Development of stencil lithography for stretchable interconnects

The growing needs for human healthcare have driven the development of wearable devices for monitoring the health status of an individual. Polymers represent a promising platform for such devices due to their ability to form an intimate contact between soft organs and rigid electronic components. Stencil lithography (SL), which uses patterned membranes as a mask to locally deposit materials on substrates, provides a simple but versatile fabrication platform for realizing wearables. However, no studies have explored the deposition of materials on polymer substrates through this technique, impeding further development. Additionally, the freestanding membranes in stencils limit the geometrical design flexibility of apertures. This thesis aims to overcome such limitations, extend the use of SL for structuring stretchable liquid metal (LM) and finally demonstrate the connection of LM to silicon nanomembrane (Si NM) for stretchable Si conductors. This thesis starts by providing a systematic study on the deposition of metals on various polymers including biocompatible and biodegradable materials using SL. The metallic patterns are successfully fabricated on the polymer substrate used by SL. The geometrical and electrical characterizations are performed and discussed. Next, we propose a concept of introducing bridges on stencil membranes to extend the geometrical design flexibility of SL. The applied bridges stabilize the suspended membrane to enable arbitrary apertures. Continuous material traces are produced by taking advantage of the blurring effect. When the stencil is lifted to an appropriate distance above the substrate, the line-of-sight evaporation process results in the material deposition under the shadowed bridge region. The proposed bridge stencils allow the realization of spiral and meandering structures at the micrometer and sub-micrometer scales. The deposition on a variety of substrate materials reveals the versatility of the proposed bridge stencil. We further extend the use of SL for the realization of stretchable LM electronics. A hybrid process combining SL and centrifugal force-assisted patterning of LM is presented. The selective wetting behaviour of oxide-removed eutectic gallium-indium (EGaIn) on metal patterns defined by SL enables micrometer LM patterns on elastomeric substrates. Microscale LM patterns are achieved by efficiently removing the excess material by the centrifugal forces experienced from spinning the substrate. The geometrical and electrical properties of LM patterns on stretchable substrates are characterized. Utilizing bridge stencils, EGaIn serpentine conductors and interdigitated capacitors are fabricated and characterized. Lastly, we translate the LM patterning technique to realize stretchable Si NM conductors using LM as interconnections. Si NM provides exceptional mechanical flexibility and electrical properties, an excellent candidate for wearable electronic applications. The use of LM significantly reduces the need for complex wiring while providing excellent stretchability. PI films are added underneath Si NMs to allow the device to sustain large mechanical deformation. The simulations detail the design strategy of the proposed device configuration. The fabrication methods, including the transfer of single-crystalline Si NMs to polymers and the integration of LM to Si NMs, are developed. The electrical characterizations under mechanical deformation are performed and the results are discussed.

Highly filled polymer nanocomposite films derived from novel nanostructured latexes

Riccardo Ruggerone

The overall aim of this thesis has been to assess the potential of latex-based technologies for the preparation of polymer/clay nanocomposites. The key feature of latex-based technologies is that they offer the possibility of improved control of the final nanocomposite morphology at significantly higher clay loadings than can be obtained with more conventional processing techniques, such as melt blending or in situ polymerization. The idea is to exploit swelling of the clay in either the aqueous or the monomer phase of a water-based latex, depending on the clay surface functionalization, to produce hybrid polymer/clay latex particles with controlled diameters of the order of 100 nm, which may then be consolidated to produce solid nanocomposite films. The materials considered in this work were based on styrenic matrices, considered to be a model system, and acrylics, which are of more interest for commercial coating applications. Two different polymerization techniques were investigated, namely conventional emulsion polymerization and miniemulsion polymerization. The thermal and mechanical properties of films produced from the resulting latexes were then studied in detail. Conventional emulsion polymerization was found to be particularly suitable for the preparation of particles with a well-defined "armoured" morphology, in which the clay formed a more or less complete shell around a matrix core, providing the focus for the remainder of the project. Clay contents of up to about 50 wt % were obtained for both the styrenic and the acrylic latexes using this approach, with excellent degrees of dispersion, the average clay aggregate thickness not exceeding 10 nm. The armoured morphology of the latex particles resulted in a cellular arrangement of the clay in the consolidated films, which became better defined as the clay content increased. The reinforcing effect of the clay on mechanical properties varied according to the physical state of the matrix. Increases in Young's modulus by a factor of 3 to 4 were observed in styrenic films with the cellular morphology in the glassy state, and the degree of exfoliation of the clay was found to be a critical parameter under these conditions, samples containing 5 to 7 wt % of clay showing increased moduli with respect to those obtained at somewhat higher clay contents, for which aggregation was more apparent. In the rubbery state, on the other hand, the Young's modulus increased by more than 2 orders of magnitude for clay contents above 20 wt % and was strongly correlated with the overall filler content. Thermal analysis showed that a significant proportion of the matrix remained immobilized in the rubbery state, i.e. did not contribute to the glass transition. This was argued to be due to strong physical confinement of regions of the matrix intercalated in the clay aggregates. While the increases in Young's modulus in the glassy state could be accounted for in terms of classical micromechanical models, such as those of Halpin-Tsai and Mori-Tanaka, the same models failed to predict the behaviour in the rubbery state. Models based on foam mechanics were therefore developed incorporating an immobilized matrix fraction in the cell walls, whose elastic properties were treated as fitting parameters. Although somewhat different values of the Young's modulus for this immobilized matrix fraction were required to fit the data, depending on the details of the model, they were consistently found to be two to three orders of magnitude greater than that of the neat matrix in the rubbery state (but not to exceed the Young's modulus of the matrix in the glassy state), providing direct evidence for the importance of this interphase for the overall nanocomposite properties. The importance of the cellular arrangement of the clay was also confirmed through comparison with nanocomposites containing non-cellular morphologies resulting from other preparation techniques or the use of mechanical deformation to break-up the initial cellular structure. Finally, it was demonstrated that the results obtained for the styrenic systems could be extended to acrylic-based nanocomposites with comparable morphologies, underlining their broader significance for formulations of potential commercial interest. A further goal of this thesis was to study the effect of the clay on the microdeformation and fracture mechanisms of the nanocomposites, and it was also of interest to compare these results with those obtained for conventional isotactic polypropylene (PP)/clay nanocomposites prepared by melt blending, whose macroscopic properties have been studied previously in our institute. In situ TEM investigation of deformation in glassy styrenic nanocomposite films of about 200 nm in thickness containing the cellular structure revealed a decrease in local matrix drawability at clay contents above 10 wt %, and extensive coarse cavitation, thought to be associated with the particle cores, which replaced crazing as the dominant deformation mechanism, accounting for the observed decrease in macroscopic tensile strength at intermediate clay contents. Moreover, at the highest clay contents, these mechanisms were replaced by failure of the particle-particle interfaces, leading to an extremely brittle macroscopic response. Above Tg, localized deformation zones were also observed, indicating the network formed by the clay and the immobilized regions of matrix to show yielding behaviour, again consistent with the macroscopic response. The conventional PP/clay nanocomposites also showed a decrease in matrix ductility and an increase in coarse cavitation with increasing clay content. In this latter case, however, the zones of cavitation were clearly identified with breakdown of the clay aggregates or the interfaces between the clay and the matrix, underlining the important role of the clay-matrix interface for the properties of the styrenics, and by inference, those of the acrylics. To provide further insight into the role of the cellular structure of the styrenics, finite element (FE) simulations were used to investigate the distribution in hydrostatic stress in the nanocomposite films under finite deformations. These showed the principal stress concentrations to appear in the clay aggregates aligned in the direction of the applied deformation, while the hydrostatic stress in the matrix remained relatively uniform, suggesting that as long as the interface and clay aggregates remain stable, cavitation may initiate anywhere within the matrix, as in the unmodified polymer. The FE simulations were also used to model the elastic properties in the rubbery state, confirming the need to assume the presence of an immobilized matrix fraction in order to account for the observed Young's moduli. The results of these calculations were found to be consistent with those obtained from the foam-based micromechanical models, confirming the applicability of these latter, and suggesting the local anisotropy of the clay aggregates to play a limited role in the overall low strain elastic properties. The main outcome of this thesis has been the establishment of a general physical basis for predicting structure-mechanical property relationships in a new range of latex-based nanocomposite materials with exceptional properties, thanks to the possibility of incorporating very high clay loadings without compromising processability, and vast potential for fine tuning of stiffness and stiffness related properties. Moreover, the work has highlighted the important role of the "nano" effect, particularly at temperatures above the glass transition of the matrix, where matrix immobilization at the clay-matrix interface is clearly demonstrated to contribute to both thermal and mechanical properties. It is consequently expected to provide solid guidelines for the choice of morphological and materials parameters for the optimization of basic mechanical properties in such materials via controlled synthesis. Moreover, with further advances in synthesis and processing and hence in film quality, it should ultimately be possible to extend the basic physical model developed here to account for the structure dependence of other important applicative properties, such as permeability and fire resistance.

EPFL2010

Application of Microstencil Lithography on Polymer Surfaces for Microfluidic Systems with Integrated Microelectrodes

Jürgen Brugger, Nao Takano, Marc Antonius Friedrich van den Boogaart

Microfluidic systems with incorporated microelectrodes are adopted in a variety of applications in life sciences. However, most of micro-electro-fluidic devices are fabricated based on silicon and silicon related materials micromachining processes, which are still expensive for disposable or semi-disposable sensing devices. Plastics and other polymer materials are advantageous to be used as substrates of microfluidic systems in order to realize a low-cost production process of devices, since these materials can be easily structured by e.g. hot-embossing, injection molding or thermal forming with excellent reproducibility. However, the lithographical process of microelectrode fabrication on plastic substrates is one of the critical issues, because many of these materials cannot withstand organic solvents which are unavoidable during conventional photolithography processes. Microstencil lithography, i.e., local deposition of micrometer scale metal patterns through small shadow masks, is a promising method for metal micropattern definition on plastic substrates that are not feasible for solvent-based photoresist technology. Since microstencil lithography is a resistless, single-step direct vacuum patterning method, it enables us to simplify the device fabrication process as well. Thus, we propose to apply microstencil lithography to fabricate microelectrodes onto plastic substrates as part of a low-cost production process of microfluidic devices. However, microstencil lithography is faced with several possible drawbacks. Therefore, we focused on two specific issues accompanying microstencil lithography, which are expected to affect on micro electro fluidic device fabrication process when considered as low-cost, reproducible alternative to standard lithography on plastic substrates: clogging and blurring. One of the main limitations of microstencil lithography is that the method is associated with gradual aperture clogging. To suppress the clogging phenomenon, coating stencils with self-assembled monolayers (SAM) is highly helpful, by which the useful life time of stencils can be extended. But in this study, we traced the clogging of micro-apertures without SAM coating by an observation of the stencil apertures and deposited microstructures after each evaporation step in order to asses the basic feasibility of microstencil lithography. From the series of observation, it was determined that approximately 50 % of the thickness of the evaporated metals (Ti, Cu) was deposited at the side walls of the stencil apertures, i. e., when 200 nm of Ti was evaporated through the microstencil, the thickness of deposited Ti at the side wall of the aperture was 0.1 um. Similar values were observed when Cu was used as a deposited material. The progress of clogging is clearly visible in an SEM image, in which the corresponding microstencil aperture was observed from the side having faced the substrate during the evaporation. These results suggest that the microstencils can be used repeatedly even without a cleaning treatment when the device specifications allow some micrometer tolerance of the electrode size. The other drawback of stencil lithography is related to the gap presence between the stencil and the substrate. We will also present quantitative values of gap induced clogging and discuss the blurring mechanism of deposited structures based on a simplified geometrical model.

2006

Nano-electromechanical systems based on ultra-thin semiconductors

Sajedeh Manzeli

Nowadays, the interest in 2D materials has gone far beyond graphene. Specially, monolayers of transition metal dichalcogenides (TMDs) offer a broad spectrum of electronic and optical properties, and show the potential to revolutionize the electronics industry. The promising electronic properties of 2D semiconductors combined with their mechanical strength and flexibility, makes them ideal candidates for nanoelectromechanical systems (NEMS). This thesis focuses on realizing NEMS based on graphene and MoS2, which is one of the most appealing TMDs. First, we present the realization of graphene NEMS by fabricating single and bilayer graphene transistors featuring a doubly clamped suspended channel. The electromechanical response of monolayer graphene nanoribbons show a strain-induced increase in their electrical resistance, making it possible to estimate an upper limit for the piezoresistive gauge factor. Surprisingly, we observe oscillations in the electromechanical response of bilayer graphene. Our numerical simulations indicate that these oscillations arise from quantum mechanical interference in the transition region induced by sliding of the two graphene layers with respect to each other. Our results report on the rare observation of room temperature electronic interference in bilayer graphene. Next, we investigate the static electromechanical response of atomically thin MoS2. MoS2 exhibits high youngâs modulus and fracture strength. Furthermore, the bandgap of MoS2 is highly strain-tunable. This coupling between electrical and mechanical properties makes MoS2 a promising material for NEMS. Here we incorporate monolayer, bilayer and trilayer MoS2 in a suspended membrane configuration with the electrodes acting as mechanical clamps. Strain-induced bandgap tuning is detected via electrical conductivity measurements and the emergence of the piezoresistive effect in MoS2 is demonstrated. We observe a reversible bandgap modulation in atomically thin MoS2 membranes with a thickness dependent modulation rate. Finite element method (FEM) simulations are used to obtain the spatially varying bandgap profile on the membrane and to quantify the rate of bandgap change. The piezoresistive gauge factor is calculated for single layer, bilayer and trilayer MoS2. Our results reveal that monolayer and bilayer MoS2 show a piezoresistive effect which is comparable to the state-of-the-art silicon strain sensors and two orders of magnitude higher than in graphene. Finally, we present the investigation of MoS2 NEMS resonators working in the VHF range and featuring piezoresistive transduction. The atomic thickness of monolayer MoS2 places it as a promising candidate for miniaturization of electromechanical devices to the limits of vertical scaling. While the small mass of MoS2 leads to an increased resonant frequency and a higher mass sensitivity, the presence of piezoresistivity offers a transduction mechanism in addition to the traditional capacitive transduction. Operating in the tension-dominated regime, monolayer MoS2 NEMS resonators not only allow tunability of the resonant frequency using an external voltage, but also show the strain-induced enhancement of their dynamic range. Furthermore, the resonators are driven into the nonlinear regime allowing the study of nonlinear effects. This work sheds light on the potential of TMD based NEMS as ultra-low power switches, sensors and resonators for applications in RF range.

EPFL2018