Tuning the spin-component scaling scheme for inverted singlet–triplet (INVEST) systems: improving CC2 and ADC(2) from excited-state energies to topological features

Abstract Inverted singlet–triplet (INVEST) materials, characterized by a negative singlet–triplet energy gap (ΔE ST ), hold great promise for next-generation Organic Light Emitting Diode (OLED) applications. However, their accurate computational description remains challenging, as conventional low-cost methods often fail to capture the subtle electron correlation effects governing the gap inversion. In this work, we present a systematic fine-tuning of the spin-component scaling (SCS) parameters for the second-order coupled cluster (CC2) and Algebraic Diagrammatic Construction (ADC(2)) methods, targeting the accurate prediction of ΔE ST in INVEST systems. Using a set of eight triangular-shaped INVEST compounds, we screened opposite- and same-spin SCS parameters against Theoretical Best Estimate (TBE) reference values from Jacquemin and collaborators.[1] The optimal parameters (c os = 1.0, c ss = 1.4/1.6 for SCS-CC2/ADC(2)) differ markedly from the conventional ones, demonstrating that reparameterization is essential for a balanced description of the ΔE ST in this class of molecules. The spin-component decomposition of the correlation energy allowed deriving a general expression to estimate TBE gaps directly from CCS excitation energies, offering a computationally inexpensive diagnostic tool. Beyond parameter optimization, SCS tuning proved to be a valuable instrument to reveal how electron correlation couples to topology-related features such as electron density redistribution. We also outlined a strategy for extending the SCS tuning to extended INVEST systems. Taken together, these results establish SCS-CC2 and SCS-ADC(2) as cost-effective and physically transparent alternatives to higher-level methods for the study of INVEST photophysics.