# Fileset

[s41467-024-45709-x.pdf](https://mdr.nims.go.jp/filesets/2691e10e-2791-4cb9-b001-0a421c09dd92/download)

## Creator

Dongyang Yang, Jing Liang, Jingda Wu, Yunhuan Xiao, Jerry I. Dadap, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Ziliang Ye

## Rights

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

## Other metadata

[Non-volatile electrical polarization switching via domain wall release in 3R-MoS2 bilayer](https://mdr.nims.go.jp/datasets/0d538563-042a-4e90-9d2e-d9c7f55d2ea4)

## Fulltext

Non-volatile electrical polarization switching via domain wall release in 3R-MoS2 bilayerArticle https://doi.org/10.1038/s41467-024-45709-xNon-volatile electrical polarization switchingvia domain wall release in 3R-MoS2 bilayerDongyang Yang1,2,5, Jing Liang 1,2,5, Jingda Wu 1,2,5, Yunhuan Xiao1,2,Jerry I. Dadap 1,2, Kenji Watanabe 3, Takashi Taniguchi 4 & Ziliang Ye 1,2Understanding the nature of sliding ferroelectricity is of fundamental impor-tance for the discovery and application of two-dimensional ferroelectricmaterials. In this work, we investigate the phenomenon of switchable polar-ization in a bilayer MoS2 with natural rhombohedral stacking, where thespontaneous polarization is coupled with excitonic effects through asym-metric interlayer coupling. Using optical spectroscopy and imaging techni-ques, we observe how a released domain wall switches the polarization of alarge single domain. Our results highlight the importance of domain walls inthe polarization switching of non-twisted rhombohedral transition metaldichalcogenides and open new opportunities for the non-volatile control oftheir optical response.The rise of two-dimensional (2D) ferroelectric materials has offeredmany new opportunities for developing novel nano-electronic andoptoelectronic applications1, such as non-volatile memories2,3, high-performance photodetectors4–6, and integrated ferroelectrictransistors7–9. Recently, an interfacial phenomenon has been dis-covered in artificial stacks of two layers of non-polar 2Dmaterials withmarginal twists10–12, where a network of domains with alternatingstacking configurations arises out of the atomic reconstruction. Thesedomains exhibit spontaneous electrical polarization due to an asym-metric inter-layer coupling at the interface and the net polarization canaccumulate over multiple layers13–17. Under an external electric field,atoms at the domain wall (DW) slide perpendicular to an externalfield18–20, in contrast to conventional ferroelectricmaterials. As a result,the domains of favorable polarization expand while the others con-tract, leading to a switch of the total polarization. This phenomenon,referred to as sliding ferroelectricity21,22, has been observed in a widerange of van der Waals materials23–25, greatly broadening the family of2D ferroelectrics. Rhombohedral stacking also exists naturally in che-mically synthesized single crystals, such as rhombohedral molybde-num disulfide (3R-MoS2), which exhibits a large polarization-inducedspontaneous photovoltaic effect4. However, the polarization switchingin these crystals proves to be more challenging compared to artificialstacks16,23,26.Here we show the challenge lies in the underlying switchingmechanism, that is, the release of pre-existing domain walls in singlecrystalline MoS2 flakes. By optically mapping the polarization dis-tribution and its variation during the switching, we reveal that a freelypropagating DW can switch the polarization of a large single domain.The polarization switch is non-volatile, as DWs become localized bypinning centers. In contrast to DW networks in artificial stacks, theDWs in 3R-MoS2 can be released from pinning centers and sweepacross nearly the entire flake under a sufficiently strong externalelectrical field. Our findings highlight the crucial role of DWs in slidingferroelectricity, suggesting a promising pathway for 3R-MoS2 to serveas fundamental building blocks for programmable optoelectronicdevices27.ResultsThe polarization switching in our experiments is achieved by applyingan out-of-plane electric field and monitored by leveraging the stronglight-matter interaction of MoS228,29. As illustrated in Fig. 1a, a bilayer3R-MoS2 is encapsulated in a dual-gated device, which allows us toapply an electric field without doping. The bilayer 3R-MoS2 has twostacking configurations: If the Mo atom in the top layer sits on top ofthe sulfide atom in the bottom layer, we define it as AB stacking, andthe opposite structure is termed BA stacking. Previously, we haveReceived: 26 June 2023Accepted: 2 February 2024Check for updates1Department of Physics and Astronomy, The University of British Columbia, Vancouver, BC, Canada. 2Quantum Matter Institute, The University of BritishColumbia, Vancouver, BC, Canada. 3Research Center for Functional Materials, National Institute for Materials Science, Tsukuba, Japan. 4International Centerfor Materials Nanoarchitectonics, National Institute for Materials Science, Tsukuba, Japan. 5These authors contributed equally: Dongyang Yang, Jing Liang,and Jingda Wu. e-mail: zlye@phas.ubc.caNature Communications |         (2024) 15:1389 11234567890():,;1234567890():,;http://orcid.org/0000-0001-6348-2068http://orcid.org/0000-0001-6348-2068http://orcid.org/0000-0001-6348-2068http://orcid.org/0000-0001-6348-2068http://orcid.org/0000-0001-6348-2068http://orcid.org/0000-0002-6783-8719http://orcid.org/0000-0002-6783-8719http://orcid.org/0000-0002-6783-8719http://orcid.org/0000-0002-6783-8719http://orcid.org/0000-0002-6783-8719http://orcid.org/0000-0001-5100-9396http://orcid.org/0000-0001-5100-9396http://orcid.org/0000-0001-5100-9396http://orcid.org/0000-0001-5100-9396http://orcid.org/0000-0001-5100-9396http://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/0000-0001-8314-6977http://orcid.org/0000-0001-8314-6977http://orcid.org/0000-0001-8314-6977http://orcid.org/0000-0001-8314-6977http://orcid.org/0000-0001-8314-6977http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45709-x&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45709-x&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45709-x&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45709-x&domain=pdfmailto:zlye@phas.ubc.cashown that the interlayer potential arising from the polarization givesrise to a finite energy offset in both conduction and valence bands at Kpoint, leading to an effective type-II band alignment (Fig. 1b)16. In sucha band structure, optically excited electrons rapidly relax to the con-duction band edge at the K point, which is localized in one of the twolayers, while holes transfer to the valence band edge at the Γ point,which is delocalized between two layers, forming a momentum-indirect interlayer exciton. Depending on the stacking configuration,the interlayer exciton exhibits an out-of-plane electric dipole momentalong either the upward (AB) or downward (BA) direction30,31.The dipole moment of interlayer excitons can be measuredthrough quantum Stark shift of their photoluminescence (PL) peaksexcited by a continuous-wave laser (2.33 eV). At zeroelectricfield and atemperature of 4 K, four distinct PL peaks are observed in the1.4–1.6 eV range. Two of them have been attributed to phonon side-bands of the Γ-K transition32, while the rest are likely originated fromtheir trion counterparts16. Under a finite electric field, all four peaksshift toward the same direction, as a result of their out-of-plane dipolemoments.When adownward (negative) electricfield is applied to aAB-stacked domain, the field is parallel to the dipole moment and thepeaks undergo a redshift (Fig. 1c). Thepeakswill blueshift if the electricfield is anti-parallel to the dipolemoment. Consequently, bymeasuringthe slope of the Stark shift, we can determine the dipole momentdirection and infer its corresponding stacking order, thus identifyingthe moment of switch.We first study the quantum Stark shift of an AB-stacked bilayer(Fig. 2a). Within a small field range (−0.068 V/nm< Eex < 0.085 V/nm),all four peaks shift positively with the field, confirming a positivedipole moment. The slope of the Stark shift corresponds to a dipolemoment of about 0.3 e⋅nm, which varies slightly between excitons andtrions (μ1-μ4 in Table 1 of SI). Such a variation can result from thedifference between interlayer excitons and interlayer trions16. Whenthe field exceeds 0.085 V/nm, the differential slope of the Stark shiftchanges from positive to negative. This change occurs because thelarge positive external field (Eex) overcomes the built-in depolarizationfield (Edep) and reverses the energyorder of conductionbandsbetweenthe two layers. As a result, optically excited electrons relax to the otherlayer, reversing the direction of the dipole moment of the interlayerexciton. Such a dipole moment switch has been observed in artificiallystacked bilayers30 — it happens when the electric field is parallel to thepolarization but does not represent a change in the stacking order.The signature of stacking order switching is clearly observed at alarge negative external field (Ec=−0.068 V/nm). As the field is con-tinuously scanned towards the negative range, an abrupt change isacVtgVbg3R-MoS2hBNGrAB BA 1.3 1.4 1.5 1.6Photon Energy (eV)00.10.20.30.40.5 PL Intensity (arb. units) IT2IX1IX2IT1Eex=0 V/nmEex=-0.04 V/nmIX1IX2IT1IT2DWbK Г K-+-+AB BA +-ωℏtopbottom- -AB (P>0) BA (P<0)DWSMoFig. 1 | Coupling interlayer excitons with stacking orders. a Left panel: sche-matic of a dual-gated device with a 3R-MoS2 bilayer composed of both AB and BAdomainswith adomainwall located at the interface.Right panel: side and top viewsof atomic structures in AB (MotopSbottom) and BA (StopMobottom) stacking config-urations. The positive polarization direction is defined as vertical upward. b Leftpanel: the electronic band structure of an AB-stacked 3R-MoS2 bilayer. The red andblue lines at the Kpoint represent the conduction and valencebands in the top andbottom layers, respectively. Themomentum-indirectΓ-K excitons are composedoflocalized electrons (orange) atKpoint and layer-sharedholes (green) at the Γpoint.The wavy line represents the photoluminescence from the Γ-K excitons. Rightpanel: the real-space charge distribution of the Γ-K excitons. c Normalized PLspectrum of AB-stacked 3R-MoS2 bilayer with zero (black) and –0.04 V/nm (blue)external electric field. The former is shifted upward by 0.1 for clarity. Two phononreplicas of Γ-K excitons are labeled as IX1 and IX2 while their trion counterparts aredenoted as IT1 and IT2, respectively.Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 2observed in the interlayer trion peak, characterized by a much stron-ger amplitude and a blueshift of tens ofmeV.We assign themajor peakas an interlayer trion since it has a similar Stark shift slope as IT2 priorto the switch (Fig. 2b). A comparable shift can be found for the twointerlayer excitons, although their emission intensity becomes veryweak after the switch. This sudden change in the Stark shift, involvingthe same dipole moment and external field, implies a change in thedepolarization field, indicating a switch of stacking order from AB toBA. The sudden change in the trion emission amplitude may also berelated to the stacking order change, as the free carrier distributesdifferently in different structures. With a fitted blueshift of 38(3) meVfor the IT2 peak,we conclude aΔEdepof 0.17(2) V/nm, corresponding toa Edep ≈0.09(1) V/nm and an interlayer potential of 58(7) meV, con-sistent with the previous observations16,23,25.The switch in stacking order is non-volatile and shows hystereticbehavior. After the stacking order is switched from AB to BA, weobserve that both the Stark shift slope near zero field and the depo-larization compensation field become negative, consistent with the BAstacking order (Fig. 2c). This switch remains persistent up to roomtemperature, as we will discuss later. When we apply a large positivefield Eex (0.073 V/nm), the stacking order and associated polarizationreverses back to AB, as indicated by another abrupt change in the peakposition. The hysteresis becomes more apparent if we plot the dipolemoment versus the training electric field Etr (Fig. 2d). To train thepolarization, we initially apply a training field for 0.5 s. Due to thelinear-stark shift under an external field less than 0.07 V/nm, theexciton dipole moment is determined by normalizing the PL peakenergy shift to the small field applied. The measured hysteresis loophas a rectangular shape, exhibiting two opposite dipole momentsalong with two coercive fields matching the abrupt features in thespectra.The coercive field is an important parameter for ferroelectricmaterials and can reveal the mechanism of the polarization switching.The coercive field reported here is asymmetric in the negative andpositive field directions. Since our optical sensing approach provides adiffraction-limited spatial resolution, we can measure the coercivefield at different locations, where we also find different coercive fields.Moreover, in another device, we observe an AB-to-BA transition atEc = −0.21 V/nm (Fig. S2 of SI), which is nearly two times larger. Suchsignificant variation in Ec suggests that the switching field is not anintrinsic property of 3R-MoS2 crystal, e.g., for nucleating newdomains.The ferroelectric polarization switching in a single-domain 3R-MoS2bilayer is challenging likely due to the large domain nucleationenergy21. The three-fold symmetry in the structure leads to three-0.2 -0.1 0 0.11.41.451.51.55Peak Energy (eV)Eex (V/nm)IX1IX2IT1IT22μ4Edep-0.15 -0.1 -0.05 0 0.05 0.1Eex (V/nm)Photon Energy (eV)Ec-0.15 -0.1 -0.05 0 0.05 0.1Eex (V/nm)1.41.451.51.55Photon Energy (eV)a bdc Ec0 0.4PL (arb. units)Dipole (e  nm)●AB-0.2 0 0.2-0.400.4Etr (V/nm)BA1.41.451.51.550 0.4PL (arb. units)Fig. 2 | Observing polarization switching through quantum Stark shift. a.Photoluminescence spectra of Γ -K transitions as a function of the external electricfield (Eex). Eex is swept towards the negative direction. IX1, IX2, IT1, and IT2 arelabeled by green, yellow, purple, and red arrows. At the coercive field Ec=0.068 V/nm, the stacking order switches from AB to BA at the focus spot. b. The fitted peakenergy is plotted as a function of Eex. The peak shift of IT2 before and after theswitch corresponds to the change of the depolarization field times the dipolemoment (μ4). c. PL spectra of Γ -K transitions versus Eex when the stacking orderswitches back to AB. The domain switching happens at Ec=0.073 V/nm. d. Thedipole moment of IT2 as a function of training electric field (Etr) in the forward(blue) and backward (red) scan directions, showing a hysteresis behavior.Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 3equivalent directions for the sliding to occur, which makes theswitching nondeterministic33. Therefore, we attribute the observedswitching behavior to the propagation of a pre-existing domain wall(DW),with the coercive field corresponding to the pinning potential ofa pinning center that localizes the DW.The domain wall in 3R-MoS2 is a ~ 10 nm wide region connectingtwo polarization domains, where the stacking order smoothly transi-tions from one to the other34,35. DWs are rare in chemically synthesizedsingle crystals, but can be found in mechanically exfoliated flakes,where shear strain can induce avalanches of interlayer sliding36. TheseDWs are usually trapped by pinning centers like defects, bubbles, oredges of the sample. When an external electric field is applied, onestacking order becomes more energetically favorable than the other,providing a driving force for the DW to propagate. In our device, as Eexbecomes larger than Ec, the free energy difference is expected toovercome the local pinning potential and release a DW. The sharptransition in Fig. 2d suggests that the DW can propagate throughoutthe focus spotwithout getting pinned. The variation in Ec is therefore aresult of the random distribution of pinning potentials. The asym-metric coercive fields in different switching directions arise from thedifference in pinning centers between the initial and final states of aswitch.Such domain-wall-propagation based polarization switching pic-ture is further supported by the polarization mapping at room tem-perature and ambient conditions. As shown in Fig. 3a, although the finepeak splitting is no longer resolvable, the overall Stark shift is stilldistinguishable at room temperature under Eex=−0.04 V/nm. The PLpeaks in AB and BA domains shift in opposite directions. Here we usethe Stark shift as a qualitative representation of the dipole moment,which yields a hysteresis loop similar to the one observed at low-temperature (Fig. 3b). No obvious temperature dependence of thecoercive fields is observed up to the room temperature, indicating thepinning potential being larger than the thermal energy. More impor-tantly, the long-term stability at ambient conditions allows us to mapthe real-space distribution of the polarization, through which weobserve different intermediate DW positions during the switching.The polarization mapping is performed in a way similar to thehysteresis loop measurement. We first apply a large training electricfield (Etr) and then map the Stark shift across the entire flake with adiffraction-limited focus spot with a small field (Eex=−0.04 V/nm). Inthe upper-left panel of Fig. 3c, we first apply a positive training field(Eex=0.1 V/nm) to prepare the sample in a single AB-stacked domain. Asa result, the entire sample exhibits a redshift, which supports ourstacking order assignment. Subsequently, we apply a negative trainingfield (Eex=−0.08 V/nm), leading to polarization switching in over half ofthe device, including the device center, which corresponds to the spotprobed in Fig. 2. Finally, an even larger negative field is applied (Eex=−0.1 V/nm), resulting in a further expansion of the BA domain. Bycomparing the two states, we can conclude that polarization switchingis achieved through the propagation of domain walls, one of which is1.3 1.4 1.5 1.6Photon Energy (eV)00.20.40.60.81PL Intensity (arb. units)ABBAEex=-0.04 V/nm1.7abcPeak shift (meV)-25 25Etr=+0.10 V/nmEtr=-0.10 V/nmEtr=+0.05 V/nmEtr=-0.08 V/nm-0.2 0 0.21.4651.471.475Peak Energy (eV)Etr (V/nm)ABBAFig. 3 | Non-volatile polarization switching enabled by domain wall release.a Interlayer PL spectra of AB and BA stacked domains at room temperature underan external electric field of Eex=-0.04 V/nm. b Room temperature hysteresis loop.Blue and red arrows denote the forward and backward scan direction of thetraining field. c Real-space mapping of the Stark shift after four different trainingfields (Etr) are applied. The negative shift corresponds to an AB-stacked domainwhile the positive shift indicates a BA-stacked domain. The dashed circle in theupper-left panel outlines a bubble area where the initial domain walls (DWs) arelikely pinned. One DW in the lower-left panel is labeled by a magenta dashed lineand its local propagation directions are indicated by the arrows. Scale bar: 5μm.Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 4highlighted in magenta in the lower-left panel of Fig. 3c. When theelectric field exceeds the local pinning field, the DW is released, whichthen moves westward, and finally becomes pinned again by strongerpinning potentials near the edge of the flake. The intermediate andfinal positions of DWs are non-volatile, as the observation is indepen-dent of the Stark shift field.To study the hysteresis effect in DW, we further apply a positivetraining field to reduce the BA domain. As shown in the upper-rightpanel of Fig. 3c, the DW moves to a new location when we apply atraining field Etr=0.05 V/nm. The newDWposition is different from theprevious intermediate state, indicating that the depinning and pinningprocess strongly depends on the history of training process. When weapply an even larger positive field (Etr=0.1 V/nm), the BA domainshrinks to a size smaller than the detection limit, and the DWs likelymove into a bubble area in the upper middle area of the flake, whichcannot be accessed by our optical probe. Hence, we complete thedemonstration of a reversal of polarization switching in a natural 3R-MoS2 bilayer over a ~ 100μm2 area. Such a series of spatial distribu-tions of polarization domains also agree with the hysteresis loopmeasured at the center spot. We note there are potentially moreintermediate states between our applied training fields— the efficiencyof the PL probe limits us on the few conditions as presented. Overall,we find 4 out of 9 fabricated devices are switchable. We think therepeatability is limited by the probability of having pre-existingdomain walls36 and large domain nucleation energy in 3R-MoS2 bilay-ers. The existence of domain walls in another switchable device(sample 4) is confirmed by Electrostatic Force Microscopy (EFM)imaging prior to the device encapsulation (Fig. S3). Compared withDWs in artificially stacked homo-bilayer, it is clear that our DWs are notalways attached to certain pinning centers and can propagate throughthe entire flake under a sufficiently large field.In addition, we report that the polarization switching in 3R-MoS2 bilayers can be optically observed not only in the Stark shift ofinterlayer exciton but also in the intensity of intralayer exciton.Despite being an indirect band gap semiconductor, the bilayer 3R-MoS2 exhibits hot PL from direct band gaps in K valleys within eachindividual layer (Fig. 4). As shown in Fig. 4a, two distinctive PL peakswith an energy separation of 11 meV near 1.9 eV are observedbecause the nonequivalent local environment of Mo atom induces asmall band gap difference between the top and bottom layers4,16.Based on the optically determined BA stacking order, the higherenergy peak (Xh) originates from the bottom layer while the lowerab--++--++ωℏ--++ωℏtop bottomωℏ-0.15 -0.1 -0.05 0 0.05 0.1Eex (V/nm)1.851.91.952Photon Energy (eV)XhXl1.00PL (arb. units)EcEdepEex > EcEex= 0Eex< EdepFig. 4 | Observing polarization switching through intralayer excitons. a PLspectra of intralayer excitons as a function of external electric field (Eex). Xh and Xlare attributed to the excitons in the bottom and top layers, respectively. The whitedashed lines indicate when the intensity ratio between the two peaks changes.b Schematics illustrating the interlayer charge transfer process that accounts forthe intensity ratio variation between Xh and Xl. The red and blue lines represent theband gaps in the top and bottom layers, respectively. Left panel: at a large negativefield, the interlayer potential is compensated, causing the conduction band in thebottom layer to behigher than that in the top layer. The photoluminescenceof Xh istherefore suppressed. Middle panel: with zero external field, the PL of Xl is quen-ched because the electrons migrate from the higher conduction band in the toplayer to the lower one in the bottom layer. Right panel: when Eex exceeds thecoercive field, the domain switches from BA to AB and the band gap in the bottomlayer becomes smaller. The large external field causes the band alignment tochange from type-II to type-I, leading to a quench of the PL from Xh. Yellow andgreen balls represent electrons and holes, respectively.Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 5energy species (Xl) emits from the top layer. Under zero externalfield, the PL intensity of Xh is much stronger than Xl, because theasymmetrical coupling in 3R-MoS2 leads to a type-II band alignmentbetween the two layers— the photoexcited electrons in the top layerquickly relax to the bottom layer, which quenches Xl (Fig. 4b). Sincethe holes in both layers relax to the Γ point at a similar rate4,5, thevalence band offsets at the K points do not contribute to theintensity imbalance.Such an intralayer PL intensity contrast can be reversed by a largeexternal electrical field applied in either direction. In the negativedirection, the externalfield is anti-parallel to the built-in depolarizationfield. When this field is sufficiently strong to compensate the depo-larizationfield, theband alignment changes to type-I as the conductionbandminimumof the top layer becomes lower than that of the bottomlayer, leading to a quenching of the bottom layer PL (Xh) (Fig. 4b). Thisreversal in PL intensity shares the same origin as the dipole momentchange in the interlayer exciton (Fig. 2c).Another sudden PL intensity change is observed when a largepositive electrical field (E = Ec) is applied. In this case, the externalfield is anti-parallel to the electrical polarization and the abruptintensity change is caused by a switch in the stacking order.When thestacking order changes from BA to AB, the band gap in the top layerbecomes larger than that in the bottom layer, and the conductionband offset should be opposite to that in the middle panel of Fig. 4b.Nevertheless, similar to the situation in the left panel of Fig. 4b, alarge positive field can change the band alignment from type-II totype-I, thus quenching the PL from the top layer with higher energy.Importantly, the switching field Ec corresponds to the coercive fieldobserved in Fig. 2c, confirming the picture that the stacking order isswitched from BA to AB. Hysteresis is also observed in the intralayerPL intensity when the stacking order reverts back to BA (Fig. S4 of SI).The comparable intensity ratio between Xh and Xl when there is noexternal field in Fig. 4 and Fig. S4 indicates that the observed hys-teresis behavior originates from intrinsic domain switching ratherthan the interfacial charge trapping effect (Fig. S6 in SupplementaryInformation).In conclusion, we have optically observed a non-volatile switch inthe electrical polarization in a natural 3R-MoS2 bilayer. By probing theStark shift of interlayer exciton with a diffraction-limited focus spot,we map the spatial distribution of polarization domains and theirvariations during the switch. Most importantly, we identify that thisswitch is enabled by the propagation of pre-existing domain walls thatare released by the external electric field. The polarization switch canalso be optically read through the relative photoluminescence inten-sity of intralayer excitons between the two layers.Our findings demonstrate the interplay between rich excitoniceffects and sliding ferroelectricity,which enables a non-volatile controlof the optical properties of 2D semiconductors. The polarization-dependent optical response of 3R-MoS2 provides a promising foun-dation for optical data storage, optical communication, and opticalcomputing applications. Currently, the formation of domain wall inour flakes is not controlled, which agrees with the observation thatonly a fraction of our bilayer devices are switchable. In the future, it willbe important to explore how to systematically generate domain wallsin homogeneous rhombohedral transitionmetal dichalcogenide films,such as by exerting shear strain near the critical point36 or applyingstrong THz field37,38, in order to improve the repeatability and scal-ability of the switching behavior for the future applications of slidingferroelectricity.MethodsSample fabricationThe dual-gated device is fabricated by the standard dry transfermethod under the ambient condition39. The electrical contact isachieved by overlapping the graphite layers with gold electrodes pre-patterned by optical lithography on heavily p-doped Si/SiO2 sub-strates. The 3R-MoS2 is exfoliated from single crystals grown by HQgraphene using chemical vapor transport method40.Optical measurementAll the optical measurements are performed in a continous-flow opti-cal cryostat (Oxford Microstat He), sitting on a x-y motorized stage(Thorlabs, PLS-XY). The base temperature is 4K. The PL spectra aremeasured by a home-built scanning microscope with a 100x objectivelens (Mitutoyo, N.A. = 0.5). An excitation laser of 532 nm is normallyincident on the sample. The diffraction-limited laser focus spot isestimated to be 0.5 μm. The signal is collected by a spectrometer(Princeton Instruments) equipped with a Blaze camera. A long passfilter (600nm) is placed before the entrance of the spectrometer toremove the reflected excitation laser. The optical power at the focus isaround 50–100μW.Electrostatic modelTo obtain the dipole moments of Γ-K excitons in the 3R-MoS2, wecalculate the external electricfield (Eex) appliedwithin the bilayer usinga parallel capacitance model. The external electric field Eex is deter-mined by the top and bottom gate voltages, Vtg and Vbg16,41.Eex =ϵhBN2ϵMoS2Vtgdt� Vbgdb� �ð1ÞHere ϵhBN ≈ 2.7 and ϵMoS2≈7 are the permittivity of hBN and 3R-MoS2bilayer. dt ≈ 18.5 nm and db ≈ 29.0 nm are the thickness of top andbottom hBN. The peak energy (hν) of Γ-K excitons shifts linearly withrespect to Eex. hν0 is the peak position at zero field.hν � hν0 = � μ× Eex ð2ÞThe dipole moments (μ1-μ4) can be extracted from the slope of theStark shift, according to equation (2).Data availabilityThe source data generated in this study have been deposited in theFigshare database (https://doi.org/10.6084/m9.figshare.23577069).Code availabilityThe source code used in this study has been deposited in the Figsharedatabase (https://doi.org/10.6084/m9.figshare.23577069).References1. Wang, C., You, L., Cobden, D. & Wang, J. Towards two-dimensionalvan der Waals ferroelectrics. Nat. Mater. 22, 542–552 (2023).2. Wang, S. et al. Two-dimensional ferroelectric channel transistorsintegratingultra-fastmemory andneural computing.Nat. Commun.12, 53 (2021).3. Wang, X. et al. Van der Waals engineering of ferroelectric hetero-structures for long-retention memory. Nat. Commun. 12,1109 (2021).4. Yang, D. et al. Spontaneous-polarization-induced photovoltaiceffect in rhombohedrally stacked MoS2. Nat. Photonics 16,469–474 (2022).5. Wu, J. et al. Ultrafast responseof spontaneousphotovoltaic effect in3R-MoS2–based heterostructures. Sci. Adv. 8, eade3759 (2022).6. Dong, Y. et al. Giant bulk piezophotovoltaic effect in 3R-MoS2. Nat.Nanotechnol. 18, 36–41 (2023).7. Chai, X. et al. Nonvolatile ferroelectric field-effect transistors. Nat.Commun. 11, 2811 (2020).8. Jiang, X. et al. Ferroelectric field-effect transistors based on WSe2/CuInP2S6 heterostructures for memory applications. ACS Appl.Electron. Mater. 11, 4711–4717 (2021).Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 6https://doi.org/10.6084/m9.figshare.23577069https://doi.org/10.6084/m9.figshare.235770699. Si, M. et al. Ferroelectric field-effect transistors based on MoS2 andCuInP2S6 two-dimensional van der Waals heterostructure. ACSNano 12, 6700–6705 (2018).10. Yasuda, K., Wang, X., Watanabe, K., Taniguchi, T. & Jarillo-Herrero, P.Stacking-engineered ferroelectricity in bilayer boron nitride. Science372, 1458–1462 (2021).11. Vizner Stern, M. et al. Interfacial ferroelectricity by van der Waalssliding. Science 372, 1462–1466 (2021).12. Woods, C. et al. Charge-polarized interfacial superlattices in mar-ginally twisted hexagonal boron nitride. Nat. Commun. 12,1–7 (2021).13. Molino, L. et al. Ferroelectric switching at symmetry-broken inter-faces by local control of dislocation networks. Adv. Mater. 35,2207816 (2023).14. Enaldiev, V. V., Ferreira, F. & Falko, V. I. A scalable networkmodel forelectrically tunable ferroelectric domain structure in twistronicbilayers of two-dimensional semiconductors. Nano Lett. 22,1534–1540 (2022).15. McGilly, L. J., Kerelsky, A., Finney, N. R., Shapovalov, K. & Shih, E. M.Visualization of moiré superlattices. Nat. Nanotechnol. 15,580–584 (2020).16. Liang, J. et al. Optically probing the asymmetric interlayer couplingin rhombohedral-stacked MoS2 bilayer. Phys. Rev. X 12,041005 (2022).17. Deb, S. et al. Cumulative polarization in conductive interfacial fer-roelectrics. Nature 612, 465–469 (2022).18. Bennett, D. Theory of polar domains in moiré heterostructures.Phys. Rev. B 105, 235445 (2022).19. Bennett, D., Chaudhary, G., Slager, R. J., Bousquet, E. & Ghosez, P.Polar meron-antimeron networks in strained and twisted bilayers.Nat. Commun. 14, 1629 (2023).20. Bennett, D. & Remez, B. On electrically tunable stacking domainsand ferroelectricity in moiré superlattices. npj 2D Mater. Appl. 6,7 (2022).21. Wu, M. & Li, J. Sliding ferroelectricity in 2D van der Waals materials:related physics and future opportunities. Proc. Natl Acad. Sci. USA118, e2115703 (2021).22. Li, L. &Wu, M. Binary compound bilayer andmultilayer with verticalpolarizations: two-dimensional ferroelectrics, multiferroics, andnanogenerators. ACS Nano 11, 6382–6388 (2017).23. Wang, X. et al. Interfacial ferroelectricity in rhombohedral-stackedbilayer transition metal dichalcogenides. Nat. Nanotechnol. 17,367–371 (2022).24. Rogée, L. et al. Ferroelectricity in untwisted heterobilayers oftransition metal dichalcogenides. Science 376, 973–978(2022).25. Weston, A. et al. Interfacial ferroelectricity in marginally twisted 2Dsemiconductors. Nat. Nanotechnol. 17, 390–395 (2022).26. Meng, P. et al. Sliding induced multiple polarization states in two-dimensional ferroelectrics. Nat. Commun. 13, 7696 (2022).27. Sun, Y. et al. Mesoscopic sliding ferroelectricity enabled photo-voltaic random access memory for material-level artificial visionsystem. Nat. Commun. 13, 5391 (2022).28. Mak, K., Lee, C., Hone, J., Shan, J. &Heinz, T. Atomically thinMoS2: anew direct-gap semiconductor. Phys. Rev. Lett. 105, 136805(2010).29. Wang, G. et al. Colloquium: excitons in atomically thin transitionmetal dichalcogenides. Rev. Mod. Phys. 90, 021001 (2018).30. Sung, J. et al. Broken mirror symmetry in excitonic response ofreconstructed domains in twisted MoSe2/MoSe2 bilayers. Nat.Nanotechnol. 15, 750–754 (2020).31. Andersen, T. et al. Excitons in a reconstructed Moiré potential intwisted WSe2/WSe2 homobilayers. Nat. Mater. 20, 480–487(2021).32. Li, Z. et al. Stacking-dependent exciton multiplicity in WSe2 bilay-ers. Phys. Rev. B 106, 045411 (2022).33. Ji, J., Yu, G., Xu, C. & Xiang, H. J. General theory for bilayer stackingferroelectricity. Phys. Rev. Lett. 130, 146801 (2023).34. Alden, J. S. et al. Strain solitons and topological defects in bilayergraphene. Proc. Natl Acad. Sci. USA 110, 11256–11260 (2013).35. Petralanda, U. &Olsen, T. Polarization switching induced by domainwall sliding in two-dimensional ferroelectric monochalcogenides.2D Mater. 10, 015001 (2022).36. Jing, L. et al. Shear-strain-induced two-dimensional slip avalanchesin rhombohedral MoS2. Nano Lett. 23, 7228–7235 (2023).37. Shi, J. et al. Intrinsic 1 T0 phase induced in atomically thin 2 H-MoTe2by a single terahertz pulse. Nat. Commun. 14, 5095 (2023).38. Park, J. et al. Ferroelectric switching in bilayer 3R MoS2 via inter-layer shear mode driven by nonlinear phononics. Sci. Rep. 9,14919 (2019).39. Wang, L. et al. One-dimensional electrical contact to a two-dimensional material. Science 342, 614–617 (2013).40. Suzuki, R. et al. Valley-dependent spin polarization in bulk MoS2with broken inversion symmetry. Nat. Nanotechnol. 9, 611(2014).41. Wang, Z., Chiu, Y. H., Honz, K., Mak, K. F. & Shan, J. Electrical tuningof interlayer exciton gases in WSe2 bilayers. Nano Lett. 18,137–143 (2018).AcknowledgementsD.Y. acknowledges Shangyi Guo for assistance on sample fabrication.Z.Y., D.Y., J.L., J.W., Y.H., J.D. acknowledge support from the NaturalSciences and Engineering Research Council of Canada, Canada Foun-dation for Innovation, New Frontiers in Research Fund, Canada FirstResearch Excellence Fund, Max Planck-UBC-UTokyo Centre for Quan-tum Materials, and Gordon and Betty Moore Foundation’s EPiQS Initia-tive (Grant GBMF11071). Z.Y. is also supported by the Canada ResearchChairs Program. K.W. and T.T. acknowledge support from JSPS KAKENHI(Grant Numbers 19H05790, 20H00354 and 21H05233).Author contributionsZ.Y. conceived the project. J.L., D.Y., and J.W. fabricated the samples.D.Y., J.W., and J.L. conducted thePLmeasurement under the supervisionof Z.Y. Y.X. and J.D. assisted on the low temperature measurements.K.W. and T.T. provided the hBN crystal. Z.Y. and D.Y. analysed the data.Z.Y. and D.Y. wrote the manuscript based on the input from all otherauthors. D.Y. J.L. and J.W. contributed equally to this work.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-024-45709-x.Correspondence and requests for materials should be addressed toZiliang Ye.Peer review information Nature Communications thanks XiangjianMeng, and the other, anonymous, reviewer(s) for their contribution tothe peer review of this work. A peer 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.Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 7https://doi.org/10.1038/s41467-024-45709-xhttp://www.nature.com/reprintsOpen 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) 2024Article https://doi.org/10.1038/s41467-024-45709-xNature Communications |         (2024) 15:1389 8http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/ Non-volatile electrical polarization switching via domain wall release in 3R-MoS2 bilayer Results Methods Sample fabrication Optical measurement Electrostatic�model Data availability Code availability References Acknowledgements Author contributions Competing interests Additional information