betaFeedback

WorksPhysics and Astronomy

Exciton Delocalization Suppresses Polariton Scattering

Abstract

Exciton-polaritons (EPs) are part-light part-matter quasiparticles that combine large exciton-mediated nonlinearities with long-range coherence, ideal for energy harvesting and nonlinear optics. Optimizing EPs for these applications is predicated on a still-elusive understanding of how disorder affects their propagation and dephasing times. Here, using cutting-edge femtosecond spatiotemporal microscopy, we directly image EP propagation at light-like speeds in systems ranging from two-dimensional semiconductors to amorphous molecular films with systematically varied exciton-phonon coupling, exciton delocalization, and static disorder. Despite possessing similar EP dispersions, we observe dramatically different transport velocities and scattering times across systems. We establish a robust scaling law linking EP scattering to exciton transfer integral, demonstrating that polaritons based on materials with large exciton bandwidths are immune to disorder even for highly excitonic EPs. This observation cannot be deduced from the systems' linear optical properties, including EP dispersion and linewidth disorder. Our work highlights the critical and often-overlooked role of the matter component in dictating polariton properties, and provides precise guidelines for simultaneously optimizing EP propagation and nonlinearities.

Hong_etal_Polariton_04012025_rev

Introduction

When light strongly couples to electronic transitions in molecular and material systems, new part-matter part-light eigenstates known as exciton-polaritons (EPs) can form.[1, 2] EP formation is often facilitated by embedding excitonic systems in photonic cavities[3] that confine the light field and reduce photonic losses. EPs can inherit both the strong nonlinearity of excitons and the long-range coherence of light.[4, 5] This hybrid nature has been leveraged to realize phenomena of high technological interest, including long-range electronic energy transport[6, 7, 8, 9, 10, 11, 12, 13, 14, 15], polariton-assisted nonlinearities[16, 17, 18, 19, 20, 21], low-threshold and high-temperature Bose-Einstein condensates[22, 23], and single-photon blockades for quantum gates[24, 25, 26]. Optimizing these phenomena requires the simultaneous maximization of nonlinear interactions that occur at high excitonic content, and propagation or coherence lengths.

EP transport velocities are normally inferred from the gradient of the polariton dispersion. The picture typically invoked to describe EPs in the collective strong coupling regime is shown in Figure 1a, in which a cavity mode couples to NN excitons to form an upper polariton (UP), lower polariton (LP), and N−1N-1 dark states (DS).[27, 28] In this picture, the polariton dispersion (and velocity) is controlled by the cavity photon dispersion, Rabi splitting ΩR\Omega_{\mathrm{R}}, and cavity-exciton detuning; the exciton dispersion contributes negligibly to the cavity polariton dispersion due to the small exciton bandwidth within the light cone. Nevertheless, recent studies have shown that the actual propagation velocities of polaritons can be much lower than their dispersions suggest, and that EPs exhibit transitions from ballistic (coherent) to diffusive (incoherent) transport at high excitonic fractions.[11, 12, 29, 30, 31, 32, 33, 34, 35, 36] These results were assigned to disorder-induced polariton scattering or transient localization, either within the polariton manifold or into DS. These observations suggest a limited ability to simultaneously maintain long-range coherence and large nonlinearities that occur for highly excitonic EPs. A crucial question that remains unanswered is whether the properties of the underlying excitonic medium can be tuned to optimize this balance between propagation and nonlinear interactions that is central to EP applications.

Dispersion diagram, microscopy schematic, and EP transport dynamics for localized, intermediate, and delocalized excitons

Figure 1. Exciton delocalization protects EP transport. a. Dispersion diagram for polaritons under strong light-matter coupling. The Rabi splitting ΩR\Omega_{\mathrm{R}} is related to the exciton transition dipole moment μ\mu, number of dipoles NN and cavity mode volume VV. b. Schematic diagram for momentum-resolved ultrafast microscopy. c-e. EP transport dynamics imaged in a TIPS-pentacene microcavity (c, localized exciton), crystalline TIPS-pentacene microcavity (d, intermediate exciton), and crystalline 2D perovskite microcavity (e, delocalized exciton). The top panels display expected (grey filled circle) and measured (red circle) velocities for EPs with ∼50%\sim50\% excitonic fractions in each cavity. The measured velocities are extracted from ultrafast optical imaging, with representative time delays shown on the bottom panel. TIPS-Pc stands for tri-isopropylsilyl-ethynyl pentacene. All data symbols are presented as the mean +/- one standard deviation derived from the fitting error. All scale bars, 1 μm1\ \mu\mathrm{m}.

Here, by directly imaging EP propagation in systematically varied materials using state-of-the-art ultrafast microscopy, we show that electronic delocalization in the underlying excitonic medium suppresses EP scattering even for high exciton fractions. We reveal a remarkably robust empirical scaling law linking the exciton transfer integral with EP scattering over a broad range of materials and temperatures. These results are not captured in prevalent models of polariton transport. We establish that materials possessing a rare combination of large exciton binding energies, oscillator strengths, and exciton transfer integrals, such as rigid two-dimensional (2D) perovskites, are ideal for simultaneously optimizing nonlinear interactions and long-range coherence at room temperature.

Delocalized excitons support delocalized polaritons

To characterize EP transport in different materials as a function of exciton fraction and temperature, we leverage momentum-resolved ultrafast polariton imaging (MUPI, Figure 1b and Supplemental Note 1), a recently-developed spatiotemporally-resolved approach[11]. In MUPI, a diffraction-limited (∼200\sim200 nm diameter) pump pulse generates a local nonequilibrium population of excitons and polaritons, and a widefield highly collimated monochromatic probe pulse impinges the sample at finite momentum kk to probe EPs at a specific point in their dispersion. The resulting pump-probe signal, collected in a reflection geometry, images the spatial evolution of polaritons with momentum-specificity through a nonlinear polariton blockade effect. In our current configuration, MUPI affords ∼150\sim150 fs temporal resolution and <30< 30 nm spatial sensitivity to EP motion (Supplemental Note 1). This powerful combination of spatiotemporal, spectral and momentum specificity allows MUPI to be uniquely selective to EPs, avoiding complications from nonlinear signals associated with dielectric shift from exciton and nonequilibrium phonon populations[11].

Figure 1 c-Figure 1e illustrate the transport properties for EPs with ~50% exciton fraction extracted from MUPI data in three microcavities. The microcavities comprise excitonic materials with progressively more delocalized excitons: an amorphous film of TIPS-pentacene molecules (Figure 1c, localized excitons), a TIPS-pentacene molecular crystal (Figure 1d, intermediate excitons), and the layered halide perovskite semiconductor PEA2(CH3NH3)Pb2I7\mathrm{PEA}_{2}(\mathrm{CH}_{3}\mathrm{NH}_{3})\mathrm{Pb}_{2}\mathrm{I}_{7}, where PEA is phenylethylammonium (Figure 1e, delocalized excitons). We return to a detailed characterization of each system below. In all cases, we observe ultrafast EP flow over more than a micrometer within a picosecond. Nevertheless, Figure 1 c-Figure 1e summarize a key finding of our study: despite each microcavity possessing similar polariton dispersions, we observe dramatic and visually salient differences in EP transport dynamics. Notably, the discrepancy between expected EP velocity (extracted from the experimentally-measured dispersion), and measured EP velocity (extracted from MUPI) depends monotonically on exciton delocalization. EPs in materials with more delocalized excitons result in propagation velocities closer to the expected group velocities. As detailed below, this trend cannot be explained by changes in EP dispersion, ΩR\Omega_{\mathrm{R}}, cavity losses, or linewidth disorder. Instead, our results suggest that EP scattering and dephasing through interactions with phonons, disorder, or DS are suppressed in systems with delocalized excitons, despite the latter having no measurable impact on EP dispersion.

Polarons renormalize polariton velocities

Solution-processed halide perovskites have emerged as ideal systems to investigate polariton photophysics owing to their highly polarizable exciton transitions and strong nonlinearities.[19, 37] To controllably evaluate the impact of exciton delocalization on EP scattering, we compare two halide perovskites with equivalent linear optical properties but distinct polaronic (non-perturbative exciton-phonon) interactions. We focus on the prototypical two-dimensional (2D) halide perovskites A2(CH3NH3)Pb2I7\mathrm{A}_{2}(\mathrm{CH}_{3}\mathrm{NH}_{3})\mathrm{Pb}_{2}\mathrm{I}_{7} with A = PEA for a rigid lattice (PEAP) or butylammonium for a soft lattice (BAP), Figure 2a and Supplemental Note 2. PEAP and BAP possess near-identical exciton energies and linewidths (Figure 2b, right inset), but the change in lattice rigidity induced by the different organic spacers leads to significant differences in excited-state exciton-phonon interactions. Using photoluminescence measurements (Supplemental Note 3), we estimate exciton–optical (OP) phonon coupling strengths of ΓOP=25\Gamma_{\mathrm{OP}} = 25 meV for PEAP in agreement with prior literature, compared to ΓOP=∼60\Gamma_{\mathrm{OP}} = \sim60 meV for BAP.[38] To realize strongly-coupled EPs, we embed mechanically-exfoliated single-crystal 2D perovskites in multimode Fabry-Pérot microcavities (Supplemental Note 2), as in our prior work that focused on BAP microcavities.[11] Angle-resolved reflectance spectra of the PEAP cavity dispersion in Figure 2b display clear LP branches, agreeing closely with transfer matrix simulations (Figure 2b, right). Note that UP branches are not visible due to strong material absorbance above the semiconductor bandgap. Using a coupled oscillator model (dashed lines in Figure 2b, and Supplemental Note 4), we extract a Rabi splitting of ΩR=260\Omega_{\mathrm{R}} = 260 meV for PEAP, comparable to ΩR=275\Omega_{\mathrm{R}} = 275 meV for BAP, with similar cavity quality (Q) factors of ~100 (Table S1).

EP transport in a 2D perovskite cavity

Figure 2. EP transport in a 2D perovskite cavity. a. Structures for BAP and PEAP 2D perovskites. b. Angle-resolved reflectance (left) and transfer-matrix simulations (right) of polariton dispersion for a PEAP microcavity. Dashed lines are coupled oscillator model fits. Right inset displays the linear absorbance for BAP and PEAP outside the cavity. c. Transport of reservoir excitons imaged at k=0k = 0 µm−1^{-1}. d. MSD plot of exciton transport in PEAP and BAP microcavities. The result for BAP is from ref. 11. e. EP transport imaged at k=10.5k = 10.5 µm−1^{-1}, corresponding to EPs with 42% excitonic fraction. f. MSD plot of EP transport from panel (e), extracted using wavefront analysis (Supplemental Note 5). g. Expected group velocity from the gradient of the polariton dispersion (solid line) vs. measured transport velocity (symbols) for the different points in the EP dispersion labeled with circles in Figure 2b. h. EP velocity renormalization as a function of exciton content for BAP (yellow square) and PEAP (purple circle) microcavities. The results for BAP are from ref.11. All pump-probe data is collected using focused 2.41 eV pump excitation with an incident peak fluence of ~5 µJ/cm2^2. All data symbols are presented as the mean +/− one standard deviation derived from the fitting error. All scale bars, 1 µm.

Γ\Gamma (meV)ΔE\Delta E (meV)ΩR\Omega_{\mathrm{R}} (meV)Q factoraΩ/Γ\Omega/\Gamma
PEAP82.3186260843.16
BAP (filled)b99.7213 (166)2751262.76
BAP (partially-filled)99.7207190831.91
Crystalline TIPS-Pc128.2216200421.56
Amorphous TIPS-Pc100.9105160381.59

Table S1. Linewidth of exciton transition (Γ\Gamma), cavity-exciton detuning (ΔE\Delta E), Rabi splitting (ΩR\Omega_{\mathrm{R}}), and Q factor for various microcavities.

a. Quality factor is obtained from the angle-resolved reflectance of the polariton dispersion, and is estimated at a point corresponding to 40% exciton fraction.

b. The values are from ref.[1]. Different numbers correspond to microcavities with different detuning.

Figure 2c displays spatiotemporal images tracking the propagation of uncoupled reservoir excitons in PEAP microcavities (probed at k=0k = 0 µm−1^{-1} filled diamond in Figure 2b). From these images we extract the mean squared displacement MSD=σ2(t)−σ2(0)\mathrm{MSD} = \sigma^2(t) - \sigma^2(0), where σ2(t)\sigma^2(t) is the variance of the population spatial distribution at pump-probe delay tt extracted from a Gaussian fit to each image (Figure 2d, analysis details in Figure S6). For uncoupled excitons, we observe trap-limited transport characterized by an exponentially-decaying diffusivity, modeled as MSD=2λ2[1−exp⁡(−Dλ2t)]\mathrm{MSD} = 2\lambda^2 \left[1 - \exp\left(-\frac{D}{\lambda^2}t\right)\right] (where DD is diffusivity and λ\lambda is the average distance between trap sites).[39] We extract D=0.21D = 0.21 cm2^2/s and a 2D trap density (1/λ2)(1/\lambda^2) of 35.7 µm−2^{-2}. Figure 2d also plots the MSD for uncoupled excitons in BAP microcavities, which show similarly trap-limited transport, but with lower diffusivity D=0.10 cm2/sD = 0.10\ \mathrm{cm}^{2}/\mathrm{s} and trap density of 29 μm−2\mathrm{\mu m}^{-2}.[11] These values concur with literature values of exciton transport in BAP and PEAP crystals outside microcavities.[39, 40, 41] The larger exciton diffusivity in PEAP compared to BAP is attributed to lower exciton-phonon coupling and a more rigid lattice in the former.[39]

Exciton transport analysis plot

Figure S6. Exciton transport analysis. Bare exciton transport for PEAP with probe at ∼580 nm. The pump fluence for this dataset is 5.0 μJ/cm².

Figure 2e shows MUPI images of EP transport in the lower polariton branch for the PEAP microcavity probed at k=10.5 μm−1k = 10.5\ \mathrm{\mu m}^{-1} (yellow circle in Figure 2b), corresponding to an excitonic fraction of 42%. All MUPI data presented herein is collected using a non-resonant (above-gap) focused pump; we previously showed that EP velocities are equivalent under resonant and non-resonant excitation conditions[11]. The MSD for EP propagation (Figure 2f) is quadratic in time, indicating ballistic (scatter-free) transport with a group velocity of vmeas=24 μm/psv_{\mathrm{meas}} = 24\ \mathrm{\mu m}/\mathrm{ps}, or ~8% of light speed. These measurements confirm that EPs realized through cavity strong coupling enable orders-of-magnitude enhancement in electronic energy transport, as previously observed.[9, 11, 12, 29, 42, 43, 44]

In Figure 2g, we plot the measured EP velocity (symbols) as a function of exciton content by leveraging the momentum-selectivity of MUPI to probe different points in the polariton dispersion. The measured EP velocities are overlaid with the expected group velocity (Figure 2g, colored lines) extracted from the gradient of the experimentally-measured dispersion in Figure 2b. The dependence of EP transport on exciton content is in agreement with previous work on BAP[11] and organic microcavities[12] – we observe: (i) An increasingly large deviation between vmeasv_{\mathrm{meas}} and the expected group velocity (vgv_{\mathrm{g}}) as the exciton content increases due to polariton scattering; (ii) A ballistic-to-diffusive transition for EPs with exciton content higher than 50% (Figure S8).

Image not extracted

(a)

Image not extracted

(b)

Figure S8. Figure S8. Additional MSD transport results for PEAP microcavity. a-b. MSD plots of EP transport at 2.03 eV (a, 53% exciton content) and 1.91 eV (b, 30% exciton content), extracted using wavefront analysis (Supplemental Note 5). All data symbols are presented as the mean +/− one standard deviation derived from the fitting error.

The key comparison of EP velocities between PEAP and BAP cavities is shown in Figure 2h, which plots the velocity renormalization (defined as νrenorm.=100×(vmeas.−vg)/vg\nu_{\mathrm{renorm.}} = 100 \times(v_{\mathrm{meas.}} - v_{\mathrm{g}}) / v_{\mathrm{g}}), as a function of exciton content. Crucially, we observe a clear difference in the degree of renormalization between the two systems, with BAP experiencing ~2.8 fold greater renormalization compared to PEAP. We emphasize that the factors typically assumed to affect EP transport (Rabi splitting, cavity Q-factor, and linewidth of the exciton transition) are within experimental error of each other for the two materials (Table S1), indicating that these features are not sufficient to capture the different EP velocity renormalizations between PEAP and BAP cavities. We recover EP velocities that are renormalized by only ~30% for EPs with 87% exciton content in PEAP, indicating that velocity renormalization can be minimal even at room temperature and in standard Fabry-Perot microcavities with an appropriate material choice.

Our results indicate that EP transport depends heavily on underlying exciton-phonon interactions. The fact that these interactions are not captured in the linear polariton dispersion or optical spectra suggest that they are non-Condon excited-state interactions. Specifically, 2D halide perovskites are known to undergo exciton-polaron formation[45, 46, 47], wherein the lattice relaxes around the exciton to form an electronic quasiparticle dressed by a lattice deformation, resulting in an increase of the exciton effective mass and reduced exciton delocalization.

Although polaron formation is likely to occur in both BAP and PEAP crystals, the lattice relaxation is greater in the more flexible BAP (Supplemental Note 7).[39] We therefore assign the different EP transport properties for BAP and PEAP observed in Figure 2h to EP–lattice interactions mediated by polaron formation. These results confirm – using a well-controlled material platform – that the details of the underlying excitonic medium strongly affects EP properties, beyond the combination of dispersive photons and non-dispersive excitations captured in static optical properties.

Exciton-phonon and polariton-phonon interactions are different

Although results presented above indicate an intimate connection between exciton-lattice and EP-lattice interactions, the question of whether the same phonons are involved in both remains an open question. To answer this question, we perform temperature-dependent measurements of EP transport in PEAP cavities. First, we find that even highly excitonic EPs that display diffusive transport at room temperature (Figure 3a for 87% exciton content) recover ballistic transport at 3.5 K (Figure 3b) with no velocity renormalization. This result reinforces our assignment of velocity renormalization to EP-phonon interactions, and indicate that these interactions give rise to decoherence at elevated temperatures for highly excitonic EPs.

EP-phonon interactions in 2D perovskite microcavities

Figure 3. EP-phonon interactions in 2D perovskite microcavities. a, b. MSD plot for EPs with 87% exciton fraction in PEAP microcavity at 297 K (a) and 3.5 K (b). c. Temperature-dependent velocity renormalization for PEAP microcavity with 87% exciton content. The inset is the scattering rate plotted as a function of temperature. All data symbols are presented as the mean +/− one standard deviation derived from the fitting error.

The temperature-dependent renormalization plotted in Figure 3c provides further insight on EP-lattice interactions. To extract an effective EP-phonon scattering time, we assume a semiclassical picture 2D=vg2×τscattering2D = v_{\mathrm{g}}^{2} \times\tau_{\mathrm{scattering}} and extract DD assuming linear MSDs at each temperature. In this picture, the scattering rate normally obeys power law behavior (1/τscattering∝Tγ)(1/\tau_{\mathrm{scattering}} \propto T^{\gamma}), where the value of the exponent characterizes the dominant phonon contribution.[48, 49] For example, in 2D, acoustic phonon scattering is characterized by γ=1\gamma= 1.[49] Outside the microcavity, PEAP exhibits characteristic values of γ\gamma between 2.1 – 2.6 (Figure S9), indicating that scattering is primarily mediated by out-of-plane non-polar optical phonons (γ>1\gamma> 1), in line with prior results.[48, 50] In contrast, under strong light-matter coupling in microcavities, we observe γ=0.79±0.04\gamma= 0.79 \pm0.04 (inset in Figure 3c) over 30 - 300 K. This value of γ\gamma indicates a substantially reduced contribution from short-range optical phonons in favor of longer-range acoustic phonon scattering. This observation suggests that the delocalization of EPs results in lattice interactions that can differ from those of bare excitons, as was recently theoretically postulated.[36] This result reinforces a picture of EPs as hybrid quasiparticles that go beyond linear sums of their underlying components, and calls for models of EP transport that go beyond projecting known exciton-phonon interactions into the polariton basis.

Image not extracted

Figure S9. Temperature dependent time-resolved photoluminescence (TRPL) of PEAP. a. Temperature-dependent phonon interactions, described below.[21] b. Temperature-dependent TRPL at a pump wavelength of 515 nm and emission wavelength of 570 nm. c-d. Rate of rise-time extracted from panel (b) vs. temperature. The plots are fit with power law k=Tγk = T^{\gamma} where TT is temperature. γ=2.58\gamma= 2.58 for the range of 30-300 K (c) and γ=2.14\gamma= 2.14 for the range of 80-300 K (d). The rise time reflects primarily intraband relaxation processes to the band edge following pump excitation. All data symbols are presented as the mean +/- one standard deviation derived from the fitting error.

A largely overlooked parameter that can affect EP transport is the extent of exciton delocalization, as shown in Figure 1. Here, we show that the exciton transfer integral JJ is likely the best predictor for the degree of EP velocity renormalization experienced by all systems investigated in this work. Identifying systems with radically different JJ to compare EP transport is challenging because other properties such as linewidth inevitably change. We identify tri-isopropylsilyl-ethynyl pentacene (TIPS-Pc) as a promising candidate because: (i) We can prepare microcavities from large single-crystal domains with J≈43.9J \approx43.9 meV (Figure 4a and Figure S12, see Supplemental Note 7 for estimation of JJ), and positionally-disordered monomeric molecules in a polymer matrix (J≈0J \approx0 meV, Figure 4b) with comparable Rabi splitting. (ii) The slight increase in exciton linewidth in the single-crystal microcavities (∼\sim128.2 meV compared to ∼\sim100.9 meV in the monomer film) allows us to rule out that larger EP renormalization in the low-JJ sample is due to larger linewidth Γ\Gamma, or a lower ratio ΩR/Γ\Omega_{\mathrm{R}}/\Gamma, which is currently understood as a good predictor of EP delocalization.[51, 52, 53]

EP propagation in organic microcavities

Figure 4. EP propagation in organic microcavities. a, b. Schematic structure (left) and angle-resolved reflectance spectrum showing the polariton dispersion (right) for a crystalline TIPS-Pc microcavity (a) and amorphous TIPS-Pc film (b). Right insets display the linear absorbance for each sample outside the cavity. c. EP velocity renormalization as a function of exciton content in the crystalline TIPS-Pc microcavity (blue diamond). The filled diamond indicates the measured EP velocity at 5 K. d. EP velocity renormalization as a function of exciton content in the amorphous TIPS-Pc microcavity (red pentagon). The filled pentagon indicates the measured EP velocity at 5 K. All pump-probe data is collected using focused 1.97 eV pump excitation with an incident peak fluence of ∼2\sim2 µJ/cm2^{2}. All data symbols are presented as the mean +/− one standard deviation derived from the fitting error.

Image not extracted

Figure S12. Optical image of crystalline TIPS-Pc domains. Scale bar, 50 μ\mum.

As shown in Figure 4a,Figure 4b, distinct upper (UP), middle (MP), and lower (LP) polaritons are observed in both single-crystal and amorphous molecular films. From coupled oscillator model fits, we extract ΩR=200\Omega_{R} = 200 meV and 160 meV for TIPS-Pc crystalline and amorphous cavities, respectively. Although the cavity–exciton detunings for the two cavities investigated here are different, we have previously demonstrated that the effect of cavity detuning on EP renormalization for a given excitonic fraction is negligible[11]; note that the exciton fraction determines the actual detuning between the EP being probed and the exciton reservoir regardless of cavity detuning. Importantly, both cavities show similar ratios ΩR/Γ=1.59\Omega_{R}/\Gamma= 1.59 for the crystalline system and 1.56 for the amorphous system.

Despite these similarities, we observe drastically different EP transport for crystalline vs amorphous samples both at room temperature and at 5 K. The EP velocity renormalization in the TIPS-Pc crystal cavity follows a similar trend to the perovskite cavities, but with a 3.2-fold steeper dependence on exciton content compared to PEAP, reaching over 50% renormalization for 49% exciton content (Figure 4c and Figure S14). In contrast, the TIPS-Pc amorphous cavity exhibits a large, relatively uniform velocity renormalization of 60-70% across exciton contents spanning 22–71% (Figure 4d and Figure S15). At 5 K, further differences emerge: the crystalline cavity recovers ballistic transport with a non-zero but small velocity renormalization (Figure 4c and Figure S16), while the amorphous cavity maintains diffusive transport with renormalization comparable to room temperature data (Figure 4d and Figure S17). The crossover from diffusive to ballistic transport as the temperature lowers indicates that dynamic disorder dominates EP scattering in the crystalline system, similarly to the perovskite cavities. In contrast, the large velocity renormalization and persistent diffusive transport at 5 K in the amorphous cavity signal the inability of EPs to overcome scattering from static disorder. We emphasize that changes in the linear absorption properties of amorphous and crystalline samples as a function of temperature are unable to explain these striking differences (Figure S18).

Image not extracted

Figure S14. MUPI results for crystalline TIPS-Pc microcavity at room temperature. a. Schematic structure for crystalline TIPS-Pc microcavity. b. EP transport imaged at k=3.0 μm−1k = 3.0\ \mu\mathrm{m}^{-1}, corresponding to EP with 15% excitonic content. c-e. MSD plots of EP transports at 1.65 eV (c, 49% exciton content), 1.59 eV (d, 22% exciton content) and 1.53 eV (e, 15% exciton content), extracted using wavefront analysis (Supplemental Note 5). f. Calculated group velocity as a function of detuning energy (E – Eexc_{\mathrm{exc}}). The measured EP velocities are overlaid with the expected group velocity. The measured group velocities are probed at the points labeled with circle in Figure 3a in main text. All data symbols are presented as the mean +/− one standard deviation derived from the fitting error. All scale bars, 1 μ\mum.

Image not extracted

Figure S15. MUPI results for amorphous TIPS-Pc microcavity at room temperature. a. Schematic structure for amorphous TIPS-Pc microcavity. b. EP transport imaged at k=4.8 μm−1k = 4.8\ \mu\mathrm{m}^{-1}, corresponding to EPs with 51% excitonic content. c-e. MSD plots of EP transport at 1.85 eV (c, 71% exciton content), 1.82 eV (d, 51% exciton content) and 1.75 eV (e, 22% exciton content), extracted using wavefront analysis (Supplemental Note 5). f. Calculated group velocity as a function of detuning energy (E – Eexc_{\mathrm{exc}}). The measured EP velocities are overlaid with the expected group velocity. The measured group velocities are probed at the points labeled with circle in Figure 3b in the main text. All data symbols are presented as the mean +/− one standard deviation derived from the fitting error. All scale bars, 1 μ\mum.

Image not extracted

Figure S16. Cryogenic temperature results for crystalline TIPS-Pc microcavity. a. Angle-resolved reflectance of polariton dispersion for a crystalline TIPS-Pc microcavity at 5 K. Dashed lines are coupled oscillator model fits. Middle polariton features are beyond the NA of the in-cryostat objective. b. Calculated group velocity as a function of detuning energy (E−EexcE - E_{\mathrm{exc}}). The measured EP velocities are overlaid with the expected group velocity. The measured group velocities are probed at the point labeled with a circle in Figure S16. c. MSD of EP transport probed at 1.65 eV, corresponding to EP with 58% exciton content, extracted using wavefront analysis (Supplemental Note 5). All data symbols are presented as the mean +/− one standard deviation derived from the fitting error.

Image not extracted

Figure S17. Cryogenic temperature results for amorphous TIPS-Pc microcavity. a. Angle-resolved reflectance of polariton dispersion for amorphous TIPS-Pc microcavity at 5 K. Dashed lines are coupled oscillator model fits. Middle polariton features are beyond the NA of the in-cryostat objective. b. Calculated group velocity as a function of detuning energy (E−EexcE - E_{\mathrm{exc}}). The measured EP velocities are overlaid with the expected group velocity. The measured group velocities are probed at the point labeled with the circle in Figure S17. c. MSD plots of EP transport probed at 1.84 eV, corresponding to EP with 64% exciton content, extracted using wavefront analysis (Supplemental Note 5). All data symbols are presented as the mean +/− one standard deviation derived from the fitting error.

Image not extracted

Figure S18. Temperature dependent absorption spectra. a. Absorbance of crystalline TIPS-Pc at 5 K (dashed line) and RT (solid line). b. Absorbance of amorphous TIPS-Pc at 5K (dashed line) and RT (solid line).

Overall, our results indicate that polariton dispersions are insufficient to predict EP transport properties in a large range of systems due to EP scattering, transient localization, and decoherence on much shorter scales than the polariton lifetime for highly excitonic EPs. Nevertheless, a clear trend emerges from our measurements: higher electronic coupling in the underlying excitonic medium leads to better EP transport, which we define as lower EP renormalization for a given exciton content. To capture this trend quantitatively, Figure 5 plots the velocity renormalization for EPs with 40% exciton character as a function of JJ, for all measured systems at room temperature (open symbols) and low temperature (solid symbols). At low temperature, JJ is larger in TIPS-Pc, BAP and PEAP crystals, which all experience renormalized masses at room temperature due to polaron formation (Supplemental Note 7). Figure 5 shows that the relationship between EP renormalization and JJ is maintained over all data collected in this work, cementing our findings as a robust empirical scaling law. We find much poorer correlations between EP renormalization and other common parameters such as Q-factor, ΩR\Omega_{\mathrm{R}} or Γ\Gamma, including when comparing BAP microcavities possessing different ΩR\Omega_{\mathrm{R}} tuned through cavity geometry (Supplemental Note 16).[54] Therefore, the key finding from this plot is that EP scattering for highly excitonic EPs can be dramatically suppressed in systems that possess delocalized excitons. Remarkably, within the systems explored here, the strongest predictor for whether EP transport reaches its expected group velocity is the underlying excitonic band structure.

Velocity renormalization as a function of exciton transfer integral

Figure 5. Figure 5. EP scattering depends on excitonic intersite coupling. Summary of velocity renormalization for EPs with 40% exciton fraction plotted as a function of intersite exciton transfer integral for PEAP (purple circles), BAP (yellow squares), crystalline TIPS-Pc (blue diamonds), and amorphous TIPS-Pc (red pentagons) microcavities. Open symbols indicate room temperature data; filled symbols indicate low-temperature data. Estimations of JJ for different temperatures are discussed in Supplemental Note 7. The shaded region indicates the 95% confidence band of a linear fit to the data. The plot displays a clear correlation between intersite excitonic coupling and EP transport.

One way to rationalize the observed scaling between EP renormalization and JJ is to invoke a picture where phonons and/or static disorder mediate EP scattering into relatively localized states such as DS, including through a recently-proposed superexchange mechanism that does not require transferring population to DS[55]. In this picture, large JJ implies more delocalized dipoles, which in turn reduces the density of DS under collective light-matter coupling and associated EP-DS scattering. Similarly, a recently-developed theoretical framework suggests that large intersite exciton coupling in 2D molecular aggregates protects polaritons against disorder through persistent DS delocalization.[56] Taken together, our observations indicate that delocalized excitons in the underlying medium allows highly excitonic EPs to maintain long-range coherence and transport. Achieving sufficient oscillator strengths in delocalized (non-Frenkel) excitonic systems to allow strong coupling at room temperature can be challenging. Layered semiconductors such as 2D halide perovskites are ideal in this respect, as they combine macroscopic crystallinity with large oscillator strengths and appreciable exciton delocalization, with the latter optimized in structures that are more rigid.

Through systematic imaging of EP transport in a variety of microcavity polaritons, we have elucidated to what extent the details of the underlying excitonic medium project into the properties of hybridized EPs. Our most important finding is that exciton-lattice interactions and excitonic intersite coupling are crucial predictors of EP transport, even when these features have no influence on the linear optical and dispersion properties of EPs. Most importantly, we reveal a robust empirical scaling law linking EP velocity renormalization to exciton delocalization. This relationship spans from disordered organic molecular films to highly crystalline 2D semiconductors over a large temperature range, and should extend to new polaritonic material platforms[57, 58]. Our results underscore the importance of the underlying matter in dictating EP properties, and provides guidelines for designing excitonic systems that maximize both nonlinear interactions and long-range coherence, a key demand of future polaritonic devices.

Resource availability

Lead contact

Further information and requests for resources should be directed to the lead contact, Milan Delor ([email protected]).

Data availability

All raw data are displayed in Figure 1–Figure 5 of the main text and Figure S1–Figure S20 of the Supplemental Information. Raw image files are available from the contact author upon request.

MUPI setup schematic

Figure S1. MUPI setup schematic.

MUPI results for partially filled BAP microcavity

Figure S20. Figure S20. MUPI results for partially filled BAP microcavity. a. Schematic structure for partially filled BAP microcavity[9]. b. Angle-resolved reflectance of polariton dispersion for a partially filled BAP microcavity. A coupled oscillator fit (Supplemental Note 4) yields ΩR=190\Omega_{\mathrm{R}}=190 meV. c. EP transport imaged at 1.95 eV, corresponding to EPs with 25% excitonic fraction. d-e. MSD plots of EP transport probed at 2.05 eV (d, 44.5% exciton content) and 1.95 eV (e, 17.3% exciton content), extracted using wavefront analysis (Supplemental Note 5). f. Calculated group velocity as a function of detuning energy (E−EexcE-E_{\mathrm{exc}}). The measured EP velocities are overlaid with the expected group velocity. The measured group velocities are probed at different points labeled with circle in Figure S20. g. Velocity renormalization plots overlapping BAP (square for fully filled and partially filled), PEAP (circle). All data symbols are presented as the mean +/− one standard deviation derived from the fitting error. All scale bars, 1 μ\mum.

Acknowledgements

This material is based upon work primarily supported by the National Science Foundation under Grant Number CHE-2203844, and by the Arnold and Mabel Beckman Foundation through a Beckman Young Investigator award. Low-temperature measurements were supported by the National Science Foundation under Grant Number DMR-2115625. Y.H. acknowledges support from the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2023-00240362).

Author contributions

Y.H. and M.D. conceived and designed the experiments. Y.H. and D.X. prepared samples. Y.H. acquired the experimental data. Y.H. and D.X. analysed the experimental results and performed the numerical simulations. M.D. supervised the project. Y.H. and M.D. wrote the paper with input from all authors.

Declaration of interests

The authors declare no competing interests.

Hong_etal_Polariton_SI_revised-refs

Note 1. Ultrafast momentum-resolved polariton imaging (MUPI)

Exciton polariton imaging of various Fabry-Pérot microcavities is performed using ultrafast momentum-resolved polariton imaging (MUPI). The systems used here are described in detail elsewhere[1, 2] and illustrated schematically in Figure S1.

Briefly, a Yb:KGW ultrafast regenerative amplifier (Light Conversion Carbide, 40 W, 1030 nm fundamental, 1 MHz repetition rate) was used as the light source for spatiotemporal measurements. The pump pulses are generated by frequency-doubling either the fundamental 1030 nm output, or the idler output of an optical parametric amplifier (OPA, Light Conversion, Orpheus-F). A white light continuum (WLC) as a probe pulse is generated in a YAG window (EKSMA). The pump and probe beams are directed to the high numerical-aperture oil-immersion objective (Leica HC Plan Apo 63x, 1.4 NA oil immersion). The pump beam is collimated before the objective lens and focuses on the sample plane. The probe beam focuses on the back focal plane of the objective by an f=300f = 300 mm widefield lens, enabling widefield illumination of the sample. A tilting mirror placed one focal length upstream of the widefield lens allows tuning the angle at which the widefield probe illuminates the sample, thus allowing probing at any momentum up to a maximum of k/k0=1.4k/k_0 = 1.4, limited by the numerical aperture of the objective. The reflected probe light at the sample interface with the back-scattered light from the sample forms images on a CMOS camera (Blackfly S USB3, BFS-U3-28S5M-C). For cryogenic measurements, the same light source is used, but the sample is placed in a closed loop Montana Instruments s100 cryostation equipped with a cryo-optic objective (Zeiss LD EC Epiplan-Neofluar 100x/0.90 DIC M27, NA = 0.9). For angle-resolved reflectance, the back focal plane of the objective is projected on the entrance slit of a home-built prism spectrometer using a pair of lenses (f1=300f_1 = 300 mm and f2=100f_2 = 100 mm), as depicted in Figure S4. For real-space MUPI imaging, an additional 150 mm lens forms an image on a CMOS camera (Blackfly S USB3, BFS-U3-28S5M-C). Both the spectrometer camera and the real-space camera are triggered at double the pump modulation rate, allowing the consecutive acquisition of images with the pump ON followed by the pump OFF. Consecutive frames are then processed according to (pump on/pump off − 1). The pump-probe temporal cross-correlation at the sample is approximately 166 fs (Figure S2).

Normalized pump-induced reflection profile

Figure S2. Normalized pump-induced reflection profile using 515 pump and 620 nm probe. The estimated instrument response function is 166 fs.

Schematic illustrations of various cavities and their corresponding light–matter matrices

Figure S4. Schematic illustrations of various cavities (a-c) and their corresponding light–matter matrices (d–f) that can be diagonalized to obtain multimode polariton dispersion. The dotted box corresponds to the materials with various thicknesses and the solid box indicates the layer of PMMA polymer.

Here, we note that the spatial precision for both exciton and polariton transport measurements is not defined by the diffraction limit, but rather by the signal-to-noise ratio (SNR) of the measurement, as is well described in the field of spatiotemporal imaging.[3, 4] Our measurements typically yield sub-30 nm spatial precision.

Sample preparation

(C6H5C2H7N)2(CH3NH3)Pb2I7(\mathrm{C}_{6}\mathrm{H}_{5}\mathrm{C}_{2}\mathrm{H}_{7}\mathrm{N})_{2}(\mathrm{CH}_{3}\mathrm{NH}_{3})\mathrm{Pb}_{2}\mathrm{I}_{7} (PEA2_{2}(MA)Pb2_{2}I7_{7}) crystal synthesis. The synthesis of large-area single-crystal flakes followed a published procedure.[1, 5] All chemicals were purchased from Sigma-Aldrich. Hydriodic acid (HI) solution was prepared by mixing 57% wt. aqueous HI (9 mL) and 50% wt. aqueous Hypophosphorous acid (H3PO2\mathrm{H}_{3}\mathrm{PO}_{2}, 1 mL). Lead (II) iodide (PbI2\mathrm{PbI}_{2}, 99.999% trace metals bases) powder (2720 mg, 5.9 mmol) was dissolved in the solution mixture under constant magnetic stirring. Methylammonium iodide (MAI, ≥99%\ge99\%, anhydrous 493 mg, 3.1 mmol) was added to the solution, and black powder precipitated instantly. The black powder redissolved quickly by heating the solution to 100 °C with stirring. After subsequent addition of phenylethylammonium iodide (PEAI, 864 mg, 4.3 mmol), the solution was heated to 105 °C under constant magnetic stirring until all precipitate dissolved. The solution was then subjected to a controlled cooling rate of 0.5 °C/h to room temperature in an oil bath, and large crystals formed on the solution surface. The crystals were collected by vacuum filtering and washed twice with toluene.

Metallic cavity fabrication. The bottom partial reflector through which light impinges the sample is a 30 nm gold film that was deposited on cover glass (Richard-Allan Scientific, 24×50 #1.5) by e-beam evaporation (Angstrom EvoVac deposition system). The deposition rate was 0.05 nm/s. The perovskite flakes were then mechanically exfoliated onto the gold-deposited cover glass with PVC tape (Nitto SPV224 PVC Vinyl Surface Protection Specialty Tape). Although gold is more lossy than silver, thin (~50nm) silver films tend to deteriorate during measurements, which rapidly leads to poor imaging quality, an important aspect for our scattering based experiments. The top reflector is a 200 nm silver film that was prepared by e-beam deposition on a silicon wafer. A layer of polyvinylpyrrolidone (PVP) solution (Sigma Aldrich, M.W. 40000, 10% wt in ethanol/acetonitrile wt 1/1) was then spin-coated on the silver film (3000 rpm, acceleration 1000rpm/s, 2min) and thermally annealed at 150 °C for 5 min. The prepared PVP/Ag was picked up with thermal release tape (semiconductor corp., release temperature 90 °C) and placed on the perovskites flakes with firm pressure, completing the cavity structure. The full structures were encapsulated between glass slides using epoxy (OG159-2, Epoxy Technology) in a nitrogen-filled glovebox to prevent sample oxidation during the measurements.

Amorphous/Crystalline TIPS-pentacene cavities. For amorphous films, TIPS-Pentacene (≥99%\ge99\%, 20 mg) and polystyrene (PS, M.W. 280000, 60 mg) were dissolved in toluene (1 mL) under constant stirring at 60°C until the TIPS-pentacene/PS is fully dissolved. The TIPS-pentacene/PS solution is then spin-coated on the gold coated cover glass at 3000 rpm for 120 seconds. In contrast, for a crystalline film, TIPS-Pentacene (5 mg) was dissolved in toluene under constant stirring at 60°C. 50 ul of TIPS-pentacene solution was then drop-cast onto the gold coated cover glass. TIPS-pentacene crystals with large domain sizes form within 20 minutes of drop-casting. For both amorphous and crystalline systems, we then deposit a 200 nm gold film by e-beam evaporation on the samples. The deposition rate was 0.1 nm/s.

Temperature dependent photoluminescence linewidths of PEAP

We perform temperature-dependent photoluminescence (PL) experiments on PEAP single crystals (Figure S3). We analyze the linewidth evolution as a function of temperature to estimate the exciton-phonon coupling strength of PEAP crystals (Figure S3). Exciton-phonon interactions are captured in the Debye-Einstein approximation[6, 7]:

Temperature dependent PL results for PEAP

Figure S3. Temperature dependent PL results for PEAP. a. Contour map of normalized PL as a function of temperature. b. Full-width half maximum (FWHM) of PL spectrum as a function of temperature. The solid line represents the Debye-Einstein fit to extract contributions from phonon-induced broadening.

Γ=Γ0+γAT+ΓOPexp⁡(E0kBT)−1(1)\Gamma= \Gamma_{0} + \gamma_{A}T + \frac{\Gamma_{\mathrm{OP}}}{\exp\left(\frac{E_{0}}{k_{B}T}\right)-1} \tag*{(1)}

where Γ0\Gamma_{0} is PL line width at 0 K, γAT\gamma_{A}T is the broadening induced by acoustic phonons, and ΓOP\Gamma_{\mathrm{OP}} is the interaction with optical phonons. By fitting the data in Figure S3, we extract coupling strengths for acoustic and optical phonons of 87 μ\mueV/K and 24.6 meV, respectively, in line with previous results[6, 8].

Analysis of polariton dispersions

Here, we detail the coupled oscillator models used to model cavity dispersion for various geometries[9, 10, 11, 12] (Figure S4).

Fully-filled multi-mode perovskite cavities in the strong-coupling regime (Figure S4) are best fit using a coupled oscillator model described by a 2N2N-dimension block-diagonal Hamiltonian for NN cavity modes. For the three modes observed in the dispersion in our case, we model the dispersion using the following coupled oscillator model:

[Uexδ0000δH1000000Uexδ0000δH2000000Uexδ0000δH3][χ1φ1χ2φ2χ3φ3]=Epol[χ1φ1χ2φ2χ3φ3](2)\begin{bmatrix} U_{\mathrm{ex}} & \delta& 0 & 0 & 0 & 0 \\ \delta& H_1 & 0 & 0 & 0 & 0 \\ 0 & 0 & U_{\mathrm{ex}} & \delta& 0 & 0 \\ 0 & 0 & \delta& H_2 & 0 & 0 \\ 0 & 0 & 0 & 0 & U_{\mathrm{ex}} & \delta\\ 0 & 0 & 0 & 0 & \delta& H_3 \end{bmatrix} \begin{bmatrix} \chi_1\\ \varphi_1\\ \chi_2\\ \varphi_2\\ \chi_3\\ \varphi_3 \end{bmatrix} = E_{\mathrm{pol}} \begin{bmatrix} \chi_1\\ \varphi_1\\ \chi_2\\ \varphi_2\\ \chi_3\\ \varphi_3 \tag*{(2)} \end{bmatrix}

where UexU_{\mathrm{ex}} is the exciton energy (2.14 eV), HnH_n is the energy for each cavity mode, δ\delta is the exciton-photon coupling (ΩR/2\Omega_{\mathrm{R}}/2) between each cavity mode and the exciton, and EpolE_{\mathrm{pol}} is the energy of polariton. Figure 2b (dashed lines) and Figure S5 display the result of such a coupled oscillator fit for PEAP cavities. The fit provides a value of δ=130\delta= 130 meV, i.e. a Rabi splitting of 260 meV.

Angle-resolved reflectance, SMM calculation, and coupled oscillator model fits

Figure S5. a. Angle-resolved reflectance of dispersion spectrum for PEAP microcavity. b. SMM calculation. c. Coupled oscillator model fits.

For TIPS-Pc single-mode cavities (Figure S4), in which the excitonic absorption shows distinct vibronic features, the lowest two vibronic bands both couple with the cavity modes. These cavities are described using:

(Uex,10δ10Uex,2δ2δ1δ2H1)(3)\begin{pmatrix} U_{\mathrm{ex},1} & 0 & \delta_1 \\ 0 & U_{\mathrm{ex},2} & \delta_2 \\ \delta_1 & \delta_2 & H_1 \tag*{(3)} \end{pmatrix}

where Uex1U_{\mathrm{ex}1} and 2 correspond to the lowest two vibronic energies (1.78 and 1.9 eV for crystalline TIPS-Pc and 1.9 and 2.05 eV for amorphous TIPS-Pc). The fits (Figure 4a and Figure 4b) provide values of δ1=80\delta_1 = 80 (100) meV for amorphous (crystalline) TIPS-Pc, i.e. a Rabi splitting of 160 and 200 meV, respectively.

Finally, for the partially-filled multimode BAP cavity (PMMA/perovskite/PMMA, Figure S4), it is necessary to distinguish the light-matter coupling with even- and odd- cavity modes. This type of cavity is best modeled using an (N+2)(N+2)(N+2)(N+2)-dimension Hamiltonian for NN cavity modes:

(Uex,1δ1δ3000δ1H10000δ30H3000000Uex,1δ2δ4000δ2H20000δ40H4)(4)\begin{pmatrix} U_{\mathrm{ex},1} & \delta_1 & \delta_3 & 0 & 0 & 0 \\ \delta_1 & H_1 & 0 & 0 & 0 & 0 \\ \delta_3 & 0 & H_3 & 0 & 0 & 0 \\ 0 & 0 & 0 & U_{\mathrm{ex},1} & \delta_2 & \delta_4 \\ 0 & 0 & 0 & \delta_2 & H_2 & 0 \\ 0 & 0 & 0 & \delta_4 & 0 & H_4 \tag*{(4)} \end{pmatrix}

The fit (Figure S20) provides a value of δ=90\delta= 90 meV, i.e. a Rabi splitting of 180 meV.

Figure S3 shows the result of a scattering matrix method (SMM) calculation, performed using the open-source S4 package,[13] using the perovskite slab thickness and the known dielectric functions of the metallic mirrors and the perovskite. The experimental dispersion, coupled oscillator models (Figure S5) and SMM calculations are all in excellent agreement with one another. Note that upper polariton branches are not visible in the experimental dispersion (Figure S5) because the material is highly absorptive above bandgap, as also captured in the SMM calculation (Figure S5).

The exciton content of the polariton for each mode (Hopfield coefficients) are obtained from eigenstate normalization.

The exciton transition linewidth, cavity-exciton detuning, Rabi splitting, and Q factors are tabulated in Table S1.

Exciton and EP transport analysis

To analyze exciton transport, we extract the mean squared displacement MSD=σ2(t)−σ2(0)\mathrm{MSD} = \sigma^{2}(t) - \sigma^{2}(0), where σ2(t)\sigma^{2}(t) is the variance of the spatial distribution extracted from Gaussian fits to the expanding population profile (Figure 2d in main text, fits in Figure S6). In general, the MSD can be described by MSD=2Dtα\mathrm{MSD} = 2D t^{\alpha}, where DD is diffusivity, and the exponent characterizes: subdiffusive transport (α<1\alpha< 1), diffusive transport (α=1\alpha= 1), superdiffusive transport (1<α<21 < \alpha< 2), and ballistic transport (α=2\alpha= 2).[3]

For EP transport at finite kk, σ2(t)\sigma^{2}(t) can be extracted from fitting the arrival time (t0t_{0}) of the EP at specific positions away from the excitation spot (Figure S7, Figure S7), as described in our previous work.[1] To fit each arrival time across positions, we used a Gaussian function convolved with a biexponential decay.

Wavepacket analysis of EP transport

Figure S7. Wavepacket analysis of EP transport. a. EP transport imaged at k=10.5 μm−1k = 10.5\ \mu\mathrm{m}^{-1} in a PEAP microcavity. Here the pump excitation is at the bottom of the image and propagates upward. Scale bar, 1 μm1\ \mu\mathrm{m}. b. ΔR/R\Delta R/R profiles for each position indicated by the dots in the image (a). c. Distance vs. position, extracted using the wavefront arrival analysis described above. All data symbols are presented as the mean ±\pm one standard deviation derived from the fitting error.

ΔRR=A1e(−tτ1+t0+w22τ12)(1+erf⁡(t−t0−w2τ12w))+A2e(−tτ2+t0+w22τ22)(1+erf⁡(t−t0−w2τ22w))(5)\frac{\Delta R}{R} = A_{1} e^{\left(-\frac{t}{\tau_{1}}+\frac{t_{0}+w^{2}}{2\tau_{1}^{2}}\right)} \left(1+\operatorname{erf}\left(\frac{t-t_{0}-\frac{w^{2}}{\tau_{1}}}{\sqrt{2}w}\right)\right) + A_{2} e^{\left(-\frac{t}{\tau_{2}}+\frac{t_{0}+w^{2}}{2\tau_{2}^{2}}\right)} \left(1+\operatorname{erf}\left(\frac{t-t_{0}-\frac{w^{2}}{\tau_{2}}}{\sqrt{2}w}\right)\right) \tag*{(5)}

where A1,A2,τ1,τ2A_{1}, A_{2}, \tau_{1}, \tau_{2} are amplitudes and decay lifetimes, ww is the IRF, and t0t_{0} is the EP arrival time. An example result plotting distance vs time is shown in Figure S7. Throughout the rest of the text, we plot MSD (rather than distance).

Additional EP transport results for PEAP microcavity

Estimation of temperature-dependent exciton transfer integrals

The zero-temperature exciton transfer integrals J=ℏ/2m0a2J = \hbar/2m_0a^2 are estimated from exciton effective masses m0=me+mhm_0 = m_{\mathrm{e}} + m_{\mathrm{h}} (where mem_{\mathrm{e}} is the conduction band mass and mhm_{\mathrm{h}} is the valence band mass) derived from band-structure calculations reported in the literature, and lattice constants aa. For room-temperature JJ, we apply a temperature-dependent renormalization to the effective mass as described below. Since monomeric molecules are isolated in a polymer matrix in the amorphous TIPS-Pc samples, we assume that J=0J = 0 meV.

There are multiple possible models to calculate temperature-dependent renormalization of effective masses, and the exact form of exciton-phonon interactions responsible for band renormalization in the systems studied here is not known. Here, we use a Holstein-Peierls phonon-induced renormalization model that was previously invoked to explain temperature-dependent changes to JJ observed in angle-resolved photoemission spectra of pentacene crystals and 2D semiconductors[14, 15].

J=J0exp⁡[−(0.5+Neff)geff2](6)J = J_0 \exp\left[-(0.5 + N_{\mathrm{eff}})g_{\mathrm{eff}}^2\right] \tag*{(6)}

where Neff=[exp⁡(ℏωeff/kBT)−1]−1N_{\mathrm{eff}} = [\exp(\hbar\omega_{\mathrm{eff}}/k_{\mathrm{B}}T) - 1]^{-1} refers to the effective phonon occupation number given by the effective phonon frequency (ωeff)(\omega_{\mathrm{eff}}) and geffg_{\mathrm{eff}} is the phonon coupling strength.

The phonon coupling strength is calculated by[14]

geff=[ln⁡(1+λ)]−1(7)g_{\mathrm{eff}} = [\ln(1 + \lambda)]^{-1} \tag*{(7)}

where λ=(2πkB)−1dΓ/dT\lambda= (2\pi k_{\mathrm{B}})^{-1} \mathrm{d}\Gamma/\mathrm{d}T refers to temperature-induced linewidth broadening.

0 K295 K
PEAPa^{\mathrm{a}}11367.2
BAPa^{\mathrm{a}}11358.6
Crystalline
TIPS-Pcb^{\mathrm{b}}90.243.9
Amorphous
TIPS-Pc00

Table S2. Exciton transfer integrals JJ in meV at zero and room temperatures. a. The values for JJ (0 K) were obtained from ref.[2, 16, 17]. The temperature-dependent renormalization uses parameters extracted in Supplemental Note 3 and ref.[7]. b. The values for JJ (0 K) were obtained from ref.[2, 18, 19]. The temperature-dependent renormalization uses parameters from pentacene crystals reported in ref.[14, 15, 20].

Time-resolved photoluminescence of PEAP crystals

Following a procedure outlined in by Huang and co-workers[21], we evaluate the main phonons responsible for exciton-phonon interactions in PEAP single crystals outside microcavities using time-resolved photoluminescence (TRPL) measurements. The PL rise-time reflects intraband scattering to the band edge following pump excitation, which is primarily mediated by exciton-phonon scattering. The temperature-dependent rate of the PL rise can be fit to a power law k=Tγk = T^{\gamma}, with γ\gamma characterizing different phonon contributions as shown in Figure S9.

By fitting the rise-time from the temperature-dependent TRPL results in Figure S9, we extract values of γ\gamma ranging from ~2 to 2.6 depending on the temperature range (Figure S9,Figure S9). These values indicate that exciton-phonon interactions are dominated by homopolar phonon scattering in our PEAP crystals, as had been previously observed in other 2D perovskites.[21]

Low temperature results of PEAP cavities

Image not extracted

Figure S10. Additional results for PEAP microcavity at 3.5 K. a. Angle-resolved reflectance of polariton dispersion for a PEAP microcavity at 3.5 K. Note the lower range of momenta available due to the lower NA objective (0.9 NA) compared to RT datasets. b. Calculated group velocity as a function of detuning energy (E−Eexc\mathrm{E} - \mathrm{E}_{\mathrm{exc}}). The measured EP velocity is overlaid with the expected group velocity.

Image not extracted

Figure S11. Temperature dependent EP transport for PEAP microcavity with ~87% exciton content. a-c. Angle-resolved reflectance of polariton dispersions of PEAP microcavity at 9 K (a), 30 K (b), and 80 K (c). Dashed lines are coupled oscillator model fits used to extract the Hopfield coefficients for each temperature. d-f. MSD plots of EP transport at 9 K (d), 30 K (e), and 80 K (f). All data symbols are presented as the mean ±\pm one standard deviation derived from the fitting error.

Amorphous and Crystalline TIPS-PC films

As shown in Figure S13, amorphous TIPS-Pc thin films show signs of aggregation at concentrations exceeding 20 mg/ml. Thus, all measurements are performed at concentrations of 20 mg/ml for the EP measurements in the text to ensure samples representative of isolated monomeric molecules in a polymer matrix. In contrast, the absorption spectrum of crystalline TIPS-Pc (Figure S13) shows character consistent with mixing of Frenkel and charge transfer excitons typical for crystalline domains of this material.[22]

Image not extracted

Figure S13. Absorbance of TIPS-Pc samples. a. Concentration dependent absorption spectra of amorphous TIPS-Pc films in polystyrene using various concentration ratios (reported in the legend as mg/mg in 1 ml toluene solutions). b. Normalized absorbance of crystalline and amorphous TIPS-Pc films.

EP transport for crystalline TIPS-Pc microcavity at RT

EP transport for amorphous TIPS-Pc microcavity at RT

Note 13. EP transport for crystalline TIPS-Pc microcavity at 5 K.

Note 14. EP transport for amorphous TIPS-Pc microcavity at 5 K

Low temperature absorbance for crystalline and amorphous TIPS-Pc

Correlating EP velocity renormalization to various parameters

Of all parameters explored in Figure S19, we find that EP renormalization correlates best with intersite coupling (Figure S19). While other parameters (S19c-d) show weak correlation, we note that the quantity ΩR/Γ\Omega_{\mathrm{R}}/\Gamma (Figure S19) also displays relatively good correlation. To rule out that this parameter governs EP renormalization in the microcavities explored here, we perform an additional control experiment that varies ΩR\Omega_{\mathrm{R}} while keeping all other parameter constant. We fabricate a partially-filled BAP microcavity using a thin BAP structure encapsulated in polymer spacers (Figure S20). This structure remains in the strong coupling regime, but reduces ΩR\Omega_{\mathrm{R}} from 275 meV in a fully-filled cavity to 190 meV in the partially-filled cavity. We observe very similar EP transport as in the fully-filled cavity (Figure S20), suggesting that ΩR\Omega_{\mathrm{R}} for such layered materials in the strong-coupling regime is not a good predictor of EP renormalization.

Image not extracted

Figure S19. Correlation of EP velocity renormalization with various parameters. All data correspond to EPs with 40% exciton fraction at room temperature. a. EP velocity renormalization as a function of intersite coupling. b. EP velocity renormalization as a function of Rabi splitting/disorder. c. EP velocity as a function of cavity-exciton detuning. d. EP velocity renormalization as a function of Q factor. The symbols correspond to respective microcavities: PEAP (purple circle), BAP (yellow square for fully-filled and yellow hexagon for partially-filled), crystalline TIPS-Pc (blue diamond), amorphous TIPS-Pc (red pentagon) cavities. The red shaded region indicates the 95% confidence band. All data symbols are presented as the mean +/− one standard deviation derived from errors in the linear fit of renormalization vs exciton content.

Supplemental References

  1. Xu, D. et al. Ultrafast imaging of polariton propagation and interactions. Nat Commun 14, 3881 (2023).

  1. Tulyagankhodjaev, J. A. et al. Room-temperature wavelike exciton transport in a van der Waals superatomic semiconductor. Science 382, 438–442 (2023).

  1. Ginsberg, N. S. & Tisdale, W. A. Spatially Resolved Photogenerated Exciton and Charge Transport in Emerging Semiconductors. Annu Rev Phys Chem 71, 1–30 (2020).

  1. Zhu, T., Snaider, J. M., Yuan, L. & Huang, L. Ultrafast Dynamic Microscopy of Carrier and Exciton Transport. Annu Rev Phys Chem 70, 219–244 (2019).

  1. Stoumpos, C. C. et al. Ruddlesden–Popper Hybrid Lead Iodide Perovskite 2D Homologous Semiconductors. Chemistry of Materials 28, 2852–2867 (2016).

  1. Paritmongkol, W., Powers, E. R., Dahod, N. S. & Tisdale, W. A. Two Origins of Broadband Emission in Multilayered 2D Lead Iodide Perovskites. Journal of Physical Chemistry Letters 11, 8565–8572 (2020).

  1. Shen, R. et al. Exciton–Phonon Coupling and Low Energy Emission in 2D and Quasi-2D BA2MAn−1PbnI3n+1\mathrm{BA}_{2}\mathrm{MA}_{n-1}\mathrm{Pb}_{n}\mathrm{I}_{3n+1} Thin Films with Improved Phase Purity. J Phys Chem Lett 12, 12336–12344 (2021).

  1. Gauthron, K. et al. Optical spectroscopy of two-dimensional layered (C6H5C2H4NH3)2PbI4(\mathrm{C}_{6}\mathrm{H}_{5}\mathrm{C}_{2}\mathrm{H}_{4}\mathrm{NH}_{3})_{2}\mathrm{PbI}_{4} perovskite. Opt Express 18, 5912 (2010).

  1. Mandal, A. et al. Microscopic Theory of Multimode Polariton Dispersion in Multilayered Materials. Nano Lett 23, 4082–4089 (2023).

  1. Richter, S. et al. Maxwell consideration of polaritonic quasi-particle Hamiltonians in multi-level systems. Appl Phys Lett 107, 231104 (2015).

  1. Georgiou, K., McGhee, K. E., Jayaprakash, R. & Lidzey, D. G. Observation of photon-mode decoupling in a strongly coupled multimode microcavity. J Chem Phys 154, 124309 (2021).

  1. Balasubrahmaniyam, M., Genet, C. & Schwartz, T. Coupling and decoupling of polaritonic states in multimode cavities. Phys Rev B 103, L241407 (2021).

  1. Liu, V. & Fan, S. S4: A free electromagnetic solver for layered periodic structures. Comput Phys Commun 183, 2233–2244 (2012).

  1. Hatch, R. C., Huber, D. L. & Höchst, H. Electron-Phonon Coupling in Crystalline Pentacene Films. Phys Rev Lett 104, 047601 (2010).

  1. Li, Q., Liu, F., Russell, J. C., Roy, X. & Zhu, X. Strong polaronic effect in a superatomic two-dimensional semiconductor. J Chem Phys 152, 171101 (2020).

  1. Ziegler, J. D. et al. Fast and Anomalous Exciton Diffusion in Two-Dimensional Hybrid Perovskites. Nano Lett 20, 6674–6681 (2020).

  1. Ziegler, J. D. et al. Excitons at the Phase Transition of 2D Hybrid Perovskites. ACS Photonics 9, 3609–3616 (2022).

References

  1. [1]Hopfield, J. J. Theory of the Contribution of Excitons to the Complex Dielectric Constant of Crystals. Physical Review 112, 1555–1567 (1958).DOI
  2. [2]Mills, D. L. & Burstein, E. Polaritons: the electromagnetic modes of media. Reports on Progress in Physics 37, 817–926 (1974).DOI
  3. [3]Weisbuch, C., Nishioka, M., Ishikawa, A. & Arakawa, Y. Observation of the coupled exciton–photon mode splitting in a semiconductor quantum microcavity. Phys Rev Lett 69, 3314–3317 (1992).pubmed.ncbi.nlm.nih.gov/10046787
  4. [4]Sanvitto, D. & Kéna-Cohen, S. The road towards polaritonic devices. Nat Mater 15, 1061–1073 (2016).DOI
  5. [5]Basov, D. N., Fogler, M. M. & García de Abajo, F. J. Polaritons in van der Waals materials. Science 354, aag1992 (2016).pubmed.ncbi.nlm.nih.gov/27738142
  6. [6]Steger, M. et al. Long-range ballistic motion and coherent flow of long-lifetime polaritons. Phys Rev B 88, 235314 (2013).DOI
  7. [7]Coles, D. M. et al. Polariton-mediated energy transfer between organic dyes in a strongly coupled optical microcavity. Nat Mater 13, 712–719 (2014).DOI
  8. [8]Hu, F. et al. Imaging exciton–polariton transport in MoSe2 waveguides. Nat Photonics 11, 356–360 (2017).DOI
  9. [9]Rozenman, G. G., Akulov, K., Golombek, A. & Schwartz, T. Long-Range Transport of Organic Exciton-Polaritons Revealed by Ultrafast Microscopy. ACS Photonics 5, 105–110 (2018).
  10. [10]Wurdack, M. et al. Motional narrowing, ballistic transport, and trapping of room-temperature exciton polaritons in an atomically-thin semiconductor. Nat Commun 12, 5366 (2021).DOI
  11. [11]Xu, D. et al. Ultrafast imaging of polariton propagation and interactions. Nat Commun 14, 3881 (2023).
  12. [12]Balasubrahmaniyam, M. et al. From enhanced diffusion to ultrafast ballistic motion of hybrid light–matter excitations. Nat Mater 22, 338–344 (2023).DOI
  13. [13]Dang, N. H. M. et al. Long-Range Ballistic Propagation of 80% Excitonic Fraction Polaritons in a Perovskite Metasurface at Room Temperature. Nano Lett 24, 11839–11846 (2024).arxiv.org/abs/2406.01271
  14. [14]Son, M. et al. Energy cascades in donor-acceptor exciton-polaritons observed by ultrafast two-dimensional white-light spectroscopy. Nat Commun 13, 7305 (2022).DOI
  15. [15]Berghuis, A. M. et al. Controlling Exciton Propagation in Organic Crystals through Strong Coupling to Plasmonic Nanoparticle Arrays. ACS Photonics 9, 2263–2272 (2022).DOI
  16. [16]Khatoniar, M. et al. Relaxing Symmetry Rules for Nonlinear Optical Interactions in Van der Waals Materials via Strong Light–Matter Coupling. ACS Photonics 9, 503–510 (2022).DOI
  17. [17]Xu, D. et al. Spatiotemporal imaging of nonlinear optics in van der Waals waveguides. Nat Nanotechnol (2025) doi:10.1038/s41565-024-01849-1.DOI
  18. [18]Sun, Y. et al. Direct measurement of polariton–polariton interaction strength. Nat Phys 13, 870–875 (2017).DOI
  19. [19]Fieramosca, A. et al. Two-dimensional hybrid perovskites sustaining strong polariton interactions at room temperature. Sci Adv 5, eaav9967 (2019).DOI
  20. [20]Quirós-Cordero, V. et al. Competitive exciton and polariton scattering inhibits condensation in two-dimensional metal-halide-semiconductor microcavities. arXiv:2404.14744 (2024).
  21. [21]Zhao, J. et al. Nonlinear polariton parametric emission in an atomically thin semiconductor based microcavity. Nat Nanotechnol 17, 396–402 (2022).DOI
  22. [22]Keeling, J. & Kéna-Cohen, S. Bose–Einstein Condensation of Exciton-Polaritons in Organic Microcavities. Annu Rev Phys Chem 71, 435–459 (2020).DOI
  23. [23]Deng, H., Haug, H. & Yamamoto, Y. Exciton-polariton Bose-Einstein condensation. Rev Mod Phys 82, 1489–1537 (2010).DOI
  24. [24]Zasedatelev, A. V. et al. Single-photon nonlinearity at room temperature. Nature 597, 493–497 (2021).
  25. [25]Muñoz-Matutano, G. et al. Emergence of quantum correlations from interacting fibre-cavity polaritons. Nat Mater 18, 213–218 (2019).
  26. [26]Delteil, A. et al. Towards polariton blockade of confined exciton–polaritons. Nat Mater 18, 219–222 (2019).
  27. [27]Ribeiro, R. F., Martínez-Martínez, L. A., Du, M., Campos-Gonzalez-Angulo, J. & Yuen-Zhou, J. Polariton chemistry: controlling molecular dynamics with optical cavities. Chem Sci 9, 6325–6339 (2018).arxiv.org/abs/1802.08681
  28. [28]Garcia-Vidal, F. J., Ciuti, C. & Ebbesen, T. W. Manipulating matter by strong coupling to vacuum fields. Science 373, eabd0336 (2021).
  29. [29]Pandya, R. et al. Tuning the Coherent Propagation of Organic Exciton-Polaritons through Dark State Delocalization. Advanced Science 9, 2105569 (2022).DOI
  30. [30]Sokolovskii, I., Tichauer, R. H., Morozov, D., Feist, J. & Groenhof, G. Multi-scale molecular dynamics simulations of enhanced energy transfer in organic molecules under strong coupling. Nat Commun 14, 6613 (2023).DOI
  31. [31]Engelhardt, G. & Cao, J. Polariton Localization and Dispersion Properties of Disordered Quantum Emitters in Multimode Microcavities. Phys Rev Lett 130, 213602 (2023).DOI
  32. [32]Chng, B. X. K., Mondal, M. E., Ying, W. & Huo, P. Quantum Dynamics Simulations of Exciton Polariton Transport. Nano Lett 25, 1617–1622 (2025).DOI
  33. [33]Suyabatmaz, E. & Ribeiro, R. F. Vibrational polariton transport in disordered media. J Chem Phys 159, 034701 (2023).DOI
  34. [34]Krupp, N., Groenhof, G. & Vendrell, O. Quantum dynamics simulation of exciton-polariton transport. arXiv:2410.23739 (2024).arxiv.org/abs/2410.23739
  35. [35]Litinskaya, M. Propagation and localization of polaritons in disordered organic microcavities. Phys Lett A 372, 3898–3903 (2008).DOI
  36. [36]Blackham, L., Manjalingal, A., Koshkaki, S. R. & Mandal, A. Microscopic Theory of Polaron-Polaritон Dispersion and Propagation. arXiv:2501.16622 (2025).DOI
  37. [37]Su, R. et al. Perovskite semiconductors for room-temperature exciton-polaritonics. Nat Mater 20, 1315–1324 (2021).DOI
  38. [38]Paritmongkol, W., Powers, E. R., Dahod, N. S. & Tisdale, W. A. Two Origins of Broadband Emission in Multilayered 2D Lead Iodide Perovskites. Journal of Physical Chemistry Letters 11, 8565–8572 (2020).DOI
  39. [39]Seitz, M. et al. Exciton diffusion in two-dimensional metal-halide perovskites. Nat Commun 11, 2035 (2020).
  40. [40]Xiao, X. et al. Ultrafast Exciton Transport with a Long Diffusion Length in Layered Perovskites with Organic Cation Functionalization. Advanced Materials 32, 2004080 (2020).
  41. [41]Deng, S. et al. Long-range exciton transport and slow annihilation in two-dimensional hybrid perovskites. Nat Commun 11, 664 (2020).
  42. [42]Freixanet, T., Sermage, B., Tiberj, A. & Planel, R. In-plane propagation of excitonic cavity polaritons. Phys Rev B 61, 7233–7236 (2000).DOI
  43. [43]Hou, S. et al. Ultralong-Range Energy Transport in a Disordered Organic Semiconductor at Room Temperature Via Coherent Exciton-Polariton Propagation. Advanced Materials 32, 2002127 (2020).DOI
  44. [44]Barachati, F. et al. Interacting polariton fluids in a monolayer of tungsten disulfide. Nat Nanotechnol 13, 906–909 (2018).
  45. [45]Lei, H. et al. Persistent Exciton Dressed by Weak Polaronic Effect in Rigid and Harmonic Lattice Dion–Jacobson 2D Perovskites. ACS Nano 18, 31485–31494 (2024).DOI
  46. [46]Biswas, S. et al. Exciton polaron formation and hot-carrier relaxation in rigid Dion–Jacobson-type two-dimensional perovskites. Nat Mater 23, 937–943 (2024).DOI
  47. [47]Thouin, F. et al. Phonon coherences reveal the polaronic character of excitons in two-dimensional lead halide perovskites. Nat Mater 18, 349–356 (2019).
  48. [48]Guo, Z., Wu, X., Zhu, T., Zhu, X. & Huang, L. Electron–Phonon Scattering in Atomically Thin 2D Perovskites. ACS Nano 10, 9992–9998 (2016).pubmed.ncbi.nlm.nih.gov/27740753
  49. [49]Fivaz, R. & Mooser, E. Electron-Phonon Interaction in Semiconducting Layer Structures. Physical Review 136, A833–A836 (1964).DOI
  50. [50]Du, Q. et al. Stacking Effects on Electron–Phonon Coupling in Layered Hybrid Perovskites via Microstrain Manipulation. ACS Nano 14, 5806–5817 (2020).DOI
  51. [51]Engelhardt, G. & Cao, J. Unusual dynamical properties of disordered polaritons in microcavities. Phys Rev B 105, 064205 (2022).DOI
  52. [52]Aroeira, G. J. R., Kairys, K. T. & Ribeiro, R. F. Coherent transient exciton transport in disordered polaritonic wires. Nanophotonics 13, 2553–2564 (2024).DOI
  53. [53]Liu, T., Yin, G. & Xiong, W. Unlocking Delocalization: How Much Coupling Strength can Overcome Energy Disorder in Molecular Polaritons? Chem Sci 16, 4676–4683 (2025).DOI
  54. [54]Mandal, A. et al. Microscopic Theory of Multimode Polariton Dispersion in Multilayered Materials. Nano Lett 23, 4082–4089 (2023).DOI
  55. [55]Ying, W., Chng, B. X. K., Delor, M. & Huo, P. Microscopic Theory of Polariton Group Velocity Renormalization. ArXiv:2411.08288 (2024).arxiv.org/abs/2411.08288
  56. [56]Bradbury, N. C., Ribeiro, R. F., Caram, J. R. & Neuhauser, D. Stochastic methodology shows molecular interactions protect two-dimensional polaritons. Phys Rev B 109, L241303 (2024).DOI
  57. [57]Dirnberger, F. et al. Magneto-optics in a van der Waals magnet tuned by self-hybridized polaritons. Nature 620, 533–537 (2023).DOI
  58. [58]Anantharaman, S. B. et al. Ultrastrong light–matter coupling in two-dimensional metal–organic chalcogenolates. Nat Photonics (2025) doi:10.1038/s41566-024-01590-0.DOI

Paper details

vlnc:W_ggu8vpyy
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents