Self-consistent Tight-binding Method Research Articles

Effect of vacancy defect on the structural and electrical properties of single-walled silicon nanotube

The effect of vacancy defects on zigzag and armchair single-walled silicon nanotubes (SiNTs) has been investigated using the self-consistent charge density functional tight-binding (SCC-DFTB) method. The findings indicate that the introduction of defects leads to the inability of the smaller-diameter tube to uphold its tubular structure and cluster. Conversely, the larger-diameter tube undergoes an elliptical transformation, yet retains tubular characteristics and hexagonal structures displaying specific symmetry. The double vacancy defect alters the atomic arrangement on the side opposite the defect. The stability of single-walled SiNTs is related to defect type and chiral index. In addition, the single vacancy defect increases the binding energy and stability of the zigzag tube, while causing a rapid decrease in stability of the armchair tube (7,7). The spin polarization phenomenon manifests within the defect tube, resulting in the splitting of the initial degenerate energy band and a significant decrease in its degeneracy. The existence of electronic interaction, results in a prominent manifestation of vacancy-induced magnetism in armchair and zigzag defect tubes. This phenomenon gives rise to localized spin-up bands situated below the Fermi level, as well as localized spin-down bands positioned above the Fermi level. In the energy band structure of most nanotubes, there are conduction bands or valence bands passing through the Fermi level, leading to a transition from semiconductor to semimetal or metal properties.

Structure and electrical properties of bilayer zigzag silicene nanoribbons depending on stone-wales and applied electric fields by SCC-DFTB method

The influence of geometry and electrical properties of Stone-Wales (SW) defects and an applied electric field on bilayer zigzag silicene nanoribbons (ZSiNRs) were investigated using the self-consistent charge density functional tight-binding (SCC-DFTB) method. The results show that concave deformation occurs at SW defects, leading to varying degrees of disruption in the original symmetrical structure. The overall stability of the marginal Stone-Wales (MSW) nanoribbons surpasses that of the central Stone-Wales (CSW) nanoribbons, SW-II type defects are prone to form in bilayer nanoribbons. The band gap of MSW defective nanoribbons expands significantly, realizing the transition from metallic to semi-metallic and semi-conductor properties. The electrons are transferred from the inner to the outer side of the nanoribbons, with a higher prevalence of defect nanoribbons compared to those that are perfect. The electric field intensity significantly alters the properties of both MSW-I and MSW-II defective nanoribbons.

Structure and electrical properties of double-walled silicon nanotubes depending on 585 defects and applied electric fields by SCC-DFTB method

The geometric structural and electronic properties of 585 defective double-walled silicon nanotubes (DWSiNTs) by using the self-consistent charge density functional tight-binding (SCC-DFTB) method. Furthermore, we investigated the impact of external electric field and intensity on the electronic properties. After the relaxation of perfect tubes, the structure of the tube changes significantly, and the winding direction of the tube wall changes. The buckling value of the tube wall is most closely related to the geometric shape and atomic arrangement order of tubes. The overall stability of zigzag tubes is higher than armchair tubes, and (6,6)@(10,10) stability is significantly affected by defects. The partially defected zigzag tubes change from direct band gap semiconductor properties to indirect band gap semi-metal properties. A transformation of the semiconductor-semimetal properties, direct-indirect band gap of DWSiNTs is realized. The charge transfer amplitude of the defect tubes is greater than the perfect tube. The applied electric field makes the stability of all tubes decrease linearly, and the stability of the defect tubes is higher than the perfect tube. Part of the electric field strength changes the defect tubes from semiconductor properties to semi-metal properties, and semi-metal properties to metal properties. These theoretical studies provide a reference for the preparation and future application of DWSiNTs.

Variation in properties of double-walled zigzag and armchair silicon nanotubes depending on SW defects and applied electric fields by SCC-DFTB method

The geometric structure and electrical properties of the zigzag DWSiNT (8,0)@(12,0) and armchair DWSiNT (5,5)@(7,7) perfect tubes with the same diameter are simulated by self-consistent charge density functional tight binding method (SCC-DFTB). We considered the defective tubes by introducing three types of Stone-Wales (SW) defects, and further investigated the impact of both strong and weak electric fields on these defective tubes. Calculations demonstrate that the atomic arrangement regularity, degree of buckling, stability, energy gap, and charge distribution strongly depend on the type of perfect tubes. Following the introduction of the SW defects, the symmetry of the tubes disappeared, and there is a obvious atomic aggregation occurring at the defects. The armchair tubes demonstrate a higher stability compared to zigzag tubes and all exhibit metallic properties. SWⅡ and SWⅢ defects in zigzag tubes result in a transition from semiconductor to semi-metallic properties. The number of charge transfer increases, and the degree of atomic aggregation is more dispersed. The effect of applied electric field strength on perfect zigzag tubes is more obvious. The weak field has minimally impacted on the stability, and the strong field makes the stability improved to some extent. The presence of a strong field caused the energy gap values of (8,0)@(12,0) tube decrease sharply, and all tubes are transformed into semi-metal or metal. The weak electric field has a more pronounced regulation of band gap within the range of 0 V/nm - 0.3 V/nm on the inner tube. The consequences of this investigation can certainly be helpful in future experimental studies.

Effects of Stone-Wales defects on structure and electrical properties of double-walled silicon nanotubes by SCC-DFTB method

In this work, the geometric structure, system stability, electrical properties and charge distribution of zigzag double-walled silicon nanotubes (DWSiNTs) (nin,0)@(nout,0) (nin = 5–7, nout = 8–13) perfect tubes and defective tubes with three types of Stone-Wales (SW) defects were studied using the self-consistent charge density functional tight-binding (SCC-DFTB) method. The change of the chirality index between the inner and outer walls was simulated. The results show that the symmetry of the structure disappears, and the warping of the tube wall changes from order to disorder after introducing SW defects. The warps of outer tube are obviously higher than that of inner tube. The perfect tubes with inner wall chirality index (6,0) are more stable than other tubes. SW defects of outer wall tubes have a higher effect on system stability than those of inner wall tubes. The difference of inner and outer wall chirality indices plays an obvious role in regulating the energy gap of the system, half of which are near-metal properties and the other half semiconductor properties, and the regulation range is 0.1075–0.8519 eV. All types of SW defects at all locations have a certain regulatory effect on the energy gap of DWSiNTs, and the effect of type III defect is the most obvious. The effect of atomic charge redistribution in the tube is related to the structure, and the charge transfer in a perfect tube is symmetrical. The number of electrons obtained at the defect is increased, and the electronegativity is stronger, due to the curvature of local atoms produced at the defect. This study provides a reference for the preparation and future application of DWSiNTs.

Island-Type Graphene-Nanotube Hybrid Structures for Flexible and Stretchable Electronics: In Silico Study.

Using the self-consistent charge density functional tight-binding (SCC-DFTB) method, we study the behavior of graphene-carbon nanotube hybrid films with island topology under axial deformation. Hybrid films are formed by AB-stacked bilayer graphene and horizontally aligned chiral single-walled carbon nanotubes (SWCNTs) with chirality indices (12,6) and 1.2 nm in diameter. In hybrid films, bilayer graphene is located above the nanotube, forming the so-called "islands" of increased carbon density, which correspond to known experimental data on the synthesis of graphene-nanotube composites. Two types of axial deformation are considered: stretching and compression. It has been established that bilayer graphene-SWCNT (12,6) hybrid films are characterized by elastic deformation both in the case of axial stretching and axial compression. At the same time, the resistance of the atomic network of bilayer graphene-SWCNT (12,6) hybrid films to failure is higher in the case of axial compression. Within the framework of the Landauer-Buttiker formalism, the current-voltage characteristics of bilayer graphene-SWCNT (12,6) hybrid films are calculated. It is shown that the slope of the current-voltage characteristic and the maximum values of the current are sensitive to the topological features of the bilayer graphene in the composition of graphene-SWCNT (12,6) hybrid film. Based on the obtained results, the prospects for the use of island-type graphene-nanotube films in flexible and stretchable electronic devices are predicted.

Computational insights into the adsorption mechanisms of anionic dyes on the rutile TiO2 (1 1 0) surface: Combining SCC-DFT tight binding with quantum chemical and molecular dynamics simulations

Metal oxides are gaining momentum rapidly for application in water pollutant remediation. In this study, the adsorption proprieties of two anionic dyes, i.e., acid yellow 36 (AY36) and acid orange 6 (AO6) on the (110) surface of rutile titanium dioxide (TiO2) in an aqueous medium were investigated using computational methods. The density functional theory (DFT) was used to determine the reactivity of organic molecules by calculating the frontier molecular orbital energies, energy gap (ΔEgap), chemical hardness (η), chemical softness (σ), electronegativity (χ), chemical potential (μ), electrophilicity (ω), the fraction of electrons transferred (ΔN), back-donation energy (ΔEback-donation), Mulliken charge, and Fukui indices. The obtained results showed that the AY36 molecule is more reactive than the AO6 molecule and may have a good adsorption capability compared to the AO6 dye. The most favorable adsorption configurations of AY36 and AO6 molecules were investigated using molecular dynamics (MD) simulation. The calculated interaction energies by MD simulation showed that the TiO2 (110) surface has a high sensitivity to interact with the two anionic dyes, with more affinity toward the AY36 molecule. Furthermore, to get deep insights into the chemistry of interactions between the anionic dyes and the TiO2 (110) surface, the self-consistent charge density functional tight-binding (SCC-DFTB) method was carried out. Results showed that anionic dyes adsorbed on the TiO2 (110) surface by forming covalent bonds between oxygen atoms of the sulfonic group and Ti atoms. Theoretical insights from this work would serve as a guide for researchers to explore the application of oxides in water pollutant remediation.

An Efficient Way to Model Complex Iron Carbides: A Benchmark Study of DFTB2 against DFT.

Iron carbides have attracted increasing attention in recent years due to their enormous potential in catalytic fields, such as Fischer-Tropsch synthesis and the growth of carbon nanotubes. Theoretical calculations can provide a more thorough understanding of these reactions at the atomic scale. However, due to the extreme complexity of the active phases and surface structures of iron carbides at the operando conditions, calculations based on density functional theory (DFT) are too costly for realistically large models of iron carbide particles. Therefore, a cheap and efficient quantum mechanical simulation method with accuracy comparable to DFT is desired. In this work, we adopt the spin-polarized self-consistent charge density functional tight-binding (DFTB2) method for iron carbides by reparametrization of the repulsive part of the Fe-C interactions. To assess the performance of the improved parameters, the structural and electronic properties of iron carbide bulks and clusters obtained with DFTB2 method are compared with the previous experimental values and the results obtained with DFT approach. Calculated lattice parameters and density of states are close to DFT predictions. The benchmark results show that the proposed parametrization of Fe-C interactions provides transferable and balanced description of iron carbide systems. Therefore, spin-polarized DFTB2 is valued as an efficient and reliable method for the description of iron carbide systems.

Toward an efficient f-in-core/f-in-valence switchable description for DFTB calculations of Ce 4f states in ceria.

A computational protocol is developed for efficient studies of partially reduced redox-active oxides using the self-consistent charge density functional tight-binding method. The protocol is demonstrated for ceria, which is a prototypical reducible oxide material. The underlying idea is to achieve a consistent (and harmonized) set of Slater-Koster (SK) tables with connected repulsive potentials that enable switching on and off the in-valence description of the Ce 4f states without serious loss of accuracy in structure and energetics. The implicit treatment of the Ce 4f states, with the use of f-in-core SK-tables, is found to lead to a significant decrease in computational time. More importantly, it allows for explicit control of the oxidation states of individual Ce atoms. This makes it possible to "freeze" the electronic configuration, thereby allowing the exploration of the energetics for various meta-stable configurations. We anticipate that the outlined strategy can help to shed light on the interplay between the size, shape, and redox activity for nanoceria and other related materials.

Carboxyl Functionalization of N-MWCNTs with Stone-Wales Defects and Possibility of HIF-1α Wave-Diffusive Delivery.

Nitrogen-doped multi-walled carbon nanotubes (N-MWCNTs) are widely used for drug delivery. One of the main challenges is to clarify their interaction with hypoxia-inducible factor 1 alpha (HIF-1α), the lack of which leads to oncological and cardiovascular diseases. In the presented study, N-MWCNTs were synthesized by catalytic chemical vapor deposition and irradiated with argon ions. Their chemical state, local structure, interfaces, Stone-Wales defects, and doping with nitrogen were analyzed by high resolution transmission electron microscopy (HRTEM), Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), and near-edge X-ray absorption fine structure (NEXAFS) spectroscopy. Using experimental data, supercells of functionalized N-MWCNTs with an oxygen content of 2.7, 4 and 6 at. % in carboxyl groups were built by quantum chemical methods. Our analysis by the self-consistent charge density functional tight-binding (SCC DFTB) method shows that a key role in the functionalization of CNTs with carboxyl groups belongs to Stone-Wales defects. The results of research in the decoration of CNTs with HIF-1α demonstrate the possibility of wave-diffusion drug delivery. The nature of hybridization and relaxation determines the mechanism of oxygen regulation with HIF-1α molecules, namely, by OH-(OH-C) and OH-(O=C) chemical bonds. The concentration dependence of drug release in the diffusion mode suggests that the best pattern for drug delivery is provided by the tube with a carboxylic oxygen content of 6 at. %.

Effects of edge hydrogenation and applied electric field on the structure and electrical properties of zigzag silicene nanoribbons by SCC-DFTB calculations

The effects of edge hydrogenation and applied electric field in different directions on the geometrical structure and electrical properties of zigzag silicene nanoribbons (ZSiNRs) with different period widths (width = 2,3,4) were investigated by using a self-consistent charge density functional tight-binding (SCC-DFTB) method. The results show that the degree of warpage of the ZSiNRs is significantly changed by hydrogenation, while the bond length and bond angle are less affected. The edge hydrogenation reduces the Si-dangling bonds in the nanoribbons, which increased the binding energy of nanoribbon and enhanced system stability. In the absence of an electric field, the unhydrogenated nanoribbons with different period widths all exhibit semiconducting properties, and the hydrogenated nanoribbons exhibit metallic or semi-metallic properties. Under the vertical external electric field in the z-direction, the stability and energy gap of the unhydrogenated nanoribbons are sensitive to the change of strength, and the oscillation fluctuation is large. With the increase of strength, the atoms in the outermost four layers of the nanoribbons show obvious charge gain and loss. When the vertical electric field is applied to the hydrogenated nanoribbons, the energy gap changes are related to the period width. Their electrical properties realized the conversion between the semi-metal properties and the direct band gap semiconductor properties. Charge transfer occurs between the adjacent two layers of atoms. Under the same electric field strength, the amount of charge transfer increases from left to right along the direction of the width of the nanoribbon; at the same time, with the increase of the electric field strength, the amount of charge transfer also presents an increasing trend.

Quasi-point versus point nodes in Sr2RuO4, the case of a flat tight binding γ sheet

We perform a numerical study of the unitary regime as a function of disorder concentration in the imaginary part of the elastic scattering cross-section for the compound Sr2RuO4 in the flat band non-disperse limit. By using a self-consistent tight binding (TB) method, we find a couple of families of Wigner probabilistic functions that help to explain macroscopically the distribution between Fermion dressed quasiparticles and Cooper pairs, and also the position of nodes in the order parameter for Sr2RuO4. Therefore, we are able to show that a TB model for the FS γ-sheet, numerically shows 4 point nodes in a flat γ sheet limit, or 4 quasi-point nodes for strong dispersion γ sheet limit in the reduced phase scattering space (RPS).

Effects of chiral index, vacancy defects and external electric field on the structures and electronic properties of DWSiNTs by SCC-DFTB calculations

As a new candidate material, silicon nanotubes (SiNTs) have been widely studied. However, various types of structural defects and external electric field are usually inescapability introduced in the experimental process of SiNTs synthesis. Structures, stability and electronic properties of the materials especially for double-walled silicon nanotubes (DWSiNTs) require to be considered. In this work, effects of vacancy defects and external electric field on properties of the geometrically structural optimization and electronic properties of DWSiNTs zigzag (3,0)@(7,0) and armchair (3,3)@(7,7) are studied via the self-consistent charge density functional tight binding (SCC-DFTB) method. The results of us show that the regular arrangement of the atoms, quantum molecular descriptors, distribution of charge, the degree of warping, stability, energy gap and Fermi energy levels intensely hinge on the chirality index of the tube and impacts of vacancy defects and electric field. Especially, the appearance of monovacancy and the change of vacancy density have the most significant effect on the band gap. For the pristine tubes, the charges transfer from the inside to the outside along the radial direction of the tube exhibiting an axisymmetric distribution. With the introduction of defects, the range of the charge transfer is broadened dramatically. The direction of charge transfer in the applied electric field is always opposite to the field direction. Our work may provide a new route to tune the electronic properties of DWSiNTs and opens a wider field of its application in nano-electronic devices.

Role of the Backbone when Optimizing Functional Groups─A Theoretical Study Based on an Improved Inverse-Design Approach.

We present an improved inverse-design approach for automatically identifying molecular (or other) systems with optimal values for prechosen properties. The new approach uses SMILES (simplified molecular input line entry system) to describe molecular structures efficiently, a genetic algorithm to optimize the molecules automatically, and the DFTB+ (self-consistent charge density functional tight-binding) method to calculate electronic properties. Thereby, almost every class of materials─even macromolecules or monomers─can be studied easily. Without crossover operators but with only mutation operators, the genetic algorithm is more adaptive to SMILES while keeping its efficiency. DFTB+ is more accurate than the DFTB method used in our previous inverse-design approach for the study of excited states and charge transfer processes. The improved approach is applied to optimize benzene, pyridine, pyridazine, pyrimidine, and pyrazine derivatives for seven electronic properties, which all are highly relevant and important for the performance of molecules in solar cells. We found that for some electronic properties, the precise composition and structure of the backbone have remarkable impacts on the value of the electronic properties and/or on the set of functional groups that leads to the best performance. On the contrary, for other properties, these effects are less pronounced. The reasonable optimal functional groups and/or substitution patterns are reported.

DNA-Mimicking Metal-Organic Frameworks with Accessible Adenine Faces for Complementary Base Pairing.

Biologically derived metal–organic frameworks (Bio-MOFs) are significant, as they can be used in cutting-edge biomedical applications such as targeted gene delivery. Herein, adenine (Ade) and unnatural amino acids coordinate with Zn2+ to produce biocompatible frameworks, KBM-1 and KBM-2, with extremely defined porous channels. They feature an accessible Watson–Crick Ade face that is available for further hydrogen bonding and can load single-stranded DNA (ssDNA) with 13 and 41% efficiency for KBM-1 and KBM-2, respectively. Treatment of these frameworks with thymine (Thy), as a competitive guest for base pairing with the Ade open sites, led to more than 50% reduction of ssDNA loading. Moreover, KBM-2 loaded Thy-rich ssDNA more efficiently than Thy-free ssDNA. These findings support the role of the Thy-Ade base pairing in promoting ssDNA loading. Furthermore, theoretical calculations using the self-consistent charge density functional tight-binding (SCC-DFTB) method verified the role of hydrogen bonding and van der Waals type interactions in this host–guest interface. KBM-1 and KBM-2 can protect ssDNA from enzymatic degradation and release it at acidic pH. Most importantly, these biocompatible frameworks can efficiently deliver genetic cargo with retained activity to the cell nucleus. We envisage that this class of Bio-MOFs can find immediate applicability as biomimics for sensing, stabilizing, and delivering genetic materials.

Stable nanotube construction conditions and electronic properties of possible Si double-walled nanotubes (nin,min)@(6,mout) (nin =3, 4) by SCC-DFTB calculations

The geometric structural optimization and electronic properties of double-walled silicon nanotubes (DWSiNTs) (nin,min)@(6,mout) (nin = 3, 4, min = 0 to nin, mout = 0 to 6) are studied in terms of the self-consistent charge density functional tight binding (SCC-DFTB) method. A suitable spacing range of initial model that can form a stable DWSiNT is explored. Most of the nanotubes present circular-like shapes of their cross-sectional configurations. When a DWSiNT has the same chiral angle, 0 or 30 in degree, in its inner and outer walls, a regular periodic structure develops. The bond length, inner and outer wall diameters, and wall spacing are also significantly affected by changes in the chiral index. The stability of the nanotube is affected by three factors: the configuration, the diameter, and the regularity of the atomic arrangements. The electrons in the inner wall are transferred to the outer wall, and the electron distribution of the atoms in the outer wall tube has the regional distribution of the interval between the electron-obtaining and the electron-losing regions. Among the seven tubes with fully commensurate inner and outer walls, DWSiNTs (3,2)@(6,4) exhibit semi-metallic characteristics; the others exhibit semiconductor-like properties with a narrow band gaps. The band gap of DWSiNTs is changed by tuning the chirality index, resulting in the transition between direct and indirect band gaps of nanotubes, which can be adjusted to match the performance needs of different devices, inducing a metal-semiconductor transition and direct-indirect band gap transition. These theoretical studies provide references for the preparation, and future application, of DWSiNTs.

A new method for determining energetically favorable landing sites of carboxyl groups during the functionalization of graphene nanomesh

In this paper, we propose a new method for the stepwise functionalization of graphene nanomesh (GNM) with carboxyl (COOH) groups. The key point of this method is the determination of landing sites for COOH groups. As a criterion for determining the most favorable arrangement of COOH groups, it is proposed to use the charge distribution over the GNM atoms. According to our idea, atoms with the largest negative charge will more easily form strong covalent bonds with functional groups. Testing of the proposed method is carried out on the example of GNM supercell with a circular hole 1.2 nm in diameter and 2.46 nm in the direction of zigzag edge and 2.55 nm in the direction of armchair edge. The self-consistent charge density functional tight-binding (SCC-DFTB) method is used to simulate the stepwise functionalization of GNM with a sequential increase in the number of COOH groups from 1 to 9. During landing, COOH groups are located at the GNM hole edges. The orbital charge distribution is analyzed according to the Mulliken. According to the binding energy calculations, the addition of COOH groups by selected GNM atoms is energetically favorable at each step of functionalization. In the course of functionalization, the energy gap of GNM practically does not change, and the Fermi level shifts downward by several tenths of electron volts. At the maximum saturation of the hole edge atoms with COOH groups, the Fermi level and the energy gap of the functionalized GNM take values ​close to the values ​of the non-functionalized GNM.

Study on the structures and electronic properties of double-walled silicon nanotubes (4,min)@(8,mout) under external electric field by SCC-DFTB calculations

The geometric structural optimization and electronic properties of double-walled silicon nanotubes (DWSiNTs) (4,min)@(8,mout) (min = 0 to 4, mout = 0 to 8) are studied in terms of the self-consistent charge density functional tight binding (SCC-DFTB) method. Calculations demonstrate that the regularity of the atomic arrangement, the diameter of the inner and outer walls, the degree of buckling, stability, energy gap, Fermi energy level, quantum molecular descriptors, and charge distribution strongly depend on the chiral index of the tube. In particular, the possibility of tuning the band structure by changing the chiral index has been demonstrated to help for the performance needs of different devices, inducing a metal–semiconductor transition and direct–indirect band gap transition. In addition, the influence of adding an external electric field on electronic properties of the zigzag DWSiNT (4,0)@(8,0) has been investigated. The tube applied electric field transforms from semiconductor to semi-metallic property. The tube becomes metallic upon the critical electric field strength of 0.7 V/nm and 1.0 V/nm. As increasing the external field strengths, the stability of the tube is reduced and the EF level is increased. The direction of charge transfer is always in the reverse of the field direction, which is different from the case of absent field.

Enhancing classical gold nanoparticle simulations with electronic corrections and machine learning

Classical simulations of materials and nanoparticles have the advantage of speed and scalability but at the cost of precision and electronic properties, while electronic structure simulations have the advantage of accuracy and transferability but are typically limited to small and simple systems due to the increased computational complexity. Machine learning can be used to bridge this gap by providing correction terms that deliver electronic structure results based on classical simulations, to retain the best of both worlds. In this study we train an artificial neural network (ANN) as a general ansatz to predict a correction of the total energy of arbitrary gold nanoparticles based on general (material agnostic) features, and a limited set of structures simulated with an embedded atom potential and the self-consistent charge density functional tight binding method. We find that an accurate model with an overall precision of 14 eV or 8.6% can be found using a diverse range of particles and a large number of manually generated features which were then reduced using automatic data-driven approach to reduce evaluation bias. We found the ANN reduces to a linear relationship if a suitable subset of important features are identified prior to training, and that the prediction can be improved by classifying the nanoparticles into kinetically limited and thermodynamically limited subsets based prior to training the ANN corrections. The results demonstrate the potential for machine learning to enhance classical molecular dynamics simulations without adding significant computational complexity, and provides methodology that could be used to predict other electronic properties which cannot be calculated solely using classical simulations.

Infrared Spectroscopy of Chemically Diverse Carbon Clusters: A Data-Driven Approach.

Carbon clusters exhibit a broad diversity of topologies and shapes, encompassing fullerene-like cages, graphene-like flakes, and more disordered pretzel-like and branched structures. Here, we examine computationally their infrared spectra in relation with these structures from a statistical perspective. Individual spectra for broad samples of isomers were determined by means of the self-consistent charge density functional-based tight-binding method, and an interpolation scheme is designed to reproduce the spectral features by regression on a much smaller subset of the sample. This interpolation proceeds by encoding the structures using appropriate descriptors and selecting them through principal component analysis, Gaussian regression or inverse distance weighting providing the nonlinear weighting functions. Metric learning is employed to reduce the global error on a preselected testing set. The interpolated spectra satisfactorily reproduce the specific spectral features and their dependence on the size and shape, enabling quantitative prediction away from the testing set. Finally, the classification of structures within the four proposed families is critically discussed through a statistical analysis of the sample based on iterative label spreading.