Skip to main content
Advanced Science logoLink to Advanced Science
. 2026 Nov 15:e78790. Online ahead of print. doi: 10.1002/advs.78790

In Situ Multimodal Insights Into Hydration‐Driven MgO–Brucite Strength Build‐Up

Dan Meng 1, Mohammadreza Izadifar 2, Flemming Bjerg Grumsen 1, Gustavo Pinzón 3, Kasper Enemark‐Rasmussen 4, Adrian Alexander Schiefler 5, Eduardus Koenders 2, Julie Villanova 3, Jakob Birkedal Wagner 6, Rajmund Mokso 5, Carsten Kuenzel 7, Navid Ranjbar 1,✉
PMCID: PMC13771413  PMID: 43035974

ABSTRACT

Decarbonizing the built environment demands scalable alternatives to Portland cement. Magnesium oxide (MgO) underpins a transformative class of low‐carbon binders, yet existing hydration mechanisms are primarily based on studies of dilute equilibrium systems, failing to capture the spatial confinement and local thermodynamic boundaries. By coupling a temporal and multimodal in situ analytical suite with density functional theory, we uncover a multiscale, non‐equilibrium hydration paradigm spanning orders of magnitude in length scale. We show that localized water activity acts as a dynamic switch: sufficient liquid water drives direct conversion to brucite, whereas water‐deficient conditions metastabilize a transient amorphous hydroxide intermediate. Water exposure triggers rapid MgO expansion, driving crystallite reorientation and particle rearrangement that reduce nucleation site spacing and promote hydration. At the single‐particle scale, inward front propagation induces sub‐surface hydrogen diffusion, where the hydrated overlayer accumulates tensile stress and generates local structural instability, sustaining cyclic renewal until progressive shell thickening arrests conversion. Similarly, at the particle‐cluster level, hydration products nucleate preferentially at the cluster perimeter. At the mesoscale, compaction organizes these core–shell architectures into an interlocked and load‐bearing brucite network. These insights into non‐equilibrium phase transformation and the structuration mechanisms of hydrating MgO pave the way for the next generation of potentially carbon‐negative cementitious materials.

Keywords: brucite, cementitious, density functional theory, hydration reaction, in situ analysis, intermediate product, multimodality, reactive magnesia, sustained contraction


Hydrating MgO binders follow a non‐equilibrium, multiscale pathway. In this study, we used space‐time‐resolved correlative in situ multimodal characterization and atomistic‐level simulations to reveal how MgO‐to‐brucite transformation generates strength from the atomic to the macroscopic scale. Structural rigidity is governed by a discontinuous transition driven by morphological interlocking, which is temporally decoupled from bulk chemical conversion.

graphic file with name ADVS-9999-e78790-g004.webp

1. Introduction

Global production of Portland cement exceeds 4000 Mt annually and accounts for ∼7.5% of anthropogenic CO2 emissions [1], arising primarily from limestone calcination and high‐temperature (∼1450°C) clinker formation. The research for sustainable complementary alternatives has therefore intensified [2, 3, 4, 5]. Magnesium (Mg)‐based binders have emerged as promising alternatives to Portland cement, owing to their potential for low‐to‐negative CO2 emissions, lower alkalinity, and rapid setting kinetics [6, 7, 8]. Mg‐bearing silicate minerals, including olivine and serpentine, are abundant in regions surrounding the Circum‐Pacific and Mediterranean regions, providing vast fossil‐CO2‐free feedstocks (Figure 1a). Further, pilot‐scale initiatives such as PILCCU and MAGNEX have demonstrated potential extraction methods for a scalable production of magnesium hydroxide (Figure 1b), which can be converted to MgO at rather low temperatures [9]. The binders span a wide spectrum of mechanical properties, enabling a wide range of applications, from general‐purpose materials to ultra‐high‐performance concretes (Figure 1c, Data S1). However, the mechanistic link between the formation of specific hydrated phases and the nucleation and growth of their precursors and transient intermediates is poorly understood, posing a fundamental challenge for the large‐scale application of Mg‐based binders. Central to this is the hydration of MgO, which constitutes the initial step in the overall transformation sequence.

FIGURE 1.

FIGURE 1

(a) Worldwide distribution of ultramafic rocks (adapted from Ref. [22], Note: The boundaries shown on the map are provided solely for geographical reference and do not imply any position regarding legal territorial status); (b) Process route for Mg extraction from silicate minerals (adapted from Ref [22]); (c) 28‐day compressive strength for magnesium‐based binder systems (MC: magnesium carbonate, MSH: magnesium silicate hydrate, MO: magnesium oxychloride, MOS: magnesium oxysulfate, MP: magnesium phosphate; source data are provided in Data S1), Note: I, II, and III refer to general purpose, high strength, and ultra‐high performance concrete, respectively [23, 24]; (d) Equilibrium Mg2+ concentration as a function of pH in MgO‐Mg(OH)2 aqueous system (adapted from Ref. [25]); (e) Electron backscatter diffraction orientation map of MgO grains (adapted from Ref. [26]); (f) Correlation between MgO crystallite size, specific surface area and reactivity; refer to Data S2 for details (adapted from Ref. [27]).

MgO hydration is generally described by a coupled dissolution‐diffusion‐precipitation mechanism [10, 11]. Surface protonation of MgO produces MgOH+ and OH− species, resulting in an increase in pH that promotes desorption and diffusion of Mg2+. This is followed by precipitation of Mg(OH)2 once supersaturation is reached [12, 13, 14]. The overall process is regulated by a dynamic equilibrium between MgO dissolution and Mg(OH)2 precipitation as a function of pH (Figure 1d). Alternative pathways have also been proposed, involving transient intermediates such as anion‐coordinated polyhedral Mg(OH)6 4− growth units [15] or Mg(OH)2 nanosheet precursors [16]. These mechanistic frameworks are largely derived from dilute aqueous systems, where reactions proceed gradually toward equilibrium and are predominantly governed by chemical driving forces. In cementitious environments, however, hydration occurs under rigorously confined boundary constraints, including limited water availability, confined pore structures, and short reaction times, while simultaneously supporting the development of bulk mechanical strength. Under such conditions, hydration kinetics and resulting structural evolution are not governed solely by chemistry, but are strongly coupled with physicochemical factors, such as particle characteristics and evolving boundary conditions.

Structurally, MgO particles are polycrystalline and composed of densely packed crystallites with heterogeneous orientations (Figure 1e). Hydration kinetics are sensitive to the thermal processing history, with light‐burnt MgO, calcined at ∼600°C–1000°C, undergoing rapid hydration owing to its high surface area and fine crystallite size, whereas hard‐ and dead‐burnt MgO (> 1000°C) exhibit substantially lower reactivity as a result of sintering and densification [17] (Figure 1f). Theoretically, hydration of MgO to Mg(OH)2 phases is accompanied by a volume increase (11.3 vs. 24.6 cm3mol−1 for MgO and Mg(OH)2, respectively) [18, 19]. This expansion arises from displacive crystal growth and pore formation, which contributes to bulk strength development through crystallization‐induced stress generation [20]. The extent of expansion strongly depends on particle reactivity and the degree of mechanical confinement. Elevated confining pressure or restricted fluid supply suppresses bulk expansion, making the behavior particularly sensitive to mixture design and pore‐scale transport [20]. The efficiency with which crystallization generates expansive stress depends on multiple coupled factors. First, the formation of crystallization‐inducing phases must be thermodynamically favorable under the chemical, pressure, and temperature conditions. Second, displacive crystal growth must occur within an optimal hydration window; premature crystallization in a paste‐like matrix limits stress transmission owing to the absence of a load‐bearing network, whereas post‐setting crystallization may induce structural damage such as cracking [12]. Third, water activity fundamentally governs both hydration kinetics and microstructural evolution [21].

Here, we uncover the mechanistic coupling between phase formation, microstructural evolution, and macroscopic strength development during MgO hydration through multimodal and multiscale characterization across particle‐to‐bulk length scales.

2. Results and Discussion

2.1. Water Activity‐Controlled Phase Evolution During MgO Hydration

In situ X‐ray diffraction (XRD) was employed to resolve the phase evolution of MgO during hydration under varying water exposure conditions, with particular emphasis on the governing role of water activity in cement‐like systems where water availability is inherently limited. MgO hydration was monitored under three conditions: MgO exposed to water vapor, water‐retentive (i.e., water‐saturated sealed environment), and moisture‐exchanging (i.e., sample openly exposed to 95% humidity environment). A H2O:MgO mass ratio of 1.0 was used for both the water‐retentive and moisture‐exchanging systems, while higher ratios of 1.2 and 1.4 were additionally used in the moisture‐exchanging system. These ratios ensure sufficient flowability while avoiding segregation.

As shown in Figure 2a, MgO exhibits no detectable phase evolution after 24 h exposure to water vapor, whereas partial transformation to brucite occurs upon direct contact with liquid water (Figure 2b,c), demonstrating that liquid water is required to initiate MgO hydration on timescales relevant for cementitious systems. In both moisture‐exchanging and water‐retentive systems, two X‐ray amorphous phases were observed in addition to the crystalline phases. A broad hump at 20°–35° (2θ), commonly associated with water [28], is progressively reduced over time, most likely due to evaporation and consumption during brucite formation. A second diffuse reflection is observed at ∼15.7° (2θ), adjacent to the (001) reflection of brucite (Figure 2b and Figure S1b), and could originate from poorly crystalline or amorphous hydrated intermediates. Similar features in the 6.1°–16.8° (2θ) range have been reported in MgO‐hydromagnesite systems [29, 30, 31]. The presence of amorphous phases is further supported by in situ transmission electron microscopy (TEM) electron diffraction (Figure 2d), which shows a transition from a well‐defined crystalline MgO spot pattern to a diffuse halo upon hydration. Moreover, two inner diffraction rings exhibit increased d‐spacings, shifting from 2.11 and 2.43 Å (MgO) to 2.26 and 2.58 Å after hydration; these values are inconsistent with either MgO or brucite, suggesting the formation of transient hydrated intermediates or structurally disordered phases. The residual faint rings indicate the persistence of crystalline domains, with the most distinct outer ring corresponding to the (220) plane of MgO. The accelerated hydration, compared to the XRD observation of bulk material, is attributed to the substantially smaller particle size examined with TEM (an order of magnitude lower), which increases reaction kinetics. This interpretation is supported by XRD analysis, which reveals an increase in average MgO crystallite size upon hydration (Figure S1c,d), indicating preferential reaction of finer crystallites and persistence of coarser ones.

FIGURE 2.

FIGURE 2

(a) In situ XRD results of dry MgO under moisture vapor condition; (b) MgO paste with H2O:MgO = 1, under moisture‐exchanging conditions; (c) MgO paste with H2O:MgO = 1 under water‐retentive condition, followed by drying and rewetting; (d) In situ TEM electron diffraction patterns; (e) Phase evolution of MgO pastes with H2O:MgO = 1.0 (f) 1.2 and (g) 1.4 under moisture‐exchanging condition; (h) Phase evolution of MgO paste with H2O:MgO = 1.0 under water‐retentive condition, followed by drying and rewetting; (i) Comparison between XRD‐derived and estimated brucite contents; (j) Time‐resolved brucite‐to‐MgO ratio.

Quantitative XRD analysis (Figure 2e–g) reveals delayed brucite formation in the moisture‐exchanging system, initiating at ∼7 h and reaching ∼ 20 wt.% after 24 h. In contrast, water‐retentive conditions accelerate the transformation, with brucite appearing at ∼4 h and attaining ∼40% by 24 h (Figure 2h). In the water‐retentive system, MgO consumption closely follows brucite formation, indicating near‐complete conversion (i.e., without a detectable transient phase). Such a trend is not observed for the moisture‐exchanging system, where the theoretical water content, excluding the stoichiometric amount required for brucite, is lower than the amorphous fraction (Figure 2e–g). This indicates the presence of ∼10 wt.% amorphous hydration products and reflects that conversion of MgO to brucite is not complete. This interpretation is further corroborated by comparison with the theoretical brucite yield assuming complete MgO conversion, which exceeds the brucite content quantified by XRD (Figure 2i).

Time‐resolved brucite‐to‐MgO trends (Figure 2j) show that sustained water availability promotes brucite formation, although the conversion rate decreases with time. This deceleration, most evident under water‐retentive conditions, is consistent with surface passivation of MgO by brucite, indicative of a self‐limiting hydration mechanism [31]. To further study this behavior, partially hydrated MgO from the water‐retentive system was subjected to subsequent drying and rewetting. However, no specific change to the MgO‐to‐brucite transformation trajectory was observed (Figure 2h,j). This indicates that the reaction tends to become kinetically constrained beyond a certain degree of conversion upon drying. This can be associated with the alteration in pore structure, medium transport, and particle contact states that may not be reversible after rewetting. Nevertheless, continued brucite formation was measured after two weeks of hydration under conditions with sufficient water availability, with higher H2O:MgO ratios resulting in greater brucite formation (Figure 2j).

2.2. Particle‐Level Pathway of Hydration and Morphological Evolution

While XRD provided a quantitative phase analysis of the hydration products, understanding the underlying reaction mechanisms requires spatially resolved characterization of phase distributions, nucleation sites, and the interplay between particle morphology and local reaction environments. The as‐received MgO (Figure 3a) exhibited a broad particle size distribution (d50 = 4.4 µm; d10 = 1.6 µm; d90 = 18.1 µm) and an agglomeration ratio of ∼4.3, consistent with substantial particle agglomeration [32, 33]. Ultrasonication in ethanol for 2 min reduced d50 to ∼2 µm and disintegrated large agglomerates (10–100 µm), indicating predominantly weak physical interparticle interactions (Figure 3b). Complementary dispersion protocols further showed strongly shear‐dependent agglomerate breakdown (Figure 3b i–vi), revealing a polydisperse particle ensemble in which accessible surface area is strongly coupled to local hydration pathways.

FIGURE 3.

FIGURE 3

(a) BSE images of unhydrated MgO particles; (b) Particle size distribution of unhydrated and hydrated MgO particles under different treatment conditions; (c) In situ TEM bright‐field images of MgO particles before and after water contact; (d) Line profiles extracted along the line indicated in Figure (c); (e) BSE images of MgO after 24 h hydration and (f) Microcracking at particle edge; (g) Raman‐SEM overlay image of MgO after 24 h hydration; (h) TEM images of unhydrated MgO and (j) brucite: (i) bright‐field imaging, (ii) electron diffraction pattern, (iii, iv) dark‐field imaging.

This intrinsic heterogeneity renders hydration inherently size‐dependent [34]. To resolve the corresponding particle‐scale dynamics, in situ liquid‐cell TEM experiments were conducted within nanochannels (3 µm × 300 nm). Following water injection, bright‐field imaging revealed a loss of distinct particle boundaries (Figure 3c), likely arising from water‐induced surface reactions, consistent with the accompanying changes in d‐spacing observed by electron diffraction (Figure 2d). Although line profile analysis suggested a slight reduction in particle dimensions (Figure 3d), the present observations do not conclusively distinguish between particle shrinkage or expansion during hydration. Previous in situ TEM observations [35] reported particle expansion; however, this discrepancy likely arises from differences in MgO reactivity. In contrast to this earlier report [35], the MgO studied here remained unreactive under comparable vapor exposure conditions (Figure 2a). Semi in situ TEM experiments nevertheless indicated the possibility of particle aggregation and expansion (Figure S4).

Owing to TEM's size limitations, morphological evolution in larger particles (>d10) was examined using cross‐sectional Backscattered electron (BSE) imaging in scanning electron microscopy (SEM). Needle‐like hydration products were observed at particle edges (Figure 3e), consistent with enhanced water accessibility and surface reactivity. Because MgO and brucite exhibit limited contrast in both morphology (Figure 3h–j) and elemental analysis, correlative Raman‐SEM imaging of hydrated MgO cross‐sections (Figure 3g, Figure S6) was used to spatially resolve the hydration products, revealing brucite at the outer surface layer. In several particles, localized microcracks were additionally observed near the particle edges (Figure 3f, Figure S5), suggesting hydration‐induced stress arising from volumetric expansion and crystal growth pressure exceeding the tensile strength of the reacted shell [36]. Such localized shell failure may facilitate renewed water ingress to the reacted MgO core, thereby sustaining continued hydration.

2.3. Mesoscale Evolution and Structural Build‐Up

Building on the particle‐resolved observations, the heterogeneous and size‐dependent hydration behavior indicates that local transformations do not occur in isolation but instead propagate through interparticle interactions. Surface nucleation, particle expansion, and product growth at individual particles continuously perturb the surrounding micro‐environment, altering local transport conditions and reactive interfaces. Hydration therefore evolves as a collective process, in which neighboring particles are both mechanically and chemically coupled. Resolving this transition from particle‐scale dynamics to mesoscale is essential to connect phase evolution with emergent reactivity and the development of bulk mechanical strength.

Particle size evolution during hydration was probed with laser diffraction. Both as‐prepared and ultrasonicated samples were measured to assess flocculation and agglomeration stability (Figure 4a). During the early hydration stages, particles in the 10–100 µm range were detected only in the as‐prepared samples and disappeared following ultrasonication up to 3.5 h. The fraction dispersed upon ultrasonication predominantly consisted of submicron particles (< 1 µm), and likely consisted of initially formed hydrated phases with low mechanical integrity [37], displaced by ultrasonic forces through a peeling‐off mechanism [38]. By ∼5 h, however, larger particles (10–100 µm) persisted after ultrasonication and increased in abundance, indicating the formation of mechanically stronger hydrated products capable of aggregation or growth (Figure S7). This transition coincided with the onset of brucite formation observed by XRD (∼4–7 h).

FIGURE 4.

FIGURE 4

(a) Evolution of particle size distribution of MgO during the first 7 h of hydration; (b) BSE images of aggregated hydrated MgO particles after 24 h; (c) In situ synchrotron nano XCT images showing the temporal evolution of hydrated MgO particles; (d) Volume expansion quantified by rheological characterization and Raman microscopy; (e) Raman spectra of hydrated MgO; (f) In situ Raman mapping during MgO hydration; (g) Normalized 1H NMR results of MgO pastes with H2O:MgO = 1.0 and 1.4, schematic of hydration and pore refinement in i); (h) Ultrasonic pulse velocity evolution of MgO pastes and rheological characterization of storage modulus.

Aggregate formation was further investigated by in situ synchrotron X‐ray computed nano tomography (Figure 4c, Figure S9). Upon exposure to water, MgO particles rapidly dispersed throughout the aqueous phase and brought nucleation sites into closer proximity, driven by colloidal inter‐particle interactions, including van der Waals attraction, electrostatic forces, and steric repulsion [39], before stabilizing. As water was gradually consumed or evaporated, gradients in local pore water distribution resulted in capillary pressures that compact the assembly of hydrating MgO particles [40]. After about 4 h, a distinct shell‐like boundary started to form. Simultaneously, interparticle bridges formed through interlocking of the newly formed magnesium hydroxide, establishing connections between neighboring particles (Figure 4c i–iii). Corresponding morphological evolution was observed at higher resolution in BSE cross‐section imaging (Figure 4b), where densely integrated particles at the outer shell enclosed a cluster of loosely packed particles. This aggregation process was accompanied by pronounced particle‐level expansion, contributing to the growth of large hydrate assemblies. At the mesoscale, MgO showed substantial bulk expansion upon exposure to excess water. Most of the expansion occurred immediately after hydration and continued within the first 4 h, after which volumetric changes largely plateaued, as shown with both rheometry and confocal Raman microscopy (Figure 4d). Here, the discrepancies in absolute expansion magnitude measured by Raman and rheometry arise from differences in measurement onset and experimental conditions. Interestingly, this expansion preceded the detection of brucite by Raman spectroscopy (Figure 4e,f). Noteworthy, it is known that the conversion of MgO to brucite involves a large increase in the molar volume of the solid phase, while, based on our observation, crystallographic‐scale volume changes do not necessarily translate into macroscopic expansion, as the solid‐volume increase may be accommodated by pore filling or microstructural rearrangement.

To elucidate the densification mechanism underlying structural build‐up, we probed pore network evolution, central to water transport, capillary stresses, and hydrate accommodation, using 1H Nuclear Magnetic Resonance (NMR) transverse relaxation measurements. The transverse magnetization decay (T2 relaxation) of water during hydration was monitored with the CPMG method. A relatively long echo delay (250 µs) rendered the experiment selective toward hydrogen species with T2 relaxation in the millisecond range, while chemically bound hydrogens with much shorter relaxation times did not contribute to the NMR signal. The resulting T2 distributions therefore primarily reflect pore water and track changes in pore size during hydration. As hydration proceeds, the main relaxation peaks progressively shift to shorter relaxation times (Figure 4g), indicating continuous pore refinement [41]. This reflects the growth of hydration products, progressively densifying the initial pore structure. Around the onset of brucite formation, a secondary peak emerges in the T2 distribution, revealing a distinct population of finer pores, likely associated with water confined within interlocking brucite networks. At higher water content (H2O:MgO = 1.4), a third peak appears at even shorter relaxation times (Figure 4g), indicating further pore refinement within more densely packed brucite assemblies, consistent with the enhanced brucite formation observed by XRD (see schematic in Figure 4g i). At the bulk scale, the formation and interlocking of brucite are reflected in the evolution of ultrasound propagation velocity (Figure 4h), indicating increasing solid‐phase connectivity. As hydration proceeds, velocities increase from values dominated by fluid‐ and air‐filled pores toward those characteristic of solid MgO [42] and brucite [43] (∼4 to > 10 km s−1). Initial velocities remain below that of water (∼1490 m s−1) due to air entrapment (∼340 m s−1), consistent with early‐stage pore discontinuity and observations in related cementitious systems [44, 45, 46]. Increasing water content accelerates the rate of velocity increase, indicating enhanced hydration kinetics. However, absolute velocities remain lower due to interparticle spacing and residual porosity. These microstructural signatures directly translate into mechanical response. The evolution of the storage modulus (Figure 4h) reveals minimal stiffness development during the early stage of the major volumetric expansion, indicating that particle rearrangement and free volumetric growth do not contribute significantly to load‐bearing capacity. In contrast, the onset of brucite formation coincides with a sharp (∼two orders of magnitude) increase in modulus, marking the transition from a loosely packed assembly to a mechanically integrated network through the coupled effects of particle agglomeration, capillary compaction, microstructural rearrangement, and interparticle interlocking. Subsequent deceleration in modulus growth follows the reduced rate of brucite formation, indicating that strengthening remains hydration‐controlled under water‐retained conditions.

2.4. Atomistic Insights Into Multiscale Hydration and Brucite Formation

To rationalize the experimentally observed multiscale evolution, MgO hydration can be described as a sequence of coupled atomistic and interfacial processes governing the transition from surface reactivity to bulk structural development. Figure 5a schematically illustrates three major steps that occur concurrently throughout the system: (i) upon exposure to water, hydration is initiated by rapid surface water adsorption and ingress through intercrystallite gaps [47], leading to bulk volumetric expansion and crystallite rearrangement; (ii) surface protonation and Mg‐OH formation, which destabilize the MgO lattice and enable Mg2+ release; (iii) ordered aggregation of Mg‐OH species drives nucleation and growth of brucite.

  • Step (i): The polycrystalline nature of MgO exposes a distribution of crystallographic facets (e.g., (100), (110), and (111)) with distinct surface energetics, giving rise to facet‐dependent water adsorption and interfacial hydroxylation. Density functional theory (DFT) calculations indicate that adsorption extends beyond localized sites into a stabilized interfacial regime, promoting formation of a hydrogen‐bonded hydroxylated layer prior to convergence toward bulk‐like water energetics (Figure 5b). Among exposed facets, the O‐terminated MgO (111) surface exhibits particularly strong exothermic interaction, with a calculated binding energy of −3.36 eV per proton. This reflects the high basicity of surface oxygen sites and intrinsic polar instability of this termination [48]. The resulting energetically heterogeneous yet cooperative adsorption promotes the formation of a percolating hydration front across crystallite interfaces (Figure 5c). This provides the thermodynamic driving force for collective water uptake, crystallite reorientation, and the onset of bulk volumetric expansion observed by confocal Raman microscopy and rheometry within the first few hours (Figure 4d).

  • Step (ii) and Step (iii): Progressive hydroxylation of the MgO surface destabilizes the lattice, initiating Mg dissolution via the formation and migration of Mg‐OH species [49]. Climbing‐image nudged elastic band calculations delineate the kinetic landscape (Figure 5d) associated with successive Mg‐extraction events. These calculated energies should not be interpreted as a cumulative free‐energy profile for the overall MgO hydration reaction. The binding energies characterize the relative stabilization of individual Mg‐containing configurations, whereas the cNEB energy differences represent activation barriers for the corresponding elementary extraction events. For the selected pathways, early surface Mg extraction proceeds over moderate barriers (∼1.5 eV), whereas extraction of subsurface Mg after formation of a hydrated overlayer becomes substantially more difficult, reaching 4.68 eV for Mg4 (Figure 5e,f). Although local vacancy formation can transiently stabilize a neighboring configuration, such as Mg5 (binding energy ≈ −1.23 eV), subsequent Mg extraction remains kinetically activated. Thus, the overall thermodynamic driving force for hydration is compatible with locally uphill elementary steps, although the present calculations do not establish that the selected energy changes correspond to elementary steps along the actual hydration pathway or exclude alternative, lower‐energy pathways. This behavior is consistent with the experimental observations and the intrinsically low dissolution rate of MgO (10−12 mol cm−2s−1) [50, 51].

FIGURE 5.

FIGURE 5

(a) Schematic atomistic overview of MgO hydration; (b) Adsorption energy as a function of distance from MgO surface; (c) Water adsorption configurations on different MgO surfaces; (d) Kinetic barrier along minimum energy path; (e) Binding energies of selected Mg‐containing configurations and cNEB activation barriers for the corresponding Mg‐extraction pathways in the hydrated MgO surface models; (f) Activation energies for interfacial hydration and barrier‐limited hydration stages; (g) Bulk modulus of MgO, Mg(OH)2, and intermediates (adapted from Ref. [52]); (h) Ideal tensile response of defect‐free Mg(OH)2 under uniaxial loading; (i) Schematic of atomic configuration at critical stress conditions.

Concomitantly, continued hydroxylation beneath this overlayer, sustained by hydrogen diffusion, induces volumetric expansion and generates tensile stress within the passivation layer. This effect is amplified by the elastic mismatch between compliant Mg(OH)2 (∼60 GPa) and stiffer MgO (∼160 GPa) (Figure 5g) [52]. DFT tensile calculations (Figure 5h,i) provide an atomistic reference for the tensile stability of defect‐free brucite. The reduction in stress beyond the maximum tensile stress in Figure 5h is accompanied by structural instability and atomic rearrangement. Although this idealized calculation does not directly reproduce fracture of the heterogeneous brucite‐rich shell, the experimentally observed localized microcracking indicates that the reacted layer undergoes local cracking during hydration. Together, these results support our interpretation that hydration‐induced volumetric expansion locally destabilizes and cracks the reacted layer, facilitating renewed water ingress and continued hydration. As dissolution‐precipitation proceeds, interfacial reorganization drives brucite nucleation via a topotactic transformation pathway, facilitated by crystallographic registry between MgO (111) planes and Mg(OH)2 (0001) planes [53]. Initially disordered Mg‐OH species progressively reorganize into ordered hydroxide layers, forming brucite nuclei that subsequently coalesce into crystalline aggregates [13].

This cyclic interplay between passivation and mechanically induced rupture sustains hydration and progressively builds a thicker, increasingly transport‐limiting shell. This is evidenced by rim‐localized brucite formation in Raman‐SEM and BSE imaging (Figures 3g and 4b). At larger scales, these brucite crystallites nucleate and interlock at particle contacts and along shell‐like boundaries around loosely packed assemblies, establishing load‐bearing connectivity and driving mechanical strength development.

The atomistic picture developed here is consistent with previous computational descriptions of MgO hydration, in which water adsorption and dissociation promote surface hydroxylation and Mg dissolution, followed by Mg(OH)2 formation through a dissolution–precipitation pathway. Previous experimental and mechanistic studies have likewise described MgO hydration as a dissolution‐diffusion‐precipitation process and have shown that Mg(OH)2 formed at or near the MgO surface can remain associated with the particle and progressively influence further hydration [12, 15]. Recent computational studies provide direct molecular‐level evidence that Mg2+ dissolution is coupled to dissociative water adsorption and proton penetration, followed by Mg(OH)2 nucleation and growth through dissolution and precipitation [54, 55]. The present study extends this molecular picture by linking the local surface chemistry to the evolving physical environment of hydration. Specifically, progressive formation of a hydrated overlayer increases the local kinetic resistance to Mg extraction, while volumetric expansion, local shell destabilization and interparticle reorganization modify transport pathways and reaction accessibility. The resulting mechanism therefore connects surface chemistry and particle‐scale evolution with mesoscale structural development.

3. Conclusion and Outlook

In situ multimodal characterization, combined with atomistic modeling, elucidates MgO hydration under non‐equilibrium conditions across multiple length scales. Upon exposure to water, MgO undergoes rapid initial expansion, accompanied by water ingress through intercrystallite gaps. This drives crystallite‐scale reorientation and mesoscale particle rearrangement, bringing nucleation sites into closer proximity and thereby facilitating subsequent hydration. The resulting hydrated products precipitate on crystal and particle surfaces and along the perimeters of particle clusters, initiating strength development. Brucite formation refines pore structure and promotes interparticle interlocking, giving rise to a load‐bearing network. These coupled processes coincide with a pronounced increase in mechanical stiffness. Time‐resolved X‐ray diffraction shows that high water availability favors direct conversion of MgO to brucite, whereas limited water slows down the transformation and renders intermediate hydrated phases more measurable. Moreover, the hydration products form a barrier that suppresses further reaction, resulting in a self‐limiting hydration mechanism. Collectively, these findings establish a mechanistic link between MgO hydration, microstructural evolution, and strength development.

As hydration often defines the primary trajectory of phase assemblage, the mechanistic insights established here provide a transferable framework linking MgO hydration chemistry to microstructural densification and structural build‐up in broader magnesium‐based systems, including magnesium silicate hydrate and hydroxycarbonate binders. In these systems, hydration is inherently coupled with concurrent processes such as in situ and ex situ carbonation or dissolution‐reprecipitation‐driven magnesium silicate hydrate (MSH) formation. Crucially, resolving the temporal sequence of phase and morphological evolution enables targeted intervention. For example, organic additives and inorganic seeds can be deployed to modulate interfacial kinetics and synchronize Mg release with secondary phase nucleation, thereby tuning functionalities like setting time, CO2 uptake, volumetric stability, and mechanical strength under realistic exposure conditions. More broadly, this framework establishes a pathway toward predictive, multiscale design of next‐generation sustainable cementitious materials with reduced environmental footprints and programmable functionality.

4. Methods

4.1. Raw Materials and Initial Characterization

MgO used in this study was supplied by Hebei Messi Biology Co., Ltd., light‐burnt at 700°C–800°C, with a d50 of 4.4 µm, bulk density of 320 kg/m3, citric acid activity of 10–15s, and high‐purity grade (97 wt.% MgO and 3 wt.% volatile materials as per X‐ray fluorescence analysis).

4.2. Physical and Morphological Analysis

PSD of MgO particles was measured using a Malvern 3000 laser diffraction particle size analyzer (Malvern Instruments Ltd., UK) in wet mode in water‐free denatured ethanol. Prior to each measurement, the powder was dispersed in ethanol with and without ultrasonication. For ultrasonicated samples, dispersion was aided using a Branson Sonifier 250 for 60 s prior to analysis. The test was performed on unhydrated and hydrated MgO particles that were manually mixed at H2O:MgO = 1 for 2 min and subjected to different hydration durations up to 7 h. Particle sizes were measured by comparing the light scattering profiles with Mie theory, using a refractive index for Mg(OH)2 = 1.57 and an absorption index = 0.01.

The agglomeration ratio of raw MgO was determined based on the ratio of the primary particle size, derived from the BET surface area measurement (GBET), to the crystallite size (28.5 nm) obtained from the Scherrer equation [56]. GBET was calculated as GBET = F/ρS [57], where F = 6 is a particle shape factor, ρ = 3595 kg/m3 is the theoretical density of MgO, and S = 11.6 m2/g is the specific surface area, yielding a primary particle size of 122.6. BET was measured using an Autosorb‐iQ analyzer, Quantachrome Instruments USA.

4.3. X‐Ray Diffraction Analysis (XRD)

In situ XRD was performed in Bragg‐Brentano geometry on a Bruker D8 Discover diffractometer equipped with a beryllium‐dome temperature stage (MRI, BTS‐Basic) using Cu‐Kα radiation (λ = 1.5406 Å) at 40 kV and 40 mA. Data were collected over 12°–64° (2θ) with a 0.019° scan step and 0.33 s per step. Measurements started ∼30 min after sample preparation and continued for 24 h with 15 min intervals. The chamber atmosphere was maintained 95% relative humidity (RH) and 21°C using a Cellkraft Humidifier P‐2S. Three sets of experiments were conducted: (i) dry MgO exposed to 95% RH water vapor; (ii) MgO premixed with water at H2O:MgO mass ratios of 1, 1.2 and 1.4 under 95% RH, referred to as moisture‐exchanging system; (iii) MgO premixed with H2O:MgO = 1, sealed with 6 µm Kapton film for 24 h, referred to as water‐retentive system, then unsealed and exposed to 95% RH for 24 h, followed by surface water replenishment to compensate for drying water loss and monitored for another 24 h. Here, anatase (titanium (IV) oxide powder, 99.8% trace metals basis, CAS Number 1317‐70‐0, Sigma–Aldrich) was measured on the same day as an external standard. Water loss was corrected by pre‐ and post‐ measurement mass comparison, consistent with mass loss trends in a climate chamber under identical exposure conditions. Quantitative phase evolution was determined by Rietveld refinement using HighScore Plus V4.9 (Malvern Panalytical) with reference structures for periclase (COD 96‐900‐6750) and brucite (COD 96‐100‐0055) from Crystallography Open Database. The weighted R profile (Rwp ) value from fitting was generally below 15.

4.4. Transmission Electron Microscopy (TEM)

In situ TEM was performed on a JEOL 2100 using four‐inlet silicon nitride membrane nanochannel liquid cells (Insight Chips, 3 µm width, 300 nm height) connected to a four‐inlet flow TEM holder. Due to dimensional constraints, MgO was deagglomerated using a vibratory micromill (Pulverisette 0, Fritsch), dispersed in ≥ 99.9% absolute ethanol, filtered through a 0.45 µm polytetrafluoroethylene syringe filter, and ultrasonicated for 60 min. The suspension was loaded into the nanochannels and left to dry. The nanochannels remain hermetically sealed until punctured for water injection. The nanochannel before and after MgO loading is shown in Figure S2. Water was slowly injected into the nanochannels through the inlets. Changes in morphology and crystallinity of MgO samples upon water addition were monitored via bright‐field imaging (Video S1) and electron diffraction imaging (Video S2). Ex situ TEM was conducted to characterize the morphology and crystallinity of unhydrated MgO and brucite formed after 2 weeks of hydration.

Environmental TEM (ETEM) was conducted using a Titan E‐Cell 80–300ST TEM (FEI Company, Hillsboro, OR) operated at 300 kV. Images were acquired using a OneView CCD camera controlled with Digital Micrograph (Gatan, Pleasanton, USA). MgO particles were ultrasonically dispersed in 99% ethanol for 30 s, and 5 µL of the suspension was deposited onto holey carbon films supported on 300 mesh Cu grids (AGS147, Agar Scientific) and allowed to settle for 2 min. The sample was affixed using a Gatan double tilt sample holder. Initial screening was first performed to locate and annotate the number of unhydrated MgO particles across the grid. Morphology and crystalline structure were characterized using multiscale bright‐field, electron diffraction and dark‐field imaging at 10−4 mbar. For post‐hydration, the TEM grid was removed from the microscope, and the particles were exposed to 5 µL deionized water atop the surface for 2 min (Figure S3c). Residual water on the opposite side of the TEM grid was blotted. The evolution of the annotated particles was monitored under a flow of H2O vapor. Later, the samples were kept at 95% RH for 24 h, after which the same particles were examined. Beam‐induced damage was mitigated by minimizing electron dose and restricting exposure time to 1 s for bright field and 0.125 s for diffraction imaging.

4.5. Scanning Electron Microscope (SEM)

The microstructure of hydrated MgO paste (H2O:MgO = 1) at different hydration times was examined using a Zeiss Super SEM operated at 15–20 kV with a working distance of 5.5–10 mm in backscattered electron mode. Samples were embedded in epoxy, cured, and ground using SiC papers of decreasing grit size from 30 to 3–5 µm, and polished using 3 and 1 µm diamond paste, with absolute ethanol as grinding medium.

4.6. Confocal Raman Microscopy (CRM) & Correlative Raman Imaging Scanning Electron Microscopy (RISE)

In situ CRM was conducted to monitor the evolution of MgO hydration over 12 h in water using an alpha300R Apyron confocal Raman microscope equipped with a 457 nm laser. The laser, with a power of 18 mW, was focused via a 63X microscope objective lens into a spot area of 50 × 50 µm. RISE imaging of 24 h hydrated MgO particles was conducted using ZEISS Sigma 300 RISE (WITec, Ulm), integrating CRM with scanning electron microscopy ZEISS Sigma 300. This technique enables spatially correlated 2D Raman‐SEM overlays for brucite phase identification. Raman spectra were acquired using a 532 nm fiber‐coupled laser operating at 54 mW, focused through a 100× objective (NA = 0.75) over a 35 × 35 µm2 area with 70 × 70 pixel resolution and 0.5s acquisition per pixel. SEM imaging was performed under low vacuum conditions at 15 kV acceleration voltage and 300 pA beam current.

4.7. Synchrotron X‐Ray Computed Nano‐Tomography (Syn‐nXCT)

MgO particles smaller than 10 µm were loaded into a fused silica capillary tubing (Molex, Polymicro) with an inner/outer diameter (ID/OD) of 100/238 µm (Figure S8a). The MgO particles were spatially confined and connected to an on‐stage DC pump that allowed vibration‐free injection through wireless control (Figure S9a,b). Two regions were selected for acquisition, including the upper region adjacent to the finer tubing and the lower region near the melt glass beads. Acquisition was first performed on dry MgO particles. Water was then pumped into the capillary at 5 V DC for 15 min, followed by a 5‐min stabilization period before initiating the first scan. Phase‐contrast Holo‐tomography was performed at the ID16B [58] beamline of ESRF using a 50 × 50 nm2 (horizontal × vertical full width at half‐maximum, FWHM) cone beam with high flux 1.47 × 1012 ph s−1 at the energy of 29.6 keV (ΔE/E ≈ 10−2) and a pixel size of 35 nm. Holo‐tomography involved scanning the specimen at four different source‐sample distances. For each distance, a full tomogram was acquired over a 360° rotation with an exposure time of 0.02s per projection, resulting in an acquisition time of 4 min per single scan. The total acquisition time for the whole holotomography process took around 20 min, considering motor movements and stabilization time. The reconstruction procedure consisted of projection alignment, phase retrieval with the PyNX software of ESRF [59] using an iterative transport of intensity based method and a delta/beta of 150, and a filter back projection step performed using the ESRF software Nabu [60]. Volumes at different stages of the experiment were aligned using SPAM to correct for sample movement and ease the analysis [61] (Video S3). To confirm water presence, the X‐ray count number inside the capillary was recorded before and after water injection. A decrease in the count number following water injection indicated successful entry, further supported by the appearance of a water meniscus after 12 min; see Figure S8h.

4.8. Rheological Characterization

The early‐stage structural build‐up and vertical expansion of hydrating MgO were monitored using an Anton Paar Physica MCR 502 rheometer with a stainless‐steel parallel plate geometry (PP25, 25 mm diameter) in a configuration permitting applying oscillatory shearing while compensating for the volume expansion of MgO during hydration. The sample was prepared ex situ by placing 1 g MgO in a confined volume and pre‐compacting it with a gravitational force of ∼100 g plate to level up the sample surface. The prepared sample was then placed onto a rheometer, after which water was added to fully submerge the sample (∼60 s). The evolution of storage modulus was measured after 30 s oscillatory prehearing of 10 times higher than the critical strain = 1% strain, to remove structuration history. Evolution of storage modulus (G’) and vertical expansion were measured simultaneously. G’ was measured at 1H oscillation and 1:10 of the critical strain, providing an index of early strength development under zero vertical force. Notably, it is assumed the material does not go through plastic changes when a strain smaller than the critical strain is applied. Vertical expansion was expressed as a relative change in sample height. Both experiments were performed at a controlled temperature of 25°C using P‐PTD200 and H‐PTD200 temperature control systems. The data were recorded with the software Rheoplus/32, version 1.23.403 (Anton Paar GmbH).

4.9. 1H Nuclear Magnetic Resonance (NMR)

1H static NMR CPMG measurements were carried out with a 14.1 T AVANCE IIIHD spectrometer equipped with a 4 mm CP/MAS probe (Bruker, DE). A standard CPMG echo train pulse sequence with continuous data acquisition was employed [62]. The experiment was set up with an echo delay of 250 µs (including a 4.8 µs π‐pulse), 16 scans per FID, and an interscan delay of 10 s. All measurements were done at 25°C. Data processing was performed with Topspin and the Topspin‐based Dynamics Center software module. Inverse Laplace Transformation with Tikhonov regularization was used to fit the CPMG decay curves and extract T2 distributions. Before the first measurement, specified amounts of water and MgO at ratios of 1 and 1.4 were mixed directly in a 4 mm o.d. Zirconia rotor using a spatula and a vortexer. The rotor was then transferred to the magnet bore, and the first measurement was started within 5 min of mixing. Each subsequent measurement was timed to start every hour, keeping the rotor and the sample in the magnet at a constant temperature of 25°C for the entire 24 h.

4.10. Ultrasonic Pulse Velocimetry (UPV)

MgO paste densification was measured via ultrasonic pulse velocity using the IP8 system (UltraTest GmbH, Germany) with stainless steel transmitter and receiver probes (E90‐12, 30 mm diameter, 35 kHz). The probe distance was set to 20 mm. Samples were mixed for 2 min, placed in a high‐acoustic‐attenuation cylindrical silicone mold (30 mm diameter), and vibrated for 5 s to de‐air and homogenize. All UPV testing was performed in a climate chamber (Climacell 111 Evo, MMM Group) at 21°C and 95% relative humidity. Ultratestlab V7.2 software (UltraTest GmbH) was used for data collection, de‐noising, and post‐processing.

4.11. Density Functional Theory (DFT) Calculations

The interaction between the interfaces of MgO (100, 110, and 111) surfaces and water molecules was investigated using DFT [63]. All electronic structure calculations were performed with the Vienna Ab initio Simulation Package [64, 65, 66], employing the projector‐augmented wave method to define electron‐ion interaction. The Perdew–Burke–Ernzerhof functional within the generalized gradient approximation [67] was used to model exchange and correlation effects. Brillouin zone sampling was carried out with a well‐converged k‐sampling equivalent to a 1 × 1 × 1 Monkhorst–Pack k‐point grid, which was found to be sufficient for the large interface systems [68]. Structural relaxations were performed with a plane‐wave cutoff energy of 400 eV, electronic self‐consistency cycles converged to 1 × 10−6 eV, and ionic forces were relaxed below 1 × 10−3 eV Å−1. Van der Waals interactions were included through the Becke–Johnson DFT‐D3 dispersion correction with damping. To characterize reaction pathways and transition states (TS) [69, 70, 71], the nudged elastic band (NEB) method [72] was applied. In this approach, the fully optimized initial and final structures are connected by a discrete set of intermediate “images” along the reaction coordinate. Images are relaxed perpendicular to the path, while spring forces maintain continuity along the path. For accurate saddle‐point determination on the potential energy surface, the climbing‐image NEB [73, 74] was employed, where the highest‐energy image is driven uphill along the path to converge precisely on the TS, providing the activation energy barrier for interfacial processes. DFT calculations were performed on MgO (100), (110) and (111) surfaces to evaluate the effect of volumetric strain on hydrated surface stability, using the Birch–Murnaghan equation of state [75]. Hydrated surface models, with one water molecule per surface Mg─O pair, were uniformly varied between 90% and 110% of their equilibrium volumes. The MgO (100) surface exhibits a minimum energy configuration at ∼2% volumetric expansion of the unit cell, whereas the (110) and (111) surfaces are most stable at their unstrained volumes; see Figure S10.

4.12. Tensile Deformation Calculations

The mechanical response of defect‐free brucite under tensile loading was investigated using DFT calculations. The equilibrium crystal structure obtained from the structural optimization was used as the reference configuration corresponding to zero strain. To determine the tensile behavior, uniaxial strain was applied along the selected crystallographic direction by incrementally expanding the corresponding lattice vector. The fractional atomic coordinates were initially kept unchanged. The strain was applied in small increments of ∼1% until mechanical instability was observed. For each strain value, the internal atomic coordinates were fully relaxed while keeping the strained lattice vector fixed. This procedure allows the atoms to reach their lowest‐energy configuration under the imposed deformation and ensures that the calculated stress corresponds to the equilibrium atomic configuration at each strain step. The stress tensor was obtained directly from the DFT calculations. The stress component corresponding to the loading direction was extracted and used to construct the stress–strain curve. By plotting the calculated stress as a function of the applied strain, the mechanical response of the crystal under tensile deformation was determined. The ideal tensile strength was defined as the maximum stress reached on the stress–strain curve before the onset of structural instability. Using this approach, the stress–strain curve of brucite was obtained by sequentially applying strain, relaxing atomic positions, extracting the corresponding stress component, and identifying the peak stress value as the theoretical tensile strength. The resulting peak stress represents the ideal tensile strength of defect‐free brucite and provides an atomistic reference for its tensile stability.

Author Contributions

Dan Meng: methodology, investigation, visualization, formal analysis, writing – original draft. Mohammadreza Izadifar: investigation, formal analysis, writing – review and editing. Flemming Bjerg Grumsen: investigation, writing – review and editing. Gustavo Pinzón: investigation, formal analysis, writing – review and editing. Kasper Enemark‐rasmussen: investigation, writing – review and editing. Adrian Alexander Schiefler: methodology, writing – review and editing. Eduardus Koenders: writing – review and editing. Julie Villanova: investigation, writing – review and editing. Jakob Birkedal Wagner: investigation, writing – review and editing. Rajmund Mokso: methodology, writing – review and editing. Carsten Kuenzel: methodology, investigation, writing – review and editing. Navid Ranjbar: conceptualization, methodology, funding acquisition, writing – review and editing, writing – original draft, visualization, validation, resources.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: advs78790‐sup‐0001‐SuppMat.docx.

Supporting File 2: advs78790‐sup‐0002‐VideoS1.mp4.

Download video file (33.1MB, mp4)

Supporting File 3: advs78790‐sup‐0003‐VideoS2.mp4.

Download video file (28.1MB, mp4)

Supporting File 4: advs78790‐sup‐0004‐DataS1.xlsx.

Supporting File 5: advs78790‐sup‐0005‐DataS2.xlsx.

Supporting File 6: advs78790‐sup‐0006‐VideoS3.zip.

ADVS-9999-e78790-s006.zip (111.4MB, zip)

Acknowledgements

The authors would like to acknowledge the research grant (VIL42098 and VIL69272) from VILLUM FONDEN, as well as the support of Aase og Ejnar Danielsens Fond. We would like to acknowledge the European Synchrotron Radiation Facility for the provision of synchrotron beamtime at beamline ID16B through proposal IHMA661 (DOI: 10.15151/ESRF‐ES‐2128660055) and the technical support for the set‐up and operation of the beamline. We acknowledge the support from the Imaging Centre at DTU. The Villum Foundation is acknowledged for access to the 600 MHz Solid‐State spectrometer at the NMR Center of DTU. Mohammadreza Izadifar and Eduardus Koenders gratefully acknowledge the financial support of the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) within the Priority Programme “Net‐Zero Concrete” (SPP 2436), under project number 541788011. Computational work was carried out using the Lichtenberg High‐Performance Computer at TU Darmstadt. Appreciation also goes to Kristian Riskær Povlsen, Kurt Kielsgaard Hansen, and Emil Christian Stillhoff Jensen for facilitating experiments.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  • 1. Dunant C. F., Joseph S., Prajapati R., and Allwood J. M., “Electric Recycling Of Portland Cement At Scale,” Nature 629, no. 8014 (2024): 1055–1061, 10.1038/s41586-024-07338-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Van Roijen E., Miller S. A., and Davis S. J., “Building Materials Could Store More Than 16 Billion Tonnes of CO2 Annually,” Science 387 (2025): 176–182, 10.1126/science.adq8594. [DOI] [PubMed] [Google Scholar]
  • 3. Meng D., Djobo J. N. Y., Segura I. P., Kaze R. C., and Kuenzel C., “Clay Activation: A Review,” Materials Today 92 (2025): 519–551, 10.1016/j.mattod.2025.11.024. [DOI] [Google Scholar]
  • 4. Wang S., Pourhaji P., Vassallo D., Heidarnezhad S., Scarlata S., and Rahbar N., “Durable, High‐Strength Carbon‐Negative Enzymatic Structural Materials Via A Capillary Suspension Technique,” Matter 9, no. 3 (2026): 102564, 10.1016/j.matt.2025.102564. [DOI] [Google Scholar]
  • 5. Beatty D. N. and Srubar W. V., “Nucleation Effects Of Coccoliths In Portland Cement,” Matter 8, no. 7 (2025): 102100, 10.1016/j.matt.2025.102100. [DOI] [Google Scholar]
  • 6. McQueen N., Kelemen P., Dipple G., Renforth P., and Wilcox J., “Ambient Weathering Of Magnesium Oxide for CO2 Removal From Air,” Nature Communications 11, no. 1 (2020): 3299, 10.1038/s41467-020-16510-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Bernard E., Nguyen H., Kawashima S., et al., “MgO‐Based Cements—Current Status And Opportunities,” RILEM Technical Letters 8 (2023): 65–78, 10.21809/rilemtechlett.2023.177. [DOI] [Google Scholar]
  • 8. Vlasopoulos N. and Cheeseman C. R., “Binder Composition,” US8496751B2, Calix Ltd. 2013.
  • 9. Kitagawa M., Misu S., Ichikawa J., and Matsuhashi H., “Preparation of Active MgO by Short‐Time Thermal Decomposition of Mg(OH)2 ,” Research on Chemical Intermediates 41, no. 12 (2015): 9463–9473, 10.1007/s11164-015-1971-z. [DOI] [Google Scholar]
  • 10. Amaral L. F., Oliveira I. R., Salomão R., Frollini E., and Pandolfelli V. C., “Temperature and Common‐Ion Effect On Magnesium Oxide (MgO) Hydration,” Ceramics International 36, no. 3 (2010): 1047–1054, 10.1016/j.ceramint.2009.12.009. [DOI] [Google Scholar]
  • 11. Rocha S. D., Mansur M. B., and Ciminelli V. S., “Kinetics And Mechanistic Analysis Of Caustic Magnesia Hydration,” Journal of Chemical Technology & Biotechnology 79, no. 8 (2004): 816–821, 10.1002/jctb.1038. [DOI] [Google Scholar]
  • 12. Pettauer M., Baldermann A., Eder S., and Dietzel M., “Hydration of MgO: Reaction Kinetics and pH Control on Brucite Crystal Morphology,” Crystal Growth & Design 24, no. 7 (2024): 3085–3092, 10.1021/acs.cgd.4c00243. [DOI] [Google Scholar]
  • 13. Scheck J., Berg J. K., Drechsler M., et al., “New Insights Into The Nucleation Of Magnesium Hydroxide And The Influence Of Poly(acrylic acid) During the Early Stages of Mg(OH)2 Crystallisation,” CrystEngComm 24, no. 44 (2022): 7718–7726, 10.1039/D2CE00896C. [DOI] [Google Scholar]
  • 14. Gevaudan J. P., Ruiz‐Agudo C., Manzano H., et al., “Dissolution, Diffusion, and Precipitation of Mg2+ in MgO‐Based Cements, a Review by the RILEM TC‐311 MBC,” Materials and Structures 59 (2026): 161, 10.1617/s11527-026-03048-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Xing Z., Bai L., Ma Y., Wang D., and Li M., “Mechanism of Magnesium Oxide Hydration Based on the Multi‐Rate Model,” Materials 11, no. 10 (2018): 1835, 10.3390/ma11101835. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Luong N. T. and Boily J.‐F., “Water Film‐Driven Brucite Nanosheet Growth and Stacking,” Langmuir 39, no. 31 (2023): 11090–11098, 10.1021/acs.langmuir.3c01411. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Walling S. A. and Provis J. L., “Magnesia‐Based Cements: A Journey of 150 Years, and Cements for the Future?,” Chemical Reviews 116, no. 7 (2016): 4170–4204, 10.1021/acs.chemrev.5b00463. [DOI] [PubMed] [Google Scholar]
  • 18. Lothenbach B., Bernard E., German A., and Winnefeld F., “MgO‐Based Binders,” ce/Papers 6 (2023): 342–356, 10.1002/cepa.2774. [DOI] [Google Scholar]
  • 19. Mundra S., Hallet V., Bernard E., et al., “Durability of MgO‐Based Binders: A Critical Review From RILEM TC 311‐MBC,” Materials and Structures 59, no. 4 (2026): 202, 10.1617/s11527-026-03076-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Wolterbeek T., Cornelissen E., Hangx S., and Spiers C., “Impact of Downhole Pressure And Fluid‐Access On The Effectiveness Of Wellbore Cement Expansion Additives,” Cement and Concrete Research 147 (2021): 106514, 10.1016/j.cemconres.2021.106514. [DOI] [Google Scholar]
  • 21. Li Z., Qian J., Qin J., Hua Y., Yue Y., and Tang H., “Cementitious and Hardening Properties Of Magnesia (MgO) Under Ambient Curing Conditions,” Cement and Concrete Research 170 (2023): 107184, 10.1016/j.cemconres.2023.107184. [DOI] [Google Scholar]
  • 22. Scott A., Oze C., Shah V., et al., “Transformation of Abundant Magnesium Silicate Minerals For Enhanced CO2 Sequestration,” Communications Earth & Environment 2, no. 1 (2021): 25, 10.1038/s43247-021-00099-6. [DOI] [Google Scholar]
  • 23. ASTM International Standard Practice for Fabricating and Testing Specimens of Ultrahigh Performance Concrete (ASTM International, 2017). [Google Scholar]
  • 24. British Standards Institution Concrete: Specification, Performance, Production and Conformity (British Standards Institution, 2000). [Google Scholar]
  • 25. Zhang T., “Development of Novel Low pH Magnesium Silicate Hydrate (MSH) Cement Systems for Encapsulation of Problematic Nuclear Wastes,” Doctor of Philosophy (Imperial College London, 2012). [Google Scholar]
  • 26. Ledoux E. E., Lin F., Miyagi L., et al., “Deformation of Polycrystalline MgO Up to 8.3 GPa and 1270 K: Microstructures, Dominant Slip‐Systems, and Transition to Grain Boundary Sliding,” Frontiers in Earth Science 10 (2022): 849777, 10.3389/feart.2022.849777. [DOI] [Google Scholar]
  • 27. Nguyen H., Dewitte C., Jin F., et al., “Unlocking Unexplained Characteristics of MgO: A Review by RILEM TC 311‐MBC,” Materials and Structures 59, no. 8 (2026): 389, 10.1617/s11527-026-03231-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. De Matos P. R., Neto J. S. A., Jansen D., De la Torre A. G., Kirchheim A. P., and Campos C. E., “In‐Situ Laboratory X‐Ray Diffraction Applied to Assess Cement Hydration,” Cement and Concrete Research 162 (2022): 106988. [Google Scholar]
  • 29. Winnefeld F., Epifania E., Montagnaro F., and Gartner E. M., “Further Studies of the Hydration of MgO‐Hydromagnesite Blends,” Cement and Concrete Research 126 (2019): 105912, 10.1016/j.cemconres.2019.105912. [DOI] [Google Scholar]
  • 30. German A., Winnefeld F., Lura P., Rentsch D., and Lothenbach B., “Hydrous Carbonate‐Containing Brucite (HCB) in MgO/Hydromagnesite Blends,” Cement and Concrete Research 173 (2023): 107304, 10.1016/j.cemconres.2023.107304. [DOI] [Google Scholar]
  • 31. Kuenzel C., Zhang F., Ferrándiz‐Mas V., Cheeseman C. R., and Gartner E. M., “The Mechanism of Hydration of MgO‐Hydromagnesite Blends,” Cement and Concrete Research 103 (2018): 123–129, 10.1016/j.cemconres.2017.10.003. [DOI] [Google Scholar]
  • 32. Jin F. and Al‐Tabbaa A., “Characterisation of Different Commercial Reactive Magnesia,” Advances in Cement Research 26, no. 2 (2014): 101–113, 10.1680/adcr.13.00004. [DOI] [Google Scholar]
  • 33. Mi T., Yang E.‐H., and Unluer C., “Investigation of the Properties of Reactive MgO‐Based Cements and Their Effect on Performance,” Cement and Concrete Composites 138 (2023): 104984, 10.1016/j.cemconcomp.2023.104984. [DOI] [Google Scholar]
  • 34. Aranda M. A. G., De la Torre A. G., Bernal S. A., and Provis J. L., “Insights Into the Nano‐ And Microstructures Of Cementitious Materials Using Synchrotron X‐Rays,” Cement and Concrete Research 202 (2026): 108150, 10.1016/j.cemconres.2026.108150. [DOI] [Google Scholar]
  • 35. Vlug W., Plümper O., Kandianis M., van Blaaderen A., and van Huis M., “In‐Situ Hydration of MgO Nanocrystals to Amorphous Mg(OH)2 Using Liquid Cell Transmission Electron Microscopy,” in European Microscopy Congress 2016: Proceedings (Wiley‐VCH, 2016), 157–158. [Google Scholar]
  • 36. Cao F., Miao M., and Yan P., “Hydration Characteristics And Expansive Mechanism of MgO Expansive Agents,” Construction and Building Materials 183 (2018): 234–242, 10.1016/j.conbuildmat.2018.06.164. [DOI] [Google Scholar]
  • 37. Chen J., Huang L., Dong L., et al., “Hydration Behavior of MgO Surface: A First‐Principles Study,” Applied Surface Science 611 (2023): 155441, 10.1016/j.apsusc.2022.155441. [DOI] [Google Scholar]
  • 38. Tang X., Nie Y., Jin Q., et al., “Kinetics and Mechanism Of Ultrasonic‐Assisted Magnesium Oxide Hydration,” Ultrasonics Sonochemistry 40 (2018): 995–1002, 10.1016/j.ultsonch.2017.08.027. [DOI] [PubMed] [Google Scholar]
  • 39. Yim H. J., Kim J. H., and Shah S. P., “Cement Particle Flocculation And Breakage Monitoring Under Couette Flow,” Cement and Concrete Research 53 (2013): 36–43, 10.1016/j.cemconres.2013.05.018. [DOI] [Google Scholar]
  • 40. Acker P., “Swelling, Shrinkage And Creep: A Mechanical Approach To Cement Hydration,” Materials and Structures 37, no. 4 (2004): 237–243, 10.1007/BF02480632. [DOI] [Google Scholar]
  • 41. Friedemann K., Stallmach F., and Kärger J., “NMR Diffusion And Relaxation Studies During Cement Hydration—A Non‐Destructive Approach For Clarification Of The Mechanism Of Internal Post Curing Of Cementitious Materials,” Cement and Concrete Research 36, no. 5 (2006): 817–826, 10.1016/j.cemconres.2005.12.007. [DOI] [Google Scholar]
  • 42. Gleason A., Marquardt H., Chen B., Speziale S., Wu J., and Jeanloz R., “Anomalous Sound Velocities in Polycrystalline MgO Under Non‐Hydrostatic Compression,” Geophysical Research Letters 38 (2011): L03304. [Google Scholar]
  • 43. Hermann A., “High‐Pressure Phase Of Brucite Stable At Earth's Mantle Transition Zone And Lower Mantle Conditions,” Proceedings of the National Academy of Sciences 113, no. 49 (2016): 13971–13976, 10.1073/pnas.1611571113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Zhu J., Kee S.‐H., Han D., and Tsai Y.‐T., “Effects of Air Voids On Ultrasonic Wave Propagation In Early Age Cement Pastes,” Cement and Concrete Research 41, no. 8 (2011): 872–881, 10.1016/j.cemconres.2011.04.005. [DOI] [Google Scholar]
  • 45. Rahemipoor S., Kuenzel C., Valdemārs Eiduks T., et al., “Surface‐Engineered Cenospheres Encapsulating Phase Change Materials for Functional Cementitious Composites,” Advanced Science 12, no. 26 (2025): 2417350, 10.1002/advs.202417350. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Tayebi M., Ilango N. K., Nguyen H., et al., “New Insights Into Hydration of MgO in the Presence Of Polycarboxylate Ether Superplasticizers,” Cement and Concrete Composites 162 (2025): 106153, 10.1016/j.cemconcomp.2025.106153. [DOI] [Google Scholar]
  • 47. Arıanpour A. Ç., “Phase and Microstructural Studies of Two Types Commercial Magnesia and Spinel Refractory Raw Materials With Emphasis on Impurities Characterization,” Karaelmas Fen ve Mühendislik Dergisi 7 (2017): 395–401. [Google Scholar]
  • 48. Zhang S., Zhao C., Reyes Y. I. A., et al., “Facile Gram‐Scale Synthesis of Size‐Tunable MgO Nanosheets Enclosed by (111) Surface With Remarkable Stability and Uniform Host Sites for Atom Dispersion,” Chemistry of Materials 36, no. 9 (2024): 4204–4214, 10.1021/acs.chemmater.3c03018. [DOI] [Google Scholar]
  • 49. Jug K., Heidberg B., and Bredow T., “Molecular Dynamics Study of Water Adsorption Structures on the MgO(100) Surface,” The Journal of Physical Chemistry C 111, no. 18 (2007): 6846–6851, 10.1021/jp067651n. [DOI] [Google Scholar]
  • 50. Ohlin C. A., Villa E. M., Rustad J. R., and Casey W. H., “Dissolution of Insulating Oxide Materials At The Molecular Scale,” Nature Materials 9, no. 1 (2010): 11–19, 10.1038/nmat2585. [DOI] [PubMed] [Google Scholar]
  • 51. Abdolhosseini Qomi M. J., Miller Q. R. S., Zare S., Schaef H. T., Kaszuba J. P., and Rosso K. M., “Molecular‐Scale Mechanisms of CO2 Mineralization In Nanoscale Interfacial Water Films,” Nature Reviews Chemistry 6, no. 9 (2022): 598–613, 10.1038/s41570-022-00418-1. [DOI] [PubMed] [Google Scholar]
  • 52. McKelvy M. J., Sharma R., Chizmeshya A. V. G., Carpenter R. W., and Streib K., “Magnesium Hydroxide Dehydroxylation:  In Situ Nanoscale Observations of Lamellar Nucleation and Growth,” Chemistry of Materials 13, no. 3 (2001): 921–926, 10.1021/cm000676t. [DOI] [Google Scholar]
  • 53. Jug K., Heidberg B., and Bredow T., “Cyclic Cluster Study on the Formation of Brucite From Periclase and Water,” The Journal of Physical Chemistry C 111, no. 35 (2007): 13103–13108, 10.1021/jp072889c. [DOI] [Google Scholar]
  • 54. Inagaki T. and Hatanaka M., “Comprehensive Molecular‐Level Understanding of MgO Hydration Through Computational Chemistry,” The Journal of Physical Chemistry C 130, no. 3 (2026): 1312–1326, 10.1021/acs.jpcc.5c06767. [DOI] [Google Scholar]
  • 55. Wang Y., Zhou S., and Wang Z., “Bridging Accuracy And Efficiency In Solid‐Liquid Interface Simulations: An Accelerated Fine‐Tuning Strategy for MgO Hydration,” Computational Materials Science 269 (2026): 114726, 10.1016/j.commatsci.2026.114726. [DOI] [Google Scholar]
  • 56. Langford J. I. and Wilson A., “Scherrer After Sixty Years: A Survey And Some New Results In The Determination Of Crystallite Size,” Journal of Applied Crystallography 11, no. 2 (1978): 102–113, 10.1107/S0021889878012844. [DOI] [Google Scholar]
  • 57. Dong H., Yang E.‐H., Unluer C., and Jin F., “Investigation of the Properties of MgO Recovered From Reject Brine Obtained From Desalination Plants,” Journal of Cleaner Production 196 (2018): 100–108, 10.1016/j.jclepro.2018.06.032. [DOI] [Google Scholar]
  • 58. Martínez‐Criado G., Villanova J., Tucoulou R., et al., “ID16B: A Hard X‐Ray Nanoprobe Beamline at the ESRF for Nano‐Analysis,” Journal of Synchrotron Radiation 23, no. 1 (2016): 344–352, 10.1107/S1600577515019839. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. Favre‐Nicolin V., Girard G., Leake S., et al., “PyNX: High‐Performance Computing Toolkit For Coherent X‐Ray Imaging Based On Operators,” Journal of Applied Crystallography 53, no. 5 (2020): 1404–1413, 10.1107/S1600576720010985. [DOI] [Google Scholar]
  • 60. High‐performance tomographic reconstruction software, (accessed: July 21, 2025) https://gitlab.esrf.fr/tomotools/nabu.
  • 61. Stamati O., Andò E., Roubin E., et al., “spam: Software for Practical Analysis of Materials,” Journal of Open Source Software 5, no. 51 (2020): 2286, 10.21105/joss.02286. [DOI] [Google Scholar]
  • 62. Zhou C., Ren F., Zeng Q., Xiao L., and Wang W., “Pore‐Size Resolved Water Vapor Adsorption Kinetics Of White Cement Mortars As Viewed From Proton NMR Relaxation,” Cement and Concrete Research 105 (2018): 31–43, 10.1016/j.cemconres.2017.12.002. [DOI] [Google Scholar]
  • 63. Kohn W. and Sham L. J., “Self‐Consistent Equations Including Exchange and Correlation Effects,” Physical Review 140 (1965): A1133–A1138, 10.1103/PhysRev.140.A1133. [DOI] [Google Scholar]
  • 64. Kresse G. and Hafner J., “Ab Initio Molecular Dynamics For Liquid Metals,” Physical Review B 47, no. 1 (1993): 558–561, 10.1103/PhysRevB.47.558. [DOI] [PubMed] [Google Scholar]
  • 65. Kresse G. and Furthmüller J., “Efficient Iterative Schemes For Ab Initio Total‐Energy Calculations Using A Plane‐Wave Basis Set,” Physical Review B 54, no. 16 (1996): 11169–11186, 10.1103/PhysRevB.54.11169. [DOI] [PubMed] [Google Scholar]
  • 66. Kresse G. and Furthmüller J., “Efficiency of Ab‐Initio Total Energy Calculations For Metals And Semiconductors Using A Plane‐Wave Basis Set,” Computational Materials Science 6, no. 1 (1996): 15–50, 10.1016/0927-0256(96)00008-0. [DOI] [PubMed] [Google Scholar]
  • 67. Perdew J. P., Burke K., and Ernzerhof M., “Generalized Gradient Approximation Made Simple,” Physical Review Letters 77, no. 18 (1996): 3865–3868, 10.1103/PhysRevLett.77.3865. [DOI] [PubMed] [Google Scholar]
  • 68. Monkhorst H. J. and Pack J. D., “Special Points for Brillouin‐Zone Integrations,” Physical Review B 13, no. 12 (1976): 5188–5192, 10.1103/PhysRevB.13.5188. [DOI] [Google Scholar]
  • 69. Izadifar M., Ukrainczyk N., and Koenders E., “Atomistic Insights Into Silicate Dissolution of Metakaolinite Under Alkaline Conditions: Ab Initio Quantum Mechanical Investigation,” Langmuir 40, no. 37 (2024): 19332–19342, 10.1021/acs.langmuir.4c00890. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70. Izadifar M., Ukrainczyk N., Schönfeld K., and Koenders E., “Activation Energy Of Aluminate Dissolution In Metakaolin: MLFF‐Accelerated DFT Study of vdW and Hydration Shell Effects,” Nanoscale Advances 7, no. 14 (2025): 4325–4335, 10.1039/D5NA00103J. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71. Salah Uddin K. M., Izadifar M., Ukrainczyk N., Koenders E., and Middendorf B., “Molecular Pathways Governing the Earliest Stages of C–S–H Nucleation,” Materials Advances 7 (2026): 9678–9694, 10.1039/d6ma00783j. [DOI] [Google Scholar]
  • 72. Tao Y., Martin P., Manzano H., and Qomi M. J. A., “Mesoscopic Mechanisms Of Dicalcium Silicate Dissolution,” Cement and Concrete Research 189 (2025): 107660, 10.1016/j.cemconres.2024.107660. [DOI] [Google Scholar]
  • 73. Henkelman G., Uberuaga B. P., and Jónsson H., “A Climbing Image Nudged Elastic Band Method For Finding Saddle Points And Minimum Energy Paths,” The Journal of Chemical Physics 113, no. 22 (2000): 9901–9904, 10.1063/1.1329672. [DOI] [Google Scholar]
  • 74. Izadifar M., Natzeck C., Emmerich K., et al., “Unexpected Chemical Activity of a Mineral Surface: The Role of Crystal Water in Tobermorite,” The Journal of Physical Chemistry C 126, no. 30 (2022): 12405–12412, 10.1021/acs.jpcc.1c10151. [DOI] [Google Scholar]
  • 75. Izadifar M., Königer F., Gerdes A., and Wöll C., “Correlation Between Composition and Mechanical Properties of Calcium Silicate Hydrates Identified by Infrared Spectroscopy and Density Functional Theory,” The Journal of Physical Chemistry C 123, no. 17 (2019): 10868–10873, 10.1021/acs.jpcc.8b11920. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting File 1: advs78790‐sup‐0001‐SuppMat.docx.

Supporting File 2: advs78790‐sup‐0002‐VideoS1.mp4.

Download video file (33.1MB, mp4)

Supporting File 3: advs78790‐sup‐0003‐VideoS2.mp4.

Download video file (28.1MB, mp4)

Supporting File 4: advs78790‐sup‐0004‐DataS1.xlsx.

Supporting File 5: advs78790‐sup‐0005‐DataS2.xlsx.

Supporting File 6: advs78790‐sup‐0006‐VideoS3.zip.

ADVS-9999-e78790-s006.zip (111.4MB, zip)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.


Articles from Advanced Science are provided here courtesy of Wiley

RESOURCES