A mathematical investigation into the silicon carbide polytype network known as SiC4-I[r,s] has established analytical closed-form expressions for eleven Sombor index-based topological indices and their corresponding graph entropy measures, demonstrating how molecular branching patterns dictate structural information. Published in Discover Chemistry, the theoretical study conducted by Narahari Narasimha Swamy, Shrikanth Chikkenahalli Krishnamurthy, Sharathkumar Hunjanalu Thimmarayappa, Nagesh Hadonahally Mudalagiraiah, and Vijaya Chandra Kumar Uppari provides explicit predictive formulas that quantify network complexity and disorder across expanding crystalline dimensions. By translating the complex atomic layout of the material into rigorous graph-theoretical representations, the authors evaluated both vertex connectivity and edge weighting frameworks. Their findings show that while the studied network dimensions expand systematically, all derived topological descriptors and entropy metrics display monotonic growth, establishing reliable mathematical foundations that connect discrete molecular geometry with statistical measures of disorder.
Silicon carbide represents an exceptionally durable compound of silicon and carbon first discovered by Edward Acheson in 1891, which has earned critical prominence as an industrial ceramic and high-performance semiconductor. Possessing remarkable thermal stability that withstands temperatures reaching up to 1400 degrees Celsius without sacrificing mechanical integrity, the material routinely outperforms traditional silicon in modern power electronic systems. Silicon carbide components deliver superior blocking voltages, rapid switching speeds, elevated junction temperature tolerance, and outstanding resistance to corrosion, oxidation, and radiation degradation. Beyond high-power circuit architectures, diodes, and field-effect devices, the compound serves in diverse demanding applications including cutting tools, clutch assemblies, bulletproof armor, optoelectronic switches, satellite systems, and catalytic water-splitting mechanisms. The foundational driver of these physical virtues is its distinctive crystalline arrangement, which permits polytypism—a specific one-dimensional polymorphism where identical planar sheets of atoms stack in distinct recurring spatial sequences.
Although more than two hundred distinct polytypes of silicon carbide have been documented since Baumhauer first classified its morphological variants, examining each specific crystalline configuration requires robust computational and analytical methodologies. Within chemical graph theory, molecular systems are modeled as simple connected graphs in which atomic nuclei correspond to vertices and covalent chemical bonds correspond to edges. Topological indices translate this intricate web of physical bonds into single numerical descriptors reflecting atomic adjacency, degree distributions, and molecular branching. These descriptors serve as fundamental parameters in quantitative structure-property relationship and quantitative structure-activity relationship models, which allow researchers to estimate physical, chemical, and thermodynamic behaviors without relying exclusively on expensive experimental synthesis or intensive quantum mechanical calculations. While previous investigations evaluated certain silicon carbide networks using classic indices like the Zagreb, Nirmala, or atom-bond connectivity metrics, the SiC4-I[r,s] network had previously remained unexplored through Sombor-based frameworks.
The Sombor index, originally introduced by Ivan Gutman, defines edge weights through the geometric Euclidean norm of the degrees of adjacent endpoints, offering an intuitive metric for local structural variation across chemical graphs. In this study, the research team analyzed the SiC4-I[r,s] network characterized by a two-dimensional grid of unit cells arranged across r columns and s rows, where both structural parameters are positive integers. Within this configuration, each unit cell incorporates distinct arrangements of carbon and silicon atoms that yield specific vertex degrees ranging strictly from a minimum degree of one to a maximum degree of three. By classifying all vertices into three distinct degree sets and partitioning every connecting edge into five separate combinatorial pairings, the investigators mapped the exact incidence counts for every degree combination across the entire molecular boundary and internal lattice framework.
Utilizing the computer algebra system wxMaxima, the investigators formulated precise closed-form polynomial expressions as functions of the lattice coordinates r and s for eleven distinct topological indices. These descriptors encompassed the classic Sombor index, Banhatti Sombor index, Euler Sombor index, F-Sombor index, modified Sombor index, delta Sombor index, Revan Sombor index, reduced Sombor index, reverse Sombor index, modified delta Sombor index, and elliptic Sombor index. For instance, the general Sombor index reduced to 63.63959rs minus 13.86205s minus 22.90003r plus 6.49282, whereas the elliptic Sombor index produced a significantly larger expression of 381.837rs minus 108.9078s minus 169.2414r plus 56.7142. Across the full spectrum of evaluated networks, the elliptic Sombor index generated the highest numerical values due to its degree-sum multiplier, demonstrating superior sensitivity to progressive structural expansion, whereas the modified Sombor index generated the lowest numerical magnitude.
Complementing the topological indices, the research team evaluated Shannon entropy measures formulated over edge-weighted graphs using the paradigm developed by Chen and collaborators in 2014. By substituting individual topological edge contributions into the total network descriptor, graph entropy quantifies the degree of randomness, information diversity, and structural disorder inherent in the molecular distribution. For each of the eleven indices, explicit logarithmic expressions were computed to define the exact associated entropy for any arbitrary two-dimensional lattice size.
A notable theoretical outcome of the investigation arose from the inherent topological properties of the SiC4-I[r,s] network, wherein the minimum vertex degree is exactly one. Because the delta degree subtracts the minimum graph degree and adds one, the delta degree of every vertex becomes mathematically identical to its standard degree. Consequently, the calculations proved that the conventional Sombor index and the delta Sombor index produce identical closed-form values across all dimensions, and the modified Sombor index perfectly coincides with the modified delta Sombor index. Similarly, because the maximum degree is three and the minimum degree is one, the formulation for Revan degree yields numerical values that match the reverse vertex degree precisely. This alignment causes the Revan Sombor index and reverse Sombor index to assume entirely identical values and formulas, demonstrating an intriguing mathematical scenario where nominally distinct molecular descriptors yield completely identical numerical behavior for an entire material class.
To evaluate the functional interrelationships between these structural descriptors and their corresponding entropy metrics, the authors performed numerical covariance and regression analyses across a square lattice domain ranging from dimension r equals s equals one up to forty unit cells. The statistical evaluation deployed logarithmic, power, and polynomial regression models, treating the topological indices as independent variables and their associated entropy measures as dependent variables. Statistical validation was assessed through eight key parameters: the correlation coefficient, coefficient of determination, adjusted coefficient of determination, root mean squared error, sum of squared errors, Fisher F-value, significance F, and hypothesis test p-values.
The regression analyses revealed remarkably strong, statistically significant curvilinear correlations across every examined mathematical model, with the absolute correlation coefficient exceeding 0.96 for all combinations. In the logarithmic regression model, the relationship between the standard Sombor index and its entropy, as well as between the delta Sombor index and its entropy, achieved an extraordinary correlation coefficient of 0.99994. Under the power regression framework, the F-Sombor index and its associated entropy demonstrated the strongest alignment with a correlation coefficient of 0.97857. In the polynomial regression analysis, the Revan and reverse Sombor indices achieved the highest agreement with their respective entropy measures at 0.9890. Across all configurations, the overall regression F-values were notably high, the significance F approached zero, and the p-values remained strictly below 0.02, confirming that individual topological indices act as powerful predictive factors for network information entropy.
These derived analytical expressions provide material scientists and computational chemists with scalable formulas that bypass the need for repeated numerical simulations whenever the dimensions of an SiC4-I[r,s] crystallite change. Because quantitative structure-property relationship approaches evaluate molecular databases rapidly without demanding high-performance supercomputing clusters, establishing verified descriptors directly advances the theoretical screening of silicon carbide networks. However, the authors emphasize that graph-theoretical models carry inherent limitations, as two-dimensional topological representations cannot fully capture spatial conformations, bond angles, heteroatom polarization, or multiple bond orders.
Subject of Research: Topological index and entropy analysis of the SiC4-I[r,s] silicon carbide network
Article Title: Study of the silicon carbide network (SiC_4-I[r,s]) using Sombor index-based topological indices and their associated entropy measures
Article References: Swamy, N. N., Krishnamurthy, S. C., Thimmarayappa, S. H., Mudalagiraiah, N. H., & Uppari, V. C. K. (2026). Study of the silicon carbide network $$SiC_4-I[r,s]$$ using Sombor index-based topological indices and their associated entropy measures. Discover Chemistry, 3(1), Article 641. https://doi.org/10.1007/s44371-026-01082-w
Image Credits: AI Generated
DOI: 10.1007/s44371-026-01082-w
Keywords: silicon carbide, topological indices, Sombor index, graph entropy, polytypism, semiconductor physics, chemical graph theory, quantitative structure property relationship, material science, molecular descriptors, covariance analysis, regression modeling, discrete mathematics, crystal structures, structural complexity, Banhatti Sombor index, Euler Sombor index