Mechanisms and Protection Strategies for Concrete Degradation Under Magnesium Salt Environment: A Review

Abstract

Concrete structures suffering from Mg²⁺ environments may suffer severe damage, which mainly has something to do with the coupled effect among Cl⁻, SO₄²⁻, and Mg²⁺. Based on a systematic review of Web of Science and Scopus database (2000–2025), we first summarized the migration behavior, reaction paths, and interaction mechanism of Cl⁻, SO₄²⁻, and Mg²⁺ in cementitious matrices. Secondly, from the perspective of Cl⁻ cyclic adsorption–desorption breaking the passivation film of steel bars, SO₄²⁻ generating expansion products leads to crack expansion, then Mg²⁺ decalcifies C-S-H and transforms into M-S-H; we analyzed the main damage mechanisms, respectively. In addition, under the coexistence conditions of three kinds of ions, the “fixation–substitution–redissolution” process and “crack–transport” coupling positive feedback mechanism further increase the development rate of damage. Then, some anti-corrosion measures, such as mineral admixtures, functional chemical admixtures, fiber reinforcements, surface coatings, and new binder systems, are summarized, and the pros and cons of different anti-corrosion technologies are compared and evaluated. Lastly, from two aspects of simulation prediction for the coupled corrosion damage mechanism and service life prediction, respectively, we have critically evaluated the advances and problems existing in the current research on the aspects of ion migration-reaction coupled models, multi-physics coupled frameworks, phase-field methods, etc. We found that there is still much work to be conducted in three respects: deepening mechanism understanding, improving prediction precision, and strengthening the connection between laboratory test results and actual projects, so as to provide theoretical basis and technical support for the durability design and anti-corrosion strategies of concrete in complex Mg²⁺ environments.

1. Introduction

Magnesium salts refer to the magnesium cations (Mg²⁺) combined with chloride ions (Cl⁻), sulfate ions (SO₄²⁻), carbonate ions (CO₃²⁻) and so on. They are commonly found in the intertidal zone of the ocean, salt lake basins, underground brine mines, permafrost regions subjected to repeated freeze–thaw cycles and other areas where large-scale concrete structures can be used for long periods [1,2,3]. Here, the corrosion damage of concrete from magnesium salts exhibits high complexity and time dependence. Magnesium salts themselves might cause corrosion damage to concrete, or they might coexist with Cl⁻ and SO₄²⁻, thereby forming multivalent ions that lead to longer-term and even greater corrosion damage. On the one hand, the characteristics of “intrusion from the outside”, “accumulation inside” and “decreased performance” occur in the long-term development of magnesium salt corrosion [4,5,6]. According to research results from site-cured concrete specimens [7,8], it was found that after long-term exposure to a magnesium salt environment, most concrete structures exhibited severe surface peeling and porous destruction (Figure 1), accompanied by crazing, deterioration of compressive strength, and steel bar corrosion. The load-carrying capacity of members was reduced, as well as their design service life.

In recent years, researchers worldwide have paid more and more attention to the clarification of the corrosion mechanisms of chloride, sulfate, and magnesium salts, which has significantly advanced our understanding of the corrosion mechanisms of individual salt species. Cl⁻ has a higher migration rate in concrete and can destroy the passivation film on the surface of reinforcement through a cycle adsorption/desorption process, thereby inducing corrosion [9,10,11,12]; SO₄²⁻ promotes the generation of expansive products such as ettringite, gypsum, and aggravate cracking [13]; Mg²⁺ reacts with cement hydration products and reduces the compactness and stability of the cement matrix [14]. Previous researchers mainly focused their studies on the conditions of a single ion environment, without a uniform opinion on the interaction mechanism between Cl⁻, SO₄²⁻, and Mg²⁺. In terms of the improvement and repair of durability properties, research on various remedial treatments (blended cementitious materials, functional cementitious systems, surface sealing treatment) is relatively mature. As for the durability prediction model, although some models have been established based on diffusion theory, coupled transport–reaction models, kinetic reaction models, etc., there are still many problems in multi-physics coupling modeling, parameter determination, and so forth.

According to the above analysis, unlike previous review articles that mainly describe single-salt corrosion qualitatively, we summarize the quantitative coupling mechanism and synergetic protection strategy under the Mg²⁺-SO₄²⁻-Cl⁻ environment. In view of the deficiency of quantitative basis in previous studies, the novelty of our work is reflected in two aspects: (1) Critical synthesis of coupled mechanisms: Different from the simple description before, we quantitatively summarize how the ion concentration ratio affects the “fixation–substitution–redissolution” process, and evaluate the critical threshold and the promotion effect of the corrosion rate under the “crack–transport” positive feedback based on the existing theoretical models. (2) Synergetic protection strategies: Rather than considering each technology individually, we analyze the compatibility among different levels of protection systems (mineral admixture, surface coating, and new binder system) to put forward a more efficient path for durability design under complex marine conditions. Finally, some current research hotspots in multiphysical simulation and lifespan prediction are discussed. In order to intuitively show the logical relationship between related contents,

Table 1. Conceptual framework of coupled deterioration mechanisms in magnesium–sulfate–chloride environments.

Stage	Specific Deterioration Mechanism	Key Chemo-Physical Process	Corresponding Section
I. Intrusion	Multi-Species Transport Coupling (Mg2+-SO42−-Cl− Interaction)	(1) Electrostatic Coupling: High-valent Mg2+ changes the pore solution potential and slows down Cl− migration. (2) Activity Correction: Ionic strength variations affect the thermodynamic activity of species.	Section 2.3.1 (Transport Models)
II. Reaction	(1) Magnesium Attack (Decalcification) (2) Sulfate Attack (Expansion) (3) Chloride Attack (Depassivation)	(1) Isomorphic Substitution: Mg2+ replaces Ca2+ in C-S-H, forming non-cementitious M-S-H (Ca/Si↓). (2) Thermodynamic Instability: Decrease in pH results in decomposition of Friedel’s salt. (3) Mineral Precipitation: Expansive Gypsum/Ettringite and pore-blocking Brucite formation.	Section 2.1 and Section 2.2 (Reaction Pathways)
III. Damage	Chemo-Mechanical Damage Evolution	(1) Confinement Crystallization: The crystal growth pressure is larger than the tensile strength of the pore wall. (2) Pore Structure Destruction: Enlargement of capillary pores leads to an increase in permeability. (3) Macro-Cracking: The coalescence of micro-cracks causes surface spalling and exposure of the reinforcement.	Section 2.2.2 and Section 2.3.2 (Damage Models)
IV. Response	Synergistic Protection and Prediction	(1) Controlled Phase Matching: Control Al/Si ratio to prevent the generation of metastable phases. (2) Calibrate Model: Bayesian update of service life models based on field inspection data.	Section 3 and Section 4 (Strategies and Outlook)

shows the conceptual framework from micro-mechanisms to structural durability.

2. Study on the Damage Mechanism of Concrete Under Magnesium Salt Attack

2.1. Magnesium Chloride Attack

Chloride ions in concrete exist mainly in two forms in concrete’s pore system (Figure 2) [15,16], i.e., they exist freely in the pore solution or they are fixed on the solid matrix because of physical adsorption at the C-S-H surface or chemical interaction with CAH and the formation of chloride-aluminate phases such as AFm-Cl (Friedel’s salt). The degradation mechanism of concrete subjected to chloride salt environments mainly results from the good mobility and chemical activity of chloride ions: the chloride ions will rapidly migrate along the interconnected pores, reach the surface of the reinforcing bar and destroy the protective passive film; simultaneously, they react with different hydrated products, change the microstructure, and induce further corrosion reactions. Compared with other types of salt corrosion damage, the characteristics of chloride corrosion damage belong to “penetration quickly” and “corrode quickly.” It is easy to cause early damage phenomena such as cover spalling, crack expansion, and rapid corrosion development. Therefore, it is important to understand the migration path and fixation behavior of Cl⁻ in cement-based materials and its effect on the passivation film of steel bars to reveal the degradation mechanisms of concrete under MgCl₂ corrosion.

2.1.1. Chloride Ion Transport Mechanism

Chloride ions (Cl⁻) are the most aggressive species in natural seawater, possessing the highest concentration of approximately 1.94% [17]. Their small ionic radius and high hydrophilicity facilitate rapid penetration into pore structures to initiate corrosion. Generally, four mechanisms describe this migration [17].

(1)

Unlike Na⁺ in NaCl solutions, Mg²⁺ reacts with C-S-H gels to form porous M-S-H. This decalcification process increases pore connectivity and reduces tortuosity. Consequently, chloride diffusivity is significantly higher in magnesium-rich environments than in standard conditions [17,18].

(2)

The development of cracks is another distinctive feature unique to Mg attack. The crystallization pressure produced by Mg salts causes internal micropores/cracks to form inside concrete; such pores/cracks then become ‘highways’ along which ingress occurs. Studies have shown [19,20] that once the crack width exceeds 0.07 mm, the rate of Cl migration increases dramatically in a nonlinear fashion compared with pure diffusion without cracking.

(3)

Under wet–dry cycles, “pump effect” is further enhanced by magnesium-induced surface degradation: the magnesium salts attack causes surface scaling/spalling (physical removal of the protective layer), which in turn decreases the resistance to capillary uptake, thus favoring a faster accumulation and inward migration of chloride ions during wet phases [21].

(4)

Permeation and electromigration are further promoted by chemical effects: unlike what happens in NaCl solutions (where the pore fluid stays highly alkaline), the hydrolysis of Mg²⁺ leads to precipitation of Mg(OH)₂, with a strong decrease in pH. Acidification destabilizes the electrostatic potential of the pore wall and decreases its capacity of repelling and binding Cl⁻, thereby enhancing their entry inside the specimen [22,23].

Table 2. Chloride (Cl⁻) transport mechanisms for different zones within marine concrete.

Exposure Zone	Structural Examples	Chloride Ion Transport Mechanisms	References
Full Submersion Zone	Structures Submerged in Seawater	Permeation, Diffusion, Electrochemical Migration	[18,22,24]
Tidal Zone of Marine Structures	Components in the Upper and Lower Tidal Zones	Capillary Absorption, Diffusion	[17,21,25]
Splash Zone	Wave-Impacted Structural Components	Capillary Absorption, Diffusion	[9,25,26]
Atmospheric Zone of Marine Structures	Coastal Structures Not Directly Exposed to Seawater	Capillary Absorption	[9,26]

shows the typical structure example of concrete in different exposure zones in the marine environment. It summarizes the main migration mechanisms of chloride ions in each zone, respectively, which provides a visual reference for understanding the paths of chloride ion invasion into concrete under different conditions of the sea environment.

2.1.2. Mechanism of Magnesium Chloride Attack

After the chloride ions enter into the matrix, the chloride ions (Cl⁻) react chemically with the hydration product of cementitious materials, change its microstructure, accumulate corrosion products, produce cracks, etc., continuously promoting their own corrosion and damage evolution. The degradation mechanism was defined as a “Binding–Damage–Feedback” process, and developed through three different stages (as shown in Figure 3):

(1)

Stage I—Ion-binding and sequestration effect. At the early stage, part of the entering Cl⁻ is fixed by the cement matrix through chemical binding with aluminates to generate Friedel’s salt (3CaO·Al₂O₃·CaCl₂·10H₂O), and physical adsorption on C-S-H gels. Such fixation effects can temporarily “fix” free Cl⁻, forming a “buffer” effect which slows the rise in Cl⁻ concentration in the pore solution [27,28,29].

(2)

Stage II-Critical threshold and depassivation. Once the binding capacity of concrete has been saturated, free chlorides will build up at the steel–concrete interface. Studies have shown that if the molar ratio of Cl⁻/OH⁻ exceeds its critical threshold (usually between 0.6 and 1) [30,31], the protective passive film of the rebars would be unstable and destroyed. The threshold is not a fixed value but changes with the porosity, as well as the density of ITZ, as was reported by Sui et al. [32].

(3)

Stage III: Propagation and Positive Feedback Loop. After corrosion is initiated, the produced rust product has a volumetric expansion of about 6–10 times higher than steel, which will produce internal tensile stresses to crack the concrete cover [33,34,35]. These cracks develop into the shortcut path for migration, greatly increase the chloride diffusion coefficient (especially at the early stage of corrosion) [27], form another self-reinforced “corrosion–crack–migration” positive feedback loop, and cause the quick collapse of structures.

To synthesize these mechanisms, we have visually summarized the complete degradation pathway in Figure 3, which shows the path-dependent progression from “outside intrusion” at the beginning to “performance degradation” in the end, with emphasis on how the microchemical changes in Stages I and II finally cause macrophysical fracture in Stage III, indicating that it is not a linear but a coupled chemophysical damage-driven positive feedback loop.

Although the degradation mechanism mentioned above seems reasonable from a conceptual point of view, there has not yet been a unified scientific view. Although it has already been agreed that the formation of Friedel’s salt and migration of Cl⁻ through the pores has occurred, there have not yet been any consistent values reported for the critical concentration of Cl⁻ leading to the breakdown of the protective passive film. Recently, it was discovered that this value may vary according to interface conditions. After all, the corrosion rate mainly depends on how much Mg²⁺ displaces Ca²⁺ and modifies the porosity.

In summary, current research has well established the formation of Friedel’s salt and the migration paths of chloride ions. However, the critical chloride threshold for steel depassivation remains debated. For engineering practice, the most critical factor is the magnesium-induced decalcification. This process coarsens the pore structure and accelerates chloride ingress, posing a greater threat than chloride diffusion alone.

2.2. Magnesium Sulfate Attack

In addition to the above-mentioned ion migration mechanisms, there is another effect that is caused by the corrosion product that should not be neglected. What exactly are its effects on the cementitious matrix? We know that magnesium sulfate is the main corrosive active species that causes damage when the concrete specimen is immersed in seawater. Compared with the corrosion effects produced when the concrete is separately immersed in Na₂SO₄ solution and MgCl₂ solution, magnesium sulfate can produce more complex chemical reactions and coupling degradation effects during the corrosion process; thus, it causes more serious damage and consequences. In order to further explain what is different among them in terms of reaction mechanism, product, and structural damage features, we compare and analyze them together as shown below in

Table 3. Comparison analysis of magnesium sulfate and common sulfate salts on concrete corrosion mechanism.

Aggressive Medium	Ions Involved in Primary Reactions	Key Reaction Products	Dominant Damage Mechanism	Characteristics of Structural Impact	References
Na2SO4	SO42−	AFt, Gypsum	Expansion and Crystallization: Formation of expansive products and salt crystallization pressure	Cracking: Surface expansion, spalling, and micro-cracking	[36,37]
MgCl2	Cl−, SO42−	Friedel’s Salt, Mg(OH)2, M-S-H	Decalcification and Binding: Transformation of C-S-H to M-S-H; Reaction of Cl− to form Friedel’s salt	Pore Coarsening: Increased porosity, surface softening, and pitting corrosion of steel	[14,27,38]
MgSO4	SO42−, Mg2+	Mg(OH)2, M-S-H, AFt	Coupled Attack: Simultaneous decalcification of C-S-H and formation of expansive products	Disintegration: Loss of binder cohesion (mushy surface) combined with internal cracking	[1,39]

. Note: for clear comparison in damage severity, “dominant damage mechanism” here means the microscopic chemical reaction and/or phase transformation (such as decarbonating reaction/crystallization), and “characteristic feature of structural effect” means the macroscopic physical damage phenomenon (such as cracking/softening/increase in porosity) it produces afterwards.

It can be understood from Table 3 that magnesium sulfate, compared with sodium sulfate and magnesium chloride, has a more complicated corrosion mechanism. There are two reasons for its destruction of concrete: First, it not only leads to the expansion of the surface products, but also results in widespread decalcification of C-S-H and the generation of M-S-H gel, so as to destroy the concrete structure layer by layer, from outside to inside. Secondly, different from sodium sulfate, which has a single expansion route, magnesium sulfate has multiple routes: expansion, decalcification, colloid transformation, etc. Compared with magnesium chloride, it emphasizes both the chemical reaction product effect and the physical effect of crystal growth. Therefore, the damage caused by magnesium sulfate to concrete is usually continuous and coupled damage, which will be explained separately below, from a chemical and physical perspective.

2.2.1. Chemical Corrosion Mechanism of Magnesium Sulfate

Hydration products of cement are mainly: calcium silicate hydrate (C-S-H), calcium hydroxide (CH), calcium aluminate hydrate (C-A-H), ettringite (AFt), monosulfate (AFm) and a small amount of other phases. The above products are mainly poorly thermally stable solids and easy to decompose under the magnesium sulfate environment with both SO₄²⁻ and Mg²⁺ [40]. Santhanam et al. [41] found that the corrosion evolution process of concrete under the attack of magnesium sulfate can be divided into several development stages. In the initial stage, the reaction was dominated by the thermodynamic stabilization reaction. Because the solubility of magnesium hydroxide [Mg(OH)₂] is lower than that of calcium hydroxide [Ca(OH)₂], Mg²⁺ in pore solution replaced Ca²⁺ and the reaction belongs to cation exchange reaction. This caused in situ precipitation of Mg(OH)₂ (Equation (1)), and then continuous outward migration of Ca²⁺ occurred, and Ca(OH)₂ was consumed. With the consumption of Ca(OH)₂ phase, the pH value of pore solution decreased, and the alkalinity of the system weakened, causing the instability of C-S-H and C-A-H, the occurrence of decalcification phenomena, and the occurrence of phase transformation.

$${Ca\left(OH\right)}_2+{MgSO}_4\to {CaSO}_4+{Mg\left(OH\right)}_2$$

(1)

In the later stage, the penetrated SO₄²⁻ reacts with Ca²⁺ and Al^3+, respectively, generating gypsum (CaSO₄·2H₂O) and ettringite (AFt) via Reactions (2) and (3). The generated hydration products will produce volume expansion, depositing inside pores and ITZ and inducing high-concentration stress. In addition, parts of AFm phases (monosulfate) are unstable and transform from AFm to AFt, further promoting the expansion effect.

$${Ca\left(OH\right)}_2+{SO}_4^{2-}+{2H}_{2}O\to {CaSO}_4·{2H}_{2}O$$

(2)

$$\begin{gathered} 3CaO·{Al}_2{O}_3·{CaSO}_4·12H_{2}O+{CaSO}_4·{2H}_{2}O+{16H}_{2}O\to \\ 3CaO·{Al}_2{O}_3·{3CaSO}_4·{32H}_{2}O \end{gathered}$$

(3)

Finally, in the last stage, the continuous replacement of Ca²⁺ in C-S-H by Mg²⁺ will form amorphous M-S-H gel (Equation (4)). And such conversion is strictly dominated by the chemistry conditions of pore solution. It has been reported in previous research that the stability of C-S-H may be destroyed as long as the pH is lower than around 12.5 because of the dissolution of portlandite. Then, it would induce the conversion into M-S-H once the pH continued decreasing to 9.0–10.5 [14,42]. In addition, the conversion rate is also affected by the ionic strength: higher concentration of MgSO₄ solution (>5 wt%) could greatly promote the speed of the decalcification reaction, resulting in a quick decrease in adhesion strength on the surface layers [39].

$$C\text{-}S\text{-}H+{MgSO}_4\to M\text{-}S\text{-}H+{CaSO}_4$$

(4)

For all reaction formulas mentioned above, gypsum and ettringite are two common expansive degradation products in a sulfuric acid corrosion environment. Magnesium hydroxide [Mg(OH)₂] and poorly bound SiO₂·xH₂O gel are unique degradation products under MgSO₄ system. Although there are still disputes on how these products (e.g., gypsum) affect the damage level of the concrete specimen, according to previous studies, it has been recognized by most researchers that these products would alter the pore characteristics and bonding property of cement paste dramatically and then accelerate concrete damage evolution [43].

From the results of the analysis from XRD and SEM-EDS by Liu et al. [39], it can be seen that both the kind and amount of corrosion products are related to the concentration of sulfates and time of immersion. If it is immersed in a low-concentration solution or for a short time, the corrosion product has a “filling–strengthening” effect on the internal structure, and if it is immersed in a high-concentration solution or for a long time, the corrosion product continues to deposit into the pore channels, expand in volume, develop cracks, and lose cementitious material successively, so that the macroscopic performance indicators of concrete specimens further decrease. The chemical corrosion behavior of MgSO₄ solution shows an interesting mechanism of staged, path-wise, coupled evolution, kinetic conversion from protective film formation to deep damage evolution, and diversity in the expansion and cementitious loss risk of some reaction products. This brings great difficulties for concrete’s durability design under marine environments.

2.2.2. Physical Erosion Mechanism of Magnesium Sulfate

In addition to the above-mentioned chemical degradation mechanism, magnesium sulfate corrosion also suffers from obvious physical damage effects when the concentration is high and there are repeated wetting/drying actions. Under this environment, when the salt repeatedly dissolves and crystallizes/recrystallizes inside the pore, the crystallization cycle pressure will produce a certain stress effect on the pore wall, leading to the occurrence and expansion of micro-cracks, and directly destroy them. Meanwhile, it also has an interactive promotion effect on chemical corrosion and promotes connection between pores and cracks. Currently, the understanding of physical damage is largely derived from crystallization pressure theories originally developed for sodium sulfate (Na₂SO₄) systems [44]. The classical hydration pressure model established by Scherer describes the stress generated when an anhydrous salt converts to a hydrated phase under humid conditions (Equation (5)),

$$P_c-P_e={\displaystyle \frac{RT}{V_H-V_A}}\left[\ln \left({\displaystyle \frac{K_A}{K_H}}\right)+{\alpha }_{H_{2}O}\ln \left(RH\right)\right]$$

(5)

where P_c is the equilibrium pressure (Pa), P_e is the generated pressure (Pa), R is the gas constant (7.314 J·mol⁻¹·K⁻¹), T is the absolute temperature (K). V_H and V_A are the molar volumes of hydrated salt and anhydrous salt, respectively; α_H2O is the number of molecules of H₂O in the hydrate crystal product; K_A/K_H represents the equilibrium constant ratio; and RH is the ambient relative humidity.

However, directly applying this Na₂SO₄-based model to magnesium sulfate (MgSO₄) requires a clear distinction.

(1)

Theoretical Limitations: Equation (5) assumes a distinct phase transition driven by humidity (like Thenardite → Mirabilite). Applying this to MgSO₄ is partially speculative. Unlike sodium sulfate, magnesium sulfate exhibits multiple metastable hydration states (e.g., kieserite, hexahydrite, epsomite) rather than a single direct transition. This complexity makes it difficult to accurately determine the specific volume change (V_H-V_A) and thermodynamic parameters in a real fluctuating environment [45].

(2)

Validated MgSO₄ Mechanisms: Distinct from the theoretical assumptions based on sodium salts, experimental evidence has confirmed specific physical damage pathways for magnesium salts. The primary validated mechanism is the multi-step hydration expansion. MgSO₄ hydrates evolve through various forms (MgSO₄·nH₂O, where n = 1, 6, 7). The transformation between these phases generates significant volume expansion. Furthermore, recent studies confirm that in the presence of sodium ions [46], the crystallization of double salts (specifically Na₂Mg(SO₄)₂·4H₂O) generates local stresses exceeding 28 MPa, acting as a key driver for surface scaling and micro-cracking in marine environments.

Recent studies show that MgSO₄ damage is more complex than traditional Na₂SO₄ mechanisms [44,47]. First, repeated crystallization and dissolution generate fluctuating pressure inside pores, initiating cracks. Second, phase transitions between hydrate forms (MgSO₄·nH₂O; n = 1, 6, 7) during wet–dry cycles induce substantial volume changes. This exerts high stress on pore walls.

Experimental results from Chen et al. [48] align with this mechanism. They found that concrete subjected to wet–dry cycles in 8% MgSO₄ solution suffered a damage coefficient 2.4 times higher than in a 3% solution, accompanied by a dense crack network. SEM analysis further revealed the crystallization of “double salts” (Na₂Mg(SO₄)₂·4H₂O) on the partially immersed surface. The growth of these crystals generates local stresses up to 28 MPa, exceeding the tensile strength required for microcracking. Consequently, crystallization stress drives rapid crack development, explaining the faster spalling rates observed in the splash zone compared to the full immersion zone. Lin et al. [49] performed wetting–drying cyclic tests and found that in the early stages of the test, the elastic modulus increased by approximately 20.6%, possibly due to the initial formation of corrosion products causing a short-term “filling effect”.

The above coupled chemical-physical mechanism continuously destroys the strength properties of concrete material and changes the pore structure characteristics, which in turn seriously affects the long-term service performance of marine concrete structures in the future. Although, from a macro perspective, it has been identified, the mechanism underneath still needs further discussion. The mechanism of C-S-H decalcification has been clearly understood; however, the direct use of the sodium-based thermodynamic model for magnesium salts, due to their complex hydrated species, is still controversial. So is the physical effect: at low concentrations, it may be temporarily filled into pores, but when it is too high, it will cause rapid expansion crazing.

To summarize, it is well established that crystallization stress causes rapid cracking and spalling, particularly in splash zones. However, the theoretical modeling of this mechanism remains uncertain. Specifically, applying sodium-based thermodynamic models to magnesium salts is controversial due to the complexity of magnesium’s hydrated states. From an engineering perspective, protection strategies must prioritize the splash zone. In this zone, physical destruction driven by wetting–drying cycles develops significantly faster than chemical corrosion in fully immersed areas.

2.3. Coupled Corrosion of Magnesium Chloride and Magnesium Sulfate

Under complex environments such as coast, salt lake, underground brine zone, etc., concrete structures are usually in the coexistence situation of several kinds of harmful ions (Cl⁻, SO₄²⁻, Mg²⁺); therefore, its corrosion rate is much higher than that of those in single-salt conditions, and has obvious non-linear property and stage-evolutionary features.

In order to reveal the characteristics of spatial distribution of coupled corrosion in mixed salts environments, Jakobsen [50] put forward the typical eroded profile model. The four characteristic corroded zones, from outside towards inside, were defined as: shell layer; magnesium salts corrosion zone [Mg-zone]; sulfate corrosion zone [S-zone]; chloride front zone [Cl-zone]. Based on this theoretical model, the schematic zonation and corresponding elemental profiles are illustrated in Figure 4.

It should be noted, though, that this zoning is not fixed; its development and thickness depend largely on the mixture proportions, e.g., a lower W/B will result in a compact matrix that will prevent further penetration of the ions [51]. Appropriate curing will provide a dense surface ‘skin,’ which would be mainly responsible for preventing the formation of the outer ‘Mg zone.’.

This layered distribution was further verified by observation under an optical microscope and SEM–EDX analysis: the migration distance of Cl⁻ is farthest, SO₄²⁻ is mainly distributed in the middle layer, and Mg²⁺ still accumulates near the surface. This kind of “zonal invasion–competitive reaction–synergetic destruction” of Cl⁻, SO₄²⁻, and Mg²⁺ in cementitious materials pores (the three ions co-intrusion and co-corrosion way) has been shown clearly. Therefore, compared with the corrosion damage caused solely by chlorides or sulfates, the joint attack of magnesium chloride and magnesium sulfate is more harmful and more difficult to study; it will bring new challenges for the durability design and evaluation of marine concrete structures.

2.3.1. Mechanisms of Coupled Corrosion

Under the coexistence conditions of Cl⁻, SO₄²⁻, Mg²⁺ in the mixed salt solution environment, due to differences in migration capacity, reaction sequence, product generation, and so on, of these three kinds of ions, they are intertwined together in spatial distribution and interaction mechanisms and finally produce coupled damage effects. Jin et al. [52] considered that when concrete was subjected to multispecies ions, Cl⁻ has a higher migration ability and rapidly diffuses into the concrete matrix in the early stage of corrosion, reacts with AFm phase first, generates Friedel’s salt, and forms local diffusion barriers. Subsequently, when SO₄²⁻ invades the system and exchanges anions with Cl⁻ fixed by the barrier film, it will cause Cl⁻ to detach from the barrier and migrate deeper, forming a “fixed-replaced-released” cycle. At the same time, Mg²⁺ continuously leaches Ca²⁺ from C-S-H gel, which may promote the formation of M-S-H phase with poor binding ability. Li et al. [51] believed that the synergistic corrosion mechanism among them is mainly dominated by concrete’s porosity, ITZ’s density, W/B ratio, etc. Then, Zhang et al. [53] proposed that the M-S-H phases produced by Mg²⁺ not only reduce the migration resistance against Cl⁻/SO₄²⁻ diffusion but also increase the number of connected micro-cracks, thereby inducing degradation processes: competition for cation coordination, disruption of the original network structure, extension of corrosion paths. The pore evolution process caused by the coupled erosion of sulfate ions and chloride ions, i.e., from the filling of expansion products to the final triggering of expansion cracking, is shown in Figure 5.

2.3.2. Modeling Analysis of Coupled Corrosion

In order to predict the coupled corrosion behaviors involving ion migration, reaction kinetics, structure evolution, and so on, many kinds of modeling methods have been proposed. Instead of taking them separately, it is necessary to compare the different modeling frameworks from the perspective of the trade-off between physical faithfulness and engineering application.

(1) Ion Transport-Reaction Coupling Models

This framework, primarily rooted in Fick’s second law, represents the most engineering-oriented approach due to its high parameter identifiability. Models like the one proposed by Chen et al. [55] introduce kinetic terms (Reaction rate R, binding coefficient K₁) into the diffusion equation to describe the “migration–adsorption–redesorption” cycle,

$${\displaystyle \frac{\partial C_f}{\partial t}}={\displaystyle \frac{\partial }{\partial x}}\left[D_c\times \left({\displaystyle \frac{1}{1+K_1\times R}}\right)\times {\left(1-K_1\times K_z\times q\right)}^2\times {\displaystyle \frac{\partial C_f}{\partial x}}\right]$$

(6)

Here, C_f is the free chloride concentration at depth x. The standard diffusion coefficient D_c is modified by the reactivity parameter R, the sulfate reaction fraction q, and the filling coefficient K_z. While this approach offers low computational cost and simple inputs suitable for rapid engineering estimation, its physical realism is limited. It often treats complex thermodynamic coupling as simplified empirical coefficients, making it less accurate for strongly coupled multi-ion environments.

(2) Multiphysics Coupled Model (PNP-Thermodynamics-Damage)

To address the theoretical limitations of Fickian models, frameworks based on the Nernst–Planck–Poisson (PNP) equations offer higher physical realism. Yu et al. [56] derived a set of equations coupling electrical charges, ion concentrations, and film potentials, incorporating a reactive damage function H_D(C_d) to dynamically characterize diffusivity changes,

$$H_D\left(C_d\right)=\left(1+{\displaystyle \frac{32}{9}}{C}_d\right)+D_p$$

(7)

where H_D acts as the diffusion coefficient multiplier, determined dynamically by the crack density C_d and the crack-width dependent diffusivity D_p. This framework provides a rigorous theoretical basis for capturing electric field coupling and thermodynamic equilibrium. However, it suffers from low parameter identifiability; input parameters such as microscopic ion mobilities and reaction rate constants are difficult to measure in situ, limiting its engineering applicability primarily to detailed mechanism analysis rather than routine service life design.

(3) Chemomechanical-Structural Feedback Model

Representing the frontier of physical realism, phase-field methods (PFM) coupled with dual-porosity mechanics, as developed by Wang et al. [57], can visually simulate the “corrosion–cracking” feedback loop. The pore evolution is explicitly modeled as,

$${\phi }_t=max\left[\left({\phi }_0-{\phi }_p,0\right)\right]$$

(8)

where the time-dependent porosity ${\phi }_t$ is updated by subtracting the volume of expansive products ${\phi }_p$ (e.g., Friedel’s salt, gypsum) from the initial porosity ${\phi }_0$. The primary strength of this approach lies in its ability to explicitly simulate the complex topology of crack propagation and its feedback on ion transport. However, this method has significant limitations. It requires high computational costs and involves complex mathematical formulations. Consequently, it is currently unsuitable for large-scale structural predictions. Nevertheless, it remains an invaluable tool for theoretical verification at the micro-scale.

(4) Solution Methods and Frontier Developments

The finite element method (FEM) and finite difference method (FDM) are still two mainstream numerical simulation technologies at present. For example, Yin et al. [37] developed a coupled finite element model of the diffusion of chlorides, corrosion of steel bars, propagation of cracks and so on based on FEM, realized the unified simulation for the chemical–mechanical coupling process; Chen [58] put forward a two-dimensional variable coefficient finite difference scheme, tracked the migration front of Friedel’s salt during migration as well as development process of the SO₄²⁻ reaction zone, and controlled the relative error between the simulation results and experimental data within 23.5%.

Although many mathematical models have been developed, the maturity level of different models is quite different: early theoretical models can give quick predictions but rarely receive strict examination when multi-ion reactions couple together; while some numerical reactive transport models show good performance. Especially those combining chemical equilibrium calculation with ion fluxes predict reasonable evolutions of phase assemblage and swelling strain rate in comparison with long-term laboratory experiments [59]. But up to now, most verifications are still based on the results from laboratory accelerated tests instead of long-term site inspection under true seashore conditions. The development of verified models according to site monitoring results is still very limited in future studies.

Recently, Peridynamic (PD) methods [59], and MPFM (Multiphysics Phasefield Model) [60,61,62], etc., were proposed for efficiently simulating the coupling evolution process of corrosion–cracking coupling processes, which will be used in further research on multi-factor coupled degradation mechanism under the coupled Cl⁻–SO₄²⁻–Mg²⁺ corrosion environment. In addition, very recently, Chen et al. [63] proposed a new “TCTD” coupled model from the point of view of the coupling relation among mechanisms: (a) Ion migration; (b) electrochemical corrosion; (c) release amount/concentration of Ca²⁺; and (d) pore formation evolution process considering the effect of temperature field. The output of this model is the spatial distribution image of Friedel’s salt, AFt, gypsum, etc. Thus, it can predict the durability performance under the coupled environment of multiple types of ions. Recently, the numerical study [64] by Bastidas-Arteaga et al. takes COMSOL Multiphysics^® as the numerical analysis tool to consider the time-dependent diffusion coefficient coupled with surface chloride accumulation, and the predicted results showed that a prediction deviation lower than 15% can be obtained.

It is crucial to distinguish “engineer tool” and “academic tool.” In terms of this point, it has been acknowledged that the analytical model is the unique and feasible option for engineer application since its input (diffusion coefficient) can be obtained through common site inspection. Otherwise, the multiphysics model is only feasible for academics because some microstructure parameters cannot be measured under real situations. Yet, the absolute credibility of any framework is still under debate due to no long-term verification up to now. Therefore, we need to choose according to different situations: using the analytical model for service life prediction or adopting the numerical tool for fundamental study. The specific comparison can be found in

Table 4. Critical evaluation of modeling frameworks for coupled corrosion.

Modeling Framework	Strengths	Limitations	Applicability	References
Analytical Models (Diffusion-Reaction)	High computational efficiency; Simple inputs; Analytical solutions available	Ignores thermodynamic equilibrium; Low accuracy for multi-ion coupling	Rapid service life estimation for simple geometries	[55,64]
Multiphysics Models (PNP-Damage)	High accuracy; Rigorous theoretical basis; Captures electric field coupling	Requires extensive input parameters; Sensitive to boundary conditions.	Detailed durability design for critical infrastructure	[56]
Chemo-Mechanical Models (Phase-Field)	Visually simulates crack propagation; Captures “corrosion-cracking” feedback	Extremely high computational cost; Difficult to simulate large-scale structures	Theoretical research and mechanism verification	[57,61]
Numerical Schemes (FEM/FDM)	Flexible for complex geometries; Mature commercial software support (e.g., COMSOL)	Mesh-dependency issues; Long-term validation data is often lacking	Engineering applications with complex boundary conditions	[37,58]

.

To provide a clear comparison of the applicability and limitations of these frameworks, a critical evaluation is summarized in Table 4. Regarding the trade-off between predictive accuracy and practical usability, it is well established that numerical methods offer superior precision for multi-field coupling, whereas analytical models provide computational efficiency for quick estimates. However, the absolute reliability of all current frameworks remains controversial, strictly limited by the ‘time-scale gap’ between short-term lab parameters and decades-long field exposure.

From a practical perspective, it is crucial to distinguish between these models based on their utility. Empirical and semi-analytical models (e.g., Fick’s law variants) are computationally efficient and robust, making them highly suitable for engineering service-life prediction and standard design codes. Conversely, complex coupled models (such as the thermodynamic and phase-field approaches discussed above) serve primarily as academic tools. Their value lies in the theoretical verification of micro-mechanisms and understanding topological evolution, rather than widespread application in large-scale structural engineering due to their high computational demand.

3. Anti-Corrosion Measures

In view of the above problems caused by the magnesium cations, in recent studies, researchers have paid more attention to how to improve the denseness of microstructures, block the entry channels of harmful media, and stabilize the interface zone. In this paper, in order to make a systematic summary of all kinds of protection measures proposed so far, we divide them into three categories: material optimal configuration, surface coating treatment, and cement matrix regulation (Figure 6). When it comes to co-corrosion of multiple cations (Cl⁻, SO₄²⁻, and Mg²⁺), these three measures cooperate with each other from the aspects of external barriers, internal corrosion inhibitors, and matrix reinforcement, respectively, to form an integrated three-dimensional protection system.

In summary, the whole protection measures mainly involve the following three aspects: material modification and reinforcement (nano-SiO₂, metakaolin, hydrophobic admixture, etc.), surface film treatment (PAC coating and organic/inorganic hybrid film), cement matrix regulation (CSA, magnesium phosphate binder system, low water/binder ratio design), which can jointly enhance concrete’s micromorphology compactness, resist ion migration capacity, and interface stability, etc., preventing Cl⁻-SO₄²⁻-Mg²⁺ coupled corrosion from occurring. According to the above idea, in the following sections, we will successively discuss the exact anti-corrosion mechanism and application performance of each measure in material aspect, surface protection aspect, and cementitious system aspect.

Thus, the protection strategies introduced hereafter are described based on three aspects according to the above degradation mechanism analysis in Section 2: material modification (in Section 3.1) against Mg²⁺-induced chemical decalcification (in Section 2.1); interface control (in Section 3.2) against coupled Cl⁻ diffusion (in Section 2.3); and matrix densification (in Section 3.3) against SO₄²⁻ crystallization expansion (in Section 2.2).

3.1. Material Anti-Corrosion Design

Mainly through the coupled use of mineral admixture improvements, Li et al. [65] found that UFA could react secondarily with Ca(OH)₂ to generate more C-S-H gel, refine pore structure, improve microstructure compactness, and restrain ettringite generated under SO₄²⁻ environment. Metakaolin (MK) has high aluminosilicate activity and may promote C-A-S-H gel formation, increase the amount of fixed Cl⁻; GGBS can regulate the Ca/Si molar ratio and solution pH value of pores, stabilize C-S-H gel, prevent the transformation of M-S-H under Mg²⁺ condition, etc. From another point of view, “MK + slag”, “UFA + MK” blended systems show better synergetic effects of resisting Cl⁻, SO₄²⁻, and Mg²⁺ attack; however, the amount of addition and pH adjustment are also quite crucial—too much MK content (>30% mass fraction) will lead to a more compacted microstructure, but too low pH will have bad effects on both strength development and subsequent long-term performance [66].

In addition, for the alumina-rich mineral admixture (high-alumina slag/Class C FA) applied under the magnesium salt environment containing sulfate: On one side, the higher aluminates content can effectively bind more chlorides by the generation of Friedel’s salt to block the migration path of chloride ion migration [67]; on the other side, it also brings about a “trade-off”. When there are MgSO₄, the aluminates in the cement matrix will react with sulfate ions to produce expansive ettringite. In addition, the initially fixed Friedel’s salts may become unstable and release free chlorides again if the pH decreases due to Mg attack or sulfate concentration increases (“chloride release”) [68]. Thus, the amount of alumina-rich admixtures should be reasonably adjusted so that the balance between the ability of fixing chlorides and resisting sulfate is reached.

The functional chemical admixture also shows good performance in multi-salt systems. It was found by Zhang et al. [69] that when SBT-TIA bifunctional hydrophilic-hydrophobic anti-corrosion agents were added into OPC-SF systems under high SO₄²⁻ conditions, the strength of mortars can still be maintained at about 79%. This might be because the hydrophilic group has a chelating effect on Ca²⁺ to produce nanoscale barrier films, thereby blocking the migration path of SO₄²⁻. Meanwhile, the hydrophobic group can reduce capillary adsorption and finally achieve physical isolation together with chemical sustained-release protection effects.

The future development strategy of new materials needs to be combined with microstructure (dense), reaction regulation, and crack control, to build a multifunctional “win-win” protection system. Careful choices must be made to achieve this balance: the effect of mineral admixture on fine-tuning the pore structure has been recognized, but there is still controversy about whether it will cause the secondary expansion when using an alumina-rich precursor under a sulfuric acid attack environment; therefore, it depends on specific situations and we need to find a good balance between the two aspects—chloride binding ability vs. anti-sulfate performance [70,71].

3.2. Interface Protection Strategies

Constructing a stable, dense, well-adhered surface protection film is another important way to improve its service performance. Polyurethane, acrylic resin and epoxy resin coatings are all high-density, deformable films with good resistance against chloride ion attack [21], as shown in Figure 7. According to Yin et al. [72], the permeability coefficient of chloride ions through the polyurethane coating could be reduced by roughly 86%, and after three years of immersion testing, the penetration depth was less than 20 mm. Due to the good adhesion properties of acrylic resins, they were more suitable for the parts where stress concentration easily occurred on bridges. Epoxy resins have high cross-linked structure characteristics, which can intercept both chloride ions and oxygen migration coupling pathways at the same time, delaying the passivation time of steel bars’ films. But under the environment of high temperature and humidity changes and so forth, cementitious coatings had better performance than organic coatings. Adding styrene-acrylates and a small amount of TiO₂ can effectively increase their adhesion strength and retain the properties after UV aging treatment [19].

Recently, the damage caused by erosion was reduced by the modification of silica. In this process [73], hydrogen methyl silane (HMDS) is used as the modifier. The -Si(CH₃)₃ groups in HMDS react chemically with the -OH groups on the surface of silica, thereby allowing the highly active -Si(CH₃)₃ groups in HMDS to bond with the surface of silica. As a result, the content of -OH decreases, and the surface of silica changes from hydrophilic to hydrophobic, increasing its service life by nearly 80%. The reaction equation is shown in Figure 8.The silane/clay nanocomposites can decrease the surface chloride content below 92% under harsh environments with good impermeability, and multilayer gradient coatings (such as ZrO₂/TiO₂/Al₂O₃) have good anti-thermal shock properties, wear-resistance, and corrosion protection performance and are applied in some special equipment like offshore wind turbines and wharf facilities [70]. And the other biomimetic superhydrophobic coating imitates some natural structures (like oyster shells). The stable film formed in this way can prevent water intrusion even when the environment is salty.

The effectiveness of the coating is not only related to its own material properties, but also closely related to the treatment of the interface before adhesion and monitoring during use. As shown in Figure 9, non-destructive testing methods, such as the acoustic emission method and the infrared thermography method are able to detect delamination and aging in advance and guarantee long-term protective effectiveness. Research on protection technology for the interface is pursuing development from the current single barrier toward the direction of densification + function + inspection integrated design and providing more stable external protection for the cementitious structure under the complex marine environment.

The barrier efficiency of coatings against ion ingress is well established. Yet, their long-term adhesion to moist concrete substrates remains controversial and prone to failure. The actual protective service life is therefore context-dependent, relying heavily on the quality of surface preparation and environmental exposure conditions.

3.3. Anti-Corrosion Cementitious Materials

In conclusion, non-silicate cementitious materials and functionally modified measures are two effective means to enhance corrosion resistance properties. Phosphate aluminate cement (PAC) and calcium sulfoaluminate cement (CSA) are two kinds of low-alkali, dense hydrated structure cements, which have excellent chemical stability in Cl⁻ and SO₄²⁻-rich environments. As reported by Liang et al. [74], after soaking in a NaCl solution at a concentration of 10% for 120 days, the PAC coating penetration depth was detected by the chloride diffusion coefficient test, which is only 33.24 μm. Compared with ordinary Portland cement systems, this value is significantly lower. Additionally, the water absorption of CSA after being soaked in water for 72 h is only 6.10%, indicating that the pores have little connectivity and water permeability.

The polymer-modified cementitious system has a good modification effect as well. Adding acrylic emulsion or epoxy resin to ordinary Portland cement does not reduce its strength, but increases its bond strength by 16.67%, meanwhile it also increases its toughness and self-healing capacity [75,76]. Adding other additives such as nanometer TiO₂ can adjust its pore structure, further improving its water impermeability performance and compressive performance, etc. The CAC system, which is an easy to produce phase transition during later periods, leads to a decrease in material’s strength with time going on. Recent studies mainly concentrate on how to stabilize hydration products. Yoo et al. [29] found that GGBFS had the best resistance against chloride penetration among all blended cements tested, and GGBFS not only had a good effect on sound specimens, but also on cold joints specimens. The results of the chloride diffusion coefficients of 91-day-old samples are shown in Figure 10. Under no load and cold joints effects, the lowest diffusion coefficient of GGBFS concrete is 6.6 × 10⁻¹² m²/s, which is only 27.8% of the diffusion coefficient of OPC concrete.

While all the above-mentioned strategies showed good performance under laboratory conditions, they differ greatly concerning practicability from an engineering point of view. This is due to different issues related to upscaling and failure mechanisms of both surface protection concepts, which may debond due to hydrostatic pressure when aged [77], and new nanomaterials, which are still costly and difficult to disperse homogeneously at larger scale [78]. In terms of “critical synthesis,” this results in clear precedence of approaches: modification of binders (low w/b mixed with SCMs) was found to be the most robust and easily scalable main barrier; whereas, the long-term reliability of external coatings/new materials alone is still under debate. Thus, in practice, we believe that the solution should be based on the circumstances: take a hierarchical approach—first form a compact cementitious bulk and then apply surface treatment—to achieve long-term resistance against attack from MgSO₄ solution. A comparison of all anti-corrosion measures is summarized in

Table 5. Comparative analysis of major anti-corrosion measures.

Category of Measures	Primary Mechanism	Advantages	Limitations
Mineral Admixtures (e.g., Fly Ash, Slag, Silica Fume, etc.)	Dilution effect and pozzolanic reaction to refine pore structure and reduce ion diffusion rate	Cost-effective, utilizes industrial by-products; significantly reduces Cl− diffusion and SO42− reactions	Limited reactivity, lower early-age strength; restricted inhibition of Mg2+-induced decalcification
Functional Cementitious Materials (e.g., CSA, Magnesium-Based Binders, etc.)	Rapid formation of ettringite/other non-expansive products to mitigate sulfate attack; strong Mg2+ binding capacity	Highly targeted; significantly enhances resistance to SO42− and Mg2+ attack	Relatively high cost; limited compatibility with ordinary cement systems
Surface Protectives/Coatings (Hydrophobic Agents, Penetrating Crystallization Agents, Polymer Coatings)	Barrier effect to reduce water and ion penetration; some coatings possess self-healing properties	Easy to apply, rapid effect; particularly effective against Cl−-induced corrosion	Limited durability; potential for cracking or delamination; requires periodic maintenance
Composite Material Systems (Nanomaterials, Graphene, Fiber-Reinforced Systems)	Enhances microstructural density, improves mechanical toughness, adsorbs/fixes part of the aggressive ions	Simultaneously improves mechanical performance and durability; suitable for complex coupled corrosion environments	High cost; large-scale application remains limited
Structural and Design Optimization (High-Performance Concrete, Low Water-to-Binder Ratio, Protective Layer Thickness Design)	Reduces ion transport pathways and penetration efficiency through design measures	Can be combined with material optimization for synergistic effects; high engineering feasibility	Requires optimization based on environmental conditions; still limited under extreme environments

.

4. Discussion and Future Directions

4.1. Discussion

Although there is a good understanding of degradation mechanisms for concrete under magnesian environments, important knowledge still needs to be bridged from basic study to practical application, i.e., existing knowledge mainly comes from research within one single salt solution instead of realistic multispecies solutions. Under actual marine or hypersaline lake conditions, chlorides, sulfates, and magnesium coexist; therefore, complicated “competition reactions” and “layer-by-layer invasion” effects occur, which cannot be found when studying only one type of interaction.

As for service life prediction, however, the current research is mostly based on the diffusion-reaction model; although some new tools, such as the phase-field method and PNP model, were proposed recently, these two models highly depend on many empirical parameters, which cannot be obtained easily in practice. In addition, they also ignore the positive feedback effect of cracks connection and corrosion products on ion migration, causing the predicted degradation rate to be much higher than the actual value.

In terms of material protection strategy, successful experience under single salt conditions cannot be directly applied into coupled environments. Taking high alumina mineral admixture for example, although it has good ability to capture Cl⁻, it can induce thaumasite sulphate attack under Mg²⁺-rich circumstance; another typical case is related to surface coating/hydrophobic treatments, which show excellent performance at early ages, but age, crack and even debond after being immersed for a long time, so it cannot be regarded as a reliable barrier alone in long-term service.

4.2. Future Directions

In order to solve the above problems, it is necessary to transform the research direction of future studies from “single-mechanism protection” to “systematic synergetic effect”; in particular, in view of how to improve the predictive foresight of the model, we need to change the current fixed laboratory fitting method into iterative online calibrations based on field monitoring information. By taking advantage of inverse analysis technology (Bayesian inversion, etc.), it will become possible to continuously calibrate important model parameters (diffusion coefficient, etc.) according to limited field detection data; at the same time, monitor-based verification can connect the numerical model with the real-time acquisition sensor signals (Figure 9), so as to correct the prediction trajectory in real time during the simulation process, thereby reducing the prediction uncertainty of long-term forecast.

In addition to model optimization, future designs should focus on building an integrated multi-functional protection system against multi-ion co-corrosion. This involves not only matching graded mineral admixtures to balance the Ca/Si ratio but also developing bio-mimetic multi-layer coatings for smart barrier performance. A multi-level defense strategy must be constructed to simultaneously resist chloride diffusion, sulfate expansion, and magnesium-induced decalcification.

In conclusion, in terms of the development of the research itself, it will continue to advance towards the construction of a cross-scale research system that can quantitatively connect chemical changes at the micrometer level with the performance of the structure at the macroscopic scale. Take the RVE analysis method based on the homogeneous model as an example, further mapping the kinetic parameters describing phase change at the micrometer level (such as M-S-H generation) into damage evolution at the mesoscale (such as crack density), which would continuously reduce the macroscopic transport tensor used for predicting the life of the structure, establishing a multi-physical field simulation analysis tool coupled with an in situ monitoring database, which could realize the prediction of concrete performance from the material to the structural level.

5. Conclusions

In summary, this review highlights that in-depth study of the coupled degradation mechanism and reasonable formulation of the synergetic protection strategy are important ways to improve the durability performance and prolong the service time of concrete material under magnesium salts environments. Based on this work, some main conclusions could be obtained, as follows:

(1)

Magnesium salts environments multi-component cation coupling damage mechanism. Concrete damage under magnesium salts action is a coupled reaction process among various cations, Cl⁻ migrates into the interior of concrete by diffusion, capillary suction, and electromigration, combines with the AFm and C-S-H phase to form Friedel’s salt, destabilizes the passivation film on steel rebars and causes local pitting corrosion. SO₄²⁻ reacts with Ca²⁺ and Al³⁺ in the pores to produce gypsum and ettringite, a kind of expansive hydration product that produces expansive stress and propagates cracks. Mg²⁺ reacts with CH and causes the decalcification transformation of C-S-H into non-cementitious M-S-H, destroying the strength skeleton. In the presence of different cations, the destruction effect will be strengthened by the “fixation–displacement–releasing” cyclic dynamic changes among them and positive feedback between crack development and penetration.

(2)

The spatial-temporal development of ion controls the damage pattern. In multi-salt environments, the spatial distribution and sequence of time of the ions controls the mode of deterioration, i.e., Cl⁻ locates at the deepest penetrating front, SO₄²⁻ concentrates at the middle layer of concrete as an expansive product, while Mg²⁺ is mainly located at the concrete surface as Mg(OH)₂ and promotes decalcification and M-S-H phases. Moreover, together with the effects of the wet–dry cycle, salt crystallization/thawing stress, and initial cracks, the coupled physicochemical damage is triggered, which has nonlinearity and coupling characteristics. Different model frameworks based on diffusion-reaction equation (D-R), Nernst-Planck-Poisson (NPP), phase field method (PFM) or multiphysical field method (such as TCTD) have been proposed for quantitative description of coupled mechanisms between ion migration, binding/substitution reactions with mineral particles, and pore-crack evolution.

(3)

Hierarchical and synergetic material strategies. The measures for mitigating magnesium sulfate corrosion damage based on materials have hierarchically interacted and produced a synergetic effect. Taking mineral admixture as an example, ultrafine fly ash, metakaolin, and ground granulated blast furnace slag can not only reduce ion diffusion rate but also improve C-S-H gel stability by pozzolanic reaction (pozzolanic effect) and fine pore structure. But attention must be paid to the proportion of mineral admixture and the total alkali degree of the system so that adverse change caused by Ca/Si ratio is avoided. For functional admixtures with dual functions of hydrophilicity, hydrophobicity, and anti-corrosion agents, such as amine and nitrite corrosion inhibitors, they can effectively inhibit SO₄²⁻ exudation amount and increase the anti-corrosion potential value of reinforcement. In addition, the indirect way of increasing fiber dosage and decreasing W/B ratio can block crack connectivity and reduce ion migration speed.

(4)

Surface/interface protection and improvement of cementitious matrix. The high-density surface coating (polyurethane, epoxy, acrylic, and nanocomposite coatings), as well as some types of water-repellent or crystallization admixtures, can effectively block surface water invasion and ion migration rates; CSA, PAC, modified polymer, or nanocomposite binder with functional groups can effectively reduce porosity and change the hydration products to resist the attack from SO₄²⁻/Mg²⁺.

(5)

Based on the complex degradation mechanisms, practical durability design in magnesium-rich environments should prioritize material optimization. Specifically, the incorporation of Supplementary Cementitious Materials (SCMs), such as fly ash or silica fume, is highly recommended. These materials consume calcium hydroxide to mitigate M-S-H formation while refining the pore structure to slow ion diffusion. Furthermore, maintaining a low water-to-binder ratio and ensuring adequate curing are critical to creating a dense surface skin that acts as the primary barrier against magnesium and chloride penetration. Finally, durability assessment must transition from single-salt models to multi-ion coupling frameworks to avoid underestimating the structural service life.