Chair for Multicomponent Materials

Memristive devices

The limitations of traditional computer electronics such as the von-Neumann bottleneck and the predicted end of scaling by Moore’s law impose severe challenges for future increase in computational capabilities. With the increasing needs for processing large amounts of data and energy efficient electronics nowadays, these limitations become more and more evident. Novel computation strategies like the emerging field of neuromorphic engineering attempt to overcome the shortcomings of traditional computing concepts by incorporation of principles of neuronal networks. In the context of the development of novel computing architectures, special attention is paid to memristors. Generally speaking, the key characteristics of memristors is their capability to undergo changes in their electrical resistance as a response to an external electrical field. By this behavior, memristors can be used to imitate biological synapses, which are able to adjust the efficiency of signal transmission and therefore open new paths for information processing hardware.

One focus in the Chair for Multicomponent Materials is the research on nanoparticle composites and their potential application in neuromorphic engineering. We pursue the concept of taking alloy nanoparticles as fundamental building units for memristive switching and apply it to memristive systems across a variety of length scales, ranging from investigations on memristive action in individual nanoparticles towards 0-3 nanocomposites. The group’s experience in a broad variety of deposition methods is a key contribution to the development of nanoparticle-based memristive devices and systems. In this context, the gas aggregation source technique is expanded by a multicomponent target approach and employed for the deposition of alloy nanoparticles of a broad range of material systems like AgPt, AgAu or CuNi. [1–3] Reactive sputtering is applied to deposit oxide thin films. [4] To uncover memristive switching on the level of individual nanoparticles, AgPt nanoparticles were encapsulated in a dielectric SiO2 matrix and the resulting stack was electrically probed by conductive AFM (Fig.1).


Fig 1. Schematic cross section of AgPt nanoparticles, encapsulated in a thin SiO2 layer. To probe the memristive switching on the level of individual nanoparticles, a conductive AFM tip is brought into contact with the sample surface, as depicted in the inset. Figure taken from [5].

Even on this scale, with each nanoparticle having a diameter around 10nm, reproducible memristive switching with diffusive switching characteristics does occur (Fig.2), which impressively proves the potential of our concept of using alloy nanoparticles as building blocks form memristive devices and systems. [2,5] Considering recent challenges in research on memristive devices and neuromorphic computing, our research pursues the idea of combining memristor and sensor functionalities for memsensor systems. As inspired from biological principles, memsensor systems exhibit memristive characteristics which can be modulated by an environmental stimulus like UV light or the presence of gas molecules. In this way, signal processing and detection are closely related. For the development of memsensor devices, we can rely on a broad experience on semiconducting metal oxide sensors for UV and gas sensing applications. [3,4,6] As a proof of concept, we observed memsensor functionalities in a single ZnO microrod, which go beyond its inherent UV sensing functionality. In this system we also found stimulus hysteresis, the occurrence of memristive switching is directly linked to the presence or absence of an UV stimulus. [5,7] Enhanced functionalities like adaptation to an environmental input, as known from biological systems, can potentially arise in memsensitive systems which could in future promote the design of intelligent sensing systems. [5,7]



Fig 2. Diffusive memristive switching as observed by c-AFM measurements. For both voltage polarities, distinct SET and RESET events (as indicated in a) are detected for multiple consecutive cycles (b). Due to the instability of the metallic filament, the device is always in its high resistance state upon zero crossing. Figure taken from [5]


Selected publications

[1] A. Vahl, J. Strobel, W. Reichstein, O. Polonskyi, T. Strunskus, L. Kienle, F. Faupel, Single target sputter deposition of alloy nanoparticles with adjustable composition via a gas aggregation cluster source, Nanotechnology. 28 (2017) 175703. doi:10.1088/1361-6528/aa66ef.

[2] A. Vahl, N. Carstens, T. Strunskus, F. Faupel, A. Hassanien, Diffusive Memristive Switching on the Nanoscale, from Individual Nanoparticles towards Scalable Nanocomposite Devices, Sci. Rep. 9 (2019) 17367. doi:10.1038/s41598-019-53720-2.

[3] V. Postica, A. Vahl, N. Magariu, M.-I. Terasa, M. Hoppe, B. Viana, P. Aschehoug, T. Pauporté, I. Tiginyanu, O. Polonskyi, V. Sontea, L. Chow, L. Kienle, R. Adelung, F. Faupel, O. Lupan, Enhancement in UV Sensing Properties Of Zno :Ag Nanostructured Films by Surface Functionalization with Noble Metalic and Bimetallic Nanoparticles, J. Eng. Sci. XXV (2018) 41–51. doi:10.5281/zenodo.2557280.

[4] A. Vahl, J. Dittmann, J. Jetter, S. Veziroglu, S. Shree, N. Ababii, O. Lupan, O.C. Aktas, T. Strunskus, E. Quandt, R. Adelung, S.K. Sharma, F. Faupel, The impact of O2 /Ar ratio on morphology and functional properties in reactive sputtering of metal oxide thin films, Nanotechnology. 30 (2019) 235603. doi:10.1088/1361-6528/ab0837.

[5] A. Vahl, On the Development of Memsensors, CAU Kiel, 2019. .

[6] V. Postica, A. Vahl, D. Santos-Carballal, T. Dankwort, L. Kienle, M. Hoppe, A. Cadi-Essadek, N.H. de Leeuw, M.-I. Terasa, R. Adelung, F. Faupel, O. Lupan, Tuning ZnO Sensors Reactivity toward Volatile Organic Compounds via Ag Doping and Nanoparticle Functionalization, ACS Appl. Mater. Interfaces. 11 (2019) 31452–31466. doi:10.1021/acsami.9b07275.

[7] A. Vahl, J. Carstensen, S. Kaps, O. Lupan, T. Strunskus, R. Adelung, F. Faupel, Concept and modelling of memsensors as two terminal devices with enhanced capabilities in neuromorphic engineering, Sci. Rep. 9 (2019) 4361. doi:10.1038/s41598-019-39008-5.