Publication Deep Dive · 2026

Long-Lived Polaritonic Coherence and Polaron Decoupling Effects in 2D Electronic Spectra

M. E. Mondal, S. Park, M. Son, A. N. Vamivakas, S. T. Cundiff, T. D. Krauss, and P. Huo · PNAS 123, e2530454123

Paper ↗PDF ↗Code ↗All publications
Collective upper and lower polaritons, a four-peak 2D spectrum, and reduced molecular vibrational displacement

From the paper

Can collective coupling protect room-temperature polariton coherence?

Molecular vibrations and cavity photon leakage normally shorten electronic coherence. This work asks whether collective exciton-cavity coupling can protect coherence between upper and lower polaritons and whether 2DES can reveal both the longer coherence and the accompanying suppression of exciton-phonon coupling. CdSe nanoplatelets in a high-quality Fabry-Perot cavity are proposed as a favorable experimental platform.

From the paper

Bright, dark, and polariton states

\[|+\rangle=\frac{|B,0\rangle+|G,1\rangle}{\sqrt2},\qquad |-\rangle=\frac{-|B,0\rangle+|G,1\rangle}{\sqrt2},\qquad \hbar\Omega_R=2\sqrt N\hbar g_c.\]

The Holstein-Tavis-Cummings model contains $N$ two-level excitons, one cavity mode, and one independent harmonic bath per exciton. At resonance the bright-state bath displacement is suppressed, giving

\[\lambda_N=\frac{\lambda}{4N}.\]

That diagonal decoupling competes with phonon-assisted upper-polariton transfer into the $N-1$ dark states and with cavity leakage.

  1. Collective hybridizationOne bright exciton and one photon form LP and UP.
  2. Reduced displacementThe bright-state reorganization energy decreases as $1/(4N)$.
  3. Residual decayUP-to-dark transfer and cavity loss limit coherence.

From the paper

How the 2D spectrum measures both effects

The L-PLDM calculation evaluates rephasing and nonrephasing ground-state bleach, stimulated-emission, and excited-state-absorption pathways through the double-excitation manifold. The upper-lower cross-peak amplitude oscillates during population time,

\[I_{+-}(t_2)\propto\cos(\Omega_Rt_2)e^{-t_2/\tau}.\]

The decay of this oscillation diagnoses polaritonic coherence. Separately, the nodal-line slope through the lower-polariton diagonal feature tracks spectral diffusion: a flatter line indicates stronger polaron decoupling and weaker frequency memory from the bath.

Synthetic 2D map for explaining diagonal and cross peaks; not numerical data from the paper.

Simplified educational example

Estimate the collective scales for four molecules

Take $N=4$, $\hbar g_c=24.8$ meV, molecular $\lambda=6.2$ meV, and loss width $\Gamma=10$ meV. Then $\hbar\Omega_R=99.2$ meV, $\lambda_N=0.3875$ meV, the cavity lifetime scale is $\hbar/\Gamma\approx65.8$ fs, and the full beat period is about 41.7 fs.

Programmatically generated teaching curve with a 42 fs period and 100 fs envelope constant.
Python scale calculation
import numpy as np

hbar = 658.2119569  # meV fs
N, gc = 4, 24.8
lambda_mol, Gamma = 6.2, 10.0
Omega_E = 2*np.sqrt(N)*gc
lambda_pol = lambda_mol/(4*N)
period = 2*np.pi*hbar/Omega_E
print(Omega_E, lambda_pol, hbar/Gamma, period)

From the paper

Reported predictions

  • The fitted cross-peak decay parameter rises from about 74 to 100 to 122 fs as collective coupling increases in the reported series; the observable coherence is discussed as persisting to roughly 200 fs.
  • Protection is attributed to reduced bright-state phonon coupling and a phonon bottleneck as the UP-dark gap moves through the bath spectrum.
  • The lower-polariton nodal line becomes very flat at the strongest coupling, indicating suppressed spectral diffusion.
  • Rephasing pathways dominate the oscillatory cross peaks, while imperfect GSB/ESA cancellation permits weaker nonrephasing cross peaks.

Scope and source cautions

What should not be overinterpreted

  • The production calculations use $N=2,4,8$ and infer behavior of larger experimental ensembles at fixed collective coupling.
  • Static energetic disorder is omitted, although it may dominate organic polariton broadening.
  • L-PLDM and Markovian Lindblad loss are approximations, and the work reports theoretical predictions rather than measured spectra.
  • The supplied sources contain a parameter inconsistency: the main paper lists $\lambda=50$ cm$^{-1}$ and characteristic bath frequency 18 cm$^{-1}$, while the SI appears to interchange them. The values should be author-confirmed before a parameter-reproduction notebook is added.