Abstract: Neuromorphic computing proposes to process information inspiring from the brain mechanisms to aim computationally cost-effective and intuitive manners to process complex data. While neural network algorithms have proven their real pote ntial in diverse pattern classification applications, their replication into hardware remains technologically challenging. An interesting solution for such hardware implementation is based on organic electrochemical transistors (O ECTs) that have shown promises in emulating synaptic plasticity at the device level, and coherent communication from one device to another. We demonstrate (i) the bottom-up fabrication of OECT micro-arrays and (ii) the pattern classification task using this technology. We achieved the electro-polymerization of a new p-type accumulation-mode conducting polymer, iteratively on top of 12 micrometric OECT devices in a honeycomb array. We characterized the rich material's morphology of the bottom-up grown organic semiconductor structures and assessed their functionality as synaptic transistor devices. The exploitation of the array for the recognition of pulse-frequency-modulated gate-voltage patterns through an aqueous electrolyte show ed that the pattern recognition can cope with the rich variability of device performances, but also that the recognition benefits from the large dispersion in device property of each individual OECTs. These results announce a new game-c hange for organic and molecular electronics. It shows that the chemical and morphological richness of these electronic materials, for which their disorder induces large property distributions, are actually enabling information processing in a neurom orphic computing context: at the image of our brain which exploits in a powerful way the natural morphological and chemical variability of its dendritic and synaptic network.

Abstract: Biological computing systems are very inspirational objects, from their structure, organization and up to there computing principles for the development of new computing paradigms. Emulating some of these basic concepts in hardware could potentially revo lutionize our way of processing information. This approach needs to consider the neuromorphic computing paradigm in its globality, from the basic sensors level to the data analysis one. In this presentation, we will put the emphasis on organic materials as a promising platform for future neurmorphic engineering solutions. In particular, we will present an innovative approach that relies on both intrinsic micro-sensors' physics and neuromorphic computing concepts to show pattern classification out of a 1 2-unit bio-sensing array. We adapt the proposition of reservoir computing to demonstrate that relevant computing can be realized based on the ionic dynamics in 400-nm channel-length organic electrochemical transistor (OECT) and the key concept of learnin g. Furthermore, we show that this approach can deal efficiently with the high level of variability obtained by bottom-up electro-polymerized OECT. We discuss the effect of the array size and variability on the performances for a simple real-time classifi cation task paving the way to new sensing and processing approaches.

Abstract: Organic diode rectifier have attracted a lot of attention recently for RF energy harvesting, and much effort has been applied toward extending the ultra-high frequency range. An important parameter that should be considered for diodes used in RF rectifie r is the turn on voltage, which should be low to overcome the problem of the low voltage generated and power extracted from energy harvesting. In this work, we focused on pentacene organic rectifier with high rectification ratio and low threshold voltage obtained by tuning the work function of gold with a self-assembled monolayer of PFBT and optimizing the thickness of the organic layer. We demonstrate a high rectification ratio up to 107 and a very low turn on voltage as low as 20 mV. Flexible rectifie r diode have been also fabricated in release paper WO84, high rectification ratio of 106 was obtained even after bending of the device. The pentacene based rectifier diodes were also demonstrated to operate at more than 1GHz. This provides a great potent ial for fabricating high-performance organic flexible diodes and opens the way for the development of high frequency response using organic materials.