Matter and Radiation at Extremes, Volume. 10, Issue 1, 017801(2025)

Macroscopic perspective on phase transition behavior of natural single-crystal graphite under different pressure environments

Xiaoshuang Yin1,*... Songyang Li2, Lijuan Wang1, Peiyuan Liu1, Zhihai Cheng2, Huiyang Gou1 and Liuxiang Yang1 |Show fewer author(s)
Author Affiliations
  • 1Center for High Pressure Science and Technology Advanced Research, Beijing 100193, China
  • 2Key Laboratory of Quantum State Construction and Manipulation (Ministry of Education), Department of Physics, Renmin University of China, Beijing 100872, China
  • show less

    Comprehensive understanding of the direct transformation pathway from graphite to diamond under high temperature and high pressure has long been one of the fundamental goals in materials science. Despite considerable experimental and theoretical progress, current experimental studies have mainly focused on the local microstructural characterizations of recovered samples, which has certain limitations for high-temperature and high-pressure products, which often exhibit diversity. Here, we report on the pressure-induced phase transition behavior of natural single-crystal graphite under three distinct pressure-transmitting media from a macroscopic perspective using in situ two-dimensional Raman spectroscopy, scanning electron microscopy, and atomic force microscopy. The surface evolution process of graphite before and after the phase transition is captured, revealing that pressure-induced surface textures can impede the continuity of the phase transition process across the entire single crystal. Our results provide a fresh perspective for studying the phase transition behavior of graphite and greatly deepen our understanding of this behavior, which will be helpful in guiding further high-temperature and high-pressure syntheses of carbon allotropes.

    I. INTRODUCTION

    Graphite, one of the most common minerals found in nature, usually has a hexagonal structure with ABAB stacking order. In the layers, carbon atoms are connected by strong sp2 bonding to form honeycomb basal planes, which are then weakly bound by van der Waals forces.1–3 When graphite is squeezed to sufficient pressure, it begins to lose its anisotropic character and emerge into a high-pressure phase at room temperature.4–6 As the temperature rises, graphite eventually stabilizes into cubic diamond.7–9

    According to the epitaxial relationship, the direct conversion of graphite into diamond through the martensitic path requires two steps: buckling and sliding.10–13 Previous studies have proposed various phase transition models along this pathway, including chair buckling, boat buckling, and wave-like buckling models.4,11,13–18 Current understanding of the phase transition behavior of graphite relies predominantly on local analyses of recovered high-temperature and high-pressure (HTHP) samples using transmission electron microscopy.18,19 However, the changes in carbon precursors and experimental conditions always lead to discrepancies in the micro-morphology and polyphase structures of recovered samples, which impose certain limitations on microscopic characterization.20–22 Obtaining all in situ information of graphite from a macroscopic perspective, including deformation patterns, stress distribution, and the continuity of the phase transition process, would effectively compensate for the limitations of microscopic characterization.

    Investigating the behavior of graphite under cold compression, which can be studied in situ by combining two-dimensional (2D) Raman spectroscopy with a diamond anvil cell (DAC), could serve as an effective way to obtain in situ macroscopic information about graphite under pressure. According to the phase diagram of carbon, graphite begins to transform into a metastable phase when compressed to 14 GPa at room temperature.23 During this process, structural transformations, such as rearrangement of the atomic configuration and the corresponding rise in internal disorder, form the basis for the emergence of various carbon phases in subsequent heat treatment. In situ information obtained throughout the entire sample at this stage could provide vital insights into the transition process from graphite to diamond under HTHP conditions. Moreover, in a recent report,24 we have noted that the phase transition pressure of highly ordered pyrolytic graphite exhibits notable variability across diverse pressure transmitting media (PTM), suggesting that the pressure environment may have a substantial impact on the phase transition behavior of graphite. A systematic investigation to illustrate the response of graphite under different pressure environments and the underlying mechanisms may facilitate regulation of the phase transition process, enabling the designed synthesis of post-graphite products.

    Therefore, in the study reported here, we employed a 2D Raman imaging technique in a DAC to systematically investigate the in situ pressure-induced phase transition process of graphite under different pressure environments at room temperature. Structurally homogeneous and highly ordered natural single-crystal graphite (SCG) was used as starting material. Three PTM materials, namely, helium (He), argon (Ar) and cubic boron nitride (cBN), with significant varying stiffness were used to generate different pressure environments. Additionally, through morphological characterization of the recovered samples via scanning electron microscopy (SEM) and atomic force microscopy (AFM), more details of the surface deformation of graphite under pressure were revealed.

    II. EXPERIMENTAL METHODS

    In situ high-pressure experiments were conducted using a DAC with rhenium (Re) gaskets. He, Ar, and cBN were used as PTM materials. Pressure calibration in runs using He and Ar as PTM materials was done by ruby fluorescence,25 whereas the pressure in the cBN experiment was calibrated utilizing the diamond Raman edge, positioned at the central region of the diamond anvil, aligning with the sample’s center.26 The culets of the anvils were 400, 400, and 500 μm, respectively. To preserve crystal integrity, a low-power laser was used to cut the SCG into cylindrical shapes from flat regions of the original bulk material, with diameters of 78, 62, and 142 μm, respectively, each having a thickness approaching 10 μm. 2D Raman data were acquired using a confocal Raman spectroscope (excitation wavelength 532 nm with 40 mW power) equipped with an objective lens (20×). The measurement parameters were set as follows: 90 × 90 lines scanned over 90 × 90 μm2 for SCG in He, 70 × 70 lines scanned over 70 × 70 μm2 for SCG in Ar, and 150 × 150 lines scanned over 150 × 150 μm2 for SCG in cBN, all with an exposure time of 2 s and a single acquisition. The morphologies of both pristine and recovered SCG samples were examined using SEM, and the 3D surface structures of these specimens were investigated by AFM with scan sizes of 10 × 10, 20 × 20, and 30 × 30 μm2, respectively.

    III. RESULTS AND DISCUSSION

    A. Characterization of the initial material

    Figure 1(a) presents an optical image of SCG, which is characterized by a smooth surface accompanied by striped structures distributed at specific angles. The SEM image in the upper left corner of Fig. 1(a) displays laser-cut SCG in a cylindrical shape, revealing a flat surface that could provide detailed information on the changes in SCG during the phase transition process. Figure 1(b) shows the Raman spectrum of the SCG, exhibiting a sharp G peak and an absence of the D (disorder) peak, indicative of a highly ordered structure and excellent crystallinity.

    (a) Optical image of SCG, with the upper left insert showing an SEM image of SCG after laser cutting. (b) Raman spectrum of laser-cut SCG with a 532 nm excitation laser under ambient conditions.

    Figure 1.(a) Optical image of SCG, with the upper left insert showing an SEM image of SCG after laser cutting. (b) Raman spectrum of laser-cut SCG with a 532 nm excitation laser under ambient conditions.

    B. Phase transition of SCG in Ar

    In the experiment with Ar as the PTM, as illustrated in Fig. 2(a), the optical images (observed under reflected light) of SCG reveal the dynamic evolution of its surface morphology under pressure. At 1.5 GPa, the sample surface remains basically unchanged. However, as the pressure reaches 6.5 GPa, the SCG surface starts to exhibit a distinctive striped feature. These interlaced stripes mainly form fixed angles of 60° and 120°, which can also be observed in graphite crystals under ambient pressure [Fig. 1(a)] owing to the twinning effect of their in-plane hexagonal symmetry.10,27 When the pressure increases to 18.0 GPa, these striped structures become denser and begin to stabilize, with no further discernible variations even as the pressure continues to increase. Instead, the optical properties of the SCG surface change. From 19.2 to 28.3 GPa, as the pressure increases, the color of the SCG gradually changes from white gray to dark gray. The decrease in reflectance of the SCG indicates a change in its bandgap under pressure, suggesting the occurrence of a phase transition.13,28

    (a) Optical photographs of SCG surface morphology under different pressures in Ar. (b) and (c) 2D FrequencyG distribution imaging maps and 2D FWHMG distribution imaging maps of SCG under different pressures in Ar, respectively. Images were acquired at 1.5, 6.5, 18.0, 19.3, 21.3, and 28.3 GPa.

    Figure 2.(a) Optical photographs of SCG surface morphology under different pressures in Ar. (b) and (c) 2D FrequencyG distribution imaging maps and 2D FWHMG distribution imaging maps of SCG under different pressures in Ar, respectively. Images were acquired at 1.5, 6.5, 18.0, 19.3, 21.3, and 28.3 GPa.

    The 2D Raman imaging maps of the G-peak frequency (FrequencyG) and the G-peak full width at half maximum (FWHMG) of SCG provide detailed information on the phase transformation process of SCG. Under pressure, the frequency shift and significant broadening of the G peak in graphite are considered indicators of pressure-induced phase transitions.15,29 As the pressure increases, the G peak shifts to higher frequencies owing to phonon hardening caused by stress. The significant broadening of the G peak reflects an increase in internal disorder during the phase transition.29,30 The combination of Raman spectral changes and optical reflectivity analysis provides important evidence for studying the phase transformation process of graphite under pressure.

    As shown in Figs. 2(b) and 2(c), by superimposing these Raman imaging maps onto optical images acquired at matching pressures, the phase transition process of the SCG under pressure is more intuitively presented. At 1.5 GPa, there is no noticeable change in the optical image of the SCG, but a wave-like structure with fluctuation of 3 cm−1 appears in the 2D Raman imaging of FrequencyG, revealing the onset of buckling in SCG at this pressure. The observed pressure-induced deformation mode is consistent with previous theoretical modeling of the mechanism of graphite-to-diamond phase transformation, which suggests that atomic graphite layers deform through wavelike buckling at the nanoscale to transform into diamond, although direct experimental evidence for this model has never been observed before.12,18 However, according to our experimental observation, from a macroscopic viewpoint, this wave-like pattern can only be clearly observed in the lower-pressure range, and at higher pressures, it may be masked by the larger frequency distribution within the specimen caused by the pressure. As the pressure reaches 19.3 GPa, 2D Raman imaging maps of FrequencyG and FWHMG reveal some distinct triangular regions [marked by dashed yellow lines in Figs. 2(b) and 2(c)] with lower FrequencyG and broader FWHMG compared with other regions. This indicates an increase in lattice disorder in these regions, suggestive of a localized phase transition.

    Through statistical analysis of FrequencyG and FWHMG of SCG under different pressures, as shown in Figs. 3(a) and 3(b), we elucidate the average behavior and trend of SCG with pressure. Below 19.3 GPa, surface folding of SCG gradually increases with rising pressure, resulting in the distributions of FrequencyG and FWHMG gradually widening, but the overall range still remains relative narrow. At 19.3 GPa, the distribution of FWHMG begins to deviate from a Gaussian pattern, with a considerable proportion widening into the range of 30–60 cm−1, corresponding to the triangular regions in Fig. 2(c). When the pressure reaches 21.3 GPa, both the FrequencyG and FWHMG distributions become notably dispersed, suggesting a significant increase in the disorder of the SCG. These results correspond with the 2D Raman imaging maps presented in Figs. 2(b) and 2(c), which depict a range of phase transition covering almost the entire sample. However, as shown in Figs. 2(b) and 2(c), there is a recognizable regularity in the distribution of phase transition regions.

    (a) and (b) Distribution diagrams of FrequencyG and FWHMG, respectively, under different pressures in Ar. The solid line represents the Gaussian fit to the statistical distribution. (c) Raman spectra of folding area (red solid lines) and planar area (blue solid lines) of SCG under different pressures. The upper right insert is an optical image of SCG at 19.3 GPa, and the red and blue dashed boxes are enlarged images of the Raman spectrum acquisition positions of the folding and planar areas, respectively.

    Figure 3.(a) and (b) Distribution diagrams of FrequencyG and FWHMG, respectively, under different pressures in Ar. The solid line represents the Gaussian fit to the statistical distribution. (c) Raman spectra of folding area (red solid lines) and planar area (blue solid lines) of SCG under different pressures. The upper right insert is an optical image of SCG at 19.3 GPa, and the red and blue dashed boxes are enlarged images of the Raman spectrum acquisition positions of the folding and planar areas, respectively.

    The surface morphology of SCG under pressure can be roughly divided into two parts: a planar area that basically retains the pristine graphite morphology, and a striped area formed by deformation. Figure 3(c) presents Raman spectra of the striped area (red solid lines) compared with those of the planar area (blue solid lines) under different pressures. It can be clearly seen that when the pressure exceeds 19.3 GPa, the G peak in the planar region exhibits a significant broadening and flattening trend. Conversely, the striped region maintains relatively sharper peaks under the same pressure. It is not until the pressure reaches 28.3 GPa that the characteristic of the G peak in the striped region begin to align with that in the planar region.

    Evidently, the phase transition pressure in striped regions lags behind that in the planar regions. This asynchronous phase transformation behavior reveals the substantial impediment that deformed structures pose to the continuity of graphite’s phase transition. It can result in inconsistencies in the post-graphite phase structures across different regions of a sample under the same pressure. Even after subsequent heating and conversion to diamond, these initial disparities can lead to a diverse microstructure in diamond and even to diversity of carbon phases. These are recurrent problems in HPHT diamond synthesis, being exemplified by the coexistence of lamellar and granular crystal structures, as well as the frequent detection of hexagonal diamond or other carbon phases in cubic diamonds.20,31 On the basis of these results, synthesizing pure cubic diamond may necessitate the application of more extreme pressures during the cold-compressed stage. Moreover, the hindrance of graphite’s phase transitions by deformation structures highlights the significance of the boundaries between planar and striped regions, which might reveal crucial information about the conversion path from graphite to diamond.

    C. Phase transition of SCG in He and cBN

    To systematically investigate the response mechanisms of SCG under different pressure environments, He and cBN were used as PTM materials for comparison. He is known for its low molecular weight and viscosity, and its high solidification pressure of up to 11.5 GPa, and it stands out as the best PTM material currently known for providing an ideal hydrostatic pressure environment.32 By contrast, cBN, as the second hardest material, could provide an extremely non-hydrostatic environment with a high stress gradient distribution.33

    Figure 4 shows that in contrast to the experiments with Ar as the PTM, where striped structures of the SCG first appear at 6.5 GPa and continue to grow with increasing pressure, in both He and cBN environments, the SCG exhibits minimal change until reaching the transition pressures of 18.8 GPa in He and 16.1 GPa in cBN. However, once the phase transition pressure has been reached, the surface evolves rapidly over time.

    (a) Optical photographs of SCG surface morphology acquired at 2.1, 11.0, 18.8, and 25.8 GPa in He. (b) Optical photographs of changes in SCG morphology with time at the pressure of 18.8 GPa, taken at 1 s, 3 s, 6 min, 17 min, and 10 h. (d) Optical photographs of SCG surface morphology acquired at 4., 12.0, 16.1, and 20.5 GPa in cBN. (d) Optical photographs of changes in SCG morphology with time at the pressure of 16.1 GPa, taken at 1 s, 5 min, 17 min, 27 min, and 10 h.

    Figure 4.(a) Optical photographs of SCG surface morphology acquired at 2.1, 11.0, 18.8, and 25.8 GPa in He. (b) Optical photographs of changes in SCG morphology with time at the pressure of 18.8 GPa, taken at 1 s, 3 s, 6 min, 17 min, and 10 h. (d) Optical photographs of SCG surface morphology acquired at 4., 12.0, 16.1, and 20.5 GPa in cBN. (d) Optical photographs of changes in SCG morphology with time at the pressure of 16.1 GPa, taken at 1 s, 5 min, 17 min, 27 min, and 10 h.

    In He, when the pressure reaches 18.8 GPa [Fig. 4(b)], lots of striped structures abruptly begin to appear on the SCG surface, and this process intensifies over time until it stabilizes within 20 min. After 10 h, the sample darkens without significant change in striped structures. The stripe patterns exhibit an interlacing structure characterized by angles of 60° and 120°, which is similar to those observed for SCG in Ar. The observed denser and more uniform stripe patterns can be attributed to the superior hydrostatic pressure conditions provided by He. As is evident from Figs. 5(c) and 5(d), the FrequencyG of the whole sample increases uniformly with rising pressure, while the FWHMG remains largely unchanged, indicating that the stress applied across the entire sample is exceedingly uniform. Notably, Fig. 4(b) shows a significant color change on the SCG surface from 17 min to 10 h, indicating that, similarly to SCG in Ar, SCG in He undergoes deformation before experiencing a phase transition as well.

    (a) and (b) 2D FrequencyG and 2D FWHMG distribution imaging maps of SCG surface acquired at 2.1, 11.0, 18.8, and 25.8 GPa in He. (c) and (d) Statistical diagrams of FrequencyG and FWHMG distributions, respectively, under different pressures in He.

    Figure 5.(a) and (b) 2D FrequencyG and 2D FWHMG distribution imaging maps of SCG surface acquired at 2.1, 11.0, 18.8, and 25.8 GPa in He. (c) and (d) Statistical diagrams of FrequencyG and FWHMG distributions, respectively, under different pressures in He.

    For the experiments employing cBN as the PTM, the situation is completely different. Figure 4(c) illustrates that at 12 GPa, some local dark spots begin to appear along the edges of the SCG. The corresponding broadened FWHMG depicted in Fig. 6(b) signifies that these black areas are the initiation sites of local phase transitions. On a further increase in pressure to 16.1 GPa, the phase transition region rapidly develops from its boundary and diffuses toward the center over time until it stabilizes. However, in contrast to the hydrostatic environments, in cBN, the wavelike and striped structures are not observed in SCG before the phase transition, indicating that the phase transition in the cBN environment is more localized. This can be attributed to the uneven distribution of high stress provided by cBN, as evidenced by Figs. 6(c) and 6(d), where the distributions of FrequencyG and FWHMG for SCG in cBN are wider than those in He, indicating a more complex and spatially variable stress distribution on the SCG surface in cBN.

    (a) and (b) 2D FrequencyG and 2D FWHMG distribution imaging maps of SCG surface acquired at 4.2, 12.0, 16.1, and 20.5 GPa in cBN. (c) and (d) Statistical diagrams of FrequencyG and FWHMG distributions, respectively, under different pressures in cBN.

    Figure 6.(a) and (b) 2D FrequencyG and 2D FWHMG distribution imaging maps of SCG surface acquired at 4.2, 12.0, 16.1, and 20.5 GPa in cBN. (c) and (d) Statistical diagrams of FrequencyG and FWHMG distributions, respectively, under different pressures in cBN.

    D. Morphological analysis of recovered samples

    It is well known that the pressure-induced phase transition of graphite is reversible at room temperature. Surface analysis of recovered specimens should provide important clues about the phase transition process. Figure 7 presents SEM and 2D/3D AFM images of pristine SCG and recovered samples from He and Ar environments. As shown in Figs. 7(b) and 7(c), even after being subjected to pressures approaching 30 GPa, the recovered SCG samples maintain their structural integrity, and the striped structures remain intact upon returning to ambient pressure. These striped patterns reflect the response of SCG to different pressure environments. The AFM results for these samples shown in Figs. 7(d)7(i) provide more details of this response process.

    (a)–(c) SEM images of pristine SCG, SCG recovered from Ar (28.3 GPa), and SCG recovered from He (25.8 GPa), respectively. The yellow dotted boxes and arrows show the AFM scanning areas and directions. (d)–(f) 2D AFM images corresponding to the positions indicated by the yellow dotted boxes in (a)–(c), respectively. The bottom left inserts how the size distributions of striped segmented regions. (g)–(i) 3D AFM images corresponding to the positions indicated by the yellow dotted boxes in (a)–(c), respectively.

    Figure 7.(a)–(c) SEM images of pristine SCG, SCG recovered from Ar (28.3 GPa), and SCG recovered from He (25.8 GPa), respectively. The yellow dotted boxes and arrows show the AFM scanning areas and directions. (d)–(f) 2D AFM images corresponding to the positions indicated by the yellow dotted boxes in (a)–(c), respectively. The bottom left inserts how the size distributions of striped segmented regions. (g)–(i) 3D AFM images corresponding to the positions indicated by the yellow dotted boxes in (a)–(c), respectively.

    Figures 7(d) and 7(g) show that the pristine SCG exhibits the expected flat surface, except for some small amounts of dust on it. In the case of the samples recovered from Ar and He, the 2D and 3D images reveal that all the striped structures on the SCG surface are elevated above the pristine basal plane. This means that the striped structures appearing on the surface before the phase transition are due to uplift of the surface. The atoms in the upwelling region are subjected to a stretching force, resulting in a weakening of the external high pressure. This observation explains the phenomenon observed in Fig. 2(b), where the striped regions exhibit lower Raman shifts than the other regions. This could also be one of the reasons for the elevation in phase transition pressure in these regions. Furthermore, it is noteworthy that in Fig. 7(h), many small and sharp bumps appear at each intersection of the different uplift lines, suggesting that the intersection areas are the weakest points for releasing internal strain during the fragmentation process. These protruding structures, which arise from the deformation of graphite under pressure, may play a pivotal role in the bonding process between different grains during the sintering of polycrystalline diamond.

    The surface morphology of the quenched sample from He is characterized by the presence of uniformly distributed vertical flakes ranging from 2 to 9 μm in size [Figs. 7(f) and 7(i)], suggesting a faster fragmentation process of SCG in He. This is in stark contrast to the SCG surface recovered from Ar, which is characterized by a predominant pattern of triangular cracks with a size of around 2 μm, indicating a more complete fragmentation. It is evident that different pressure environments have a significant impact on the deformation patterns and rate of graphite under high pressure, which in turn affects its phase transition process. These findings are of great significance for the design and synthesis of diamond or other allotropes of carbon.

    IV. CONCLUSIONS

    We have discussed and compared the phase transition behaviors of SCG under different PTM materials by a series of characterization techniques. Our findings indicate that graphite under a hydrostatic environment initially undergoes regular deformation before the phase transformation. We have provided the corresponding deformation patterns and analyzed their impact on the phase transformation process. These macroscopic-level results give new insights into the phase transition process of graphite. They not only serve as a reference for subsequent microscopic structural studies, enabling more precise targeting of the pivotal features of the phase change process of graphite, but also provide guidance for the design and synthesis of new materials by manipulating the phase transition pathways of graphite. We hope that our results can be extended to explore the phase transition mechanisms of other layered materials, and further promote the development of phase transition theory in advanced materials science and the innovative design of new materials under HTHP.

    ACKNOWLEDGMENTS

    Acknowledgment. We are grateful to Dr. Cheng Ji and Dr. Xinwei Li for their advice on the content of this article, and to Dr. Wanghua Wu for assistance with 2D Raman imaging data processing. The authors acknowledge funding support from the National Science Fund for Distinguished Young Scholars (Grant No. T2225027), the NSAF (Grant No. U1930401), the National Key R&D Program of China (MOST) (Grant No. 2023YFA1406500), and the National Natural Science Foundation of China (NSFC) (Grant No. 61674045).

    Tools

    Get Citation

    Copy Citation Text

    Xiaoshuang Yin, Songyang Li, Lijuan Wang, Peiyuan Liu, Zhihai Cheng, Huiyang Gou, Liuxiang Yang. Macroscopic perspective on phase transition behavior of natural single-crystal graphite under different pressure environments[J]. Matter and Radiation at Extremes, 2025, 10(1): 017801

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Aug. 22, 2024

    Accepted: Sep. 25, 2024

    Published Online: Feb. 21, 2025

    The Author Email:

    DOI:10.1063/5.0234582

    Topics