# Fileset

[s41467-022-32708-z.pdf](https://mdr.nims.go.jp/filesets/babe8c96-cce1-4296-9781-f5b856c77db0/download)

## Creator

Hae Yeon Lee, Soumya Sarkar, Kate Reidy, Abinash Kumar, Julian Klein, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), James M. LeBeau, Frances M. Ross, Silvija Gradečak

## Rights

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

## Other metadata

[Strong and Localized Luminescence from Interface Bubbles Between Stacked hBN Multilayers](https://mdr.nims.go.jp/datasets/2798cd93-9427-422c-9096-c59c79ca0212)

## Fulltext

Strong and Localized Luminescence from Interface Bubbles Between Stacked hBN Multilayersnature communicationsArticle https://doi.org/10.1038/s41467-022-32708-zStrong and Localized Luminescence fromInterface Bubbles Between Stacked hBNMultilayersHae Yeon Lee 1, Soumya Sarkar2, Kate Reidy 1, Abinash Kumar1, Julian Klein1,Kenji Watanabe 3, Takashi Taniguchi 4, James M. LeBeau 1,Frances M. Ross 1 & Silvija Gradečak 1,2Extraordinary optoelectronic properties of van der Waals (vdW) hetero-structures can be tuned via strain caused bymechanical deformation. Here, wedemonstrate strong and localized luminescence in the ultraviolet region frominterface bubbles between stacked multilayers of hexagonal boron nitride(hBN). Compared to bubbles in stacked monolayers, bubbles formed bystacking vdW multilayers show distinct mechanical behavior. We use thisbehavior to elucidate radius- and thickness-dependent bubble geometry andthe resulting strain across the bubble, from which we establish the thickness-dependent bending rigidity of hBN multilayers. We then utilize the polymericmaterial confined within the bubbles to modify the bubble geometry underelectron beam irradiation, resulting in strong luminescence and formation ofoptical standing waves. Our results open a route to design and modulatemicroscopic-scale optical cavities via strain engineering in vdW materials,which we suggest will be relevant to both fundamentalmechanical studies andoptoelectronic applications.Van der Waals (vdW) layered materials and their heterostructuresexhibit extraordinary physical properties while readily allowing out-of-plane deformation, which is of great interest forflexible and conformalelectronics1,2. A unique approach to understand the role ofmechanicaldeformation on optoelectronic properties of vdW materials is to takeadvantage of the bubbles that are formed during the fabrication ofvertical vdW heterostructures. During the stacking process, vdW for-ces squeeze out and trap a material, such as hydrocarbons3–5, air6, orwater7,8, adsorbed on the surface resulting in formation of bubbles atthe interface between the stacked layers. Potential applications ofbubbles in vdW heterostructures have been recently reported due totheir unique structure, the resulting strain, and the pressure in thebubble (vdW pressure)9,10. For example, the ability to trap materialsunder extremely high pressure inside bubbles leads to unusualphenomena such as nano-confined hydrophobic ice7 or chemicalreactions that would not occur under ambient conditions11. Further-more, bubbles between non-permeable graphene membranes havebeen used to trap material for liquid cell electron microscopyapplications12. For optoelectronics and photonics, bubbles in mono-layer transition metal dichalcogenides (TMDs) can serve as strain-induced local emitters13 and optical cavities14. However, the use ofbubbles and induced strain to modify optical properties has so farbeen limited to monolayers. Strain engineering of vdW multilayersshould present new opportunities in optoelectronics sincemultilayersof vdWmaterials exhibit robust optical performance15–17, their bendingis different from that seen in a monolayer or in classical plate theory18,and multilayers circumvent issues related to exfoliation and manip-ulation of monolayers.Received: 29 December 2021Accepted: 12 August 2022Check for updates1Department of Materials Science and Engineering, Massachusetts Institute of Technology, 77 Massachusetts Ave, Cambridge, MA 02141, USA. 2Departmentof Materials Science and Engineering, National University of Singapore, 9 Engineering Drive 1, 117575 Singapore, Singapore. 3Research Center for FunctionalMaterials, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan. 4International Center for Materials Nanoarchitectonics, NationalInstitute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan. e-mail: gradecak@nus.edu.sgNature Communications |         (2022) 13:5000 11234567890():,;1234567890():,;http://orcid.org/0000-0003-2712-0726http://orcid.org/0000-0003-2712-0726http://orcid.org/0000-0003-2712-0726http://orcid.org/0000-0003-2712-0726http://orcid.org/0000-0003-2712-0726http://orcid.org/0000-0003-1178-0009http://orcid.org/0000-0003-1178-0009http://orcid.org/0000-0003-1178-0009http://orcid.org/0000-0003-1178-0009http://orcid.org/0000-0003-1178-0009http://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/0000-0002-7726-3533http://orcid.org/0000-0002-7726-3533http://orcid.org/0000-0002-7726-3533http://orcid.org/0000-0002-7726-3533http://orcid.org/0000-0002-7726-3533http://orcid.org/0000-0003-0838-9770http://orcid.org/0000-0003-0838-9770http://orcid.org/0000-0003-0838-9770http://orcid.org/0000-0003-0838-9770http://orcid.org/0000-0003-0838-9770http://orcid.org/0000-0003-4148-4526http://orcid.org/0000-0003-4148-4526http://orcid.org/0000-0003-4148-4526http://orcid.org/0000-0003-4148-4526http://orcid.org/0000-0003-4148-4526http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-32708-z&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-32708-z&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-32708-z&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-32708-z&domain=pdfmailto:gradecak@nus.edu.sgHere, we study the mechanical behavior of bubbles betweenstacked multilayers of hexagonal boron nitride (hBN) that we findshow a strong localized optical emission. We focus on hBN as a pro-mising material for optoelectronic and quantum optics applicationsbecause it exhibits rich optical properties in the ultraviolet region atroom temperature15,19 with a large band gap of approximately 6 eV20.Importantly for applications, hBN emits strong near-band edge emis-sion with large excitonic binding energy19,21; its intensity increases withhBN thickness, in contrast to TMDs that are luminescent only in themonolayer limit22. The optical emission of hBN can be furtherenhanced and systematically tuned by the twist angle between twostacked multilayers as a result of moiré superlattice effects15. More-over, single photon emitters attributed to structural defects have beenobserved in hBN multilayers in a wide spectral range23,24.By combining experimental measurements and theoretical mod-eling, we show that bubbles formed by bending of vdW multilayersexhibit radius- and thickness-dependent geometry and strain acrossthe bubble, unlike the constant values seen in the case of monolayers.These results provide insights into the bubble geometry as well as thethickness-dependent bending rigidity of vdW multilayers, an impor-tant parameter for strain engineering. We observe a strong phonon-coupled luminescence from the hBN bubble regions and discuss itsorigin and spatial variation. Finally, we demonstrate an approach totune the bubble geometry and optical properties by utilizing decom-position of the trapped material via electron beam irradiation. Thisstrategy to design effective optical emitters in the ultraviolet regionand cavities with controlled dimensions is promising for biological,medical, and stretchable electronic and photonic applications.ResultsInterface bubble formationStacks of multilayer hBN/multilayer hBN (referred to as hBN double-multilayers) were formed by mechanical exfoliation from high-qualitysingle crystal hBN20, followed by vertical stacking onto an Si/SiO2substrate using a modified dry transfer method6. During this process,surface adsorbents are trapped between the two stacked multilayers,forming bubbles at the interface within the double-multilayer. Thesamples were then annealed at 250 °C for 6 h, during which the initialbubbles merged to create fewer but larger bubbles25. The annealingtime was selected to reach an equilibrium26 after which no furtherchanges in size or shape of bubbles were observable. After theannealing, the bubbles formed between the twomultilayers aremostlycircular (Fig. 1a), unless they are near a line defect or an edge. In Fig. 1a,the sixfold pattern in the type II secondary electron (SE2) image isobserved due to the presence of a radially symmetric strain fieldcombinedwith electron channeling effects.As the electronbeamscansacross the bubble, the relative tilt of the lattice with respect to thebeam changes due to the bubble curvature (Supplementary Fig. S1).Because the penetration depth and spreading of the electron beamdepends on the orientation of the incident beam with respect to thelattice27, the yield of backscattered electrons—and hence intensity ofthe SE2 signal—varies across the bubble following the symmetry of theimaged crystal28.Enhanced luminescence from the bubblesTo study optical properties of the bubbles, we use cathodolumines-cence (CL) attached to either a scanning electronmicroscope (SEM) ora scanning transmission electronmicroscope (STEM). CLutilizes a highenergy electron beam to excite optical transitions in a materialthroughout the interaction volume, therefore enabling high resolutionoptical characterization of nanomaterials such as encapsulatedmonolayer TMDs and multilayer hBN15,29,30. Prior to the CL measure-ments, stacked hBN samples were transferred and suspended on atransmission electron microscopy (TEM) grid31 to eliminate any effectof the Si/SiO2 substrate.A panchromatic SEM-CL intensity map of a representative hBNbubble excited using 5 kV electrons and integrated across all thephoton energies (Fig. 1b) shows a strong and localized emission with acircular interference pattern indicative of radially-symmetrical lightinterference behavior. A CL linescan across the bubble (Fig. 1c) showsdramatic changes both in the CL intensity and the dominant opticaltransitions, compared to an unstacked bottom hBNmultilayer and theflat hBN/hBNdouble-multilayer regionwithout bubbles. From all threeregions, we observe intrinsic near-band-edge excitonic recombinationat 215 nm (5.77 eV), demonstrating the high crystalline quality of thehBN20 even after stacking and transferring. The spectrum from the flatdouble-multilayer exhibits an additional symmetric peak at 320 nm(3.9 eV) due to the atomic misalignment between the twomultilayers15with ameasured twist angle of 29°. Finally, a strong emission spanning300–400nm(3–4 eV) is observedonly in thebubble region. This sharpand asymmetric emission with zero-phonon line centered at 303 nm(4.09 eV) and phonon replicas with a constant spacing of 180meV(Supplementary Fig. S2) indicates a distinct origin compared to theluminescence from the flat double-multilayer. The emission band at300–400nm in hBN has been previously attributed to carbon-relateddefects, such as substitutional carbon impurities20,32, which we latershow is likely the origin of the emission in the bubble region. As thebubble effectively acts as a radially-symmetric optical cavity, the con-structive and destructive interference of the emitted light thatdepends on the distance between two multilayers within the bubbleresult in the radially-dependent of the CL intensity across the bubble.For example, in Fig. 1b, CL from hBN multilayers interferes con-structively near the edge of the bubble and interferes destructivelynear the center, leading to the brighter edge. Interference patterns forFig. 1 | Cathodoluminescence frommultilayer hBN bubbles. a SE2 image ofrepresentative hBN (100nm)/hBN (50 nm) bubbles measured at 5 kV.b Panchromatic SEM-CL intensity map (5 kV) of a bubble in a suspended hBN(100nm)/hBN (150 nm) double-multilayer structure, forming an optical cavity. Thearrow shows the line scan direction for (c). c CL line scan across the bubble along ydirection. For reference, the band-edge CL spectrum of an unstacked bottom hBNmultilayer is shown in blue.Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 2bubbles of different sizes are shown in Supplementary Fig. S3. Toelucidate the origin of the strong phonon-coupled luminescence andpotentially control the cavity modes by changing the bubble size andshape, it is essential to understand the mechanical properties of themultilayer bubbles as well as the nature of thematerial trapped inside.Mechanical properties of multilayer bubblesMultilayers of vdW layered materials are expected to show differentbending behavior compared to a monolayer since the bending rigiditycannot be neglected in the case of multilayers, unlike the case of asingle layer. We now show that the bubbles buried inside vdW multi-layers can be used as an experimental tool to extract the bendingrigidity of vdWmultilayers. As shown in Fig. 2a, the bubble geometry isdescribed by the bubble height h and radius r, and the aspect ratioa=h=r determines the maximum strain at the center of a bubble,ε0 / a2,33 which can also be measured by Raman spectroscopy (Sup-plementary Fig. S4). As the stack is supported by the Si/SiO2 substrate,only the top hBN multilayer is expected to deform. The bubbledimensions and shapes as well as the thickness of multilayers weremeasured using atomic force microscopy (AFM) in tapping mode. Fora bubble with the thickness of the topmultilayer t = 70nm,we observethat the bubble aspect ratio varies between 0.01 and 0.02 (Fig. 2b),which is an order of magnitude smaller than the reported values forhBN monolayer bubbles (0.11 is measured for monolayer hBN onhBN26). Moreover, the aspect ratio is a function of the radius r, unlikethe aspect ratio of monolayer bubbles that is independent of r7,26,34,35.From Fig. 2b, it can be seen that in addition to the magnitude ofthe aspect ratio, the deflection profile (i.e., the shape) of the bubbles isdifferent for the multilayer bubbles compared to monolayer bubbles(Fig. 2c). The deflection profile determines the strain distribution fromthe bubble center to the edge33,36 and can be expressed as follows37:yðxÞh= 1� x2r2� �αð1Þwhere x is the distance from the center of the bubble and y(x) is thedeflection. In a conventional plate with a finite thickness, the exponentα =2, whereas α = 1 is assumed for membranes with negligiblethickness37,38. The fundamental difference between the plate andmembrane theory is in the assumed sample thickness. In the classicalplate theory, a finite thickness of a plate results in a finite bendingrigidity, whereas the membrane analysis assumes negligible thicknessand consequently ignores the bending rigidity36. Due to the extremethinness of 2D monolayers, the membrane model (α = 1) has beensuccessfully applied formonolayer bubbles33,34,36,39, and is confirmed inour control case of an hBNmonolayer (blue circles in Fig. 2c).However,in the case of a representative hBN multilayer bubble (red squares inFig. 2c),fitting our experimental results yields α = 1:5, which is betweenthese two theoretical limits. This result indicates that multilayerdeformation cannot be fully explained by the membrane theory usedfor individual monolayers because of the non-negligible bendingrigidity of multilayers, nor by the classical plate theory used forconventional thin films due to the weak vdW interaction betweenlayers.As the bending rigidity is a dominant parameter that governs theout-of-plane deformation, we next calculate how it determines theFig. 2 | Multilayer vs. monolayer bubble geometry. a Schematic of a bubbleformed between twomultilayers of hBN/hBNon a Si/SiO2 substrate. The radius andheight of the bubble are r andh respectively, and the thickness of the topmultilayeris t.b The aspect ratio a as a function of the bubble radius rmeasured from an hBN/hBN multilayer bubble (t = 70 nm, Supplementary Fig. S5). The dashed line indi-cates the aspect ratio reported previously for hBN monolayer bubbles, which is aconstant independent of r26. The shaded region shows the experimentallymeasured aspect ratio for multilayer bubbles, which is r-dependent and sig-nificantly lower than for the monolayer counterpart. Error bars represent onestandard deviation above and below the mean. c Experimental deflection profilesexperimentally measured from a representative hBN/hBN multilayer bubble(squares) and a hBN/hBN monolayer bubble (circles). The red solid curve fitted toEq. (1) yields α = 1:5. Analytic solutions are also shown for the elastic plate (dashedgreen line, α = 2) and membrane (dashed blue line, α = 1) models.Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 3aspect ratioa in the caseofmultilayer compared to a singlemonolayer.Wedo this by considering a theoreticalmodel of amultilayer bubble atequilibriumwhere its total energy (Etot) isminimizedwith respect to itsheight and radius26,33,36. Etot can be described by four energy termswhen no external strain is applied: (i) in-plane elastic energy (Eel), (ii)bending energy (Ebend), (iii) adhesion energy between the bent multi-layer and the substrate multilayer (Eadh), and (iv) free energy of thematerial inside the bubble (Eb Vð Þ):Etot = Eel + Ebend + Eadh + Eb Vð Þ ð2ÞEach energy term can be further described as Eel = c1Yh4=r2,Ebend = c2κh2=r2, Eadh = c3γr2, ∂Eb Vð Þ=∂V =�P, V = c4hr2, where Y is in-plane stiffness, κ is bending rigidity, γ is adhesion energy, V is thevolume of the bubble, P is the pressure inside the bubble, andc1,c2,c3,c4 are constants26. The aspect ratio a is then calculated byminimizing Etot with respect to h and r.For amonolayer, the bending energy term is neglected (κ =0) dueto its extreme thinness (the membrane model), and the aspect ratio issimplified to Eq. (3) below.26 Details of the calculation are provided inSupplementary Note 1.a=γc1Ymono� �1=4ð3ÞEquation (3) shows that the aspect ratio of a monolayer bubbledepends only on the ratio between the adhesion and in-plane elasticenergies of the 2D crystal, resulting in a universal aspect ratio for aspecific material that is independent of the bubble size. For a mono-layer of hBN, the aspect ratio can be calculated using the values of thein-plane stiffness for a monolayer Ymono ≈ 22 ± 6ð ÞeV _A�2,γ ≈0:005eV _A�2and c1 ≈ 1:55,26,40 which leads to the aspect ratioa=0:110 (±0:008), consistent with the experimental reports26. Thevalues of Ymono reported in other literature18,41 are within 20% of thevalue used here, which does not result in a substantial change in theaspect ratio a. The parameters used here are summarized in Supple-mentary Table 1.We next extend the model to the case of a multilayer, where thebending rigidity cannot be neglected (κ ≠0):a=γh2c1Ymultih2 + c2κ !1=4ð4ÞYmulti is the in-plane stiffness of the multilayers (Ymulti = Ymono�t). Wenote that κ also scales with thickness, however, we do not assume aspecific functional dependence at this point and allow κ ∼ tx . γ isindependent of the thickness and c1 and c2 are dimensionlesscoefficients that depend on the in-plane and out-of-plane displace-ment, respectively. Therefore, only c2 varies with the deflection profileof the bubbles, and hence their thickness. As κ and c2 both changewiththickness, the term c2κ, the geometry-dependent bending rigidity,provides a more comprehensive relationship between the aspect ratioand the thickness by capturing the combined effect of bending rigidityand deflection profile on the aspect ratio. Figure 3a shows thecalculated aspect ratios as a function of the bubble radius r fordifferent values of the geometry-dependent bending rigidity c2κ from0 to 200 × 104eV. As shown inFig. 3a, for non-zeroc2κ, the aspect ratioof multilayer bubbles increases with their radius. Furthermore, as thec2κ increases, the slope of the curve decreases, extending the linearregime in which the aspect ratio is approximately a linear function ofthe radius and the effect of the thickness dependence of Ymulti onaspect ratio is negligible (Supplementary Fig. S6). An importantconsequence of this analysis is that unlike monolayer bubbles, wherethe strain is consistent, the aspect ratio and strain of multilayerbubbles can be engineered systematically through the radius and themultilayer thickness. For example, a change in strain of an order ofmagnitude can be induced by changing the bubble radius from0.2μmto 0.3μm for t = 7 nm, or by changing the multilayer thickness from10nm to 30nm for r =0:3μm. Our experimental results are inexcellent agreement with the theoretical prediction: the aspect ratioof hBN bubbles decreases with increasing multilayer thickness and foreach thickness, the aspect ratio increases with radius (Fig. 3b,Supplementary Fig. S7). It is worth noting that this model is applicableto other vdW layered materials, as we demonstrate for WS2 inSupplementary Fig. S8.By comparing the experimental data with the calculatedresults for different c2κ, we can now extract the c2κ that corre-sponds to each thickness, which enables us to unravel the rela-tionship between thickness and bending rigidity, as shown inFig. 3c. In the classic linear plate theory, bending rigidity is pro-portional to thickness cubed (κ ~ t3)37, whereas in the case of freeshear interaction (i.e., no interaction) between adjacent layers,bending rigidity is linearly proportional to thickness (κ ~ t)18. Theseare the two limiting cases where the first ignores the possibility ofsliding (infinite friction of sliding), whereas the second ignoresinterlayer interaction (zero friction of sliding). In the case of vdWlayered materials, however, the thickness dependence of thebending rigidity is between the two cases, as we show in Fig. 3c andalso corroborated by other studies18. This behavior can be under-stood by considering the anisotropic in-plane and out-of-planemechanical properties of an individual layers18,42–44 and weak vdWinteractions between the layers, which may subject the vdW mul-tilayer to interlayer sliding.Electron beam-induced modulation of the bubble geometryWe next turn to the question of whether the bubble geometry can bemodified, both to increase our understanding of the mechanicalproperties and to eventually tailor themultilayer optical properties. Asmentioned previously, the bubbles are filled with polymeric materialssuch as hydrocarbons, air, or water. We find that this confinement ofmaterials can be used to expand the bubble size using electron beamirradiation. The bubbles were irradiated inside an SEM, using electronenergy between 1 and 10 kV, which is far below the threshold energy ofhBN. Their geometry was subsequently measured by AFM (see Sup-plementary Note 2). Figure 4a shows 3D AFMmaps of a representativehBN bubble before and after irradiation at 3 kV, demonstrating a 5×volume increase and 1.6× increase in the aspect ratio. We suggest thatdecomposition of organic compound inside thebubble by the electronbeam is a likely cause of the expansion and provide supporting evi-dence in Fig. 5 below. It is interesting to note that the aspect ratio stillfollows the theoretical model even after irradiation, as shown inFig. 4b, indicating that Eq. (2) is still applicable and can be used topredict the geometry of the expanded bubbles. Measurements onmultiple bubbles show that all the extracted values κ before and afterirradiation remain similar (2≤ c2κ ≤4), implying that the electronbeamdoes not dramatically change the intrinsicmechanical properties (Y, κ,and γ) of the multilayer.While the mechanical parameters are preserved, the deflectionprofiles show a transition between the two limits (α = 1 and α =2),approaching the membrane limit as the bubble expands (Fig. 4c),likely due to the increasing maximum strain45. We note a sharpchange in curvature near the edge in the transition from α =2 to α = 1,which can cause strain concentration at the edge (SupplementaryFig. S9). The deflection profiles from all bubbles in Fig. 4b were col-lected and the values of α obtained by fitting to Eq. (1) are shown inFig. 4d. The exponent approaches 1 with increasing radius, and thereciprocal of the aspect ratio 1/a calculated by the theoretical modelmatches our experimental results. Therefore, in addition to its aspectratio, the bubble shape can also be predicted by the model, whichArticle https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 4enables us to calculate both the strain at the bubble center and thestrain distribution across the bubble including the stress con-centration at the edge. This result broadens our understanding ofvdWmultilayer bending, since previous literature has suggested thatthe deflection profiles of multilayer bubbles collapse onto one curve(α = 1 or α =2)18.Origin of the enhanced luminescence from the bubblesBy combining the tunability of the bubble geometry using an electronbeam while simultaneously measuring optical properties using CL, wecan examine the origin of the strong and localized optical emissionshown in Fig. 1 and also obtain information regarding the compositionof the material within the bubble. CL measurement at different accel-erating voltages enables us to investigate the depth profile of opticalemission of a sample because of the difference in electron penetrationdepth (Supplementary Fig. S10). We therefore irradiated the sample at80 keV, measuring STEM-CL, then irradiated at 5 keV while measuringSEM-CL (Fig. 5a and Supplementary Fig. S10). Comparing the STEM-CLand SEM-CL panchromatic maps in Supplementary Fig. S10a and d,only SEM-CL shows strong emission from the bubble region, whereasthe bubbles appear dark in STEM-CL. For the 5 keV electrons used inSEM, energy is absorbedmainlywithin the tophBNmultilayer, whereasfor the 80 keV electrons used in STEM, energy is absorbed throughoutthe entire sample thickness within a narrow (<5 nm) interactionvolume(Supplementary Fig. S10b–f). The strongCLemissionobservedin the bubble region that is only measured at low accelerating voltagetherefore originates from the top strained hBN multilayer. This is fur-ther confirmed by comparing CL measurement at 5 keV and 10 keV(Supplementary Fig. S11).To further understand the origin and localization of this lumi-nescence, we next focus on the optical interference and show that thebubbles are likely filled by organic compounds such as hydrocarbons.For the bubble in Fig. 1b, c, we show the measured CL peak intensityacross the bubble (Fig. 5b, left) at several wavelengths in the range300–330 nm. Since peak shifts of the zero-phonon line and its phononreplicas are not observed, we can attribute changes in emissionintensity across the bubble to interference. We take advantage of ourmechanical model to determine the shape and height of the bubblefrom the experimentally measured bubble radius (r =4μm) and layerthickness (t = 150nm). Based on this calculated shape (h= 100nm andα = 1:7), we conduct transfer-matrix method reflection simulations46that assume optical properties of several likely candidates: poly-carbonate (Fig. 5c), air, and water (Supplementary Fig. S12). For thesecalculations, we measured the optical constants (refractive index andextinction coefficient) of the hBN multilayer using ellipsometry whilethe wavelength-dependent optical constants of polycarbonate47 andwater48 were taken from literature. By comparing the experimentalresults with the simulations, the closest match is for polycarbonate,fromwhichwe conclude that the refractive index of thematerial insidethe bubble is n≥ 1:65 at 300nm. We note that this simulation is con-sistent with data from bubbles of different sizes (SupplementaryFig. S3). Taken together, these results suggest that organic compoundsof density similar to polycarbonate fill the bubble. This material pre-sumably originates from the dry transfer process. As polymericmaterials including polycarbonate readily decompose under an elec-tron beam by beam-induced ionization and dissociation, producinguncombined free carbon species49, this likely accounts for the electronbeam-induced expansion.a00.20.4124102040100200c2bThickness (nm)1507501000.000.020.040.060.080.100.120.0 0.5 1.0 1.5 2.0 2.5 3.0c0.0 0.5 1.0 1.5 2.0 2.5 3.00.000.010.020.030.040.05(×10 )25Aspect ratio (a) Aspect ratio (a)0 20 40 60 80 120 160Thickness (nm)100 140020406080100120140160 t³tc2(× 10⁴ eV)multilayerplatevdWmultilayerno interactionmonolayer (ref.26) �4Radius (     )m��platevdW interactionno interactionRadius (     )m�Fig. 3 | Mechanical properties of multilayer hBN bubbles. a Calculated aspectratio a as a function of bubble radius for various values of the geometry-dependentbending rigidity. The shaded area indicates the region magnified in b where theaspect ratio is an approximately linear function of radius. b Experimentally mea-sured aspect ratio of hBN/hBNmultilayer bubbles (solid circles) as a function of thebubble radius and the multilayer thickness (represented by the symbol color). Thesolid lines are theoretically calculated values shown in a. c The experimentallydetermined bending rigidity. Error bars represent one standard deviation aboveand below the mean or are smaller than the symbols and the solid red curve is aguide to the eye. Two dashed curves show ~ t3 and ~t dependency, which corre-spond to the case of an elastic plate and multilayers with no interlayer interaction,respectively. The schematics show atomic configurations for each case with dif-ferent interlayer interaction. The color gradient along the y axis corresponds tobending rigidity in (a).Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 5We finally investigate the structural features of carbon inside thebubbles qualitatively by comparing electron energy loss spectroscopy(EELS) measured from the bubbles and from flat regions near thebubbles (Fig. 5c). The carbon K edge peak from the bubbles shows aclear feature of sp2 2 bonding (285 eV) and a relatively narrow peak(295 eV),whereas the spectrummeasured from flat regions shows onlybroad features that have been observed in amorphous carbon50. Wespeculate that this change in carbon bonding state may be caused bystructural modification of the strained hBN multilayer due to incor-poration of carbon into hBN.DiscussionAn asymmetric luminescence with phonon replicas in the spectralrange of 300–400 nm, similar to what we observe in the bubble region(Fig. 1c), hasbeen reportedpreviously in hBN20,32. It hasbeen attributedto carbon-related defects such as the substitutional carbon impurities,which formdeep impurity levelswithin the hBNbandgap20,32. Basedonour findings that (1) the emission originates from a deformed top hBNmultilayer, (2) polycarbonate is confined inside the bubble and dis-sociated by electron beam irradiation resulting in modulation of thebubble geometry, and (3) the carbon bonding state changes inside theexpanded bubble (Fig. 5d), we speculate that the trappedmatter insidethe bubble induces a structural modification of the strained hBNmultilayer, resulting in strong and local luminescence from thebubble.Possible carbon doping is also consistent with the previous reportsthat show reduced radiation hardness of hBN under strain51 and elec-tron beam-induced carbon doping in hBN52. Furthermore, the opticalstanding waves formed within the expanded bubble depend on itsgeometry (Supplementary Fig. S3), which can be calculated by ourmechanical model and modulated by the electron beam.Our results suggest that nanoscale bubbles can be created in acontrolled manner with properties of interest for optoelectronic andphotonic applications, for example photonic crystals formed bycreating arrays of nanoscale bubbleswhere each component plays roleas an ultraviolet emitter. As strain across the bubble enhances theincorporation of dopants, bubbles can also serve as an effective tool tostudy strain-related optical properties within a band gap. It is alsopossible to tune the emission wavelength further by changing thefilling materials inside the hBN bubble: for example, Ni or Cu in hBNcan result in luminescence in the visible regime53,54.We have observed strong and localized luminescence in theultraviolet region from the bubbles formed between two hBN multi-layers and have suggested a model for this luminescence based onanalysis of the mechanical characteristics of strained vdWmultilayers.Because deformation of a vdWmultilayer is dominated by its bendingrigidity, in contrast to a monolayer, we show that bubbles formed invdW multilayers show a distinct geometry both in aspect ratio andshape. The geometry of bubbles and the thickness-dependent bendingFig. 4 | Electron-beam-inducedmodulation of hBN bubbles. a 3D AFMmap of amultilayer bubble (t = 50nm) before and after electron beam irradiation. vdWpressure of the bubbles are calculated in Supplementary Note 1. b Aspect ratio ofthe bubbles before (black) and after (red) irradiation with the theoretically calcu-lated aspect ratio for different c2κ values in Fig. 2a. All data points lie betweenc2κ = 2 and c2κ =4. The star symbol indicates the bubble shown in (a). c Fittedcurves of the normalized deflection profiles of a representative multilayer bubbleto Eq. (1). The deflection profiles of the same bubble are measured three times:before irradiation, α = 1:53 (red), after the first dose (α = 1:36, dark red) and aftersecond dose (α = 1:10, darker red). Two dashed curves are the two limits shown inFig. 1d. d The values of exponent (α) obtained by fitting the deflection profilesmeasured from pre-irradiated (black) and irradiated (red) bubbles in b. The reci-procal of theoretically calculated aspect ratio (1/a) is plotted, showing that the datapoints are still between c2κ = 2 and c2κ =4. Error bars represent one standarddeviation above and below the mean.Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 6rigidity can be calculated using the theoretical model developed here,which is a precondition for strain engineering. As a pathway to modifythe geometry of the bubbles, we find that irradiation by an electronbeam can decompose the trapped materials inside the bubbles,increasing their luminescence. These results open a route towardcreation of strong ultraviolet optical emitters and suggest opportu-nities for designing cavities for luminescence utilizing the bending ofvdW multilayers and confinement of materials within. Intentionalintroduction and confinement of other materials within bubbles madefrom various vdW multilayers may further expand the possibilities ofthis strategy to induce local luminescence at different frequencyranges.MethodsMaterials and stackingOur starting materials were bulk crystals of hBN synthesized by thehigh-temperature-high-pressure method20 and bulk crystals of WS2from HQ Graphene. hBN and WS2 multilayers were mechanicallyexfoliated onto an Si/SiO2 substrate by the scotch-tape method, thenvertically stacked via a modified dry transfer method using a polymerstamp (polydimethylsiloxane (PDMS)/polypropylene carbonate(PPC)). For the PDMS mask, a solution of 20:1 ratio of Sylgard 184prepolymer to curing agent was kept at ambient condition for ∼24 h.After oxygen plasma treatment of PDMS (18W for 5min), 15%of PPC inanisole was spin-coated at 3000 rpm on the PDMS. The solvent wasremoved by heating the PDMS/PCC mask at 160 °C for 10min. Thepolymer stampwas contacted to a topmultilayer for 1min at 50 °C forpick-up and then dropped onto a bottommultilayer at 100 °C followedby increasing the temperature to 120 °C for 2min. To remove PPCresidue, the sample was immersed in acetone for 5min and iso-propanol for 10 s. The resulting double-multilayers were annealed at250 °C in an Ar environment for 6 h.Topographic characterizationHeight profiles of bubbles on Si/SiO2 substrate were obtainedbefore and after electron beam irradiation using a Veeco Dimen-sions 3100 AFM operated in tapping mode with 300 kHz tip. A ZeissMerlin high resolution SEM was used for electron beam irradiationa2.01.61.20.80.40.0Intensity (arb. units)280 300 320 340Energy (eV)flatbubbleEELSc dπ0.80.60.40.201.01.52.02.53.03.5e-beamSEM-CL Panchromatic )R( ytivitcelfeR 330nm300nm315nmbNormalized CL1.00.80.60.40.20Wavelength (nm)330300Optical microscopy Optical microscopyafter ebeamirradiation-1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 1.0x / r x / r280 285 290intensity (a.u.)m10 m10CL intensity (arb. units)0.5Fig. 5 | Origin of the enhanced cathodoluminescence from hBN bubbles.aOpticalmicroscope imagebefore (hBN (70 nm)/hBN (70 nm)on Si/SiO2) and after(suspended hBN/hBN) CL measurement, scale bar 10μm, and SEM-CL panchro-matic map at 5 kV. The STEM-CL map is shown in Supplementary Fig. S10. b Left:Normalized CL intensity at a specific wavelength (300–330 nm) across the bubblein Fig. 1. Right: Reflectivity at specific wavelengths (300, 315, 330 nm) calculatedusing a transfermatrixmethod across a bubble withh= 100 nmassumedfilled withpolycarbonate. c EELS of carbonK edge peaksmeasured on the bubble and at a flatregion nearby. The inset shows the spectrum in the shaded region after linearsubtraction. d Schematic showing the expanding bubble and standing waves.Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 7with current density of 100–500 pA and dwell time of 25 ms/pixel(104�105electrons=nm2) at 1–10 kV.Optical characterizationFor CL measurements the samples were transferred to a C-flat TEMgrid (Protochips Inc., 200mesh, 2 μmholes) by a wet transfer methodusing poly(methylmethacrylate) (PMMA, 495 K, A4 Microchem) and1M potassium hydroxide. STEM-CL wasmeasured first at 80 kV (usinga GatanMonoCL 3+ attached to a JEOL 2011 TEM) followed by SEM-CLat 5 kV (AttolightTM Allalin) with dwell time of 20ms/pixel.The refractive index n of an exfoliated hBN flake with a thicknessof 5 nm was measured with a spectroscopic imaging nulling ellips-ometer EP4 (Accurion Gmbh, Göttingen, Germany) in ambient condi-tions and at room temperature.Interference simulationThe regions of constructive and destructive interference in the hBN/hBN stacks were calculated using a conventional transfer matrixmethod calculation46, implemented in MATLAB. The reflectiondepends on the spacing d between two multilayers, and the inter-ference pattern can be predicted by calculating the reflection coeffi-cient (R) through each layer using establishedn and koptical constantsfrom literature and ellipsometry (Supplementary Fig. S13).EELS measurementSTEM-EELS measurements with energy resolution ~0.3 eV wereobtained at 200 kV using a probe-corrected Thermo Fisher ScientificTitan G3 60–300 kV with monochromator and a Gatan Continuumspectrometer. High energy-loss spectra were aligned using the zero-loss shift determined using the simultaneously collected low-lossspectra. The carbon K-edge background was fitted to a power law andsubtracted from the raw spectrum. The effects of plural scatteringwere removed from the spectrum using the Fourier ratio method.Data availabilityThe datasets generated and analyzed during this study are included inthis paper and its Supplementary Information, and are also availablefrom the corresponding author upon reasonable request.References1. Akinwande, D., Petrone, N. & Hone, J. Two-dimensional flexiblenanoelectronics. Nat. Commun. 5, 1–12 (2014).2. Chen, P., Liu, M., Wang, Z., Hurt, R. H. &Wong, I. Y. From Flatland toSpaceland: Higher Dimensional Patterning with Two‐DimensionalMaterials. Adv. Mater. 29, 1605096 (2017).3. Haigh, S. J. et al. Cross-sectional imaging of individual layers andburied interfaces of graphene-based heterostructures and super-lattices. Nat. Mater. 11, 764–767 (2012).4. Kretinin, A. V. et al. Electronic properties of graphene encapsulatedwith different two-dimensional atomic crystals. Nano Lett. 14,3270–3276 (2014).5. Schwartz, J. J. et al. Chemical Identification of Interlayer Con-taminants within van der Waals Heterostructures. ACS Appl. Mater.Interfac. 11, 25578–25585 (2019).6. Pizzocchero, F. et al. The hot pick-up technique for batch assemblyof van der Waals heterostructures. Nat. Commun. 7, 11894 (2016).7. Bampoulis, P., Teernstra, V. J., Lohse, D., Zandvliet, H. J. W. &Poelsema, B. Hydrophobic ice confined between graphene andMoS2. J. Phys. Chem. C. 120, 27079–27084 (2016).8. Ghorbanfekr-Kalashami, H., Vasu, K. S., Nair, R. R., Peeters, F. M. &Neek-Amal, M. Dependence of the shape of graphene nanobubbleson trapped substance. Nat. Commun. 8, 1–11 (2017).9. Vasu, K. S. et al. Van der Waals pressure and its effect on trappedinterlayer molecules. Nat. Commun. 7, 12168 (2016).10. Tan, B. H. et al. DirectMeasurement of theContents, Thickness, andInternal Pressure of MolybdenumDisulfide Nanoblisters.Nano Lett.20, 3478–3484 (2020).11. Lim, C. H. Y. X., Nesladek, M. & Loh, K. P. Observing High‐PressureChemistry in Graphene Bubbles. Angew. Chem. Int. Ed. 53,215–219 (2014).12. Ross, F. M. Opportunities and challenges in liquid cell electronmicroscopy. Science 350, aaa9886 (2015).13. Tyurnina, A. V. et al. Strained bubbles in van der Waals hetero-structures as local emitters of photoluminescence with adjustablewavelength. ACS Photon. 6, 516–524 (2019).14. Jia, Z. et al. Photoluminescence and Raman Spectra OscillationsInduced by Laser Interference in Annealing‐Created MonolayerWS2 Bubbles. Adv. Opt. Mater. 7, 1801373 (2019).15. Lee, H. Y. et al. Tunable Optical Properties of Thin Films Controlledby the Interface Twist Angle. Nano Lett. 21, 2832–2839 (2021).16. Sarkar, S. et al. Direct Bandgap-like Strong Photoluminescencefrom Twisted Multilayer MoS2 Grown on SrTiO3. ACS Nano 14,16761–16769 (2020).17. Gupta, S., Shirodkar, S. N., Kutana, A. & Yakobson, B. I. In pursuit of2D materials for maximum optical response. ACS Nano 12,10880–10889 (2018).18. Wang, G. et al. Bending ofmultilayer van derWaals materials. Phys.Rev. Lett. 123, 116101 (2019).19. Watanabe, K., Taniguchi, T. & Kanda, H. Direct-bandgap propertiesand evidence for ultraviolet lasing of hexagonal boron nitride singlecrystal. Nat. Mater. 3, 404–409 (2004).20. Taniguchi, T. & Watanabe, K. Synthesis of high-purity boron nitridesingle crystals under highpressurebyusingBa–BN solvent. J. Cryst.Growth 303, 525–529 (2007).21. Schué, L. et al. Bright Luminescence from Indirect andStrongly Bound Excitons in h-BN. Phys. Rev. Lett. 122,067401 (2019).22. Mak, K. F., Lee, C., Hone, J., Shan, J. & Heinz, T. F. Atomically thinMoS 2: a new direct-gap semiconductor. Phys. Rev. Lett. 105,136805 (2010).23. Bourrellier, R. et al. Bright UV single photon emission at pointdefects in h-BN. Nano Lett. 16, 4317–4321 (2016).24. Mendelson, N. et al. Identifying carbon as the source of visiblesingle-photon emission from hexagonal boron nitride. Nat. Mater.20, 321–328 (2021).25. Rosenberger, M. R. et al. Nano-“squeegee” for the creation of clean2D material interfaces. ACS Appl. Mater. Interfac. 10,10379–10387 (2018).26. Khestanova, E., Guinea, F., Fumagalli, L., Geim, A. K. & Grigorieva, I.V. Universal shape and pressure inside bubbles appearing in vander Waals heterostructures. Nat. Commun. 7, 1–10 (2016).27. Joy, D. C., Newbury, D. E. & Davidson, D. L. Electron channelingpatterns in the scanning electron microscope. J. Appl. Phys. 53,R81–R122 (1982).28. Rehn, S. M., Gerrard-Anderson, T. M., Qiao, L. & Jones, M. R.Mechanical Reshaping of Inorganic Nanostructures with WeakNanoscale Forces. Nano Lett. 21, 130–135 (2020).29. Singh, A., Lee, H. Y. & Gradečak, S. Direct optical-structure corre-lation in atomically thin dichalcogenides and heterostructures.Nano Res. 13, 1363–1368 (2020).30. Lim, S. K. et al. Direct correlation between structural and opticalproperties of III−Vnitride nanowire heterostructureswith nanoscaleresolution. Nano Lett. 9, 3940–3944 (2009).31. Schué, L. et al. Dimensionality effects on the luminescence prop-erties of hBN. Nanoscale 8, 6986–6993 (2016).32. Museur, L., Feldbach, E. & Kanaev, A. Defect-related photo-luminescence of hexagonal boron nitride. Phys. Rev. B 78,155204 (2008).Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 833. Wang, P., Gao, W., Cao, Z., Liechti, K. M. & Huang, R. Numericalanalysis of circular graphene bubbles. J. Appl. Mech. 80,040905 (2013).34. Sanchez, D. A. et al. Mechanics of spontaneously formed nanobl-isters trapped by transferred 2D crystals. Proc. Natl Acad. Sci. 115,7884–7889 (2018).35. Tedeschi, D. et al. Controlled Micro/Nanodome Formation in Pro-ton‐Irradiated Bulk Transition‐Metal Dichalcogenides. Adv. Mater.31, 1903795 (2019).36. Yue, K., Gao, W., Huang, R. & Liechti, K. M. Analytical methods forthe mechanics of graphene bubbles. J. Appl. Phys. 112,83512 (2012).37. Timoshenko, S. P. & Woinowsky-Krieger, S. Theory of plates andshells. (McGraw-hill, 1959).38. Freund, L. Ben & Suresh, S. Thin film materials: stress, defect for-mation and surface evolution. (Cambridge university press, 2004).39. Dai, Z. et al. Interface-governed deformation of nanobubbles andnanotents formed by two-dimensional materials. Phys. Rev. Lett.121, 266101 (2018).40. Han, T., Luo, Y. &Wang, C. Effects of temperature and strain rate onthe mechanical properties of hexagonal boron nitride nanosheets.J. Phys. D. Appl. Phys. 47, 25303 (2013).41. Gao, E. &Xu, Z. Thin-shell thickness of two-dimensionalmaterials. J.Appl. Mech. 82, 121012 (2015).42. Tapasztó, L. et al. Breakdown of continuum mechanics fornanometre-wavelength rippling of graphene. Nat. Phys. 8,739–742 (2012).43. Zhang, D.-B., Akatyeva, E. & Dumitrică, T. Bending ultrathin gra-phene at themargins of continuummechanics. Phys. Rev. Lett. 106,255503 (2011).44. Wei, Y., Wang, B., Wu, J., Yang, R. & Dunn, M. L. Bending rigidity andGaussian bending stiffness of single-layered graphene. Nano Lett.13, 26–30 (2013).45. Sheploak, M. & Dugundji, J. Large deflections of clamped circularplates under initial tension and transitions to membrane behavior.Trans. ASME, J. Appl. Mech. 65, 107–115 (1998).46. Mackay, T. G. & Lakhtakia, A. The Transfer-Matrix Method in Elec-tromagnetics andOptics. Synth. Lect. Electromagn. 1, 1–126 (2020).47. Sultanova, N., Kasarova, S. & Nikolov, I. Dispersion proper ties ofoptical polymers. Acta Phys. Pol. A Gen. Phys. 116, 585 (2009).48. Hale,G.M. &Querry,M. R.Optical constantsofwater in the 200-nmto 200-μm wavelength region. Appl. Opt. 12, 555–563 (1973).49. Ding, W. et al. Mechanics of hydrogenated amorphous carbondeposits from electron-beam-induced deposition of a paraffinprecursor. J. Appl. Phys. 98, 14905 (2005).50. Garvie, L. A. J., Craven, A. J. & Brydson, R. Use of electron-energyloss near-edge fine structure in the study of minerals. Am. Mineral.79, 411–425 (1994).51. Peng, Q., Ji, W. & De, S. First-principles study of the effects ofmechanical strains on the radiation hardness of hexagonal boronnitride monolayers. Nanoscale 5, 695–703 (2013).52. Wei, X., Wang, M.-S., Bando, Y. & Golberg, D. Electron-beam-induced substitutional carbon doping of boron nitride nanosheets,nanoribbons, and nanotubes. ACS Nano 5, 2916–2922 (2011).53. Ikram, M. et al. 2D chemically exfoliated hexagonal boron nitride(hBN) nanosheets dopedwith Ni: synthesis, properties and catalyticapplication for the treatment of industrial wastewater. Appl.Nanosci. 10, 3525–3528 (2020).54. Ikram, M. et al. Evaluation of antibacterial and catalytic potential ofcopper-doped chemically exfoliated boron nitride nanosheets.Ceram. Int. 46, 21073–21083 (2020).AcknowledgementsThis work was supported by the National Research Foundation (NRF)of Singapore through the Singapore−MIT Alliance for Research andTechnology Centre (SMART). K.R. acknowledges funding from an MITMathworks Engineering Fellowship and support provided by Exxon-Mobil Research and Engineering Company through the MIT EnergyInitiative. This work made use of Shared Experimental Facilitiessupported in part by the MRSEC program of the National ScienceFoundation under award DMR-1419807 and the use of the MIT Char-acterization.nano facility. K.W. and T.T. acknowledge support fromthe Elemental Strategy Initiative conducted by the MEXT, Japan(Grant Number JPMXP0112101001) and JSPS KAKENHI (Grant Num-bers JP19H05790 and JP20H00354). We thank AttolightTM for helpwith SEM-CL measurements. K.R. would like to thank Prof. ArturDavoyan and Prof. Giulia Tagliabue for helpful discussions on thetransfer matrix method code implementation.Author contributionsH.Y.L. fabricated samples, performed characterization, data analysis andsimulation, and developed the model. S.S. performed certain AFMmeasurements. K.R. implemented the transfermatrixMATLABcode. A.K.measured EELS supervised by J.L. J.K. measured the optical constant ofhBN. K.W. and T.T. synthesizedbulk hexagonal boron nitridecrystal. S.G.and F.M.R. supervised the project. H.Y.L., S.G. and F.M.Rwrote the paperand all authors provided their comments.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-022-32708-z.Correspondence and requests for materials should be addressed toSilvija Gradečak.Peer review information Nature Communications thanks the anon-ymous reviewer(s) for their contribution to the peer review of this work.Reprints and permission 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 license, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons license, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2022Article https://doi.org/10.1038/s41467-022-32708-zNature Communications |         (2022) 13:5000 9https://doi.org/10.1038/s41467-022-32708-zhttp://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/ Strong and Localized Luminescence from Interface Bubbles Between Stacked hBN Multilayers Results Interface bubble formation Enhanced luminescence from the bubbles Mechanical properties of multilayer bubbles Electron beam-induced modulation of the bubble geometry Origin of the enhanced luminescence from the bubbles Discussion Methods Materials and stacking Topographic characterization Optical characterization Interference simulation EELS measurement Data availability References Acknowledgements Author contributions Competing interests Additional information