Abstract
Liquid metals are promising unconventional solvents for dissolving metals and growing their crystals. However, how the crystals grow within these metallic solvents cannot be observed easily due to the opacity of liquid metals. Using X-ray micro-computed tomography, the three-dimensional formations of crystals inside liquid metal solvents, under different conditions, are presented. In this work, we explore the use of different liquid metals solvents for growing metallic crystals, while imaging the crystals shape and sizes inside liquid metals. We use gallium-based solvents and platinum, as the solutes and demonstrate the use of liquid metals for forming metallic crystals through modulating environmental boundary effects, controlled cooling rates, and solvent selection. It is observed that the cooling rate and composition of the reaction media will influence crystals morphologies and intermetallic phases. As a proof-of-concept, we perform hydrogen evolution reaction using metallic crystals extracted from the solvents, showcasing the potential of liquid metal-based synthesis in tailoring metallic crystals properties for enhancing catalytic activity demonstrating a strategy for engineering metallic crystals in liquid metal solvents.
Similar content being viewed by others
Introduction
Liquid metals neither resemble solid metals nor typical liquids1. Liquid metals possess relatively high density and a shimmering surface, while also displaying a perplexing fluidity2,3. They maintain many of the metallic and physicochemical properties of solid metals, while offering liquid-like characteristics4,5,6. Liquid metals embody some core qualities from its constituent metallic solid and liquid phases7,8, while exhibiting strong interactions with solid metals when in direct contact with them when kept at the right temperatures9. Such conditions are innately favorable for dissolution10, interdiffusion11,12, and intermetallic13 formation of metals inside liquid metals11,13,14. Therefore, liquid metals present promising possibilities as reaction media for dissolving metallic elements and synthesizing a wide variety of metallic crystals.
Gallium-based alloys, characterized by their low melting points, are amongst the most extensively researched liquid metals15,16. Gallium (Ga), as a p-block post-transition metal, is able to form covalent bonds with other elements, and exhibits a low melting point due to the relatively weak intermolecular forces4,5. Liquid Ga forms intermetallic phases with many metals, particularly post-transition and zinc group metals, at or near room temperature17,18. Additionally, considering the evidence of local structuring within liquid Ga at low temperatures19,20,21, liquid Ga offers the potential to direct the growth of solute metal structures by favoring particular symmetries and or orientations during crystal growth, post-nucleation14,22,23.
While liquid metals offer promising potential as the growth media for metallic solutes24, crystal growth observation within liquid metal solvents has remained a challenge25. One reason for this is that liquid metals are non-transparent to low- and mid- energy electromagnetic waves, and even high energy electron beams struggle to penetrate them, leading to a tendency to overlook phenomena that cannot be directly observed4. Investigating crystal growth within bulk liquid metals provides fundamental insights into the growth mechanisms, elucidating the underlying principles governing solidification kinetics and phase transformations inside liquid metal systems. The possible access to observe inside liquid metal solvents allows for the more precise control of crystal morphology, size, and distribution, leading to the exploration of novel pathways for fabricating crystalline structures. It also enables more accurate predictions of material behavior under diverse conditions for the synthesis and processing of advanced materials using this category of solvents.
Altogether, controlling the dimensions and facets of the formed crystals requires real-time observational access during synthesis. Furthermore, no information on the location or growth direction within the liquid metal solvent has been reported so far, due to the limitations of existing observational tools, which rely on low-energy waves that cannot penetrate the bulk of the metallic solvents. This means, how the environmental conditions alter the crystal growth and morphologies, during the growth of crystals inside a liquid metal solvent, could not be determined. Hence, there is a need to introduce the use of instrumental techniques and procedures for realizing the elaborate growth mechanisms inside liquid metals26.
In this study, X-ray micro-computed tomography (XCT) analysis was employed as the primary technique to observe the synthesized crystals within the liquid metal under different conditions. While the XCT technique has been used in metal additive manufacturing during development processes and certification in various sectors27,28, such as aerospace29, tooling30, and biomedical applications31,32, their utilization in the analysis of liquid metal–based materials remain rare. XCT analysis offers distinct advantages over other visualization techniques, including the ability to focus on the evaluation of internal structures, orientation, size, and distribution, without the need for sample preparation33. Additionally, the sample can be fully recovered once the analysis is complete. The XCT technique provides a comprehensive three-dimensional (3D) representation, enabling detailed visualization of internal features that are otherwise unobservable.
The novelty of this research lies in the ability to observe crystal growth and morphologies inside liquid metal using XCT. This method allows crystal growth to be visualized without disrupting the surrounding environment or the boundary (i.e., the liquid metal as a solvent). To the best of the authors’ knowledge, there is no existing literature discusses the direct observation of crystals forming inside liquid metal.
Here, platinum (Pt) is dissolved as a solute metal in liquid Ga and liquid eutectic Ga-indium (EGaIn), which serves as the solvents, representing our model systems. This process leads to the formation of intermetallic compounds comprising two or more metals with an ordered crystal structure and defined stoichiometry. We demonstrate that the morphology and crystal formation processes of compounds inside liquid metals can be tailored under different growing conditions applied to the metallic solvents. The influence of the rate of cooling and reaction media type on this system are specifically investigated.
Liquid metals, such as Ga and EGaIn, have low melting points (29.6 °C and 15 °C, respectively), allowing them to remain liquid at low temperatures, making them accessible and easy to handle in laboratory settings34. Post-transition elements, in such liquid metals, exhibit covalent characteristics and also directional bonding, enabling their use as solvents for metallic solutes35,36. Particularly, a recent study by Liu et al. have investigated the crystallization of liquid Ga, showing the presence of two different crystalline phases (α-Ga and β-Ga), distinguished by directional bonding37. In particular, directional bonding has been predicted at the surface of Ga-based liquid metals38,39. The low viscosity of Ga-based liquid metal40 facilitates manipulation and processing, such as extraction. Additionally, liquid metals function as electronic liquids with intrinsic conductivity, distinguishing them from conventional solvents. Unlike highly reactive low-melting-point metals like alkali metals, post-transition liquid metals are stable in air and moisture, enhancing their practicality.
Pt is chosen since it results in a high X-ray contrast with respect to Ga and Indium (In). XCT analysis is utilized to provide observations of the spatial arrangement of formed crystals inside the liquid metal droplets in a non-destructive manner. By voltage-based zeroing of the surface tension of the liquid metals, the synthesized metallic crystals are extracted and characterized as complementary investigations to XCT. The binding energy calculations and ab initio molecular dynamic (AIMD) simulations are also performed to support the data analysis. Finally, as a proof-of-concept application, we utilize the crystals extracted from inside the liquid metal droplets as electrocatalysts in a hydrogen evolution reaction (HER) study, demonstrating the viability of using the method for creating functional metallic crystals.
Results
Growth of crystals within liquid metal droplets
In this study, the changes to the crystallinity, microstructure, and spatial arrangement of intermetallic crystals within liquid metal droplets, due to the influence of environmental boundaries and solvent contents, were explored. Various metallic crystals inside the liquid metal bulk were grown at different conditions which will be discussed subsequently.
Two different solvent systems, Ga and EGaIn, were used in a liquid state to serve as the reaction media. The solute element of 2 weight percentage (wt%) of Pt, was thermally dissolved into these two different solvent systems. Pt (density of 21,450 kg/m3) was used as it provides a stark contrast with Ga (density of 5905 kg/m3) and In (density of 7290 kg/m3) for X-ray imaging, owing to their significantly different densities and X-ray attenuation profiles41. The higher X-ray attenuation behavior of Pt maximizes the likelihood of observing the formed crystals inside the liquid metal systems.
After homogeneous liquid alloys were produced, these alloy systems were subjected to two different cooling rates: (a) fast-cooling, and (b) slow-cooling. The cooling allowed the solute metals to become oversaturated and precipitate inside the liquid metal solvents in crystalline forms. Fast-cooling is defined as leaving the liquid metal droplet to cool naturally from 500 °C to ambient temperature, after ~5 min, during which the temperature remained nearly constant. This results in an approximate average cooling rate of ~95 °C/min. Moderate-cooling is defined as a controlled temperature reduction from 500 °C to 20 °C at an intermediate rate of 10 °C/min, between the fast- and slow-cooling conditions, and is achieved over 48 min. In contrast, slow-cooling is defined as a controlled temperature reduction from 500 °C to 20 °C, achieved in over 480 min, at a rate of 1 °C/min. EGaIn is naturally in liquid form at room temperature. In our experiments, bulk Ga also remained in its liquid state owing to the supercooling effect42. By tuning the growth conditions, the choice of liquid metal solvent, and the rate of cooling, diverse crystal morphologies were produced.
Figure 1a depicts a schematic of the synthesized process for growing metallic crystals with distinct crystalline shapes using Ga or EGaIn liquid metals as the reaction media. These liquid metal droplets were then analyzed using XCT to observe the crystals formed within the bulk of liquid metals in their original state as illustrated in Fig. 1b. Afterward, the crystal extraction process from the liquid metal droplets (Fig. 1c) was carried out to examine their physical dimensions and crystalline structure in detail.
a Alloying procedure for growing metal crystals inside a droplet of liquid metal solvent, with inset showing crystalline precipitates within liquid metal bulk formed during the cooling process. b Experimental setup for XCT scanning of liquid metal droplets. c Extraction process of crystals from inside liquid metals electrochemically aided by a vacuum-driven force. Figures were created by the authors using Autodesk Fusion 360, version 2604.0.316 and composited in Adobe Illustrator, version 29.0.
3D XCT observation of crystallites inside liquid metals
Supplementary Fig. 1 shows a photo of the XCT instrument used in this experiment, with the sample holder, a brass pin mounted on a rotational stage, positioned between the source (left) and the detector with filter wheel (right). The liquid metal droplet was drop-cast onto a silica wafer, mounted onto the sample holder, as shown in the zoomed-in photo in Supplementary Fig. 1b (left). Our XCT collected a series of X-ray projection images through a sample rotating around a central vertical axis. These projections were then reconstructed to produce a stack of cross-sectional images that could be visualized in 3D. The intensity of each pixel in a cross-sectional image indicated the beam’s attenuation at that point within the sample. When two-dimensional (2D) image stacks, as illustrated in Supplementary Fig. 1c, were reconstructed into 3D volumes, the pixels were represented as voxels (volume elements)43.
A schematic diagram of the experimental design showing the overall crystal formation process, cooling duration, and cooling rate under different cooling conditions, is presented in Fig. 2a–c. Additionally, the photo of the sample, illustrating the opacity and surface reflectivity of a liquid metal droplet, is shown in Supplementary Fig. 1b (right). By utilizing XCT analysis, the crystal inside liquid metal droplet can be observed, while preserving their original environment, as presented in Fig. 2d–i. The XCT images for Pt-Ga droplets are shown in Fig. 2d, f, h, while the observations for Pt-EGaIn systems are presented in Fig. 2e, g, i, under different rates of cooling (fast-, moderate-, and slow- cooling). Based on many observations, most crystals grew near the interface of droplets that likely acted as a substrate for seeding.
Diagram of the experimental design showing a overall crystal formation process, b cooling duration, and c cooling rate values under different cooling conditions. Formation of crystals under fast-cooling within d Ga liquid metal droplet, and e EGaIn liquid metal droplet. Formation of crystals under moderate-cooling within f Ga liquid metal droplet, and g EGaIn liquid metal droplet. Formation of crystals under slow-cooling within h Ga liquid metal droplet, and i EGaIn liquid metal droplet.
The crystals formed in the liquid Ga system (Fig. 2d, f, h) were predominantly rod-shaped, whereas those grown in the liquid EGaIn medium (Fig. 2e, g, i) were mainly polyhedral. A decrease in cooling rate, from fast- to slow-cooling, resulted in progressively larger crystals. Under fast-cooling conditions (Fig. 2d, e), small crystals were obtained. At a moderate-cooling (Fig. 2f, g), the crystals exhibited dimensions that were intermediate between those formed under fast- and slow-cooling conditions. Under the slow-cooling conditions (Fig. 2h, i), much larger crystals were observed. Specifically, crystals formed in Ga-based solvents, fast-cooling (Fig. 2d) produced small crystal, while moderate- (Fig. 2f) and slow-cooling (Fig. 2h), led to the progressive development of elongated crystal, rod-like morphologies. In contrast, crystals formed in EGaIn-based solvents tended to be more equiaxed, with sizes increasing from small crystals under fast-cooling (Fig. 2e) to progressively larger, polyhedral-shaped crystals formed under moderate- (Fig. 2g) and slow-cooling (Fig. 2i). Other supporting observations can be seen in Supplementary Videos 1–6 and Supplementary Figs. 4–7.
The size distribution of metallic crystals analyzed by XCT is depicted in Fig. 3a–f. The crystals formed under fast-cooling conditions tend to have a smaller size distribution, with dimensions ranging from 0 to 650 µm (median of 85 µm) and from 0 to 500 µm (median of 73 µm) for Ga and EGaIn solvent-based crystals, respectively. Under moderate-cooling condition, the size distribution of crystals grown in Ga-based solvents ranges up to 860 µm (rod-shaped crystals with a median size of 290 µm), while for those grown in EGaIn-based solvents, the distribution reaches up to 560 µm (polyhedral-shaped crystals with a median size of 130 µm). In contrast, crystals obtained under slow-cooling conditions resulted in broader size distributions, with Ga solvent-based crystals measuring up to 2800 µm (long rod shaped crystals - median of 664 µm for the length of rods only) while EGaIn solvent-based crystals ranged in size up to 1100 µm (polyhedron - median 459 µm), respectively.
Size distribution (histogram) of crystals formed in a Ga-based liquid metal droplet under a fast-cooling, b moderate-cooling, and c slow-cooling conditions. Size distribution (histogram) of crystals formed in a EGaIn-based liquid metal droplet under d fast-cooling, e moderate-cooling, and f slow-cooling conditions. Crystal growth speed distribution presented as a box plot under different liquid metal compositions and cooling rates: g fast-cooling, h moderate-cooling, and i slow-cooling conditions. Data are obtained from a single droplet with each dot representing a single crystal, and the error bars correspond to outliers 1.5× the interquartile range (IQR).
The crystal growth speed varies depending on the chemical composition of the liquid metal solvent and the cooling rate, as illustrated in Fig. 3g–i. Crystals formed in Ga-based liquid metal solvents exhibit faster growth compared to those in EGaIn-based solvents. Regarding cooling rates, the fast-cooling conditions results in the fastest crystal growth relative to the moderate- and slow-cooling conditions. The crystal growth speed distribution in Ga-based solvents ranges from 13 to 120 µm/min, with a median of 17 µm/min under fast-cooling; from 3 to 18 µm/min, with a median of 6 µm/min under moderate-cooling; and from 0.5 to 5.7 µm/min, with a median of 1.4 µm/min under slow-cooling conditions. For crystals formed in EGaIn-based solvents, the growth speed distribution ranges from 8.3 to 90.5 µm/min with a median of 14.6 µm/min; from 1.5 to 11.6 µm/min with a median of 2.8 µm/min; and from 0.3 to 2.2 µm/min with a median of 0.95 µm/min under fast-, moderate-, and slow-cooling respectively.
For a more comprehensive comparison, we have also conducted additional experiments to observe crystal growth in other liquid metal solvent systems, specifically eutectic Ga-tin (EGaSn) and the eutectic alloy of Ga, In, and Sn (EGaInSn), as illustrated in Supplementary Figs. 2, 3, and Supplementary Note 1. In short, crystals formed in Ga-based solvent exhibit the greatest length and the fastest growth rate. The addition of other components, In and Sn, reduces both the crystal length and growth rate. Among the cooling conditions, the fast-cooling condition lead to the shortest crystal length and the fastest crystal growth rate.
Altogether, we illustrated that XCT offers a reliable way to overcome the challenge of observing crystal growth within the bulk of liquid metal, an opaque material that is impossible to penetrate with light and electrons44. The combination of using Ga, as the liquid metal solvent, with fast-cooling provides the highest crystal growth rate among all conditions. Under fast-cooling conditions, the temperature decreases rapidly (95 °C/ min), creating a strong driving force for crystallization to begin. Crystal growth kinetics change, limiting the time for crystal growth and leading to the formation of many nucleation sites for small crystals to emerge. By contrast, under slow-cooling conditions, the temperature decreases gradually (1 °C/ min), allowing the growth environment to remain in saturated condition for Pt for a longer time. This provides the system with sufficient time for the formation of fewer but larger crystals. The broader size distributions observed under slow-cooling conditions can also be attributed to the longer time available for crystal growth. The phenomenon is similar to conventional crystal growth45,46.
Crystals extraction and characterization
To further investigate the detailed properties of the crystals formed within the liquid metal droplet, both the bulk sample containing the crystals and the extracted crystals were examined to characterize their physical dimensions and crystalline structure. The characterization of the bulk sample aims to explore the crystal properties in their original environment, while the post-extraction analysis focuses on examining the extracted crystals without interference from the background (solvent). To extract the crystal from the liquid metal droplet, the extraction process was carried out by removing the liquid metal solvent, starting with the filtration of the metallic colloidal solvents while applying voltage and a vacuum-driven force to the liquid metal droplets. (Fig. 1c and Supplementary Note 2)
Crystallographic structures of the bulk sample and the formed intermetallic crystals were assessed using X-ray diffraction (XRD) to distinguish various phases in the crystals generated at from different conditions. As can be seen, the impact of compositions of the reaction media and boundaries led to the formation of different intermetallic phases. XRD analysis revealed that Pt dissolved in the liquid Ga system resulted mostly in Ga2Pt intermetallic phase (Fig. 4a, c), while Pt dissolved in the liquid EGaIn resulted in the predominance of Ga7Pt3 (Fig. 4b, c). Additionally, some In7Pt3 phases were also found in crystals from the EGaIn solvent. Additional peaks corresponding to Ga and In are observed in the XRD of the bulk sample, confirming the presence of the solvents used as the growth medium (Supplementary Fig. 9a, b). The pre- and post-extraction XRD results are consistent, suggesting that the extraction process does not alter the characteristics of the formed intermetallic crystals. The crystallographic parameters of the formed crystals in two different liquid metal solvents are summarized in Supplementary Table 1. The phase diagrams governing the formation of crystals are presented in Supplementary Fig. 8.
X-ray diffraction (XRD) patterns of the extracted crystals from a Ga liquid metal, and b EGaIn liquid metal under the fast-, moderate-, and slow-cooling conditions. c The ratios of extracted intermetallic crystal phases formed inside the liquid metal. d Growth preferences for rod-shaped crystal, with the inset illustrating the tetragonal Ga2Pt unit cell (top), and polyhedral-shaped crystal, with the inset showing the cubic Ga7Pt3 unit cell (bottom). Ga atoms are shown in green and Pt atoms in blue. Scanning electron microscopy (SEM) images of typical crystals inside Ga liquid metal under e fast-, f moderate-, and g slow-cooling conditions. SEM images of typical crystals inside EGaIn liquid metal under h fast-, i moderate-, and j slow-cooling conditions.
The surface morphology and the surface elemental mapping of the intermetallic crystals, both within the bulk sample and extracted from the liquid droplet Ga vs EGaIn systems, were also observed using scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS). The schematic of crystal growth preferences is illustrated in Fig. 4d. Altogether, the crystals exhibited various morphologies and dimensions, as shown by SEM-EDS imaging in Fig. 4e–j. When Pt was dissolved in liquid Ga as the reaction medium, rod-shaped crystals were predominant (Fig. 4e–g), whereas polyhedral crystals were more frequent in liquid EGaIn reaction medium (Fig. 4h–j). The SEM observations of the surface of the bulk samples reveal similar morphologies to those observed in the extracted crystals at the surface boundaries, with the crystals still surrounded by their solvent (Ga or EGaIn), as shown in Supplementary Fig. 9c–f. In the fast-cooling process, the crystals tend to reside at the interface of the droplets, emerging from their solvent. In contrast, in the slow-cooling samples, crystals tend to be located beneath the interface, likely due to their larger size. Overall, these SEM observations align with the results obtained from the XCT analysis.
Table 1 summarizes the intermetallic crystals formed under different liquid metal solvent systems and growth conditions, highlighting the key aspects where the solvent composition influences the resulting crystal morphology and phase composition of the formed crystals. Crystals grown in Ga-based solvent generally exhibit a rod-like morphology, predominantly forming the Ga2Pt intermetallic phase. In contrast, crystals formed in EGaIn-based solvent systems present a polyhedral crystal morphology, with the majority composition corresponding to the Ga7Pt3 intermetallic phase.
Measurements of dimensions from the SEM observations also suggest that the slow-cooling process induced more intense growth around a few seeds, promoting the formation of larger crystal structures (Fig. 4g, j) ranging from 0 to 1400 µm (median 206 µm for rods) for Ga solvent-based crystals (length of rods) and 0 to 850 µm (median 115 µm for clusters) for EGaIn solvent-based polyhedral crystals (Supplementary Figs. 10c, f). Crystals obtained under moderate-cooling conditions exhibit sizes intermediate between those formed under slow- and fast-cooling (Fig. 4f, i), ranging from 0 to 650 µm, with median sizes ~70 µm and 50 µm for crystals formed in Pt-Ga and Pt-EGaIn solvent, respectively (Supplementary Fig. 10b, e). In comparison, fast-cooling, results in smaller crystals (Fig. 4e, h) observed to range between 0 to 200 µm (median 52 µm) and 0 to 160 µm (median 32 µm) for Ga and EGaIn solvent-based crystals, respectively (Supplementary Fig. 10a, d). These size distributions are similar above approximately a 25 µm threshold (XCT imaging limit in our work) to those presented in Fig. 3. The imaging rate of our XCT instrument is a limitation and more advanced XCT instruments can result in a much better matching of the median value.
The larger crystal sizes formed during the slow-cooling process result from the increased time available for solute molecules to arrange into the growing crystal lattice. Besides, the slow-cooling condition provides more time for the Ostwald ripening process to occur where the smaller crystals dissolve, allowing the solute to redistribute, and contribute to the growth of larger crystals47.
Different parameters of the environmental boundaries, the cooling rates, and the chemical compositions of solvent are crucial factors governing material formation and properties, thereby shaping the growth behavior of crystals. How the addition of other parameters, such as magnetic field48,49, can influence crystals characteristics remain as interesting topics for future studies.
DFT and AIMD simulations for defining crystals growth
The effect of intermetallic Pt crystals on the liquid metal properties was determined using ab initio molecular dynamics simulations. The first model system (Fig. 5a–e), the Pt particles were represented using a Pt(001) surface, interfacing with liquid metal (Ga, EGaIn, and In) to provide insight into the fundamental atomic distribution and interaction between the growing Pt crystals and liquid metals (Ga, EGaIn, and In). The second model system of Pt dopant in 6-layer Ga matrix was intended to examine the availability of Pt at the liquid metal interface, in order to better understand possible growth mechanisms.
a Snapshots of the Ga(l) (upper), EGaIn(l) (middle) and In(l) (lower) on the Pt(001) surface structures after a 10 ps simulation time. b Radial distribution functions (RDFs) between the liquid metal (Ga and In) and Pt atoms. c RDFs between the liquid metal atoms. d, e Density profile analysis of the liquid metal atoms: d Ga and e In. f Mean square displacement (MSD) of the liquid metal atoms. g MSD of Pt atoms perpendicular to the surface in a Ga or EGaIn liquid metal matrix.
The simulations represent fast-cooling (Fig. 5) and slow-cooling (Supplementary Fig. 12) conditions. The analysis of system pressure during the AIMD simulations is provided in the Supplementary Note 3. The radial distribution functions (RDFs) for the three systems (Fig. 5b) show that the main interaction between the Ga and Pt atoms is at a distance of ca. 2.5 Å in the Ga(l)/Pt system, while the dominant peak between In and Pt is at a greater distance of ca. 3.0 Å in the In(l)/Pt system. The dominant peak in the RDF between Ga and Pt is higher in intensity than that between In and Pt, suggesting that the interaction between Ga and Pt is stronger than that between In and Pt. In the EGaIn(l)/Pt system, the intensity of the first-shell peak at 2.5 Å, corresponding to a Ga-Pt interaction increases in intensity compared to the pure Ga system, while the peak at 3.0 Å, corresponding to an In-Pt interaction, decreases. This indicates that the addition of In to liquid Ga enhances the interaction between Ga and Pt and weakens the interaction between In and Pt. The RDFs for the liquid metals themselves (Fig. 5c) indicate that they retain their liquid properties, as evidenced by the absence of distinct peaks representing long-range ordering.
The density profiles of Ga and In along the z-axis (starting from the outermost Pt surface layer at ca. 9.3 Å) for the simulation representing fast-cooling (Fig. 5d, e) show distinct patterns. In the Ga(1)/Pt system, Ga reaches its highest peak at ca. 9.8 Å, with multiple lower-intensity peaks at greater distances. In contrast, for the In(l)/Pt system, the In is mostly located a bit further from the Pt surface, at a distance of ca. 10.1 Å. In the EGaIn system, the main Ga peak is again at ca. 9.7 Å, however, the In is more evenly distributed throughout the system as shown by peaks of more uniform intensities. Also, the first peak for In shifts to ca. 10.8 Å. This also indicates that the solid Pt has a less pronounced effect on the distribution of In, consistent with the weaker interactions between In and Pt as also indicated by the RDFs.
The mean square displacement (MSD) analysis (Fig. 5f) shows that the liquid metals maintain their diffusive behavior, under these conditions. Notably, Ga in the Ga(l)/Pt system exhibits a higher diffusion rate than In in the In(l)/Pt system. However, the addition of In to Ga significantly impacts the mobility of Ga, resulting in a lower diffusion rate in the EGaIn(l)/Pt system. Hu et al.50 have conducted an extended theoretical investigation of classical nucleation possibilities, highlighting the roles of interfacial pre-ordering and interfacial energy in influencing crystallization. A high degree of geometric structuring is apparent within 8.5 Å of liquid Ga interfaces (surface of liquid metal) and can be engineered by the addition of oxide layers or other elements51. The preferential mobility of Pt atoms toward the bulk liquid Ga, combined with the strong Ga-Pt interaction, limits Pt’s diffusion and affects their interfacial behavior52. The observation of different intermetallic compositions depending on cooling rate and solvent suggests a kinetic explanation for the change of composition, based on the relative availability of the Pt dopant as the crystal forms in solution.
DFT calculations of the mobility of individual Pt atoms were performed for slabs of liquid Ga or EGaIn, 6 atomic layers thick, at 500 K. This temperature was chosen to ensure the surface in the simulations remained liquid, as the melting point of a finite slab of Ga is well-understood to be higher than that of bulk Ga in experiment53 and maximizes the chance of migration of the dopant being observed. This offset in temperature can be thought of as a systematic shift made to ensure that the liquid Ga in the simulations maintains a similar mobility to experimentally observed Ga.
Pt was doped into the cell in place of a Ga atom at approximately the mid-point of the model. The duration of the simulation was 100 ps with a 2 fs timestep (Supplementary Note 3). The average MSD of the dopant atom in the z-direction–i.e. towards the surface of the slab–are pictured in Fig. 5g. Notably, the MSD for the Pt atom in Ga is significantly higher than in EGaIn, reflecting a higher mobility of the dopant. This will be reflected in a higher availability of Pt as the intermetallic crystal phase grows and thus provides an explanation for the higher proportion of the Pt-deficient Ga7Pt3 intermetallic phase in the EGaIn solvent.
Proof-of-concept HER performance
To demonstrate a utilization of the findings, a proof-of-concept HER study of the synthesized materials, using different metal solvent systems and cooling, was evaluated in both acidic (0.5 M H₂SO₄) and alkaline (0.1 M KOH) media. In this study, the synthesized crystals were first loaded onto a polished glassy carbon electrode (GCE) with a surface area of 0.196 cm2, which served as the working electrode. A graphite rod and a saturated calomel electrode (Hg/Hg2Cl2, SCE) were used as the counter and reference electrodes, respectively. The schematic illustrations of HER samples preparation and experimental setup are shown in Fig. 6a, b. The potential was applied to the system and calculated according to the Nernst equation54 (Supplementary Eq. (2a, b)), and the Tafel plots were derived from the polarization curves using the Tafel equation55 (Supplementary Eq. (3a, b)).
a Catalyst powder preparation process. b HER experimental setup. (Figures a and b were created by the authors using Autodesk Fusion 360, version 2604.0.316 and composited in Adobe Illustrator, version 29.0). Polarization (linear sweep voltammograms) curves for the formed crystals performed in (c) an acidic electrolyte (0.5 M H₂SO₄), and (d) an alkaline electrolyte (0.1 M KOH).
The variation in HER performance among the samples can be attributed to the solvent used, the extracted solid crystals, their phase, and their surface porosity. A detailed discussion of these differences is presented as follows and illustrated in Fig. 6. Based on linear sweep voltammograms- LSV (Fig. 6c, d) and Tafel analysis (Supplementary Fig. 14), crystals obtained from the liquid Ga system showed superior HER activity compared to those from the liquid EGaIn system across both media, regardless of cooling rate during synthesis.
Among all samples, the electrode containing crystals from the liquid Ga system formed during fast-cooling exhibited the highest HER catalytic activity (Fig. 6c, d). The Pt-Ga crystals produced under fast-cooling conditions were primarily composed of Ga₂Pt, which contained smaller crystals than those synthesised during slow-cooling. These crystals displayed high HER performance, achieving a low overpotential of −102 mV vs RHE at a current density of −10 mA/cm² in acidic medium (Fig. 6c). In alkaline medium (0.1 M KOH), the same crystals required a slightly more negative potential of −171 mV vs RHE at −10 mA/cm² (Fig. 6d). Supplementary Fig. 13 illustrates the LSV of Pt/C serves as the benchmark. Since Ga is much cheaper than Pt, achieving good HER performance with less Pt represents a technological advancement.
Figure 6c, d and Supplementary Fig. 14 demonstrate that the catalytic activity for all of the systems are better in the acidic environment compared to alkaline conditions. In alkaline electrolytes, it is known that HER activity depends on the speed of water dissociation. The additional energy barrier introduced by water dissociation determines the rate of alkaline HER. This is the reason the benchmark of Pt displays sluggish alkaline HER56,57. Crystals extracted from a liquid Ga-based solvent formed Ga2Pt intermetallic crystals, and it is well known that Pt plays the catalytic role in HER. It is worth mentioning that the usage of Ga in this system also proved advantageous, as we have previously shown that each Pt atom can be covered by several Ga atoms, with Ga atoms exhibiting activation comparable to that of Pt atoms44,52,58.
Something also important is the strong performance of Ga2Pt, regardless of the crystal size dimensions. This phenomenon may stem from small pores on the surface of Ga₂Pt crystals, formed as Ga was etched away during extraction, as shown in our SEM images (Fig. 7a–d). The extraction process produced porous surfaces with varying porosity due to crystal orientation and surface properties under different growth solvents and cooling conditions. The crystals obtained after fast- and slow-cooling have different surface properties, which result in varying pore sizes after initial NaOH washing. Specific phases of the Ga-Pt alloys have a higher tendency to form porous structures due to their crystal orientation and surface properties. It appears that the nano- and meso-porous surface of samples, obtained through fast-cooling, leads to a more porous surface, enhance a higher surface area, contributing to better HER performance compared to the slow-cooling samples, where only sparce etched areas are present, resulting in a less porous surface. The pore size distribution is illustrated in Fig. 7e–h. The pore size distributions of crystals obtained from Ga- and EGaIn-based solvents, under fast-cooling condition, generally show comparable pore sizes, ranging from 10 to 300 nm. Under slow-cooling conditions, the pore sizes are typically larger than those formed under fast-cooling conditions. It ranges from 15 to 500 nm for Ga-based crystals, and 40 to 1020 nm for EGaIn-based crystals. In addition, regarding the porous morphology (Fig. 7a–d), the porous structure that emerged from Ga-based solvent crystals tend to be more elongated, whereas those from EGaIn-based solvent crystals are generally more circular.
Scanning electron microscope (SEM) images showing the porosity of typical crystals formed in Ga-based liquid metal solvents under a fast-cooling, and b slow-cooling, and in EGaIn-based liquid metal solvents under c fast-cooling, and d slow-cooling conditions. e–h Statistical pore size distribution, measuring the Feret’s diameter of individual pores in a single crystal using ImageJ software. The data were fitted with a lognormal distribution curve, where each point represents a single pore formed in Ga-based liquid metal solvents under e fast-cooling (243 points), f slow-cooling (529 points), and those formed in EGaIn-based liquid metal solvents g fast-cooling (204 points), h slow-cooling (371 points) conditions.
The crystals obtained from EGaIn solvent systems, exhibited inferior HER activity compared to those obtained from Ga solvent samples. The inclusion of In in liquid metal solvents, EGaIn (l), led to the presence of In on the surface of and within crystals (Fig. 4h–j). The presence of In could potentially hamper the active catalytic sites of Pt (Supplementary Note 4). In addition, it is also seen that pores in In-containing samples are not as widespread across the surface of crystals (Fig. 7c, d). This may either be due to the presence of surface In or the effect of crystal orientation that hampered the production of surface pores.
Overall, we showed that In inclusion and surface structural changes affect the catalytic performance of the extracted crystals for HER activity. Altogether, the solvents’ properties and cooling rates contributed to the modulation of crystal phases, their interfaces, and subsequently altered catalytic performance of the materials extracted from solvents. Fast-cooling of Ga leads to the highest HER activity of the extracted crystals, compared to those extracted from EGaIn and the slow-cooling.
Discussion
This study presents the 3D spatial arrangement of the formed intermetallic crystals within liquid metal droplets in their original condition using XCT, providing insights into the influence of environmental boundaries and conditions on crystal growth dynamics. By controlling the cooling rate and the chemical composition of the reaction media (solvents), the crystal morphologies and phases can be tuned. The presentation of the 3D images promises novel insights for in -situ tailoring and verification of crystal formed within liquid metal bulk, opening avenues for utilizing liquid metal-based material in additive manufacturing and metallurgy. The use of liquid metals as reaction media offers a unique, streamlined approach to metallic crystal growth, unlike conventional methods relying on reactive soluble precursors. The agreement between experimental and simulations further provided theoretical insights for the influence of solvent system composition and cooling conditions on crystal growth. This study also demonstrates that the formed crystals within liquid metal droplets can be successfully extracted from their solvents via electrocapillary modulation and vacuum filtration. The application of the extracted crystals showcased the catalysts’ performance as a proof-of-concept experiment in HER. This approach can be implemented in laboratories and manufacturing sites without sophisticated equipment, reducing energy costs and supporting sustainable technologies and industries. The outcomes will advance applications of liquid metal for crystallization pathways.
Methods
Materials and liquid metal droplet preparation
Gallium (Ga shots, 99.999%, Indium Corporation, USA), Indium (In shots, 99.999%, Indium Corporation, USA), and Platinum (Pt beads, ≤3 mm, 99.999%, Indium Corporation, USA) were used as received for alloy preparation. Sodium hydroxide (NaOH pellets, Chem-Supply, Pty., Ltd) were used in extraction experiments. To prepare the liquid metal droplets, pre-weighed liquid metal solvents (Ga and In) were first melted and mixed in a glass vial on a hotplate at 300 °C for 2 h to create a homogeneous solvent. Pt beads (2% by wt) were dissolved into the liquid metal solvents on a hotplate at 500 °C for 6 h to ensure successful alloying was accomplished and homogeneous liquid alloys were obtained. The liquid metal droplets were prepared by drop-casting the liquid melts onto a silicon wafer. For fast-cooling process, the liquid metal droplets were left as is. For slow-cooling, the liquid metal droplet was placed into a LINKAM stage (LINKAM HFS600E-PB4) programmed to pre-condition at 500 °C for 1 h, followed by cooling from 500 °C to 20 °C at a rate of 1 °C per minute. All of the experiments were conducted in an ambient environment.
X-ray micro-CT scanning for liquid metal droplet
The Bruker X-ray Micro-CT Skyscan 2214 (micro-XCT) scanner was used for scanning the liquid metal droplet sample for analyzing the microstructure and spatial arrangement of the formed crystal within liquid metal bulk. The X-ray source was set to a beam energy of 140 kV with a current of 65 μA. A Cu 0.5 mm filter was used to minimize the effects of beam hardening. The samples were rotated over 180° during collection of the projections. The projections were then reconstructed with NRecon software (Version 2.2.0.6). CTVox (from the Bruker 3D.Suite, Version 3.3.1) software was utilized for visualizing the 3D reconstructed dataset, while the Avizo (Thermo Fisher Scientific, Version 2022.2) software was employed to analyze the size distribution of the formed crystals, which can be used for measuring the 3D length of the crystals by determining the maximum Feret diameter of each object. The samples were scanned at the practical limit of resolution for our XCT, given the dimensions of the droplets. As a result, the minimum resolvable feature was determined to be 25 µm or greater.
Extracting and washing the formed crystals
The extraction process utilized a 20 µm pore size nylon membrane filter (Merck Millipore), placed in a funnel assembled with a vacuum system to accelerate the filtration process and allow the liquid metal to diffuse through the membrane. The liquid metal alloy droplet was then deposited on the surface of the membrane and immersed in an excess of 1.0 M sodium hydroxide solution as an electrolyte. A positive gold electrode was connected to the liquid alloy, while the negative gold electrode was immersed in the electrolyte surrounding the liquid metal droplet. A DC voltage of +5 V was applied to the system using a programmable linear power supply (Singlet SPD 3303x) to reduce the surface tension of liquid metal alloy to near zero. The liquid metal passed through the membrane, leaving behind the unfiltered portion containing crystal formed on top of the membrane. The crystals were obtained for further washing process. The formed crystals from the extraction process were washed with 15 mL of 0.50 M NaOH to remove the excess Ga/EGaIn inclusions. Subsequently, the product was washed with Milli-Q water to remove excess NaOH. Finally, the crystals were collected for further characterization and utilization.
Theoretical calculations
DFT calculations were carried out using the Vienna ab initio simulation package (VASP, Version 6.4.2)59. The core electrons were treated utilizing the projector augmented wave method (PAW) method60 and the Perdew−Burke−Ernzerhof (PBE) exchange correlation functional was used61. The kinetic energy cutoff value for the wave functions was set to 550 eV.
Three systems were created to represent the different solvents used experimentally and to investigate their interaction with the Pt particles. In each cell, a layer of amorphous Ga, GaIn or In were placed on the top of a 4-layer Pt(001)-[4 × 4] slab. A (001) plane of Pt crystal was chosen to represent the Pt particles. The Pt(001) surface was cleaved from bulk Pt and expanded to a 4 × 4 supercell with optimized lattice parameters of 11.1 × 11.1 Å2. The upper amorphous Ga and In layers, each comprising 72 atoms, were generated using the Amorphous Cell module in Materials Studio 2022 and placed on the top of the Pt(001) surface. Similarly, the amorphous EGaIn system was constructed with 54 Ga atoms and 18 In atoms (atomic ratio of Ga/In = 3:1). A vacuum region greater than 15 Å in the direction perpendicular to the slabs was implemented to eliminate interactions between periodic cells. The three models, Ga(l)/Pt, EGaIn(l)/Pt and In(l)/Pt all had the same supercell dimensions of 11.1 × 11.1 × 40.0 Å3. All systems were sampled with 1 × 1 × 1 k-points in consideration of the efficiency, given the large number of electrons in Pt atoms and the dynamics of liquid metals. All atoms in the systems were relaxed until the Helman-Feynman force on each relaxed atom reached less than 0.03 eV Å−1 and the total energy was converged to 1 × 10−4 eV.
To simulate a fast-cooling process, an ab initio molecular dynamics simulation (as implemented in VASP) was performed for a simulation time of 10 ps (using a 2 fs timestep and a total of 5000 steps) using an NVT ensemble, and a Nose-Hoover thermostat62 to control the temperature, decreasing from 773 K to 293 K. A slow-cooling process was simulated by performing an AIMD simulation for 15 ps (using a 2 fs timestep for 7500 steps) and the same NVT ensemble and temperature control algorithm. The density profile analysis was obtained by extracting the distribution of Ga atoms along the z-axis from the last 500 frames of the AIMD trajectories. The MSD analysis was obtained from the last 800 frames of the AIMD trajectories.
Materials characterizations
The morphology and elemental distribution of the extracted crystals were characterized using a SEM/EDX instrument (JEOL InTouchScope, JSM-IT500HR, paired with a JEOL, Ex-74600U4L2Q model, EDX detector). ImageJ 1.8.0 software was used to quantify the pore size distribution in single crystal. The crystallographic structures and different phases of samples were characterized using the X’pert Multipurpose X-ray diffraction system (PANalytical, λ = 1.5418 Å, Cu-Kα radiation) with the acquisition parameters for voltage of 45 kV, current of 40 mA, and a step size of 0.02°.
Hydrogen evolution reaction studies
Sulfuric acid (H2SO4, 98% solution, Sigma-Aldrich), potassium hydroxide (KOH pellets, Chem-Supply, Pty., Ltd), a commercially available Pt/C catalyst (10 wt% Pt loading, Fuel Cell Store, USA), and Nafion solution (99.99%, Sigma-Aldrich) were used in electrochemical studies. Electrochemical measurements were carried out with a CHI760E workstation (CH Instruments Inc., USA) integrated with a rotating ring disk electrode apparatus (PINE Research, USA) coupled to a three-electrode cell. A glassy carbon electrode (GCE, 5 mm in diameter with a surface area of 0.1963 cm2) loaded with different catalysts was used as the working electrode. A graphite rod was used as the counter-electrode throughout the measurements, and a saturated calomel electrode (SCE) electrode was used as the reference electrode. Prior to the sample deposition, the GCE was finely polished with 0.05 m alumina oxide particles on a cloth and rinsed with Milli-Q water. To prepare the catalyst ink, the catalyst powder was uniformly dispersed in a mixture of 450 µL ethanol and 50 µL Nafion (0.5% wt.%). The suspension was sonicated for 30 min to achieve a homogenous catalyst ink. Subsequently, 10 μL of the catalyst ink was deposited onto a polished working electrode and dried at room temperature before testing. In this study, the catalyst powder for the sample consisted of 1 mg crystals, containing ~0.008 mg of Ga and 0.011 mg of Pt, while the control utilized 10 mg of 10% Pt/C, corresponding to 0.02 mg of Pt. Polarization (LSV) curves were carried out in 0.5 M H2SO4 (giving an electrolyte pH of 0) and 0.1 M KOH (giving an electrolyte pH of 13) which was degassed with nitrogen gas for 10 min prior to any measurement, at 5.0 mV/s. All polarization curves were automatically corrected for the iR contribution from the cell. The working electrode was rotated at a speed of 1600 rpm to alleviate the accumulation of evolved hydrogen bubbles on the GCE surface. The recorded potential was converted to a RHE according to the Nernst equation54. The electrochemical measurements were performed in a controlled laboratory environment of 25 °C average room temperature.
Data availability
Source data are provided with this paper. The crystallographic information was obtained from the Cambridge Crystallographic Data Centre (CCDC) under deposition numbers 1619886, 1661856, 1689832, 1725961, free of charge via https://www.ccdc.cam.ac.uk/structures/. Source data are provided with this paper.
References
Handschuh-Wang, S., Stadler, F. J. & Zhou, X. Critical Review on the Physical Properties of Gallium-Based Liquid Metals and Selected Pathways for Their Alteration. J. Phys. Chem. C. 125, 20113–20142 (2021).
Handschuh-Wang, S. et al. The Liquid Metal Age: A Transition From Hg to Ga. Adv. Mater. 36, 2408466 (2024).
Lu, Y. et al. Dynamic Leakage-Free Liquid Metals. Adv. Funct. Mater. 33, 2210961 (2023).
Kalantar-Zadeh, K., Rahim, M. A. & Tang, J. Low Melting Temperature Liquid Metals and Their Impacts on Physical Chemistry. Acc. Mater. Res. 2, 577–580 (2021).
Daeneke, T. et al. Liquid metals: fundamentals and applications in chemistry. Chem. Soc. Rev. 47, 4073–4111 (2018).
Tang, S.-Y., Tabor, C., Kalantar-Zadeh, K. & Dickey, M. D. Gallium Liquid Metal: The Devil’s Elixir. Annu. Rev. Mater. Res. 51, 381–408 (2021).
Zhou, Y., Li, N., Zhao, X. & Liu, J. Reversal Tuning of Liquid Metal Motor under Rotating Magnetic Field. Adv. Funct. Mater. 34, 2402832 (2024).
Wang, Y. et al. Biomimetic Liquid Metal Mechatronic Devices. Adv. Funct. Mater. 33, 2303312 (2023).
Chi, Y. et al. Liquid-Metal Solvents for Designing Hierarchical Nanoporous Metals at Low Temperatures. ACS Nano 17, 17070–17081 (2023).
Nguyen, C. K. et al. Instant-in-Air Liquid Metal Printed Ultrathin Tin Oxide for High-Performance Ammonia Sensors. Adv. Funct. Mater. 34, 2309342 (2023).
Jung, W. et al. Giant Decrease in Interfacial Energy of Liquid Metals by Native Oxides. Adv. Mater. 36, 2406783 (2024).
Zhang, M. et al. Bio-Inspired Differential Capillary Migration of Aqueous Liquid Metal Ink for Rapid Fabrication of High-Precision Monolayer and Multilayer Circuits. Adv. Funct. Mater. 33, 2215050 (2023).
Liu, H., Wang, K., Aasmundtveit, K. E. & Hoivik, N. Intermetallic Compound Formation Mechanisms for Cu-Sn Solid–Liquid Interdiffusion Bonding. J. Electron Mater. 41, 2453–2462 (2012).
Yuan, R. et al. Liquid Metal Memory. Adv. Mater. 36, 2309182 (2024).
Wang, D. et al. Liquid Metal Combinatorics toward Materials Discovery. Adv. Mater. 35, 2303533 (2023).
Khan, H. et al. Liquid metal-based synthesis of high performance monolayer SnS piezoelectric nanogenerators. Nat. Commun. 11, 3449 (2020).
Liu, S. et al. Intermetallic formation mechanisms and properties in room-temperature Ga soldering. J. Alloy. Compd. 826, 154221 (2020).
Xing, Z. et al. Liesegang Phenomenon of Liquid Metals on Au Film. Adv. Mater. 35, 2209392 (2023).
Lu, H. et al. Nanoengineering Liquid Metal Core–Shell Nanostructures. Adv. Funct. Mater. 34, 2311300 (2023).
Tang, J. et al. Dynamic configurations of metallic atoms in the liquid state for selective propylene synthesis. Nat. Nanotechnol. 19, 306–310 (2024).
Tang, J. et al. Oscillatory bifurcation patterns initiated by seeded surface solidification of liquid metals. Nat. Synth. 1, 158–169 (2022).
Steenbergen, K. G. & Gaston, N. A Two-Dimensional Liquid Structure Explains the Elevated Melting Temperatures of Gallium Nanoclusters. Nano Lett. 16, 21–26 (2016).
González, L. E. & González, D. J. Structure and dynamics of bulk liquid Ga and the liquid-vapor interface: Anab initiostudy. Phys. Rev. B Condens. Matter 77, 064202 (2008).
Chen, X. et al. Humidity-Responsive Liquid Metal Core-Shell Materials for Enduring Heat Retention and Insulation. Adv. Mater. 36, 2404705 (2024).
Idrus-Saidi, S. A. et al. Liquid metal synthesis solvents for metallic crystals. Science 378, 1118–1124 (2022).
Krishnamurthi, V. et al. A toolbox for investigating liquid metal systems. Cell. Rep. Phys. Sci. 5, 101820 (2024).
Khanzadeh, M. et al. Dual process monitoring of metal-based additive manufacturing using tensor decomposition of thermal image streams. Addit. Manuf. 23, 443–456 (2018).
Gobert, C. et al. Porosity segmentation in X-ray computed tomography scans of metal additively manufactured specimens with machine learning. Addit. Manuf. 36, 101460 (2020).
Blakey-Milner, B. et al. Metal additive manufacturing in aerospace: A review. Mater. Des. 209, 110008 (2021).
Rubtsov, V. et al. The Regularities of Metal Transfer by a Nickel-Based Superalloy Tool during Friction Stir Processing of a Titanium Alloy Produced by Wire-Feed Electron Beam Additive Manufacturing. Metals 14, 105 (2024).
Vidakis, N. et al. Interpretation of the optimization course of Silicon Nitride nano-powder content in biomedical resins for vat photopolymerization additive manufacturing. Ceram. Int. 50, 14919–14935 (2024).
Wu, C. et al. Machine learning-based design for additive manufacturing in biomedical engineering. Int. J. Mech. Sci. 266, 108828 (2024).
Behnsen, J. G., Black, K., Houghton, J. E. & Worden, R. H. A Review of Particle Size Analysis with X-ray CT. Materials 16, 1259 (2023).
Wang, Y. et al. Liquid state of post-transition metals for interfacial synthesis of two-dimensional materials. Appl. Phys. Rev. 9, 021306 (2022).
Brown, T. L. The Chemistry of Metallic Elements in the Ionosphere and Mesosphere. Chem. Rev. 73, 645 (1973).
Fatima, S. S. et al. Current state and future prospects of liquid metal catalysis. Nat. Catal. 6, 1131–1139 (2023).
Liu, J. et al. Size Dependent Phase Transformation of Liquid Gallium. Small 20, e2305798 (2024).
Steenbergen, K. G., Lambie, S. & Gaston, N. Discerning order from chaos: characterising the surface structure of liquid gallium. Mater. Horiz. 12, 1314–1322 (2025).
Lambie, S., Steenbergen, K. G. & Gaston, N. Resolving decades of debate: the surprising role of high-temperature covalency in the structure of liquid gallium. Mater. Horiz. 11, 4201–4206 (2024).
Li, J., Wang, J., Li, X. & Qin, J. Composition-dependent diffusion and viscosity behavior in liquid Ti–Al–Ni ternary alloys. J. Mater. Res. Technol. 33, 3864–3873 (2024).
Gunn G. Critical Metals Handbook (John Wiley & Sons, Ltd, 2014).
Ma, J. et al. Shaping a Soft Future: Patterning Liquid Metals. Adv. Mater. 35, 2205196 (2023).
Lechowski, B. et al. Laboratory X-ray Microscopy of 3D Nanostructures in the Hard X-ray Regime Enabled by a Combination of Multilayer X-ray Optics. Nanomaterials 14, 233 (2024).
Kalantar-Zadeh, K., Daeneke, T. & Tang, J. The atomic intelligence of liquid metals. Science 385, 372–373 (2024).
Galenko, P. K. & Ankudinov, V. Local non-equilibrium effect on the growth kinetics of crystals. Acta Mater. 168, 203–209 (2019).
Ma, K., Zhao, Y., Xu, X. & Hou, H. The effect of undercooling on growth velocity and microstructure of Ni95Cu5 alloys. J. Cryst. Growth 513, 30–37 (2019).
Kabalnov A. Ostwald Ripening and Related Phenomena. J. Dispers. Sci. 22, 1 (2001).
Wei, J. et al. Growth of CdZnTe crystal under high magnetic field and its photon-counting detector performance. J. Alloy. Compd. 1011, 178390 (2025).
Shen, L. Q. et al. Shear-band affected zone revealed by magnetic domains in a ferromagnetic metallic glass. Nat. Commun. 9, 4414 (2018).
Hu, Y. C. & Tanaka, H. Revealing the role of liquid preordering in crystallisation of supercooled liquids. Nat. Commun. 13, 4519 (2022).
Steenbergen, K. G., Lambie, S., Ruffman, C. & Gaston, N. Atomic-Scale Dynamics at the Interface of Doped Liquid Gallium: Contrasting Effects of Gallium Oxide and Vacuum. Small Sci. 5, 2500153 (2025).
Rahim, M. A. et al. Low-temperature liquid platinum catalyst. Nat. Chem. 14, 935–941 (2022).
Lambie, S., Steenbergen, K. G. & Gaston, N. Modulating the thermal and structural stability of gallenene via variation of atomistic thickness. Nanoscale Adv. 3, 499–507 (2021).
Niu, S., Li, S., Du, Y., Han, X. & Xu, P. How to Reliably Report the Overpotential of an Electrocatalyst. ACS Energy Lett. 5, 1083–1087 (2020).
Allen, J. Bard L. R. F. Kinetics of Electrode Reactions, 2 edn. (John Wiley & Sons, Inc., 2001).
Zhai, W. et al. Recent progress on the long-term stability of hydrogen evolution reaction electrocatalysts. InfoMat 4, 12357 (2022).
Wang, X. et al. Strategies for design of electrocatalysts for hydrogen evolution under alkaline conditions. Mater. Today 36, 125–138 (2020).
Allioux, F. M. et al. Atomic Dispersion via High-Entropy Liquid Metal Alloys. Small Struct. 5, 2400294 (2024).
Hafner, J. Ab-initio simulations of materials using VASP: Density-functional theory and beyond. J. Comput. Chem. 29, 2044–2078 (2008).
Kresse, G. & Furthmuller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B Condens. Matter 54, 11169–11186 (1996).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
Evans, D. J. & Holian, B. L. The Nose–Hoover thermostat. J. Chem. Phys. 83, 4069–4074 (1985).
Acknowledgements
This work was supported by the Australian Research Council (ARC) Laureate Fellowship grant (FL180100053/K.K.-Z.) and the ARC Discovery Project grant (DP230102813/ K.K.-Z. and DP240101215/M.J.S.S.). The authors also would like to acknowledge the technical and scientific assistance of the Sydney Microscopy & Microanalysis, the University of Sydney node of Microscopy Australia. The computational work was supported by the Australian Government’s National Collaborative Research Infrastructure Strategy (NCRIS), with access to computational resources provided by the National Computational Infrastructure Facility and the Pawsey Supercomputing Centre, through the National Computational Merit Allocation Scheme. The authors thank Fernando Valiente Dies for valuable discussions about the imaging software. The authors also thank Zeno Rizqi Ramadhan for his advice on electrochemistry.
Author information
Authors and Affiliations
Contributions
M.S.W. made the preliminary experimental observations. M.S.W. and K.K.-Z. conceived and designed the experiments. M.S.W. conducted the experiments, characterisations, analyzed the data with the assistance from M.F., F.-M.A., S.A.I.-S., M.K. (Mohamed Kilani), Z.P., P.K., M.K. (Minkyung Kang), S.-H.C., N.-A.N.-A., R.B.K. and J.T. The molecular dynamics simulations were performed by J.Z. C.R., N.G., M.J.S.S. and T.D. The first manuscript was drafted by M.S.W. and K.K.-Z. with inputs from all authors.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Qinglei Meng and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Widjajana, M.S., Foley, M., Zheng, J. et al. Observing growth of metallic crystals inside liquid metal solvents. Nat Commun 16, 10044 (2025). https://doi.org/10.1038/s41467-025-66249-y
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41467-025-66249-y
