In-situ STM analysis of neuromorphic computing via clusters produced by the matrix assembly cluster source (MACS)
Lead Research Organisation:
Swansea University
Department Name: College of Engineering
Abstract
This project will investigate using size-controlled clusters of nano-particles deposited on to patterned supports to create novel electronic devices. In order to deposit the clusters, a smaller version of a novel method of cluster production will be built and used - the Matrix Assembly Cluster Source (MACS), more specifically - the MiniMACS, which is an environmentally friendly method of cluster production when compared to chemical approaches. Our arrangement will allow in-situ study of produced samples, and the potential of scale up to an industrial scale of cluster production by linking in with related projects based at Swansea to achieve this. Using such structures this project will investigate Neuromorphic Computing - alternative architectures for computing. Networks of randomly-deposited clusters will be connected from two contacts, with a field applied. The electro-migration of atoms within clusters will form and break filaments connecting the network. The percolation threshold (the formation of long-range connectivity in random systems) will be investigated. Two resistances states are expected to be observed.
Aims and objectives - To construct the MiniMACS cluster deposition tool within an existing ultra high vacuum chamber which will be re-purposed for this project. To develop a process to deposit metal nanoparticles on to substrates. To investigate the potential neuromorphic computing applications of clusters arrays formed in to filament networks. The research methodology, including new knowledge or techniques in engineering and physical sciences that will be investigated. Scanning Tunnelling Microscopy to image the filaments, PVD to pattern electrodes, cluster beam deposition to create clusters, four point probe to analyse electrical properties/potentially use as a gate (regarding transistors), SEM and AFM to initially observe deposition, and potentially TEM to get atomic resolution imaging, and use of the assembled mini-MACS to create and deposit clusters. The mini-MACS will be unique, and the use of a multi-probe system to drive, perturb and interrogate the filament network may be the first time this has been done.
Alignment to EPSRC's strategies and research areas. Clusters are widely used as catalyst materials and this project will add knowledge in to Swansea's wider EPSRC-funded projects to develop commercial level solvent-free production of nanoparticles - EPSRC priority Sustainable Industries. The investigation of neuromorphic computing offers the potential to develop novel electronic computing devices - EPSRC priority 21st Century Products.
Aims and objectives - To construct the MiniMACS cluster deposition tool within an existing ultra high vacuum chamber which will be re-purposed for this project. To develop a process to deposit metal nanoparticles on to substrates. To investigate the potential neuromorphic computing applications of clusters arrays formed in to filament networks. The research methodology, including new knowledge or techniques in engineering and physical sciences that will be investigated. Scanning Tunnelling Microscopy to image the filaments, PVD to pattern electrodes, cluster beam deposition to create clusters, four point probe to analyse electrical properties/potentially use as a gate (regarding transistors), SEM and AFM to initially observe deposition, and potentially TEM to get atomic resolution imaging, and use of the assembled mini-MACS to create and deposit clusters. The mini-MACS will be unique, and the use of a multi-probe system to drive, perturb and interrogate the filament network may be the first time this has been done.
Alignment to EPSRC's strategies and research areas. Clusters are widely used as catalyst materials and this project will add knowledge in to Swansea's wider EPSRC-funded projects to develop commercial level solvent-free production of nanoparticles - EPSRC priority Sustainable Industries. The investigation of neuromorphic computing offers the potential to develop novel electronic computing devices - EPSRC priority 21st Century Products.
Organisations
People |
ORCID iD |
| Richard Cobley (Primary Supervisor) | |
| James Mccormack (Student) |
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509553/1 | 01/10/2016 | 30/06/2022 | |||
| 2022228 | Studentship | EP/N509553/1 | 01/01/2018 | 30/09/2021 | James Mccormack |
| Description | Electron Microscopy has been employed to image clusters made from both the Matrix Assembly Cluster Source method and direct sputtering. These sources have produced a full range of nanoparticle structures demonstrating growth from single atoms up to several hundreds of atoms in one image. From here, we can use a simulation of structures as a reference guide in order to compare different rotations of shapes in order to identify the correct structure. Currently, the images have to be analysed manually, which can take hours if the image has hundreds of nanoparticles. In the coming months, these images will be used to teach a machine learning program to automatically identify the structure of the shape. The large amounts of data will be useful to see the larger picture, rather than very small areas. |
| Exploitation Route | With the large amount of data produced, simulations of the nanoparticle growth can be tweaked and modified to hopefully allow for more accurate simulations. These simulations are important as these are probes into how nanoparticles behave as a material, be that catalysts or electronics. |
| Sectors | Chemicals,Electronics,Energy,Environment |
| Title | MiniMACS |
| Description | The MiniMACS is a system that is bases upon the principles of the Matrix Assembly Cluster Sources (MACS), but has been developed to create Nanoparticles using more readily available methods of cooling. By trapping evaporated target material of metals in a cold gas condensate, injection of energy through the ion beam can be used to sputter out nanoparticles. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2018 |
| Provided To Others? | No |
| Impact | Discovery of the formation of nanoparticles through direct sputtering. A full range of cluster sizes ranging from single atoms to several hundred atoms |