Abstract: Conducting polymer dendrite (CPD) morphogenesis is an electrochemical process that enables in materio evolving intelligence in wetware devices. During CPD morphogenesis, voltage transients drive the physical evolution of electrically conductive structure s, thereby programming their filtering properties as nonlinear analog devices. Either studied as an electrochemical experiment or as neuromorphic devices, the dependence of the electrical properties of the electrogenerated structures on the chemical comp osition of their growth environment is still unreported. In this study, we report the existing intrication between the nature and concentration of the electrolytes, electroactive compounds and co-solvents and the electrical and the electrochemical proper ties of CPDs in an aqueous electrolyte. CPDs exhibit various chemical sensitivities in water: Their morphology is highly dependent on the nature of the chemical resources available in their environment. The selection of these resources therefore critical ly influences morphogenesis. In addition, the concentration of the different electrochemical species have varying impacts on growth dynamics, conditioning the balance between thermodynamic and kinetic control on polymer electrosynthesis. By correlating t he dependencies of these evolving objects with the availability of the chemical resources in an aqueous environment, this study offers guidelines to tune the degree of evolution of electronic materials in water and highlights potential avenues for their application. Such evolving hardware is envisioned to exploit the chemical complexity of real-world environments as part of information processing technologies.

Abstract: Neuromorphic engineering is emerging as a promising approach for the next generation of Artificial Intelligence. While advances in bio-inspired algorithms and computing mechanisms strive to reproduce biology's capacity to adapt, learn continuously, and e volve within unstructured environments, significant breakthroughs are also needed at the hardware level. The diversification of materials and circuits, particularly through organic electronics, offers alternative solutions to current hardware, which prim arily relies on parallelization and increased computing resources. In this talk, we will present how electropolymerization of PEDOT-based materials can emulate biology's ability to define optimal computing topologies. We will demonstrate how a bottom-up approach to growing interconnections between nodes enables the discovery of optimal networks for specific tasks. In the second part of the presentation, we will discuss the integration of these dendritic networks onto 2D substrates. We will show how inte grated PEDOT networks and iono-electronic processes in organic electrochemical dendritic transistors can result in complex, non-linear signal transformations-supporting key computing functions used extensively in neural networks.

Abstract: Electropolymerization under an alternating-current results in the formation of conducting polymers dendrites (CPDs), that conduct both ionic matter and electronic charges simultaneously, offering features from both the worlds of electronics and electroch emistry. Versatile, they can be grown in various electrolytes to develop classes of electronic components that are evolvable and process information using mass-transfer mechanisms. By self-healing or resorbing, CPD have the potential to enable new functi onalities in conventional electronic systems with low material and energy costs, making them a promising avenue for bio-inspired information processing. They also offer a simple, low-voltage alternative to address the ongoing problem of high manufacturin g costs in the microelectronics industry. In this work, we investigate the control of poly(3,4-ethylenedioxythiophene) (PEDOT) based CPD morphology through electrolyte chemistry and its impact on impedance patterns in a two-electrode system, particularly in relation to their observed constant phase element (CPE) behavior. We also explore how morphology influences the charge/discharge dynamics when the dendritic connection is not yet completed. Specifically, it is shown that the electrical parameters of the CPDs, extracted by fitting the transient curves using the Mittag-Leffler function, are defined early during the growth, and that thicker CPDs will allow longer relaxation times. By changing the voltage pulse duration in the growth signal, one has the refore the ability to tune both the characteristic times and the non-ideality of a CPD charge. Ultimately, we aim to demonstrate the applicability of these concepts for programming sensors and integrating neuro-inspired functionalities into electronic no ses, which exploit electrochemistry for the recognition of complex environmental patterns.

Abstract: Process variation has always been a challenge to mitigate in electronics. This especially holds true for organic semiconductors, where reproducibility concerns hinder industrialization. Challenging this concept, we show AC-electropolymerization to be a p owerful platform for the development of morphology-dependent computing hardware, thanks precisely to its intrinsic stochasticity. Our findings reveal that electropolymerized polymer dendrite networks exhibit a complex structure-operation relationship tha t allows to implement nearly linear to nonlinear functions. Moreover, dendritic networks can integrate a limitless number of inputs from their environment, which can be used to our advantage in the context of in materio computing to discriminate between different spatiotemporal inputs. These results position electropolymerization as a pivotal technique for the bottom-up implementation of computationally powerful objects. We anticipate this study to help shifting the negative perception of variability in the material science community and promote the electropolymerization framework as a foundation for the development of a new generation of hardware defined by its topological richness.

Abstract: Contributions of organic semiconducting materials to electronics are particularly hard to assess: As macromolecular organizations, they have low enthalpy so they can be processed in soft conditions and they have resilience to deformation. However, for th e same reason, they have also broader density of energy states and more instabilities than silicon in ambient. Controlling matter's order at low scale and its properties for as long as possible were always golden standards for microelectronics. Neverthel ess, in a time where brain functioning rises even more as a source of inspiration, shall it still be so? Here are presented clues on how physical property dispersions may be relevant features for information generator nodes to recognize patterns. In a co ntext where the information to recognize is not trivial to physically define, no model can rule sensors' classification a priori. Despite this, broadening the conducting polymer temporal responses in a sensing array allows recognizing dynamical voltage p atterns, or broadening conducting polymer's chemistry in a sensing array enlarges a classifier's perception field to recognize solvent vapors in air. By the nature of property dispersions in regards to the information to recognize, physical variabilities (structural and chemical) can be assets to exploit for pattern recognition and not necessarily drawbacks to bypass for hardware manufacturing. The brain architecture is also transient: a part of the processed information is engraved in its topology, sho wing that a hardware classifier can make use of physical instabilities as part of its programing, by forming new connections in a nodal architecture. Some evidences are also presented here, on how dendritic morphogenesis of a conducting polymer can be a mean to store past voltage experiences in the impedance between nodes in a topology. Very distinct electrochemical features appear in the readout impedance information after growth and these features are to be associated with the shape of a voltage wave inputted on the junction. By the physical implementation of materials' disorder and transience in electronics devices, it is expected that organic semiconductors will integrate essential ingredients in future-emerging information generator nodes beyond s ensors: from embedded random information generating resources to evolving abilities in information classification architectures.

Abstract: In the quest to change the way we envision computing, the ability to create connections in between computing nodes on demand offers the very exciting possibility to explore bottom-up strategies when it comes to thinking the design of electronic devices a nd circuits. Recently, the growth of conductive polymer fibers via electropolymerization (most notably poly(3,4-ethylenedioxythiophene) doped with poly(styrenesulfonate), abbreviated as PEDOT:PSS) was employed in the realm of neuromorphic hardware, first as a way to tune the resistance of a connection, and then to perform computing tasks, such as biosignal classification trough reservoir computing, thus establishing the advantages of the method.In this work, we propose to demonstrate that electropolymer ization is a useful tool for the creation of sensors that are able to realize computing tasks in an aqueous environment. By taking advantage of the morphology of the fibers, we show that PEDOT:PSS dendrites can discriminate between different types of vol tage pulses emitted in their vicinity by a local gate electrode, thus performing in materio classification.In addition, we discuss the growth of networks of polymer fibers on 2D substrates. The ability to create structures that cover several microelectro des allows us to study the behavior of the whole system instead of a single dendrite, and it highlights the relationship that relates the morphology of the object and its electrical properties. In particular, we show that dendrites present an asymmetric non-linear behavior due to the volume of polymer that increases the capacitance of the device albeit not participating in conduction. Moreover, because of the ionoelectronic coupling that exists within an electrolyte, two active devices working concomita ntly will influence one another. We demonstrate that it is a blessing for the realization of computing tasks, such as logic, and that it can also be used to program the dendrites into non-volatile conductance states. Therefore, dendritic networks could c onstitute a new building block for non-conventional information processing that fits into the larger framework of brain-inspired computing.

Abstract: Neuromorphic computing and engineering is capitalizing heavily on the new physical properties offered by nantechnologies to engineer biological processes. At the frontiers in between bio-mimetism and bio-inspiration, various solutions have been proposed for synaptic plasticity or neuronal features based on discrete memory elements, bistable switches or transistors circuits. One missing element that has been missing in the neuromorphic toolbox is the ability to reproduce the complex 3D interconnections o bserved in biological neural networks. Here, we propose to take advantage of bipolar electropolymerization of PEDOT dendritic fibers to reproduce the ability of neural networks to generate complex topologies. The electropolymerization mechanism is used t o realize structural plasticity based on Hebbian-like plasticity rules. We explore how such bottom-up process can find optimal topologies for specific computing tasks. We demonstrate that such optimal topologies results in a drastic reduction of intercon nects for classification and reconstruction tasks, thus offering an interesting option for neural network design.

Abstract: In the quest to change the way we envision computing, the ability to create connections in between computing nodes on demand offers the very exciting possibility to explore bottom-up strategies when it comes to thinking the design of electronic devices a nd circuits. Recently, the growth of conductive polymer fibers via electropolymerization (most notably poly(3,4-ethylenedioxythiophene) doped with poly(styrenesulfonate), abbreviated as PEDOT:PSS) was employed in the realm of neuromorphic hardware, first as a way to tune the resistance of a connection, and then to perform computing tasks, such as biosignal classification trough reservoir computing, thus establishing the advantages of the method. In this work, we propose to demonstrate that electropolyme rization is a useful tool for the creation of sensors that are able to realize computing tasks in an aqueous environment. By taking advantage of the morphology of the fibers, we show that PEDOT:PSS dendrites can discriminate between different types of vo ltage pulses emitted in their vicinity by a local gate electrode, thus performing in materio classification. In addition, we discuss the growth of networks of polymer fibers on 2D substrates. The ability to create structures that cover several microelect rodes allows us to study the behavior of the whole system instead of a single dendrite. Complex interactions through iono-electronic coupling arise when several fibers are interconnected. These mechanisms (e.g. self- / inter-gating, ionic relaxation,...) offer new possibilities for computing signals and could potentially open the way for new applications.

Abstract: Over the past few years, organic electronics - and especially organic mixed ionic electronic conductors (OMIECs) - has taken bio sensing and neuromorphic applications to a whole new level. However, one of the major limitations of the mainstream technolog ies today is that electronic circuits need to be pre-shaped according to the intended use and the expected outcome. This top-down approach, far from being flexible/adaptive, does not really make the most of the resources at hand, as it is hard to predict precisely where cells will be located. To counter that, we can either choose to increase the density and the number of electrodes, so that the entire area would be mapped, or shift from a top-down to a bottom-up approach which would allow for a more enl ightened decision-making process. Recently, the electrodeposition of PEDOT:PSS has been explored as a novel technique to grow conducting polymer films and fibers on non-conductive substrates. The work of Janzakova and coworkers took that concept a step f urther by using electropolymerization of EDOT as a way to create freestanding dendritic-like conductive fibers in a 3D environment, paving the way for in operando material modification, and in fine bottom-up fabrication routes that would be more adaptive and allow for more flexibility. Moreover, it was lately showed that these objects could work as Organic Electrochemical Transistors (OECTs). Here, we explore the possibility of growing dendritic-like PEDOT fibers on Multielectrode Arrays (MEAs) via elec tropolymerization of EDOT. Electrophysiological measurements are based on the capacitive coupling between cells and the electrode material. In comparison with local electrodes, the dendritic objects present spatially distributed impedance due to the exte nsions of their dendritic branches interacting with the biological environment. We investigate the relation between morphology and impedance in these dendritic-like fibers by using a non-conventional Electrochemical Impedance Spectroscopy (EIS) setup tha t will allow us to apply a potential difference between the two ends of the dendrites, thus studying how biasing them can affect their behavior. Moreover, it appears that dendritic fibers can be considered both as passive electrodes as well as active dev ices. We explore the use of these two strategies in the context of electrophysiological measurements. Finally, the ability to record biological signals results from the interaction between cells and an electrode. Unconventional objects such as dendrites present spatio-temporal filtering properties that could affect the recording of such signals. We investigate how tuning the impedance of a dendrite might be used to record efficiently bio-signals.

Abstract: The development of electronic devices such as microelectrode arrays (MEAs), used to record extracellularly simultaneous electrical activity of large populations of neurons is blooming. To enhance the quality of the recordings, the use of electrode made o f conducting polymer such as PEDOT has recently emerged for optimizing the performance of microelectrodes due to its mixed ionic electronic conduction, biocompatibility and low impedance. However, the extent to which these new interfaces can help the alg orithmic pipelines of spike sorting, i.e. turning extracellular potentials into individual spike trains remains unexplored. To address this issue, we checked if the physical positions of the neurons could be reliably inferred from extracellular electrica l recordings obtained by MEAs, and thus be exploited by downstreams spike sorting algorithms. To do so, we combine high resolution images of neuronal tissues and dense recordings performed via high performant electropolymerized electrodes based MEAs. Fir stly, we report the use of EDOT electropolymerization to tune post-fabrication material and geometrical parameters of passive microelectrodes. The process optimizes the cell/electrode interface by decreasing its impedance and improving its affinity with neurons: results demonstrate a better biocompatibility and improved signal-to-noise ratio (SNR) (up to 40 dB). Thanks to the higher SNR, we were able to detect more cells in comparison with gold electrodes from the same neural network by using spike sort ing. Hence, the higher number of cells detected will lead into more accurate analysis of the localization of the active neurons on top of MEA. Secondly, by using these high performant MEAs, we investigated the possibility to accurately estimate the posit ions of the neurons solely from extracellular recordings by studying the correlation between electrical activity (obtained via spike sorting), optical imaging (Fluorescent) and Scanning Electron Microscopy (SEM) of neural networks cultured on MEAs. By us ing the SpykingCircus software to spike sort the extracellular recordings, we estimated the positions of the neurons either by using the center of mass of their electrical signatures, or by inferring the positions assuming cells would behave as monopoles . By superposition of the fluorescent and the SEM images, we compared the observed physical positions of the neurons with the ones predicted by the two aforementioned methods. This approach showed the high accuracy of the monopole hypothesis compared to the center of mass. In this work, we showed how the use of a material engineering technique for optimizing state of art MEAs can enhance the quality of the recordings. Hence, the correlation of these high quality recordings with optical imaging paves the way towards new algorithmic possibilities for spike sorting algorithms that could make use of a more reliable estimation of neuronal positions.

Abstract: The development of electronic devices for neurosensing is leading to fundamental discoveries in communication setups for interfacing and computing the brain's electrical activity that is still a demanding task in the 21st century. One of the greatest cha llenges for efficient neurosensing is to ensure that detection/transduction between biochemically rich systems and tools is fully mastered to reliably gather relevant information. In extracellular devices such as microelectrode arrays (MEAs), the discord ance lies at the interface between ions and the electrodes. Engineering chemically/morphologically the electrode's materials by decreasing its impedance, improving its affinity with neurons, and boosting its biocompatibility ensures better cell/electrode interface conditions to find the right materials that detect ionic signals from neurons and transduce them into electronic signals with the lowest information loss. Hence, the use of conducting polymers (PEDOT) has emerged for optimizing the performance of microelectrodes in neurosensing due to its mixed ionic electronic conduction, biocompatibility and low impedance. In parallel to the development of passive microelectrode, organic electrochemical transistors (OECTs) have received lots of attention in the biosensing field since they exhibit high coupling with cells and signal amplification. Notably, the transconductance represents an important parameter that depends on geometrical and material parameters that rules largely OECTs performances in biose nsing. In this direction, we explore the use of EDOT electropolymerization to tune post-fabrication material and geometrical parameters of passive microelectrodes for optimizing the cell/electrode interface by decreasing its impedance and improving its a ffinity with neurons (increasing the resistance "Rseal" that represents the cell/electrode cleft). For electropolymerized PEDOT MEAs, we demonstrate long term and stable extracellular recording of primary cortical neurons with a record signal-to-noise ra tio (SNR) up to 37 dB (with ultra-low noise down to 2.1 μV RMS). Secondly, for active sensing with OECTs, this strategy exploits the concept of adaptive sensing where both transconductance and impedance are tuned simultaneously or independently. This approach shows an improvement of OECTs transconductance by 150-percent, volumetric capacitance by 300-percent, and a reduction in array's variability by 60-percent in comparison with standard spin-coated OECTs. The cytotoxicity of the electropolymerized EDOT was assessed for primary neural cells culture and no detrimental effect of electropolymerized EDOT on cell viability was observed. To extract the impedance and transconductance values for both MEAs and OECTs, we combine DC electrical measurements w ith electrochemical impedance spectroscopy (EIS). To show the cell/electrode morphology and neurite outgrowth to electropolymerized microelectrodes, Scanning Electron Microscopy (SEM) was performed. To correlate the morphological changes of the material with the enhancement of its electrical and electrochemical performances, Atomic Force Microscopy (AFM) in liquid and Raman Spectroscopy were achieved. Finally, in-vitro extracellular recorded signals from entorhinal cortex cultured slices and primary cor tical neurons using both MEAs and OECTs are presented. The key novelty of this technique is to propose a post-fabrication material engineering technique that can be used to optimize both passive (MEAs) and active (OECTs) devices for extracellular recordi ng and promote new exploratory sensing strategies to ensure high quality neurosensing alternatives.

Abstract: Iono-electronic materials and devices are suscitating lots of interest from both bio-electronics and neuromorphic research communities. In the one hand, iono-electronic materials are offering attractive features such as bio-compatibility, water environme nt operation and efficient ionic to electronic signals transduction. In the other hand, functional devices based on such materials (organic electrochemical transistors, for instance) have shown multiple neuromorphic features from synaptic plasticity to d endritic integration. This talk will present how electropolymerization of PEDOT:PSS materials can be used in both a bio-electronic and a neuromorphic perspective. Electropolymerization of OECT sensors can indeed be advantageously used for optimizing / tu ning the iono-electronic responses of organic electrochemical transistors, thus paving the way to plastic electrophysiological sensors. Notably, we will show how transconductance and volumetric capacitance are evolving with potenstiostatic electropolymer ization. Secondly, bipolar AC electropolymerization can be used to engineer dendritic-like fibers of PEDOT:PSS. Such unconventional structures can implement various neuromorphic concept such as structural plasticity and synaptic plasticity. Finally, comp uting task taking advantage of both electropolymerized sensors and neuromorphic computing will present temporal patterns classification with reservoir computing approach.