Vibrational strong coupling — where molecular vibrations and cavity photons hybridize into polaritons — is typically probed by nonlinear spectroscopy. Simulating these spectra requires two models: one to evolve the molecular dynamics and another to compute the spectroscopic signal from the resulting trajectories.
The paper on two-dimensional IR-Raman spectroscopy of vibrational polaritons (arXiv: 2603.24521) reveals that using different dipole surface models for these two stages — dynamics and post-processing — produces severe artifacts in the 2D spectra, even when the linear polariton spectrum is barely affected.
The linear spectrum is forgiving because it depends on first-order correlations. The 2D spectrum depends on higher-order correlations where inconsistencies between the interaction model and the observation model compound. Wide frequency regions of the 2D spectrum become distorted when the dipole surface used to compute the optical response doesn't match the one used to evolve the cavity molecular dynamics.
With a consistent dipole-induced-dipole model, the cavity 2D-IIR spectrum cleanly splits the OH stretch band into polariton branches along the IR axis (but not the Raman axis), while fading molecular signals elsewhere. The correct physics emerges only from self-consistent simulation.
The through-claim: consistency between the model that generates behavior and the model that observes it is not a refinement — it's a prerequisite. Linear observables tolerate inconsistency because they average over the discrepancies. Nonlinear observables amplify them. The error is invisible in the simple measurement and catastrophic in the detailed one.
2603.24521. Chemical physics / vibrational polaritons / 2D spectroscopy / cavity molecular dynamics / model consistency.