# Fileset

[s44214-025-00075-7.pdf](https://mdr.nims.go.jp/filesets/f9bbf0aa-ad9e-481e-924f-684bd0b19d8d/download)

## Creator

Feng Kai, Xiong Wang, Yiqin Xie, Yuhui Yang, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Hongyi Yu, Wang Yao, Xiaodong Cui

## Rights

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

## Other metadata

[Distinct moiré exciton dynamics in WS2/WSe2 heterostructure](https://mdr.nims.go.jp/datasets/e9a4784b-76a8-4d8e-aaa3-b0b09c555871)

## Fulltext

Distinct moiré exciton dynamics in WS2/WSe2 heterostructureKai et al. Quantum Frontiers             (2025) 4:2 https://doi.org/10.1007/s44214-025-00075-7O R I G I N A L A R T I C L E Open AccessDistinct moiré exciton dynamics inWS2/WSe2 heterostructureFeng Kai1, Xiong Wang1, Yiqin Xie1, Yuhui Yang1, Kenji Watanabe4, Takashi Taniguchi4, Hongyi Yu2,3,Wang Yao1 and Xiaodong Cui1*AbstractThis letter reports a time resolved pump-probe reflectance spectroscopic study on moiré excitons in a twistedmonolayer WS2/WSe2 heterostructure. By probing at the resonant energies of intralayer excitons, we observed theirdistinct temporal tracks under the influence of interlayer excitons, which we attribute to the discrepancy in spatialdistribution of the intralayer excitons in different layers. We also observed that intralayer moiré excitons in WSe2 layerdiffer at decay rate, which reflects different locations of Wannier-like and charge-transfer intralayer excitons in amoiré cell. We concluded that the interlayer moiré excitons form within a few picoseconds and have the lifetimeexceeding five nanoseconds. Our results provide insights into the nature of moiré excitons and the strain’s significantimpact on their behaviour in twisted heterostructures, which could have important implications for thedevelopment of novel optoelectronic devices.Keywords: Exciton dynamics, Moiré exciton, 2D semiconductor, van der Waals heterostructure1 IntroductionThe emergent van der Waal (vdW) heterostructures pro-vide a platform that allows manipulation of various de-grees of freedoms and creation of artificial lattices. Thedeliberate lattice mismatch between individual vdW layersengenders a periodic interference pattern, known as themoiré pattern. This tailored moiré pattern offers a greatflexibility of building artificial superlattices and has in-voked fantastic quantum states including superconduct-ing, fractional quantum anomalous Hall, strongly cor-related states, etc [1–11]. One of the intriguing topicsis the exciton physics in moiré superlattices, particularTMD moiré superlattices which feature direct bandgapin monolayers and rich degrees of freedoms. Predomi-nantly, TMD heterostructures display a type-II quantumwell band alignment, with conduction and valence bandedges residing in disparate layers. Consequently, photo-*Correspondence: xdcui@hku.hk1Physics Department and HK Institute of Quantum Science & Technology, TheUniversity of Hong Kong, Pokfulam road, Hong Kong, ChinaFull list of author information is available at the end of the articlecarriers, upon optical excitation, swiftly relax to the dis-parate band edges through interlayer charge transfer, cul-minating in the formation of interlayer excitons (IX). Thespatial separation in these structures renders the inter-layer excitons an extended lifetime, often reaching severalnanoseconds [12, 13] and ultralong valley lifetime around40 μs [14, 15]. This long-life significantly lowers the bar-riers towards quantum phase transitions. In moiré super-lattices these interlayer excitons modulated by the peri-odic moiré potentials display exotic features originatingfrom the unique dispersion in the superlattice. Besides themoiré potential from the heterostructure stacking, the in-plane strain in individual layer arising from the lattice mis-match also contributes to the moiré potential modulation[16–24]. Given the disparity in Young’s modulus and frac-ture strength [25], the strain within each heterostructurelayer varies in intensity. For interlayer excitons, this resultsin electrons and holes in opposing layers experiencing dis-tinct in-plane moiré potential modulations. These multi-faceted moiré potentials complicate the microscopic un-derstanding of moiré excitons. The pump-probe study hasflexibility to tune the probed energy, enabling the study© The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, whichpermits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate creditto the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. Theimages or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwisein a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is notpermitted by statutory regulation or exceeds the permitted use, 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/.https://doi.org/10.1007/s44214-025-00075-7https://crossmark.crossref.org/dialog/?doi=10.1007/s44214-025-00075-7&domain=pdfhttps://orcid.org/0000-0002-2013-8336mailto:xdcui@hku.hkhttp://creativecommons.org/licenses/by/4.0/Kai et al. Quantum Frontiers             (2025) 4:2 Page 2 of 8of interactions between IX and various excitons in themoiré superlattices, providing insightful microscopic un-derstanding. As yet, the pump-probe study of moiré exci-tons remain largely unexplored.The lack of pump-probe study of the interlayer moiré ex-citons partially results from the relatively weak oscillatorstrength of the interlayer excitons. The spatial separationreduces the electron-hole wave function overlap and weak-ens the oscillator strength of interlayer excitons. Conse-quently, interlayer exciton is usually invisible in the reflec-tion/absorption spectra which violates the prerequisite ofpump-probe measurement. In this letter, we report a time-resolved pump-probe reflection spectroscopic study onmoiré excitons of the twisted monolayer WSe2/WS2 het-erostructure. We probed the reflectance spectra at the res-onant energies of WSe2 and WS2 intralayer excitons andobserved the different carrier dynamic patterns. These dif-ferent dynamics reflect the coupling between intralayer ex-citons and interlayer excitons in a moiré cell, respectively.Our findings reveal a significant in-plane moiré strain inWSe2 layer and the difference in WSe2 and WS2 intralayerexcitons under the strain modulation.2 FabricationMonolayer WSe2 and WS2 were mechanically exfoli-ated from single crystal and were vertically stacked as aheterostructure with a twist angle of around 0.8◦. Theheterostructure was encapsulated single crystal hexag-onal boron nitride (h-BN) nanoflakes, then the wholestack was transferred onto a silicon substrate with 90 nmsilicon dioxide cap-layer (Fig. 1a). The twist angle be-tween the monolayer WS2 and WSe2 was identified bythe polarization-dependent second harmonic generation(SHG) measurement (Data shown in Fig. 1). The twoslightly mis-aligned six-fold SHG patterns were clearly ob-served (three folds is shown here), and the six-fold fittingrevealed the relative twist angle of 0.8°. The heterostruc-ture yields a stronger SHG intensity (Fig. 1b) than bothindividual monolayer WSe2 and WS2, indicating the het-erostructure has AA stacking configuration (R-stacking).3 Interlayer excitonFigure 1c shows the photoluminescence spectra collectedon the heterostructure and monolayer WSe2 at the basetemperature of 20K. The photoluminescence peaks (red)of neutral excitons at 1.734 eV and trions (charged ex-citon) at 1.698 eV of monolayer WSe2 fade in the het-erostructure as a consequence of interlayer charge transfer.This is consistent with the type-II band alignment (Fig. 1d)in which the minimum of conduction band is located atK (K’) valley of WS2 and the maximum of valence bandat WSe2 layer. Under optical excitation, most intralayerelectron-hole pairs are disassociated and drift to oppo-site layers, namely, holes transferring to WSe2 while elec-trons to WS2. This interlayer charge transfer takes place inaround hundreds of femtoseconds [26–30] which is a feworders of magnitude faster than the intralayer exciton life-time, consequently, quenches the PL of intralayer excitons.The band edge electron-hole pairs at different layers arebonded to form interlayer excitons, identified as the peaksaround 1.395 eV (Fig. 1c, black).4 Ladder-like interlayer exciton emissionTo further investigate the interlayer exciton in our device,we did a series of intensity-dependent photoluminescencemeasurements. Figure 2a presents photoluminescence ofinterlayer excitons at different excitation intensities. Dueto the interference of our CCD device at the specific wave-length, the raw data (grey lines) carries significant inter-ference patterns. To remove artificial device noise, we usemulti-peak Lorentz fitting to extract the PL spectra. Weobserved that there’s only one interlayer exciton emissionpeak (IX1) at 1.388 eV under low excitation intensity of0.1 μW/μm2 (Fig. 2a, black line). As the excitation inten-sity increases to 17 μW/μm2, the second interlayer emis-sion peak (IX2) arises at the energy 37 meV higher than IX1(blue line). Under low excitation intensity, the interlayerexcitons density, or the filling factor (the number of exci-tons per moiré unit cell), is low, and the exciton-excitoninteraction is negligible. As the excitation intensity in-creases, the interlayer excitons populates and the exciton-exciton interaction arises. When the filling factor is above1, namely, every moiré cell has one exciton and some moirécells has to accommodate two excitons, the exciton energyis lifted owing to the onsite dipole-dipole repulsion. Thisis well described by the exciton Hubbard model [31]. Ourphotoluminescence experiment indicates an onsite dipole-dipole repulsion energy (U) of 37 meV, which manifestsas the energy jump from IX1 to IX2 (Fig. 2b), qualitativelyconsistent with the previous reports [31]. With such un-derstanding, the thirdly and fourthly occupied moiré cellcorrespond to IX3 and IX4, with the energy jump of ΔE32 =29 meV and ΔE43 = 49 meV, shown in Fig. 2b. Besides theenergy jump between IXs, all the four peaks blue-shift withthe increasing excitation intensity from 0.1 to 362 μW.The blue shift originates from long-range dipolar repul-sion between interlayer excitons trapped in the neighbor-ing moiré cell [31].5 Reflection contrastWe then characterized the absorption by static reflectioncontrast spectroscopy. Figure 2c shows the background-free reflection contrast (RC) of the heterostructure deviceand monolayers. Here, the RC is obtained by R/R0 – Rbg ,where R and R0 are the reflection spectrum of the het-erostructure device and its hBN-only region, and Rbg isthe background in RC due to substrate interference. Wefound the A exciton absorption peak (1.727 eV, red line)of monolayer WSe2 splits into three peaks (with the ener-gies of 1.683 eV, 1.742 eV and 1.784 eV, labelled as I, II,Kai et al. Quantum Frontiers             (2025) 4:2 Page 3 of 8Figure 1 a, Optical microscopic image of the device. Heterostructure, monolayer WSe2 and WS2 region are outlined by orange, yellow and blackdash line, respectively. b, Polarization-resolved SHG of individual monolayers WSe2 (yellow), WS2 (green) and heterostructure (orange). The relativetwist angle is extracted to be 0.8° ± 0.5° by fitting. c, Photoluminescence of monolayer WSe2 (red) and WSe2/WS2 heterostructure device (black)under 2.087 eV (594 nm) excitation at intensity of 5 μW. The interlayer exciton peak appears at 1.395 eV, labelled as IX . d, Diagram of the electron andhole transfer in type-II band alignmentIII, respectively) in the heterostructure. These emergentthree peaks were previously observed and attributed to theunique moiré exciton feature in WSe2/WS2 heterostruc-ture [32]. It was interpreted with the continuum model thatthe emergent periodic moiré potential modifies the exci-tonic dispersion, splitting the edge into three sub bands,contributing to the observed three peaks [32–34].However, we also note that the WS2 layer, sharing themoiré pattern with WSe2, does not experience splitting.This could be attributed to the different in-plane strainpatterns in the two layers under the lattice reconstruc-tion [23]. It has been shown that, in a rigid moiré land-scape without the lattice reconstruction, the moiré poten-tial experienced by intralayer excitons in either layer isvery weak [35]. When structure relaxation is taken intoaccount, the resultant periodic strains can give rise to pro-nounced modulation of electronic band structures and theenergy landscape of the intralayer exciton [36]. Drivenby the stacking-dependent interlayer van der Waals inter-action, the WS2/WSe2 heterostructure, once fabricated,would experiences a structural reconstruction [23]. Giventhe WS2 has higher Young’s modulus (302 GPa) and frac-ture strength (47 GPa) than WSe2 (258 GPa and 38 GPa)[25], WSe2 and WS2 experience different extent of latticereconstruction and the in-plane strain modification. Thediscrepancy in intralayer exciton splitting demonstratesthat a stronger strain pattern is generated in the WSe2layer upon the lattice reconstruction, giving rise to thestrong moiré potential and miniband splitting for WSe2exciton, while the strain in WS2 layer is not strong enoughto split its intralayer exciton dispersion. We further inferthat the WS2 intralayer exciton remains largely similar tothat in pristine WS2 monolayer due to its strong mechan-ical strength.6 Pump probe spectroscopy on moiré superlatticeAfter characterising the effects of moiré landscape onthe intralayer excitons in our device from the photolu-Kai et al. Quantum Frontiers             (2025) 4:2 Page 4 of 8Figure 2 Photoluminescence and reflection contrast spectra of WSe2/WS2 moiré superlattice at 20K. a, Excitation intensity-dependent interlayerexciton photoluminescence: raw data (grey lines) and the corresponding Lorentz fitting curves (colour lines). The obtained four interlayer excitonpeaks, labelled as IX1 to IX4, are outlined by dashed black lines, respectively. b, Extracted IX peak energies vs excitation intensity from Fig. 2a. c,Background-free reflection contrast (RC) of the WSe2/WS2 device (top, black). In the bottom, RC of monolayer WSe2 (red) and WS2 (blue). Theintrinsic WSe2 A exciton at 1.727 eV (718 nm) splits into three n resonances in the heterostructure, labelled as I, II, III. Three resonances have energiesof 1.784 eV (695 nm), 1.742 eV (712 nm), 1.683 eV (737 nm), respectivelyminescence and reflection contrast spectra, we investi-gate the moiré exciton dynamics by pump probe spec-troscopy. Since the interlayer exciton has very weak os-cillator strength, the pump probe spectroscopy that basedon reflection change can’t capture it directly. Instead, weprobe around the resonant energies of the intralayer exci-tons of WSe2 and WS2. Here we pump the heterostructurewith a beam of 3.1 eV (400 nm) at 100 μW/μm2 that is en-ergetic enough to excite all low-energy transitions. The hotphoto-carriers, electron-hole pairs, thermally relax in thefirst hundreds of femtosecond [37, 38], followed by chargetransfer and interlayer exciton formation in around 1 ps[39]. Owing to the type-II band alignment in WSe2/WS2,the IX’s electrons (holes) reside in the conduction band ofWS2 layer (valence band of WSe2 layer), which share thesame band as the intralayer excitons of individual layers.By probing at the intralayer excitons, we can monitor thedynamics of the electron or hole components of moiré in-terlayer excitons, namely, we monitor the impact of pump-induced interlayer excitons on the intralayer exciton tran-sitions.7 Probe around resonance of WS2 intralayerexcitonsFigure 3a shows the pump-induced reflection change(dR/R) as a function of the delay time and the probe en-ergy across the range of the WS2’s intralayer exciton tran-sition. The transient reflectance of dR/R > 0 represented bybright-colored region implies that the sample reflectanceincreases under the increased pump excitation intensity,and vice versa. To analyze the dR/R map, we horizontallyslice the dR/R map at different delay time (Fig. 3b) to getthe profile of dR/R versus energy. We notice that the dR/Rprior to the pump-probe coincidence t = 0 shows a non-zero value, significantly higher than the background noise.As the laser pulse repeats with a time interval of 13.2 ns(76 MHz), the non-zero dR/R at t < 0 indicates that a sig-nificant portion of interlayer excitons survive longer than13.2 ns. Since the profile of dR/R vs energy is influenced bythe resonant energy shift and the suppression of neutralexciton (X0) and trion (X–) oscillator strength owing tostate filling. To quantitatively extract the energy shift, weappend the pump induced reflection change (dR) to thestatic reflection contrast (R/R0), constructing the dynamicreflection contrast at various time (Fig. 3c). With Lorentzfitting, we extract the evolution of the energy shifts for theabsorption peaks (Fig. 3d). The energy shifts of both reso-nances show flat decay pattern in the probe time rangingfrom t = 20 ps to 80 ps (Fig. 3d), which is another evidencefor interlayer excitons’ long lifetime. In contrast, the exci-ton lifetime in monolayer WS2 determined by the pump-probe spectroscopy is around 102 ps [40]. This long life-time probed at the resonant energy of A exciton doesn’treflect the dynamics of intralayer excitons in WS2 layer.Instead, this probe at intralayer exciton energy reflects thelifetime of the interlayer excitons as the interlayer exci-tons and delocalized intralayer excitons of WS2 layer sharethe same electron orbits. This provides an easy approachto address the dynamics of interlayer exciton with pump-probe spectroscopy.8 Probe around the resonance energy of WSe2intralayer excitonsWe probed the WSe2’s intralayer exciton transition toinvestigate the influence of moiré interlayer excitons onWSe2. We observed a negative dR/R region from 1.65 to1.68 eV (Fig. 4a, dark region) accompanied by a strong pos-itive region from 1.68 to 1.72 eV (Fig. 4a, bright region),a trend that is clearly visible in the cross-cutting diagramKai et al. Quantum Frontiers             (2025) 4:2 Page 5 of 8Figure 3 Pump probe spectroscopy for WS2 intralayer transitions. a, Pump-induced reflection change map (dR/R) as a function of delay and probeenergy, where R is the reflection spectra of twisted WSe2/WS2 heterostructure. b, dR/R vs Probe energy at –3, 0, 3, 80 ps. The neutral exciton (X0) andTrion (X–) is labelled with dash line for reference. c, Reflection contrast at different delays. d, Energy red shift vs delays for the neutral exciton (X0) andTrion (X–)(Fig. 4b). This can be explained by the red shift and sup-pression of moiré resonance I, as diagramed in Fig. 4c.Another positive region appears in the range from 1.73to 1.8 eV, covering both resonance II and III. To quantita-tively extract the energy shift and absorption suppression,we use the same method as we did in WS2 to get the reflec-tion contrast at various delays, (dR + R)/R0, as presentedin Fig. 4d, and then extract the energy shifts and the am-plitude suppression with Lorentz line shape fitting.We found that peak I and II red shift, and in contrast, thepeak III blue shifts upon pump (Fig. 4e). Here we considerseveral possible mechanisms for the energy shift: (1) Thebandgap renormalization [41, 42] caused by Coulombscreening of free carriers or excitons on Coulomb re-pulsion in lattice, resulting in electronic bandgap shrink.(2) The screening effect [43–45] caused by exciton or freecarrier’s Coulomb screening on exciton which reduces theexciton binding energy and makes the exciton resonant en-ergy blue shift. Since the moiré exciton III has larger Bohrradius (around 5 nm) than moiré exciton I (tightly bound)[36], exciton III is more sensitive to the dielectric envi-ronment and consequently experiences stronger charge-exciton and exciton-exciton screening effect. The signifi-cantly higher Coulomb screening on large Bohr radius ex-citons might be the reason for the blue shift. The sensitivityof large Bohr radius exciton to Coulomb screening is wellknown and one has utilized the energy shift of 2 s exci-ton with large Bohr radius (∼6 nm) as a sensor to monitormetal-insulator Mott transition in WSe2/WS2 moiré su-perlattice [7]. (3) The exciton-exciton interaction [12, 35],which includes a dipole-dipole interaction term and an ex-change interaction term. The dipole-dipole interaction be-tween intralayer and interlayer excitons is close to zero, asthe intralayer exciton has no static electric dipole. Mean-while, the exciton-exciton exchange interaction relies onwavefunction spatial overlap [31, 46]. Particularly, the in-tralayer moiré excitons could have sizable spatial overlapwith the interlayer moiré excitons when their hole con-Kai et al. Quantum Frontiers             (2025) 4:2 Page 6 of 8Figure 4 Pump probe spectroscopy for WSe2 intralayer transitions. a, Pump-induced reflection change map (dR/R) as a function of delay time andthe probe energy. b, dR/R vs Probe energy at 0, 3, 10, 80 ps. The moiré excitons I, II and III are labelled with dash line for reference. c, Schematic of thedR/R curve as a result of energy shift and suppression of absorption peak. d, Dynamics of the reflection contrast, (dR + R)/R0, at various delay times.The Lorentz line shape fitting (colour lines) gives the evolution of three moiré excitons. e, Extracted energy red shift as a function of delay time forthree moiré excitons. f, Extracted absorption amplitude, A(τ )/A0, as a function of delay time for three moiré excitons. A(τ ) is the absorptionamplitude at delay τ , A0 is a constant represents the amplitude before the pump excitation. The vertical axis is in log scale and red straight linespresents single exponential fitting. The lifetimes probed at the resonance energies of three excitons are estimated through single exponential fitting.g, Schematic of the Spatial distribution for interlayer exciton and intralayer moiré excitons [36, 47]stituents are both in WSe2. The lowest-energy interlayerexcitons generated by the pump beam would have theirmaximum intensity reside around RXh site (Fig. 4g); For in-tralayer moiré exciton I, its electron and hole both residearound Rhh site that has small spatial overlap with inter-layer excitons (Fig. 4g, moiré exciton I), so exchange inter-action is negligible here. By contrast, moiré exciton III hasits hole in RXh site, occupying the same site with the holeof interlayer exciton (Fig. 4g, moiré exciton I), where theexchange interaction can arise from the exchange of holes[35]. The exchange interaction between lowest-energy in-terlayer exciton and intralayer moiré exciton III tends toincrease their total energy [31, 46], which is one possiblereason for the blue shift of moiré exciton III. Our obser-vations suggest that the bandgap renormalization effectdominates for moiré exciton I and II, while exciton-excitonscreening and exchange interaction effects may dominatefor exciton III.Moiré exciton I, II and III’s absorption are all suppresseddue to state filling of the interlayer excitons, and their de-cay rate are approximately within the same orders of mag-nitude, with minor variations (Fig. 4f ). The moiré exciton I(black) shows a faster decay rate than the moiré exciton III(blue). To understand the difference in decay rate, we fol-Kai et al. Quantum Frontiers             (2025) 4:2 Page 7 of 8low the spatial overlap picture of intralayer and interlayerexcitons [36, 47]. In our understanding, the spatial overlapbetween the holes of intralayer and interlayer excitons de-termines the extent of absorption suppression. The morespatial overlap between the two holes of intralayer and in-terlayer excitons, the more suppression for the intralayertransition. After the pump excitation, the hot interlayerexcitons would thermalize and follow a thermal distribu-tion near the band edge within sub-picosecond [48]. In realspace picture, the lowest-energy interlayer exciton has itshole localized in the RXh site [47], which barely overlapswith intralayer moiré exciton I that has hole in the Rhh site,but overlap significantly with moiré exciton III that hasits hole also in RXh site. As for the hot interlayer excitonthat has extra kinetic energy, it tends to be more delocal-ized, contributing to overlap with the hole of exciton I inRhh site. As the assembly of interlayer excitons further cooldown, their spatial distribution gradually shrinks to the RXhsite, leading to a gradual decrease of the spatial overlapbetween the holes, accompanied by the recovery of moiréresonance I, which would account for the difference in de-cay rate.9 DiscussionWe observed a noticeable discrepancy between the dy-namics probed at the resonant energies of WS2 and WSe2intralayer excitons. The lifetime probed at resonant in-tralayer excitons of WS2 is longer than that of WSe2. Asthe photo carriers exist in the form of interlayer excitons,the probe at the intralayer exciton energy measures thewavefunction overlap between intralayer excitons (probe)and occupied interlayer excitons. Specifically, the datareflects the wavefunction overlap between the electron(hole) component of intralayer excitons in the WS2 (WSe2)layer and its counterpart in the interlayer excitons. Notethat the interlayer exciton has its electron and hole local-ized around the RXh site [47]. The photoluminescence andreflection contrast measurements (Fig. 2) indicate that dueto the in-plane periodic strain introduced by the lattice re-construction, the intralayer exciton in WSe2 layer splitsinto three moiré excitons (I, II and III). On the other hand,the intralayer exciton in WS2 layer is not much differentfrom that in a pristine layer, spreading uniformly acrossthe moiré cell. The reflection probed at intralayer exci-ton resonance in WS2 layer well captures the interlayerexciton lifetime. By contrast, the intralayer moiré excitonsin WSe2 layer are localized with distinct spatial profileswithin a moiré cell [36]. Especially, exciton I resides at RXhsite, away from the lowest-energy interlayer exciton in spa-tial distribution. For excited interlayer excitons carry extraenergies [46], they can have extended distributions andtherefore build a sizable spatial overlap with the hole com-ponent of the intralayer exciton I. As interlayer excitonsfurther cool down and their wave function intensities con-centrate around the RXh site, the overlap with intralayer ex-citons I gradually shrinks and it displays a fast decay rate.Therefore, the signal probed at intralayer exciton I energyprimarily reflects the cooling process of interlayer exci-tons. Intralayer exciton III, however, tends to reside at RXhsite and shares the similar hole component spatial distri-bution as the lowest-energy interlayer exciton. Therefore,the probe at intralayer exciton III shows a slower decay ratethan that at intralayer exciton I. Nonetheless, its decay pat-tern is still different from that probed at WS2 intralayer ex-citon. This could be attributed to the weak signal vs. noiseratio data at WSe2 intralayer exciton III owing to it weakoscillator strength. Or the hole components at interlayerexciton and WSe2 intralayer exciton III are located at dif-ferent sub-bands. This needs further investigation.Yet we cannot quantitatively extract the interlayer exci-ton lifetime in probing at WS2 intralayer exciton energy,for the nearly zero decay pattern in the detection timerange. The lower bound that we can determine is approx-imately from the exciton III’s decay pattern. Given excitonIII occupies the same site with lowest-energy interlayer ex-citons, exciton III’s lifetime reflects interlayer exciton life-time. Therefore, we infer that the interlayer exciton life-time exceeding 5 ns.10 ConclusionIn this work, we observed that a near-zero twisted WSe2/WS2 heterostructure exhibits spectroscopic features mod-ulated by a signature moiré potential: the ladder-like in-terlayer photoluminescence and the splitting of moiré in-tralayer excitons in WSe2 layer. The former reveals the on-site exciton-exciton repulsion between moiré interlayerexcitons localized at the same site, while the latter reflectsthe influence of moiré potential and in-plane strain mod-ification on exciton dispersion. The distinct splitting be-haviours in absorption spectra and the slow decay patternin the pump-probe spectra indicate that the intralayer ex-citon is delocalized in WS2 layer and the intralayer moiréexcitons in WSe2 layer tend to be localized at different sitesin the moiré cell, as a result of different in-plane strainand reconstruction. Our pump-probe data confirms thatthe moiré interlayer exciton experiences a few picosecondscooling-down time and a lifetime exceeding five nanosec-onds. Our study provides a comprehensive approach to in-vestigate the dynamics of moiré interlayer excitons.AcknowledgementsThe work was supported by the National Key R&D Program of China(2020YFA0309600), Guangdong-Hong Kong Joint Laboratory of QuantumMatter, the University Grants Committees/Research Grants Council of HongKong SAR (AoE/P-701/20, 17300520, 17301223) and the Theme ResearchT45-406/23-R, MOST 2023ZD0300600. K.W. and T.T. acknowledge support fromthe Elemental Strategy Initiative conducted by the MEXT, Japan (Grant NumberJPMXP0112101001) and JSPS KAKENHI (Grant Numbers 19H05790, 20H00354and 21H05233). H.Y. acknowledges support by NSFC under Grant No.Kai et al. Quantum Frontiers             (2025) 4:2 Page 8 of 812274477 and the Department of Science and Technology of GuangdongProvince in China (No. 2019QN01X061).Author contributionsX.D.C and K.F conceived the research. K.F, X.W, Y.Q.X and Y.H.Y fabricated vander Waals heterostructures. K.W. and T.T. grew hexagonal boron nitride crystals.K.F carried out optical measurements. X.D.C, K.F., X.W performed data analysisand interpreted the results. X.D.C, K.F., W.Y and H.Y.Y wrote the paper with inputfrom all authors. All authors discussed the results.Data availabilityThe authors confirm that the data supporting this study are available withinthe manuscript. Raw data are available upon the request.DeclarationsCompeting interestsThe authors declare no competing interests.Author details1Physics Department and HK Institute of Quantum Science & Technology, TheUniversity of Hong Kong, Pokfulam road, Hong Kong, China. 2GuangdongProvincial Key Laboratory of Quantum Metrology and Sensing & School ofPhysics and Astronomy, Sun Yat-sen University, Zhuhai Campus, Zhuhai,519082, China. 3State Key Laboratory of Optoelectronic Materials andTechnologies, Sun Yat-sen University, Guangzhou Campus, Guangzhou,510275, China. 4Advanced Materials Laboratory, National Institute for MaterialsScience, 1-1 Namiki, Tsukuba, 305-0044, Japan.Received: 6 September 2024 Revised: 12 December 2024Accepted: 13 January 2025References1. H. Park, et al., Observation of fractionally quantized anomalous Hall effect.Nature 622(7981), 74–79 (2023)2. Y. Cao, et al., Unconventional superconductivity in magic-angle graphenesuperlattices. Nature 556(7699), 43–50 (2018)3. Y. Cao, et al., Correlated insulator behaviour at half-filling in magic-anglegraphene superlattices. Nature 556(7699), 80–84 (2018)4. Z. Lian, et al., Valley-polarized excitonic Mott insulator in WS2/WSe2 moirésuperlattice. Nat. Phys. 20(1), 34–39 (2024)5. E.C. Regan, et al., Mott and generalized Wigner crystal states in WSe2/WS2moiré superlattices. Nature 579(7799), 359–363 (2020)6. Y. Tang, et al., Simulation of Hubbard model physics in WSe2/WS2 moirésuperlattices. Nature 579(7799), 353–358 (2020)7. Y. Xu, et al., Correlated insulating states at fractional fillings of moirésuperlattices. Nature 587(7833), 214–218 (2020)8. X. Huang, et al., Correlated insulating states at fractional fillings of theWS2/WSe2 moiré lattice. Nat. Phys. 17(6), 715–719 (2021)9. C. Jin, et al., Stripe phases in WSe2/WS2 moiré superlattices. Nat. Mater.20(7), 940–944 (2021)10. S. Miao, et al., Strong interaction between interlayer excitons andcorrelated electrons in WSe2/WS2 moiré superlattice. Nat. Commun. 12(1)(2021)11. R. Xiong, et al., Correlated insulator of excitons in WSe2/WS2 moirésuperlattices. Science 380(6647), 860–864 (2023)12. P. Rivera, et al., Observation of long-lived interlayer excitons in monolayerMoSe2–WSe2 heterostructures. Nat. Commun. 6(1), 6242 (2015)13. J. Kim, et al., Correlation-driven nonequilibrium exciton site transition in aWSe2/WS2 moiré supercell. Nat. Commun. 15(1) (2024)14. J. Kim, et al., Observation of ultralong valley lifetime in WSe2/MoS2heterostructures. Sci. Adv. 3(7), e1700518 (2017)15. C. Jin, et al., Imaging of pure spin-valley diffusion current in WS2-WSe2heterostructures. Science 360(6391), 893–896 (2018)16. L.J. McGilly, et al., Visualization of moiré superlattices. Nat. Nanotechnol.15(7), 580–584 (2020)17. J. Sung, et al., Broken mirror symmetry in excitonic response ofreconstructed domains in twisted MoSe2/MoSe2 bilayers. Nat.Nanotechnol. 15(9), 750–754 (2020)18. V.V. Enaldiev, et al., Stacking Domains and Dislocation Networks inMarginally Twisted Bilayers of Transition Metal Dichalcogenides. Phys. Rev.Lett. 124(20) (2020)19. A. Weston, et al., Atomic reconstruction in twisted bilayers of transitionmetal dichalcogenides. Nat. Nanotechnol. 15(7), 592–597 (2020)20. S. Shabani, et al., Deep moiré potentials in twisted transition metaldichalcogenide bilayers. Nat. Phys. 17(6), 720–725 (2021)21. S. Zhao, et al., Excitons in mesoscopically reconstructed moiréheterostructures. Nat. Nanotechnol. 18(6), 572–579 (2023)22. M.R. Rosenberger, et al., Twist angle-dependent atomic reconstructionand moiré patterns in transition metal dichalcogenide heterostructures.ACS Nano 14(4), 4550–4558 (2020)23. H. Li, et al., Imaging moiré flat bands in three-dimensional reconstructedWSe2/WS2 superlattices. Nat. Mater. 20(7), 945–950 (2021)24. D.A. Cosma, et al., Moiré pattern as a magnifying glass for strain anddislocations in van der Waals heterostructures. Faraday Discuss. (2014)25. A. Falin, et al., Mechanical properties of atomically thin tungstendichalcogenides: WS2, WSe2, and WTe2. ACS Nano 15(2), 2600–2610(2021)26. H. Zhu, et al., Interfacial charge transfer circumventing momentummismatch at two-dimensional van der Waals heterojunctions. Nano Lett.17(6), 3591–3598 (2017)27. C. Jin, et al., Ultrafast dynamics in van der Waals heterostructures. Nat.Nanotechnol. 13(11), 994–1003 (2018)28. X. Hong, et al., Ultrafast charge transfer in atomically thin MoS2/WS2heterostructures. Nat. Nanotechnol. 9(9), 682–686 (2014)29. F. Ceballos, et al., Ultrafast charge separation and indirect excitonformation in a MoS2–MoSe2 van der Waals heterostructure. ACS Nano8(12), 12717–12724 (2014)30. J.E. Zimmermann, et al., Ultrafast charge-transfer dynamics in twistedMoS<sub>2</sub>/WSe<sub>2</sub> heterostructures. ACS Nano15(9), 14725–14731 (2021)31. H. Park, et al., Dipole ladders with large Hubbard interaction in a moiréexciton lattice. Nat. Phys. 19(9), 1286–1292 (2023)32. C. Jin, et al., Observation of moiré excitons in WSe2/WS2 heterostructuresuperlattices. Nature 567(7746), 76–80 (2019)33. F. Wu, T. Lovorn, A.H. Macdonald, Topological Exciton Bands in MoiréHeterojunctions. Phys. Rev. Lett. 118(14) (2017)34. S. Brem, et al., Tunable phases of moiré excitons in van der Waalsheterostructures. Nano Lett. 20(12), 8534–8540 (2020)35. H. Yu, et al., Moiré excitons: from programmable quantum emitter arrays tospin-orbit–coupled artificial lattices. Sci. Adv. 3(11), e170169636. M.H. Naik, et al., Intralayer charge-transfer moiré excitons in van der Waalssuperlattices. Nature 609(7925), 52–57 (2022)37. Z. Nie, et al., Ultrafast carrier thermalization and cooling dynamics infew-layer MoS2. ACS Nano 8(10), 10931–10940 (2014)38. T. Yan, et al., Probing the exciton k-space dynamics in monolayer tungstendiselenides. 2D Mater. 6(2), 025035 (2019)39. V.R. Policht, et al., Time-domain observation of interlayer exciton formationand thermalization in a MoSe2/WSe2 heterostructure. Nat. Commun. 14(1)(2023)40. Z.E. Eroglu, et al., Ultrafast dynamics of exciton formation and decay intwo-dimensional tungsten disulfide (2D-WS(2)) monolayers. Phys. Chem.Chem. Phys. 22(30), 17385–17393 (2020)41. P.D. Cunningham, et al., Photoinduced bandgap renormalization andexciton binding energy reduction in WS2. ACS Nano 11(12), 12601–12608(2017)42. E.A.A. Pogna, et al., Photo-induced bandgap renormalization governs theultrafast response of single-layer MoS2. ACS Nano 10(1), 1182–1188 (2016)43. G. Aivazian, et al., Many-body effects in nonlinear optical responses of 2Dlayered semiconductors. 2D Mater. 4(2), 025024 (2017)44. K. Xiao, et al., Exciton-exciton interaction in monolayer MoSe2 frommutualscreening of Coulomb binding. ACS Nano 18(46), 31869–31876 (2024)45. K. Xiao, et al., Exciton-exciton Interaction in Monolayer MoSe2 from MutualScreening of Coulomb Binding (2023). https://doi.org/10.48550/arXiv.2308.14362. arXiv:2308.1436246. P. Rivera, et al., Valley-polarized exciton dynamics in a 2D semiconductorheterostructure. Science 351(6274), 688–691 (2016)47. X. Wang, et al., Intercell moiré exciton complexes in electron lattices. Nat.Mater. 22(5), 599–604 (2023)48. H. Chen, et al., Ultrafast formation of interlayer hot excitons in atomicallythin MoS2/WS2 heterostructures. Nat. Commun. 7(1), 12512 (2016)Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.https://doi.org/10.48550/arXiv.2308.14362https://doi.org/10.48550/arXiv.2308.14362https://arxiv.org/abs/2308.14362 Distinct moire exciton dynamics in WS2/WSe2 heterostructure Abstract Keywords Introduction Fabrication Interlayer exciton Ladder-like interlayer exciton emission Reflection contrast Pump probe spectroscopy on moiré superlattice Probe around resonance of WS2 intralayer excitons Probe around the resonance energy of WSe2 intralayer excitons Discussion Conclusion References