From salt water to bioceramics: Mimic nature through pressure-controlled hydration and crystallization

Abstract

Modern synthetic technology generally invokes high temperatures to control the hydration level of ceramics, but even the state-of-the-art technology can still only control the overall hydration content. Magically, natural organisms can produce bioceramics with tailorable hydration profiles and crystallization traits solely from amorphous precursors under physiological conditions. To mimic the biomineralization tactic, here, we report pressure-controlled hydration and crystallization in fabricated ceramics, solely from the amorphous precursors of purely inorganic gels (PIGs) synthesized from biocompatible aqueous solutions with most common ions in organisms (Ca²⁺, Mg²⁺, CO₃²⁻, and PO₄³⁻). Transparent ceramic tablets are directly produced by compressing the PIGs under mild pressure, while the pressure regulates the hydration characteristics and the subsequent crystallization behaviors of the synthesized ceramics. Among the various hydration species, the moderately bound and ordered water appears to be a key in regulating the crystallization rate. This nature-inspired study offers deeper insights into the magic behind biomineralization.

A biocompatible approach producing ceramics with tailorable hydration characteristics and crystallization behaviors is reported.

INTRODUCTION

Life builds stunning ceramic architectures, such as coccoliths, corals, and bones, with superior structural control, under mild conditions (1–6). Biogenic ceramics were known to arise from hydrated amorphous precursors (7), e.g., amorphous calcium carbonate (ACC) in invertebrates and amorphous calcium phosphate (ACP) in vertebrates (8). These precursors were observed to dehydrate before crystallization, indicating that the water species are important in preserving amorphousness (9–11). Even for minerals of similar compositions (e.g., ACC), its amorphous precursors of different organisms (e.g., seashells, sea stars, or corals) often feature not only distinct total water contents but also distinct proportions of the various water species (e.g., loosely, moderately, and tightly bound water) as revealed by their thermogravimetric responses (1). Hydration characteristics as such affect the properties of the bioceramic products (10, 12, 13), including not only crystallization behaviors but also density, mechanical performance, fracture behaviors, optical properties, chirality, electric and ionic conductivities, dielectric functions, and ferroelectric responses (14–16). For example, the crystallization or fusion behaviors of the ACC particles closely depend on the structural and mobile water contents (12, 17). Conventionally, thermal annealing (e.g., above 100°C in the air) (17, 18) can be used to control the water content, but not the relative amounts of the various water species (as the loosely bound ones will be the first to go), to help induce and preserve the amorphous state; organic solvents and vacuum were also applied in laboratory procedures (19). Life, on the other hand, is delicate and water-based, so the puzzle is: How does life modulate the hydration profile (the water content and the relative proportion of various water species) for controlling amorphous minerals and for biomineralization in its aqueous matrices?

Here, we report a facile and biocompatible construction of monolith ceramics: from aqueous salt solutions and by way of stable amorphous purely inorganic gels (PIGs), wherein the hydration characteristics of the ceramics can be easily modified by external pressure to affect the crystallization traits of the ceramic solid products (Fig. 1A). For background information, the phospholipid-based membranes of cells or vesicles are generally permeable to water, feature tailorable rigidity, and can withstand high stress up to the gigapascal level (e.g., mechanical compression or osmotic pressure) (20). Although the exact content may vary for different cell types, water is the most abundant component of cells, and so are the ionic species of Ca²⁺, Mg²⁺, K⁺, Na⁺, Cl⁻, CO₃²⁻, and PO₄³⁻ (20). Moreover, the ceramics studied in this work are also closely related to mineralogy. For example, besides being the most abundant biomineral, CaCO₃ is a major component of the lithosphere, comprising approximately 4% of the Earth’s crust (21). Therefore, the findings reported here may help to understand the hydration and crystallization behaviors of both natural minerals in geology (e.g., geopolymers) (22) and biogenic ceramics of life.

Fig. 1. Characterizations of the Mg-ACCP gel.

(A) Schematic diagram from salt water to bioceramics: stable amorphous gels, fusion under mild pressure, and control over hydration and crystallization. (B and C) TEM images, with the inset in (C) showing a selected-area electron diffraction pattern. (D) Particle size distribution. (E) Scanning TEM image and elemental mapping of Ca, Mg, P, and O.

RESULTS

In this study, the four aqueous solutions of CaCl₂, MgCl₂, Na₂CO₃, and K₂HPO₄ were mixed at room temperature to afford different products of variable amorphous stabilities (figs. S1 and S2 and table S1). If the four ions were mixed with equal concentration (e.g., 0.8 M) and volume, an ultra-stable PIG (denoted as “Mg-ACCP gel”) can remain to be amorphous for over 4 months under ambient conditions (figs. S3 to S5). The stable gelatinous state likely arises from the “high entropy” forged by the multiple ionic components and various water species. Kinetically, the complex mixture traps the gelatinous amorphous state through spatial hindrance, viscosity, or internal stress; thermodynamically, the high entropy lowers the free energy to help disfavor crystallization (23). In the same vein, additives (such as Mg²⁺ in ACC and ACP, PO₄³⁻ in ACC, and CO₃²⁻ in ACP) and/or volume restriction were found to stabilize the amorphous biominerals. We recently proposed a mechanism to explain why the volume restriction at the microscale suppresses crystallization (3, 24, 25). To enable the high entropy, it is crucial to involve many (e.g., >3) types of coprecipitating ions. For example, the following simpler mixtures, with only three coprecipitating ions, do not form as stable amorphous gels: Mg-ACC (by mixing Mg²⁺, Ca²⁺, and CO₃²⁻ ions), Mg-ACP (by mixing Mg²⁺, Ca²⁺, and PO₄³⁻ ions), and ACCP (by mixing Ca²⁺, CO₃²⁻, and PO₄³⁻ ions)—these products crystallized within 1 day, while that of Mg²⁺ and CO₃²⁻ stayed amorphous longer but also crystallized within 7 days (table S1 and fig. S2D) (7).

The ultra-stable Mg-ACCP gel appeared milky white to the naked eye (fig. S5) and featured an amorphous framework of mesoporous nanoparticles with uniform elemental distribution (Fig. 1, B to E, and fig. S6). The thermal gravimetric analysis (TGA) measurement showed water content of 83.2 wt % that can be removed below 400°C (fig. S7). The gel sample was resistant against thermal crystallization up to 660°C (with a small exothermal peak suggestive of crystallization at this temperature). X-ray diffraction (XRD) monitoring from 100° to 900°C (fig. S8) confirmed the high thermal stability with no diffraction peaks observed up to 600°C. The crystalline phase emerging at 700°C matched Ca₇Mg₂(PO₄)₆.

Another puzzle regards the assembly mechanism of bioceramics. Biogenic ceramics often appear nanogranular, indicating that they are likely assembled from smaller nanoparticle building blocks (26). But how do they coalesce into a monolith ceramic object in aqueous environments at mild temperatures? As a key clue, we found that the amorphous multi-ionic wet gels can serve as a stable precursor that directly transforms into monolithic bioceramic objects upon compression under ambient conditions (24). A more recent study by Tang and coworkers on dry compression of annealed pure ACC powders further revealed that the fusion behaviors of ACC powders can be controlled by their hydration content (which was adjusted by thermal annealing) and the pressure exerted, offering important insights into the biomineralization mechanism (17).

Adopting the simple procedure used for pressing salt pellets for infrared spectroscopy, the wet Mg-ACCP gel was directly compressed in a mold at room temperature. The Mg-ACCP tablets produced under different pressures displayed similar compositions (fig. S9). Consistent with the gradual enhancement of density, optical transparency, and mechanical properties (Fig. 2, F and G), the scanning electron microscope (SEM) characterization (Fig. 2, A to E) showed more thoroughly fused morphologies for the tablets compressed under higher pressures: 50 MPa caused appreciable coalescence of the neighboring particles, and complete fusion was enabled at 400 MPa, substantially lower than the pressure (at least 2 GPa) needed for complete fusion of dry ACC nanoparticles (17). The easier fusion of Mg-ACCP gel is possibly due to the following factors. First, the nanoparticles in the gel offer a high surface energy that promotes the nanoparticles to fuse. Second, the large amount of free and bound hydration water in the gels bestows plasticity to the mineral frameworks and provides pathways to facilitate the mass transfer and structural rearrangements, possibly through pressure-driving deformation and dissolution-reprecipitation (24, 27). The simple compression treatment of the Mg-ACCP gel resulted in compact tablets with high mechanical performance (e.g., modulus/hardness of 19.1/1.0 GPa) and transparency (Fig. 2, F and G, and fig. S10), which may help elucidate the origin of the biogenic optical ceramics (e.g., calcite lens) (24, 28). The hardness and modulus of prepared Mg-ACCP ceramics are comparable to some common biogenic structural materials (fig. S11 and table S2) (29, 30). The differential scanning calorimetry (DSC) analysis (fig. S12A) shows that the compressed Mg-ACCP tablets can withstand a high temperature over 600°C without crystallization. The thermal conductivities and water contact angles of the compressed Mg-ACCP tablets were also measured (fig. S12, B and C), displaying moderate thermal conductivities (0.4 to 0.6 Wm⁻¹ g⁻¹) and hydrophilicity.

Fig. 2. Characterizations of the Mg-ACCP tablets compressed using different pressures.

(A to E) SEM images and corresponding binary-processing images (blue line highlights the particle boundaries). Scale bars, 250 nm. (F) Density and transparency with the optical photographs shown in the insets. (G) Modulus and hardness.

Raman spectral analysis was performed to further investigate the pressed Mg-ACCP tablets [Fig. 3, A to C; the Fourier transform infrared (FTIR) spectra shown in fig. S13 for comparison]. The most prominent peak at 958 cm⁻¹ is attributed to v₁ stretching of P─O bonds, and the peak at 1081 cm⁻¹ corresponds to v₁ symmetric stretching of the C─O bonds (fig. S14). The peaks at 436 and 576 cm⁻¹ are assigned to v₂ symmetric bending modes and v₄ antisymmetric bending modes of P─O bonds, respectively (31, 32). The broad spectra response around 3400 cm⁻¹ offers important information of the hydration species (33). Compared with pure water, the broad band shifts to lower wave number for the Mg-ACCP gel and tablets (fig. S15), indicating the presence of stronger hydrogen bonds between the Mg-ACCP framework and the water molecules (34, 35). Gaussian fitting further resolved it into three peaks representing the O─H stretching vibrations with strong (3020 cm⁻¹), medium (3270 cm⁻¹), and weak (3480 cm⁻¹) hydrogen bonds mainly originating from the single donor–double acceptor (DAA; tied to minerals), double donor–double acceptor (DDAA; “ice-like” and ordered), and dynamic single donor–single acceptor (DA) hydrogen-bonded networks, respectively (Fig. 3, B and C) (36). Peak area calculation revealed that the medium hydrogen-bonded or the ordered “DDAA” water was the majority, and, importantly, the hydration characteristics can be adjusted by the pressure applied during the compression treatment: With higher pressures, the percentage of the strong hydrogen-bonded or the DAA water rose but content of the weak hydrogen-bonded or the DA water dropped, while the DDAA percentage peaked (65.5%) at 200 MPa (Fig. 3D).

Fig. 3. Water analysis of the Mg-ACCP tablets compressed using different pressures.

(A) Full Raman spectra. (B) Zoom-ins of the Raman water signals (Gaussian fits of the three stretching modes of O─H). (C) Typical analysis of the Raman signals of water. (D) Peak area percentages of the different hydration species, calculated from (B). (E) TGA curves of the freeze-dried tablets. (F) Representative TGA curve showing the three dehydration steps. (G) Weight percentages of the different hydration species obtained from (E).

The above trends were found to be highly consistent with the hydration features revealed by the TGA (Fig. 3, E to G) on the tablets that were freeze-dried and then stored in the air under ambient conditions. Roughly three weight loss steps (before 50°C, 50° to 200°C, and 200° to 400°C) were recorded (Fig. 3F), corresponding to the loosely, moderately, and tightly bound water molecules in the tablets, respectively (37–39). Similar to the Raman results, with increasing compression pressure (Fig. 3G), the content of “loosely bound water” declined, the content of “tightly bound water” rose, whereas the content of “moderately bound water,” the dominant hydration species that accounts for over 55.2% of the overall hydration level for all the samples pressed at different pressures, peaked at 200 MPa. Notably, the Raman and thermal observations together revealed the pressure dependency of the hydration states in the fabricated biomineral tablets (40). The strong hydrogen bonds and tightly bound water originate from the water molecules bound tightly with the ionic mineral frameworks (DAA-type water); the medium hydrogen bonds and moderately bound water reside further away from the mineral frameworks (DDAA-type water) (41); the weak hydrogen bonds and loosely bound water are ascribed to the outmost water layers in the bulk water or interfaced with the air (36). Molecular dynamics simulation study on the water layer interfaced with ACC under pressures ranging from 20 to 800 MPa also showed that the content of DDAA-type water peaked at 200 MPa (figs. S16 and S17).

It was found that, among all the Mg-ACCP tablets manufactured under different pressures, the 200 MPa–pressed one crystallized first (within 7 days)—indicating that the ordered water layers are most critical to facilitate crystallization. The crystallization behaviors of the as-pressed Mg-ACCP tablets in the air under ambient conditions closely depended on the compression pressure (Fig. 4 and fig. S18). Once initiated, the crystallization progressed steadily to form cattiite [Mg₃(PO₄)₂·22H₂O] (figs. S19 to S21), a recently discovered natural biomineral that enhances osteogenic (bone growth) activity (42). Upon exposure to the focused electron beam under the SEM, some crystals were observed to exhibit cracking or form pores (Fig. 4F). This is possible due to loss of hydration under the strong electron beam in the vacuum chamber of SEM. The relatively low (≤20 MPa) or high (≥1000 MPa) pressure resulted in stable amorphous states that persisted over 28 days (Fig. 4A), whereas pressures ranging from 50 to 800 MPa induced crystallization within 28 days, with the fastest one observed within 7 days on the 200 MPa–pressed tablet (fig. S22). As mentioned above, these crystallization behaviors are consistent with the Raman and TGA results on the contents of the DDAA-type water (Fig. 3). TGA measurements of the as-pressed samples (fig. S23) revealed that the total water content declined monotonically with the applied pressure and was thus not a determined factor to control the crystallization rate. These observations indicate that the DDAA-type water in the amorphous mineral microstructures is the most essential hydration species to facilitate crystallization, possibly because it features an ordered structure that may serve as a template for initial orientation or nucleation of the mineral ions and a relatively large specific volume that may more easily accommodate the ion migration and structural rearrangement during crystallization.

Fig. 4. Post-compression evolution of the Mg-ACCP tablets stored under ambient conditions.

(A) Pressure- and time-dependent crystallization and fusion behaviors. (B to F) SEM images of the representative samples corresponding to points B to F in (A).

The above experimental observations indicate that the hydration profiles of bioceramics can be manipulated by varying the pressure applied, whereas the hydration species strongly affect the crystallization behaviors. It is possible that, by tuning the compression pressure during the manufacturing process, life gains control over the hydration states and amorphous stability of biogenic ceramics. Notably, the supervariate (multi-ionic, hydrated, and amorphous) strategy (43) demonstrated here offers more versatility, e.g., the ionic compositions of the PIGs are highly adjustable (fig. S1), which, in turn, modulates the properties of the press-produced bioceramics. For example, the tablets obtained by compressing a gel with higher Ca²⁺ and CO₃²⁻ contents (point “B-7373” in fig. S1) displayed enhanced amorphous stability than the Mg-ACCP tablets (fig. S24). More concentrated Na⁺, K⁺, and Cl⁻ in the Mg-ACCP gel precursors effectively delay crystallization of the compressed tablets (fig. S25). Another test using the identical synthetic conditions of the Mg-ACCP tablets but with HPO₄²⁻ replaced by H₂PO₄⁻ led to amorphous bioceramics, which subsequently produced calcite and cattiite within 21 days after the compression treatment (fig. S26). These interesting observations demonstrate the flexibility and accessibility offered by the supervariate approach reported here.

Furthermore, compression-induced fusion was achieved for PIGs of other compositions (e.g., points “A-8282,” “C-7337,” “D-3737,” and “E-3773” in fig. S1) to deliver considerable hardness and modulus (figs. S27 to S29). Taking the PIG corresponding to point A-8282 in fig. S1 as an example, the impacts of the compression pressure on the hydration species and the crystallization behaviors were systematically investigated (figs. S30 and S31). Similar traits to the Mg-ACCP gel were observed: In particular, the DDAA-type water was present as the dominant hydration species whose percentage reached maximum at 200 MPa (fig. S30); again, the tablet compressed under 200 MPa enabled the highest crystallization rate to render monohydrocalcite (MHC) (CaCO₃·H₂O) within 11 days (fig. S31).

DISCUSSION

Compared with the production of ceramics by dry compression of annealed powders, fabricating ceramics by compressing the amorphous PIGs has manifold advantages. First, the PIGs can be easily synthesized (e.g., by simply mixing the four aqueous solutions of CaCl₂, MgCl₂, Na₂CO₃, and K₂HPO₄ at room temperature) and compressed at the wet state, eliminating any drying treatment. The entire synthesis procedure is fully biocompatible, with no need for high temperature, vacuum, or organic solvents. Second, the hydration level of the amorphous precursor is critical to controlling the fusion and crystallization behaviors of the pressed ceramics. For the dry press method, a thermal pretreatment is applied to adjust the hydration level of the amorphous precursors before compression (17), whereas for the PIG approach, the hydration level can be directly tuned through compression. Third, the pressure level to induce mineral fusion is dramatically reduced by using the PIG precursors. Last, the PIG approach offers great convenience and flexibility in incorporating different ions or dopants and in tuning the composition and phase of the resultant ceramic materials.

The amorphous Mg-ACCP gel reported here consists of multiple ionic and hydration species. The high entropy helps to lower the free energy of the amorphous gel, while the multiple ionic and hydration species set space hindrances and barriers for kinetic structural rearrangement, both preventing crystallization. On the other hand, for simulation, this complex material system is extremely difficult to calculate, and we had to use much simplified model of fewer components and thermally quenched disordered structures (44). Inorganic gels reported in literature are mainly oxide frameworks containing alcohols or water, synthesized from alkoxide hydrolysis using the sol-gel method (45). By contrast, the “supervariate gels” reported here are synthesized in a fully biocompatible manner, capable of hosting a large amount of water (fig. S32).

Besides the pressure-mediated dissolution-reprecipitation mechanism (as commonly discussed in the field of cold sintering for inducing the fusion and densification effect) that may produce nuclei to trigger crystallization, the Mg-ACCP gels in this study may take another crystallization route. From the water phase diagram, crystalline ice (e.g., I_VI and I_VII) indeed exists at room temperatures under high pressures: The lowest pressure to crystallize liquid water (to I_VI) is approximately 1 GPa. Here, in this study, the compression treatments of the PIGs were carried out under pressures no higher than 1 GPa. On the other hand, the presence of a notable amount of residue ordered water in the ceramic tablets (e.g., the 200 MPa–pressed ones) under ambient conditions indicates that the ordered alignments of water molecules (e.g., crystallization) may be enabled under milder pressure with the aid of PIGs: The interfacial water molecules are probably oriented in the surface electrical double layers on the negatively charged PIG colloids, which may help lower the pressure needed to form ordered water (i.e., ice) at room temperature. The ice thus produced under pressure may in turn align the disordered mineral surface atoms in a more periodic manner and thus initiate the mineral crystallization by serving as an ordered template that can melt away upon pressure relief. In the subsequent stage, a sufficient amount of hydration water in the ceramic tablet is critical to allow further mass transfer and structural rearrangements for the crystallization and phase separation (e.g., calcite or cattiite grown out of the supervariate ceramics) to proceed.

Achieving the necessary stress levels for the compression treatment can be accomplished with a relatively small force. For instance, when pressing a gel over a 10-nm-wide disk, a minimal force of approximately 8 nN (equivalent to roughly 800 ng) would generate a substantial pressure of 100 MPa. On the other hand, characterizing the biomineralization process within living cells, especially quantifying the applied pressure, poses great challenges for current techniques. To our knowledge, there have been no reports on such characterization. Nevertheless, a noteworthy milestone study on slices of frozen coccolithophore cells observed tiny ACC-containing vesicles secreted from “Ca-rich body” reservoirs, indicating the possible application of forces in expelling the biominerals. This observation aligns well with the fact that both cells and cell membranes can withstand high stress (46, 47).

It is also important to clarify that the synthesis experiments in this study were conducted on a macroscopic scale (e.g., 200-ml solutions and 4-mm-wide tablets) with simplified treatment conditions (pressure was the only variable). These conditions certainly do not reflect the complexity and precision of natural biomineralization processes that occur at the micro- and nanoscales. Moreover, life may use various strategies to control the hydration and crystallization behaviors of its ceramics at lower stress levels, such as via manipulation of ionic strengths, additives, strain rates, the application of electric or magnetic fields, or squeezing through tiny channels (e.g., the transmembrane ones). As a demonstration, the Mg-ACCP gel was pushed through a syringe needle and the content of structural water was measured before and after the treatment: A notable increase by 36.6% (from 4.1 to 5.6 wt%) was enabled by simply squeezing the gel through a needle with an inner diameter of 0.4 mm (fig. S33).

In conclusion, we have accomplished bottom-up biocompatible mineralization from the multi-ionic solutions to the supervariate PIGs and then to the sintered bioceramic objects. The supervariate PIG incorporating water and multiple species of ions (Mg²⁺, Ca²⁺, CO₃²⁻, and PO₄³⁻) boasts a persistently amorphous state to facilitate various treatments. For example, it can be compressed under mild conditions to coalesce into bulk ceramic objects that are dense, transparent, and mechanically strong. In addition, the various water species and their proportions can be easily adjusted by varying the pressure applied, allowing for control over the structure and properties of the resultant ceramic products. To regulate the crystallization rate, the moderately bound, ordered water in the compressed tablets was found to outweigh other hydration characteristics such as the total water content and the tightly bound water molecules. This study opens an avenue for producing amorphous ceramic materials with customizable hydration levels and enables control over their crystallization behaviors, all under fully biocompatible conditions.

MATERIALS AND METHODS

Materials

The chemicals, such as analytical grade anhydrous calcium chloride (CaCl₂), anhydrous magnesium chloride (MgCl₂), anhydrous sodium carbonate (Na₂CO₃), dipotassium phosphate (K₂HPO₄), and potassium dihydrogen phosphate (KH₂PO₄), were purchased from Sigma-Aldrich without further purification.

Synthesis

The MgCl₂ (0.8 M) and CaCl₂ (0.8 M) solutions each of 50 ml were mixed and labeled as solution A, and the Na₂CO₃ (0.8 M) and K₂HPO₄ (0.8 M) solutions each of 50 ml were mixed as solution B. Solution B was slowly added to solution A under magnetic stirring and then continuously stirred for 30 min. The mixture was centrifugally washed five times with the deionized water and finally centrifuged at 8000 rpm for 4 min to afford the Mg-ACCP gel. All experiments were performed at room temperature (25°C).

A mold with an inner diameter of 4 mm (Tianjin Jingtuo Instrument Technology Co. Ltd.) was used. The wet gel was transferred to the mold and pressed in a hydraulic machine (Power Team SPX Hand Pump 10000 PSI P59 Model B) for 5 min. The pressure was increased to the target level at a slow rate to reduce leakage, and monitored using a digital pressure sensor (SBT710, Guangzhou SIMBA TOUCH Technology Co. Ltd.). The thickness of the pressed tablets was approximately 1.5 mm.

Freeze-drying for 24 hours was performed on tablet samples to remove the free mobile water while preserving most of the hydrated water. Regarding sample storage, the tablets were stored in the air at room temperature of 25 ± 2°C with a humidity of 80 ± 10%. The annealing of the Mg-ACCP gel was carried out in a tube furnace in the air at the specific temperature (100° to 900°C) for 10 min with a heating rate of 20°C min⁻¹. The samples were naturally cooled in the air afterward.

Characterizations

The density (ρ) of the tablet was calculated using ρ = m/v, where m is the mass of the tablet, measured on a Sartorius ENTRIS224I-1S electron analysis scale, and v represents the occupied volume of specimens. The values and error bars were obtained by measuring five samples prepared using identical conditions. SEM images were taken using a Thermo Fisher Scientific Quattro ESEM system operated at an accelerating voltage of 15 kV. The samples were sputtered with Au for 30 s in advance. The SEM instrument was equipped with an energy-dispersive x-ray spectroscope (EDS) (Apollo X-SDD). To examine the fusion effect, the SEM images were processed using the ImageJ software, with the potential edges determined by the Canny edge detector and the binary images generated through the Otsu threshold function. Transmission electron microscope (TEM) study was carried out using an FEI Talos F200X equipped with a Talos 200S EDS. XRD patterns were collected on a BRUKER SRD-D2 Phaser with Cu Kα radiation (λ = 1.54 Å) and a Rigaku SmartLab equipped with Cu Kα radiation (λ = 1.54186 Å). The diffraction intensity data were recorded with a step of 0.01° in the 2θ range from 10° to 80°. The PDF cards were from the Jade software database. TGA and DSC tests were performed with TGA-DSC3+ Synchronous Thermal Analyzer (METTLER Instruments) with a heating rate of 10°C min⁻¹ under a constant Ar flow rate of 20 ml min⁻¹. The temperature range was from 25° to 800°C. The FTIR spectra were recorded on a PerkinElmer Spectrum II Spectrometer from 4000 to 400 cm⁻¹ with 32 scans at a resolution of 4 cm⁻¹. Light microscopy images and Raman spectra were obtained using a WITec alpha300 R Raman System from 150 to 3900 cm⁻¹ for four accumulations under a laser beam (wavelength of 532 nm; power of 10 mW). The optical transparencies were obtained by averaging the values between 400 and 780 nm measured on a UV/Vis Spectrometer Lambda 2S. The transparency values were normalized to that of the 1 GPa–pressed tablet (set as 100%). Three specimens were tested to give the average values. X-ray photoelectron spectroscopy (XPS) measurements were made on a PHI model 5802 using monochromatized Al Kα radiation. The hardness and elastic modulus were evaluated with a force of 100 mN and a loading time of 20 s using a Fischer HM2000XY micro-hardness tester. The hardness was characterized using the Martens hardness measurement method with a Vickers indenter tip. The zeta potential tests of Mg-ACCP gel were performed using Brookhaven Instruments with Smoluchowski mode. Thermal conductivities of ceramic samples were evaluated by a TPS2500S Hot Disk Instrument. The wetting performance was carried out by a DataPhysics Contact Angle Tester.

Molecular dynamics simulation

All the molecular dynamics simulations were performed on GROMACS 5.1.4 platform (48), and trajectories were analyzed in VMD 1.9.3 (49). The OPLSAA force field (50) was used to depict the L-J parameters of MHC/ACC (51), and the TIP5P model was applied to depict the water molecules. A cutoff of 12 Å was used to calculate the short-range electrostatic interactions and van der Waals interactions. The long-range electrostatic interactions, considering the periodic boundary condition of the simulation cell, were treated with the particle mesh Ewald method (52). The V-rescale thermostat and the Parrinello-Rahman barostat were applied to control the temperature at around 300 K and the pressure at designated values, respectively. The time step was set to 1 fs, and the trajectories were output every 1000 steps. To achieve a simulated structure of ACC, a supercell of MHC containing 1152 CaCO₃ was first annealed from 300 to 3000 K and then cooled to 300 K within 10 ns. After the annealing, another cell of the water molecules was added onto the ACC cell. The combined ACC-water system was first energy minimized for 3000 steps, followed by equilibrations in NVT (constant volume and temperature ensemble) for 4 ns and NPT (constant pressure and temperature ensemble) for 5 ns, respectively. The production simulations were run for 20 ns with different applied forces to investigate the relations between ACC-water interfaces and the applied pressures. The trajectories of the last 5 ns were used for data analysis. To simulate the process of pressurization, the periodic boundary conditions were applied along the x and y directions but removed along the z direction (fig. S16). Consequently, two walls at the two ends of the simulation cell (z direction) were added, and a graphene piston on the top of the simulation cell was applied to press the water and generate the designated pressures. The CA atom types in the CHARMM force field were used to depict the carbon atoms in the piston. The piston was initially placed at around 1 nm below the wall on the top. For comparison, the pure water systems were also simulated as the control groups. The applied pressing forces on the piston were 1.821, 182.1, 364.2, 901.5, 1821.1, 3642.2, 7284.4, 10926.6, and 14568.8 kJ mol⁻¹ nm⁻¹, respectively, to generate the applied pressures of 0.1, 10, 20, 50, 100, 200, 400, 600, and 800 MPa, respectively. The initial box of the combined ACC-water system was 5.6 nm by 5.4 nm by 8.5 nm, and that of the pure water system was 5.6 nm by 5.4 nm by 7.2 nm. The hydrogen bond was identified by the criteria of the O···O distance of 3.5 Å and donor-hydrogen-acceptor angle of 140°. Different donor-acceptor types were identified under the different applied pressures in both the ACC-water system and the pure water system.

Supplementary Materials

This PDF file includes:

Figs. S1 to S33

Tables S1 and S2