Corrosion-resistant cladding strategies for molten salt reactor components: Ni-based alloys and laser cladding on 316H stainless steel

1 Introduction

Molten salt reactors, one of the six Generation IV systems, use fluoride or chloride melts as coolant or as a combined fuel-coolant at 550 °C–700 °C and near-atmospheric pressure (Serp et al., 2014). The chromium oxide films that stainless steels and nickel superalloys rely on are thermodynamically unstable in halide melts, so protection must come from the alloy itself or an engineered surface (Guo et al., 2018). The Molten Salt Reactor Experiment (MSRE) answered in the 1960s with a purpose-built alloy, Hastelloy N; however, the program closed with unresolved tellurium embrittlement (McCoy et al., 1970), and 6 decades later Hastelloy N still lacks an established Class A design basis under ASME Section III, Division 5 (Wright and Sham, 2018).

Current developers have responded by balancing corrosion performance, material availability, and nuclear-code qualification. Kairos Power’s approach, accepted by the U.S. NRC in a final safety evaluation in 2023, justifies the use of bare 316H for FLiBe-wetted structures through tight beryllium-based redox control (Kairos Power, 2021). By contrast, the U.S. DOE is pursuing the long and costly path of generating the creep and creep-fatigue data needed to bring Hastelloy N into the Code (Wright and Sham, 2018). A third strategy, already well known in conventional pressure-vessel practice, lies between these poles: the component is built in 316H, and the salt-wetted surface is clad with a corrosion-resistant Ni-based layer.

Existing reviews cover molten-salt corrosion science (Guo et al., 2018; Zheng and Sridharan, 2018; Wei et al., 2025), irradiation–corrosion synergy (Liu et al., 2024), and laser-cladding process technology (Liang et al., 2023). To our knowledge, previous reviews have not explicitly connected bimetallic fabrication with the ASME design rule pathway or the manufacturing, inspection, and supply chain constraints of MSR components. Accordingly, this Mini Review integrates these previously separate considerations within a single framework for clad 316H components. The literature was selected through focused, non-systematic academic searches and citation tracking, prioritizing primary experimental studies and authoritative technical documents relevant to molten-salt corrosion, cladding and fabrication, and structural qualification, including literature available through August 2026. Section 2 summarizes corrosion mechanisms and the landscape of candidate alloys, Section 3 assembles the evidence for clad 316H components across chemistry, fabrication, and design rules; and Section 4 discusses the remaining industrialization gaps.

2 Corrosion mechanisms and the candidate-alloy dilemma

2.1 Mechanisms

Corrosion in halide melts occurs through the selective dissolution of the least noble alloying constituent, principally chromium and then iron, because no stable passive film can form; depending on the alloy, salt chemistry, and impurity level, the damage appears as relatively uniform depletion, intergranular penetration, or more localized attack (Guo et al., 2018; Zheng and Sridharan, 2018). Several accelerants are superimposed on this thermodynamic driving force. Impurities (e.g., H₂O, O₂, HF, and their reaction products) dominate early-life attack: a meta-analysis of 6 decades of heterogeneous corrosion data found that salt purity is the strongest single correlate of corrosion rate, ahead of temperature and alloy identity (Raiman and Lee, 2018), consistent with mechanistic simulations and impurity–temperature tests (Schneider et al., 2024; Li et al., 2023). Thermal gradients and dissimilar-material couples drive continuing mass transfer even in pure salt: alloy-activity gradients alone can move chromium from hot to cold legs or between alloys that share the same melt (Falconer et al., 2022), and graphite in the circuit roughly doubles the attack depth on 316 stainless steel in FLiBe through carbide formation (Zheng et al., 2015). In fuel salt, fission-product tellurium adds intergranular embrittlement, unresolved at MSRE shutdown and recently mapped thermodynamically in E–pTe form (Surenkov et al., 2020; Ge et al., 2025). Purification and redox buffering, for example, through the U(IV)/U(III) ratio in fuel salt or contact with reducing metals, act most directly on the impurity- and redox-driven components of attack, while gradient-driven mass transfer and mechanically assisted attack persist (Zhang et al., 2018; Surenkov et al., 2020; Ding et al., 2019). One further consequence deserves emphasis: without a passive film, no repassivation arrests damage once it has begun; restoring purity and redox lowers the subsequent rate, but lost material and established grain-boundary paths remain (Guo et al., 2018; Falconer et al., 2022; Gu et al., 2022), so protection has to be built in from the start.

2.2 Candidate alloys and the Division 5 design-basis axis

Measured against this mechanism set, the candidate alloys fall along two axes: corrosion performance and the availability of a Class A design basis under ASME Section III, Division 5 (Figure 1A; Table 1). Hastelloy N (Ni-Mo-Cr, ∼7% Cr) remains the fluoride benchmark. However, its allowable stresses fall steeply above approximately 600 °C–650 °C; it is susceptible to tellurium cracking and helium embrittlement, and it does not have such an established basis (Wright and Sham, 2018). China developed a domestic equivalent, GH3535, as its thorium MSR structural alloy (Liu et al., 2024), illustrating the cost of the monolithic-alloy route. Commercial Ni superalloys with higher chromium content carry more of the very element that the salt removes: in FLiNaK screening of Hastelloy N and X, Haynes 230, Inconel 617, and Incoloy 800H, weight loss scaled with alloy Cr content (Olson et al., 2009), and Inconel 625 develops Cr-depleted subsurface zones in ternary chloride melts at rates strongly dependent on cover-gas chemistry (Xu et al., 2022). Stainless steel 316H occupies the opposite position. It has an established Class A design basis under Division 5, including time-dependent design data, and a viscoplastic constitutive model was added to the 2023 Code edition (Messner, 2022); this basis does not automatically qualify a clad component, fabrication process, or joint configuration. It is inexpensive and widely available; however, with 16%–18% Cr, it dealloys in both fluorides and chlorides (Zheng et al., 2015; Li et al., 2023).

TABLE 1. Candidate materials and clad architectures for MSR salt-wetted service. Bulk alloys are listed together with the two explicitly identified clad architectures. Evidence labels distinguish between direct evidence for the specified material/product form (Direct), indirect or inferred evidence (Indirect), and candidate-level assessments with limited direct data (Candidate). Nominal Cr contents are from manufacturer datasheets; see the text and cited studies for test conditions and limitations.

Therefore, the dilemma is as institutional and economic as it is metallurgical: the alloy with adequate corrosion resistance lacks an established Division 5 Class A design basis, while 316H has such a basis but is not corrosion-resistant. A bimetallic architecture that assigns the structural function to 316H and the salt exposure to a resistant surface layer provides a potential approach for addressing both constraints (Figure 1B).

3 The cladding-based bimetallic strategy

3.1 Chemical feasibility

The chemical stability of the barrier layer is necessary but not sufficient: the clad must also retain continuity, adhesion, and thermomechanical compatibility with the substrate throughout the design life, which is where the industrialization questions of Section 4 (Discussion) arise. For fluorides, the evidence is promising but currently limited to electroplated materials: pure-Ni layers on 316H suppressed Cr dissolution through 1,000 h of immersion in 700 °C FLiNaK (Weinstein et al., 2021). The chloride evidence is mixed. In ORNL screening of cladding candidates in 700 °C NaCl-MgCl₂, refractory metals (W, Mo, Ta) lost almost nothing, while wrought Ni-200 showed the most severe degradation in the set (Raiman et al., 2019); however, Ni clads deposited on 316H outperformed bare 316H and Incoloy 800H after 502 h in NaCl-MgCl₂ at 650 °C, gaining weight by Fe deposition from co-immersed Fe-base alloys (Yoon et al., 2025). The evidence therefore does not support classifying Ni as universally fluoride-compatible but chloride-incompatible: chloride performance depends on salt composition and purification, redox conditions, temperature, exposure configuration, and product form. Nickel has direct protective evidence in FLiNaK; Mo- and W-rich materials have shown promise in chloride screenings, and any deposited clad system requires validation under the target salt chemistry and operating conditions. IN625, the most industrially available option, carries the very chromium the salt removes and shows condition-dependent chloride performance (Xu et al., 2022; Yang et al., 2024). Two further cautions apply: the deposited clad is not wrought material, exposing a dendritic, segregated microstructure whose corrosion behavior differs from plate data (Yang et al., 2024), and every clad edge, weld transition, or through-clad defect creates a dissimilar-metal couple in the same melt, activating the activity-gradient transport discussed above (Falconer et al., 2022; Massimi et al., 2024).

The electroplating results establish only the chemical feasibility. Electrodeposited layers lack metallurgical bonding, and their thickness uniformity, adhesion under thermal cycling, and defect-free coverage have not been demonstrated at the scale of vessel components or complex internal geometries. Therefore, translating this chemistry into hardware requires metallurgically bonded layers, which is where the fabrication question begins.

3.2 Fabrication: Arc overlay, strip cladding, and laser deposition

The available fabrication routes fall into two families (Figure 1C): (i) arc overlay, including its high-deposition electroslag and submerged-arc strip variants, and (ii) laser-based directed energy deposition (DED), including extreme high-speed laser cladding (EHLA). Both are demonstrated for the material pair of interest: Ni cladding was applied to 316H by gas-tungsten-arc weld overlay and by laser DED, characterized, bend-tested without cracking, and corrosion-tested in molten chloride, as discussed in Section 3.1 above (Yoon et al., 2025). This is the closest antecedent to the strategy reviewed here. For large pressure vessels, strip cladding is the established industrial route, covering square meters of shell economically; its limitation is metallurgical rather than economic. Arc processes melt substantial substrate, so iron dilutes the corrosion-resistant layer, and multiple layers are typically needed to meet deposit-chemistry limits (Fe capped at 10% for 625/C-276 overlay deposits; IOGP, 2020).

Laser processes address exactly this limitation. Melt-pool control in laser DED confines dilution far below arc levels, and the acceptable window is quantified for Ni-superalloy-on-austenitic-steel deposition: IN625-to-304L interfaces remain crack-free below roughly 40 wt% steel in the melt pool, with solidification cracking appearing above ∼50 wt% (Stair et al., 2024). This window was measured on 304L; transfer to 316H remains to be confirmed. EHLA extends this, reaching deposition rates an order of magnitude above those of conventional laser cladding, with dilution below 1% and heat-affected zones of tens of micrometers (Liang et al., 2023); high-speed laser-clad IN625 coatings on inexpensive structural steel have shown molten-salt corrosion resistance approaching that of bulk IN625, although validation in purified reactor-grade salt chemistries is still needed (Wang et al., 2025a). Compositional grading between 316-type substrates and IN625 offers a further tool for managing interface chemistry and expansion mismatch (Mehrabi et al., 2023).

Therefore, the two route families are complementary rather than competitive, although their industrial maturity differs: weld-overlay processes are already established among pressure-equipment fabricators, whereas component-scale laser DED and EHLA would require additional equipment and nuclear-specific process qualification. Strip and arc overlay remain the economic choice for large, accessible shell surfaces where multi-layer deposition is tolerable, whereas laser DED and EHLA are preferable where dilution must be minimal in a thin single layer, where distortion must be limited, or where geometry is small and complex (e.g., valve bodies, nozzles, and tube-to-tubesheet joints). Nuclear deployment will require process-specific qualification covering bonding defects, cracking, residual stress, repair procedures, and deposit variability. Head-to-head data on identical MSR-relevant geometries, including cost per protected area, are not yet available and are needed before route selection is codified.

3.3 Design rules for clad components

Implementation of a bimetallic strategy requires a design code framework that accounts for the clad layer. The existing Class A design basis for 316H under Division 5 does not, by itself, qualify specific clad materials, deposition procedures, clad/substrate interfaces, joint configurations, inspection methods, or completed components. A less widely known development reviewed here is the Argonne draft methodology for 316H Class A components clad with nickel or tungsten: a set of rules in Code Case format for treating a thin, non-structural, corrosion-resistant layer in Division 5 creep and creep-fatigue evaluations (Barua et al., 2021). Together with the metallurgical assessment of clad/base interdiffusion (Kirkendall porosity and brittle-phase risk; Young and Sham, 2018), this methodology provides the basis for a future Code Case pathway that could turn cladding from a shop practice into a code-recognized design option (Figure 1D). Currently, the methodology covers nickel and tungsten clads only, and its extension to other clad alloys has not been established; it addresses structural evaluation, while salt-specific corrosion performance and regulatory approval of components require separate demonstrations. These reports remain rarely cited in the corrosion literature; connecting these two research areas is one purpose of this review.

4 Discussion: industrialization gaps and outlook

The assembled evidence supports a clad 316H architecture as a promising candidate for integrated qualification of MSR salt-wetted components, rather than an already near-term solution. The strategy also hedges a practical vulnerability of chemistry-control-only approaches: because impurity level is the dominant corrosion variable (Raiman and Lee, 2018), these approaches must keep cover-gas and salt purity discipline intact for the plant’s lifetime, while the number of penetrations, seals, and maintenance operations grows with system scale. A continuous clad layer adds a line of defense that does not depend on continuous active control; it complements, rather than replaces, purification and redox management, and it can be compromised by local defects or off-normal conditions. The support assembled here also comes from independent studies with different materials, salts, and exposure conditions; no single study has demonstrated a bonded clad on 316H in purified reactor salt under flowing, thermally cycled conditions with representative joints, irradiation, and inspection included. Five gaps stand between demonstration coupons and licensed hardware.

4.1 Irradiation response of clad layers and interfaces

Irradiation–corrosion coupling in salts is not intuitive: proton irradiation decelerated intergranular corrosion of Ni-Cr alloys in fluoride salt through radiation-enhanced healing of grain-boundary depletion (Zhou et al., 2020), while defect and grain-boundary kinetics can push the synergy in either direction (Zhu et al., 2021; Liu et al., 2024). No comparable data exist for deposited clad microstructures or diluted, graded interfaces. Furthermore, the thermal-neutron chain ⁵⁸Ni(n,γ)⁵⁹Ni(n,α) generates He and can embrittle Ni-rich materials at high fluence (Stopher, 2017). This introduces an additional composition-selection trade-off for pure-Ni and Ni-base clads. Mo- or W-rich candidates reduce reliance on Ni and have shown promise in chloride screening (Raiman et al., 2019); however, their fabrication, joining, irradiation, and qualification evidence are less mature. Grading can manage dilution and thermomechanical mismatch, but it does not inherently remove irradiation concerns on a Ni-rich surface. Because comparative irradiation data are unavailable, no universal ranking is justified. Initial deployment is therefore more defensible in lower-flux components such as vessels, valves, and piping.

4.2 Inspectability

Volumetric examination of columnar-dendritic Ni clad on an austenitic substrate is acoustically difficult: beam skewing and attenuation in dissimilar-metal structures are well documented (Hwang et al., 2020), and code-compliant ultrasonic testing conventionally requires machined surfaces, which has kept inspection out of the deposition process. In-process phased-array approaches are only now emerging (Nicolson et al., 2024), and machine-learning-assisted ultrasonic analysis is a plausible route to reliable clad-bond and dilution QA (Sun et al., 2023). Current codes contain general provisions for examining clad and weld overlay, but MSR-specific acceptance criteria that link bond integrity, surface chemistry, and permissible defect size to design life do not yet exist; developing them is a near-term standardization task.

4.3 Joining and assembly

Clad components must eventually be joined to one another, and every joint interrupts the cladding. The most critical case is the girth weld between internally clad spools or between a clad nozzle and piping: welding locally consumes the clad layer, and in small-bore or long runs, the inner surface of a completed field weld cannot be reached for repair cladding. Offshore clad-pipe practice answers this with nickel-alloy root passes and pipe-end overlays and can hold iron dilution in the internal weld bead to a few percent (Alvarez et al., 2024; IOGP, 2020). Orbital welding studies on austenitic piping add a further variable set: welding position and pulse-current parameters measurably alter bead geometry, microstructure, and mechanical properties in 304L girth welds (Widyianto et al., 2022), and a clad girth-weld qualification will need to control these variables. Two caveats prevent direct transfer to MSR service. The first is temperature: oil-and-gas clad joints operate far below the creep regime, whereas an MSR girth weld is a high-temperature dissimilar joint, and the carbon-migration and expansion-mismatch creep failures documented for ferritic-austenitic dissimilar welds in fossil plants are an instructive, if partial, precedent for its qualification (DuPont, 2012). The second caveat is salt chemistry at the root: weld metal can be the preferential attack site in flowing chloride salt (Wang et al., 2025b), and a GH3535/316H dissimilar joint in FLiNaK exhibits closely related galvanic interfacial attack of the kind a Ni-clad-to-316H closure weld could create (Li et al., 2026), although matched-filler Ni-Mo-Cr weld metal can corrode no faster than the parent material (Danon et al., 2020). Therefore, joint design belongs in the qualification scope from the outset, and the inspectability problem discussed above applies with added weight to one-sided, unmachined field girth welds (Hwang et al., 2020; Nicolson et al., 2024).

4.4 Thermomechanical integrity

The thermal-expansion mismatch between Ni-based clads and 316H loads the interface with every thermal cycle, adding to residual deposition stresses (Young and Sham, 2018). Stress is known to accelerate intergranular attack even in Hastelloy N (Gu et al., 2022). Long-term interdiffusion and carbon migration at the interface also require data collection at component-relevant times (Young and Sham, 2018; Barua et al., 2021).

4.5 Supply chain

An INL assessment of powder-bed-fusion Ni alloys for advanced reactors favored Alloys 625 and 282 largely because commercial powder feedstock exists, while Haynes 244 (a newer Ni-Mo-Cr-W alloy attractive for salt corrosion resistance) has no commercial powder feedstock (McMurtrey et al., 2023). Qualification pathways for additively manufactured nuclear components are likewise still maturing (Hensley et al., 2021). A clad 316H architecture inverts the problem: the structural material is a commodity, and the specialty alloy is only needed as a thin surface layer, deposited by processes that pressure-vessel fabricators already operate. For countries building MSR supply chains, this is a practical industrial rationale for the bimetallic route.

4.6 Outlook

The near-term agenda follows directly: (i) salt-loop testing of deposited (not wrought) clad chemistries, matched to salt chemistry, beyond 10,000 h; (ii) irradiation and irradiation–corrosion testing of clad layers and diluted interfaces; (iii) clad-specific NDE procedures and acceptance criteria aligned with the ANL draft clad design methodology; and (iv) thermal-cycling and creep-fatigue qualification of representative clad mock-ups (e.g., nozzles, tube-to-tubesheet joints, girth welds, and valve bodies). Items (iii) and (iv) are the engineering consolidation of existing knowledge, whereas items (i) and (ii) require new data that only salt loops and irradiation facilities can provide. Based on the evidence reviewed here, the bimetallic clad component, rather than a monolithic ideal alloy, is a practical unit for the integrated qualification of corrosion performance, fabrication, inspection, and structural integrity; the literature establishes its technical plausibility, and the integrated demonstration is the step that remains.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. The manuscript was edited for English grammar with the assistance of an AI-based tool.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Antonio Viscusi, Universitas Mercatorum Dipartimento di Ingegneria e scienze, Italy

Reviewed by: Agus Widyianto, Yogyakarta State University, Indonesia