Nano AI-Driven

Concepts

Molecular nanotechnology (MNT) is a technology based on the ability to build structures to complex, atomic specifications by means of mechanosynthesis. This is distinct from nanoscale materials. Based on Richard Feynman's vision of miniature factories using nanomachines to build complex products (including additional nanomachines), this advanced form of nanotechnology (or molecular manufacturing) would make use of positionally-controlled mechanosynthesis guided by molecular machine systems. MNT would involve combining physical principles demonstrated by biophysics, chemistry, other nanotechnologies, and the molecular machinery of life with the systems engineering principles found in modern macroscale factories.

A molecular assembler, as defined by K. Eric Drexler, is a "proposed device able to guide chemical reactions by positioning reactive molecules with atomic precision". A molecular assembler is a kind of molecular machine. Some biological molecules such as ribosomes fit this definition. This is because they receive instructions from messenger RNA and then assemble specific sequences of amino acids to construct protein molecules. However, the term "molecular assembler" usually refers to theoretical human-made devices. A nanofactory is a proposed system in which nanomachines (resembling molecular assemblers, or industrial robot arms) would combine reactive molecules via mechanosynthesis to build larger atomically precise parts. These, in turn, would be assembled by positioning mechanisms of assorted sizes to build macroscopic (visible) but still atomically-precise products.

Atomically precise manufacturing is manufacturing of artificial artifacts with at least topological atomic precision has a dedicated focus towards more advanced technology like gemstone metamaterial technology and its production devices gemstone metamaterial on-chip factories (featuring positional atomic precision all the way up to the macroscale). With atomic precision one refers to structures where the positions of all the included atoms are known in a topological sense meaning one knows which atom connects with which. An atomically precise structure may well be floppy such that thermal movement makes the actual positions of the atoms completely unknown. Many base structures for self assembly (in technology level 0 and technology level I) are examples for floppy AP structures e.g. short DNA half strands (oglionucleotides). In technology level I whole sturdy structures out of sturdy AP-building blocks are assembled in a digital fashion. One is dealing with atomically precise structures but one only needs sub block size positioning precision for positional assembly. In technology level II and technology level III diamondoid materials are the main building material. They allow not only the topological position but also the position in three dimensional space to be known (positional atomic precision). Atoms do roughly behave like a construction set with elastic linkages only if the right set of atoms is chosen. Metals with their undirected bonding tend to diffuse at room temperature destroying topological order and thus often do not preserve Atomic precision (AP) making them unsuitable for nanomachinery.

Papers about the use of AI for nanoscale and scanning microscope:

Introduction: Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems

https://arxiv.org/abs/2102.06321

1. Atomic-Level structural engineering of graphene: A convolutional neural network (CNN) is used for identification of the atom positions, their element-specific contrast, and the resulting topology:

https://pubs.acs.org/doi/10.1021/acs.nanolett.1c01214#

2. The Atomic Simulation Environment (ASE) is a set of tools and Python modules for setting up, manipulating, running, visualizing and analyzing atomistic simulations:

https://wiki.fysik.dtu.dk/ase/index.html https://gitlab.com/ase/ase

3. Machine learning for laser-induced electron diffraction imaging of molecular structures: Combining electron diffraction with machine learning presents new opportunities to image complex and larger molecules in static and time-resolved studies:

https://www.nature.com/articles/s42004-021-00594-z

4. Embedding human heuristics in machine-learning-enabled probe microscopy: The very small number of machine learning approaches to probe microscopy published to date, however, involve classifications based on full images. Given that data acquisition is the most time-consuming task during routine tip conditioning, automated methods are thus currently extremely slow in comparison to the tried-and-trusted strategies and heuristics used routinely by probe microscopists (Philip Moriarty):

https://iopscience.iop.org/article/10.1088/2632-2153/ab42ec/meta

5. Automated Searching and Identification of Self-Organized Nanostructures: use a combination of Monte Carlo simulations, general statistics, and machine learning to automatically distinguish several spatially correlated patterns in a mixed, highly varied data set of real AFM images of self-organized nanoparticles:

https://pubs.acs.org/doi/10.1021/acs.nanolett.0c03213

6. Improving the segmentation of scanning probe microscope images using convolutional neural networks. Segmentation strategy using the U-Net convolutional neural network has some benefits over traditional automated approaches and has particular potential in the processing of images of nanostructured systems:

https://iopscience.iop.org/article/10.1088/2632-2153/abc81c/meta

7. Automation method for the identification of defects prior to atomic fabrication via hydrogen lithography using deep learning. We trained a convolutional neural network to locate and differentiate between surface features of the technologically relevant hydrogen-terminated silicon surface imaged using a scanning tunneling microscope. (Robert Wolkow):

https://iopscience.iop.org/article/10.1088/2632-2153/ab6d5e

8. Automated methods based on machine learning to automatically detect and recondition the quality of the probe of a scanning tunneling microscope:

https://arxiv.org/abs/1803.07059

9. TOP [REVIEW] Machine learning at the (sub)atomic scale: next generation scanning probe microscopy: We discuss the exciting prospects for a step change in our ability to map and modify matter at the atomic/molecular level by embedding machine learning algorithms in scanning probe microscopy (with a particular focus on scanning tunnelling microscopy, STM). This nano-AI hybrid approach has the far-reaching potential to realise a technology capable of the automated analysis, actuation, and assembly of matter with a precision down to the single chemical bond limit. >>>>>> Are the nanobots nigh? "Extending the Millipede methodology to the atomic level, combined with autonomous control and correction of the apices of the tip array, would enable a dramatic step change in our control of matter with atomic precision":

https://iopscience.iop.org/article/10.1088/2632-2153/ab7d2f/meta

10. TOP! Autonomous robotic nanofabrication with reinforcement learning: Here, we present a strategy to work around both obstacles and demonstrate autonomous robotic nanofabrication by manipulating single molecules. Our approach uses reinforcement learning (RL), which finds solution strategies even in the face of large uncertainty and sparse feedback. We demonstrate the potential of our RL approach by removing molecules autonomously with a scanning probe microscope from a supramolecular structure. Our RL agent reaches an excellent performance, enabling us to automate a task that previously had to be performed by a human.

https://www.science.org/doi/10.1126/sciadv.abb6987

From: Momalab http://momalab.org/index.php/?action=index

NEW: Atomic Boltzmann machine

Physical implementation of a BM directly in the stochastic spin dynamics of a gated ensemble of coupled cobalt atoms on the surface of semiconducting black phosphorus. Implementing the concept of orbital memory utilizing scanning tunnelling microscopy, we demonstrate the bottom-up construction of atomic ensembles whose stochastic current noise is defined by a reconfigurable multi-well energy landscape. Exploiting the anisotropic behaviour of black phosphorus, we build ensembles of atoms with two well-separated intrinsic time scales that represent neurons and synapses. By characterizing the conditional steady-state distribution of the neurons for given synaptic configurations, we illustrate that an ensemble can represent many distinct probability distributions.

https://arxiv.org/abs/2005.01547

Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems: This Review is written for new and experienced researchers working at the intersection of both fields. We first provide concise tutorials of computational chemistry and machine learning methods, showing how insights involving both can be achieved. We follow with a critical review of noteworthy applications that demonstrate how computational chemistry and machine learning can be used together to provide insightful (and useful) predictions in molecular and materials modeling, retrosyntheses, catalysis, and drug design.

https://pubs.acs.org/doi/10.1021/acs.chemrev.1c00107

Quantum machine learning using atom-in-molecule-based

We combine atom-in-molecule-based fragments, dubbed ‘amons’ (A), with active learning in transferable quantum machine learning (ML) models. The efficiency, accuracy, scalability and transferability of the resulting AML models is demonstrated for important molecular quantum properties such as energies, forces, atomic charges, NMR shifts and polarizabilities and for systems including organic molecules, 2D materials, water clusters, Watson–Crick DNA base pairs and even ubiquitin. Conceptually, the AML approach extends Mendeleev’s table to account effectively for chemical environments, which allows the systematic reconstruction of many chemistries from local building blocks.

https://www.nature.com/articles/s41557-020-0527-z

Bonus:

Perspectives of Molecular Manipulation and Fabrication (Book): https://www.springerprofessional.de/en/perspectives-of-molecular-manipulation-and-fabrication/13280370

Control on a molecular scale: A perspective: https://ieeexplore.ieee.org/document/7525387

Machine Learning for Molecular Simulation Annual Review of Physical Chemistry: https://www.annualreviews.org/doi/10.1146/annurev-physchem-042018-052331

Multiscale Dynamics Simulations: Nano and Nano-bio Systems in Complex Environments (Book): https://pubs.rsc.org/en/content/ebook/978-1-83916-178-0

Autonomous Scanning Probe Microscopy in Situ Tip Conditioning through Machine Learning: https://pubs.acs.org/doi/10.1021/acsnano.8b02208

Oculus Rift virtual reality goggles to hand-controlled manipulation with 3D visual feedback that displays the currently executed tip trajectory and the position of the SPM tip during manipulation in real time:

1. Hand controlled manipulation of single molecules via a scanning probe microscope with a 3D virtual reality interface by P. Leinen, M. F. B. Green, T. Esat, C. Wagner, F. S. Tautz, and R. Temirov, J. Vis. Exp. 116, e54506 (2016)

2. Virtual reality visual feedback for hand-controlled scanning probe microscopy manipulation of single molecules by P. Leinen, M. F. B. Green, T. Esat, C. Wagner, F. S. Tautz, and R. Temirov, Beilstein J. Nanotechnol. 6, 2148 (2015)

[email protected]: Maltimore/robotic_nanofabrication.git

Biomolecular Machines: From diatoms to silica-based biohybrids and Self-Assemebly:

0. Diatoms with Invaluable Applications in Nanotechnology, Biotechnology, and Biomedicine: Recent Advances: The micro- to nanoscale properties of the diatom frustules have garnered a great deal of attention for their application in diverse areas of nanotechnology and biotechnology, such as bioimaging/biosensing, biosensors, drug/gene delivery, photodynamic therapy, microfluidics, biophotonics, solar cells, and molecular filtrations. https://pubs.acs.org/doi/10.1021/acsbiomaterials.1c00475

1. From diatoms to silica-based biohybrids: Diatom nanotechnology is becoming a new field of research where biologists and materials scientists are working together! https://pubs.rsc.org/en/content/articlelanding/2011/cs/c0cs00122h/unauth

2. Diatom Nanotechnology provides a comprehensive overview of the material and its uses. The first part of the book looks at the distinctive porous silica structure of diatoms, the mechanism of their formation and their properties. Individual chapters then explore the broad range of their applications in nanotechnology including nanofabrication, optical biosensors, gas sensors, water purifications, photonics, drug delivery, batteries, solar cells, supercapacitors, new adsorbents and composite materials. https://www.amazon.com.br/Diatom-Nanotechnology-Progress-Applications-Nanoscience-ebook/dp/B078KZWQ1K

3. Biomimetic and bioinspired silica: recent developments and applications https://pubs.rsc.org/en/content/articlelanding/2011/cc/c0cc05648k/unauth

4. All New Faces of Diatoms: Potential Source of Nanomaterials and Beyond https://www.frontiersin.org/articles/10.3389/fmicb.2017.01239/full

5. Diatoms with Invaluable Applications in Nanotechnology, Biotechnology, and Biomedicine: Recent Advances (2021) https://pubs.acs.org/doi/abs/10.1021/acsbiomaterials.1c00475

6. Diatoms: Self assembled silica nanostructures, and templates for bio/chemical sensors and biomimetic membranes https://pubs.rsc.org/en/content/articlelanding/2011/an/c0an00602e

7. Nanoscale Self-Assembly for Therapeutic Delivery https://www.frontiersin.org/articles/10.3389/fbioe.2020.00127/full

8. A Matter of Size Book (Philip Moriarty) Chapter: 5 Molecular Self-Assembly https://www.nap.edu/read/11752/chapter/1

9. Nanofabrication by self-assembly https://www.sciencedirect.com/science/article/pii/S1369702109701567

10. Nanofabrication approaches for functional three-dimensional architectures by self-assembly https://www.sciencedirect.com/science/article/abs/pii/S174801321930502X

Simulators:

Atomic Simulation Environment (ASE) - Python https://gitlab.com/ase/ase.git

AtomAI Python package for machine learning-based analysis of experimental atomic-scale and mesoscale data from electron and scanning probe microscopes https://github.com/pycroscopy/AICrystallographer.git

PiNN: A Python Library for Building Atomic Neural Networks of Molecules and Materials https://github.com/Teoroo-CMC/PiNN#:~:text=PiNN%20is%20a%20Python%20library,Yunqi%20Shao%20at%20Uppsala%20Unversiy. https://pubs.acs.org/doi/10.1021/acs.jcim.9b00994