Early detection and targeted elimination of cancer cells remain priorities because many chemotherapeutic agents cause severe side effects. Luminescent transition metal complexes, and in particular Pt(II) pincer complexes, are attractive as biomolecular probes because their photophysical properties depend on the local chemical environment. That environmental sensitivity can enable differentiation of abnormal DNA environments and thus support cancer-related detection. However, computational protocols that reliably predict optical properties of transition metal intercalators are limited, motivating method benchmarking for accurate absorption spectra calculations to support candidate screening.
The benchmark focused on a Pt(II) pincer model complex studied both in isolation and intercalated into a short double-helical DNA fragment that represents probes designed to target DNA-associated abnormalities. Time-dependent density functional theory (TD-DFT) was used to calculate UV-Vis absorption spectra. The study compared different computational approximations, geometry optimization approaches, and exchange-correlation functionals to assess trade-offs between computational cost and spectral accuracy.
Key approximations and methods evaluated include the Tamm–Dancoff approximation (TDA) and the resolution of identity (RI) for TD-DFT, which are known to reduce computational cost. Geometry optimization approaches compared conventional DFT with the more approximate composite method PBEh-3c, and further acceleration via tight-binding methods (GFN1-xTB and GFN2-xTB). For TD-DFT calculations, different exchange-correlation functionals were tested, including PBE0 and range-separated variants. The effects of spin–orbit coupling (SOC) and basis set choices were also examined in the generation of UV-Vis spectra.
The authors report that geometry optimization is often the dominant computational expense. PBEh-3c emerges as an efficient alternative to conventional DFT geometry optimizations, introducing errors comparable in magnitude to those from the TDA approximation applied in TD-DFT. Tight-binding methods (GFN-xTB variants) provide further acceleration but result in larger structural deviations relative to higher-level DFT geometries. These larger deviations from GFN-xTB-optimized structures translate into more pronounced differences in calculated UV-Vis spectra compared with PBEh-3c or conventional DFT structures.
Applying the TDA and RI approximations to TD-DFT produces a substantial increase in computational speed while causing only a modest loss of accuracy in the calculated spectra for both the isolated and intercalated Pt(II) complex. The figures presented compare spectra calculated with and without TDA and the RI, using a consistent smearing (FWHM 0.3 eV) for visualization. Spin–orbit coupling (SOC) was also evaluated: inclusion of SOC affects spectral features, and comparisons were made for spectra computed with and without SOC. Basis set and relativistic treatment choices (for example, x2c-SVPall) were considered in the spectra reported. Overall, the TDA and RI approximations are recommended when computational efficiency is needed without large sacrifices in spectral fidelity.
The largest source of uncertainty identified in the study arises from the choice of exchange-correlation functional used in the TD-DFT calculation. Among tested functionals, the authors obtained good results using PBE0 applied to structures optimized with PBEh-3c, suggesting this combination provides a favorable balance between accuracy and cost for UV-Vis predictions of the Pt(II) probe. Range-separated functionals and other alternatives were compared in the figures, indicating functional choice can materially shift spectral predictions, and thus should be considered carefully depending on the precision required.
For screening workflows where many candidate structures must be evaluated, the study supports the following practical approach based on the benchmark:
These recommendations reflect a balance between computational efficiency and the need for reliable absorption spectra in the context of designing or screening luminescent Pt(II) probes targeted to DNA-associated cancer markers.
The authors declare no conflicts of interest.