Hydro-Dynamic (EHD) effects could exhibit interesting performances when a moving fluid is needed, as is the case with mono or diphasic thermal management systems [1] or for soft robotics actuators [2]. Indeed, EHD pumps are non-mechanical devices that allow inducing flow motion of dielectric fluids through a channel under the influence of a high electric field (typically 10V/mu m). This phenomenon arises from the unbalance between dissociation and recombination of neutral species in the polarized media [3]. In conduction mode, the efficiency of such pumping system is highly depending on the fluid properties and especially the mobility of dissociated species [4]. In this paper we present the theory and a methodology associated with a specific test bench that allows investigating in the mobility of dissociated species in various fluids. This makes possible to assess their respective effectiveness for the obtention of high performances EHD pumping microsystems.

Using heat pipe solutions for the thermal management of embedded missiles is all the more complex than military electronic systems have to sustain very harsh environments (accelerations, vibrations and temperatures) for long-term applications. However, Electro-Hydro-Dynamic (EHD) pumps have demonstrated promising performances for draining of fluids. Their main advantage is that no mechanical part is involved during operation. Moreover, compatible with MEMS technologies, they could be miniaturized for very dense electronics. The main goal of this paper was to determine influent design parameters for EHD pumps (made with Cu planar footprints on epoxide FR4 substrates). Various fluids were also tested, showing the most efficient solutions compatible with heat transfer applications.

System-in-a-Package (SiP) aims to integrate a functional sub-system with one or more semiconductor chips along with passive components onto a substrate. SiP is transferred molded to the OEM for second level assembly using industry standards and high volume equipment. In most recent applications (high I/O density, improved electrical performance), LGA (Land Grid Array) is a very practical solution for the 2nd solder joint level. LGA packages reduce integrated circuit mounted height by eliminating the solder balls that are used for mounting BGA (Ball Grid Array) packages to system circuit boards. Instead, LGA packages are reflow-mounted using solder that has been applied to the board. This paper addresses the thermomechanical lifetime reliability of a 104 lead SiP LGA with lead-free solder joints submitted to thermal cycles in environmental chamber. The goal is to determine the impact of the internal IC (Integrated Circuits) on the 2nd level solder joint reliability. It firstly explains how to conduct preliminary technological analyses (RX views, chemical attacks, layout re-design) and material property characterization (CTE in the 3 axis on very small samples) before the construction of an accurate 3D model. Then it focuses on simulation tools (boundary conditions and post-processing cartography methods) necessary for unsymmetrical geometries. Simulation results (life-time prediction and damage localization) lead to quite satisfactory data when compared with first experimental data. Even at the 2nd solder joint level, the first failures (not located in the periphery of the LGA device) are highly linked with the internal components and with the layout of the SiP. The more realistic are the geometry and the material properties, the better are the numerical results. Such 3D simulation models can then be used to predict the life-time as well as to define microsection axis in complex failure analyses. The paper also shows that dummy compon- - ents (daisy-chained) are not convenient for the process assessment of SiP LGA.

This paper presents in details the fatigue life prediction models developed for lead-free assemblies in harsh environments during the 3-year European LIFE project GEAMCOS. The approach uses advanced finite element modelling as well as experimental fatigue data. It is also based on a complete material characterisation for a wide range of stress conditions. It focuses on the most critical packages of the full lead-free GEAMCOS test board and of the functional demonstrator developed for military communication.

This study introduces a computation method to assess damage in electronic solder joints under random vibration. It addresses full 3D dynamic behaviour of electronic board. Finite element modelling (FEM) of electronic BGA (ball grid array) and CGA (column grid array) packages assemblies are developed and adjusted with experimental modal identification of the test board. Vibration FEM simulations are performed to calculate stress transfer functions of critical solder joints. Then, solder joint time-stress responses due to an input random excitation of the board are generated. Stress range distributions are established from rainflow counting. Finally, linear damage computation is done and compared with experimental results by using an empirical damage law. The potential and the limitation of this method are discussed.

As part of the European LIFE project GEAMCOS, this paper presents the simulation evaluation of lead-free electronic assemblies submitted to harsh environments encountered in aeronautical and military communication applications. It addresses the following questions: (1) the determination of an optimal accelerated test condition in the temperature range of -40degC/100degC for assemblies with SnAgCu alloy, (2) the evaluation of the life-time of five assemblies with SnAgCu and SnPb, (3) a sensitivity analysis on design and serigraphy parameter variations, (4) the determination of a test protocol in vibration environments. All the results give relevant life-time predictions. The CSP and PQFP components should have longer time-to- failure with SnAgCu solder joints than with SnPb joints. The contrary is expected for 1st and 2nd level interconnects of large BGA packages. In any case, the factor is found to remain less than about 2. The PCB thermal expansion coefficient has the highest impact, followed by the solder joint height. The first level interconnect reliability is not impacted by any of the parameter changes. Simulation work in GEAMCOS project shall go on in 2007-2008 with an experimental validation of the thermo- mechanical simulations. This will require to identify a new constitutive material model for SnAgCu and to calibrate the simulations with experimental test results.