# Fileset

[SS08f_manuscript_SSuzuki.pdf](https://mdr.nims.go.jp/filesets/42930bc5-f6d6-4149-ab54-f358a51b60a1/download)

## Creator

[Seiya Suzuki](https://orcid.org/0000-0003-2445-7462), Yoshihiro Nemoto, Natsumi Shiiki, Yoshiko Nakayama, [Masaki Takeguchi](https://orcid.org/0000-0002-0282-6020)

## Rights

This is the peer reviewed version of the following article: In Situ Transmission Electron Microscopy Observation of Melted Germanium Encapsulated in Multilayer Graphene, which has been published in final form at https://doi.org/10.1002/andp.202300122 . This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. This article may not be enhanced, enriched or otherwise transformed into a derivative work, without express permission from Wiley or by statutory rights under applicable legislation. Copyright notices must not be removed, obscured or modified. The article must be linked to Wiley’s version of record on Wiley Online Library and any embedding, framing or otherwise making available the article or pages thereof by third parties from platforms, services and websites other than Wiley Online Library must be prohibited.[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[In situ Transmission Electron Microscopy Observation of Melted Germanium Encapsulated in Multilayer Graphene](https://mdr.nims.go.jp/datasets/a494d997-552e-4264-a9e9-79f05296db8d)

## Fulltext

Microsoft Word - SS08f_manuscript_SSuzuki   1 In situ Transmission Electron Microscopy Observation of Melted 1 Germanium Encapsulated in Multilayer Graphene 2 Seiya Suzuki1-3,*, Yoshihiro Nemoto4, Natsumi Shiiki2, Yoshiko Nakayama4, and Masaki Takeguchi4 3 1Advanced Science Research Center (ASRC), Japan Atomic Energy Agency (JAEA), 2-4 Shirakata, Tokai, 4 Ibaraki 319-1195, Japan 5 2International Center for Young Scientists (ICYS), National Institute for Materials Science (NIMS), 1-1 6 Namiki, Tsukuba, Ibaraki 305-0044, Japan 7 3PRESTO, Japan Science and Technology Agency (JST), 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, 8 Japan. 9 4Electron Microscopy Analysis Station, National Institute for Materials Science (NIMS), 1-2-1 Sengen, 10 Tsukuba, Ibaraki 305-0047, Japan 11 *suzuki.seiya35@jaea.go.jp and seiya09417@gmail.com 12  13 ABSTRACT 14 Germanene is a two-dimensional (2D) germanium (Ge) analogous of graphene, and its unique topological 15 properties are expected to be a material for next-generation electronics. However, no germanene 16 electronic devices have yet been reported. One of the reasons for this is that germanene is easily oxidized 17 in air due to its lack of chemical stability. Therefore, growing germanene at solid interfaces where it is 18 not oxidized is one of the key ideas for realizing electronic devices based on germanene. In this study, 19 the behavior of Ge at the solid interface at high temperatures was observed by transmission electron 20 microscopy (TEM). To achieve such in situ heating TEM observation, we fabricated a 21 graphene/Ge/graphene encapsulated structure. In situ heating TEM experiments revealed that Ge like 22 droplets moved and coalesced with other Ge droplets, indicating that Ge remained as a liquid phase 23 between graphene layers at temperatures higher than the Ge melting point. It was also observed that 24 droplet Ge incorporates the surrounding amorphous Ge as Ge nuclei, thereby increasing its size (domain 25    2 growth). These results indicated that Ge crystals can be grown at the interface of van der Waals materials, 26 which will be important for future germanene growth at solid interfaces. 27  28 Introduction 29 Germanene is a two-dimensional (2D) germanium (Ge) analogous to graphene.[1] Theoretical studies have 30 predicted that germanene has linear band dispersion,[1-2] a size-tunable band gap by applying an electric 31 field,[3] and topological properties.[1, 4] Owing to those unique physical properties, electronic device 32 applications of germanene, such as field effect transistors (FETs),[5] has been highly demanded as well as 33 silicene,[2, 6] stanene,[7] and plumbene.[8] Nevertheless the growth of germanene has already been reported 34 on various surfaces, such as Ag(111),[9] Au(111),[10] Cu(111),[11] Al(111),[12] Pt(111),[13] graphite,[14] and 35 MoS2,[15] the electronic devices of germanene have not been reported yet. Even though by taking into 36 account other the related group-14 2D monoatomic layers, the reports of their electronic devices are 37 limited except the hydrogenated germanene (germanane)[16] and silicene FETs.[17] One of the reasons for 38 this is chemical instability of group-14 based 2D monoatomic layers including germanene. Once high-39 quality germanene is grown on a surface in an ultra-high vacuum chamber, but it is easily degraded by 40 oxidation in air. Thus, the fabrication of electronic devices of germanene and other group-14 2D 41 monoatomic layers is difficult, in contrast to chemically stable 2D materials, such as graphene and 42 transition metal dichalcogenides (TMD). 43 As one approach, we have come up with the idea of growing germanene directly at the solid interface, 44 where it is not oxidized in air.[9f] Based on this concept, we have grown germanene at the interfaces of 45 graphene/Ag(111) and hexagonal boron nitride (hBN)/Ag(111).[9a] This work proved that the high gas 46 barrier properties of graphene and hBN allow the germanene at such interfaces to be stable in air. However, 47    3 this growth method still remains challenges for the fabrication of germanene devices since the germanene 48 is not electrically isolated due to the contacting Ag surface. This is a common challenge for germanene 49 grown on metal surfaces. In order to realize germanene FETs from growing it on a metal surface, the 50 physical or electrical separation of germanene from the grown metal surface will be required. 51 On the other hand, if germanene can be formed directly at the interface of insulators, the issue of contact 52 between germanene and metal surfaces does not arise, and the development of germanene devices should 53 be achieved. To maximize the potential of germanene, hBN would be a very promising encapsulation 54 material among insulators. The advantages of hBN encapsulation are that it provides an atomically 55 smooth surface (or interface), free of dangling bonds, and screening substrate effects such as scattering 56 for preserving (modulating, or enhancing) the intrinsic properties of the encapsulated 2D monatomic layer. 57 These have already been well demonstrated in the carrier transport measurements of graphene[18] and 58 TMD.[19] 59 The direct growth of germanene between hBN layers, as well as the growth of other group 14 monoatomic 60 2D materials, has not been achieved to date. In order to form germanene between 2D materials, we need 61 to insert Ge sources and crystalize them. Crystal growth from molten state (Czochralski method)[20] and 62 phase transition from amorphous state[21] are typical ways to form bulk and crystalline thin films of Ge, 63 respectively, and we thought that employing these methods would be a reasonable way to challenge the 64 growth of germanene between 2D materials. In this study, we fabricated a Ge encapsulated structure with 65 graphene layers and observed the behavior of the encapsulated Ge at high temperatures using in situ 66 heating transmission electron microscopy (TEM). Graphene, which produces van der Waals (vdW) 67 interfaces similar to those of hBN, was used as an alternative material to hBN. From the TEM observation, 68 we found that Ge between graphene were able to exist at more than the melting point of Ge (MPGe), 69 indicating that the graphene layers successfully encapsulate the Ge layer. After several heating processes, 70    4 amorphous Ge and crystalline Ge coexisted between the graphene layers. The shape of crystalline Ge is 71 almost spherical (named droplet Ge), indicating three-dimensional nucleation. At higher temperatures 72 than MPGe, the movement of Ge droplets and their coalescence with other Ge droplets in the surrounding 73 area were observed. Atomic resolution TEM observations showed that some droplets were liquid at high 74 temperatures because they did not have clear lattice fringes. Since crystalline Ge particles were observed 75 at room temperature (RT), the droplets crystallized when they cooled down. Our observations suggest 76 that it is important to avoid the formation of droplet when growing planar Ge crystals (germanene) 77 between vdW materials, which can provide hints for the direct growth of germanene at solid interfaces in 78 the future. 79  80 Results and Discussion 81 1. Preparation and observation of graphene/Ge/graphene encapsulated structure 82 Figure 1 shows the sample preparation of encapsulated Ge between graphene layers for in situ heating 83 TEM observation. The sample preparation consists of chemical vapor deposition (CVD) growth of 84 graphene, transfer graphene, Ge deposition, and encapsulation of Ge by graphene. CVD growth was done 85 onto a commercial copper (Cu) foil by using a homemade CVD furnace. The grown graphene is 86 polycrystalline and has a high surface coverage. Figure 1(a) illustrates a transfer process of monolayer-87 graphene (1L-graphene) onto another CVD graphene/Cu. The grown graphene/Cu (1) was first spin-88 coated by poly(methyl methacrylate) (PMMA) as a mechanical support layer (2). After drying, the Cu 89 substrate was etched away with metal etchant (3) followed by multiple deionized (DI) water rinses (4). 90 The PMMA/graphene was then scooped up with another graphene/Cu (5), and finally the PMMA layer 91 was removed with acetone (6). This graphene transfer was repeated to fabricate 4L-graphene. Figure 1(b) 92 shows the encapsulation of Ge by graphene and the transfer of the encapsulated Ge onto a Nano-Chip 93    5 (Wildfire, DENSsolutions) designed for in situ heating TEM observation. A 4L-graphene/Cu (7) and a 94 Ge thin film deposited 4L-graphene/Cu (8) were prepared for the Ge encapsulation. The Ge thin film was 95 patterned using a metal shadow mask during the electron beam deposition of Ge. By patterning Ge, the 96 contact area between the top and bottom 4L-graphene is increased, which improves their adhesion. The 97 transfer of the 4L-graphene onto the Ge/4L-graphene/Cu was also performed in the same way (9,10) as 98 the transfer process in Fig. 1(a). Finally, the 4L-graphene/Ge/4L-graphene was transferred by scooping 99 up with a Nano-Chip (11). 100 We also tried to encapsulate Ge with two sheets of 1L-graphene, but the Ge disappeared after metal 101 etching. This is probably due to the metal etchant permeation into Ge across the graphene, indicating the 102 imperfections in graphene, such as defects, grain boundaries, wrinkles, insufficient adhesion to Ge, and 103 their occurrence during the transfer process. By increasing the number of graphene layers, the loss of Ge 104 was strongly suppressed, and 4Ls were found to be thick enough to fabricate a graphene/Ge/graphene 105 encapsulation structure. 106    6  107 Figure 1. Sample preparation of encapsulated Ge between graphene layers onto a Nano-Chip (Wildfire, 108 DENSsolutions) designed for in situ heating TEM observation. (a) Transfer 1L-graphene onto 1L-109 graphene/Cu for increasing the number of layers. (b) Encapsulation of Ge and transfer onto a Nano-Chip. 110  111 Figure 2(a) shows a photograph of the Nano-Chip. The size of the Nano-Chip is approximately 4 mm × 112 10 mm. The graphene/Ge/graphene was placed inside the broken-line rectangular area where the heater 113 of the Nano-Chip is located.  Figure 2(b) shows the SEM image of the heater part after the 114 graphene/Ge/graphene transfer. The heater has a spiral shape. Figure 2(c) shows a magnified SEM image 115 and the Raman spectrum recorded inside the lower hole (red arrow). In the Raman spectrum, D, G, and 116 G' (2D) peaks were observed at ~1360, ~1583, and 2725 cm-1, respectively. The small intensity of the D 117 peak and the narrow bandwidth of the G peak (about 16 cm-1 at full width half maximum) indicate high 118    7 crystallinity.[22] The G'/G peak intensity ratio (~0.8) indicates multilayer graphene,[22b, 22c] which is 119 consistent with our sample preparation process. Figure 2(d) shows the schematic cross-sectional view of 120 the fabricated sample. Figure 2(e) shows an energy dispersive X-ray spectroscopy (EDS) spectrum 121 obtained at the center of the lower hole (red arrow). Ge, Cu, oxygen (O), nitrogen (N), carbon (C), and 122 silicon (Si) were detected. The detected Cu is the residue from the substrate of CVD graphene, and the Si 123 and N are from the Nano-Chip. The detection of the Ge signal suggests that the sample preparation of the 124 Ge encapsulation by graphene was successful. 125  126 Figure 2. (a) Photograph of the Nano-Chip. (b) The SEM image of the heater part after the 127 graphene/Ge/graphene transfer. (c) Magnified SEM image and the Raman spectrum recorded inside the 128    8 lower hole (red arrow). (d) Schematic cross-sectional views of the fabricated sample. (e) EDS spectrum 129 obtained at the center of the lower hole. 130  131 Figures 3(a) and 3(b) are high-angle annular dark field scanning TEM (HAADF-STEM) images of the 132 graphene/Ge/graphene sample at RT before and after EDS analysis, respectively. Figures 3(c-h) show the 133 EDS maps of C, N, O, Si, Cu, and Ge. The distributions of N, Si, and Cu were not clearly defined in the 134 EDS maps due to the very low abundance or uniform distribution of these elements. On the other hand, 135 the distribution of Ge was clearly detected by the EDS map (Fig. 3(h)). Since the signal obtained by 136 HAADF-STEM is proportional to the α power of the atomic number (Z) [Zα (α = 1.3 ~ 2)],[23] the bright 137 contrasts in Figs. 3(a) and (b) indicate Ge morphology. The observed Ge was agglomerated, which was 138 probably due to the insufficient adhesion between Ge and graphene. As can be seen in Figs. 3(a) and 3(b), 139 the morphology of Ge changes from lumpy to smoothly connected aggregates after long-time electron 140 beam irradiation (>1 hour) during the EDS mapping measurement. Although the detailed mechanism of 141 the morphological change is unclear, unnecessary electron beam irradiation was avoided in order to 142 minimize damage to the sample hereafter. 143    9  144 Figure 3. HAADF-STEM images and EDS maps of the graphene/Ge/graphene sample at RT before 145 annealing processes. HAADF-STEM images (a) before and (b) after EDS analysis. EDS maps of (c) C, 146 (d) N, (e) O, (f) Si, (g) Cu, and (h) Ge. 147  148 149    10 Subsequently, we started to heat the sample using the in situ heating holder followed by TEM observation. 150 The sample was annealed many times with simple heating profiles of different temperatures (T) and 151 heating time duration (tHT) as shown in Fig. 4. The typical heating rate was approximately +100 °C/min 152 and typical cooling rate was approximately -100 °C/min. All annealing processes for this sample are 153 summarized in Table 1. 154  155 Figure 4. Schematic of the heating profile for the graphene/Ge/graphene sample for each annealing 156 process. tHT is heating time duration. 157  158  159  160  161  162  163    11 Table 1. Heating temperature (T) and time (tHT) for all annealing processes for the graphene(4L)/Ge(20 164 nm)/graphene(4L) sample in this study. 165 process # T (°C) tHT (min) 1 RT - 2 650 10 3 700 10 4 750 10 5 800 10 6 900 0.5 7 950 0.5 8 1000 0.5 9 1025 0.5 10 1050 0.5 11 1050 3 12 1025 0.5 13 1025 10 14 1050 20 15 1075 12 16 1075 10  166 To achieve the purpose of this study, we need to observe the crystallization of Ge between graphene layers 167 by TEM. However, the sample is not uniform as can be seen from Fig. 3. Therefore, we attempted to find 168 as complete a graphene/Ge/graphene region as possible for in situ observation. Figure 5(a) shows a TEM 169 image of the sample after annealing at 700 °C (process #3 in Table 1); the annealing at mild temperature 170 was performed to clean the sample. The band-like structure (from top left to bottom right) and dark 171 contrast in Fig. 5(a) are likely graphene and Ge, respectively. To determine the number of graphene layers, 172 HRTEM images and corresponding fast Fourier transform (FFT) images (Figs. 5(b-d)) were taken at the 173 arrowed points (b), (c), and (d) in Fig. 5(a). The FFT images (Figs. 5(b-d)) show several six-fold 174 symmetry spots corresponding to graphene. The transfer of graphene from one graphene to another 175 without intentional alignment in their crystalline direction, such as in the present sample preparation, 176 results in random stacking of those graphene layers at a high probability. Therefore, the number of six-177    12 fold symmetry spots guarantees that the same number of graphene layers present at the point where the 178 FFT image were taken. Counting the number of spots in a 60-degree range in Figs. 5(b-d), 4, 7, and 8 six-179 fold symmetry spots were observed, respectively, so that the numbers of graphene layers at these points 180 are 4, 7, and 8L. According to the numbers of those graphene layers at (b-d) and the TEM contrast in Fig. 181 5(a), the expected arrangement of graphene and Ge is shown in Fig. 5(e), where red (4L) and yellow (7-182 8L) are being superimposed as the number of graphene layers on the TEM image (Fig. 5(a)). Since there 183 are a lot of dark contrast in the yellow region in Fig. 5(e), it is expected that graphene/Ge/graphene is 184 locally remaining. Figure 5(f) shows a possible cross-sectional structure on the line of α to β in Fig. 5(a). 185 Although it is possible that the Ge is not covered by graphene layers, as shown in Fig. 5(g), uncovered 186 Ge would evaporate at temperatures higher than MPGe (938°C) because of vapor pressure. 187    13  188 Figure 5. (a) TEM image of the sample after annealing at 700 °C (process #3). (b-d) FFT images obtained 189 in (a). The relatively clear 6-fold spots corresponding to graphene are encircled by ellipses as eye guides. 190 The white numbers indicate the numbers of spots counted in a 60-degree range. (e) Expected arrangement 191 of graphene and Ge. Red (4L) and yellow (7-8L) colors are superimposed as the number of graphene 192 layers on (a). (f,g) Possible cross-sectional structure on the line of α to β in (a). 193    14  194 2. Suppression of Ge evaporation by graphene encapsulation 195 To confirm the encapsulation of Ge by graphene layers, we performed several heating processes (#4 ~ 196 12) including the heating temperatures higher than MPGe. Figures 6(a) and 6(b) show TEM images 197 obtained at RT after the annealing #12. The band-like structure expected to be a graphene/Ge/graphene 198 region seems to still present. Subsequently, additional annealing was performed at 1025 °C for 10 min 199 (process #13), and TEM observation was performed at RT in the exact same area as in Figs. 6(a) and 6(b). 200 Figures 6(c) and 6(d) show the TEM images after annealing at 1025 °C (process #13). It was observed 201 that the different contrast changes depending on the region by the annealing at 1025 °C. While the dark 202 contrasts were maintained or darkened more between the dashed lines in Fig. 7(c), the dark contrasts 203 disappeared or faded in the other regions. This indicates that graphene/Ge/graphene area exists between 204 the broken lines and that graphene encapsulation suppresses Ge evaporation at higher temperatures than 205 the MPGe. In the graphene-encapsulated regions, it was also observed that the number of black dots 206 increased and became more clearly visible. Figure 6(e) shows a HRTEM image of a black dot. Several 207 lattice fringes were observed in the HRTEM image. Figures 6(f) and 6(g) show the FFT images obtained 208 within the upper and lower circles enclosed by the yellow broken lines in Fig. 6(e), respectively. The FFT 209 images agree well with Ge (111) and Ge (110), indicating that the black dots are crystalline particles of 210 Ge. On the other hand, the majority of the surrounding area showed no spots in the FFT image. This 211 indicates that the majority of Ge between graphene layers remains amorphous state even after annealing 212 above MPGe. 213  214    15  215 Figure 6. (a,b) TEM images obtained at RT after several heating processes (#4 ~ 12).  (c,d) TEM images 216 obtained at the same area as in (a) and (b) after annealing at 1025 °C for 10 min (process #13). Expected 217 graphene covered region is between the broken yellow lines in (c). (e) HRTEM image of a black dot. (f,g) 218 FFT images obtained within the upper and lower circles enclosed by the yellow broken lines in (e), 219 respectively. The FFT images agree well with Ge (111) and Ge (110), indicating that the black dots are 220 crystalline Ge. 221    16  222 3. Crystal phase of Ge at high temperatures 223 Figure 7 shows the TEM images of black dots at high temperatures (>MPGe) extracted from the movie 224 (S1 “fusion”) obtained by in situ TEM observation at 1050 °C (process #14). The time shown in the upper 225 right corner for each TEM image correspond to the time after the start of the video (S1 “fusion”) recording 226 in real time (The video plays at ~5x speed). We observed several spherical dark contrasts, which we will 227 refer to hereafter as "droplet Ge". In Fig. 7, we mainly focused on the droplet Ge indicated by the arrow 228 in the first TEM image (20s). As can be seen, the droplet Ge moved, deformed, and finally merged with 229 the other droplet Ge on the right side. This flexible movement and coalescence of Ge indicates that the 230 droplet Ge is in a liquid state at 1050 °C. It is noted that bare Ge (not covered by graphene) is not expected 231 to exist in a liquid or solid state at 1050 °C, since the vapor pressure of Ge at 1050 °C is about 10 Pa, 232 much higher than the pressure in the TEM chamber (< 2×10-5 Pa). On the other hand, the droplets showed 233 a crystalline structure at RT (Figs. 6(e-g)), suggesting that they crystallized during cooling. 234  235    17 Figure 7. TEM images of black dots at high temperatures (>MPGe) extracted from the movie (S1 “fusion”) 236 obtained by in situ TEM observation at 1050 °C. The times shown in the upper right corner for each TEM 237 image correspond to the time after the start of the video (S1 “fusion”) recording in real time (not video 238 time). 239  240 Next, we performed HRTEM observation of droplet Ge at 1075 °C. Figure 8(a) shows an HRTEM image 241 extracted from the movie obtained by in situ TEM observation at 1075 °C (process #16). Figures 8(b-d) 242 show the FFT images in Fig. 8(a). We observed lattice fringes in the area shown in (b) of Fig. 8(a) (it is 243 more clearly seen in the area encircled by the ellipse in the lower center). The corresponding FFT image 244 (Fig. 8(b)) shows several spots, indicating the presence of ordered structure. On the other hand, some 245 droplets Ge do not have clear lattice fringes, such as (c) and (d) in Fig. 8(a). It is noted that the area (d) 246 may not be a droplet Ge but simply a vacant area. Those FFT images (Figs. 8(c) and 8(d)), which is 247 different from Fig. 8(b), shows a few faint and unclear spots. In addition, the obtained three FFT spot 248 patterns (Fig. 8(b-d)) do not agree with that of Ge (totally different from Figs. 6(f) and 6(g)), graphene, 249 or even meatal impurities such as Cu. We assumed that the observed FFT spot patterns are originated 250 from the structure of liquid Ge. It is widely reported that liquid Ge has short-range ordered structure.[24] 251 The reported average nearest-neighbor distances of Ge-Ge at high temperatures (2.66 Å at 1175-1650 252 K,[24b] 2.68-2.69 Å at 1120-1534 K,[24c] 2.74-2.76 Å at 1270-1870 K[24d]) are ~8% to ~13% larger than 253 crystalline Ge at RT (2.45-2.46 Å[24e]). Also, by taking into account the self-diffusion coefficient of Ge 254 atoms in liquid phase (1.3 × 10-8 m2 s-1 at the temperature of MPGe[25]), Ge atoms in a droplet Ge move 255 extremely fast in the size of the HRTEM image. Therefore, the short-range ordered structure in liquid Ge 256 would result in different spot patterns in the FFT image as compared to diamond cubic crystal of Ge, and 257    18 the fast movement of Ge atoms in liquid Ge would cause fluctuations and instability in the FFT spot 258 patterns. 259  260    19 Figure 8. (a) HRTEM image of droplets Ge extracted from the movie obtained by in situ TEM observation 261 at 1075 °C (process #16). (b-d) FFT images obtained within (b), (c) and (d) in (a). 262  263 Figure 9 shows the STEM-EDS results obtained at RT after all annealing processes (after process #16). 264 Figure 9(a) shows the EDS spectrum. Figures 9(b) and 9(c) show HAADF-STEM images before and after 265 the EDS analysis, respectively. Figures 9(d-i) show EDS maps of C, N, O, Si, Cu, and Ge. From Fig. 9(a), 266 Ge and a little amount of Mo and Cu were detected. The Mo signal is from the material of heating wire 267 of the Nano-Chip. The Cu would be contamination from the sample preparation process (substrate of 268 graphene grown by CVD). Since the amount of Cu detected was very small and no agglomeration was 269 observed (Fig. 9(h)), it was confirmed that Cu did not affect the TEM observations and the 270 aforementioned discussion of droplet Ge. On the other hand, the contrast distribution in the HAADF-271 STEM images (Figs. 9(b) and 9(c)) is consistent with the distribution of Ge (Fig. 9(i)). Thus, the particles 272 (droplet Ge) observed in Figs. 7 and 8 are surely made of Ge. Furthermore, no oxygen was detected at 273 the Ge positions (Fig. 9(f)), indicating that droplet Ge is pure germanium, not Ge oxides such as GeO2. 274 These results suggest that Ge can exist as a liquid phase between vdW layers even above MPGe. 275    20  276 Figure 9. STEM-EDS obtained at RT after all annealing processes (after process #16). (a) EDS spectrum. 277 (b,c) HAADF-STEM images (b) before and (c) after the EDS analysis. (d-i) EDS maps of (d) C, (e) N, 278 (f) O, (g) Si, (h) Cu, and (i) Ge. 279  280 4. Nucleation and domain growth of Ge encapsulated by graphene layers 281 Figures 10(a-f) show the TEM images extracted from the in situ TEM movie for the process # 13 (S2 282 “growth”).  The time and temperature are shown in the lower right corner of each TEM image. During 283 the heating process, contrast changes were observed at various regions. Here, we focus on the circled 284 region near the center of the TEM images. The contrast in the circled region remained dark during the 285 temperature increase from RT to 1000 °C (Fig. 10(a-c)). Thereafter, the dark contrast faded while the 286 temperature was kept constant at 1025 °C (Fig. 10(d) and 10(e)). 287    21 Figures 11(a-d) show magnified TEM image around the circled region in Fig. 10. The time and 288 temperature are shown in the upper left corner of each TEM image. At 750 °C during the temperature 289 increase, a tiny droplet Ge was observed (indicated by the arrow in Fig. 11(a)). When the temperature 290 reached 1025 °C, a new droplet Ge appeared (indicated by the left arrow in Fig. 11(b)). While the 291 temperature was maintained at 1025 °C, the size of the droplets Ge increased (Figs. 11(b-d)). At the same 292 time, the contrast in the circled broken line in Fig.11(a) gradually changed from dark to faint. The fading 293 contrast probably means that the condensed amorphous Ge diffused into the surroundings, resulting in 294 the reduction its concentration. These indicates that Ge crystal growth occurred and a Ge nucleus 295 incorporated the surrounding amorphous Ge. This situation is similar to the bulk crystal growth of Si and 296 Ge, such as Czochralski method,[20] and our TEM observations suggests that crystal growth of Ge between 297 graphene layers is possible. However, the crystal growth in this experiment (Fig. 11) seems to be three-298 dimensional because the shape of the droplet Ge was spherical and no information about the two-299 dimensionality could be obtained from the crystallized Ge. Therefore, more unique and new ideas are 300 needed to achieve the growth of 2D germanium (germanene) between vdW layers. For example, limiting 301 the direction of crystal growth by thinning the initial Ge source film to make it more truly 2D, or creating 302 a monolayer vdW/metal (111) heterostructure to induce remote epitaxy, as has been reported for graphene 303 growth on GaAs.[26] 304    22  305 Figure 10. (a-f) TEM images extracted from the in situ TEM movie for the process #13. The time and 306 temperature are shown in the lower right corner of each TEM image. 307  308  309 Figure 11. (a-d) Magnified TEM image around the circled region in Fig. 10. The time and temperature 310 are shown in the upper left corner of each TEM image. Ge droplets are indicated by arrows. 311    23  312 Conclusion 313 In this study, we fabricated a graphene/Ge/graphene encapsulated structure using transfer process of CVD 314 graphene. The fabricated encapsulated Ge was observed by in situ TEM to investigate how the Ge in the 315 vdW layer behaves at high temperatures, especially at temperatures higher than MPGe. The results showed 316 that graphene encapsulation played a role in suppressing Ge evaporation even above MPGe. At higher 317 temperatures than MPGe, the movement of Ge droplets and their coalescence with other Ge droplets in 318 the surrounding area were observed. In situ HRTEM observation at high temperatures revealed that some 319 droplets Ge do not have clear lattice fringes, indicating that Ge can exist as a liquid phase between vdW 320 layers. It was also observed that droplet Ge incorporates the surrounding amorphous Ge as Ge nuclei, 321 thereby increasing its domain size. These results indicate that the Ge encapsulation structure with vdW 322 materials can be used to grow Ge crystals, such like Czochralski method, between vdW layers. Although 323 more unique and new ideas will be needed to achieve the growth of germanene between vdW layers, the 324 present results may provide clues for the future direct growth of germanene between hBN layers. 325  326 Methods 327 CVD growth of graphene: CVD graphene was grown on a commercial Cu sheet (99.96% in purity, 100 328 µm thickness, Nilaco) in a homemade CVD system in NIMS. Polycrystalline monolayer graphene with 329 high coverage was targeted in the present CVD. The growth process was as follows: i) increasing the 330 temperature from RT to 800 °C in Ar; ii) increasing the temperature from 800 to 950 °C in H2/Ar (~15% 331 H2); iii) H2/Ar (~10% H2) annealing for 100 min at 1000 °C; iv) O2/Ar (~20 ppm O2) annealing for 35 332 min; v) graphene growth by introducing CH4 (~50 ppm) with H2 and Ar for 90 min; vi) additional 333    24 graphene growth in H2/Ar and high concentrations of CH4 (~0.1 %) for 5 min; vii) cooling to RT in H2/Ar 334 (~2.5% H2). The H2/Ar and O2/Ar annealing were performed to reduce nucleation density of graphene.[27] 335 The additional growth, so-called two-step growth,[28] was performed to increase the coverage of graphene. 336 The mechanisms related to this CVD growth are presented elsewhere.[27b, 27c, 29] 337 Transfer graphene: The transfer process (Fig. 1) basically follows the standard wet transfer process for 338 graphene.[30] PMMA was dissolved in anisole and then the suspension was spin-coated onto graphene/Cu. 339 (NH4)2S2O8 was used to etch Cu. To begin, Cu was etched for a short time (5-10 min). The sample was 340 taken out from the etchant and the backside graphene was removed by softly rubbing the backside of the 341 Cu with a commercially available cotton swab. After replacing the etchant with a new one, etching of Cu 342 restarted again. The short etching and rubbing were repeated twice in total, followed by the main etching 343 (usually several hours). After multiple DI rinses, the PMMA/graphene was then scooped up with another 344 graphene/Cu, and finally the PMMA layer was removed with acetone. 345 Ge deposition: Ge thin film of 20 nm thickness was deposited at RT by an e-beam evaporator (RDEC Co., 346 with the Ltd., RDEB-1206K). The deposition rate was ~1.0 Å s-1. 347 SEM-EDS: SEM and EDS analysis were performed using a commercial SEM (Hitachi TM4000). 348 Secondary electron images and EDS spectra were obtained at an acceleration voltage of 15 kV. 349 Raman spectroscopy: A commercial Raman microscope (Horiba, LabRam HR Evolution) with an 350 excitation wavelength of 458 nm was used for obtaining Raman spectra. 351 In situ heating TEM experiments: In situ TEM experiments on graphene-encapsulated Ge were conducted 352 using a DENSsolutions lightning sample holder in a spherical aberration (Cs)-corrected TEM, JEM-353    25 ARM200F, operated at 200kV. OneView camera and drift correction were used. EDS mapping after a 354 heating process was performed after the sample temperature was cooled to RT. 355  356 References 357 [1] A. Acun, L. Zhang, P. Bampoulis, M. Farmanbar, A. van Houselt, A. Rudenko, M. Lingenfelder, G. 358 Brocks, B. Poelsema, M. Katsnelson, J. Phys.: Condens. Matter 2015, 27, 443002. 359 [2] S. Cahangirov, M. Topsakal, E. Aktürk, H. Şahin, S. Ciraci, Phys. Rev. Lett. 2009, 102, 236804. 360 [3] Z. Ni, Q. Liu, K. Tang, J. Zheng, J. Zhou, R. Qin, Z. Gao, D. Yu, J. Lu, Nano Lett. 2011, 12, 113. 361 [4] M. Ezawa, J. Phys. Soc. Jpn. 2015, 84, 121003. 362 [5] G. Fiori, F. Bonaccorso, G. Iannaccone, T. Palacios, D. Neumaier, A. Seabaugh, S. K. Banerjee, L. 363 Colombo, Nat. Nanotechnol. 2014, 9, 768. 364 [6] P. Vogt, P. De Padova, C. Quaresima, J. Avila, E. Frantzeskakis, M. C. Asensio, A. Resta, B. Ealet, 365 G. Le Lay, Phys. Rev. Lett. 2012, 108, 155501. 366 [7] a) J. Yuhara, Y. Fujii, K. Nishino, N. Isobe, M. Nakatake, L. Xian, A. Rubio, G. Le Lay, 2D Mater. 367 2018, 5, 025002; b) F.-f. Zhu, W.-j. Chen, Y. Xu, C.-l. Gao, D.-d. Guan, C.-h. Liu, D. Qian, S.-C. 368 Zhang, J.-f. Jia, Nat. Mater. 2015, 14, 1020; c) Y. Xu, B. Yan, H.-J. Zhang, J. Wang, G. Xu, P. Tang, 369 W. Duan, S.-C. Zhang, Phys. Rev. Lett. 2013, 111, 136804. 370 [8] J. Yuhara, B. He, N. Matsunami, M. Nakatake, G. Le Lay, Adv. Mater. 2019, 31, 1901017. 371 [9] a) S. Suzuki, T. Iwasaki, K. K. H. De Silva, S. Suehara, K. Watanabe, T. Taniguchi, S. Moriyama, 372 M. Yoshimura, T. Aizawa, T. Nakayama, Adv. Funct. Mater. 2021, 31, 2007038; b) L. Kesper, J. A. 373 Hochhaus, M. Schmitz, M. G. Schulte, U. Berges, C. Westphal, Sci. Rep. 2022, 12, 1; c) J. Yuhara, 374 H. Shimazu, K. Ito, A. Ohta, M. Araidai, M. Kurosawa, M. Nakatake, G. Le Lay, ACS Nano 2018, 375 12, 11632; d) C.-H. Lin, A. Huang, W. W. Pai, W.-C. Chen, T.-Y. Chen, T.-R. Chang, R. Yukawa, C.-376    26 M. Cheng, C.-Y. Mou, I. Matsuda, C. T. C., J. H. T., T. S. J., Phys. Rev. Mater. 2018, 2, 024003; e) 377 A. Ohta, K. Yamada, H. Sugawa, N. Taoka, M. Ikeda, K. Makihara, S. Miyazaki, Jpn. J. Appl. Phys. 378 2021, 60, SBBK05; f) S. Suzuki, K. K. H. De Silva, M. Yoshimura, T. Nakayama, Jpn. J. Appl. Phys. 379 2020, 59, SN1004; g) K. Ito, A. Ohta, M. Kurosawa, M. Araidai, M. Ikeda, K. Makihara, S. Miyazaki, 380 Jpn. J. Appl. Phys. 2018, 57, 06HD08. 381 [10] a) M. Dávila, L. Xian, S. Cahangirov, A. Rubio, G. Le Lay, NJPh 2014, 16, 095002; b) M. E. Dávila, 382 G. Le Lay, Sci. Rep. 2016, 6, 20714. 383 [11] Z. Qin, J. Pan, S. Lu, Y. Shao, Y. Wang, S. Du, H. J. Gao, G. Cao, Adv. Mater. 2017, 29, 1606046. 384 [12] a) Y. Sato, Y. Fukaya, M. Cameau, A. K. Kundu, D. Shiga, R. Yukawa, K. Horiba, C.-H. Chen, A. 385 Huang, H.-T. Jeng, T. Ozaki, H. Kumigashira, M. Niibe, I. Matsuda, Phys. Rev. Mater. 2020, 4, 386 064005; b) J. Yuhara, H. Muto, M. Araidai, M. Kobayashi, A. Ohta, S. Miyazaki, S.-i. Takakura, M. 387 Nakatake, G. Le Lay, 2D Mater. 2021, 8, 045039; c) M. Derivaz, D. Dentel, R. Stephan, M.-C. Hanf, 388 A. Mehdaoui, P. Sonnet, C. Pirri, Nano Lett. 2015, 15, 2510; d) S. Endo, O. Kubo, N. Nakashima, 389 S. Iwaguma, R. Yamamoto, Y. Kamakura, H. Tabata, M. Katayama, Appl. Phys. Express 2017, 11, 390 015502; e) O. Kubo, S. Kinoshita, H. Sato, K. Miyamoto, R. Sugahara, S. Endo, H. Tabata, T. Okuda, 391 M. Katayama, Phys. Rev. B 2021, 104, 085404; f) M. Kobayashi, A. Ohta, M. Kurosawa, M. Araidai, 392 N. Taoka, T. Simizu, M. Ikeda, K. Makihara, S. Miyazaki, Jpn. J. Appl. Phys. 2020, 59, SGGK15. 393 [13] L. Li, S. z. Lu, J. Pan, Z. Qin, Y. q. Wang, Y. Wang, G. y. Cao, S. Du, H. J. Gao, Adv. Mater. 2014, 394 26, 4820. 395 [14] L. Persichetti, F. Jardali, H. Vach, A. Sgarlata, I. Berbezier, M. De Crescenzi, A. Balzarotti, J. Phys. 396 Chem. 2016, 7, 3246. 397 [15] a) L. Zhang, P. Bampoulis, A. Rudenko, Q. v. Yao, A. Van Houselt, B. Poelsema, M. Katsnelson, H. 398 Zandvliet, Phys. Rev. Lett. 2016, 116, 256804; b) Z. Jiao, Q. Yao, A. Rudenko, L. Zhang, H. 399 Zandvliet, Phys. Rev. B 2020, 102, 205419. 400    27 [16] a) B. Madhushankar, A. Kaverzin, T. Giousis, G. Potsi, D. Gournis, P. Rudolf, G. Blake, C. Van Der 401 Wal, B. Van Wees, 2D Mater. 2017, 4, 021009; b) Q. Chen, L. Liang, G. Potsi, P. Wan, J. Lu, T. 402 Giousis, E. Thomou, D. Gournis, P. Rudolf, J. Ye, Nano Lett. 2019, 19, 1520; c) Y. Katayama, R. 403 Yamauchi, Y. Yasutake, S. Fukatsu, K. Ueno, Appl. Phys. Lett. 2019, 115, 122101. 404 [17] L. Tao, E. Cinquanta, D. Chiappe, C. Grazianetti, M. Fanciulli, M. Dubey, A. Molle, D. Akinwande, 405 Nat. Nanotechnol. 2015, 10, 227. 406 [18] a) C. R. Dean, A. F. Young, I. Meric, C. Lee, L. Wang, S. Sorgenfrei, K. Watanabe, T. Taniguchi, P. 407 Kim, K. L. Shepard, J. Hone, Nat. Nanotechnol. 2010, 5, 722; b) J. Xue, J. Sanchez-Yamagishi, D. 408 Bulmash, P. Jacquod, A. Deshpande, K. Watanabe, T. Taniguchi, P. Jarillo-Herrero, B. J. LeRoy, Nat. 409 Mater. 2011, 10, 282; c) L. Banszerus, M. Schmitz, S. Engels, J. Dauber, M. Oellers, F. Haupt, K. 410 Watanabe, T. Taniguchi, B. Beschoten, C. Stampfer, Sci. Adv. 2015, 1, e1500222; d) T. Iwasaki, Y. 411 Morita, K. Watanabe, T. Taniguchi, Phys. Rev. B 2022, 106, 165134; e) T. Iwasaki, S. Nakaharai, Y. 412 Wakayama, K. Watanabe, T. Taniguchi, Y. Morita, S. Moriyama, Nano Lett. 2020, 20, 2551. 413 [19] a) M. Y. Chan, K. Komatsu, S.-L. Li, Y. Xu, P. Darmawan, H. Kuramochi, S. Nakaharai, A. 414 Aparecido-Ferreira, K. Watanabe, T. Taniguchi, K. Tsukagoshi, Nanoscale 2013, 5, 9572; b) G. H. 415 Lee, Y. J. Yu, X. Cui, N. Petrone, C. H. Lee, M. S. Choi, D. Y. Lee, C. Lee, W. J. Yoo, K. Watanabe, 416 T. Taniguchi, C. Nuckolls, P. Kim, J. Hone, ACS Nano 2013, 7, 7931; c) Y. Shimazaki, I. Schwartz, 417 K. Watanabe, T. Taniguchi, M. Kroner, A. Imamoğlu, Nature 2020, 580, 472. 418 [20] a) G. K. Teal, M. Sparks, E. Buehler, Phys. Rev. 1951, 81, 637; b) G. K. Teal, J. B. Little, Phys. Rev. 419 1950, 78, 647. 420 [21] a) P. Germain, K. Zellama, S. Squelard, J. Bourgoin, A. Gheorghiu, J. Appl. Phys. 1979, 50, 6986; 421 b) N. A. Blum, C. Feldman, J. Non-Cryst. Solids 1976, 22, 29; c) K. Chik, P.-K. Lim, Thin Solid 422 Films 1976, 35, 45. 423    28 [22] a) A. C. Ferrari, J. Robertson, Philosophical Transactions of the Royal Society of London A: 424 Mathematical, Physical and Engineering Sciences 2004, 362, 2477; b) L. Malard, M. Pimenta, G. 425 Dresselhaus, M. Dresselhaus, PhR 2009, 473, 51; c) A. Ferrari, J. Meyer, V. Scardaci, C. Casiraghi, 426 M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. Novoselov, S. Roth, Phys. Rev. Lett. 2006, 97, 187401. 427 [23] a) M. M. Treacy, Microscopy and Microanalysis 2011, 17, 847; b) N. Tanaka, Scanning transmission 428 electron microscopy of nanomaterials: basics of imaging and analysis, World Scientific,  2014. 429 [24] a) L. C. Kelsall, M. Peña-Alvarez, M. Martinez-Canales, J. Binns, C. J. Pickard, P. Dalladay-430 Simpson, R. T. Howie, E. Gregoryanz, The Journal of Chemical Physics 2021, 154, 174702; b) H. 431 Weis, D. Holland-Moritz, F. Kargl, F. Yang, T. Unruh, T. Hansen, J. Bednarčik, A. Meyer, Phys. Rev. 432 B 2021, 104, 134108; c) I. Pozdnyakova, O. Roik, J. W. Drewitt, A. Bytchkov, F. Kargl, S. Jahn, S. 433 Brassamin, L. Hennet, J. Phys.: Condens. Matter 2021, 33, 244002; d) Y. Naito, M. Inui, T. Anai, K. 434 Tamura, J. Non-Cryst. Solids 2007, 353, 3376; e) G. Kresse, J. Hafner, Phys. Rev. B 1994, 49, 14251; 435 f) N. Ashcroft, Il Nuovo Cimento D 1990, 12, 597. 436 [25] H. Weis, F. Kargl, M. Kolbe, M. Koza, T. Unruh, A. Meyer, J. Phys.: Condens. Matter 2019, 31, 437 455101. 438 [26] Y. Kim, S. S. Cruz, K. Lee, B. O. Alawode, C. Choi, Y. Song, J. M. Johnson, C. Heidelberger, W. 439 Kong, S. Choi, Nature 2017, 544, 340. 440 [27] a) I. Vlassiouk, M. Regmi, P. Fulvio, S. Dai, P. Datskos, G. Eres, S. Smirnov, ACS nano 2011, 5, 441 6069; b) S. Suzuki, K. Kiyosumi, T. Nagamori, K. Tanaka, M. Yoshimura, e-J. Surf. Sci. 442 Nanotechnol. 2015, 13, 404; c) S. Suzuki, Y. Terada, M. Yoshimura, Coatings 2017, 7, 206. 443 [28] X. Li, C. W. Magnuson, A. Venugopal, J. An, J. W. Suk, B. Han, M. Borysiak, W. Cai, A. 444 Velamakanni, Y. Zhu, L. Fu, E. M. Vogel, E. Voelkl, L. Colombo, R. S. Ruoff, Nano Lett. 2010, 10, 445 4328. 446    29 [29] Y. Hao, M. S. Bharathi, L. Wang, Y. Liu, H. Chen, S. Nie, X. Wang, H. Chou, C. Tan, B. Fallahazad, 447 H. Ramanarayan, C. W. Magnuson, E. Tutuc, B. I. Yakobson, K. F. McCarty, Y. W. Zhang, P. Kim, 448 J. Hone, L. Colombo, R. S. Ruoff, Science 2013, 342, 720. 449 [30] a) X. Li, Y. Zhu, W. Cai, M. Borysiak, B. Han, D. Chen, R. D. Piner, L. Colombo, R. S. Ruoff, Nano 450 Lett. 2009, 9, 4359; b) H. J. Jeong, H. Y. Kim, S. Y. Jeong, J. T. Han, K.-J. Baeg, J. Y. Hwang, G.-451 W. Lee, Carbon 2014, 66, 612; c) S. Suzuki, M. Yoshimura, Sci. Rep. 2017, 7, 14851. 452  Acknowledgements 453 This study was supported in part by JSPS KAKENHI Grant number  20K15134 (Grant-in-Aid for 454 Early-Career Scientists) from the Ministry of Education, Culture, Sports, Science and Technology 455 (MEXT), Japan; Japan Science and Technology Agency (JST), Precursory Research for Embryonic 456 Science and Technology (PRESTO) Grant Number JPMJPR21B7; NIMS Nanofabrication Platform 457 supported by “Nanotechnology Platform Program” of the MEXT, Grant Number JPMXP09F19NMN010; 458 Materials Analysis Station in NIMS; NIMS Electron Microscopy Analysis Station, Nanostructural 459 Characterization Group; the Public/Private R&D Investment Strategic Expansion Program (PRISM) from 460 Cabinet Office, Japan; the Center for Functional Sensor and Actuator (CFSN) from NIMS. 461  462 Author contributions statement 463 S.S. designed the research. S.S. and N.S. conducted sample preparation, SEM-EDS, and Raman 464 spectroscopy. Y.Ne. conducted TEM observation. S.S., N.S., Y.Na, and Y.Ne discussed many times to 465 improve the sample preparation and TEM observation. S.S. and Y.Ne analyzed the results. S.S., Y.Ne., 466 Y.Na and M.T. discussed interpretation of the results. S.S. mainly wrote the manuscript. All authors 467 reviewed the manuscript. 468  469    30 Conflict of Interest 470 The authors declare no conflict of interest. 471  472 Keywords 473 germanene, germanium, graphene, van der Waals materials, interface, crystal growth, nucleation, 474 transmission electron microscopy, graphene cell for transmission electron microscopy, high-angle annular 475 dark field scanning transmission electron microscopy, in situ heating measurement 476