1. Introduction
Large-volume presses (LVPs) have been widely applied in various fields of material science, chemistry, physics, and Earth science for decades owing to its use of large sample volumes under well-controlled high-pressure and high-temperature (HPHT) conditions. Remarkably, single crystals of diamond have been synthesized for industrial applications under pressures above 5 GPa at high simultaneous temperatures in belt, tetrahedral, and cubic LVPs
[1],
[2],
[3]. However, the pressure generation in these devices is commonly limited to below 10 GPa, especially at high temperatures. Kawai-type LVPs can generate HPHT conditions above 20 GPa by introducing a multistage compression system
[4],
[5]. This apparatus induces six split-sphere hardened steel anvils to compress eight truncated tungsten carbide (WC) cubes that converge on a solid octahedral pressure medium, referred to as the Kawai cell
[6]. Such LVPs have been successfully used to synthesize large-sized (ranging from cubic millimeter to cubic centimeter) transparent nanopolycrystalline diamond, cubic boron nitride, and garnet at pressures of 10–25 GPa and various temperatures
[7],
[8],
[9],
[10],
[11],
[12], and to allow the
in situ characterization of samples, including electrical conductivity, thermal conductivity, and sound velocity
[13],
[14],
[15],
[16].
Expansion of the pressure range at high temperatures is considered a technical advancement in LVPs. High pressure is generated by concentrating the applied load on small truncated edge length (TEL) of the second-stage anvils in LVPs
[17]. Thus, the mechanical properties and design of the first-stage and second-stage anvils, cell assemblies, and gaskets play critical roles in the pressure generation capacity at high temperatures. The achievable pressures of Kawai-type LVPs are limited to approximately 25 GPa at high temperatures, owing to technical difficulties such as limitations of hardness of anvils and low pressure efficiencies of assemblies. Ultrahigh pressures (> 25 GPa) at high temperatures have been generated in customizable Osugi (DIA)-type LVPs by using a Kawai-type module and precisely aligning the first-stage anvils via the implementation of hard WC anvils with a small TEL and well-designed cell assemblies
[18]. An ultrahigh pressure of 44 GPa at 2000 K was achieved by adapting hard TF05 WC anvils with tapered surfaces to yield an ∼1 mm
3 sample volume
[19]. A new record pressure of 52 GPa at 2000 K was recently reported for harder TJS01 WC anvils (Table S1 in Appendix A)
[20],
[21].
The Walker-type LVP was developed as a simplification of the original geometry of the Kawai-type LVP, and it is driven by a uniaxial ram that compresses the Kawai cell in the <111> direction
[22],
[23],
[24]. This approach required only
of the applied load in an Osugi (DIA)-type LVP to produce the same force on a Kawai cell. Thereby, the Walker-type LVP has significantly contributed to the wide distribution of high-pressure technology worldwide owing to its user-friendly and cost-effective setup and its remarkable capacity for self-organization upon compression
[2],
[22],
[23]. Despite the great success of commercial Walker-type LVPs, further technical developments are needed to reach pressures above 25 GPa at high temperatures and enable further exploration of novel materials and Earth’s interior. Very recently, our group generated pressures of up to 36 GPa at 1700 K by adapting hard ZK01F WC anvils in a Walker-type LVP to yield a sample volume of ∼2 mm
3 [25]. However, pressure generation above 36 GPa at high temperatures remains a challenge owing to blowouts and unstable heating. Further development of practical techniques for routine experiments is required.
In this study, we adopted newly developed ZK01F WC anvils (hereafter referred to as new-ZK01F anvils) with the same hardness as the former type used in our recent study
[23]; however, the new type possesses a superior distribution of sintered particles to improve its performance in HPHT experiments. The relationship between the mechanical properties and pressure generation efficiency of the new-ZK01F and two other types of WC anvils was also investigated. Hard materials such as alumina and ZrO
2 have been introduced into the cell assembly to further expand the pressure range at high temperatures and overcome the previous unstable heating problems. The Walker-type LVP was evaluated under routine experimental conditions at approximately 30 GPa and ultrahigh pressure generation above 40 GPa. These new techniques are shown to successfully generate pressures above 40 GPa at simultaneous temperatures of 1900–2100 K.
2. Methods
2.1. Walker-type LVP configurations
The experiments were performed on a JLUHC Walker-type LVP by using a maximum load of 10 MN, which was achieved at a hydraulic oil pressure of 690 bar (1 bar = 10
5 Pa; Fig. S1(a) in Appendix A). The press settings were identical to those stated in Ref.
[25]. The apparatus configuration and some minor adjustments are described in detail below.
The six removable, split, first-stage anvils were laterally supported by a tool steel cylindrical containment ring and vertically compressed by a uniaxial force (Figs. S1(b) and (c) in Appendix A). The implementation of a highly symmetrical initial position of the anvils enabled full utilization of the self-organization capacity of the Walker-type guide block system and increased the success rate of ultrahigh pressure generation. Two layers of polyester sheets were pasted on the internal surface of the containment ring and outer surfaces of the first-stage anvils to provide electrical insulation and enhance the lateral alignment precision, which could also contribute to lubricate the anvils in the vertical direction. The total thickness of the polyester sheets was adjusted from 0.200 to 0.176 mm to facilitate self-organization. The optional application of a hexagonal boron nitride (h-BN) lubricant spray between polyester sheets was found to improve the lubrication in this study. The six first-stage anvils were designed to compress a cubic Kawai cell composed of eight second-stage WC anvils with an edge length of 25.4 mm. The second-stage anvils were protected and fixed by six square epoxy glass sheets with an edge length of 48.0 mm and thickness of 0.4 mm, which were pasted onto each face of the cubic chamber between the first- and second-stage anvils. Eight truncated second-stage WC anvils converged on a solid octahedral pressure-transmitting medium to generate quasi-hydraulic pressure, within which the sample was accommodated. High temperatures were obtained by applying electrical heating with a maximum heating power of 6 kW, and a spiral cooling module was added to the apparatus to prevent anvil overheating.
2.2. Second-stage WC anvils
In our recent study, we adopted three types of WC anvils for high-pressure experiments and demonstrated that the hardest ZK01F anvils achieved the highest pressure efficiency
[25]; however, their survival rate was limited to pressures above 36 GPa. In this study, we employed new-ZK01F anvils (Heyuan Zhengxin Hard-metal Carbide Pte Limited Company, China), which have the same hardness as the previous version but a better particle distribution (
Table 1)
[25],
[26]. The new-ZK01F anvils were compared with two other types of commercial hard carbide anvils with similar mechanical properties, namely, Fujilloy F08 (Fuji Die Co., Ltd., Japan) and Hawedia (Hawedia, Germany). The microstructures and compositions of the WC anvils (
Fig. 1) were characterized by using field-emission scanning electron microscopy (FE-SEM; Magellan 400, FEI Company, USA) combined with an energy-dispersive spectroscopy (EDS) module (X-Max
N 150, Oxford Instruments, UK). All anvils were sintered by using a cobalt (Co) binder, and the Co contents in F08, new-ZK01F, and Hawedia were estimated to be approximately 10.0, 4.0, and 2.4 weight percent (wt%), respectively. The average particle size of the new-ZK01F and F08 anvils was approximately 0.35 μm, whereas that of the Hawedia anvils was slightly larger (
Fig. 1). The F08 anvils exhibited the best homogeneity in the distribution of both grain size and composition among these anvils, yielding a higher survival rate in our experiments (
Table 1). A lower Co binder content and smaller particle size generally afforded higher hardness when the other parameters were the same. The Hawedia anvils had the highest Rockwell hardness (95.2 HRA) among these anvils, whereas new-ZK01F (93.5 HRA) possessed nearly the same hardness as F08 (93.4 HRA)
[25],
[26].
Another critical factor influencing pressure efficiency is the anvil yield near the truncated edge, especially under very high pressures
[21],
[27]. We further altered the anvil geometry by applying mild, one-degree negative tapering to three surfaces around the 1.5 mm truncated edge of each new-ZK01F anvil. Tapering can reduce the yielding effect, thereby increasing the pressure. This tapering technology was pioneered by Ito and was recently adopted for ultrahigh pressure generation
[19],
[28]. The three faces around the truncated edges of the anvil were tapered, and the anvil was shaped as a wedge, as illustrated in
Fig. 2(a) [21], allowing the press load to be easily applied to the pressure medium. Pyrophyllite-sealing gaskets were pasted around the truncated edges to minimize the extrusion of cell materials, which reduced the stress concentration in the anvils. Balsa wood spacers were positioned between the anvils to maintain the initial alignment, and a heat-resistant tape was pasted behind the gaskets for electrical insulation (
Fig. 2(b)). Generally, the pressure generation in LVPs is also dependent on the TEL of the WC anvils. In principle, the efficiency of pressure generation increases with a decreasing TEL
[29]. In this study, WC anvils with TELs of 8 and 5 mm were applied for pressures below 23 GPa, whereas TELs of 4, 3, and 1.5 mm were used for pressures above 23 GPa.
2.3. Cell assemblies
The cell assembly design determines how the samples are accommodated within the octahedral pressure medium and is particularly crucial for the simultaneous generation of HPHT conditions. An improper design may frequently induce blowouts and heating failures, as has been observed in previous studies
[6]. We adopted six different cell assemblies to test both the heating stability and pressure efficiency, that is, 14/8, 10/5, 10/4, 8/3, 7/3, and 6/1.5 cell assemblies, where the assembly name was assigned according to the classical octahedral edge length (OEL)/TEL method.
Parts of the 14/8, 10/5, 10/4, and 8/3 cell assemblies were supplied by COMPRES
[30], and a MgAl
2O
4 ceramic octahedron was applied as the pressure-transmitting medium. A thin graphite tube was applied as a resistance heater in the 14/8 cell assembly, whereas rhenium foils folded into straight cylindrical metal furnaces were adopted in the other COMPRES cell assemblies. LaCrO
3 and ZrO
2 were processed into different sizes and shapes to encapsulate the heater for heating insulation. The temperature was measured by using a W
97Re
3–W
75Re
25 (Type D) or W
95Re
5–W
74Re
26 (Type C) thermocouple positioned close to the center of the heater. The thermocouples were protected by copper wires and electrically insulated by using mullite tubes and four-hole alumina plugs. The three-dimensional (3D) axial cross-sections of the 14/8, 10/5(4), and 8/3 assemblies are shown in Fig. S2 in Appendix A.
Several alterations were made to improve the experimental performance. Rhenium foil heaters were replaced with cheaper tantalum foils for the experiments using 10/5 cell assemblies. The thermocouples were connected in different ways and at different positions to determine the most effective and accurate measurements. Additionally, pyrophyllite gaskets are typically designed to perfectly fit the gap between anvils that results from the combination of a relatively large octahedron and relatively small truncated edge size. In this study, the pressure efficiency of the 10/4 cell assemblies was found to be significantly higher than that obtained in our earlier study upon reducing the gasket thickness from 2.83 to 2.38 mm
[25]. Parameter details for the gaskets are listed in Table S2 in Appendix A.
Earlier studies have shown that smaller assemblies (e.g., 7/3 and 6/1.5) are more promising for reaching pressures above 30 GPa because the applied force is concentrated in a smaller area
[25],
[27]. Therefore, further optimization of these assemblies was the focus of this study. The original 6/1.5 and 7/3 designs had the same structure, only differing in size, and both have been proven effective for high-pressure research (see the sketch map in
Fig. 3(a) and the 3D version in Fig. S3(a) in Appendix A)
[25]. Semi-sintered Cr
2O
3-doped (5 wt%) MgO octahedra (Vickers hardness: 400 MPa) supplied by Mino Ceramic Co., Ltd. (Japan) were applied as the pressure medium. The octahedra were bore-drilled in-house, ultrasonicated in acetone, and refired at 1273 K to eliminate possible impurities. The following new alterations were applied to each assembly to optimize the working performance.
The 7/3 assemblies are generally used for routine experiments at pressures below 30 GPa. The original 7/3 assembly design was able to recover large samples at high temperatures under a diameter of ∼1.5 mm and height of ∼2.0 mm
[25]. However, the heater was in direct contact with the sample and thermocouples, which increased the likelihood of eutectic melting or chemical reactions. We replaced the double-folded enclosed box furnace with a straight rhenium tube furnace (
Fig. 3(b); Fig. S3(b) in Appendix A). The thickness of the metal heater was 50 μm. The samples were not directly accommodated in the heaters, but were further protected by MgO or dense alumina tubes. The thermocouples were also protected from the rhenium heaters by using alumina tubes, thereby minimizing the risk of unstable heating. Two samples, each with a height of 1.0–1.3 mm and diameter of 1.0–1.4 mm, was recovered from a single experiment.
When the tapered anvils were used, the 6/1.5 assemblies were able to achieve higher pressures without sacrificing a relatively large sample volume (∼2 mm
3). Mo pillars with a diameter of 0.5 mm were adopted as electrodes, as shown in
Figs. 3(a) and S3(a). Hard ZrO
2 ceramic (Vickers hardness: 600 MPa) was used to envelop the heater. Rhenium foils with 25 μm thickness were doubly folded into an enclosed cylinder heater to encapsulate the sample. Thermocouples were bent at the end between the thermal insulator and rhenium heater, establishing direct electrical connection to the metal heater. The metal heater was only ∼50 μm thick, and severe deformation is known to be able to occur under HPHT conditions; however, a simple adjustment to the assembly was applied to rectify this problem. Specifically, because dense alumina has a higher bulk modulus (∼240 GPa) than MgO (∼160 GPa)
[31], a thin layer of alumina tubing was applied to prevent brittle deformation of the metal heaters at temperatures above 1500 K (
Figs. 3(a) and S3(a)). The sample volume was approximately 2.2 or 1.0 mm
3 with or without the dense alumina tube, respectively. To prevent chemical reactions between the alumina and sample, in some experiments, the sample was encapsulated in an additional layer of platinum or rhenium foil.
Ultrahigh pressure generation requires high symmetric accuracy. The uniform size of the pyrophyllite gaskets is also important for the proper alignment of a cubic Kawai cell for HPHT experiments. The gaskets used for the 6/1.5 assembly were sloped on the top (
Fig. 2(b)) and processed in-house to have a more consistent size than that achieved in our previous study; particularly, the new milling program reduced the thickness errors from ±0.05 to ±0.02 mm.
2.4. Pressure calibration methods
In this study, the pressure at room temperature was calibrated for the above-described cell assemblies and WC anvils based on the electrical resistance discontinuity associated with the phase transitions of ZnTe (I–II, 6.6 GPa; II–III, 8.9 GPa; III–VI, 12.9 GPa), GaAs (semiconductor–metal transition, 18.3 GPa), and GaP (semiconductor–metal transition, 23 GPa) upon compression, as illustrated in
Fig. 4 [25],
[32],
[33],
[34],
[35],
[36]. A two-wire method was used to measure electrical resistance. Boreholes were drilled into the octahedra at a diameter of 1.0 mm, and the samples were inserted between copper pillar electrodes with a height of 0.8–1.0 mm (
Fig. 4(b), inset).
As mentioned earlier, the new-ZK01F WC anvils possess the same hardness as the ZK01F used in our recent study
[25]. In principle, the efficiency of pressure generation using these new WC anvils was the same as that of the former anvils for the same assembly. Therefore, for the 6/1.5 assembly, we adopted the room-temperature pressure calibration results from the former ZK01F. Experiments to measure the electrical resistance jump of Zr (ω–β, 34.5 GPa) were performed by applying the four-wire method to the former ZK01F anvils
[37]. Each octahedron was cut into two equivalent parts by using a diamond saw, a thin zirconium foil (thickness ∼100 μm) was fixed at the middle of the exposed surface, and four copper wires led the electrical signal down to the center of the four faces of the octahedron (Fig. S4 in Appendix A). Further details of the four-wire method have been provided in our previous report
[25]. In this study, we focused on pressure calibration at high temperatures by using the new-ZK01F WC anvils.
The lack of an appropriate room-temperature calibrant above 23 GPa tends to negatively impact calibration at higher pressures. Fortunately, the Al
2O
3 content in the MgSiO
3 perovskite (hereafter referred to as bridgmanite) is dependent on the pressure and temperature conditions. Generally, the highest attained pressure at high temperatures can be determined by the Al
2O
3 content in bridgmanite in the quenched samples based on the phase diagram of the MgSiO
3–Al
2O
3 system
[38],
[39]. In this study, Mg
3Al
2Si
3O
12 glass or a MgAl
2SiO
6 oxide mixture was adopted as the starting material for the HPHT experiments. The texture and chemical compositions of the recovered samples were analyzed by using an FE-SEM system and electron probe microanalyzer (EPMA; JXA-8230, JEOL Ltd., Japan) operating under an acceleration voltage of 15 kV and a beam current of 2 nA with forsterite as the Mg and Si standards, respectively, and pyrope as the Al standard.
3. Pressure calibration at room temperature
To determine the mechanism by which a certain assembly design and anvil type influence the pressure generation, we evaluated some representative phase transition points as a function of the applied hydraulic oil pressure (
Fig. 5). Each pressure curve was fitted and extrapolated based on the results of at least 3–4 room-temperature phase transitions. Under the condition of the same 8/3 cell assembly (
Fig. 5(a)), the pressure efficiency of the new-ZK01F anvils was comparable to that of the F08 anvils because their levels of hardness were very similar. These two anvil types generated pressures above 25 GPa at oil pressures above 500 bar by the extrapolation of the load–pressure relationship. For the 7/3 cell assembly, the new-ZK01F anvils were found to have a lower pressure generation efficiency than the harder Hawedia anvils and required an oil pressure approximately 56 bar higher than that of the latter anvils to reach the GaP phase transition pressure (23 GPa) (
Fig. 5(b)). The highest pressure limit generally had a positive relationship with the OEL/TEL value. However, the curves show that the 8/3 cell assembly tended to have a slightly lower pressure efficiency in the high-pressure region (> 20 GPa) than the 7/3 assembly, regardless of the anvil type. This result is partially due to the thicker and wider gaskets applied in the 8/3 assembly experiments (Table S2)
[27],
[28]. To quantitatively determine the influence of gaskets in a Walker-type LVP, we improved the 10/4 assembly, as described above. Notably, the modified 10/4 assembly was reproducible without blowouts and reached 23 GPa at 470 bar (∼7.0 MN), which is considerably lower than that required in our previous study (∼9.0 MN)
[25]. The calibration results for the large-cell assemblies for routine experiments under pressures below 23 GPa are presented in Fig. S5 in Appendix A, along with an overall comparison among all tested assemblies.
Anvil yielding is a primary factor limiting the pressure efficiency. The average plastic deformation of the WC anvils was measured at the 3 mm truncated edges after decompression under 23 GPa. The Hawedia anvils exhibited slightly milder deformation near the truncated edge (
Fig. 6), whereas the extents of deformation of other two anvil types were negligible. This milder plastic deformation indicates that the anvils are less likely to yield under pressure, which, in turn, provides a higher pressure efficiency. Our calibration results were consistent with the mechanical properties of anvils; particularly, Hawedia anvils were slightly harder and less likely to yield under high pressure, resulting in higher pressure efficiency at 23 GPa, whereas the new-ZK01F and F08 anvils exhibited only negligible differences in pressure generation efficiency (
Figs. 5(a) and
(b)).
Furthermore, tapered anvils exhibited substantially higher pressure efficiency than flat anvils (
Fig. 5(d))
[25]. Additionally, in this study, pressures higher than 30 GPa at room temperature were achieved only in the 6/1.5 assemblies with tapered anvils. Based on previous reports, compression of the octahedron by a small truncated edge (< 1.5 mm) and hard carbide anvils is necessary to realize such pressure conditions
[21],
[27].
4. Pressure calibration at high temperatures
The heating efficiencies of the rhenium heaters in the various cell assemblies are summarized in
Fig. 7. As shown in
Fig. 7(a), the heating efficiency generally decreased with increasing pressure for the rhenium heaters in the same 10/5 assemblies. Temperatures above 2000 K were achieved for different pressures based on the relationship between the temperature and heating power.
Fig. 7(b) shows the heating performance of all tested cell assemblies that generated high temperatures above 2000 K at pressures above 18 GPa. The occurrence of a nonlinear increase in the power–temperature relationship in the 10/4 assembly is attributable to the heating power exhibiting invariance as the temperature continued to increase owing to the fast heating rate. The same “nonlinear” temperature increase phenomenon was also observed in the 8/3 and 7/3 assemblies, as shown in
Fig. 7(b).
The 10/5-assembly heating behaviors of tantalum and rhenium were compared at an identical cell pressure of 20 GPa (Fig. S2(b)). Laser cutting was used to shaped the tantalum heater to the same size as the commercial COMPRES rhenium heater; therefore, the original heater could be replaced without modifying the cell setup. As shown in Fig. S6 in Appendix A, the thermocouples failed at ∼2500 K, but the heating power remained stable, implying that both metal heaters could potentially reach higher temperatures in response to an increase in heating power.
The Al
2O
3 content in the bridgmanite in the recovered samples was examined by using both EPMA and FE-SEM coupled with EDS to calibrate the pressure at high temperatures
[38],
[39]. The uncertainty in the measured Al
2O
3 content for these two analytical methods was less than 2% in our calibrations. Here, we adopted the composition results from EPMA for the three recovered samples by using the 8/3 and 6/1.5 cell assemblies. The results for the 7/3 cell assembly were derived by using FE-SEM with EDS. The backscattered electron images of the quenched samples are shown in
Fig. 8. Note that bridgmanite grains coexisted with corundum in the 8/3 and 7/3 cell assemblies recovered at 650 bar and 2000 K, as shown in
Figs. 8(a) and
(b).
Fig. 8(c) shows the same phase assemblage as in the 6/1.5 cell assembly recovered at 450 bar and ∼2100 K. We observed a nearly single phase of bridgmanite and did not find corundum in this sample when it was quenched at 600 bar and ∼1950 K (
Fig. 8(d)). The absence of corundum grains in this sample may be attributable to them existing as tiny grains that are beyond the resolution of FE-SEM; alternatively, these tiny interstitial grains may have been polished out.
For the 8/3 and 7/3 assemblies, the samples were pre-pressed to a tentative 650 bar (∼9.7 MN) and then stably heated to 2000 K for a duration of 1 h. Samples recovered from the recovered 8/3 assembly yielded a pressure of (25.9 ± 0.1) GPa, as determined from the (8.67 ± 0.08) wt% Al2O3 in bridgmanite (Table S3 in Appendix A). The Al2O3 content in bridgmanite in the recovered 7/3 cell assembly was (14.09 ± 1.21) wt%, which indicates that a pressure of (29.4 ± 1.7) GPa was reached at 2000 K.
Blowouts often occurred when applying the 6/1.5 cell assembly and former ZK01F WC anvils at pressures above 36 GPa, and this is attributable to the quality of the WC anvils and rhombohedral deformation of the Kawai cell. As shown in
Fig. 1 and Table S1, the size distribution of the sintered particles in the new-ZK01F anvils was more desirable due to a smaller standard deviation. The uniform size of the machined gaskets, hard Al
2O
3 tube, and ZrO
2 cylinder further improved the cell assembly. The initial alignments of the first-stage and second-stage anvils were enhanced by adjusting the thicknesses of the polyester sheets and cell assemblies. All of these technical improvements helped the assembly to compress above 400 bar with minimal blowouts. Thus, ultrahigh-pressure and high-temperature experiments could be successfully conducted by using the Walker-type LVP. The Al
2O
3 content in bridgmanite recovered at 450 bar (∼6.7 MN) and ∼2100 K was (22.64 ± 0.54) wt%, indicating that a pressure of (37.3 ± 0.9) GPa was achieved (Table S3). Experiments above 500 bar (∼7.5 MN) resulted in higher blowout rates, likely owing to the asymmetric load on the pressure cell. In the sample successfully recovered at 600 bar (∼9.0 MN) and 1950 K, we observed (20.95 ± 0.21) wt% Al
2O
3 to be doped into bridgmanite, indicating that an ultrahigh pressure of at least (40.4 ± 0.4) GPa was obtained (Tables S1 and S3).
Compared with the pressure generation in specially designed and adjusted Osugi (DIA)-type LVPs, which afforded precise alignment of the guide block system, it was difficult for Walker-type LVPs to achieve pressures higher than 35 GPa at high temperatures owing to the uniaxial compression-induced rhombohedral distortion of the Kawai cell. By adopting tapered new-ZK01F WC anvils, improved cell assemblies and gaskets, and adjusted anvil alignments, we increased the maximum pressure of Walker-type LVPs to approximately 40.4 GPa at temperatures of 1950–2100 K, which is significantly higher than that of other commercial LVPs; we also set a new record for Walker-type LVPs (
Fig. 9)
[17],
[19],
[20],
[21],
[25],
[40],
[41]. The ultrahigh pressures of 48–52 GPa at 2000 K observed for Osugi-type LVPs were achieved at a hydraulic ram pressure of 15 MN by using TF05 and TJS01 anvils
[19],
[20],
[21], exceeding the maximum press load of ∼10 MN in our Walker-type LVP. A detailed comparison between this and previous studies is presented in Table S1. In summary, we suggest that, according to the load–pressure relationship shown in
Fig. 5(d), pressures higher than 40 GPa can be obtained by a more general technology using our new-ZK01F anvils and optimized cell assemblies if a larger press load (> 10 MN) is applied in commercial Walker-type LVPs.
5. Applications
Ultrahigh pressure generation at high temperatures in Walker/Osugi (DIA)-type LVPs has a wide range of applications in materials science, chemistry, physics, and Earth science. In this section, we summarize some representative applications of the LVP technology based on earlier studies and our recent research.
Carbon has been extensively studied under high pressure owing to its versatile bonding structure and hundreds of possible allotropes
[42]. Pressures above 35 GPa have proven effective for tuning the hybridization and phase transitions of different carbon precursors
[43],
[44],
[45],
[46],
[47]. Compared to conventional crystalline carbon (e.g., diamond and graphite), amorphous carbon (AC) has increasingly attracted attention owing to its tunable properties and importance in various applications
[48],
[49]. Using the ultrahigh-pressure techniques developed for our Walker-type LVPs, we explored the phase diagram of C
60 fullerene under pressures ranging from 20 to 37 GPa and temperatures ranging from 723 to 1900 K (
Fig. 10(a))
[47]. A series of millimeter-sized bulk AC samples with different sp
3-hybridization contents were synthesized under these extreme conditions. A representative sample, AC-7, quenched at 37 GPa and 1273 K and shown in
Fig. 10(b) [47], is a composite material in the presence of nanocrystalline diamond (diameter ≈ 10 nm) in an AC matrix. Interestingly, sample AC-3, which was recovered at 27 GPa and 1273 K (a pressure close to the cage collapse boundary), is a previously unreported ultrahard bulk sample of nearly pure sp
3-hybridized AC. The bulk AC samples were sufficiently large for detailed characterization, further confirming them to be a group of ultrahard (Vickers hardness up to 102 GPa), thermally conductive (26 W·m
−1·K
−1), and band-gap-tunable (1.85–2.79 eV) materials (
Fig. 10(b))
[47].
In addition to pure carbon allotropes, advanced LVP technology has paved the way for many other applications. These include the study of superhard and superconducting B–C binary systems such as boron-doped diamond, BC
3, and BC
5, as well as complex compounds and composites such as Re
2(N
2)(N
2)
[50],
[51],
[52],
[53],
[54],
[55],
[56],
[57]. Our ultrahigh-pressure LVP technology was recently applied in a study of core–shell structured nanocrystals. Such composite materials have been found to possess outstanding properties, such as a high energy-conversion efficiency and high photoluminescence quantum yield. We also synthesized a bulk sample of B31-phase core–shell MnSe/MnS nanorods at 35 GPa by LVP to ensure that the quantity of samples are large enough for an accurate magnetic performance measurement. The as-synthesized sample exhibited antiferromagnetic behavior at a Néel temperature of 132 K. This Néel temperature is relatively high in Mn-based semiconductor nanomaterials, which have potential applications in information storage, spintronics, and sensing
[58],
[59],
[60].
Exploration of the composition and structure of Earth’s interior is a scientific matter of chief importance. We applied the proposed ultrahigh-pressure technology to study the solubility of water and oxygen vacancies in the lower mantle dominated by Al-bearing MgSiO
3 bridgmanite at up to 40 GPa at 2000 K. High-quality and large-sized Al-bearing MgSiO
3 bridgmanite single crystals were successfully synthesized under the top-lower-mantle conditions at a depth of 750 km (∼26 GPa, 1700–2000 K). These crystals do not contain water, even in the presence of oxygen vacancies (
Fig. 11)
[39],
[61]. Regarding the consideration of a potential host for water derived from oxygen vacancies, as suggested by earlier studies, the solubility of oxygen vacancies in bridgmanite decreased significantly to nearly zero as the depth increased to 1000 km (∼40 GPa, 2000 K) (
Fig. 11(a))
[39]. Therefore, we concluded that most of the pyrolytic lower mantle was nearly dry because of its dominance (80 volume percent) of a nearly dry form of bridgmanite in this region. The rapid decrease in oxygen vacancies in bridgmanite with depth can reduce element diffusivity
[39], which may lead to high viscosity, thus explaining the mid-lower-mantle slab stagnation observed at 660–1000 km
[62]. The ultrahigh-pressure LVP techniques presented here enable measurements of sound velocity and electrical and thermal conductivity of minerals recovered under deeper mantle conditions, which is expected to provide additional physical and chemical insights for a better understanding of Earth’s interior.
6. Conclusions
We simultaneously generated an ultrahigh pressure and a high temperature of ∼40 GPa and ∼2000 K, respectively, in a Walker-type LVP. The use of hard WC anvils with tapered surfaces, optimized cell assemblies composed of hard materials, and uniform gasket sizes, as well as the high-precision initial alignment of the Kawai cell, can significantly improve pressure generation performance at high temperatures. This general technical approach can be used to perform routine experiments under pressures of 35–40 GPa and at high temperatures. These ultrahigh-pressure techniques were applied in LVPs at high temperatures to synthesize and characterize millimeter-sized ultrahard AC and high-pressure phases of core–shell nanocrystals. This technique was also applied to investigate water storage in lower-mantle minerals to better understand Earth’s interior water cycle. The broader range of feasible pressures and temperatures in LVPs expands the possibilities for future applications in various fields of science and industry.
Acknowledgments
This work was financially supported by the National Key Research and Development Program of China (2022YFB3706600 and 2023YFA1406200), and the National Natural Science Foundation of China (42272041, 52302043, 12304015, 41902034, and 12011530063), the Jilin University High-level Innovation Team Foundation, China (2021TD-05), and the National Major Science Facility Synergetic Extreme Condition User Facility Achievement Transformation Platform Construction (2021FGWCXNLJSKJ01).
Compliance with ethics guidelines
Xuyuan Hou, Yuchen Shang, Luyao Chen, Bingtao Feng, Yuanlong Zhao, Xinyu Zhao, Kuo Hu, Qiang Tao, Pinwen Zhu, Zhihui Li, Ran Liu, Zhaodong Liu, Mingguang Yao, and Bingbing Liu declare that they have no conflict of interest or financial conflicts to disclose.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2023.03.023.
PDF
(2900KB)
