Bi-directional AC electrothermal micropump for on-chip biological applications

The ability to control and pump high ionic strength fluids inside microchannels forms a major advantage for clinical diagnostics and drug screening processes, where high conductive biological and physiological buffers are used. Despite the known potential of AC electro-thermal (ACET) effect in different biomedical applications, comparatively little is known about controlling the velocity and direction of fluid inside the chip. Here, we proposed to discretize the conventional electrodes to form various asymmetric electrode structures in order to control the fluid direction by simple switching the appropriate electric potential applied to the discretized electrodes. The ACET pumping effect was numerically studied by solving electrical, thermal and hydrodynamic multi-physic coupled equations to optimize the geometrical dimensions of the discretized system. PBS solutions with different ionic strength were seeded with 1 μm sized fluorescent particles and electrothermally driven fluid motion was observed inside the channel for different electrode structures. Experimental analyses confirm that the proposed micropump is efficient for a conductivity range between 0.1 and 1 S/m and the efficiency improves by increasing the voltage amplitude. Behavior of the proposed electrode–electrolyte system is discussed by lumped circuit model. Frequency response of system illustrated that the optimal frequency range increases by increasing the conductivity of medium. For 0.18 S/m PBS solution, the constant pumping effect was observed at frequency range between 100 kHz and 1 MHz, while frequency range of 100 kHz to 5 MHZ was observed for 0.42 S/m. The characteristics of experimental results were in good agreement with the theoretical model.

Cartographie et spectroscopie des propriétés mécaniques à l'échelle du nanomètre par spectrométrie acoustique locale à fréquence variable

This work aims at completing the development of local mechanical spectroscopy as initiated by F. Oulevey during his thesis. More precisely, the measurement of the amplitude and phase lag of the strain as a function of the applied stress frequency bas to be settled. Parallel to the development of the technical aspects making such an acquisition possible, the model used to analyze the measured quantities is modified in order to describe the studied system more realistically. The local mechanical spectrometer is based on an Atomic Force Microscope (AFM) using an optical method to detect the deflection of the lever. The stress is applied locally on the sample by a transducer fixed at the base of the cantilever. The bandwidth of the microscope' s electronics does not allow one to detect the cantilever movement at frequencies higher than the MHz. Two different approaches are followed to overcome this limit. The first approach exploits the nonlinearity of the contact to down-convert, via mechanical mixing, the high-frequency signals into low-frequency ones. This work demonstrates the feasibility of this technique, as well as its limits, mainly due to the difficulty to analyze confidently the obtained results. The second approach takes advantage of the stroboscopic detection of the movement of the cantilever. The intensity of the laser used to detect the deflection of the cantilever is modulated at a frequency close to the applied stress frequency. The amplitude of the signal recorded at the difference frequency corresponds then to the high-frequency amplitude of the cantilever movement. The frequency range attainable with this method is only limited by the bandwidth of the transducer providing the excitation. The model connecting the measured quantities (strain amplitude and phase lag) to the desired mechanical properties (elasticity and damping) consists in a beam clamped at one end, with a tip attached at some point along its length. The tip itself is connected to the sample via a spring and a dashpot in parallel, representing the normal interaction, and a spring and a dashpot in parallel representing the interaction in the plane of the sample. Combined with the stroboscopic detection, this model enables one to quantitatively determine the elastic modulus of stiff samples. Many results obtained on a wide range of samples exhibiting very diverse mechanical properties are presented. They demonstrate the validity of the stroboscopic approach, which combines simplicity of use and efficiency.

EPFL2000

Biosystem for the culture and electrical characterisation of epithelial cell tissues

Serge Hediger

The aim of this work was to develop a microchamber system for performing electrophysiological measurements on epithelial surface on culture membranes in the mm2 range. This miniaturised system permits the use of small quantities of epithelial cells, which in relatively short time grow into a tight layer that can be used in electrophysiological (ion transport) experiments. Availability of cells is an issue, for example when the cells originate from human biopsies or from (expensive!) transgenic mice. Apart from having reduced culture surfaces (for use with scarce biological tissues), there are tremendous advantages in having a cell culture mini-chamber. Our systems are with integrated electrical electrodes, micro-fluidic channels and feed-throughs, making them extremely compact and easy to use, thereby avoiding cell perturbation by manipulations. The structures facilitate control of the cell layer growth, the measurement of the cell layer resistance and the transport and diffusion of biological or pharmacological molecules through the cell layer. Moreover we have chosen cheap and easy-to-tool materials for the realisation of disposable devices. We have also fabricated modular devices, in which the cell culture membranes can be reversibly placed within or removed from the system, thereby offering flexibility and economic interest. These microsystems (for biological applications ("biosystems")) are realised using photolithography, various etching procedures (among which powder-blasting), thin film deposition, electrochemical deposition, polydimethylsiloxane (PDMS) moulding and gluing technologies. Both electrical and fluidic characterisation of the biosystems has been performed. Also, a specific study of microelectrode properties of different electrode materials, such as Pt, Ag and Ag/AgCl has been done using various electrochemical experiments and models. The devices are finally tested in real biological experiments. These experiments were carried on with the collaboration of three different biological academic work groups: the group of Prof. W. Hunziker from the Institute of Biochemistry at the University of Lausanne, the group of Prof. Van der Goot from the Department of Biochemistry at the University of Geneva and Prof. J.D. Horisberger from the Institute of Pharmacology and Toxicology at the University of Lausanne. The functionality of our devices has been tested and their potential for the study of transport and diffusion of biological or pharmacological molecules through the cell layer via accurate measurement of (bio-) chemically induced resistance variations, has been demonstrated.

EPFL2002

Microfabricated solid oxide fuel cells

Samuel Rey-Mermet

Micro-fabricated solid oxide fuel cells (µSOFCs) are finding an increasing interest as potential power sources for portable devices such as MP3 players or laptops. The aim of this work was to fabricate a µSOFC demonstrator that works at 500°C and is fuelled by hydrogen. This thesis was divided into two parts. The first one was devoted to the development of an electrolyte and electrodes in form of sputtered thin films with electrical and mechanical properties suitable for the implementation in a real cell. YSZ and CGO electrolyte have been reactively sputtered from metallic targets. Both films are dense and have a columnar microstructure. The ionic conductivity of these films was of 0.5 S/m at 550°C for the CGO and of 5.5 x 10-3 S/m at 500°C for the yttria stabilized zirconia (YSZ) . Albeit the ceria doped gadolinia (CGO) was a better ionic conductor at low temperature, it was not possible to obtain an open circuit voltage (OCV) with a CGO electrolyte film. Most likely, the reduction of the Ce+4 ion into Ce+3 in a hydrogen atmosphere creates an electrical leakage. Better results were obtained with YSZ layers. Single, (111) textured columnar films showed OCV's of 200 mV. Best results were obtained with a double layer of two different microstructures. The first one exhibited a dense, columnar microstructure with (111) texture. The second layer was porous with nanocrystalline grains and preferential (200) orientation. The improved properties are ascribed to the absence of film crossing grain boundaries. Of special interest is the mechanical stress behaviour upon heating to the operation temperature. The stress was investigated as a function of temperature up to 700°C. An anomalous, hysteretic behaviour was found during the first heating cycle in YSZ as well as CGO thin films. This phenomenon could be modelled as an oxygen uptake to fill up excess oxygen vacancies created during the sputtering process. The model allowed to derive a diffusion activation energy of 0.6 eV for these excess vacancies in YSZ. Annealing in air at 700 °C permits to reduce stress and to stabilize the YSZ membrane. As electrode materials, sputter deposited, porous platinum, porous Ni-CGO composites and dense LaxSr1-xCoO3-y (LSCO) thin films were developed and characterized. The PEN (Positive electrode-Electrolyte-Negative electrode) layer processes were combined with Micro Electro Mechanical System (MEMS) process technology to fabricate µSOFC test devices. The PEN membranes were liberated by deep silicon dry etching. The cell diameter was varied between 0.5 and 5 mm, the electrolyte thickness between 500 and 700 nm. A nickel grid grown by electroplating was used to support the electrolyte layer and to serves as current collector for the anode. The cell with a 5 mm diameter shows a very good mechanical stability up to 600°C in SOFC operating conditions and for several heating cycles. The functionality of the fuel cell with two 20 nm thick porous platinum electrodes and a YSZ bilayer electrolyte (500 nm) has been demonstrated. An OCV of 850 mV was measured at 500°C with hydrogen fuel. Unfortunately, a too high cathode contact resistance reduced the current to very low values. The achieved maximal power density was only 19 µW/cm2. A simple design change should remedy the problem.

EPFL2008