{
  "abstract": "Light-harvesting efficiency hinges on how long an excitation stays coherent, yet in natural antennae the environment that destroys that coherence, the protein scaffold, cannot be redesigned at will. Here we invert the architecture and place the bath inside the chromophore: a C60 cage acts as the absorber while an encapsulated guest (CH4, N2, O2 or H2O) supplies a tunable internal environment. Combining B3LYP/6-31G* excitation energies, transition-electrostatic-potential (TrESP) couplings and hierarchical equations of motion (HEOM), we find that the choice of guest alone dictates the character of the dynamics. The guest acts through three parameters: the excitonic coupling to the neighbouring cage, the shift of the cage excitation energy and the reorganisation energy (λ). The damping of quantum beats grows in proportion to λ, coherent exchange gives way to incoherent hopping once λkBT/ℏγ becomes comparable to the coupling, and the guest-induced shifts of the excitation energy set the direction of net energy flow. Pairs containing the polar H2O guest, with λ above 40 cm−1, lie in the incoherent regime, whereas O2@C60, 5750 cm−1 below the other cages, is isolated from them. Two-dimensional electronic spectra of the all-system endohedral fullerenes show beats at the exciton gaps and spectral diffusion on the 100 fs time scale of the bath. We further demonstrate that this mapping can be inverted. A feed-forward neural network trained on a corpus of 3000 HEOM trajectories recovers a dimer’s site-energy gap, excitonic coupling and λ from a single 2 ps trace, with mean absolute errors of 3.45, 1.87 and 1.13 cm−1 on 600 unseen realisations. The coherence must accompany the population, because the sign of the coupling is reversed by a gauge transformation that leaves the populations unchanged; supplied with ρ12(t), the network recovers it in every held-out case. Internalising the bath thus renders coherence both tunable by design and readable from the dynamics.",
  "abstract_shown": true,
  "affiliations": [
    "Department of Chemistry",
    "SRM Institute of Science and Technology"
  ],
  "article_type": null,
  "arxiv": null,
  "author_total": 3,
  "authors": [
    {
      "affs": [
        0,
        1
      ],
      "name": "S Chandrasekaran",
      "orcid": "0000-0002-0898-4193"
    },
    {
      "affs": [
        0,
        1
      ],
      "name": "P Madhavan",
      "orcid": "0009-0009-4048-623X"
    },
    {
      "affs": [
        0,
        1
      ],
      "name": "R Pavithra",
      "orcid": "0009-0007-5746-7129"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/13hrgbafthk9b.bib",
  "cited_by": 0,
  "doi": "10.26434/chemrxiv.15010743/v1",
  "fulltext_url": null,
  "id": "13hrgbafthk9b",
  "issns": [],
  "landing_url": "https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010743/v1",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "3107|Spectroscopy and Quantum Chemical Studies"
    ]
  ],
  "openalex": "W7223081435",
  "pdf_url": "https://chemrxiv.org/doi/pdf/10.26434/chemrxiv.15010743/v1",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-20",
  "publisher": "American Chemical Society (ACS)",
  "ris": "https://research.draconitic.com/paper/13hrgbafthk9b.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:17:24Z",
      "native_id": "10.26434/chemrxiv.15010743",
      "source": "chemrxiv",
      "url": "https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010743/v1"
    }
  ],
  "subjects": [
    {
      "field": "physics",
      "origin": "openalex:3107|Spectroscopy and Quantum Chemical Studies",
      "subfield": "amo"
    }
  ],
  "title": "Quantum Coherence in Guest-Controlled Endohedral Fullerenes: Hierarchical Equations of Motion and a Neural-Network Model.",
  "title_text": "Quantum Coherence in Guest-Controlled Endohedral Fullerenes: Hierarchical Equations of Motion and a Neural-Network Model.",
  "type": "preprint",
  "updated_at": "2026-10-30T12:17:24Z",
  "url": "https://research.draconitic.com/paper/13hrgbafthk9b",
  "venue": "ChemRxiv",
  "versions": []
}