# Fileset

[s41467-024-45554-y.pdf](https://mdr.nims.go.jp/filesets/779e5e10-e772-411a-82d2-7294a0e62958/download)

## Creator

Zehua Hu, Tanjung Krisnanda, Antonio Fieramosca, Jiaxin Zhao, Qianlu Sun, Yuzhong Chen, Haiyun Liu, Yuan Luo, Rui Su, Junyong Wang, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Goki Eda, Xiao Renshaw Wang, Sanjib Ghosh, Kevin Dini, Daniele Sanvitto, Timothy C. H. Liew, Qihua Xiong

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Energy transfer driven brightening of MoS2 by ultrafast polariton relaxation in microcavity MoS2/hBN/WS2 heterostructures](https://mdr.nims.go.jp/datasets/122a35ec-4d1c-4dff-a2b5-73ecdf50d40a)

## Fulltext

Energy transfer driven brightening of MoS2 by ultrafast polariton relaxation in microcavity MoS2/hBN/WS2 heterostructuresArticle https://doi.org/10.1038/s41467-024-45554-yEnergy transfer driven brightening ofMoS2 by ultrafast polariton relaxation inmicrocavity MoS2/hBN/WS2heterostructuresZehua Hu 1,13 , Tanjung Krisnanda2,13, Antonio Fieramosca3,13, Jiaxin Zhao4,Qianlu Sun1, Yuzhong Chen5, Haiyun Liu 5, Yuan Luo6, Rui Su 4,Junyong Wang 7, Kenji Watanabe 8, Takashi Taniguchi 9, Goki Eda 7,Xiao Renshaw Wang 4,10, Sanjib Ghosh5, Kevin Dini4 , Daniele Sanvitto3,Timothy C. H. Liew 4 & Qihua Xiong 5,6,11,12Energy transfer is a ubiquitous phenomenon that delivers energy from a blue-shifted emitter to a red-shifted absorber, facilitating wide photonic applica-tions. Two-dimensional (2D) semiconductors provide unique opportunitiesfor exploring novel energy transfer mechanisms in the atomic-scale limit.Herein, we have designed a planar optical microcavity-confined MoS2/hBN/WS2 heterojunction, which realizes the strong coupling among donor exciton,acceptor exciton, and cavity photon mode. This configuration demonstratesan unconventional energy transfer via polariton relaxation, brightening MoS2with a record-high enhancement factor of ~440, i.e., two-order-of-magnitudehigher than the data reported to date. The polariton relaxation features a shortcharacteristic time of ~1.3 ps, resulting from the significantly enhanced intra-and inter-branch exciton-exciton scattering. The polariton relaxation dynam-ics is associated with Rabi energies in a phase diagram by combining experi-mental and theoretical results. This study opens a newdirection ofmicrocavity2D semiconductor heterojunctions for high-brightness polaritonic light sour-ces and ultrafast polariton carrier dynamics.Monolayer semiconducting transition metal dichalcogenides (TMDs)have recently emerged as a promising system in fundamental physicsand technology-related studies of electronics, optoelectronics, val-leytronics, and twistronics. Their prominent properties are enabled bytheweakdielectric screening, breaking of the inversion symmetry, andstrong spin-orbit coupling1–4. Despite being direct bandgap semi-conductors, most monolayer TMDs exhibit poor photoluminescence(PL) yield due to the low absolute optical absorption of <10% as well asthe low external quantum yield of <1%, which limits the developmentof nanophotonic devices such as light-emitting diodes (LEDs), lasers,display devices, and optical on-chip networks5–7.One prospective way to achieve a homogeneous enhancement ofemission intensity is to deliver the excitation energy from the donor tothe acceptor by nonradiative energy transfer. Conventional energytransfer comprises Förster resonance energy transfer (FRET) andDexter-type energy transfer (DET); the former is based on dipole-dipole coupling (where the dipoles in the semiconductors are exci-tons), while the latter is based on charge transfer and spinconservation8–11. To evaluate the effect of energy transfer, anenhancement factor (η) is usually defined asη= IDA=IA, where IDA and IAare the PL intensities of the acceptor in the heterojunction and on thebare substrate, respectively. For both FRET and DET in 2DReceived: 9 January 2024Accepted: 29 January 2024Check for updatesA full list of affiliations appears at the end of the paper. e-mail: zehuahu@nju.edu.cn; kdini@ntu.edu.sg; qihua_xiong@tsinghua.edu.cnNature Communications |         (2024) 15:1747 11234567890():,;1234567890():,;http://orcid.org/0000-0002-1185-2992http://orcid.org/0000-0002-1185-2992http://orcid.org/0000-0002-1185-2992http://orcid.org/0000-0002-1185-2992http://orcid.org/0000-0002-1185-2992http://orcid.org/0000-0002-3584-4778http://orcid.org/0000-0002-3584-4778http://orcid.org/0000-0002-3584-4778http://orcid.org/0000-0002-3584-4778http://orcid.org/0000-0002-3584-4778http://orcid.org/0000-0002-2808-0327http://orcid.org/0000-0002-2808-0327http://orcid.org/0000-0002-2808-0327http://orcid.org/0000-0002-2808-0327http://orcid.org/0000-0002-2808-0327http://orcid.org/0000-0003-4653-6671http://orcid.org/0000-0003-4653-6671http://orcid.org/0000-0003-4653-6671http://orcid.org/0000-0003-4653-6671http://orcid.org/0000-0003-4653-6671http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1575-8020http://orcid.org/0000-0002-1575-8020http://orcid.org/0000-0002-1575-8020http://orcid.org/0000-0002-1575-8020http://orcid.org/0000-0002-1575-8020http://orcid.org/0000-0002-5503-9899http://orcid.org/0000-0002-5503-9899http://orcid.org/0000-0002-5503-9899http://orcid.org/0000-0002-5503-9899http://orcid.org/0000-0002-5503-9899http://orcid.org/0000-0003-2568-7294http://orcid.org/0000-0003-2568-7294http://orcid.org/0000-0003-2568-7294http://orcid.org/0000-0003-2568-7294http://orcid.org/0000-0003-2568-7294http://orcid.org/0000-0002-2555-4363http://orcid.org/0000-0002-2555-4363http://orcid.org/0000-0002-2555-4363http://orcid.org/0000-0002-2555-4363http://orcid.org/0000-0002-2555-4363http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45554-y&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45554-y&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45554-y&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45554-y&domain=pdfmailto:zehuahu@nju.edu.cnmailto:kdini@ntu.edu.sgmailto:qihua_xiong@tsinghua.edu.cnsemiconductors, η is limited to ~2 to date8,10,12–14. High-performancenanophotonic devices call for a new structure or mechanism forenergy transfer with a colossal η to brighten 2D semiconductors.Alternatively, for organic-semiconductor heterojunctions withFrenkel excitons, exciton-photon polaritons due to the strong cou-pling of excitons with microcavity provide a powerful platform toachieve unconventional energy transfer by polariton relaxation15–20. Incontrast, TMDs feature a more delocalized Wannier-Mott excitonnature21,22. Their polaritons inherit the perfect in-plane dipole orien-tation, strong quantum confinement, and valley polarization, whichprovides more degree of freedoms to engineer new functionalities inthe polariton field. Thus, TMDs polaritons have attracted great atten-tion for the study of Bose-Einstein condensates23–25, nonlinear opticalprocesses26,27, valley properties28–30, LEDs31 as well as Moiréheterojunctions32. However, it remains elusive to realize the strongcoupling of TMDs heterojunctions with microcavity for unconven-tional energy transfer due to the considerable challenge to avoid thePL quenching by ultrafast charge transfer and meanwhile to achievethe mode match among donor exciton, acceptor exciton, and cavityphoton.Here, by embedding the hBN/MoS2/hBN/WS2 heterojunctioninto a Fabry-Pérot (FP) microcavity, we have realized the strongcoupling among the donor exciton, acceptor exciton, and cavityphoton mode, which leads to the brightening of MoS2 with a two-order-of-magnitude enhancement of PL yield (η ≈ 440) based on theexciton-photon polariton relaxation. The custom-designed k-spacetransient-reflectivity spectroscopy, which features a better spectralresolution than the real-spacemeasurement formicrocavity samples,was applied to evaluate the polariton relaxation dynamics, and acharacteristic time as short as ~1.3 ps was extracted. The efficient andultrafast polariton relaxation is associated with the significantlyenhanced intra- and inter-branch exciton-exciton scattering, whichovercomes the hot phonon bottleneck effect. Moreover, the phasediagram has been established to correlate the polariton relaxationefficiency with the Rabi energies. This study demonstrates a newbranch of microcavity-confined 2D semiconductor heterojunctionswith a great potential for ultrafast highly-bright polaritonic lightsources.ResultsSample fabrication and optical characterizationFor the optical microcavity-confined heterojunction, the strong cou-pling among the acceptor’s (Eex1) and donor’s (Eex2) excitons as well asthe cavity photonmode (Ec)will form the exciton-photonpolaritonwiththree anti-crossing eigenstates denoted by the upper (UPB), middle(MPB), and lower (LPB) polariton branches (Fig. 1a)15–19. Under off-resonance optical excitation, the polariton tends to relax from the UPBto MPB and finally to LPB (i.e., from high energy and high in-planemomentum (k//) to low energy and low k//). The relaxation process is anunconventional energy transfer since the UPB and LPB are mostly Eex2-Upper polaritonMiddle polaritonLower polaritonk//EnergyPhotonAcceptor exciton(Eex1)Cavity photon    (Ec)Donor exciton(Eex2)+ + =Strongcoupling-+-+pumpemissionWS2  XA(Eex2)MoS2  XBMoS2  XA(Eex1)Polariton emissionacbNormalized PL intensity (arb. units)-ΔR/R (arb. units)1800 1900 2000 21000.00.51.0Energy (meV)PL het@cavity− PL het@SiO20.00.30.6reflectivityhet@SiO2Fig. 1 | Design, construction, and properties of FP microcavity-confined hBN/MoS2/hBN/WS2 heterojunction. a Schematic formation and relaxation of exciton-photon polariton. The black and red arrows in the right panel represent intra- andinter-branch scattering, respectively. b Schematic illustration of microcavity-confined hBN/MoS2/hBN/WS2 heterojunction. c Normalized real-space PL spectraof het@SiO2 (black) and het@cavity− (blue). The differential reflection spectrumofhet@SiO2 (dashed red) is presented for reference. XA and XB are A and B exciton,respectively.Article https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 2like and Eex1-like, respectively, as determined by the Hopfieldcoefficients15–19.The configuration of microcavity-confined heterojunction inthis study is illustrated in Fig. 1b. The hBN/MoS2/hBN/WS2 het-erojunction was fabricated by a pick-up method (SupplementaryFig. 1)32. WS2 (donor) and MoS2 (acceptor) are monolayers with adirect bandgap, while the top and middle hBN are 15 and 2 nm,respectively. The middle thin hBN layer prevents the ultrafastinterlayer charge transfer from nonradiative recombination whilekeeping the FRET almost intact13,33,34. Then, the heterojunctionwas transferred onto a bottom-distributed Bragg reflector (DBR)composed of 6.5 pairs of SiO2 and TiO2 dielectric layers, with itsphotonic stopband covering the heterojunction exciton energies(Supplementary Fig. 2). Subsequently, a poly(methyl methacry-late) (PMMA) spacer and a silver mirror were deposited to finalizethe microcavity device. The thickness of PMMA spacer wasadjusted to achieve the negative (positive) acceptor exciton-photon detuning, with the microcavity heterojunction labeled ashet@cavity− (het@cavity+), respectively. Supplementary Table 1summarizes the thickness information of het@cavity−, whichensures an energy match between the cavity mode and hetero-junction excitons, i.e., the heterojunction placed in the maximumof the electromagnetic field. For quantitative comparison, anidentical heterojunction was fabricated on a SiO2(90 nm)-coatedSi substrate to avoid the cavity effect, defined as het@SiO2(Supplementary Fig. 3a).Figure 1c compares the normalized real-space PL spectra ofhet@cavity− and het@SiO2, with the differential reflection spectrumof het@SiO2 as a reference. The real-space PL spectrum of het@SiO2features two emission peaks, i.e., A exciton (XA) emission from WS2at 2002 meV and MoS2 at 1890 meV, with a Stokes shift of ~10meVrelative to the corresponding reflectivity spectrum. The PL intensityof WS2 XA is much stronger than that of MoS2 XA, which results fromthe higher intrinsic exciton dipole oscillator strength (f) and quan-tum yield of the former, and the interlayer FRET (η ≈ 3), as detailed inSupplementary Fig. 3. On the contrary, the PL spectrum ofhet@cavity− is only composed of a single peak at 1873 meV, with theeven smaller energy than the PL of MoS2 XA. The significant changeof PL spectra, in both shape and energy, points out the crucial effectof the optical microcavity confinement on energy transfer.The energy transfer in het@cavity− is further understood bymeans of the k-space energy-resolved reflectivity and PL map-pings, in comparison with the cases of microcavity-confined hBN/MoS2/hBN (MoS2@cavity) and hBN/WS2/hBN (WS2@cavity), assummarized in Fig. 2a–f. The polariton dispersions for het@-cavity− are fitted with a three coupled oscillator model15–17, i.e.,Ec θð Þ Ω2=2 Ω1=2Ω2=2 Eex2 0Ω1=2 0 Eex10B@1CAαcαex2αex10@1A= Epol θð Þαcαex2αex10@1A ð1Þwhere Ec kð Þ= E0ð1� sin2ðθÞ=n2Þ�1=2is the cavity photon mode energy,θ is the angle of incidence, E0 is the cavity cut off energy, and n is thecavity effective refractive index. Eex1, Eex2 and Epol(θ) are the energies ofacceptor exciton, donor exciton, andpolariton, respectively.Ω1 andΩ2are the corresponding Rabi splitting energies. αex1, αex2 and αc are thecorresponding Hopfield coefficients (Supplementary Fig. 4). The angledependence of the three eigenvalues corresponds to three dispersionbranches termed UPB, MPB, and LPB, respectively (the white dottedlines in Fig. 2a, b). Similarly, the polaritondispersions forMoS2@cavityand WS2@cavity are fitted with a two-coupled oscillator model, whichgives two branches, termed UPB and LPB (see Method).The reflectivity mappings of MoS2@cavity and WS2@cavity showtwo anti-crossing branches with Rabi splitting energy (Ω) of 18 and41meV, respectively (Fig. 2c, e), consistent with the smaller excitondipole oscillator strength of MoS2 than WS2 (Ω /ffiffiffifp, SupplementaryFig. 3b)35,36. In contrast, het@cavity− features three branches with twoanti-crossing points, demonstrating the strong coupling among MoS2exciton, WS2 exciton, and cavity photon mode, with Ω1 (Ω2) ≈ 26 (26)meV and a negative (negative) exciton-photon detuning (Δ) of −15(−150)meV forMoS2 (WS2) (Fig. 2a).Ω1 of het@cavity− is larger thanΩof MoS2@cavity and Ω2 of het@cavity− is smaller than Ω ofWS2@cavity, which straightforwardly indicates the transfer of dipoleoscillator strength from Eex2 to Eex137. A similar transfer is also reflectedin het@cavity+ with aΩ1 (Ω2) ≈ 42 (24) meV and a positive (negative) Δof +13 (−132) meV for MoS2 (WS2) (Supplementary Fig. 5a). These1800 1900 2000 2100050001000015000Energy (mV)het@cavity−MoS2@cavity110100100010000MoS2@cavityhet@cavity−LPB MPB UPB1800 1900 2000 2100Energy(meV)k (μm-1)k (μm-1)900hBN/MoS2/hBN (MoS2@cavity)Energy(meV)Ω 1810Theory        Exp.cdTheory        Exp.ΔMoS2 = -1039000Energy(meV)Energy(meV)k (μm-1)k (μm-1)hBN/WS2/hBN (WS2@cavity)efΩ 4110Theory        Exp.Theory        Exp.ΔWS2 = -10539000Energy(meV)k (μm-1)hBN/MoS2/hBN/WS2 (het@cavity−)abEnergy(meV)k (μm-1)10Theory        Exp.Theory        Exp.Ω2=26ΔWS2 = -150ΔMoS2 = -15Ω1=26Intensity (counts)ghReal-space spectraIntensity  (counts)Energy (meV)-3 -2 -1 0 1 2 2-3-3 -1 0 1 2 2-3-3 -1 0 1 2 3-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 31850190019502000205018501900195020002050185019001950200020501850190019502000205019001950200020501900195020002050Fig. 2 | Comparison of polariton relaxations in het@cavity−, MoS2@cavity andWS2@cavity. a, b k-space energy-resolved reflectivitymapping (a), PLmapping (b)for het@cavity − . c, d The corresponding results for MoS2@cavity. e, f The cor-responding results for WS2@cavity. The Rabi splitting (Ω) and exciton-photondetuning (Δ) energies are labeled in the unit of meV. Note: data range of the colorbar in (d) is much smaller than that in (b) and (f). g, h The real-space PL spectra ofhet@cavity− and MoS2@cavity in linear (g), and log (h) scale under 10 μWexcitation.Article https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 3results suggest a general feature of the transfer of dipole oscillatorstrength in the strongly coupled microcavity-confined heterojunc-tions, in sharp contrast to the negligible transfer in the cavity-freehet@SiO2 counterpart (Supplementary Fig. 3b). The experimentalreflectivity mappings are well-reproduced by theoretical calculationwith the transfer matrix method (Fig. 2a, c, e, detailed in Supplemen-tary part III)38–40.The PL mapping of MoS2@cavity suffers from a weak polaritonemission due to the weak dipole oscillator strength and low quantumyield of MoS2 (Fig. 2d)5, while that of WS2@cavity suffers from a hotphononbottleneck effect (thebroad emission centred at ~1.7μm−1) dueto the large negativeΔ (−105meV, Fig. 2f)29,41,42. On the contrary, the PLmapping of het@cavity− integrates the advantages of MoS2@cavityandWS2@cavity, i.e., inherits the narrow linewidth fromMoS2@cavityand high intensity fromWS2@cavity (Fig. 2b); so does the PL mappingof het@cavity+ (Supplementary Fig. 5b). The PL intensity of het@-cavity− in real-space photoluminescence excitation spectrum increa-ses monotonously without resonance peak when decreasing theexcitation energy towards LPB (Supplementary Fig. 6). The precedingresults confirm the energy transfer mechanism of polaritonrelaxation, rather than FRET or DET, in the strongly coupledmicrocavity-confined heterojunctions. In the linear-scale real-space PLspectrum, het@cavity− and MoS2@cavity show bright and negligibleemission under 10μWexcitation, respectively (Fig. 2g). From Fig. 2b, itis learned that the polariton emission of het@cavity− comes from theLPB and MPB with |k//| < 1.6 μm−1, which is composed of MoS2 excitonand cavity photon but negligible WS2 exciton (Supplementary Fig. 4a).Hence,η is obtained to be ~440 from the ratio of ILPB +MPBhet@cavity�=ILPB +UPBMoS2@cavity.Here ILPB+MPBhet@cavity� and ILPB+UPBMoS2@cavity are the integrated polariton emissionintensity from LPB +MPB (red+cyan region) of het@cavity− andLPB +UPB of MoS2@cavity (black-outlined region) in real-space,respectively (Fig. 2h, detailed in Supplementary Note 1). As summar-ized in Supplementary Table 2, the giant η of 440 for het@cavity− isover 146 times larger than the η of 3 with FRET for het@SiO2 coun-terpart, also two orders ofmagnitude higher than the typical η of FRETor DET processes in literature, demonstrating a powerful strategy tosignificantly increase the energy transfer efficiency by polaritonrelaxation.Supplementary Figs. 7 and 8 show the optical characteristicsof the additional microcavity-confined heterojunction samples,specifically identified as het8@cavity (8 nm hBN), het2@no-de@cavity (heterojunction with 2 nm hBN at the node), andhet60@cavity (60 nm hBN). These diverse configurations, char-acterized by different heterojunction’s positions within the cav-ity and/or different thicknesses of hBN, give rise to distinct Rabienergies Ω1 and Ω2. Consequently, the distinct Rabi energies leadto a substantial difference in polariton relaxation efficiency. Thisresult demonstrates the crucial role of Rabi energies in polariton-mediated energy transfer, which has not been recognized to dateand will be further theoretically clarified later.Theoretical simulationWe simulated the k-space PL mappings with the coupled rateequations43–46, which describe the dynamics of polariton populationsin different branches. The different points with the different relativeexcitonic and photonic fractions in k-space for each branch are iden-tified by the coupled oscillator model and the Hopfield coefficients(Supplementary Fig. 4). The coupled rate equations for het@cavity−with three polariton branches readdNΞjdt= P1XΞ1,j +P2XΞ2,j � γCΞj� �NΞj +XΞ0SΞΞ0 ð2Þwhere NΞj is the polariton population, Ξ= U,M,Lf g corresponds to theUPB,MPB, and LPB, 1 (2) indicatesMoS2 (WS2), and index j denotes therespective point in k-space. P1ð2Þ stands for an effective pumping forthe exciton, and γ for the decay rate of the photonmode. Here, P2>P1due to the much brighter polariton emission of WS2@cavity thanMoS2@cavity (Fig. 2d, f). The XΞ1ð2Þ and CΞ denote jα1ð2Þj2 and αc�� ��2(Supplementary Fig. 4). The relaxation processes, described by the lastterm in (2), consist of the intra- and inter-branch transitions originatingfrom the phonon-mediated exciton scattering47,48. In this formalism,the preferred transition with a high rate happens between initial andfinal destinations with high XΞ1 XΞ01 ,XΞ2XΞ02 and phonon populationnphðEphÞ. According to Bose distribution function, i.e., nph =1expEphkBT� ��1(here Eph = jEΞ0 � EΞj), nph is large for transitions with close energies.The photonic part of the numerical steady-state solution for NΞj ,together with an added Gaussian broadening, agrees with ourexperimental result (Fig. 2b, d, f; Supplementary part III for details).Specifically, for MoS2@cavity, P1 excites the population in thelow- (high-) k// region of UPB (LPB) with high exciton fractions XU1 ðXL1 Þ(Supplementary Fig. 4b). The population will relax by the inter-branch scattering (from UPB to LPB) due to the high XU1 XL1 at the anti-crossing point and the high nph (Eph ≈Ω ≈ 18meV), followed by theintra-branch scattering (within LPB) due to the non-negligible XL1XL01and the high nph (flat dispersion)47,48. However, the polariton emis-sion is considerably weak due to the weak oscillator strength and lowquantum yield of MoS2 exciton (Supplementary Fig. 3). ForWS2@cavity, P2 also excites the population in the low- (high-) k//region of UPB (LPB) with high exciton fractions XU2 (XL2) (Supple-mentary Fig. 4c). However, only a tiny portion of the population willrelax to the bottomof LPBdue to the poor inter-branch scattering forthe low nph (Eph ≈ Ω ≈ 41meV) and the poor intra-branch scatteringfor the negligible XL2XL02 , presenting the troublesome hot phononbottleneck (Fig. 2f)29,41.Based on the above experimental and theoretical results, thepolariton relaxation in het@cavity− could be well elucidated (Fig. 2b).Due to P2 >P1, P2 excites themost population in the k// < 2.5 (>2.5)μm−1region of UPB (MPB) with high XU2 (XM2 ) (Supplementary Fig. 4a). Thepopulation will relax by the inter-branch scattering (fromUPB toMPB)due to the high XU2 XM2 at the top anti-crossing point and the high nph(Eph ≈ Ω2 ≈ 26meV). The subsequent intra-branch scattering withinMPB is faster than that within LPB for WS2@cavity due to the non-negligibleXM2 XM 02 andXM1 XM 01 for the former compared to the negligibleXL2XL02 for the latter. The synergismof the two scatterings leads to a fastrelaxation of almost all the population to the bottom anti-crossingpoint. The population will further relax towards the bottomof LPB in away similar to that in MoS2@cavity, as reflected in Fig. 1a. Such a fastpolariton relaxation dynamics leads to the bright PL of het@cavity− atthe bottom of LPB (Fig. 2b).Dynamics and phase diagramWith a custom-designed microscopic k-space transient-reflectivityspectroscopy (see Methods), we have simultaneously unravelled theevolution of the polariton population in momentum, energy and timedomains. After being pumped by a 200-fs laser pulse of ~2 μW at580nm, the change of reflectivity ð4RR =Rpump on�Rpump offRpump offÞ of the broad-bandwhite light is measured after a specific time delay, where Rpump onand Rpump off are the reflectivity mappings with or without opticalpumping, respectively. Such an excitation fluence is low enough toavoid exciton-exciton annihilation49,50. The k-space transient-reflectivity spectroscopy results of WS2@cavity and het@cavity− areshown in Fig. 3, with the corresponding movies in SupplementaryMaterial. For the WS2@cavity, the laser pulse excites the excitonreservoir to quickly form polaritons, which induces a net photo-bleaching signal at 0 ps at k// ≈ 2.3 μm−1 (Fig. 3a), also demonstrated inthe integrated 4R=R at k// from 2.1 to 2.7 μm−1 (Supplementary Fig. 9).Then, a derivative signal emerges at k// ≈ 0 μm−1 region, as clarified byArticle https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 4the integrated 4R=R at k// =0 ~ 0.33 μm−1 (Fig. 3a, b and the movie inSupplementaryMaterial). Sucha signal represents the blue-shift of LPBdue to the polariton-polariton repulsive interaction or the phase-space-filling effect of exciton (Supplementary Figs. 10, 11)51. The signalintensity gradually increases, reaches the maximum at ~6 ps, andslowly decays afterward (Fig. 3a). By fitting the evolution of the deri-vative signal with the rising function (y =A +B erf (t/τ), whereerfðzÞ= 2ffiffiffiπpR z0e�t2dt), a characteristic rising time (τ) of ~2.8 ± 0.4 ps isobtained (Fig. 3c)52. In contrast, for het@cavity−, the intensity of thederivative signal at k// ≈ 0 μm−1 region increases much faster andreaches the maximum at ~2 ps (Fig. 3d, e). With the same fittingmethod, a much shorter τ of ~1.3 ± 0.2 ps is obtained (Fig. 3f), whichapproximates to the τ of 1.2 ± 0.2 ps in MoS2@cavity (SupplementaryFig. 12). Therefore, het@cavity− demonstrates the ultrafast energytransfer processes accompanied by the highest enhancement factor(440) to date (Supplementary Table 2).Based on the preceding qualitative and quantitative analysis, wehave formulated the phase diagram that correlates the polaritonrelaxation efficiency and characteristic rising time with the Rabienergies Ω1 and Ω2, as shown in Fig. 4 (see Supplementary part III fordetails). Here we define a measure of efficiency as the sum of thephotonic part of LPB population in the steady-state, i.e., N τs� � �PNLj τs� �CLj only for low-k// points (τs denotes the time at steady state).In general,N τs� �monotonously increases with decreasingΩ2 (exceptforΩ2 ! 0) due to the enhanced interbranch scattering between UPBand MPB. For a specific Ω2, N τs� �depends on the trade-off betweenthe intra-branch scattering in MPB and the inter-branch scattering(MPB to LPB), hence a suitable Ω1 is needed for the optimalN τs� �asIntensity(arb. units)k (μm-1) k (μm-1) k (μm-1)0 ps 6 ps 55 psEnergy(meV)55 ps6 ps1.5 ps0 psEnergy (meV)ΔR/Ra bWS2@cavityhet@cavity−0 ps 2 ps 53 psd53 ps2 ps1 ps0 psk (μm-1)k (μm-1) k (μm-1) Energy (meV)ΔR/ReEnergy(meV)-0.080.08-0.080.08cfTime (ps)Time (ps)Intensity(arb. units)0.0 0.5 1.0 1.5 2.00.00.20.40.60.81.01.21.4raw datafitting0 1 2 3 4 5 60.00.20.40.60.81.0 raw datafitting0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.00.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.01850190019502000205018501900195020002050Fig. 3 | Time-resolved polariton relaxation dynamics. a–c k-space transient-reflectivity spectroscopy mapping (a), the integrated 4R=R in 0 ~ 0.33 μm−1 attypical time delays (b), and the intensity evolution of the derivative signal (c) ofWS2@cavity. d–f The corresponding data of het@cavity−. Note: The steady-statepolariton branches are shown in dashed black to guide the reading. The error barsin (c) and (f) correspond to the standard error of the data points.a bEfficiency (a.u.) Rising time (ps)WS @cavity2MoS�@cavityhet@cavity-het@cavity+het@cavity-het8@cavity WS @cavity2MoS @cavity2het60@cavity het2@node@cavityFig. 4 | Phase diagram of the polariton relaxation dynamics versus Rabi ener-gies. a, b Polariton relaxation efficiency (a) and characteristic rising time (b) withrespect to Rabi energies Ω1 and Ω2. The seven specific experimental cases aremarked. In a, the dashed-dotted curve indicates the optimal efficiency concerningΩ1 andΩ2, and the dashed blue boxes show regions whereΩ1,Ω2 ! 0. Note: In thedark area of b, the population in the low-k// region of LPB does not rise initially.Article https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 5marked by the dashed-dotted curve (Fig. 4a). For theN τs� �of specificcases, the het@cavity − , het@cavity+ and het60@cavity feature highefficiency. In contrast, N τs� �is low for Ω2 ! 0 or Ω1 ! 0 (such asMoS2@cavity, WS2@cavity, and het2@node@cavity), orΩ1 andΩ2 areboth large (such as het8@cavity), in agreement with the experimentalresults. Such an agreement offers a deep insight into the polariton-mediated energy transfer, which is inherently dominated by the Rabienergies.Different from theN τs� �, the characteristic rising time is closelyrelated to the scattering rate rather than the final polariton populationin the low-k// region of LPB (Fig. 4b). After a Gaussian effective pumpwith amplitudes fP1,P2g=50γ, the evolution of the photonic part of theLPB population, i.e.,N tð Þ � PNLj tð ÞCLj , is fitted by the rising function.For (Ω1,Ω2) located in the dark region, the negligible polariton fractionin the low-k// region of LPB prevents N tð Þ from growing. For Ω1≳8meV, the rising time monotonously decreases with increasing Ω1,which can be attributed to the dominant intra-branch scattering forthe flatter LPB dispersion. Although the larger Ω1 means the lesstransition from MPB to LPB, once the population relaxes to LPB, thestrong intra-branch scattering leads to the faster relaxation to low-kpoints. It is worth noting that the faster rising time might not end upwith a higher N τs� �, as evidenced by the comparison in Fig. 4a andFig. 4b. The dependence of the rising time onΩ2 is associated with thesupply of population that will later end up in the low-k// region of LPB.Smaller Ω2 results in more population relaxation from UPB to MPB,which allows faster growth of N tð Þ, as supported by the longer risingtime of WS2@cavity than het@cavity− and MoS2@cavity (Fig. 4b), inagreement with the experimental result (Fig. 3).DiscussionIn summary, we have successfully designed and constructed a planaroptical microcavity-confined MoS2/hBN/WS2 heterojunction, in whichthe middle insulating hBN prevents the charge transfer-induced PLquenching. Such a configuration realizes the strong coupling amongdonor exciton, acceptor exciton and cavity photon mode, whichresults in the unconventional energy transfer mechanism of polaritonrelaxation in 2D material heterojunctions for the first time. Conse-quently, we have brightened the MoS2 with the record-high energytransfer enhancement factor of ~440, which is two-order-of-magnitudehigher than the data reported to date. A short characteristic time of~1.3 ps is extracted for the ultrafast polariton relaxation dynamics,resulting from the significantly enhanced intra- and inter-branchexciton-exciton scattering to overcome the hot phonon bottleneckeffect. The formulated phase diagram in this study establishes thecorrelation between the polariton relaxation dynamics and Rabienergies, which deepens the understanding on the underlying physics.This studydrives the collaborativedevelopment of energy transfer andpolariton fields toward the new topics of dark exciton energy transfer,valley-polarized energy transfer, and high-brightness ultrafast polari-tonic light sources.MethodsSample fabricationThe bottom DBR was deposited on a sapphire substrate by using ane-beam evaporator (Cello, Ohmiker-50B), and was composed of 6.5alternating pairs of titanium dioxide (TiO2, n = 2.498) and silicondioxide (SiO2, n = 1.478). The hBN/MoS2/hBN/WS2 heterojunctionswere fabricated using a dry-transfer method with a polypropylenecarbonate (PPC) stamp. Monolayer WS2, monolayer MoS2, and thinhBN flakes were exfoliated onto silicon substrates with a 90-nm SiO2layer. A PPC stamp was used to pick up the flakes in sequence by ahome-built micro-transfer stage (Supplementary Fig. 1). Then, theheterojunction was released to the DBR substrate, immersed in acet-one overnight to remove the PPC, and annealed in a high vacuum(<10−6 mbar) at 200 °C for 2 hours. The top PMMA layer was spin-coated and the top silver mirror was deposited by a thermalevaporator.Two coupled oscillator modelSimilar to het@cavity−, the polariton dispersion in MoS2@cavity andWS2@cavity can be fitted with two coupled oscillator model,Ec θð Þ Ω=2Ω=2 Eex�  αcαex�  = EpolðθÞαcαex�  where Eex is the energy of exciton, Ω the corresponding Rabi splittingenergy. αc�� ��2, αex�� ��2 are the Hopfield coefficients describing the pho-tonic and excitonic weightings of the polaritons.Optical characterizationThe real-space steady-state PL and reflectivity spectra were conductedin a confocal spectrometer (Horiba Evolution 800) by using a CW laser(532 nm). K-space energy-resolved reflectivity and PL mapping weremeasured in a home-built setupwith the Fourier imaging configurationwith a high numerical aperture 100× microscope objective (NA =0.9).The PL mappings are excited by a 532 nm continuous-wave (CW) laserof ~100 μW. The emission from the microcavity was collected throughthe narrow entrance slit of the spectrometer (Horiba iHR550) andfinally onto the 2D charge-coupled device (CCD) array (Horiba,Symphony II). In the k-space transient-reflectivity spectroscopy mea-surement, the excitation pulsed laser was taken from a Ti:Sapphirelaser equipped with an ultrafast amplifier (Spectra-Physics) and acomputer-controlled Optical Parametric Amplifier (1 kHz repetitionrate, with a roughly 200 fs pulsewidth). The output beamwas split intotwopaths.Onebeamexcited the sample almost at normal incidence toserve as the pumpbeam. The secondbeamwent through amechanicaldelay stage (Newport, M-ILS 150CC DC Servo Linear Stage) and asapphire crystal to generate a delayed continuum light to serve as theprobe pulse. The spot size of the pump and probe beamswas around 2μm and 3 μm, respectively. PLE spectra were obtained with a super-continuum light acting as the excitation source which is coupled to amonochromator. The excitation intensity was kept below 10 μW. A655 nm long-pass filter was used to cut off the excitation photon thatbackscattered from the sample. All themeasurementswere carried outin a reflection configuration.Data availabilityAll data needed to evaluate the conclusions in the paper are present inthe paper and/or the SupplementaryMaterials. Additional data relatedto this paper may be requested from the authors.References1. Chhowalla, M., Jena, D. & Zhang, H. Two-dimensional semi-conductors for transistors. Nat. Rev. Mater. 1, 1–15 (2016).2. Liu, Y., Huang, Y. & Duan, X. Van der Waals integration beforeand beyond two-dimensional materials. Nature 567,323–333 (2019).3. Novoselov, K., Mishchenko, A., Carvalho, A. & Neto, A. C. 2Dmaterials and van der Waals heterostructures. Science 353,aac9439 (2016).4. Seyler, K. L. et al. Signatures of moire-trapped valley excitons inMoSe2/WSe2 heterobilayers. Nature 567, 66–70 (2019).5. Amani, M. et al. Near-unity photoluminescence quantum yield inMoS2. Science 350, 1065–1068 (2015).6. Wang, Z. et al. Giant photoluminescence enhancement in tungsten-diselenide–gold plasmonic hybrid structures. Nat. Commun. 7,1–8 (2016).7. Mak, K. F., Lee, C., Hone, J., Shan, J. & Heinz, T. F. Atomically thinMoS2: a new direct-gap semiconductor. Phys. Rev. Lett. 105,136805 (2010).Article https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 68. Kozawa, D. et al. Evidence for fast interlayer energy transfer inMoSe2/WS2 heterostructures. Nano Lett. 16, 4087–4093 (2016).9. Jin, C. et al. Ultrafast dynamics in van der Waals heterostructures.Nat. Nanotechnol. 13, 994–1003 (2018).10. Wu, L., Chen, Y., Zhou, H. & Zhu, H. Ultrafast energy transfer of bothbright and dark excitons in 2D van der Waals heterostructuresbeyond dipolar coupling. ACS Nano 13, 2341–2348 (2019).11. Hu, Z. et al. Interfacial charge and energy transfer in van der Waalsheterojunctions. InfoMat 4, e12290 (2022).12. Zhang, L. et al. Efficient and layer-dependent exciton pumpingacross atomically thin organic-inorganic type-I heterostructures.Adv. Mater. 30, e1803986 (2018).13. Hu, Z. et al. Trion-mediated Forster resonance energy transfer andoptical gating effect in WS2/hBN/MoSe2 heterojunction. ACS Nano14, 13470–13477 (2020).14. Meng, Y. et al. Electrical switching between exciton dissociation toexciton funneling in MoSe2/WS2 heterostructure.Nat. Commun. 11,2640 (2020).15. Coles, D. M. et al. Polariton-mediated energy transfer betweenorganic dyes in a strongly coupled optical microcavity. Nat. Mater.13, 712–719 (2014).16. Zhong, X. et al. Non-radiative energy transfer mediated byhybrid light-matter states. Angew. Chem. Int. Ed. 55,6202–6206 (2016).17. Zhong, X. et al. Energy transfer between spatially separatedentangled molecules. Angew. Chem. Int. Ed. 56,9034–9038 (2017).18. Lidzey, D. G., Bradley, D. D., Armitage, A., Walker, S. & Skolnick, M.Photon-mediated hybridization of Frenkel excitons in organicsemiconductor microcavities. Science 288, 1620–1623 (2000).19. Georgiou, K., Jayaprakash, R., Othonos, A. & Lidzey, D. G. Ultralong-range polariton-assisted energy transfer in organic microcavities.Angew. Chem. Int. Ed. 60, 16661–16667 (2021).20. Garcia-Vidal, F. J. & Feist, J. Long-distance operator for energytransfer. Science 357, 1357–1358 (2017).21. Liu, X. et al. Control of coherently coupled exciton polaritons inmonolayer tungsten disulphide. Phys. Rev. Lett. 119, 027403 (2017).22. Zhao, J., Zhao, W., Du, W., Su, R. & Xiong, Q. Dynamics of excitonenergy renormalization in monolayer transition metal disulfides.Nano Res 13, 1399–1405 (2020).23. Anton-Solanas, C. et al. Bosonic condensation of exciton-polaritonsin an atomically thin crystal. Nat. Mater. 20, 1233 (2021).24. Wurdack, M. et al. Motional narrowing, ballistic transport, andtrapping of room-temperature exciton polaritons in an atomically-thin semiconductor. Nat. Commun. 12, 5366 (2021).25. Zhao, J. et al. Ultralow threshold polariton condensate in a mono-layer semiconductor microcavity at room temperature. Nano. Lett.21, 3331–3339 (2021).26. Zhao, J. et al. Nonlinear polariton parametric emission in an atom-ically thin semiconductor based microcavity. Nat. Nanotechnol. 17,1–7 (2022).27. Zhao, J. et al. Exciton polariton interactions in Van der Waalssuperlattices at room temperature. Nat. Commun. 14, 1512 (2023).28. Chen, Y.-J., Cain, J. D., Stanev, T. K., Dravid, V. P. & Stern, N. P.Valley-polarized exciton–polaritons in amonolayer semiconductor.Nat. Photon. 11, 431–435 (2017).29. Sun, Z. et al. Optical control of room-temperature valley polaritons.Nat. Photon. 11, 491–496 (2017).30. Dufferwiel, S. et al. Valley-addressable polaritons in atomically thinsemiconductors. Nat. Photon. 11, 497–501 (2017).31. Gu, J., Chakraborty, B., Khatoniar, M. & Menon, V. M. A room-temperature polariton light-emitting diode based on monolayerWS2. Nat. Nanotechnol. 14, 1024–1028 (2019).32. Zhang, L. et al. VanderWaals heterostructurepolaritonswithmoire-induced nonlinearity. Nature 591, 61–65 (2021).33. Liu, X. et al. Manipulating charge and energy transfer between 2DAtomic layers via heterostructure engineering. Nano Lett. 20,5359–5366 (2020).34. Hong, X. et al. Ultrafast charge transfer in atomically thinMoS2/WS2heterostructures. Nat. Nanotechnol. 9, 682–686 (2014).35. Skolnick, M., Fisher, T. &Whittaker, D. Strong coupling phenomenain quantum microcavity structures. Semicond. Sci. Technol. 13,645 (1998).36. Agranovich, V. M., Litinskaia, M. & Lidzey, D. G. Cavity polaritons inmicrocavities containing disordered organic semiconductors. Phys.Rev. B 67, 085311 (2003).37. Slootsky, M., Liu, X., Menon, V. M. & Forrest, S. R. Room tem-perature Frenkel-Wannier-Mott hybridization of degenerateexcitons in a strongly coupled microcavity. Phys. Rev. Lett. 112,076401 (2014).38. Vladimirova, M., Kavokin, A. & Kaliteevski, M. Dispersion of bulkexciton polaritons in a semiconductor microcavity. Phys. Rev. B 54,14566 (1996).39. Kavokin, A. V., Baumberg, J. J., Malpuech, G. & Laussy, F. P. Micro-cavities. (Oxford University Press, 2017).40. Krisnanda, T. et al. Room temperature light‐mediated long‐rangecoupling of excitons in perovskites. Adv. Opt. Mater. 9,2001835 (2021).41. Coles, D. M., Grant, R. T., Lidzey, D. G., Clark, C. & Lagoudakis, P. G.Imaging the polariton relaxation bottleneck in strongly coupledorganic semiconductor microcavities. Phys. Rev. B 88,121303 (2013).42. Müller, M., Bleuse, J., Andre, R. & Ulmer-Tuffigo, H. Observation ofbottleneck effects on the photoluminescence from polaritons inII–VI microcavities. Phys. B: Condens. 272, 476–479 (1999).43. Amthor, M. et al. Optical bistability in electrically driven polaritoncondensates. Phys. Rev. B 91, 081404 (2015).44. Solnyshkov, D., Terças, H., Dini, K. & Malpuech, G. HybridBoltzmann–Gross-Pitaevskii theory of Bose-Einstein condensationand superfluidity in open driven-dissipative systems. Phys. Rev. A89, 033626 (2014).45. Zhang, Q. et al. Electric field modulation of 2D perovskite exci-tonics. J. Phys. Chem. Lett. 13, 7161–7169 (2022).46. Lyons, T. P. et al. Giant effective Zeeman splitting in a monolayersemiconductor realized by spin-selective strong light–matter cou-pling. Nat. Photon 16, 632–636 (2022).47. Tassone, F., Piermarocchi, C., Savona, V., Quattropani, A. &Schwendimann, P. Bottleneck effects in the relaxation and photo-luminescence of microcavity polaritons. Phys. Rev. B 56,7554 (1997).48. Tassone, F. & Yamamoto, Y. Exciton-exciton scattering dynamics ina semiconductor microcavity and stimulated scattering into polar-itons. Phys. Rev. B 59, 10830 (1999).49. Sun, D. et al. Observation of rapid exciton-exciton annihilation inmonolayer molybdenum disulfide. Nano Lett. 14,5625–5629 (2014).50. Yuan, L. & Huang, L. Exciton dynamics and annihilation in WS2 2Dsemiconductors. Nanoscale 7, 7402–7408 (2015).51. Gu, J. et al. Enhanced nonlinear interaction of polaritons via exci-tonic Rydberg states in monolayer WSe2. Nat. Commun. 12,2269 (2021).52. Giovanni, D. et al. Coherent spin and quasiparticle dynamics insolution-processed layered 2D lead halide perovskites. Adv. Sci. 5,1800664 (2018).AcknowledgementsZ.H. gratefully acknowledges the National Key R&D Program of China(Grant No. 2022YFA1204301). Q.X. gratefully acknowledges fundingsupport from the National Key Research and Development Program ofChina (Grant No. 2022YFA1204700), strong funding support from theArticle https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 7National Natural Science Foundation of China (Grant No.12250710126), and strong support from the Tsinghua UniversityInitiative Scientific Research Program. H.L. gratefully acknowledgesfunding support from the National Natural Science Foundation ofChina (Grant No. 92056204). S.G. acknowledges the funding supportfrom the National Natural Science Foundation of China (Grant No.12274034). D.S. and A.F. acknowledge the funding support from:“QuantumOptical Networks based on Exciton-polaritons”, (Q-ONE, N.101115575, HORIZON-EIC-2022-PATHFINDER CHALLENGES EU pro-ject), “National Quantum Science and Technology Institute” (NQSTI,N. PE0000023, PNRR MUR project), “Integrated Infrastructure Initia-tive in Photonic and Quantum Sciences” (I-PHOQS, N. IR0000016,PNRR MUR project). X.R.W. acknowledges support from AcademicResearch Fund Tier 2 (Grant No. MOE-T2EP50120-0006 and MOE-T2EP50220-0005) and Tier 3 (Grant No. MOE2018-T3-1-002) fromSingapore Ministry of Education. K.D., T.K., and T.C. H.L., were alsosupported by Tier 3 (Grant No. MOE2018-T3-1-002) from the Singa-pore Ministry of Education. K.W. and T.T. acknowledge support fromJSPS KAKENHI (Grant Numbers 19H05790, 20H00354, and 21H05233)and A3 Foresight by JSPS.Author contributionsZ.H. conceived the ideas and designed the experiments. Z.H. preparedthe microcavity samples with the help of J.Z. and Q.S. Z.H. performedthe optical spectroscopy with the help of J.Z., A.F., Y.C., H.L., Y.L., andR.S. T.K. and K.D. performed the theoretical calculation. T.K. mainlyperformed the coupled rate equation part. J.W. and G.E. performed thePLE measurement. K.W. and T.T. prepared the hBN crystal. Z.H., T.K.,A.F., K.D., X.R.W., S.G., D.S., T.C.H.L., andQ.X.wrote themanuscriptwithinput from all authors. Q.X. supervised the project.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-024-45554-y.Correspondence and requests for materials should be addressed toZehua Hu, Kevin Dini or Qihua Xiong.Peer review information Nature Communications thanks the anon-ymous reviewers for their contribution to the peer review of this work. Apeer review file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, aslong as you give appropriate credit to the original author(s) and thesource, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons licence, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons licence and your intendeduse is not permitted by statutory regulation or exceeds the permitteduse, you will need to obtain permission directly from the copyrightholder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 20241National Laboratory of Solid State Microstructures, School of Electronic Science and Engineering, and Collaborative Innovation Center of AdvancedMicrostructures, Nanjing University, Nanjing 210093, China. 2Centre for Quantum Technologies, National University of Singapore, Singapore 117543, Sin-gapore. 3CNR NANOTEC Institute of Nanotechnology, Lecce 73100, Italy. 4Division of Physics and Applied Physics, School of Physical and MathematicalSciences, Nanyang Technological University, Singapore 637371, Singapore. 5Beijing Academy of Quantum Information Sciences, Beijing 100193, P.R. China.6State Key Laboratory of Low-DimensionalQuantum, Department of Physics Physics, TsinghuaUniversity, Beijing 100084, P.R. China. 7Department of Physics,National University of Singapore, Singapore 117542, Singapore. 8Research Center for Functional Materials, National Institute forMaterials Science, 1-1 Namiki,Tsukuba 305-0044, Japan. 9International Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044,Japan. 10School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore. 11Frontier Science Center forQuantum Information, Beijing 100084, P.R. China. 12Collaborative Innovation Center of Quantum Matter, Beijing, P.R. China. 13These authors contributedequally: Zehua Hu, Tanjung Krisnanda, Antonio Fieramosca. e-mail: zehuahu@nju.edu.cn; kdini@ntu.edu.sg; qihua_xiong@tsinghua.edu.cnArticle https://doi.org/10.1038/s41467-024-45554-yNature Communications |         (2024) 15:1747 8https://doi.org/10.1038/s41467-024-45554-yhttp://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/mailto:zehuahu@nju.edu.cnmailto:kdini@ntu.edu.sgmailto:qihua_xiong@tsinghua.edu.cn Energy transfer driven brightening of MoS2�by ultrafast polariton relaxation in microcavity MoS2/hBN/WS2 heterostructures Results Sample fabrication and optical characterization Theoretical simulation Dynamics and phase diagram Discussion Methods Sample fabrication Two coupled oscillator�model Optical characterization Data availability References Acknowledgements Author contributions Competing interests Additional information