NANOSTRUCTURED COMPOSITE FILMS FOR BRAIN-MACHINE INTERFACE APPLICATIONS

Biohybrid electronic technologies, fostering information flow and processing between natural and artificial components [1] are essential for developing energy-efficient, compact interfaces that integrate bioelectrical activity for personalized medicine. In this frame, advances in thin-film sensors and non-invasive signal processing enable wearable systems to monitor brain and peripheral neural activity [2] in view of the development of efficient Brain-Machine Interfaces (BMIs) [3]. While the quality and quantity of signal recordings have improved [4], managing and processing large data volumes in real-time necessitates significant hardware and software advancements. Crucial concern is also the mechanical matching and biocompatibility of the materials used in Micro Electrode Arrays (MEAs) and other recording devices, as well as the proper device geometries and scaling for in vivo brain activity. Nanostructured Cluster Assembled Films (NCAFs) are of particular interest for the development of artificial systems showing neuromorphic behaviors and as the basis for BMIs. Au NCAFs show nonlinear electrical properties and resistive switching behavior [5], [6]. ZrOx NCAFs are biocompatible, and they have been used as substrates for the growth of neural cells. Thanks to their controlled nanoscale roughness, they can be used to induce phenotypic modifications via mechanotransductive signaling [7], [8], [9], [10]. My work was focused on combining these two aspects by fabricating and characterizing a nanocomposite Au-ZrOx system showing resisting switching behavior and capable of enabling a suitable information coding/decoding protocol in view of the development of BMIs. I developed planar layered Au/ZrOx two terminal systems, on glass substrates [11]. An electrical characterization highlighted hysteretic I-V curves, and two separate resistance states explored with alternated polarity pulses. This system showed neuromorphic properties such as short-time memory with stretched-exponential relaxation, and potentiation/depotentiation for subsequent applied voltage pulses. I characterized the bilayered NCAF conduction processes, suggesting a memristive-like model to explain the system's hysteresis characteristics taking into account the competition of a Poole-Frenkel (PF) and a Schottky conduction mechanisms [12]. Layered Au/ZrOx devices can be interfaced with biological systems, but their electrical resistance is very high (around 1 GΩ), being prohibitive for noise and readout operations. Furthermore, their electrical properties are strongly dependent on the characteristics of the 2-D interface between the two films, making it hard to control tuning geometry and deposition parameters. For these reasons, I worked on the development of a deposition system able to produce composite NCAFs where Au and ZrOx clusters are interspersed. I have demonstrated a two-source Supersonic Cluster Beam Deposition Apparatus [Filippo Profumo et al., submitted] and I fabricated by this technique nanocomposite Au/ZrOx NCAFs with different zirconia contents. Their electrical properties, monitored in-situ during fabrication, resulted strongly related to zirconia concentrations and ranged from insulating characteristics for high ZrOx content to those of Au NCAFs. By the ex-situ electrical characterization of nanocomposite devices, I highlighted a “soft” activation of resistive switching with a gradual resistance transition, enhanced resistance retention, and the presence of Negative Differential Resistivity (NDR). Through an in-situ thermal TEM analysis of the morphological rearrangements of composite NCAFs with varying zirconia contents, combined with molecular dynamics simulations, I have addressed the improved stability of composite electrical properties to zirconia clusters acting as an insulating matrix and reducing the thermally driven diffusivity of gold atoms. [Andrea Falqui et al., submitted]. Since NCAFs exhibit complicated quasi-random electrical activity, a different neural signal processing approach than those described for other memristive devices [13] is needed to proceed toward a BMI application. The Inter-Switch Interval (ISI) analysis [14], [15] has been used to analyze the switching features of Au NCAFs under constant applied voltage [5], [16]. However, I mainly focused on pulsed stimuli, producing current traces with non-Gaussian distributions if applied on composite NCAF, and incompatible with an ISI analysis. I thus investigated the autocorrelation function of current traces following a power law decrease, as reported for similar gold nanoparticle systems [15], and sensitive to voltage and alternated polarity of pulses capable of varying its decay trend. To identify distinctive parameters of current recordings and their relationship with input pulse parameters, I introduced the use of a time-series analysis capable of comparing a large number of statistical methods at once [17]. I evinced that the voltage and polarity of applied pulses influence significantly other statistical features of the current trace recorded from the nanocomposite devices and that general voltage input signals (input-timeseries) are mapped into current recording (output-timeseries) with new statistical features. Exploiting this input/output map, I have demonstrated the possibility of classifying, in real-time, in-vivo neural recordings of rats according to their registration location and activity (normal or abnormal) [Filippo Profumo, submitted], representing a step towards the decoding of neuronal signals by the intrinsic processing of neuromorphic devices and thus towards the development of NCAF-based BMI. References [1] S. Vassanelli and M. Mahmud, ‘Trends and Challenges in Neuroengineering: Toward “Intelligent” Neuroprostheses through Brain-“Brain Inspired Systems” Communication’, Front. Neurosci., vol. 10, Sep. 2016, doi: 10.3389/fnins.2016.00438. [2] A. Mikhaylov et al., ‘Neurohybrid Memristive CMOS-Integrated Systems for Biosensors and Neuroprosthetics’, Front. Neurosci., vol. 14, p. 358, Apr. 2020, doi: 10.3389/fnins.2020.00358. [3] L. F. Nicolas-Alonso and J. Gomez-Gil, ‘Brain Computer Interfaces, a Review’, Sensors, vol. 12, no. 2, pp. 1211–1279, Jan. 2012, doi: 10.3390/s120201211. [4] E. Musk, ‘An integrated brain-machine interface platform with thousands of channels’, 2019. [5] M. Mirigliano et al., ‘Complex electrical spiking activity in resistive switching nanostructured Au two-terminal devices’, Nanotechnology, vol. 31, no. 23, p. 234001, Mar. 2020, doi: 10.1088/1361-6528/ab76ec. [6] M. Mirigliano and P. Milani, ‘Electrical conduction in nanogranular cluster-assembled metallic films’, Advances in Physics: X, vol. 6, no. 1, p. 1908847, Jan. 2021, doi: 10.1080/23746149.2021.1908847. [7] C. Schulte et al., ‘Conversion of nanoscale topographical information of cluster-assembled zirconia surfaces into mechanotransductive events promotes neuronal differentiation’, J Nanobiotechnol, vol. 14, no. 1, p. 18, Dec. 2016, doi: 10.1186/s12951-016-0171-3. [8] C. Schulte, A. Podestà, C. Lenardi, G. Tedeschi, and P. Milani, ‘Quantitative Control of Protein and Cell Interaction with Nanostructured Surfaces by Cluster Assembling’, Acc. Chem. Res., vol. 50, no. 2, pp. 231–239, Feb. 2017, doi: 10.1021/acs.accounts.6b00433. [9] A. Previdi et al., ‘Micropatterning of Substrates for the Culture of Cell Networks by Stencil-Assisted Additive Nanofabrication’, Micromachines, vol. 12, no. 1, p. 94, Jan. 2021, doi: 10.3390/mi12010094. [10] A. Previdi et al., ‘Nanotopography and Microconfinement Impact on Primary Hippocampal Astrocyte Morphology, Cytoskeleton and Spontaneous Calcium Wave Signalling’, Cells, vol. 12, no. 2, p. 293, Jan. 2023, doi: 10.3390/cells12020293. [11] F. Profumo, F. Borghi, A. Falqui, and P. Milani, ‘Potentiation and depression behaviour in a two-terminal memristor based on nanostructured bilayer ZrO x /Au films’, J. Phys. D: Appl. Phys., vol. 56, no. 35, p. 355301, Aug. 2023, doi: 10.1088/1361-6463/acd704. [12] D. Cipollini, F. Profumo, L. Schomaker, P. Milani, and F. Borghi, ‘Conduction mechanisms in a planar nanocomposite resistive switching device based on cluster-assembled Au/ZrOx films’, Front. Mater., vol. 11, May 2024, doi: 10.3389/fmats.2024.1385792. [13] I. Gupta, A. Serb, A. Khiat, R. Zeitler, S. Vassanelli, and T. Prodromakis, ‘Real-time encoding and compression of neuronal spikes by metal-oxide memristors’, Nat Commun, vol. 7, no. 1, p. 12805, Sep. 2016, doi: 10.1038/ncomms12805. [14] M. Karsai, K. Kaski, A.-L. Barabási, and J. Kertész, ‘Universal features of correlated bursty behaviour’, Sci Rep, vol. 2, no. 1, p. 397, Dec. 2012, doi: 10.1038/srep00397. [15] J. B. Mallinson, S. Shirai, S. K. Acharya, S. K. Bose, E. Galli, and S. A. Brown, ‘Avalanches and criticality in self-organized nanoscale networks’, Science Advances, vol. 5, no. 11, p. eaaw8438, Nov. 2019, doi: 10.1126/sciadv.aaw8438. [16] G. Nadalini, F. Borghi, T. Košutová, A. Falqui, N. Ludwig, and P. Milani, ‘Engineering the structural and electrical interplay of nanostructured Au resistive switching networks by controlling the forming process’, Sci Rep, vol. 13, no. 1, Art. no. 1, Nov. 2023, doi: 10.1038/s41598-023-46990-4. [17] B. D. Fulcher, M. A. Little, and N. S. Jones, ‘Highly comparative time-series analysis: the empirical structure of time series and their methods’, Journal of The Royal Society Interface, vol. 10, no. 83, p. 20130048, Jun. 2013, doi: 10.1098/rsif.2013.0048.