Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Light-induced irreversible structural phase transition in trilayer graphene

Light-induced irreversible structural phase transition in trilayer graphene A crystal structure has a profound influence on the physical properties of the corresponding material. By synthesizing crystals with particular symmetries, one can strongly tune their properties, even for the same chemical configuration (compare graphite and diamond, for instance). Even more interesting opportunities arise when the structural phases of crystals can be changed dynamically through external stimulations. Such abilities, though rare, lead to a number of exciting phenomena, such as phase-change memory effects. In the case of trilayer graphene, there are two common stacking configurations (ABA and ABC) that have distinct electronic band structures and exhibit very different behaviors. Domain walls exist in the trilayer graphene with both stacking orders, showing fascinating new physics such as the quantum valley Hall effect. Extensive efforts have been dedicated to the phase engineering of trilayer graphene. However, the manipulation of domain walls to achieve precise control of local structures and properties remains a considerable challenge. Here, we experimentally demonstrate that we can switch from one structural phase to another by laser irradiation, creating domains of different shapes in trilayer graphene. The ability to control the position and orientation of the domain walls leads to fine control of the local structural phases and properties of graphene, offering a simple but effective approach to create artificial two-dimensional materials with designed atomic structures and electronic and optical properties. Introduction the bottom layer (denoted as rhombohedral or ABC 9,10 The stacking configuration of layered materials plays stacking) . Due to different interlayer electron an important role in determining their electronic and interactions and distinct crystal symmetry, it has been optical properties. Fascinating phenomena, such as shown that ABA-stacked TLG and ABC-stacked TLG Hofstadter’s butterfly, Mott insulators, ferromagnetism, exhibit significantly different physical properties. From and unconventional superconductivity, can also emerge the perspective of the electronic band structure, ABA- in van der Waals heterostructures by carefully con- stacked TLG is a semimetal with a gate-tunable band 1–8 trolling the layer stacking sequence .Inthe case of overlap between the valence and conduction bands, trilayer graphene (TLG), there are two common stack- whereas the ABC-stacked TLG is a semiconductor with 11–15 ing configurations: the top layer may lie directly above an electrically tunable band gap .InTLG flakes the bottom layer (denoted as Bernal or ABA stacking) containing both ABA and ABC stacking, there are or may instead lie above the center of the hexagon of domain walls between the phases, consisting of a localized strain soliton in which the carbon atoms of one graphene layer shift by the carbon–carbon bond Correspondence: Jiayu Dai (jydai@nudt.edu.cn)or 16,17 distance . Such domain walls in TLGs have attracted Mengjian Zhu (zhumengjian11@nudt.edu.cn) much interest because of their intriguing physical Department of Physics, National University of Defense Technology, 410073 Changsha, China properties. For example, optically, soliton-dependent College of Advanced Interdisciplinary Studies, National University of Defense reflection of graphene plasmons at the domain walls Technology, 410073 Changsha, China has been experimentally observed . Electrically, the Full list of author information is available at the end of the article These authors contributed equally: Jianyu Zhang, Jinsen Han, Gang Peng © The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Zhang et al. Light: Science & Applications (2020) 9:174 Page 2 of 11 domain walls are predicted to host topological edge between the topmost layers in the two allotropes . Gra- states and ballistic transport and can also produce in- phene trilayer was obtained by mechanical exfoliation and plane metal–semiconductor (ABA–ABC) homojunc- confirmed by optical contrast, and the thickness 19–21 tions in TLG . (~1.2 nm) was determined by AFM, as shown in Fig. 1b, c. Previous reports have shown that applying molecular Raman spectroscopy has been demonstrated to be an absorption or an external electric field can drive the accurate and effective method to distinguish the ABA and 10,24–26 stacking order transition and generate domain wall ABC stacking structures in TLG . TLG flakes show 17,22,23 motion in graphene layers . There is inevitable very uniform optical contrast without any visible domain residue on graphene with molecular doping, which hin- walls or wrinkles. However, the Raman mapping exhibits ders the properties of graphene. Applying an electrical two distinct domains with significant contrast due to the field or strain usually leads to global control of the different stacking orders in TLG, as shown in Fig. 1b 10,25 stacking order phase and hinders precise manipulation of (here, integrated G band intensity is presented) . The the local structure. An alternative way to change the darker domain was identified as ABA-stacked TLG, and stacking configuration is by applying a local mechanical the brighter domain was identified as ABC stacking. force. For example, a previous study demonstrated that Raman spectra of the two different domains are plotted domain walls in TLGs can be moved by mechanical stress for comparison in Fig. 1d. The spectra are different from exerted through an atomic force microscopy (AFM) tip . one another in at least three ways: first, the 2D band of The domain walls are invisible in conventional AFM ABC-stacked TLG shows more asymmetric features with topography, and studies must rely on near-field infrared an enhanced peak and shoulder compared with the nanoscopy measurements. However, a simple and con- symmetric feature shown in ABA-stacked TLG; second, trollable approach to engineer the stacking phase and the G band of the ABC-stacked domain is redshifted by −1 domain walls into designed atomic structures is still ~1 cm compared with that of the ABA-stacked domain; lacking. and third, the G’ band of the ABC-stacked TLG domain Here, we experimentally demonstrate that the stacking also exhibits more asymmetric features than its ABA- 10,25,27 order in TLG can be switched from ABC to ABA by local stacked counterpart . heating enabled through laser irradiation. The light- We will focus on the integrated G band intensity induced stacking phase transition in TLG is directly mappings (Fig. 1b), which show the lowest noise level visualized using Raman mapping and near-field nano- (mappings of bandwidth of G bands and 2D yield con- scopy imaging. By controlling the movement of the laser sistent shape of domains, as shown in Fig. S1). In this beam with considerable flexibility and precision, we are work, we prepared 211 TLG flakes. Among them, 147 able to reshape the domains and manipulate the position flakes are pure ABA-stacked TLG, and the remaining 64 and orientation of the domain walls in the TLG. We flakes have coexisting ABA- and ABC-stacked domains, as attribute the laser-induced local heating effect as the main shown in Fig. S2. The proportion of ABC stacking in TLG 10,28 driving force of the ABC-to-ABA phase transition. The is ~15%, consistent with previous reports . activation energy is determined by Raman spectroscopy An attractive target for optical materials is to find a measurements and thermal annealing experiments and is system that shows the structural phase transition trig- 29–31 consistent with the calculated energy barrier height of gered by external stimulation of light . Laser irradia- approximately 40 meV determined by density functional tion has been demonstrated as an effective method to theory (DFT) calculations. The electronic and optical induce structural phase transitions in two-dimensional properties of TLG strongly depend on the stacking con- materials, for instance, laser-driven 2H-to-1T’ phase 32–34 figuration. Therefore, the ability to achieve fine control of transitions in few-layer MoTe . Here, we extend this the local stacking configuration and manipulate the methodology to control the stacking order transformation domain walls by a simple and clean approach opens the in TLG. A continuous laser beam was scanned over the way to new devices with fascinating functionalities, such TLG sample under ambient conditions, as schematically as multilevel optical switch and phase-change memory. shown in Fig. 2a (see the “Materials and methods” section for more details). The sample was illuminated by a laser Results and discussion with different powers from 1 to 20 mW. First, the laser The schematics in Fig. 1a shows the crystalline struc- beam moved from left to right to finish one line scan. tures of TLG with ABA and ABC stacking orders. The After that, the laser returned to the left and moved atoms of the topmost layer in ABA-stacked TLG lie downward to start the next line scan until all scans are exactly above of those of the bottom layer, whereas in completed. After finishing each laser scan, the sample was ABC-stacked TLG, the sublattice of the top layer lies again characterized by Raman mapping with a laser power above the center of the hexagons in the bottom layer. of 1 mW. Figure 2b summarizes the Raman mappings of There is a parallel shift of exactly one carbon honeycomb the integrated G band intensity, showing the lapsed Zhang et al. Light: Science & Applications (2020) 9:174 Page 3 of 11 A B C A B bc 010 20 0 X (μm) GG′ 2D ABA ABC 1560 1580 1600 2400 2450 2500 2600 2700 2800 1400 1600 1800 2000 2200 2400 2600 2800 –1 Raman shift (cm ) Fig. 1 Characterizations of mechanically exfoliated TLG. a Schematics of graphene trilayers with ABA stacking (left) and ABC stacking (right) configurations. The bottom, middle, and top layers are labeled with different colors. b Optical microscopy image of TLG sample #2 and the corresponding Raman mapping of the integrated G band intensity. The darker domain indicates the ABA-stacked TLG, and the brighter domain was defined as ABC stacking. The scale bars are 4 μm, and the color bar shows the integrated Raman intensity. c AFM height profile of TLG measured along the green dashed line in the optical image in (b). d Raman spectra of TLG taken from different regions marked in the Raman mapping image in −1 ’ −1 −1 (b). The insets show magnified spectra of TLG: G band (1560–1600 cm ), G band (2400–2500 cm ), and 2D band (2600–2800 cm ) dynamic process of the phase transition from ABC geometric shapes of the domain walls and the angle stacking to ABA stacking. The domain wall started to between the laser scanning direction and the domain move from the ABA-stacked domain to the ABC-stacked walls, as shown in Fig. S5. We noticed that the movement domain under 10 mW laser irradiation. As the laser power of the ABA/ABC domain walls in TLG was similar to the increased, the domain wall gradually shifted from left to bilayer graphene case. In bilayer graphene, the AB/AC right, showing a reduced ABC stacking area and an stacking boundaries were observed as nanometer-wide expanded ABA stacking region. The Raman mappings of strained channels, mostly in the form of ripples, produ- the integrated 2D band intensity show the same trans- cing smooth low-energy transitions between the two dif- formation process, as shown in Fig. S3. Furthermore, we ferent stacks . found that if the laser scan zone contained the domain Figure 2cshows thesignificant changes in the Raman walls, then the ABC-to-ABA stacking order transition spectra of TLG before and after laser irradiation. It was always initiated from the domain wall rather than ran- evident that the 2D band of the ABC domain became domly occurring in the TLG. However, the ABC-to-ABA more symmetric after laser irradiation, which agrees phase transition can also occur in the pure ABC-stacked with the 2D band features of the initial ABA-stacking TLG region, as shown in Fig. S4. Notably, the light- domain. To further confirm thenatureofthe ABC-to- induced ABC-to-ABA phase transition was highly repro- ABA structural transition, we carried out optical SHG ducible in many other TLG samples, regardless of the measurements (see “Materials and methods”), which Intensity (a.u.) Z (nm) Zhang et al. Light: Science & Applications (2020) 9:174 Page 4 of 11 ABC 1 mW 5 mW 10 mW 15 mW 20 mW ABA c d SHG ABA ABC Irradiated ABC 400 600 800 1000 1500 1600 2600 2800 –1 Wavelength (nm) Raman shift (cm ) Fig. 2 Light-induced ABC-to-ABA structural phase transition in TLG. a Artistic view of the laser-driven stacking order transformation in TLG. The ABA-stacked domain (left) and ABC-stacked domain (right) are separated by a domain wall (middle). b Raman mappings of the integrated G band intensity of TLG sample #2 after laser irradiation at various laser powers from 1 to 20 mW. The exposure time was 12 min for each laser scan. The white dashed lines indicate the gradual movement of the ABA/ABC domain wall under laser irradiation. The laser scan direction is from left to right and then from top to bottom. The scale bar is 4 μm. c Raman spectra and optical SHG responses (d) of the ABA-stacked domain, ABC-stacked domain and laser-irradiated ABC-stacked domain with a power of 20 mW. The dashed vertical line in (d) marks the SHG response of TLG at ~790 nm have been shown to be a reliable characterization ABC-stacked domains. Versatile manipulation of the method for crystal structures of two-dimensional domain walls is accomplished by our technique, including materials lacking inversion symmetry, thus being very reshaping and erasure of the domain walls, as well as sensitive to the stacking sequence. A previous study creation of closed-loop domain walls, as shown in Fig. 3. demonstrated a strong SHG response in ABA-stacked The laser-irradiated ABC-stacked domain was found to non-centrosymmetric TLG, while this response van- transform to an ABA-stacked domain, while the non- ished in ABC-stacked TLG, which preserves the inver- irradiated region retains the initial ABC stacking phase sion symmetry . We observed a similar SHG response without change. By area scanning over the desired region, for the initial ABA- and ABC-stacked domains, as the shape of the domain wall is redefined by laser irra- showninFig. 2d. The SHG peak appears in the spec- diation (Fig. 3a–c). A similar execution area scan of the trum of the area where the ABC-stacked domain is laser is employed to erase the domain walls in the TLG located after laser irradiation, suggesting light-induced (Fig. 3d, f). We can also create closed-loop domain walls disruption of the inversion symmetry due to the ABC- with an ABC-stacked domain inside by cutting through an to-ABA-stacking order transformation in the TLG. existing domain (Fig. 3d, e). Based on this technique, one Laser irradiation further enables phase patterning in can create new domains with arbitrary shapes and can TLG by local control over the geometries of the ABA- and manipulate the position and orientation of the domain Intensity (a.u.) Intensity (a.u.) Zhang et al. Light: Science & Applications (2020) 9:174 Page 5 of 11 ab c ABC 1 mW 15 mW 25 mW ABC ABA ABA ABA de f ABC ABC ABA ABA ABA 15 mW 15 mW 1 mW Fig. 3 Versatile manipulation of domain walls in TLG. a–c Reshaping the ABA/ABC domain walls in TLG. Raman mappings of the integrated G band intensity of TLG sample #63 under laser irradiation with different powers from 1 to 25 mW. The exposure time is 11 min for each laser scan. d–f Creation and erasure of ABA/ABC domain walls in TLG. Raman mappings of the integrated G band intensity of sample #14. The exposure times are 6 min in e and 11 min in f. The white dashed rectangles represent the area scan of the laser, and the arrow indicates the line scan of the laser. The scan direction of the laser is from left to right and then from top to bottom. The scale bars are 4 μm walls. Such ability to control the geometry of domain After Raman mapping, the whole flake was scanned by a walls in a desired area with a submicron resolution laser beam. After laser illumination, we again plotted the (determined by the diameter of the laser spot) will lead to Raman map of the integrated G peak intensity, as shown fine control over the structural phases and topological in Fig. 4d. There are two significant changes in the irra- states in graphene and other two-dimensional quantum diated MLG. First, the domain wall moved, and the area of materials. the ABC domain shrank. Second, a new region of mixed In addition to TLG, thicker multilayer graphene (MLG) ABA and ABC stacking formed after laser irradiation, also exhibits ABA and ABC stacking configurations. We marked by purple dots in Fig. 4d. To further understand exfoliated MLG flakes onto an oxidized silicon substrate the origin of these three regions, we analysed the Raman and combined optical contrast measurements, AFM, and spectrum of each region in more detail, as shown in Fig. Raman spectroscopy to determine the number of layers. 4h. It is evident that a part of the ABC domain has been The optical microscopy image of MLG sample #125 is completely transformed into ABA stacking (red dot) and shown in Fig. 4a. Despite the uniform thickness (~2.5 nm, that another part of ABC-stacked graphene transformed 6 ± 1 graphene layers) and featureless morphology into mixed ABA and ABC stacking (purple dot). (Fig. 4b), the Raman map of the integrated G peak Although we performed the Raman measurements with intensity (laser power: 2 mW) exhibits two regions with great care, the resolution is <1 μm but still larger than strikingly different contrast, as shown in Fig. 4c. According 500 nm due to the limit of the laser spot size (~0.6 μm). to previous reports, these distinct regions are thought to To achieve better resolution of the domain walls in gra- also arise from the different stacking sequences in the phene, we employed scattering-type scanning near-field MLG. We further probe the details of the Raman spectrum optical microscopy (s-SNOM) to directly image the of each region in the MLG, as shown in Fig. 4g. The 2D stacking structure and domain walls in the graphene peaks clearly show the line shape characteristics of ABA samples (see “Materials and methods”). In trilayer or (black) and ABC (red) stacking. In addition, the G peak is MLG, ABA- and ABC-stacked domains give different −1 ~4 cm lower than that in the ABA-stacking domain, infrared responses due to their different electronic band which is also a characteristic of ABC-stacked MLG. structures, resulting in different contrast in the s-SNOM Zhang et al. Light: Science & Applications (2020) 9:174 Page 6 of 11 ab c ABC Before irradiation 0 ABA 0 10 20 Distance (μm) II 10 μm 10 μm 10 μm ABC d e f III MAX After irradiation III III II ABA I MIN II II III 10 μm 10 μm 5 μm gh ABA ABA ABC-to-ABA ABC ABC After Mixed Before irradiation irradiation 1500 1600 2700 2900 1500 1600 2700 2900 –1 –1 Raman shift (cm ) Raman shift (cm ) Fig. 4 Raman mapping and s-SNOM imaging of the light-induced structural phase transition in MLG. a Optical microscopy image of MLG −1 −1 sample #125. b AFM image and height profile of graphene. c Raman maps of the integrated G peak intensity (position: 1576 cm , width: 5 cm ) before laser irradiation and (d) after laser irradiation. The laser power is 20 mW, and the exposure time is 34 min. e s-SNOM image of graphene after laser irradiation. f Magnified s-SNOM image of graphene. Graphene domains with different stacking orders show different contrasts in the s-SNOM image. The marked regions I, II, and III correspond to ABC stacking, ABA stacking and mixed ABC+ ABA stacking domains, respectively. The red arrows in (e, f) highlight the additional mixed ABC + ABA stacking domains that were not resolved in the Raman maps. g Raman spectra of different graphene regions taken from the marked solid dots before laser irradiation and (h) after laser irradiation image, as shown in Fig. 4e. Domain walls are observed in heating and strain, we plotted ω vs. ω . In contrast to G 2D the transitional regions between different stacking the reported upshift of ω and ω due to strain relaxa- G 2D domains. The s-SNOM image of irradiated MLG shows tion, both ω and ω downshift under laser irradiation in G 2D 37,38 features of the domain walls that are highly consistent our experiments . This result implies that laser-induced with the Raman maps but exhibits a higher resolution of local heating is essential for the phase transition in TLG, approximately tens of nanometers. Additional mixed and thus, the stacking order switch is thermal. To further ABC + ABA stacking domains and domain walls are exclude the effect of local strain, we performed laser clearly resolved in the detailed s-SNOM image (Fig. 4f), irradiation experiments in TLG on Al O . The thermal 2 3 −6 −1 which was not observed in the Raman measurements. The expansion coefficient of Al O is ~5 × 10 K , an order 2 3 s-SNOM imaging of domain walls after additional laser of magnitude higher than that of SiO , which may lead to irradiation is shown in Fig. S6. different local strains in laser-irradiated graphene. How- To understand the origin of the laser-induced ABC-to- ever, our results show that the light-induced ABC-to-ABA ABA phase transition in graphene, we summarized the structural phase transition also occurs in TLG on the positions of the G peaks (ω ) and 2D peaks (ω ) of TLG, Al O substrate, as shown in Fig. S8. In addition, we G 2D 2 3 as shown in Fig. S7. Both ω and ω undergo downshifts observed consistent light-induced stacking order transi- G 2D under laser irradiation with power ranging from 1 to tions in graphene with different exposure times 50 mW. To analyse the effect of laser-induced local (Fig. S9) and different laser wavelengths (Fig. S10). Intensity (a.u.) Heigth (nm) Intensity (a.u.) Zhang et al. Light: Science & Applications (2020) 9:174 Page 7 of 11 a c ABC ABA b d 900 °C 1100 °C Before Before After After 900 °C 1100 °C 2600 2800 2600 2800 –1 –1 Raman shift (cm ) Raman shift (cm ) Fig. 5 Thermal annealing-induced ABC-to-ABA structural phase transition in TLG. a, b Raman mappings of the integrated G band intensity of sample #98 before (a) and after (b) annealing at 900 °C for 8 h. The scale bars are 6 μm. c, d Raman mappings of the integrated G band intensity of sample #96 before (c) and after (d) annealing at 1100 °C for 8 h. The scale bars are 9 μm. The white dashed zones highlight the ABC-stacked domains in (a–c), which disappear in (d) after annealing at 1100 °C. e, f Raman spectra showing 2D bands of ABC-stacked domains before annealing and after annealing at 900 °C and 1100 °C, respectively We now consider the laser-induced local heating effect d. The comparison between the Raman spectra and 2D to be the main driving force of the structural phase bands before and after thermal annealing further confirms transition in TLG. To verify the role of thermal activation, the ABC-to-ABA phase transition at 1100 °C, as shown in we performed annealing of TLG in an argon atmosphere Fig. 5e, f. (see “Materials and methods”). In fact, Laves and Baskin During laser illumination, light absorption significantly were able to produce ABC graphite initiated by uni- increases the lattice temperature of TLG, which was directional pressure associated with shear force half a measured in situ by Raman G band shifts. Figure 6a plots century ago . They also observed transformation from a color map of the G band position as a function of laser ABC to ABA stacking in graphite by heating the samples power, showing a continuous redshift with increasing at 1300 °C for 4 h. However, the stacking order transfor- laser power. We estimated the steady-state temperature T mation has not yet been observed in TLG by thermal of TLG under laser irradiation using an established 10 −1 −1 −1 −1 heating up to 800 °C . Here, we demonstrated the coefficient γ (0.011 cm K < γ < 0.016 cm K ) annealing-induced ABC-to-ABA structural phase transi- between the G band shift (Δv) and lattice temperature of tion in TLG at 1100 °C for 8 h. As shown in Fig. 5a, b, the TLG: T = 300 + Δω /γ, where Δω is the laser heating- G G 40–43 Raman mappings do not show any significant change after induced downshift of the Raman G peak . As shown annealing at 900 °C for 8 h, whereas the ABC-stacked in Fig. 6b, T monotonically increases as a function of laser domain completely transforms to the ABA-stacked power, reaching ~600 K under 20 mW laser irradiation. domain after annealing at 1100 °C, as shown in Fig. 5c, According to the Raman mappings in Fig. 2b, the phase Intensity (a.u.) Intensity (a.u.) Zhang et al. Light: Science & Applications (2020) 9:174 Page 8 of 11 a c Configuration 0.00 0.25 0.50 0.75 1.00 SCAN+rVV10 LDA 2000 40 1500 1550 1600 1650 –1 10 Raman shift (cm ) ABC ABX ABA ABA ABC Structrual phase transition 0 1020304050 Laser power (mW) Fig. 6 Determining the energy barrier of the structural phase transition in TLG. a 2D color plot of Raman shifts of G bands as a function of applied laser power. b The change in Raman shifts of the G bands (left axis) and the determined lattice temperature (right axis) of TLG. The data are −1 −1 −1 −1 from sample #2. The error bars are determined by the uncertainty of the coefficient γ (0.011 cm K ~ 0.016 cm K ). The red solid curve is the lattice temperature of TLG calculated by the heat diffusion equation. The horizontal black dashed line marks the threshold temperature T ~ 430 K for ts the structural transition in TLG under 10 mW laser irradiation. c DFT calculations of the minimum energy transition path of TLG from ABC stacking to ABA stacking using the NEB method. The red squares represent the calculation performed by means of the SCAN + rVV10 functional, while the blue dots represent the calculation performed by means of the LDA functional. The dashed curves are visual guides obtained by Gaussian fitting. The schematics at the bottom illustrate three different stacking configurations of TLG and the process of ABC-to-ABA structural phase transition 47,48 transition initiates at 10 mW, and the corresponding laser irradiation . In addition, the Raman spectrum of temperature is approximately 430 K (see Fig. S11). We irradiated graphene still shows the characteristics of TLG defined this temperature as the threshold temperature T without any signature of bilayer or monolayer graphene. ts of the ABC-to-ABA phase transition in TLG. Interest- We attribute the absence of laser-induced defects and ingly, the deduced value of T ~ 430 K in this work is thinning effects in graphene to the relatively low power ts consistent with a previous report that the stacking tran- density of the laser and the presence of the SiO /Si sub- sition in rhombohedral graphite (7.5 nm thickness) starts strate. The substrate plays a crucial role as a heat sink for to occur at ~500 K by Joule heating . We notice that the graphene, providing additional channels for heat T value obtained from Raman spectroscopy is lower than dissipation. ts that determined from the annealing experiments. We To gain a further understanding of the physical attribute this discrepancy to the defects induced by high- mechanism of the ABC-to-ABA structural phase transi- temperature annealing, which might pin the ABC stacking tion in TLG, we performed climbing image nudged elastic phase and raise the energy barrier for the phase band (CI-NEB) calculations based on DFT (see “Materials 45,46 49 transition . and methods”) . Figure 6c shows the calculated mini- We can simulate the lattice temperature of TLG under mum energy pathway of TLG from ABC stacking to ABA- laser irradiation using the three-dimensional finite ele- stacking configurations. There are two energy minima, ment method to solve the heat diffusion equation (see corresponding to the two stable phases of TLG: ABA- and “Materials and methods”). The calculated lattice tem- ABC-stacked TLG. It is also evident that ABA-stacked perature is in good agreement with that determined from TLG exhibits a lower energy (~2 meV per unit cell) than Raman spectroscopy, as shown in Fig. 6b. Notably, we did ABC-stacked TLG. This is consistent with previous not observe any Raman signature of defects in TLG after reports that ABA stacking is more thermodynamically 13,22 laser irradiation below 20 mW. As shown in Fig. S12, the stable than ABC stacking . There is a significant energy D band only appears at the edges of TLG under 50 mW barrier during the ABC-to-ABA phase transition. It –1 Laser power (mW) G band shift (cm ) Intensity (a.u.) T (K) Energy (meV) Zhang et al. Light: Science & Applications (2020) 9:174 Page 9 of 11 should be noted that in the process of CI-NEB calcula- initial ABC stacking structure and an intermediate state. tions, we did not fix any position of the atom, and the Our results reveal the physical mechanism of the light- search for the barrier is promised by the CI-NEB induced structural phase transition in TLGs, which sheds method . It is interesting that in the final transition light on the realization of reversible stacking transitions, as path, the configuration at the highest energy corresponds well as polymorphism engineering of two-dimensional to a particular stacking structure: the topmost layer lattice material devices with new functionalities, including optical storage media, optically configurable metasurfaces, and is parallelly shifted by half a carbon–carbon bond length compared with the top layer of ABC-stacked TLG, as photonic devices. shown in the schematic of Fig. 6c bottom. We named this unstable structure ABX stacking. Materials and methods The simulation demonstrates that there is a relative Sample preparation conservation of energy when shifting only the topmost Graphene flakes were mechanically exfoliated from layer in the transition from ABC to ABA stacking. How- graphite bulk crystals (flaggy graphite, purchased from ever, the change in the structure during the whole tran- NGS company, Germany) onto a substrate with a 290 nm sition path is complicated, as the interlayer distance and SiO capping layer on top of heavily doped silicon. We the structures of the total three layers can be different. verified the TLG structure by optical microscopy, Raman The final calculated energy difference ΔE between ABX spectroscopy and AFM. After characterization, we irra- and ABC stacking is approximately 40 meV (~508 K), as diated the TLG with a continuous laser beam using the determined by the SCAN + rVV10 functional, which scan mapping functional of a Raman spectrometer (Witec considers the nonlocal interactions, including van der Alpha 300R). The wavelength of the laser was 532 nm, Waals forces, between different layers in the TLG. Con- and the laser power was adjusted from 0 to 50 mW. For sidering the thermal activation energy as E = k T (k the the thermal annealing experiments, TLG samples were a B B Boltzmann constant), the above values are quite con- placed in a quartz tube of a furnace and heated from room sistent with the estimated threshold temperature T ~ temperature to 1100 °C at a heating rate of 4 °C/min. The ts 430 K for the light-induced phase transition in TLG samples were held at 1100 °C for 8 h and then cooled to determined in this work and with the T ~ 500 K for the room temperature at a rate of 5 °C/min. During annealing, ts Joule heating-induced stacking transition in rhombohe- the samples were protected by a pure argon atmosphere dral graphite reported in a previous study . The total under a low pressure of 30 Pa. energy of TLGs with different stacking was also calcu- lated, as shown in Table S1 in the Supplementary Infor- Characterization mation. Based on the above analyses, we can attribute the Raman spectroscopy of TLG was performed by means of physical mechanism of the light-induced phase transition a commercial confocal Raman spectrometer (Witec Alpha in TLG to the thermally activated parallel slipping and 300R). The laser wavelength was 532 nm, and the diameter rearrangement of carbon atoms of the topmost layer of of the laser spot was ~0.6 μm. The spectral resolution of the −1 TLG caused by the laser local heating effect. Raman spectra was 1 cm using a grating of 600 grooves −1 per mm. We calibrated the Raman spectra by the 520 cm Conclusions Raman peak of the silicon substrate. To avoid any heating In summary, this work explores whether laser irradiation effect, the laser power was fixedat1mW.Wemeasuredthe of TLG results in a structural phase transition from ABC nonlinear optical response of the TLG sample using a stacking to ABA stacking. The stacking order transforma- home-built setup. A femtosecond laser was coupled to the tion was confirmed by significant changes in the Raman sample by a single-mode fiber with a spot diameter of spectra and nonlinear optical SHG response of TLG. The ~4 μm. Theoutputpower of thefemtosecond laserwas light-induced phase transition was found to be a gradual 10 mW, the center wavelength was 1.57 μm, and the pulse changing process and to always initiate at the ABA/ABC width of the laser was 130 fs. The light emission from TLG domain walls. Versatile manipulation of the domain walls was collected by an objective lens and then focused to a was accomplished by our technique, including reshaping single-mode fiber and finally coupled to an optical spectrum and erasure of the domain walls, as well as creation of analyser (Yokogawa AQ6370). To achieve better resolution closed-loop domain walls. We were also able to observe the of the domain walls in graphene, we employed scattering- ABC-to-ABA structural phase transition by thermal type scanning near-field microscopy (s-SNOM, neaspec annealing, highlighting the laser heating effect as the major GmbH) to directly image the stacking structure and domain driving force of the stacking order transition in TLG. The walls in the graphene samples. s-SNOM enables imaging in DFT simulations considering the van der Waals interaction the infrared regions at a spatial resolution of ~10 nm. The s- suggested an energy barrier of ~40 meV for the structural SNOM measurement was based on tapping-mode AFM. phasetransitiondue to theenergydifferencebetween the An infrared incident light beam (λ= 10.6 μm) was focused Zhang et al. Light: Science & Applications (2020) 9:174 Page 10 of 11 onto the apex of a conductive AFM tip (Arrow NCPt, Monkhorst–Pack grid case calculation was conducted, nanoWorld). To collect the scattered light that carries local yielding the same results as 6 × 6 × 1. The optimized optical information of graphene samples, we used a cooled atomic positions with the maximum force on any atom HgCdTe detector placed in the far field. During the mea- <0.001 a.u. was implemented with an initial interlayer surements, we recorded the near-field images simulta- separation at the experimental value of 3.36 Å along the z- neously with the topography information. axis for TLG. To determine the energy barrier between ABA and ABC structures, we also performed a CI-NEB Temperature simulation calculation. A 2 × 2 supercell (including 24 carbon atoms) To investigate the temperature of TLG under laser was calculated with both LDA and SCAN + rVV10 irradiation, we use the three-dimensional finite element exchange-correlation models in the CI-NEB process. It method to solve the heat diffusion equation −∇·(κ∇T) = should be mentioned that the LDA functional will over- q, where κ is the thermal conductivity and q ¼ Iα  estimate the interactions regarding the van der Waals 2 2 exp ðÞ 2r = r is the heat inflow per unit area owing to dispersion, but the long-range correlation will be lost, laser excitation, where I is the laser intensity and α is the which is included in the SCAN functional . absorptance of TLG (6.9%). r is the radius of the laser spot (0.3 µm). In our simulation, the thermal con- Acknowledgements This research was supported by the National Key R&D Program of China (no. ductivities of SiO and Si were 1.4 and 50 W/(m K), 2018YFA0306900). M.Z. acknowledges the financial support from the National respectively . The thermal conductivity of graphene was Key R&D Program of China (no. 2018YFA0306900) and the National Natural determined by the temperature of acoustic phonon κ(T ) Science Foundation of China (no. 11804386). J.D. acknowledges the financial ap support from the National Key R&D Program of China (no. 2017YFA0403200), = κ(T )(T /T ) , where κ(T ) = 450 W/(m K), γ = 1, with 0 0 ap 0 the National Natural Science Foundation of China (no. 11774429), and the T = 300 K . The interface between graphene and the NSAF (no. U1830206). J.C. acknowledges the financial support from the underlying SiO was modeled with a thermal resistance of 2 National Key Research and Development Program of China (grant no. −8 2 51 2016YFA0203500), the National Natural Science Foundation of China (grant no. 2×10 m K/W . A convective heat flux boundary 11874407), and the Strategic Priority Research Program of Chinese Academy of condition was used in our model to describe the heat Science (grant no. XDB 30000000). We thank Dr. Fang Luo, Dr. Gongjin Qi, Long transfer between air and the upper boundary. The heat Fang, and Dr. Lu Wang for their kind help with sample preparation and Raman and SNOM measurements. transfer coefficient was set as 5 W/m K. A fixed tem- perature (300 K) boundary condition was used at the Author details boundary of the substrate. The temperature measured by Department of Physics, National University of Defense Technology, 410073 Raman spectroscopy was a weighted average of the tem- Changsha, China. College of Advanced Interdisciplinary Studies, National perature inside the laser spot. In our simulation, we 3 University of Defense Technology, 410073 Changsha, China. Hunan Key defined the average temperature as : Laboratory of Super Micro-structure and Ultrafast Process, School of Physics and Electronics, Central South University, 410083 Changsha, China. Quantum Design China (Beijing) Co., Ltd, 100015 Beijing, China. Beijing National TrðÞqrðÞrdr Laboratory for Condensed Matter Physics, Institute of Physics, Chinese T  : 6 Academy of Sciences, 100190 Beijing, China. School of Physical Sciences, qrðÞrdr University of Chinese Academy of Sciences, 100049 Beijing, China. Songshan Lake Materials Laboratory, Dongguan 523808 Guangdong, China. Chongqing 2D Materials Institute, Liangjiang New Area, 400714 Chongqing, China. DFT calculations Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, 110016 Shenyang, China. School of To determine the energy barrier between ABA and ABC Material Science and Engineering, University of Science and Technology of stacking structures, we performed a CI-NEB calculation China, 230026 Anhui, China. State Key Laboratory of Quantum Optics and based on DFT with two different types of exchange- Quantum Optics Devices, Institute of Opto-Electronics, Shanxi University, 030006 Taiyuan, China. Department of Materials Science and Engineering, correlation functionals: local density approximation National University of Singapore, Singapore 117575, Singapore (LDA) and meta-GGA (SCAN) + rVV10 in the Quantum- 54–57 ESPRESSO package . The SCAN + rVV10 functional Author contributions considers the nonlocal interactions, including van der J.Z., G.P., and M.Z. prepared all the samples and carried out all the Waal forces between difference layers and many-body measurements; J.H., W.X., and J.D. carried out the DFT simulations; K.L. and Z.Z. measured the SHG response; X.Yang, X.Yuan, and Z.H. helped with the sample effects in electrons, resulting in more accurate energies preparation and data analysis. Y.L. and J.C. performed the s-SNOM imaging and structures compared with the traditional GGA measurements. W.C. carried out AFM measurements. S.Q. and K.N. supervised functional. A norm-conserved pseudopotential was this project; M.Z. wrote the paper with the help of all authors. All authors have approved the final version of the paper. implemented in the calculation, and the kinetic energy cutoff was set as 50 Ry, while the density cutoff was set as Conflict of interest 200 Ry . The Brillouin zone of a 2 × 2 supercell with 20 Å The authors declare that they have no conflict of interest. vacuum separation was sampled using a 6 × 6 × 1 Monkhorst–Pack grid with Methfessel–Paxton smearing Supplementary information is available for this paper at https://doi.org/ 59–61 of 0.01 Ry and the 2D cutoff . A 12 × 12 × 1 10.1038/s41377-020-00412-6. Zhang et al. Light: Science & Applications (2020) 9:174 Page 11 of 11 Received: 25 March 2020 Revised: 27 September 2020 Accepted: 28 31. Yue,Y.F. etal. Light-inducedmechanical response in crosslinked liquid- September 2020 crystalline polymers with photoswitchable glass transition temperatures. Nat. Commun. 9, 3234 (2018). 32. Cho, S. et al. Phase patterning for ohmic homojunction contact in MoTe . Science 349,625–628 (2015). 33. Tan, Y. et al. Controllable 2H-to-1T′ phase transition in few-layer MoTe . References Nanoscale 10,19964–19971 (2018). 1. Ponomarenko, L. A. et al. Cloning of Dirac fermions in graphene superlattices. 34. Zhang, M. Y. et al. Light-induced subpicosecond lattice symmetry switch in Nature 497,594–597 (2013). MoTe . Phys. Rev. X 9, 021036 (2019). 2. Hunt, B. et al. Massive Dirac fermions and Hofstadter butterflyinavander 35. Lin, J. H. et al. AC/AB stacking boundaries in bilayer graphene. Nano Lett. 13, Waals heterostructure. Science 340, 1427–1430 (2013). 3262–3268 (2013). 3. Dean,C.R.etal. Hofstadter’sbutterfly and the fractal quantum Hall effect in 36. Shan, Y. W. et al. Stacking symmetry governed second harmonic generation in moire superlattices. Nature 497, 598–602 (2013). graphene trilayers. Sci. Adv. 4, eaat0074 (2018). 4. Cao, Y. et al. Unconventional superconductivity in magic-angle graphene 37. Papasimakis, N. et al. Strain engineering in graphene by laser irradiation. Appl. superlattices. Nature 556,43–50 (2018). Phys. Lett. 106, 061904 (2015). 5. Cao, Y. et al. Correlated insulator behaviour at half-filling in magic-angle 38. Alexeev, E., Moger, J. & Hendry, E. Photo-induced doping and strain in exfo- graphene superlattices. Nature 556,80–84 (2018). liated graphene. Appl. Phys. Lett. 103, 151907 (2013). 6. Chen,G.R.etal. Signatures of tunablesuperconductivity in a trilayer graphene 39. Laves, F. & Baskin, Y. On the formation of the rhombohedral graphite mod- moiré superlattice. Nature 572,215–219 (2019). ification. Z. für. Kristallographie Cryst. Mater. 107,337–356 (1956). 7. Chen, G. R. et al. Evidence of a gate-tunable Mott insulator in a trilayer 40. Calizo, I. et al. Temperature dependence of the Raman spectra of graphene graphene moiré superlattice. Nat. Phys. 15, 237–241 (2019). and graphene multilayers. Nano Lett. 7, 2645–2649 (2007). 8. Chen,G.R.etal. Tunablecorrelatedchern insulator and ferromagnetism in a 41. Calizo, I. et al. Variable temperature Raman microscopy as a nanometrology moiré superlattice. Nature 579,56–61 (2020). tool for graphene layers and graphene-based devices. Appl. Phys. Lett. 91, 9. Warner,J.H., Mukai, M. &Kirkland, A. I. Atomic structure of ABC rhombohedral 071913 (2007). stacked trilayer graphene. ACS Nano 6,5680–5686 (2012). 42. Son, S. K. et al. Graphene hot-electron light bulb: incandescence from hBN- 10. Lui,C.H.etal. Imagingstacking order in few-layer graphene. Nano Lett. 11, encapsulated graphene in air. 2D Mater. 5, 011006 (2018). 164–169 (2011). 43. Luo, F. et al. Graphene thermal emitter with enhanced joule heating and 11. Lui,C.H.etal. Observationofanelectrically tunable band gap in trilayer localized light emission in air. ACS Photonics 6,2117–2125 (2019). graphene. Nat. Phys. 7,944–947 (2011). 44. Latychevskaia, T. et al. Stacking transition in rhombohedral graphite. Front. 12. Bao, W. et al. Stacking-dependent band gap and quantum transport in trilayer Phys. 14, 13608 (2019). graphene. Nat. Phys. 7,948–952 (2011). 45. Lin, Y. C. et al. Graphene annealing: how clean can it be? Nano Lett. 12, 13. Aoki, M. & Amawashi, H. Dependence of band structures on stacking and field 414–419 (2012). in layered graphene. Solid State Commun. 142,123–127 (2007). 46. Hong, J. et al. Origin of new broad Raman D and G peaks in annealed 14. Craciun, M. F. et al. Trilayer graphene is a semimetal with a gate-tunable band graphene. Sci. Rep. 3, 2700 (2013). overlap. Nat. Nanotechnol. 4,383–388 (2009). 47. Cançado, L. G. et al. Quantifying defects in graphene via Raman spectroscopy 15. Zou, K. et al. Transport studies of dual-gated ABC and ABA trilayer graphene: at different excitation energies. Nano Lett. 11, 3190–3196 (2011). band gap opening and band structure tuning in very large perpendicular 48. Eckmann, A. et al. Probing the nature of defects in graphene by Raman electric fields. Nano Lett. 13,369–373 (2013). spectroscopy. Nano Lett. 12, 3925–3930 (2012). 16. Jiang, L. L. et al. Manipulation of domain-wall solitons in bi- and trilayer 49. Hohenberg, P. & Kohn, W. Inhomogeneous electron gas. Phys. Rev. 136, graphene. Nat. Nanotechnol. 13,204–208 (2018). B864–B871 (1964). 17. Yankowitz, M. et al. Electric field control of soliton motion and stacking in 50. Henkelman, G., Uberuaga,B.P.&Jónsson, H. A climbing image nudged elastic trilayer graphene. Nat. Mater. 13,786–789 (2014). band method for finding saddle points and minimum energy paths. J. Chem. 18. Jiang, L. L. et al. Soliton-dependent plasmon reflection at bilayer graphene Phys. 113,9901–9904 (2000). domain walls. Nat. Mater. 15, 840–844 (2016). 51. Freitag, M. et al. Energy dissipation in graphene field-effect transistors. Nano 19. Ju, L. et al. Topological valley transport at bilayer graphene domain walls. Lett. 9,1883–1888 (2009). Nature 520, 650–655 (2015). 52. Kim, Y. D. et al. Bright visible light emission from graphene. Nat. Nanotechnol. 20. Vaezi, A. et al. Topological edge states at a tilt boundary in gated multilayer 10,676–681 (2015). graphene. Phys. Rev. X 3, 021018 (2013). 53. Lee, J. U. et al. Thermal conductivity of suspended pristine graphene mea- 21. Semenoff, G. W., Semenoff, V. & Zhou, F. Domain walls in gapped graphene. sured by Raman spectroscopy. Phys. Rev. B 83, 081419 (2011). Phys. Rev. Lett. 101, 087204 (2008). 54. Sabatini, R., Gorni, T. & De Gironcoli, S. Nonlocal van der Waals density func- 22. Zhang, W. J. et al. Molecular adsorption induces the transformation of tional made simple and efficient. Phys.Rev.B 87, 041108 (2013). rhombohedral- to Bernal-stacking order in trilayer graphene. Nat. Commun. 4, 55. Vydrov, O. A. & Van Voorhis, T. Nonlocal van der Waals density functional: the 2074 (2013). simpler the better. J. Chem. Phys. 133, 244103 (2010). 23. Li, H. Y. et al. Global control of stacking-order phase transition by doping and 56. Giannozzi, P. et al. QUANTUM ESPRESSO: a modular and open-source software electric field in few-layer graphene. Nano Lett. 20, 3106–3112 (2020). project for quantum simulations of materials. J. Phys. Condens. Matter 21, 24. Hao, Y. F. et al. Probing layer number and stacking order of few-layer gra- 395502 (2009). phene by Raman spectroscopy. Small 6,195–200 (2010). 57. Sun,J.W., Ruzsinszky, A.& Perdew, J.P.Strongly constrained andappropriately 25. Cong, C. X. et al. Raman characterization of ABA- and ABC-stacked trilayer normed semilocal density functional. Phys.Rev.Lett. 115, 036402 (2015). graphene. ACS Nano 5,8760–8768 (2011). 58. Hamann, D. R. Optimized norm-conserving Vanderbilt pseudopotentials. Phys. 26. Ferrari,A.C.& Basko, D. M. Ramanspectroscopy as a versatile tool for studying Rev. B 88, 085117 (2013). the properties of graphene. Nat. Nanotechnol. 8,235–246 (2013). 59. Sohier, T., Calandra, M. & Mauri, F. Density functional perturbation theory for 27. Yang, Y. P. et al. Stacking order in graphite films controlled by van der Waals gated two-dimensional heterostructures: theoretical developments technology. Nano Lett. 19, 8526–8532 (2019). and application to flexural phonons in graphene. Phys. Rev. B 96, 075448 28. Wilhelm, H. A., Croset, B. & Medjahdi, G. Proportion and dispersion of rhom- (2017). bohedral sequences in the hexagonal structure of graphite powders. Carbon 60. Methfessel, M. & Paxton, A. T. High-precision sampling for Brillouin-zone 45,2356–2364 (2007). integration in metals. Phys. Rev. B 40,3616–3621 (1989). 29. Qu, Y. R. et al. Thermal camouflage based on the phase-changing material 61. Monkhorst,H.J.&Pack,J.D.Special points forBrillouin-zone integrations. Phys. GST. Light Sci. Appl. 7, 26 (2018). Rev. B 13,5188–5192 (1976). 30. Muskens, O. L. et al. Antenna-assisted picosecond control of nanoscale phase 62. Dai,J. Y., Yuan,J.M.&Giannozzi, P. Gasadsorptionongraphenedoped with B, transition in vanadium dioxide. Light Sci. Appl. 5, e16173 (2016). N, Al, and S: a theoretical study. Appl. Phys. Lett. 95, 232105 (2009). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Light: Science & Applications Springer Journals

Loading next page...
 
/lp/springer-journals/light-induced-irreversible-structural-phase-transition-in-trilayer-2sRm9MK0wZ

References (63)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
eISSN
2047-7538
DOI
10.1038/s41377-020-00412-6
Publisher site
See Article on Publisher Site

Abstract

A crystal structure has a profound influence on the physical properties of the corresponding material. By synthesizing crystals with particular symmetries, one can strongly tune their properties, even for the same chemical configuration (compare graphite and diamond, for instance). Even more interesting opportunities arise when the structural phases of crystals can be changed dynamically through external stimulations. Such abilities, though rare, lead to a number of exciting phenomena, such as phase-change memory effects. In the case of trilayer graphene, there are two common stacking configurations (ABA and ABC) that have distinct electronic band structures and exhibit very different behaviors. Domain walls exist in the trilayer graphene with both stacking orders, showing fascinating new physics such as the quantum valley Hall effect. Extensive efforts have been dedicated to the phase engineering of trilayer graphene. However, the manipulation of domain walls to achieve precise control of local structures and properties remains a considerable challenge. Here, we experimentally demonstrate that we can switch from one structural phase to another by laser irradiation, creating domains of different shapes in trilayer graphene. The ability to control the position and orientation of the domain walls leads to fine control of the local structural phases and properties of graphene, offering a simple but effective approach to create artificial two-dimensional materials with designed atomic structures and electronic and optical properties. Introduction the bottom layer (denoted as rhombohedral or ABC 9,10 The stacking configuration of layered materials plays stacking) . Due to different interlayer electron an important role in determining their electronic and interactions and distinct crystal symmetry, it has been optical properties. Fascinating phenomena, such as shown that ABA-stacked TLG and ABC-stacked TLG Hofstadter’s butterfly, Mott insulators, ferromagnetism, exhibit significantly different physical properties. From and unconventional superconductivity, can also emerge the perspective of the electronic band structure, ABA- in van der Waals heterostructures by carefully con- stacked TLG is a semimetal with a gate-tunable band 1–8 trolling the layer stacking sequence .Inthe case of overlap between the valence and conduction bands, trilayer graphene (TLG), there are two common stack- whereas the ABC-stacked TLG is a semiconductor with 11–15 ing configurations: the top layer may lie directly above an electrically tunable band gap .InTLG flakes the bottom layer (denoted as Bernal or ABA stacking) containing both ABA and ABC stacking, there are or may instead lie above the center of the hexagon of domain walls between the phases, consisting of a localized strain soliton in which the carbon atoms of one graphene layer shift by the carbon–carbon bond Correspondence: Jiayu Dai (jydai@nudt.edu.cn)or 16,17 distance . Such domain walls in TLGs have attracted Mengjian Zhu (zhumengjian11@nudt.edu.cn) much interest because of their intriguing physical Department of Physics, National University of Defense Technology, 410073 Changsha, China properties. For example, optically, soliton-dependent College of Advanced Interdisciplinary Studies, National University of Defense reflection of graphene plasmons at the domain walls Technology, 410073 Changsha, China has been experimentally observed . Electrically, the Full list of author information is available at the end of the article These authors contributed equally: Jianyu Zhang, Jinsen Han, Gang Peng © The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Zhang et al. Light: Science & Applications (2020) 9:174 Page 2 of 11 domain walls are predicted to host topological edge between the topmost layers in the two allotropes . Gra- states and ballistic transport and can also produce in- phene trilayer was obtained by mechanical exfoliation and plane metal–semiconductor (ABA–ABC) homojunc- confirmed by optical contrast, and the thickness 19–21 tions in TLG . (~1.2 nm) was determined by AFM, as shown in Fig. 1b, c. Previous reports have shown that applying molecular Raman spectroscopy has been demonstrated to be an absorption or an external electric field can drive the accurate and effective method to distinguish the ABA and 10,24–26 stacking order transition and generate domain wall ABC stacking structures in TLG . TLG flakes show 17,22,23 motion in graphene layers . There is inevitable very uniform optical contrast without any visible domain residue on graphene with molecular doping, which hin- walls or wrinkles. However, the Raman mapping exhibits ders the properties of graphene. Applying an electrical two distinct domains with significant contrast due to the field or strain usually leads to global control of the different stacking orders in TLG, as shown in Fig. 1b 10,25 stacking order phase and hinders precise manipulation of (here, integrated G band intensity is presented) . The the local structure. An alternative way to change the darker domain was identified as ABA-stacked TLG, and stacking configuration is by applying a local mechanical the brighter domain was identified as ABC stacking. force. For example, a previous study demonstrated that Raman spectra of the two different domains are plotted domain walls in TLGs can be moved by mechanical stress for comparison in Fig. 1d. The spectra are different from exerted through an atomic force microscopy (AFM) tip . one another in at least three ways: first, the 2D band of The domain walls are invisible in conventional AFM ABC-stacked TLG shows more asymmetric features with topography, and studies must rely on near-field infrared an enhanced peak and shoulder compared with the nanoscopy measurements. However, a simple and con- symmetric feature shown in ABA-stacked TLG; second, trollable approach to engineer the stacking phase and the G band of the ABC-stacked domain is redshifted by −1 domain walls into designed atomic structures is still ~1 cm compared with that of the ABA-stacked domain; lacking. and third, the G’ band of the ABC-stacked TLG domain Here, we experimentally demonstrate that the stacking also exhibits more asymmetric features than its ABA- 10,25,27 order in TLG can be switched from ABC to ABA by local stacked counterpart . heating enabled through laser irradiation. The light- We will focus on the integrated G band intensity induced stacking phase transition in TLG is directly mappings (Fig. 1b), which show the lowest noise level visualized using Raman mapping and near-field nano- (mappings of bandwidth of G bands and 2D yield con- scopy imaging. By controlling the movement of the laser sistent shape of domains, as shown in Fig. S1). In this beam with considerable flexibility and precision, we are work, we prepared 211 TLG flakes. Among them, 147 able to reshape the domains and manipulate the position flakes are pure ABA-stacked TLG, and the remaining 64 and orientation of the domain walls in the TLG. We flakes have coexisting ABA- and ABC-stacked domains, as attribute the laser-induced local heating effect as the main shown in Fig. S2. The proportion of ABC stacking in TLG 10,28 driving force of the ABC-to-ABA phase transition. The is ~15%, consistent with previous reports . activation energy is determined by Raman spectroscopy An attractive target for optical materials is to find a measurements and thermal annealing experiments and is system that shows the structural phase transition trig- 29–31 consistent with the calculated energy barrier height of gered by external stimulation of light . Laser irradia- approximately 40 meV determined by density functional tion has been demonstrated as an effective method to theory (DFT) calculations. The electronic and optical induce structural phase transitions in two-dimensional properties of TLG strongly depend on the stacking con- materials, for instance, laser-driven 2H-to-1T’ phase 32–34 figuration. Therefore, the ability to achieve fine control of transitions in few-layer MoTe . Here, we extend this the local stacking configuration and manipulate the methodology to control the stacking order transformation domain walls by a simple and clean approach opens the in TLG. A continuous laser beam was scanned over the way to new devices with fascinating functionalities, such TLG sample under ambient conditions, as schematically as multilevel optical switch and phase-change memory. shown in Fig. 2a (see the “Materials and methods” section for more details). The sample was illuminated by a laser Results and discussion with different powers from 1 to 20 mW. First, the laser The schematics in Fig. 1a shows the crystalline struc- beam moved from left to right to finish one line scan. tures of TLG with ABA and ABC stacking orders. The After that, the laser returned to the left and moved atoms of the topmost layer in ABA-stacked TLG lie downward to start the next line scan until all scans are exactly above of those of the bottom layer, whereas in completed. After finishing each laser scan, the sample was ABC-stacked TLG, the sublattice of the top layer lies again characterized by Raman mapping with a laser power above the center of the hexagons in the bottom layer. of 1 mW. Figure 2b summarizes the Raman mappings of There is a parallel shift of exactly one carbon honeycomb the integrated G band intensity, showing the lapsed Zhang et al. Light: Science & Applications (2020) 9:174 Page 3 of 11 A B C A B bc 010 20 0 X (μm) GG′ 2D ABA ABC 1560 1580 1600 2400 2450 2500 2600 2700 2800 1400 1600 1800 2000 2200 2400 2600 2800 –1 Raman shift (cm ) Fig. 1 Characterizations of mechanically exfoliated TLG. a Schematics of graphene trilayers with ABA stacking (left) and ABC stacking (right) configurations. The bottom, middle, and top layers are labeled with different colors. b Optical microscopy image of TLG sample #2 and the corresponding Raman mapping of the integrated G band intensity. The darker domain indicates the ABA-stacked TLG, and the brighter domain was defined as ABC stacking. The scale bars are 4 μm, and the color bar shows the integrated Raman intensity. c AFM height profile of TLG measured along the green dashed line in the optical image in (b). d Raman spectra of TLG taken from different regions marked in the Raman mapping image in −1 ’ −1 −1 (b). The insets show magnified spectra of TLG: G band (1560–1600 cm ), G band (2400–2500 cm ), and 2D band (2600–2800 cm ) dynamic process of the phase transition from ABC geometric shapes of the domain walls and the angle stacking to ABA stacking. The domain wall started to between the laser scanning direction and the domain move from the ABA-stacked domain to the ABC-stacked walls, as shown in Fig. S5. We noticed that the movement domain under 10 mW laser irradiation. As the laser power of the ABA/ABC domain walls in TLG was similar to the increased, the domain wall gradually shifted from left to bilayer graphene case. In bilayer graphene, the AB/AC right, showing a reduced ABC stacking area and an stacking boundaries were observed as nanometer-wide expanded ABA stacking region. The Raman mappings of strained channels, mostly in the form of ripples, produ- the integrated 2D band intensity show the same trans- cing smooth low-energy transitions between the two dif- formation process, as shown in Fig. S3. Furthermore, we ferent stacks . found that if the laser scan zone contained the domain Figure 2cshows thesignificant changes in the Raman walls, then the ABC-to-ABA stacking order transition spectra of TLG before and after laser irradiation. It was always initiated from the domain wall rather than ran- evident that the 2D band of the ABC domain became domly occurring in the TLG. However, the ABC-to-ABA more symmetric after laser irradiation, which agrees phase transition can also occur in the pure ABC-stacked with the 2D band features of the initial ABA-stacking TLG region, as shown in Fig. S4. Notably, the light- domain. To further confirm thenatureofthe ABC-to- induced ABC-to-ABA phase transition was highly repro- ABA structural transition, we carried out optical SHG ducible in many other TLG samples, regardless of the measurements (see “Materials and methods”), which Intensity (a.u.) Z (nm) Zhang et al. Light: Science & Applications (2020) 9:174 Page 4 of 11 ABC 1 mW 5 mW 10 mW 15 mW 20 mW ABA c d SHG ABA ABC Irradiated ABC 400 600 800 1000 1500 1600 2600 2800 –1 Wavelength (nm) Raman shift (cm ) Fig. 2 Light-induced ABC-to-ABA structural phase transition in TLG. a Artistic view of the laser-driven stacking order transformation in TLG. The ABA-stacked domain (left) and ABC-stacked domain (right) are separated by a domain wall (middle). b Raman mappings of the integrated G band intensity of TLG sample #2 after laser irradiation at various laser powers from 1 to 20 mW. The exposure time was 12 min for each laser scan. The white dashed lines indicate the gradual movement of the ABA/ABC domain wall under laser irradiation. The laser scan direction is from left to right and then from top to bottom. The scale bar is 4 μm. c Raman spectra and optical SHG responses (d) of the ABA-stacked domain, ABC-stacked domain and laser-irradiated ABC-stacked domain with a power of 20 mW. The dashed vertical line in (d) marks the SHG response of TLG at ~790 nm have been shown to be a reliable characterization ABC-stacked domains. Versatile manipulation of the method for crystal structures of two-dimensional domain walls is accomplished by our technique, including materials lacking inversion symmetry, thus being very reshaping and erasure of the domain walls, as well as sensitive to the stacking sequence. A previous study creation of closed-loop domain walls, as shown in Fig. 3. demonstrated a strong SHG response in ABA-stacked The laser-irradiated ABC-stacked domain was found to non-centrosymmetric TLG, while this response van- transform to an ABA-stacked domain, while the non- ished in ABC-stacked TLG, which preserves the inver- irradiated region retains the initial ABC stacking phase sion symmetry . We observed a similar SHG response without change. By area scanning over the desired region, for the initial ABA- and ABC-stacked domains, as the shape of the domain wall is redefined by laser irra- showninFig. 2d. The SHG peak appears in the spec- diation (Fig. 3a–c). A similar execution area scan of the trum of the area where the ABC-stacked domain is laser is employed to erase the domain walls in the TLG located after laser irradiation, suggesting light-induced (Fig. 3d, f). We can also create closed-loop domain walls disruption of the inversion symmetry due to the ABC- with an ABC-stacked domain inside by cutting through an to-ABA-stacking order transformation in the TLG. existing domain (Fig. 3d, e). Based on this technique, one Laser irradiation further enables phase patterning in can create new domains with arbitrary shapes and can TLG by local control over the geometries of the ABA- and manipulate the position and orientation of the domain Intensity (a.u.) Intensity (a.u.) Zhang et al. Light: Science & Applications (2020) 9:174 Page 5 of 11 ab c ABC 1 mW 15 mW 25 mW ABC ABA ABA ABA de f ABC ABC ABA ABA ABA 15 mW 15 mW 1 mW Fig. 3 Versatile manipulation of domain walls in TLG. a–c Reshaping the ABA/ABC domain walls in TLG. Raman mappings of the integrated G band intensity of TLG sample #63 under laser irradiation with different powers from 1 to 25 mW. The exposure time is 11 min for each laser scan. d–f Creation and erasure of ABA/ABC domain walls in TLG. Raman mappings of the integrated G band intensity of sample #14. The exposure times are 6 min in e and 11 min in f. The white dashed rectangles represent the area scan of the laser, and the arrow indicates the line scan of the laser. The scan direction of the laser is from left to right and then from top to bottom. The scale bars are 4 μm walls. Such ability to control the geometry of domain After Raman mapping, the whole flake was scanned by a walls in a desired area with a submicron resolution laser beam. After laser illumination, we again plotted the (determined by the diameter of the laser spot) will lead to Raman map of the integrated G peak intensity, as shown fine control over the structural phases and topological in Fig. 4d. There are two significant changes in the irra- states in graphene and other two-dimensional quantum diated MLG. First, the domain wall moved, and the area of materials. the ABC domain shrank. Second, a new region of mixed In addition to TLG, thicker multilayer graphene (MLG) ABA and ABC stacking formed after laser irradiation, also exhibits ABA and ABC stacking configurations. We marked by purple dots in Fig. 4d. To further understand exfoliated MLG flakes onto an oxidized silicon substrate the origin of these three regions, we analysed the Raman and combined optical contrast measurements, AFM, and spectrum of each region in more detail, as shown in Fig. Raman spectroscopy to determine the number of layers. 4h. It is evident that a part of the ABC domain has been The optical microscopy image of MLG sample #125 is completely transformed into ABA stacking (red dot) and shown in Fig. 4a. Despite the uniform thickness (~2.5 nm, that another part of ABC-stacked graphene transformed 6 ± 1 graphene layers) and featureless morphology into mixed ABA and ABC stacking (purple dot). (Fig. 4b), the Raman map of the integrated G peak Although we performed the Raman measurements with intensity (laser power: 2 mW) exhibits two regions with great care, the resolution is <1 μm but still larger than strikingly different contrast, as shown in Fig. 4c. According 500 nm due to the limit of the laser spot size (~0.6 μm). to previous reports, these distinct regions are thought to To achieve better resolution of the domain walls in gra- also arise from the different stacking sequences in the phene, we employed scattering-type scanning near-field MLG. We further probe the details of the Raman spectrum optical microscopy (s-SNOM) to directly image the of each region in the MLG, as shown in Fig. 4g. The 2D stacking structure and domain walls in the graphene peaks clearly show the line shape characteristics of ABA samples (see “Materials and methods”). In trilayer or (black) and ABC (red) stacking. In addition, the G peak is MLG, ABA- and ABC-stacked domains give different −1 ~4 cm lower than that in the ABA-stacking domain, infrared responses due to their different electronic band which is also a characteristic of ABC-stacked MLG. structures, resulting in different contrast in the s-SNOM Zhang et al. Light: Science & Applications (2020) 9:174 Page 6 of 11 ab c ABC Before irradiation 0 ABA 0 10 20 Distance (μm) II 10 μm 10 μm 10 μm ABC d e f III MAX After irradiation III III II ABA I MIN II II III 10 μm 10 μm 5 μm gh ABA ABA ABC-to-ABA ABC ABC After Mixed Before irradiation irradiation 1500 1600 2700 2900 1500 1600 2700 2900 –1 –1 Raman shift (cm ) Raman shift (cm ) Fig. 4 Raman mapping and s-SNOM imaging of the light-induced structural phase transition in MLG. a Optical microscopy image of MLG −1 −1 sample #125. b AFM image and height profile of graphene. c Raman maps of the integrated G peak intensity (position: 1576 cm , width: 5 cm ) before laser irradiation and (d) after laser irradiation. The laser power is 20 mW, and the exposure time is 34 min. e s-SNOM image of graphene after laser irradiation. f Magnified s-SNOM image of graphene. Graphene domains with different stacking orders show different contrasts in the s-SNOM image. The marked regions I, II, and III correspond to ABC stacking, ABA stacking and mixed ABC+ ABA stacking domains, respectively. The red arrows in (e, f) highlight the additional mixed ABC + ABA stacking domains that were not resolved in the Raman maps. g Raman spectra of different graphene regions taken from the marked solid dots before laser irradiation and (h) after laser irradiation image, as shown in Fig. 4e. Domain walls are observed in heating and strain, we plotted ω vs. ω . In contrast to G 2D the transitional regions between different stacking the reported upshift of ω and ω due to strain relaxa- G 2D domains. The s-SNOM image of irradiated MLG shows tion, both ω and ω downshift under laser irradiation in G 2D 37,38 features of the domain walls that are highly consistent our experiments . This result implies that laser-induced with the Raman maps but exhibits a higher resolution of local heating is essential for the phase transition in TLG, approximately tens of nanometers. Additional mixed and thus, the stacking order switch is thermal. To further ABC + ABA stacking domains and domain walls are exclude the effect of local strain, we performed laser clearly resolved in the detailed s-SNOM image (Fig. 4f), irradiation experiments in TLG on Al O . The thermal 2 3 −6 −1 which was not observed in the Raman measurements. The expansion coefficient of Al O is ~5 × 10 K , an order 2 3 s-SNOM imaging of domain walls after additional laser of magnitude higher than that of SiO , which may lead to irradiation is shown in Fig. S6. different local strains in laser-irradiated graphene. How- To understand the origin of the laser-induced ABC-to- ever, our results show that the light-induced ABC-to-ABA ABA phase transition in graphene, we summarized the structural phase transition also occurs in TLG on the positions of the G peaks (ω ) and 2D peaks (ω ) of TLG, Al O substrate, as shown in Fig. S8. In addition, we G 2D 2 3 as shown in Fig. S7. Both ω and ω undergo downshifts observed consistent light-induced stacking order transi- G 2D under laser irradiation with power ranging from 1 to tions in graphene with different exposure times 50 mW. To analyse the effect of laser-induced local (Fig. S9) and different laser wavelengths (Fig. S10). Intensity (a.u.) Heigth (nm) Intensity (a.u.) Zhang et al. Light: Science & Applications (2020) 9:174 Page 7 of 11 a c ABC ABA b d 900 °C 1100 °C Before Before After After 900 °C 1100 °C 2600 2800 2600 2800 –1 –1 Raman shift (cm ) Raman shift (cm ) Fig. 5 Thermal annealing-induced ABC-to-ABA structural phase transition in TLG. a, b Raman mappings of the integrated G band intensity of sample #98 before (a) and after (b) annealing at 900 °C for 8 h. The scale bars are 6 μm. c, d Raman mappings of the integrated G band intensity of sample #96 before (c) and after (d) annealing at 1100 °C for 8 h. The scale bars are 9 μm. The white dashed zones highlight the ABC-stacked domains in (a–c), which disappear in (d) after annealing at 1100 °C. e, f Raman spectra showing 2D bands of ABC-stacked domains before annealing and after annealing at 900 °C and 1100 °C, respectively We now consider the laser-induced local heating effect d. The comparison between the Raman spectra and 2D to be the main driving force of the structural phase bands before and after thermal annealing further confirms transition in TLG. To verify the role of thermal activation, the ABC-to-ABA phase transition at 1100 °C, as shown in we performed annealing of TLG in an argon atmosphere Fig. 5e, f. (see “Materials and methods”). In fact, Laves and Baskin During laser illumination, light absorption significantly were able to produce ABC graphite initiated by uni- increases the lattice temperature of TLG, which was directional pressure associated with shear force half a measured in situ by Raman G band shifts. Figure 6a plots century ago . They also observed transformation from a color map of the G band position as a function of laser ABC to ABA stacking in graphite by heating the samples power, showing a continuous redshift with increasing at 1300 °C for 4 h. However, the stacking order transfor- laser power. We estimated the steady-state temperature T mation has not yet been observed in TLG by thermal of TLG under laser irradiation using an established 10 −1 −1 −1 −1 heating up to 800 °C . Here, we demonstrated the coefficient γ (0.011 cm K < γ < 0.016 cm K ) annealing-induced ABC-to-ABA structural phase transi- between the G band shift (Δv) and lattice temperature of tion in TLG at 1100 °C for 8 h. As shown in Fig. 5a, b, the TLG: T = 300 + Δω /γ, where Δω is the laser heating- G G 40–43 Raman mappings do not show any significant change after induced downshift of the Raman G peak . As shown annealing at 900 °C for 8 h, whereas the ABC-stacked in Fig. 6b, T monotonically increases as a function of laser domain completely transforms to the ABA-stacked power, reaching ~600 K under 20 mW laser irradiation. domain after annealing at 1100 °C, as shown in Fig. 5c, According to the Raman mappings in Fig. 2b, the phase Intensity (a.u.) Intensity (a.u.) Zhang et al. Light: Science & Applications (2020) 9:174 Page 8 of 11 a c Configuration 0.00 0.25 0.50 0.75 1.00 SCAN+rVV10 LDA 2000 40 1500 1550 1600 1650 –1 10 Raman shift (cm ) ABC ABX ABA ABA ABC Structrual phase transition 0 1020304050 Laser power (mW) Fig. 6 Determining the energy barrier of the structural phase transition in TLG. a 2D color plot of Raman shifts of G bands as a function of applied laser power. b The change in Raman shifts of the G bands (left axis) and the determined lattice temperature (right axis) of TLG. The data are −1 −1 −1 −1 from sample #2. The error bars are determined by the uncertainty of the coefficient γ (0.011 cm K ~ 0.016 cm K ). The red solid curve is the lattice temperature of TLG calculated by the heat diffusion equation. The horizontal black dashed line marks the threshold temperature T ~ 430 K for ts the structural transition in TLG under 10 mW laser irradiation. c DFT calculations of the minimum energy transition path of TLG from ABC stacking to ABA stacking using the NEB method. The red squares represent the calculation performed by means of the SCAN + rVV10 functional, while the blue dots represent the calculation performed by means of the LDA functional. The dashed curves are visual guides obtained by Gaussian fitting. The schematics at the bottom illustrate three different stacking configurations of TLG and the process of ABC-to-ABA structural phase transition 47,48 transition initiates at 10 mW, and the corresponding laser irradiation . In addition, the Raman spectrum of temperature is approximately 430 K (see Fig. S11). We irradiated graphene still shows the characteristics of TLG defined this temperature as the threshold temperature T without any signature of bilayer or monolayer graphene. ts of the ABC-to-ABA phase transition in TLG. Interest- We attribute the absence of laser-induced defects and ingly, the deduced value of T ~ 430 K in this work is thinning effects in graphene to the relatively low power ts consistent with a previous report that the stacking tran- density of the laser and the presence of the SiO /Si sub- sition in rhombohedral graphite (7.5 nm thickness) starts strate. The substrate plays a crucial role as a heat sink for to occur at ~500 K by Joule heating . We notice that the graphene, providing additional channels for heat T value obtained from Raman spectroscopy is lower than dissipation. ts that determined from the annealing experiments. We To gain a further understanding of the physical attribute this discrepancy to the defects induced by high- mechanism of the ABC-to-ABA structural phase transi- temperature annealing, which might pin the ABC stacking tion in TLG, we performed climbing image nudged elastic phase and raise the energy barrier for the phase band (CI-NEB) calculations based on DFT (see “Materials 45,46 49 transition . and methods”) . Figure 6c shows the calculated mini- We can simulate the lattice temperature of TLG under mum energy pathway of TLG from ABC stacking to ABA- laser irradiation using the three-dimensional finite ele- stacking configurations. There are two energy minima, ment method to solve the heat diffusion equation (see corresponding to the two stable phases of TLG: ABA- and “Materials and methods”). The calculated lattice tem- ABC-stacked TLG. It is also evident that ABA-stacked perature is in good agreement with that determined from TLG exhibits a lower energy (~2 meV per unit cell) than Raman spectroscopy, as shown in Fig. 6b. Notably, we did ABC-stacked TLG. This is consistent with previous not observe any Raman signature of defects in TLG after reports that ABA stacking is more thermodynamically 13,22 laser irradiation below 20 mW. As shown in Fig. S12, the stable than ABC stacking . There is a significant energy D band only appears at the edges of TLG under 50 mW barrier during the ABC-to-ABA phase transition. It –1 Laser power (mW) G band shift (cm ) Intensity (a.u.) T (K) Energy (meV) Zhang et al. Light: Science & Applications (2020) 9:174 Page 9 of 11 should be noted that in the process of CI-NEB calcula- initial ABC stacking structure and an intermediate state. tions, we did not fix any position of the atom, and the Our results reveal the physical mechanism of the light- search for the barrier is promised by the CI-NEB induced structural phase transition in TLGs, which sheds method . It is interesting that in the final transition light on the realization of reversible stacking transitions, as path, the configuration at the highest energy corresponds well as polymorphism engineering of two-dimensional to a particular stacking structure: the topmost layer lattice material devices with new functionalities, including optical storage media, optically configurable metasurfaces, and is parallelly shifted by half a carbon–carbon bond length compared with the top layer of ABC-stacked TLG, as photonic devices. shown in the schematic of Fig. 6c bottom. We named this unstable structure ABX stacking. Materials and methods The simulation demonstrates that there is a relative Sample preparation conservation of energy when shifting only the topmost Graphene flakes were mechanically exfoliated from layer in the transition from ABC to ABA stacking. How- graphite bulk crystals (flaggy graphite, purchased from ever, the change in the structure during the whole tran- NGS company, Germany) onto a substrate with a 290 nm sition path is complicated, as the interlayer distance and SiO capping layer on top of heavily doped silicon. We the structures of the total three layers can be different. verified the TLG structure by optical microscopy, Raman The final calculated energy difference ΔE between ABX spectroscopy and AFM. After characterization, we irra- and ABC stacking is approximately 40 meV (~508 K), as diated the TLG with a continuous laser beam using the determined by the SCAN + rVV10 functional, which scan mapping functional of a Raman spectrometer (Witec considers the nonlocal interactions, including van der Alpha 300R). The wavelength of the laser was 532 nm, Waals forces, between different layers in the TLG. Con- and the laser power was adjusted from 0 to 50 mW. For sidering the thermal activation energy as E = k T (k the the thermal annealing experiments, TLG samples were a B B Boltzmann constant), the above values are quite con- placed in a quartz tube of a furnace and heated from room sistent with the estimated threshold temperature T ~ temperature to 1100 °C at a heating rate of 4 °C/min. The ts 430 K for the light-induced phase transition in TLG samples were held at 1100 °C for 8 h and then cooled to determined in this work and with the T ~ 500 K for the room temperature at a rate of 5 °C/min. During annealing, ts Joule heating-induced stacking transition in rhombohe- the samples were protected by a pure argon atmosphere dral graphite reported in a previous study . The total under a low pressure of 30 Pa. energy of TLGs with different stacking was also calcu- lated, as shown in Table S1 in the Supplementary Infor- Characterization mation. Based on the above analyses, we can attribute the Raman spectroscopy of TLG was performed by means of physical mechanism of the light-induced phase transition a commercial confocal Raman spectrometer (Witec Alpha in TLG to the thermally activated parallel slipping and 300R). The laser wavelength was 532 nm, and the diameter rearrangement of carbon atoms of the topmost layer of of the laser spot was ~0.6 μm. The spectral resolution of the −1 TLG caused by the laser local heating effect. Raman spectra was 1 cm using a grating of 600 grooves −1 per mm. We calibrated the Raman spectra by the 520 cm Conclusions Raman peak of the silicon substrate. To avoid any heating In summary, this work explores whether laser irradiation effect, the laser power was fixedat1mW.Wemeasuredthe of TLG results in a structural phase transition from ABC nonlinear optical response of the TLG sample using a stacking to ABA stacking. The stacking order transforma- home-built setup. A femtosecond laser was coupled to the tion was confirmed by significant changes in the Raman sample by a single-mode fiber with a spot diameter of spectra and nonlinear optical SHG response of TLG. The ~4 μm. Theoutputpower of thefemtosecond laserwas light-induced phase transition was found to be a gradual 10 mW, the center wavelength was 1.57 μm, and the pulse changing process and to always initiate at the ABA/ABC width of the laser was 130 fs. The light emission from TLG domain walls. Versatile manipulation of the domain walls was collected by an objective lens and then focused to a was accomplished by our technique, including reshaping single-mode fiber and finally coupled to an optical spectrum and erasure of the domain walls, as well as creation of analyser (Yokogawa AQ6370). To achieve better resolution closed-loop domain walls. We were also able to observe the of the domain walls in graphene, we employed scattering- ABC-to-ABA structural phase transition by thermal type scanning near-field microscopy (s-SNOM, neaspec annealing, highlighting the laser heating effect as the major GmbH) to directly image the stacking structure and domain driving force of the stacking order transition in TLG. The walls in the graphene samples. s-SNOM enables imaging in DFT simulations considering the van der Waals interaction the infrared regions at a spatial resolution of ~10 nm. The s- suggested an energy barrier of ~40 meV for the structural SNOM measurement was based on tapping-mode AFM. phasetransitiondue to theenergydifferencebetween the An infrared incident light beam (λ= 10.6 μm) was focused Zhang et al. Light: Science & Applications (2020) 9:174 Page 10 of 11 onto the apex of a conductive AFM tip (Arrow NCPt, Monkhorst–Pack grid case calculation was conducted, nanoWorld). To collect the scattered light that carries local yielding the same results as 6 × 6 × 1. The optimized optical information of graphene samples, we used a cooled atomic positions with the maximum force on any atom HgCdTe detector placed in the far field. During the mea- <0.001 a.u. was implemented with an initial interlayer surements, we recorded the near-field images simulta- separation at the experimental value of 3.36 Å along the z- neously with the topography information. axis for TLG. To determine the energy barrier between ABA and ABC structures, we also performed a CI-NEB Temperature simulation calculation. A 2 × 2 supercell (including 24 carbon atoms) To investigate the temperature of TLG under laser was calculated with both LDA and SCAN + rVV10 irradiation, we use the three-dimensional finite element exchange-correlation models in the CI-NEB process. It method to solve the heat diffusion equation −∇·(κ∇T) = should be mentioned that the LDA functional will over- q, where κ is the thermal conductivity and q ¼ Iα  estimate the interactions regarding the van der Waals 2 2 exp ðÞ 2r = r is the heat inflow per unit area owing to dispersion, but the long-range correlation will be lost, laser excitation, where I is the laser intensity and α is the which is included in the SCAN functional . absorptance of TLG (6.9%). r is the radius of the laser spot (0.3 µm). In our simulation, the thermal con- Acknowledgements This research was supported by the National Key R&D Program of China (no. ductivities of SiO and Si were 1.4 and 50 W/(m K), 2018YFA0306900). M.Z. acknowledges the financial support from the National respectively . The thermal conductivity of graphene was Key R&D Program of China (no. 2018YFA0306900) and the National Natural determined by the temperature of acoustic phonon κ(T ) Science Foundation of China (no. 11804386). J.D. acknowledges the financial ap support from the National Key R&D Program of China (no. 2017YFA0403200), = κ(T )(T /T ) , where κ(T ) = 450 W/(m K), γ = 1, with 0 0 ap 0 the National Natural Science Foundation of China (no. 11774429), and the T = 300 K . The interface between graphene and the NSAF (no. U1830206). J.C. acknowledges the financial support from the underlying SiO was modeled with a thermal resistance of 2 National Key Research and Development Program of China (grant no. −8 2 51 2016YFA0203500), the National Natural Science Foundation of China (grant no. 2×10 m K/W . A convective heat flux boundary 11874407), and the Strategic Priority Research Program of Chinese Academy of condition was used in our model to describe the heat Science (grant no. XDB 30000000). We thank Dr. Fang Luo, Dr. Gongjin Qi, Long transfer between air and the upper boundary. The heat Fang, and Dr. Lu Wang for their kind help with sample preparation and Raman and SNOM measurements. transfer coefficient was set as 5 W/m K. A fixed tem- perature (300 K) boundary condition was used at the Author details boundary of the substrate. The temperature measured by Department of Physics, National University of Defense Technology, 410073 Raman spectroscopy was a weighted average of the tem- Changsha, China. College of Advanced Interdisciplinary Studies, National perature inside the laser spot. In our simulation, we 3 University of Defense Technology, 410073 Changsha, China. Hunan Key defined the average temperature as : Laboratory of Super Micro-structure and Ultrafast Process, School of Physics and Electronics, Central South University, 410083 Changsha, China. Quantum Design China (Beijing) Co., Ltd, 100015 Beijing, China. Beijing National TrðÞqrðÞrdr Laboratory for Condensed Matter Physics, Institute of Physics, Chinese T  : 6 Academy of Sciences, 100190 Beijing, China. School of Physical Sciences, qrðÞrdr University of Chinese Academy of Sciences, 100049 Beijing, China. Songshan Lake Materials Laboratory, Dongguan 523808 Guangdong, China. Chongqing 2D Materials Institute, Liangjiang New Area, 400714 Chongqing, China. DFT calculations Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, 110016 Shenyang, China. School of To determine the energy barrier between ABA and ABC Material Science and Engineering, University of Science and Technology of stacking structures, we performed a CI-NEB calculation China, 230026 Anhui, China. State Key Laboratory of Quantum Optics and based on DFT with two different types of exchange- Quantum Optics Devices, Institute of Opto-Electronics, Shanxi University, 030006 Taiyuan, China. Department of Materials Science and Engineering, correlation functionals: local density approximation National University of Singapore, Singapore 117575, Singapore (LDA) and meta-GGA (SCAN) + rVV10 in the Quantum- 54–57 ESPRESSO package . The SCAN + rVV10 functional Author contributions considers the nonlocal interactions, including van der J.Z., G.P., and M.Z. prepared all the samples and carried out all the Waal forces between difference layers and many-body measurements; J.H., W.X., and J.D. carried out the DFT simulations; K.L. and Z.Z. measured the SHG response; X.Yang, X.Yuan, and Z.H. helped with the sample effects in electrons, resulting in more accurate energies preparation and data analysis. Y.L. and J.C. performed the s-SNOM imaging and structures compared with the traditional GGA measurements. W.C. carried out AFM measurements. S.Q. and K.N. supervised functional. A norm-conserved pseudopotential was this project; M.Z. wrote the paper with the help of all authors. All authors have approved the final version of the paper. implemented in the calculation, and the kinetic energy cutoff was set as 50 Ry, while the density cutoff was set as Conflict of interest 200 Ry . The Brillouin zone of a 2 × 2 supercell with 20 Å The authors declare that they have no conflict of interest. vacuum separation was sampled using a 6 × 6 × 1 Monkhorst–Pack grid with Methfessel–Paxton smearing Supplementary information is available for this paper at https://doi.org/ 59–61 of 0.01 Ry and the 2D cutoff . A 12 × 12 × 1 10.1038/s41377-020-00412-6. Zhang et al. Light: Science & Applications (2020) 9:174 Page 11 of 11 Received: 25 March 2020 Revised: 27 September 2020 Accepted: 28 31. Yue,Y.F. etal. Light-inducedmechanical response in crosslinked liquid- September 2020 crystalline polymers with photoswitchable glass transition temperatures. Nat. Commun. 9, 3234 (2018). 32. Cho, S. et al. Phase patterning for ohmic homojunction contact in MoTe . Science 349,625–628 (2015). 33. Tan, Y. et al. Controllable 2H-to-1T′ phase transition in few-layer MoTe . References Nanoscale 10,19964–19971 (2018). 1. Ponomarenko, L. A. et al. Cloning of Dirac fermions in graphene superlattices. 34. Zhang, M. Y. et al. Light-induced subpicosecond lattice symmetry switch in Nature 497,594–597 (2013). MoTe . Phys. Rev. X 9, 021036 (2019). 2. Hunt, B. et al. Massive Dirac fermions and Hofstadter butterflyinavander 35. Lin, J. H. et al. AC/AB stacking boundaries in bilayer graphene. Nano Lett. 13, Waals heterostructure. Science 340, 1427–1430 (2013). 3262–3268 (2013). 3. Dean,C.R.etal. Hofstadter’sbutterfly and the fractal quantum Hall effect in 36. Shan, Y. W. et al. Stacking symmetry governed second harmonic generation in moire superlattices. Nature 497, 598–602 (2013). graphene trilayers. Sci. Adv. 4, eaat0074 (2018). 4. Cao, Y. et al. Unconventional superconductivity in magic-angle graphene 37. Papasimakis, N. et al. Strain engineering in graphene by laser irradiation. Appl. superlattices. Nature 556,43–50 (2018). Phys. Lett. 106, 061904 (2015). 5. Cao, Y. et al. Correlated insulator behaviour at half-filling in magic-angle 38. Alexeev, E., Moger, J. & Hendry, E. Photo-induced doping and strain in exfo- graphene superlattices. Nature 556,80–84 (2018). liated graphene. Appl. Phys. Lett. 103, 151907 (2013). 6. Chen,G.R.etal. Signatures of tunablesuperconductivity in a trilayer graphene 39. Laves, F. & Baskin, Y. On the formation of the rhombohedral graphite mod- moiré superlattice. Nature 572,215–219 (2019). ification. Z. für. Kristallographie Cryst. Mater. 107,337–356 (1956). 7. Chen, G. R. et al. Evidence of a gate-tunable Mott insulator in a trilayer 40. Calizo, I. et al. Temperature dependence of the Raman spectra of graphene graphene moiré superlattice. Nat. Phys. 15, 237–241 (2019). and graphene multilayers. Nano Lett. 7, 2645–2649 (2007). 8. Chen,G.R.etal. Tunablecorrelatedchern insulator and ferromagnetism in a 41. Calizo, I. et al. Variable temperature Raman microscopy as a nanometrology moiré superlattice. Nature 579,56–61 (2020). tool for graphene layers and graphene-based devices. Appl. Phys. Lett. 91, 9. Warner,J.H., Mukai, M. &Kirkland, A. I. Atomic structure of ABC rhombohedral 071913 (2007). stacked trilayer graphene. ACS Nano 6,5680–5686 (2012). 42. Son, S. K. et al. Graphene hot-electron light bulb: incandescence from hBN- 10. Lui,C.H.etal. Imagingstacking order in few-layer graphene. Nano Lett. 11, encapsulated graphene in air. 2D Mater. 5, 011006 (2018). 164–169 (2011). 43. Luo, F. et al. Graphene thermal emitter with enhanced joule heating and 11. Lui,C.H.etal. Observationofanelectrically tunable band gap in trilayer localized light emission in air. ACS Photonics 6,2117–2125 (2019). graphene. Nat. Phys. 7,944–947 (2011). 44. Latychevskaia, T. et al. Stacking transition in rhombohedral graphite. Front. 12. Bao, W. et al. Stacking-dependent band gap and quantum transport in trilayer Phys. 14, 13608 (2019). graphene. Nat. Phys. 7,948–952 (2011). 45. Lin, Y. C. et al. Graphene annealing: how clean can it be? Nano Lett. 12, 13. Aoki, M. & Amawashi, H. Dependence of band structures on stacking and field 414–419 (2012). in layered graphene. Solid State Commun. 142,123–127 (2007). 46. Hong, J. et al. Origin of new broad Raman D and G peaks in annealed 14. Craciun, M. F. et al. Trilayer graphene is a semimetal with a gate-tunable band graphene. Sci. Rep. 3, 2700 (2013). overlap. Nat. Nanotechnol. 4,383–388 (2009). 47. Cançado, L. G. et al. Quantifying defects in graphene via Raman spectroscopy 15. Zou, K. et al. Transport studies of dual-gated ABC and ABA trilayer graphene: at different excitation energies. Nano Lett. 11, 3190–3196 (2011). band gap opening and band structure tuning in very large perpendicular 48. Eckmann, A. et al. Probing the nature of defects in graphene by Raman electric fields. Nano Lett. 13,369–373 (2013). spectroscopy. Nano Lett. 12, 3925–3930 (2012). 16. Jiang, L. L. et al. Manipulation of domain-wall solitons in bi- and trilayer 49. Hohenberg, P. & Kohn, W. Inhomogeneous electron gas. Phys. Rev. 136, graphene. Nat. Nanotechnol. 13,204–208 (2018). B864–B871 (1964). 17. Yankowitz, M. et al. Electric field control of soliton motion and stacking in 50. Henkelman, G., Uberuaga,B.P.&Jónsson, H. A climbing image nudged elastic trilayer graphene. Nat. Mater. 13,786–789 (2014). band method for finding saddle points and minimum energy paths. J. Chem. 18. Jiang, L. L. et al. Soliton-dependent plasmon reflection at bilayer graphene Phys. 113,9901–9904 (2000). domain walls. Nat. Mater. 15, 840–844 (2016). 51. Freitag, M. et al. Energy dissipation in graphene field-effect transistors. Nano 19. Ju, L. et al. Topological valley transport at bilayer graphene domain walls. Lett. 9,1883–1888 (2009). Nature 520, 650–655 (2015). 52. Kim, Y. D. et al. Bright visible light emission from graphene. Nat. Nanotechnol. 20. Vaezi, A. et al. Topological edge states at a tilt boundary in gated multilayer 10,676–681 (2015). graphene. Phys. Rev. X 3, 021018 (2013). 53. Lee, J. U. et al. Thermal conductivity of suspended pristine graphene mea- 21. Semenoff, G. W., Semenoff, V. & Zhou, F. Domain walls in gapped graphene. sured by Raman spectroscopy. Phys. Rev. B 83, 081419 (2011). Phys. Rev. Lett. 101, 087204 (2008). 54. Sabatini, R., Gorni, T. & De Gironcoli, S. Nonlocal van der Waals density func- 22. Zhang, W. J. et al. Molecular adsorption induces the transformation of tional made simple and efficient. Phys.Rev.B 87, 041108 (2013). rhombohedral- to Bernal-stacking order in trilayer graphene. Nat. Commun. 4, 55. Vydrov, O. A. & Van Voorhis, T. Nonlocal van der Waals density functional: the 2074 (2013). simpler the better. J. Chem. Phys. 133, 244103 (2010). 23. Li, H. Y. et al. Global control of stacking-order phase transition by doping and 56. Giannozzi, P. et al. QUANTUM ESPRESSO: a modular and open-source software electric field in few-layer graphene. Nano Lett. 20, 3106–3112 (2020). project for quantum simulations of materials. J. Phys. Condens. Matter 21, 24. Hao, Y. F. et al. Probing layer number and stacking order of few-layer gra- 395502 (2009). phene by Raman spectroscopy. Small 6,195–200 (2010). 57. Sun,J.W., Ruzsinszky, A.& Perdew, J.P.Strongly constrained andappropriately 25. Cong, C. X. et al. Raman characterization of ABA- and ABC-stacked trilayer normed semilocal density functional. Phys.Rev.Lett. 115, 036402 (2015). graphene. ACS Nano 5,8760–8768 (2011). 58. Hamann, D. R. Optimized norm-conserving Vanderbilt pseudopotentials. Phys. 26. Ferrari,A.C.& Basko, D. M. Ramanspectroscopy as a versatile tool for studying Rev. B 88, 085117 (2013). the properties of graphene. Nat. Nanotechnol. 8,235–246 (2013). 59. Sohier, T., Calandra, M. & Mauri, F. Density functional perturbation theory for 27. Yang, Y. P. et al. Stacking order in graphite films controlled by van der Waals gated two-dimensional heterostructures: theoretical developments technology. Nano Lett. 19, 8526–8532 (2019). and application to flexural phonons in graphene. Phys. Rev. B 96, 075448 28. Wilhelm, H. A., Croset, B. & Medjahdi, G. Proportion and dispersion of rhom- (2017). bohedral sequences in the hexagonal structure of graphite powders. Carbon 60. Methfessel, M. & Paxton, A. T. High-precision sampling for Brillouin-zone 45,2356–2364 (2007). integration in metals. Phys. Rev. B 40,3616–3621 (1989). 29. Qu, Y. R. et al. Thermal camouflage based on the phase-changing material 61. Monkhorst,H.J.&Pack,J.D.Special points forBrillouin-zone integrations. Phys. GST. Light Sci. Appl. 7, 26 (2018). Rev. B 13,5188–5192 (1976). 30. Muskens, O. L. et al. Antenna-assisted picosecond control of nanoscale phase 62. Dai,J. Y., Yuan,J.M.&Giannozzi, P. Gasadsorptionongraphenedoped with B, transition in vanadium dioxide. Light Sci. Appl. 5, e16173 (2016). N, Al, and S: a theoretical study. Appl. Phys. Lett. 95, 232105 (2009).

Journal

Light: Science & ApplicationsSpringer Journals

Published: Oct 13, 2020

There are no references for this article.