Online proceedings - EDA Publishing Association

24-26 September 2008, Rome, ItalyDesign of a static TIM testerV.Székely, G.Somlay, P.G.Szabó, M.RenczBudapest University of Technology & EconomicsDepartment of Electron Devicesszekely|somlay|szabop|rencz@eet.bme.huAbstract- Nowadays the quality of thermal interfacematerials (TIM) has a growing importance as the increasingdissipation level of ICs requires more and more sophisticatedsolutions to reduce the R th along the heat-flow path. The recentapproach of TIM manufacturers is to use nanoparticles asfillings in TIM materials, in order to enhance considerably theTIM thermal conductivity. On the other hand this solutionraises difficulties in the characterization of these materials.Namely, the resolution of the conventional methods is not highenough to measure resistance values as low as 0.01-0.05 Kcm 2 /W. This is the reason why we developed static TIMtester equipment using some new concepts. The main ideabehind our design is to use the capabilities of microelectronicsin order to make small sized sensors both for temperature andheat flux sensing. This way it is possible to place these sensorsin the closest proximity of the measured sample. The status ofthis work is presented in the paper.I. INTRODUCTIONFor the measurement and characterization of the TIMsproperties many methods have been designed, however mostof them are based on the ASTM standard D-4570. This testprocedure is a standard method to measure thermalresistance and bulk conductivity for TIMs such as greases,pads, tapes and phase change materials. The inspected TIMsample is placed between a hot and a cold meter bar and aconstant heat flux is applied. The ASTM test defines thermalresistance per unit area (θ) to include the thermal resistanceof the material (θ material ) plus the interfacial contact resistanceof the TIM to the substrates (θ interface ). The heat resistance ofthe tested material is calculated from the known heat fluxforced through the tester and from the measured heat drop.The ASTM standard also prescribes that the temperaturemeasurements should be done at a clamping force of 3 MPato reduce the heat resistance between the calorimeters or heatflux meters and the specimen.This standard is only valid under the followingassumptions:• One dimensional heat flow• Thickness does not change during measurement• Contact resistance does not depend on thicknessThe most severe shortcoming of the ASTM method is itsuse of higher pressure than used in real applications [1]. Thehigh pressure reduces the contact resistance between thespecimen and the meter bars. In cases of greases and phasechange materials the high pressure will result a lower gap,therefore the total θ total will be lower than in an actualapplication.Tester based on this standard was used in the measurementof CNT (carbon nanotube) array based TIM for the thermalmanagement of microelectronic packages and highbrightness light emitting diode (HB-LED) packages [2].To obtain a profile of the temperatures across the interfacea thermal test vehicle has been developed and presented in[3]. This information provides insight into uniformity of thethermal resistance and provides information about theperformance of TIMs over their life cycles.At Sun Microsystems, Inc. the ASTM standard tester hasbeen modified in both the hardware and methodology toallow measurement of high-performance TIMs underconditions that more closely resemble the use-condition inapplication [4].Due to the high (10-20%) reproducibility error of thestandard a different method has been presented in [5]. TheInterface Thermal Resistance Tester (IRT) consists of twosilver cylinders between which the sample is clamped. Thetemperature of the lower block can be changed very fast witha jet impingement technique using two water baths. Thetemperature response of the upper block is a measure for theinterface thermal resistance.The resolution and the reproducibility of the conventionalmethods is not high enough to measure thermal resistancevalues as low as 0.01-0.05 Kcm 2 /W. This is the reason whywe decided to develop static TIM tester equipment usingsome new concepts.The main idea behind our design is to use the capabilitiesof microelectronics in order to make small sized sensorsboth for temperature and heat flux sensing [6]. This way it ispossible to place these sensors in the closest proximity of themeasured sample. Further new idea is to build a symmetricalstructure with reversible direction of the heat flow, as certaininaccuracies of the measurement (as offset errors of thesensors) can be dropped out by making measurements inboth direction of heat flow.The principal parts of the measurement arrangement areshown in Fig.1. The heat flow is driven by two,symmetrically positioned Peltier cells. Both the heat flux andthe temperature drop are measured by the two silicon dicesconstituting the grip area of the mount.©EDA Publishing/THERMINIC 2008 132ISBN: 978-2-35500-008-9