assessment methodologies used [4], or uncertainties in the product life cycle loads [1]. Traditional reliability prediction methods based on the use of handbooks have been shown to be misleading and provide erroneous life predictions [1], [4]. Although the use of appropriate stress and damage models permits a more accurate account of the physics of failure (PoF) [5], their application to long-term reliability prediction based on extrapolated short-term life testing data or field data is typically constrained by insufficient knowledge of the actual application environment of the product [3]. Temperature, in terms of either spatial or temporal gradients, or absolute temperature, is a parameter that can influence the reliability of electronic products [4]. Many integrated circuit (IC) packaging failure mechanisms have been found to have multiple temperature dependencies [4], while semiconductor die circuit electrical performance can be operating temperature dependant [6]. While the need to control electronics operational temperature is well recognized, thermal design strategies still focus on absolute temperature as the dominant thermal stress parameter, and a design for continuous operation [7] that is optimized based on an accumulation of worst-case assumptions [8], [9]. This leads to overly conservative and potentially costly [10] thermal designs, particularly for products characterized by transient peak heat loads, as a result of their usage profile and environmental conditions. Furthermore, conservative thermal designs are at odds with both the demand for miniaturization, and the need to reduce the energy consumption associated with the manufacturing, transportation and operation of thermal management hardware [11], such as central processing unit (CPU) heat spreaders, heat sinks, and fans, which consume scarce battery power in forced air-cooled portable electronics [12]. Among consumer products, notebook computers illustrate well the potential impact of uncertainties in the life cycle environment on thermal management. In the past, passive air-cooling solutions that incorporated heat spreaders [13] or heat pipes [14], forced-air cooled designs using heat sink-fan assemblies for localized CPU cooling [15], and hybrid solutions [16] were employed. Advanced CPU cooling technologies, such as pulsating synthetic micro-jets [17] and laminar air jet impingement [18] have also been investigated. To address sustainability issues, Solbrekken [19] used thermoelectric modules for converting waste heat from a microprocessor into electricity that could be used to operate a cooling fan. Similar energy /04$ IEEE

2 VICHARE et al.: IN SITU TEMPERATURE MEASUREMENT OF A NOTEBOOK COMPUTER 659 recycling approaches are likely to be adopted in the future for operating more advanced thermal management hardware. Indicating that the limits of CPU air-cooling may be reached given current miniaturization and performance constraints, liquid-cooled designs [20] have been recently introduced on the market. Typically, the cooling strategies have been optimized based on anticipated peak thermal loads and temperature usage conditions specified by environmental standards for the product category, as opposed to actual application conditions. Emerging computer architecture-based thermal management techniques [9], [10] involve an autonomous, runtime reduction of the processor chip power dissipation in response to rising temperature. Techniques include parallelism, remapping and rescheduling of computation, and clustered micro-architectures to shift computation from hot to colder units and/or schedule threads according to their expected thermal behavior [9], [10]. Although there is thermal load monitoring for integrated circuit thermal management, no record of the load history is maintained for possible use in subsequent reliability and usage analyses, that could enable significant design improvements. Electronics reliability prediction has been hampered by the lack of methods to accurately predict electronics operational temperature [21], in terms of spatial and temporal gradients, and absolute temperature [22], [23]. Based on a range of experimental benchmarks, Eveloy et al. [23] concluded that computational fluids dynamics (CFD) analysis can be valuable to parametrically assess thermal design strategies in the early design phase, but that accurate temperature boundary conditions to be used for reliability prediction [24] may only be obtained experimentally. This highlights the need to experimentally measure temperature in the actual environmental, usage, and operational conditions of the product. Health monitoring is a method that permits the reliability of a product to be evaluated in its actual application conditions [2], [3]. A product s health is defined as the extent of deviation or degradation from its expected normal operating conditions. Normal operation refers to the physical or performance-related conditions expected from the product. Health monitoring techniques combine sensing, recording, and interpretation of environmental, operational, usage and performance-related parameters indicative of a system s health. While health monitoring methodologies have been routinely employed in mechanical systems, civil structures and aircrafts, their application to electronics reliability assessment has been extremely limited. This arises in part from the small scale of electronic parts, their complex functionality, and electronics reliability assessment practices. This study is part of on-going efforts to develop and implement health monitoring methodologies in next generation electronic equipment. An example of product life cycle load monitoring process is presented. The thermal loads in a notebook computer are dynamically monitored in situ and statistically analyzed. The effects of power cycles, usage history, CPU computing resources usage and external thermal environment on internal peak transient thermal loads are characterized. The application of monitored life cycle data in a health monitoring methodology, to estimate the remaining life of electronics product, is discussed. II. PRODUCT HEALTH MONITORING Health monitoring techniques are used to provide advance warning of failure, prevent catastrophic failure, assess reliability, reduce unscheduled maintenance, identify faults efficiently, and improve both qualification methods and the design of future products [25]. Product health monitoring can be implemented through the use of various techniques to sense and interpret the parameters indicative of: 1) Performance degradation, such as deviation of operating parameters from their expected values; 2) Physical or electrical degradation, such as material cracking, corrosion, interfacial delamination, increase in electrical resistance or threshold voltage; 3) Changes in a life cycle environment, such as usage duration and frequency, ambient temperature and humidity, vibration, and shock. Based on the product s health, determined from the monitored life cycle conditions, maintenance procedures can be developed. Health monitoring therefore permits new products to be concurrently designed for a life cycle environment known through monitoring. Environmental and usage data collection has been undertaken in a variety of industries and services [26]. However, as previously noted, substantially less efforts have been invested to apply such concepts to electronic equipment. Mishra et al. [3] and Ramakrishnan and Pecht [2] monitored the temperature, humidity, vibration and shock loads experienced by an electronic board operated in two different automotive underhood environments. These data were applied in conjunction with physics-of-failure models for damage and remaining life prediction. The health monitoring methodology was shown to effectively predict remaining life, in contrast to traditionally-used electronics reliability prediction methods. A form of health monitoring is currently employed in computing equipment for hard disk drives (HDD), referred to as self-monitoring analysis and reporting technology (SMART) [27]. HDD operating parameters, including flying height of the head, error counts, variations in spin time, temperature, and data transfer rates, are monitored to provide advance warning of failures. This is achieved through an interface between the computer s start-up program (BIOS) and the hard disk drive. Focusing on the impact of computer user behavior on thermal loading, Searls et al. [8] undertook in situ temperature measurements in both notebook and desktop computers used in different parts of the world, and found that notebooks experienced more severe temperature cycling. This study demonstrates a more detailed method of environmental and usage data collection that can enable health monitoring of electronics. While a comprehensive health monitoring plan may involve multiple life cycle conditions, such as humidity, vibration, shock, radiation, and contamination, this study focuses on temperature measurement. A detailed analysis of the thermal loads in a notebook computer, including those experienced during usage, storage, and transportation is undertaken. The data collected are analyzed statistically using data simplification and cycle counting algorithms, and converted

3 660 IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY, VOL. 4, NO. 4, DECEMBER 2004 Fig. 1. Location of RTD temperature sensor on CPU heat sink base. into a format that can be used in physics-of-failure models, for both damage estimation and remaining life prediction due to specific failure mechanisms. Fig. 2. Measured absolute temperature profiles of CPU heat sink, hard disk drive, and external ambient air. III. EXPERIMENTAL CHARACTERIZATION The notebook computer characterized was a primarily-passively-cooled design, that incorporated a heat sink-fan assembly for managing peak transient loads. The processor, a Pentium II (233 MHz), had a worst-case power dissipation of 35 W, and a maximum base plate temperature rated at 75 C. In such designs, the CPU heat sink fan is automatically activated when the processor base plate temperature exceeds its maximum rated temperature. The microprocessor heat sink and HDD housing were found to experience the largest absolute surface temperatures. Their surface temperatures were dynamically monitored using thin film RTD sensors calibrated to an accuracy of C. Fig. 1 shows the microprocessor temperature recorded at the center of the heat sink base. Due to thermal contact resistance, the temperature drop between the heat sink base and base plate, was measured around 7 C for the maximum CPU power dissipation. This was in line with the corresponding vendor specification [28]. The HDD housing temperature was recorded at the center of its external top surface. Measurements were recorded using an external battery powered portable data logger, having an integrated temperature sensor for external ambient air temperature measurement. The data logger had no physical interaction with the notebook computer. All data were recorded at a rate of one sample per minute. The experiments were conducted in College Park, MD, from October to December The recorded data were converted into a sequence of peaks and valleys using the ordered overall range (OOR) method [29]. The screening level was specified at 0% to include all peaks or valleys for cycle counting. The computed sequence of peaks and valleys (time-temperature history) was converted to temperature cycles using software developed in previous work [30], that utilize the 3-parameter Rainflow cycle counting algorithm. The software calculates temperature cycle magnitude, cycle mean temperature, and cycle temperature ramp rate. IV. RESULTS AND DISCUSSION The time-temperature history recorded for the CPU heat sink, HDD, and external ambient air temperature are shown in Fig. 2. Three forms of temperature cycling are observed; that represent Fig. 3. Measured CPU usage and CPU heat sink absolute temperature. Event A: notebook is powered on. Events B to C: numerical simulation is executed. Event D: Notebook is powered off. the notebook on-off cycles, variations in power dissipation associated with different usage intensity of the computing resources, and the external ambient air temperature variations. The effect of computing resource usage on the CPU heat sink temperature is illustrated in Fig. 3. As a computationally intensive numerical simulation is executed, CPU usage increases up to 80%, resulting in a 10 C rise in CPU heat sink temperature. Such data can be valuable for the design of hybrid thermal management solutions, that address peak transient thermal loads. The cooling effect of the fan on the CPU heat sink temperature is shown in Fig. 4, where regions A-B and C-D denote the operating periods of the fan. The recorded temperatures presented in Fig. 2 range from 2 C, to 63 C for the HDD and 71 C for the CPU heat sink. The 2 C corresponded to the minimum external ambient air temperature recorded when the notebook was exposed to outdoor environments. The product usage profile observed in Fig. 2 is characterized by a large number of on-off cycles, relative to typical desktop computer usage [8], as well as temperature cycles of larger magnitude. The latter observation is attributed to exposure to large variations in external ambient temperature, covering both indoor and outdoor environments. Using the data simplification methodology summarized in Section II, the distributions of absolute temperatures, temperature cycle magnitudes, and temperature ramp rates for both

4 VICHARE et al.: IN SITU TEMPERATURE MEASUREMENT OF A NOTEBOOK COMPUTER 661 Fig. 4. Cooling fan operation in time periods A-B and C-D. Fig. 6. Occurrence of CPU heat sink and hard disk drive temperature cycles as a function of cycle magnitude. Fig. 5. Distributions of measured absolute temperature for the CPU heat sink and hard disk drive. the CPU heat sink and HDD were extracted from the measurements in Fig. 2. These analyses are presented in Figs Fig. 5 shows the fraction of total time during which the product experienced a given range of absolute temperature. Fig. 6 represents the number of occurrences of a range of temperature cycle magnitudes. Fig. 7 provides the fraction of total time during which the product experienced a given range of temperature ramp rate. Fig. 5 shows that the CPU heat sink base temperature exceeds its maximum rating, 68 C, over approximately 1% of the monitored time period. This may be due to the fact that the fan was not accurately synchronized with the CPU base plate temperature in the system design. However, the heat sink temperature does not exceed 55 C and 60 C over 90% and 95% of the time. This highlights the potential conservativeness of thermal management solutions optimized based on worst-case operating conditions, 68 C. Fig. 6 indicates that about 97% of the temperature cycles experienced by either the CPU heat sink or HDD have an amplitude of less than 5 C. In addition, the CPU heat sink experiences about 25% more such cycles than the HDD. While it is generally perceived that temperature cycles of small amplitude may not significantly impact on the reliability of electronic packaging interconnections, the potential damage induced also depends on variables such as the mean cyclic temperature, ramp rate, and dwell temperature, as highlighted by [31], [32] Fig. 7. Distribution of measured temperature cycle ramp rates for the CPU heat sink and hard disk drive. R refers to temperature cycle ramp rate. for solder joint fatigue. Pump-out of the thermally-conductive grease at CPU-heat sink interfaces, which lead to increased interface thermal resistance, has been attributed to temperature cycling having an amplitude comparable to those measured in this study, around 5 C [33]. The monitored temperature data therefore suggests that the effect of such multiple mini-cycles on interconnection reliability should be further investigated for the range of measured mean cyclic temperatures and ramp rates. This highlights the importance of recording such data. On the other extreme, temperature cycle magnitudes of up to 50 C were also measured (see Fig. 6). This exceeds the worst-case use condition specified by standard IPC SM-785 [34] for consumer and computer products, namely 30 C and 20 C respectively. Such a discrepancy between standardized and actual conditions provides a strong motivation for monitoring actual product application environments. The measured temperature ramp rate distribution shown in Fig. 7 would also permit more accurate predictions of solder joint fatigue life to be obtained than using worst-case ramp rate specifications [34]. For example, Fig. 7 shows that over 80% of the total time period, the temperature cycles experienced by the CPU heat sink have a ramp rate of less that 4 C/min, while temperature ramp rates exceeding 10 C/min are observed over less than 3% of the total time.

5 662 IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY, VOL. 4, NO. 4, DECEMBER 2004 Monitored life cycle temperature data, such as presented in Figs. 5 7, could be applied in a life consumption monitoring methodology [2], [3], to provide both damage estimation and remaining life prediction due to specific failure mechanisms influenced by temperature. For example, measured heat sink base temperature data could be used to assess the reliability of component-to-heat sink adhesive attachments. Examples include [35] [37] and [38], which evaluated the impact of differential thermal expansion and creep-induced degradation on component-heat sink adhesive joints, respectively. The measured data could also be used to determine the stress levels to be imposed in accelerated testing, refining product specifications, and setting product warranties. While a comprehensive health monitoring process would involve other environmental variables, such as humidity, vibration and shock, apart from temperature, a line of approach similar to that presented in this study could be applied to monitor such life cycle loads. Similarly, the proposed approach could be extended to other electronic products. V. CONCLUSION A method of environmental, operational and usage data collection for enabling health monitoring of electronic products was presented. The thermal loads in a notebook computer were dynamically monitored in situ for typical usage, transportation and storage environments, and statistically analyzed. The effects of power cycles, usage history, CPU computing resources usage and external thermal environment on internal peak transient thermal loads were characterized. The CPU heat sink temperature was found to be 13 C and 8 C lower than its maximum rating over 90% and 95% of the monitored time period, respectively. This highlights the potential conservativeness of thermal management solutions optimized based on worst-case operating conditions that rarely occur. Such findings could contribute to the design of more sustainable, least-energy consumption thermal management solutions. About 97% of the temperature cycles experienced by either the CPU heat sink or HDD had an amplitude of less than 5 C. However, the maximum temperature cycle amplitudes measured were found to exceed those specified by environmental standards for computer and consumer equipment. This provides a strong motivation for monitoring actual product application environments. REFERENCES [1] M. Pecht, D. Das, and A. Ramakrishnan, The IEEE standards on reliability program and reliability prediction methods for electronic equipment, Microelectron. Reliabil., vol. 42, pp , [2] A. Ramakrishnan and M. Pecht, A life consumption monitoring methodology for electronic systems, IEEE Trans. Compon. Packag. Technol., vol. 26, no. 3, pp , Sep [3] S. Mishra, M. Pecht, T. Smith, I. McNee, and R. Harris, Remaining life prediction of electronic products using life consumption monitoring approach, in Proc. Eur. Microelectronics Packaging and Interconnection Symp., Cracow, Poland, Jun. 2002, pp [4] P. Lall, M. Pecht, and E. Harkim, Influence of Temperature on Microelectronics and System Reliability. New York: CRC Press, [5] A. Dasgupta, The physics-of-failure approach at the University of Maryland for the development of reliable electronics, in Proc. 3rd Int. Conf. Thermal and Mechanical Simulation in (Micro)Electronics (EuroSimE), 2002, pp [6] R. K. Kirschman, Low temperature electronics, IEEE Circuits Dev. Mag., pt. 2, vol. 6, pp , [7] J. D. Parry, J. Rantala, and C. J. M. Lasance, Enhanced electronic system reliability challenges for temperature prediction, IEEE Trans. Compon. Packag. Technol., vol. 25, no. 4, pp , Dec [8] D. Searls, T. Dishongh, and P. Dujari, A Strategy for enabling data driven product decisions through a comprehensive understanding of the usage environment, presented at the Pacific RIM/ASME International Intersociety Electronic and Photonic Packaging Conf. (IPACK 01) Advances in Electronic Packaging, Kauai, HI, Jul [9] K. Skadron, A computer-architecture approach to thermal management in computer systems: challenges and opportunities, in Proc. 5th Int. Conf. Thermal and Mechanical Simulation and Experiments in Micro- Electronics and Micro-Systems (EuroSimE), Brussels, Belgium, May 2004, pp [10], The importance of computer architecture in microprocessor thermal design, in Proc. 9th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), Las Vegas, NV, Jun. 2004, pp [11] A. Bar-Cohen and M. Iyengar, Least-energy optimization of air-cooled heat sinks for sustainable development, IEEE Trans. Compon. Packag. Technol., vol. 26, no. 1, pp , Mar [12] W. K. Coxe, G. L. Solbrekken, K. Yazawa, and A. Bar-Cohen, Experimental modeling of the passive cooling limit of notebook computers, in Proc. 8th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 2002, pp [13] T. Kobayashi, T. Ogushi, N. Sumi, and M. Fujii, Thermal design of a ultra-slim notebook computer, in Proc. 6th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 1998, pp [14] H. Xie, A. Ali, and R. Bhatia, The use of heat pipes in personal computers, in Proc. 6th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 1998, pp [15] M. Saini and R. L. Webb, Heat rejection limits of air cooled plane fin heat sinks for computer cooling, in Proc. 8th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 2002, pp [16] T. Nguyen, M. Mochizuki, K. Mashiko, Y. Saito, I. Sauciuc, and R. Boggs, Advanced cooling system using miniature heat pipes in mobile PC, in Proc. 6th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 1998, pp [17] J. S. Campbell, W. Z. Black, A. Glezer, and J. G. Hartley, Thermal management of a notebook computer with synthetic air microjets, in Proc. 6th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 1998, pp [18] J. R. Guarino and V. P. Manno, Characterization of laminar jet impingement cooling in portable computer applications, in Proc. 17th IEEE Semiconductor Thermal Measurement and Management Symp. (SEMI-THERM XVII), 2001, pp [19] G. L. Solbrekken, K. Yazawa, and A. Bar-Cohen, Thermal management of portable electronic equipment using thermoelectric energy conversion, in Proc. 9th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM), 2004, pp [20] O. Suzuki, Design and manufacturing challenges of liquid heat sink for next generation liquid cooling, presented at the JSME Project Meeting, Design and Manufacturing of High Performance Heat Sinks for Microelectronic Equipment,, Tokyo, Japan, Oct [21] C. J. M. Lasance, The European project profit: prediction of temperature gradients influencing the quality of electronic products, in Proc. 17th IEEE Semiconductor Thermal and Temperature Measurement Symp. (SEMI-THERM XVII), 2001, pp [22] P. Rodgers, V. Eveloy, and M. Davies, An experimental assessment of numerical predictive accuracy for electronic component heat transfer in forced convection: Parts I and II, Trans. ASME, J. Electron. Packag., vol. 125, no. 1, pp , [23] V. Eveloy, P. Rodgers, and M. S. J. Hashmi, Numerical prediction of electronic component operational temperature: a perspective, IEEE Trans. Compon. Packag. Technol., vol. 27, no. 2, pp , Jun [24] C. J. M. Lasance, The need for a change in thermal design philosophy, Electron. Cooling, vol. 1, no. 2, pp , 1995.

6 VICHARE et al.: IN SITU TEMPERATURE MEASUREMENT OF A NOTEBOOK COMPUTER 663 [25] A. Ramakrishnan and M. Pecht, Load characterization during transportation, Microelectron. Reliabil., vol. 44, no. 2, pp , January [26] J. Xie and M. Pecht, Application of in situ health monitoring and prognostic sensors, presented at the 9th Pan Pacific Microelectronics Symposium Exhibits and Conf., Oahu, HI, Feb [27] PC Guide (2004). [Online]. Available: ref/hdd/perf/qual/featuressmart-c.html [28] Intel Pentium II Processor Thermal Management (2004). [Online]. Available: htm [29] R. M. Wetzel, Fatigue Under Complex Loading: Analysis and Experiments. Warrendale, PA: Soc. Automotive Eng. Inc., [30] H. Anzai, Algorithm of rainflow method, in The Rainflow Method in Fatigue. Oxford, U.K.: Butterworth-Heinmann, 1991, pp [31] P. Sharma and A. Dasgupta, Micro-mechanics of creep-fatigue damage in PB-SN solder due to thermal cycling Part II: mechanistic insights and cyclic durability predictions from monotonic data, Trans. ASME, J. Electron. Packag., vol. 124, no. 3, pp , [32] C. Zhai, C. Sidharth, and R. Blish, Board level solder reliability versus ramp rate and dwell time during temperature cycling, IEEE Trans. Device Materials Reliab., vol. 3, no. 4, pp , Dec [33] R. Mahajan, C. P. Chiu, and R. Prasher, Thermal interface materials: a brief review of design characteristics and materials, Electron. Cooling, vol. 10, no. 1, pp , [34] IPC-SM-785 Guidelines for Accelerated Reliability Testing of Surface Mount Attachments (1992, Nov.). [Online]. Available: [35] S. Bosworth, S. C. Hsu, and J. Polcari, Exceptional performance from the development, qualification, and implementation of a silicone adhesive for bonding heatsinks to semiconductor packages, IEEE Trans. Compon., Packag. Manufact. Technol. Part A, vol. 18, no. 1, pp , [36] M. F. Ben Achour and A. Bar-Cohen, Mechanical stress reduction in heat sink bond layers a numerical feasibility study, in Proc. Electronic Components Technology Conf. (ECTC), 1999, pp [37] A. Kaisare, Z. Han, D. Agonafer, and K. Ramakrishna, Prediction of effect of heat sink adhesive on mechanical reliability of a wire bonded plastic ball grid array package, in Proc. 8th Intersociety Conf. Thermal and Thermomechanical Phenomena in Electronics Systems (ITHERM 2002), 2002, pp [38] V. Eveloy, P. Rodgers, and M. Pecht, Reliability of pressure sensitive adhesive tapes for heat sink attachment in air-cooled electronic assemblies, IEEE Trans. Device Materials Reliab., vol. 4, no. 4, pp , Dec Nikhil Vichare (S 03) received the B.Sc. degree in production engineering from the University of Mumbai, India, in 1997, and the M.Sc. degree in industrial engineering from the State University of New York at Binghamton in He is currently working toward the Ph.D. degree in mechanical engineering at the University of Maryland, College Park. He is a Graduate Research Assistant at the CALCE Electronic Products and Systems Center, working in the area of electronics prognostics. Mr. Vichare is a student member of the International Microelectronics And Packaging Society (IMAPS) and the Surface Mount Technology Association (SMTA). Peter Rodgers received the Ph.D. degree in mechanical engineering from the University of Limerick, Limerick, Ireland. He is an Assistant Research Professor at the University of Maryland, College Park, where he supports the thermofluid research of the CALCE Electronic Products and Systems Center. Formerly, he was with the Nokia Research Center, Finland, and Electronics Thermal Management Ltd., Ireland, where he consulted on a wide range of aspects in electronics cooling. Apart from health monitoring of electronics, his research interests cover both conventional and advanced cooling technologies for electronic equipment. He has authored or co-authored approximately 50 conference and journal publications on a broad range of topics in this area. He has been an invited lecturer, keynote speaker, panelist and session chair at international electronics thermal management conferences. Dr. Rodgers received the 1999 Harvey Rosten Memorial Award for his publications on the application of computational fluid dynamics analysis to electronics thermal design. He is a member of the EuroSimE, SEMI-THERM and THERMINIC conference program committees, and is program co-chair for EuroSimE Valérie Eveloy received the M.Sc. degree in physical engineering from the National Institute of Applied Science (INSA), Rennes, France, and the Ph.D. degree in mechanical engineering from Dublin City University, Dublin, Ireland. She has been involved in the thermal management, packaging and reliability of electronic equipment for ten years, and is an Assistant Research Scientist at the CALCE Electronic Products and Systems Center, University of Maryland. She was previously with the Nokia Research Center, Finland, and Electronics Thermal Management Ltd., Ireland. Her current research interests focus on extending the limits of air-cooling, the application of computational fluid dynamics to electronic system thermal design, biomedical monitoring and health monitoring of electronics. She has authored or co-authored over 35 conference and refereed journal publications. Dr. Eveloy is a member of several international conference program committees focused on thermal phenomena in electronic systems. Michael G. Pecht (S 82 M 83 SM 90 F 92) received the B.S. degree in acoustics, the M.S. degree in electrical engineering, and the M.S. and Ph.D. degrees in engineering mechanics from the University of Wisconsin at Madison. He is the founder and the Director of the CALCE Electronic Products and Systems Center at the University of Maryland and a Chair Professor. He has consulted for over 50 major international electronics companies, providing expertise in strategic planning, design, test, IP and risk assessment of electronic products and systems. He has written 18 books on electronic products development, use and supply chain management. He has also edited a series of books on the Asian electronics industry, including a recent book titled The Chinese Electronics Industry 2004 Edition. Dr. Pecht has received the 3M Research Award, the IEEE Undergraduate Teaching Award, and the IMAPS William D. Ashman Memorial Achievement Award for his contributions. He is chief editor for Microelectronics Reliability and an Associate Editor for the IEEE TRANSACTIONS ON COMPONENTS AND PACKAGING TECHNOLOGY. He served as chief editor of the IEEE TRANSACTIONS ON RELIABILITY for eight years and has served on the advisory board of IEEE Spectrum. He is a Professional Engineer and an ASME Fellow.