CuTCNQ for Enzyme-Free Glucose Oxidation Modeled with Deep Learning LSTM Networks

Conductive, metal-organic complex, specifically a copper 7,7,8,8-tetracyanoquinodimethane (CuTCNQ) structure, have emerged as a suitable catalyst for electrochemical oxidation reactions. Herein, CuTCNQ is explored as an electrocatalyst for directly oxidizing glucose molecules in alkaline media. The copper-centered organic complex offers a synergy of redox-chemistry (Cu (II/I)) and conductivity (TCNQ-), enabling amperometric non-enzymatic electroanalysis of glucose from 3.0 to 39.0 mM with a LOD of 0.15 μM(S/N = 3). The interaction of CuTCNQ with glucose is evaluated via DFT where a calculated binding energy of −0.21 Ha, alongside a reduced HOMO-LUMO energy gap of 0.873 eV confirms favorability of Cu-TCNQ-glucose complex, and enhanced electron transfer potential. Differential pulse voltammetry (DPV) based assessment confirms catalyst suitability for higher concentration range where adaptation of machine learning (ML) algorithm confirms Long short-term memory (LSTM) network superiority in modeling concentration dependencies and sequential glucose oxidation patterns. The LSTM’s relatively lower MSE (0.1430), MAE (0.0207), and RMSE (0.1439) compared to traditional ML models (Linear Regression, Random Forest, and LightGBM) confirm their effectiveness for validating electrocatalyst performance.