# Fileset

[s41467-019-10660-9.pdf](https://mdr.nims.go.jp/filesets/dcaff955-e86a-4273-83bd-a777c0588fb0/download)

## Creator

Li He, Huishan Wang, Lingxiu Chen, Xiujun Wang, Hong Xie, Chengxin Jiang, Chen Li, Kenan Elibol, Jannik Meyer, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Zhangting Wu, Wenhui Wang, Zhenhua Ni, Xiangshui Miao, Chi Zhang, Daoli Zhang, Haomin Wang, Xiaoming Xie

## Rights

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

## Other metadata

[Isolating hydrogen in hexagonal boron nitride bubbles by a plasma treatment](https://mdr.nims.go.jp/datasets/f810d5f1-d3ea-4903-853e-af7f302bd48a)

## Fulltext

Isolating hydrogen in hexagonal boron nitride bubbles by a plasma treatmentARTICLEIsolating hydrogen in hexagonal boron nitridebubbles by a plasma treatmentLi He1,2,3, Huishan Wang2,3,4, Lingxiu Chen 2,3, Xiujun Wang2,3,4, Hong Xie2,3, Chengxin Jiang2,3,5, Chen Li6,11,Kenan Elibol7,12,13, Jannik Meyer 7,14, Kenji Watanabe 8, Takashi Taniguchi8, Zhangting Wu9,Wenhui Wang9, Zhenhua Ni9, Xiangshui Miao1, Chi Zhang1, Daoli Zhang 1, Haomin Wang 2,3,10 &Xiaoming Xie2,3,5Atomically thin hexagonal boron nitride (h-BN) is often regarded as an elastic film that isimpermeable to gases. The high stabilities in thermal and chemical properties allow h-BNto serve as a gas barrier under extreme conditions. Here, we demonstrate the isolation ofhydrogen in bubbles of h-BN via plasma treatment. Detailed characterizations reveal that thesubstrates do not show chemical change after treatment. The bubbles are found to withstandthermal treatment in air, even at 800 °C. Scanning transmission electron microscopyinvestigation shows that the h-BN multilayer has a unique aligned porous stacking nature,which is essential for the character of being transparent to atomic hydrogen but impermeableto hydrogen molecules. In addition, we successfully demonstrated the extraction of hydrogengases from gaseous compounds or mixtures containing hydrogen element. The successfulproduction of hydrogen bubbles on h-BN flakes has potential for further application in nano/micro-electromechanical systems and hydrogen storage.https://doi.org/10.1038/s41467-019-10660-9 OPEN1 School of Optical and Electronic Information, Huazhong University of Science and Technology, 430074 Wuhan, China. 2 State Key Laboratory of FunctionalMaterials for Informatics, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, 865 Changning Road, 200050Shanghai, China. 3 CAS Center for Excellence in Superconducting Electronics (CENSE), 200050 Shanghai, China. 4 Graduate University of the ChineseAcademy of Sciences, 100049 Beijing, China. 5 School of Physical Science and Technology, ShanghaiTech University, 319 Yueyang Road, 200031 Shanghai,China. 6 Department of Lithospheric Research, University of Vienna, Althanstraße 14, 1090 Vienna, Austria. 7 Faculty of Physics, University of Vienna,Boltzmanngasse 5, 1090 Wien, Austria. 8 National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan. 9Department of Physics, SoutheastUniversity, 211189 Nanjing, China. 10 Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, 100049Beijing, P.R. China. 11Present address: Electron Microscopy for Materials Research (EMAT), University Antwerpen, Groenenborgerlaan 171, 2020 Antwerpen,Belgium. 12Present address: School of Chemistry, Trinity College Dublin, Dublin 2, Ireland. 13Present address: Advanced Microscopy Laboratory, Centrefor Research on Adaptive Nanostructures and Nanodevices, Dublin 2, Ireland. 14Present address: Institute for Applied Physics and Natural and MedicalSciences Institute, University of Tübingen, Tübingen D-72076, Germany. Correspondence and requests for materials should be addressed toD.Z. (email: zhang_daoli@hust.edu.cn) or to H.W. (email: hmwang@mail.sim.ac.cn)NATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 |www.nature.com/naturecommunications 11234567890():,;http://orcid.org/0000-0002-0820-7175http://orcid.org/0000-0002-0820-7175http://orcid.org/0000-0002-0820-7175http://orcid.org/0000-0002-0820-7175http://orcid.org/0000-0002-0820-7175http://orcid.org/0000-0003-4023-0778http://orcid.org/0000-0003-4023-0778http://orcid.org/0000-0003-4023-0778http://orcid.org/0000-0003-4023-0778http://orcid.org/0000-0003-4023-0778http://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/0000-0003-0646-1572http://orcid.org/0000-0003-0646-1572http://orcid.org/0000-0003-0646-1572http://orcid.org/0000-0003-0646-1572http://orcid.org/0000-0003-0646-1572http://orcid.org/0000-0001-6388-2432http://orcid.org/0000-0001-6388-2432http://orcid.org/0000-0001-6388-2432http://orcid.org/0000-0001-6388-2432http://orcid.org/0000-0001-6388-2432mailto:zhang_daoli@hust.edu.cnmailto:hmwang@mail.sim.ac.cnwww.nature.com/naturecommunicationswww.nature.com/naturecommunicationsHexagonal boron nitride (abbreviated as h-BN in the fol-lowing) is a remarkable two-dimensional mesh consistingof alternating sp2-bonded boron and nitrogen atoms1. Asa wide bandgap insulator2, h-BN is regarded as an ideal substrateand tunneling barrier for graphene devices because of its atom-ically smooth surface free of dangling bonds and charge traps3. Inaddition, the exceptional thermal and chemical stabilities of h-BNenable it to function as an antioxidation layer, even at tempera-tures as high as 1100 °C4. Monolayer h-BN sheets remains stableup to 800 °C in air, whereas graphene oxidizes at temperatureabove 500 °C5. This makes h-BN a top choice for ultrathin anti-oxidization coatings. In addition to the potential for applicationin electronics and corrosion-proof coatings, h-BN holds promisefor selective membrane applications under extreme conditionsdue to its ultimate thinness, flexibility, and mechanicalstrength4,6–8.Similar to graphene9, h-BN is elastic and impermeable to allgases. High-quality h-BN sheets have a Young’s modulus of~0.85 TPa and an excellent fracture strength of ~70 GPa8. Thesheets can withstand stretching of up to 20%8. This feature makesh-BN a superb material for application in micro-electromechanical systems (MEMSs). A perfect h-BN crystal,even when just a monolayer thick, is typically impermeable tomost atoms and molecules under ambient conditions. Onlyaccelerated hydrogen atoms possess sufficient kinetic energy topenetrate through h-BN planes without damaging thelattices10,11. These properties allow bubbles with various shapes tobe produced on h-BN12. Recently, several reports have describedthe trapping of gases by graphene12–16. By forming bubblestructures, the adhesion energy between graphene sheets andsubstrate can be precisely measured17,18. A voltage-controlledgraphene micro-lens with variable focuses19 and a laser-drivennano-engine20 have been realized, as the curvature and shape ofthe bubbles can be controlled to a great extent by variation in thevoltage. Although h-BN has a similar honeycomb lattice to gra-phene, little investigation of h-BN bubbles for gas isolation hasbeen reported. Compared to graphene, h-BN is insulating andmore chemically stable, which may greatly extend its applicationin extreme conditions.There have been several reports about plasma treatment on 2Dmaterials. Most of them are about plasma applications in surfaceetching21,22, low-temperature growth23,24 and surface hydro-genation25. In this work, we demonstrate the isolation ofhydrogen in h-BN bubbles on the microscale via plasma treat-ment. The dimension of the bubbles can be regulated by alteringthe treatment conditions. The diameter of the bubbles rangesfrom tens of nanometers to several micrometers. The masses ofH2 molecules inside the bubbles are believed to have been con-verted to atomic hydrogen by the plasma, which permeatethrough the h-BN plane and is then captured inside the h-BNinterlayers. Re-formation of H2 molecules deforms the h-BNstructure to form bubbles, as the H2 molecules cannot escapefrom the h-BN net. Our demonstration provides an effectivemethod for fabricating bubbles on h-BN and may also be anapproach for extracting/storing hydrogen in h-BN.ResultsBubble formation on h-BN. Figure 1a shows a schematic illus-tration of the experimental design for the bombardment of theh-BN surface by plasma in different gaseous environments. Asshown in Fig. 1, three different gases (Ar, O2, and H2) were fedinto the chamber to examine the ability of different elementsto penetrate h-BN. The flow rate was kept at 3 sccm, whilethe pressure was kept at ~3 Pa. Plasmas were then generated bya RF generator with a power of 100W. All h-BN flakes weretreated with the plasma at 350 °C for 150 min. Atomic forcemicroscopy (AFM) images of the h-BN flakes were taken afterthe plasma treatment and are given in Fig. 1b–d (the corre-sponding optical images are given in Supplementary Fig. 1).As shown in Fig. 1b–d, the h-BN flakes treated with argon oroxygen plasma appear to be intact, while a large amount ofbubbles appeared on the h-BN surface that was treated with thehydrogen plasma (H-plasma). A series of experiments includingstructural characterization by transmission electron microscopy(TEM) (see Supplementary Figs. 2 and 3), chemical character-ization by X-ray photoelectron spectroscopy (XPS) (see Supple-mentary Figs. 4 and 5 and Supplementary Table 1), Energy-dispersive X-ray spectroscopy (EDX) (see Supplementary Figs. 6and 7), Fourier Transform infrared spectroscopy (FTIR) (seeSupplementary Fig. 8) and mass spectrometry (see Supplemen-tary Figs. 9 and 10) have been conducted to verify the h-BNh-BN bubbleArO2H2Substrateh-BNAfter treatmentPlasma treatment at 350 °Cb500.0 pm–500.0 pmc 1.0 nm–1.0 nmd150.0 nm–150.0 nmaAr atom O atom H atom H2Fig. 1 Production of bubbles on h-BN flakes via plasma treatment. a Schematic depicting the plasma treatment of h-BN flakes in different atmospheres.All AFM height images of the h-BN flakes were taken after plasma treatment in an atmosphere of b argon, c oxygen or d hydrogen. The h-BN flakesappear to remain intact after the argon and oxygen plasma treatments, while obvious bubbles are observed on the h-BN surface after treatment withthe H-plasma. All samples were obtained under similar conditions. Scale bars in (b–d) are 4 μmARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-92 NATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 | www.nature.com/naturecommunicationswww.nature.com/naturecommunicationsbubble structure intact and the presence of hydrogen gases insidethe bubbles.To understand the mechanism behind the interesting bubbleformation is important. It is well known that gaseous moleculessubjected to a strong electromagnetic field decompose into aplasma state, producing masses of charged ions whose particleradii are considerably decreased relative to the original atoms.The hexagonal honeycomb lattice of h-BN possesses a latticeconstant (ah-BN) of ~250.4 pm26. A hydrogen atom has a diameterof ~240 pm (with a van der Waals radius of ~120 pm27,28), whichis less than the value of ah-BN and could pass through the valenceelectrons concentrated around the BN atoms. Moreover, largeamounts of protons, which have a diameter of ~1.68 fm29,30, aregenerated in the plasma state. These extremely small particlescan easily penetrate the BN. The accelerated energetic protonsand electrons could then recombine to form hydrogen molecules.These hydrogen molecules are not easily able to pass throughthe h-BN honeycomb due to their large kinetic diameter of~289 pm31, which is substantially larger than ah-BN. Thus, thepressure from the trapped gas causes bubbles to form on thesurface of the h-BN flakes. In contrast, the atomic diameters ofboth argon (~376 pm) and oxygen (~304 pm) (with the van derWaals radii of ~188 pm and ~152 pm, respectively27,28) areconsiderably larger than ah-BN, and as a result, they cannotpenetrate the h-BN honeycomb barrier.The mechanism of unimpeded penetration of atomic hydrogeninto the h-BN interlayers can also be ascribed to two additionalreasons. The first one is the polarization of the electron densitydistribution in this kind of 2D material. Recent experiments havedemonstrated that h-BN possesses the best porosity among 2Dmaterials11. The concentration of valence electrons around theN atoms results in strong polarization of the BN bonds, makingh-BN much more porous than graphene and MoS2. The secondreason is the stacking order. It has previously been confirmed in theformer work that bi- and tri-layered graphene are also nonconduc-tive to protons due to their staggered lattices (AB stacked), inwhich the electron cloud of the first layer clogs the pores of thesecond11. In contrast, multilayered h-BN exhibit excellent protonconductivity probably due to its favoring the AA′ stacking structure,which has been found to be the most stable one among thefive possible stacking structures (AA′/AA/AB/AB′/A′B) in recentcalculations32. The aligned pores in the out-of-plane directionalso allow protons to move unimpededly throughout the space.This characteristic structure of multilayer h-BN could facilitate itsuse in hydrogen storage applications.To study the stacking order of the h-BN multilayer, a scanningtransmission electron microscopy (STEM) investigation wascarried out. The h-BN flakes are firstly exfoliated onto a quartzsubstrate, and then annealed to remove the residue of resist in anoxygen flow at 800 °C. After cooling down to room temperature,a group of wrinkles are found on the surfaces. The wrinkles,which are found along the armchair direction of h-BNcrystals33,34, help us determine the cutting direction of thefocused ion beam. The specimen obtained for the STEMinvestigation is similar to that shown in Supplementary Fig. 3a.Figure 2a illustrates that the incident electron beam (e-beam) forthe STEM investigation is parallel with the ½11�20� crystallographicaxis of h-BN. The cross-sectional schematic of the armchairedge-on view of h-BN multilayer is presented in Fig. 2b. Figure 2cis a Wiener filtered STEM medium angel annular-dark-field (MAADF) image showing the edge-on view configurationof h-BN multilayer, which reveals the layer stacking of the h-BN.The intensity profiles along two selected lines (red/blue) in Fig. 2care showed in Fig. 2d, exhibiting apparent synchronism of the peakcHAADFb0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4Intensity (a.u.)Distance (nm)0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4Intensity (a.u.)Distance (nm)daElectron beamFig. 2 STEM image showing the edge-on view configuration of an h-BN multilayer. a Schematic of a STEM investigation of the h-BN multilayer, the incidentelectron beam was parallel with its ½11�20� crystallographic axis (the zigzag edge of the h-BN layer). b Cross-sectional schematic of the armchair edge-onview of h-BN multilayer. The red atoms represent boron atoms while the blue ones represent nitrogen atoms. c STEM-MAADF image, postprocessed by aWiener Filter, showing the edge-on view configuration γ of an h-BN multilayer. Scale bar in (c) indicates 1 nm. d Profiles of image intensity along the redand blue arrows marked in (c), showing the stacking sequence in h-BN. Their average peak-to-peak distances are ~2.26 Å, which are in agreement with thelattice spacing of h-BN. The synchronism of the peak intensity along the arrows demonstrates the stacking structure of h-BN must be either AA′ or AANATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-9 ARTICLENATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 |www.nature.com/naturecommunications 3www.nature.com/naturecommunicationswww.nature.com/naturecommunicationspositions. Besides, the average peak-to-peak distances measured inred and blue profiles are both ~2.26 Å, which are close to 2.166 Åexpected in h-BN. The minor deviation of the spacing in STEMmeasurement might be due to the limitation of the scale calibrationof the instrument. This result shows that the stacking order of ourh-BN is either AA′ or AA in our h-BN multilayer, which is in linewith earlier predictions. Beside the cross-sectional STEM image, wealso carried out a plan-view STEM investigation (shown inSupplementary Fig. 11) of another h-BN multilayer. The imageconfirms that the h-BN multilayer we fabricated is in AA′ stacking.These images provide direct evidence that the multilayered planesin h-BN are not staggered so that the electron cloud of each layerdoes not block the pores of the successive layer. As such, atomichydrogen could tunnel through multilayered h-BN.Control on size of h-BN bubbles. Both the duration and theenvironmental temperature of plasma treatment have prominenteffects on the density and sizes of bubbles, which provide a fea-sible way to fabricate h-BN bubbles with controllable dimensions.To understand the dependence of the bubble dimensions on theduration of the H-plasma treatment, we varied the treatmentduration while holding the other conditions constant (100W RF,350 °C, 3 sccm H2, ~3 Pa) and examined the variations in thedimensions of the generated bubbles. The results of this experi-ment are given in Fig. 3. As shown in Fig. 3a, the h-BN samplesobtained in different duration of H-plasma treatment possessedbubbles with obviously different densities and size distributions.The bubbles treated for 90 min were generally smaller than 100nm in diameter and had a higher density in the distribution, whileplasma treatment for 150 min led to a lower bubble density and alarger bubble size with diameters greater than 3 μm. Thedependence of diameter distribution on treatment duration isplotted in Fig. 3b. Extended plasma treatment favors the forma-tion of larger bubbles. An H-plasma treatment for 500 min could0 2 40100200Diameter of bubble (µm)0.00.20.40 500 10000102030CountsDiameter of bubbles (nm)0 600 1200 1800 2400 3000 36000102030Counts0 800 1600 2400 3200 4000 4800010203090 min120 minCounts150 min90 120 150050100150200d > 3 µm1 µm < d < 3 µmDensity of bubbles (counts 10 µm–2)Treatment duration (min)0 µm < d < 1 µm30 150 250 3500100200300400500Density of bubbles (counts 10 µm–2)d > 2 µm1 µm < d < 2 µm100 nm < d < 1000 nmd < 100 nmhmax2RTreatment temperature (°C)h max (nm)h max/Ra bcdFig. 3 Influence of the plasma treatment duration and sample temperature on the dimensions of the bubbles. a Density and relative fractions of h-BNbubbles within three diameter ranges (0–1 μm, 1–3 μm, and >3 μm) produced by different H-plasma treatment durations; all samples were treated underthe typical experimental conditions (100W RF, 350 °C, 3 sccm H2, ~3 Pa). b Dimension distributions of the h-BN bubbles produced after H-plasmatreatment for 90, 120, and 150min. c Density and relative fractions of h-BN bubbles within four diameter ranges (<100 nm, 0.1–1 μm, 1–2 μm, and >2 μm)produced from treatment at different sample temperatures (30, 150, 250 and 350 °C); the other conditions were not changed (100W RF, 120minutes,3 sccm H2, ~3 Pa). d hmax of the bubbles produced by the H-plasma and the aspect ratio of the bubbles with respect to the bubble diameter. The insetshows an illustration of hmax and RARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-94 NATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 | www.nature.com/naturecommunicationswww.nature.com/naturecommunicationsfabricate larger bubbles. As shown in Supplementary Fig. 12a–c,bubbles with diameters of ~20 μm were obtained on h-BN.In addition to the treatment duration, the environmental/sample temperature is another key factor that influences thedimensions of the bubbles. Thus, the sample temperature was alsoadjusted while keeping the other conditions constant (100W RF,120 min, 3 sccm H2, ~3 Pa) to investigate the variations in thebubble dimensions. Figure 3c presents the density and diameterfractions of the h-BN bubbles for different sample temperatures.Almost all of the bubbles produced at room temperature (30 °C)were <100 nm in diameter, and the relative fraction of largebubbles gradually increased when the sample temperature waselevated from 30 to 350 °C. Meanwhile, the density of the bubblesdramatically decreased with increasing sample temperature.Furthermore, the bubble shown at the bottom-left of Supple-mentary Fig. 13d formed by merging three individual bubbles.The observation reveals that the bubbles can migrate and mergeat higher sample temperature. Similar phenomena were alsoobserved in Fig. 3a and Supplementary Fig. 14. The hydrogenmolecules inside smaller bubbles, normally possessing higherinner pressure than relatively large bubbles, move more easily athigher sample temperatures in-between h-BN interlayers15.Hydrogen gas in small bubbles is more likely to pass throughthe inter-planar channels and combine to form a larger bubblewith a decreased inner pressure. Such combination leads to adramatic decrease in bubble density at high sample temperatures.Detailed investigation in the stability of bubbles on h-BN is givenin Supplementary Figs. 15–20.We then further explored the relationship between the height/aspect ratio of the bubbles and their diameter. The maximumheight, hmax, and the quotient of hmax to the radius R (called theaspect ratio) of each bubble are taken from all observed h-BNbubbles by AFM. The results are presented in Fig. 3d. It is foundthat hmax approximately increases linearly with bubble diameter,while the aspect ratio remains almost unchanged.Mechanisms of the plasma-driven effect. In order to understandthe mechanism behind the plasma process, exploring hydrogenplasma itself is essential35. We carried out modeling of hydrogenplasma in a tube using COMSOL Multiphysics. The spatialdistribution of different parameters in the tube are plotted inSupplementary Fig. 21. As shown in Supplementary Fig. 21e, f, k,l, the ion density at the sample area decreases while the thermalvelocity increases, when the sample was heated from room tem-perature to 350 °C. The results indicate that protons in H-plasmacan be injected much faster into h-BN interlayers at a higherenvironmental temperature with comparatively lower spatialdensity. This simulation exhibits the variation tendency ofdimension/density when elevating the environmental tempera-ture (heated by the furnace) of the sample (as shown in Fig. 3c).The furnace temperature (Supplementary Fig. 21a, b), electrondensity (Supplementary Fig. 21c, d), electron temperature (Sup-plementary Fig. 21g, h), and pressure (Supplementary Fig. 21i, j)are also given to help in giving images shown plasma-drivenkinetics.It is noticed that the electron density (Supplementary Fig. 21c, d)is two orders in magnitude lower than the ion density(Supplementary Fig. 21e, f) at the sample position according tosimulation. This phenomenon indicates that the quasi-neutrality ofplasma at the sample position is broken. According to the setupshowed in Supplementary Fig. 22a–d, h-BN sample was placed30 cm away from the RF coil (the core region of the plasma). Alongthe axial direction of tube furnace, there is a steep electron gradientwhich leads to the unbalance between electron density and iondensity, and thus, the quasi-neutrality of plasma is not sustained.To understand this phenomenon further, we then carried outmodeling in Debye length when the furnace was set at 350 °C inSupplementary Fig. 21q. Normally, Debye length depends ondensity of ion number and environmental temperature. Debyelength increases when ion number density decreases or when theenvironmental temperature increases. It is also found that theDebye length, which is much higher than the diameter of the tube(0.042m), reached its maximum of ~0.5 m at the sample position. Itindicates the destruction of quasi-neutrality at the sample position.This also interprets why the electron density is two orders lowerthan the ion density at the sample position.Besides the sample temperature and treatment duration, theposition where the h-BN sample was placed also brings differentsize/distribution of bubble on h-BN. The schematic of experi-mental setup as well as the results of bubble distribution areexhibited in Supplementary Figs. 23 and 24, respectively. It isobvious that bubbles in the plasma core area are small butdense. while bubbles at other area show relatively large but thin.The electric potential difference (Supplementary Fig. 21m, n) andthe ion number density distribution (Supplementary Fig. 21o, p)are also simulated by COMSOL Multiphysics to analyze thebubble distributions in the supporting information. Obviously,the formation of bubbles is a plasma-driven effect. The sizeand distribution of bubbles on h-BN are generally determinedby the comprehensive influence of ion density, Debye lengthand potential gradient at the position where the h-BN samplesare placed.Low-temperature AFM measurement. To determine the type ofgases inside the h-BN bubbles, low-temperature AFM measure-ments were carried out to explore their temperature evolution. Theexperiment was inspired by a very recent literature about bubblestructure on bulk transition metal dichalcogenides36. Details aboutmeasurement are given in the Methods section. Figure 4a shows anoptical image of bubbles on an h-BN flake. After transferring it intoa vacuum chamber equipped with an AFM, the sample was cooleddown gradually. Figure 4b shows topographic AFM images of thesame area measured at 34 and 33K respectively. It is clearly shownthat the bubbles deflate at 33 K while they are inflated with gases at34 K. As shown in Fig. 4c, the height profiles of line-scans across abubble (indicated by dashed lines in Fig. 4b) clearly show that thebubble remains at ~34K and disappears at ~33K. The deflating/swelling processes were reproducible via cooling/heating the sam-ple. The highest temperature at which the bubbles become flat wasrecognized as the transition temperature (Ttransition). We measured atotal of 58 bubbles. Three other typical samples are given in Sup-plementary Figs. 25–27. The transition temperature of each bubblewas recorded when they become flat during cooling. The number ofbubbles which have the same Ttransition is counted. The histogram ofTtransition as function of temperature at an interval of 1 K is plottedin Fig. 4d. As shown in Fig. 4d, Ttransition of all the bubbles measuredcan be described as a Gaussian distribution, and exhibits an averagevalue equal to 33.2 ± 3.9 K. This value is very close to Ttransition of H2(33.18 K) among all possible gases (see Supplementary Table 2).The result strongly indicates that the gases inside the bubbles areindeed hydrogen molecules.Determining the penetration depth of atomic hydrogen in theh-BN multilayer. Increasing the power of the RF source may leadto bursting of the bubbles, which leaves many pits on the h-BNsurface. These pits make it possible to study the penetration depthof atomic hydrogen. Figure 5a shows an optical micrograph ofan h-BN flake treated for 180 min at 350 °C with an RF powerof 400W. Numerous pits are observed on the h-BN surface.Figure 5b, c depict the AFM image of a deep pit highlighted byNATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-9 ARTICLENATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 |www.nature.com/naturecommunications 5www.nature.com/naturecommunicationswww.nature.com/naturecommunicationsthe red box in Fig. 5a. The AFM height image in Fig. 5b providesdepth information about the pit, whose depth is given in theinsert. The amplitude error image provided in Fig. 5c shows aclearer picture of the morphology around the pit. The bottom ofthe pit is smooth, which indicates that no atomic hydrogenpenetrated the bottom layer. Furthermore, folds are observed inthe h-BN flaps around the outside of the pit, as illustrated inFig. 5d. The inset in Fig. 5b shows that the rupture depth of thefolded border step is approximately 15.22 nm, which indicatesthat ~23 layers of h-BN were penetrated by atomic hydrogen inthe 400W treatment. Figure 5e presents the statistics of thepenetration depth of atomic hydrogen in the h-BN sample shownin Fig. 5a. The penetration depth of the h-BN flake ranges from~11 to ~23 layers. The histogram shown in Fig. 5e reveals thatthe 400W RF power provided the H-plasma with enoughkinetic energy to penetrate at most ~23 atomic layers underneaththe h-BN surface.Characterizations of Raman spectroscopy. Raman spectroscopywas also performed to characterize the h-BN bubble. As shownin Fig. 6a, a typical bubble with a round shape was measured. Thebubble had a diameter of ~2.36 μm and a height of ~122 nm. TheRaman spectrum was measured at three spots from the flatarea next to the bubble to the center of the bubble. The resultsRuptured depthh -BN bubbleh -BN layer11 13 15 17 19 21 230510152040.0 nm–40.0 nm2.0 nm–2.0 nmCountsThe number of layersa bd15.22 nmeSubstrate SubstratecFig. 5 Penetration depth of atomic hydrogen into h-BN multilayers. a Optical image of an h-BN flake on which the bubbles generated by a 400W H-plasmatreatment ruptured. h-BN flaps peeled off the adhered macroscopic films, and the multilayer h-BN sheets tore and then folded. b AFM height andc amplitude error images taken of the red area shown in (a). The bottom of the pit is atomically flat, and no bubble filled with hydrogen molecules wasfound. The insert in (b) is the AFM depth profile along the yellow dashed line. The deviation in the height is approximately 15.2 nm, indicating that ~23atomic layers were ruptured and folded over the top layer of h-BN. The scale bars in (a–c) represent 1 μm. d Illustration of the ruptured depth for the foldingsteps depicted in (b, c) caused by the bursting of the h-BN bubble. e Depth distribution of penetration into the surface of the h-BN flake shown in (a)25 30 35 40 450481216Cumulative function (arb. units)Bubbles countsBubbles countsGauss fitCumulative function0102030405060a cbd33 K34 K0 2 4 6 810.6510.7010.7510.8010.85 34 K33 KHeight (µm)Ttransition (K)x (µm)Fig. 4 Swelling and deflating processes of the h-BN bubbles containing hydrogen. a An optical image of bubbles on an h-BN flake, taken under ambientcondition, scale bar: 20 μm. b Topographic AFM image of a bubble pointed-out by an arrow in (a) was measured at 34 and 33 K respectively, scalebars: 3 μm. c The height profiles of line-scan at the same place (indicated by dashed lines in (b)) where the bubble remains at ~34 K and disappearsat ~33 K. d Histogram of the transition temperature (Ttransition) at which bubbles collapse. The red line is a Gaussian fit to the data. The yellow line is thehistogram cumulative function (right axis)ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-96 NATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 | www.nature.com/naturecommunicationswww.nature.com/naturecommunicationsare shown in Fig. 6b. The Raman spectra of h-BN clearly showonly one peak near 1366 cm–1, which corresponds to the E2gvibration mode of h-BN. As shown in the inset of Fig. 6b, the E2gpeak position and full width at half maximum (FWHM) of eachspectrum were extracted and plotted as a function of the mea-surement site. The peak position in the Raman spectrumappeared at the center of the bubble redshifted (~3 cm–1) relativeto that of the position outside of the bubble, with almost nochange in the FWHM. The observed redshift in the peak positionis due to the presence of stretching strain on the bubble caused byexpansion37,38. Further investigation in the temperature depen-dence of Raman spectra of h-BN bubble is given in Supplemen-tary Fig. 28.Hydrogen extraction. The demonstrated fabrication of hydrogenbubbles on h-BN via plasma treatment inspired us to consider thepossibility of extracting hydrogen from hydrocarbons and mix-tures of gases. Three different gases (C2H2, CH4 and Ar+H2,5% of H2) were fed into the chamber for plasma treatmentunder the same experimental conditions (350 °C, 100W, 120 min,~3 sccm flow rate, ~3 Pa). Excitingly, all three samples formedbubble structures on the h-BN surface after the plasma treatment.The experimental illustration and AFM images of the h-BNflakes are given in Fig. 7a–f. Considering that the diameters ofcarbon and argon atoms are much larger than the lattice constantof h-BN, it is likely that only atomic hydrogen could penetrate theh-BN layers. The results indicate that hydrogen could be sepa-rated from hydrocarbons and from the mixture of argon andhydrogen.As shown in Fig. 7d–f, the bubbles formed after the ethyneplasma treatment had diameters of predominately <300 nm and ahigh density, while the bubbles produced by the methane plasmapossessed a much larger diameter range from 2 to 4 μm with amuch lower density. The exact causes for the phenomena needcomprehensive investigation in future. At this moment, the mainreason, we believe, could be that the dissociation energy ofmethane is actually lower than that of the ethyne39. When allexperiments were carried out under similar conditions, moreatomic hydrogen in CH4 plasma can be generated and then moreefficiently get into h-BN bubbles than the case in ethyne plasma.The plasma treatment with 5% hydrogen in argon also led to theformation of bubbles on the h-BN surface but with an extremelylow density. To characterize the bubbles further, the plots of hmaxand the aspect ratio with respect to the bubble diameter arepresented in Fig. 7g–i. Similar to the bubbles fabricated by theplasma of pure hydrogen, in all three plots, hmax had a nearlylinear increasing relationship with the bubble diameter, while theaspect ratio remained nearly unchanged and thus was indepen-dent of the bubble diameter. Interestingly, although the bubblesproduced from CH4 had much larger diameters (2–4 μm) thanthose produced from C2H2 (<300 nm), they had almost the samehmax (<10 nm), resulting in an extremely low aspect ratio (hmax/R= ~0.005) for the bubbles produced from methane.DiscussionWe demonstrate an approach for isolating hydrogen in h-BNbubbles by plasma treatment. The dimensions of the bubbles rangefrom nanometers to micrometers and depend on the durationand sample temperature of the plasma treatment. Most of thebubbles undergo minimal changes over 40 weeks. Additionally,the penetration depth of atomic hydrogen reaches approximately23 h-BN layers. The Raman spectra of a bubble reveal thatstretching occurs on the bubble surface. Finally, we demonstratethat hydrogen gas can be extracted from hydrocarbons and mixedgases by the plasma treatment and isolated between h-BN layersin bubbles. The demonstrated ability to fabricate hydrogen bubbleson h-BN flakes might be used to extract hydrogen from hydro-carbon/gas mixtures or to fabricate nano/micro-electromechanicalsystems.MethodsFabrication and characterization of hydrogen bubbles on h-BN. The experi-mental setup is given in Supplementary Fig. 22. First, h-BN flakes were producedby micromechanical exfoliation of single-crystal h-BN and then deposited onquartz substrates that were previously cleaned with oxygen plasma. The substrateswith h-BN flakes were loaded into a furnace equipped with an RF generator(13.56 MHz, MTI Corporation) with a tunable power from 100 to 400W. Thesample temperature in the sample chamber could be controlled from room tem-perature to 1000 °C. An advanced pump (GX100N Dry Pumping System, Edwards)was connected to the chamber to control the flow rate. After the samples wereheated to the preset temperature, hydrogen (oxygen or argon) gas with a flow rateof 3 sccm (~3 Pa) was introduced into the chamber, and then the plasma wasformed. The plasma treatment normally took 90, 120 or 150 min. The h-BNflakes were finally removed from the sample chamber for characterization underan optical microscope (Eclipse LV150, Nikon), an atomic force microscope(Dimension Icon, Bruker), and a Raman spectrometer (532 nm, WITec).Low-temperature AFM measurement of the bubbles. After bubbles were pro-duced on h-BN flakes, optical microscope and AFM were used to locate theirpositions. Subsequently, the substrate with the h-BN flakes was transferred into thevacuum chamber of a variable temperature SPM (Scanning probe microscope)system (attoDRY 1100 by Attocube). The chamber was filled with 4He in a pressureof 5 mbar for heat exchange. The sample temperature can be controlled in therange from 4 to 300 K. All the AFM scanning was conducted in contact mode.CenterMarginRaman intensity (a.u.)Wavenumber (cm–1)Flat areab136213641366CenterMarginFlat areaRaman spectrum151821FWHM (cm–1)2.36 µm122 nma150.0 nm–110.0 nmCharacteristicpeak (cm–1)1000 1500 2000Fig. 6 Raman spectra taken at different positions of an h-BN bubble. a AFM image of a typical h-BN bubble; the white curve shows the profile across thecenter of bubble and gives information about its height and diameter. Scale bar, 600 nm. b Raman spectra taken at the positions indicated in the AFMimage. The inset shows the variation in the E2g peak position and FWHM between positions. A redshift in the E2g peak position from the flat area to thecenter of the bubble is observed, while the corresponding FWHM is nearly unchangedNATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-9 ARTICLENATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 |www.nature.com/naturecommunications 7www.nature.com/naturecommunicationswww.nature.com/naturecommunicationsIn order to minimize the damage to the bubbles, AFM tips with a force constant of0.02–0.77 N m−1 were used. The sample was cooled at a rate of 0.5 K h−1 whileAFM scanning was carried on the h-BN bubbles. The highest temperature at whichthe bubbles became flat was recorded as the transition temperature. Afterwards, thesamples were warmed up for several Kelvins to observe the expansion of bubbles inorder to exclude the possible case of gas leakage.Scanning transmission electron microscopy. Electron microscopy experimentswere conducted using a Nion UltraSTEM100 scanning transmission electronmicroscope, operated at 60 kV in near-ultrahigh vacuum (2 × 10−7 Pa). The beamcurrent during the experiments varied between 8 and 80 pA, corresponding to doserates of ∼5–50 × 107 e Å−2 s−1. The beam convergence semi-angle was 35 mradand the semi-angular range of the medium-angle annular-dark-field detector was60–200 mrad.Data availabilityThe data that support the findings of this study are available from the correspondingauthors on request.Received: 22 April 2018 Accepted: 13 May 2019References1. Corso, M. et al. Boron nitride nanomesh. Science 303, 217–220 (2004).2. Kubota, Y., Watanabe, K., Tsuda, O. & Taniguchi, T. Deep ultravioletlight-emitting hexagonal boron nitride synthesized at atmospheric pressure.Science 317, 932–934 (2007).3. Dean, C. R. et al. Boron nitride substrates for high-quality grapheneelectronics. Nat. Nanotechnol. 5, 722–726 (2010).4. Liu, Z. et al. Ultrathin high-temperature oxidation-resistant coatings ofhexagonal boron nitride. Nat. Commun. 4, 2541 (2013).5. Liu, L. et al. Graphene oxidation: thickness-dependent etching and strongchemical doping. Nano Lett. 8, 1965–1970 (2008).6. Xu, M., Liang, T., Shi, M. & Chen, H. Graphene-like two-dimensionalmaterials. Chem. Rev. 113, 3766–3798 (2013).7. Li, L. H., Cervenka, J., Watanabe, K., Taniguchi, T. & Chen, Y. Strongoxidation resistance of atomically thin boron nitride nanosheets. ACS Nano 8,1457–1462 (2014).8. Falin, A. et al. Mechanical properties of atomically thin boron nitride andthe role of interlayer interactions. Nat. Commun. 8, 15815 (2017).9. Bunch, J. S. et al. Impermeable atomic membranes from graphene sheets.Nano Lett. 8, 2458–2462 (2008).10. Lozada-Hidalgo, M. et al. Sieving hydrogen isotopes through two-dimensionalcrystals. Science 351, 68–70 (2016).11. Hu, S. et al. Proton transport through one-atom-thick crystals. Nature 516,227–230 (2014).12. Khestanova, E., Guinea, F., Fumagalli, L., Geim, A. K. & Grigorieva, I. V.Universal shape and pressure inside bubbles appearing in van der Waalsheterostructures. Nat. Commun. 7, 12587 (2016).13. Boddeti, N. G. et al. Graphene blisters with switchable shapes controlledby pressure and adhesion. Nano Lett. 13, 6216–6221 (2013).14. Liu, X. et al. Observation of pull-in instability in graphene membranesunder interfacial forces. Nano Lett. 13, 2309–2313 (2013).15. Stolyarova, E. et al. Observation of graphene bubbles and effective masstransport under graphene films. Nano Lett. 9, 332–337 (2009).16. Zabel, J. et al. Raman spectroscopy of graphene and bilayer under biaxialstrain: bubbles and balloons. Nano Lett. 12, 617–621 (2012).17. Koenig, S. P., Wang, L., Pellegrino, J. & Bunch, J. S. Selective molecularsieving through porous graphene. Nat. Nanotechnol. 7, 728–732 (2012).120 160 200 240246h max (nm)h max (nm)h max/Rh max (nm)h max/Rh max/RDiameter of bubble (nm)0.00.10.22468Diameter of bubble (µm) Diameter of bubble (µm)0.000.010.020.5010200.000.050.10BubbleBubblea bd eg h ih -BNSubstrate3.0 nm 8.0 nm 2.0 nm–8.0 nm –2.0 nm–3.0 nmSubstrate SubstrateC atom H atom Ar atomc3 4 1.0 1.5fFig. 7 Extracting hydrogen from hydrocarbon and mixture gases. a–c Schematic illustration of the separation of atomic hydrogen from ethyne (C2H2),methane (CH4) and a mixture of hydrogen and argon (Ar+H2, 5% H2) by an h-BN membrane via plasma treatment. d–f AFM height images of the h-BNsurfaces treated by a typical plasma process (350 °C, 100W, 120min, ~3 sccm, ~3 Pa) in an environment of ethyne, methane and a mixture of hydrogenand argon. Scale bars: d 600 nm, e, f 4 μm. g–i hmax of the bubbles fabricated via the plasma treatment in C2H2, CH4 and Ar+H2 (5% of H2), respectively,and their aspect ratios as a function of the bubble diameterARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-98 NATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 | www.nature.com/naturecommunicationswww.nature.com/naturecommunications18. Koenig, S. P., Boddeti, N. G., Dunn, M. L. & Bunch, J. S. Ultrastrong adhesionof graphene membranes. Nat. Nanotechnol. 6, 543–546 (2011).19. Georgiou, T. et al. Graphene bubbles with controllable curvature. Appl. Phys.Lett. 99, 093103 (2011).20. Lee, J. H. et al. Nanometer thick elastic graphene engine. Nano Lett. 14,2677–2680 (2014).21. Xie, L., Jiao, L. & Dai, H. Selective etching of graphene edges by hydrogenplasma. J. Am. Chem. Soc. 132, 14751–14753 (2010).22. Xiao, S. et al. Atomic-layer soft plasma etching of MoS2. Sci. Rep. 6, 19945(2016).23. Malesevic, A. et al. Synthesis of few-layer graphene via microwaveplasma-enhanced chemical vapour deposition. Nanotechnology 19, 305604(2008).24. Kim, J. et al. Low-temperature synthesis of large-area graphene-basedtransparent conductive films using surface wave plasma chemical vapordeposition. Appl. Phys. Lett. 98, 091502 (2011).25. Elias, D. C. et al. Control of graphene’s properties by reversible hydrogenation:evidence for graphane. Science 323, 610–613 (2009).26. Lynch, R. W. & Drickamer, H. G. Effect of high pressure on the latticeparameters of diamond, graphite, and hexagonal boron nitride. J. Chem. Phys.44, 181–184 (1966).27. Bondi, A. Van Der Waals volumes and radii. J. Phys. Chem. 68, 441–451(1964).28. Rowland, R. S. & Taylor, R. Intermolecular nonbonded contact distances inorganic crystal structures: comparison with distances expected from van derWaals radii. J. Phys. Chem. 100, 7384–7391 (1996).29. Beyer, A. et al. The Rydberg constant and proton size from atomic hydrogen.Science 358, 79–85 (2017).30. Vassen, W. The proton radius revisited. Science 358, 39–40 (2017).31. Mehio, N., Dai, S. & Jiang, D. E. Quantum mechanical basis for kineticdiameters of small gaseous molecules. J. Phys. Chem. A 118, 1150–1154(2014).32. Constantinescu, G., Kuc, A. & Heine, T. Stacking in bulk and bilayerhexagonal boron nitride. Phys. Rev. Lett. 111, 036104 (2013).33. Chen, L. et al. Oriented graphene nanoribbons embedded in hexagonal boronnitride trenches. Nat. Commun. 8, 14703 (2017).34. Tang, S. et al. Silane-catalysed fast growth of large single-crystalline grapheneon hexagonal boron nitride. Nat. Commun. 6, 6499 (2015).35. Felizardo, E. et al. Energetic hydrogen atoms in wave driven discharges.Appl. Phys. Lett. 99, 041503 (2011).36. Tedeschi, D. et al. Proton-driven patterning of bulk transition metaldichalcogenides. Preprint at https://arXiv.org/abs/1803.09825 (2018).37. Gorbachev, R. V. et al. Hunting for monolayer boron nitride: optical andRaman signatures. Small 7, 465–468 (2011).38. Cai, Q. et al. Raman signature and phonon dispersion of atomically thin boronnitride. Nanoscale 9, 3059–3067 (2017).39. Partridge, H. & Bauschlicher, C. W. The dissociation energies of CH4 andC2H2 revisited. J. Chem. Phys. 103, 10589–10596 (1995).AcknowledgementsThe work was partially supported by the National Key R&D program (Grant No.2017YFF0206106), the Strategic Priority Research Program of Chinese Academy ofSciences (Grant No. XDB30000000), the National Science Foundation of China(Grant Nos. 51772317, 51302096), the Science and Technology Commission ofShanghai Municipality (Grant No. 16ZR1442700), the Hubei Provincial NaturalScience Foundation of China (Grant No. ZRMS2017000370), and the FundamentalResearch Funds of Wuhan City (No. 2016060101010075). K.W. and T.T. acknowledgesupport from the Elemental Strategy Initiative conducted by the MEXT, Japan andJSPS KAKENHI Grant Numbers JP15K21722. C.L. acknowledges support from theEuropean Union’s Horizon 2020 research and innovation program under the MarieSkłodowska-Curie Grants No. 656378—Interfacial Reactions. L.H. acknowledges finan-cial support from the program of China Scholarships Council (No. 201706160037).H.W. and D.Z. thank Y. Gu, Y. Ma, X. Chen (Shanghai Institute of TechnicalPhysics, Chinese Academy of Sciences) for FTIR spectra measurement. L.C. and L.H.thank Q. Liu and Z. Liu (Shanghai Institute of Microsystem and InformationTechnology, Chinese Academy of Sciences) for measurement in XPS spectra andmass spectra.Author contributionsH.W. conceived and designed the research. L.H. fabricated the samples and performedthe measurements. L.C., C.Z. and Huishan W. assisted with the transfer techniqueand AFM measurements. C.J. carried out low-temperature AFM measurements. X.W.and H.X. provided assistance with substrate fabrication. Z.W., W.W. and Z.N. performedthe Raman measurements. C.L., K.E., J.M. carried out STEM measurements. K.W.and T.T. fabricated the h-BN crystals. L.H., C.L. X.X., X.M., D.Z. and Haomin W.prepared the manuscript. All authors discussed the results and commented on themanuscript.Additional informationSupplementary Information accompanies this paper at https://doi.org/10.1038/s41467-019-10660-9.Competing interests: The authors declare no competing interests.Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/Peer review information: Nature Communications thanks Marcelo Lozada-Hidalgo andother anonymous reviewer(s) for their contribution to the peer review of this work. Peerreviewer reports are available.Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims inpublished 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, as long as you giveappropriate credit to the original author(s) and the source, provide a link to the CreativeCommons license, and indicate if changes were made. The images or other third partymaterial in this article are included in the article’s Creative Commons license, unlessindicated otherwise in a credit line to the material. If material is not included in thearticle’s Creative Commons license and your intended use is not permitted by statutoryregulation or exceeds the permitted use, you will need to obtain permission directly fromthe copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2019NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-10660-9 ARTICLENATURE COMMUNICATIONS |         (2019) 10:2815 | https://doi.org/10.1038/s41467-019-10660-9 |www.nature.com/naturecommunications 9https://arXiv.org/abs/1803.09825https://doi.org/10.1038/s41467-019-10660-9https://doi.org/10.1038/s41467-019-10660-9http://npg.nature.com/reprintsandpermissions/http://npg.nature.com/reprintsandpermissions/http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/www.nature.com/naturecommunicationswww.nature.com/naturecommunications Isolating hydrogen in hexagonal boron nitride bubbles by a plasma treatment Results Bubble formation on h-BN Control on size of h-BN bubbles Mechanisms of the plasma-driven effect Low-temperature AFM measurement Determining the penetration depth of atomic hydrogen in the h-BN multilayer Characterizations of Raman spectroscopy Hydrogen extraction Discussion Methods Fabrication and characterization of hydrogen bubbles on h-BN Low-temperature AFM measurement of the bubbles Scanning transmission electron microscopy References References Acknowledgements Author contributions Competing interests ACKNOWLEDGEMENTS