Hierarchical strengthening and toughening design of superalloy joints via a chemically complex intermetallic alloy-modified diffusion bonding strategy

Abstract

The design and fabrication of advanced interlayer materials are pivotal for the diffusion bonding (DB) of precision and slip-structure turbine components in aero-engines. Here, we developed a chemically complex intermetallic alloy (CCIMA) interlayer exhibiting exceptional mechanical properties at both room and elevated temperatures. Acting as the core of a tailored “BNi-2/CCIMA/BNi-2” sandwich interlayer, the CCIMA plays a critical role in joining powder metallurgy superalloys via the multi-interlayer composite bonding (MICB) strategy. This innovative approach leads to the formation of a robust hetero-structured joint architecture comprising alternating diffusion-affected zones (DAZs), isothermally solidified zones (ISZs), and a central CCIMA region. The CCIMA core, featuring recrystallized L1₂-ordered grains, critically promotes the extensive precipitation of cuboidal L1₂-structured nanoparticles within the ISZs. Under the optimized bonding conditions (1150 °C, 2 h), detrimental Nb–Ta-rich borides in the CCIMA region are completely suppressed, while a high-density (∼57 vol%, ∼367 nm diameter) of L1₂ precipitates forms in the ISZ. The resulting joints achieve an ultimate tensile strength of ∼1325.7 MPa (>93% of the base metal (BM) strength) and elongation of ∼27.7%, which is comparable to the BM's ductility. This outstanding performance is attributed to hierarchical strengthening and toughening mechanisms induced by the CCIMA interlayer, including macroscale strain delocalization enabled by the hetero-structured architecture, microscale dislocation pinning via grain boundaries/serrated bonded interfaces and L1₂ precipitates, and atomic-level bonding enhancement through tailored diffusion control. This study highlights the critical role of CCIMA as a functional interlayer material, establishing a new paradigm for manufacturing/repairing high-performance turbine components in next-generation aero-engines.

---

New concepts

This study introduces a conceptually novel multi-interlayer composite bonding (MICB) strategy that redefines the conventional diffusion bonding paradigm for superalloys. Rather than relying on a single interlayer, we construct a “BNi-2/CCIMA/BNi-2” sandwich architecture, where the core chemically complex intermetallic alloy (CCIMA) layer is deliberately enriched with L1₂-stabilizing elements (Al, Ti, Ta, and Nb). During bonding, this design triggers in situ precipitation of high-density, coherent L1₂ nanoparticles directly within the joint region—an unprecedented approach that transforms the joint from a simple bonding zone into a hierarchically hetero-structured material. This built-in architecture creates alternating soft/hard domains that fundamentally alter the mechanics of deformation: macroscopic strain is delocalized across the joint, while at the micro- and nano-scales, dislocation pinning and atomic-scale bonding enhancement operate synergistically. The result is a simultaneous breakthrough in strength and ductility, overcoming the long-standing strength–ductility trade-off in diffusion-bonded superalloys. This study thus establishes a new conceptual framework of “hierarchical strengthening and toughening via interlayer-intrinsic precipitation” that shifts the focus from post-bonding treatment to proactive, multi-scale architectural design of the joint itself.

1. Introduction

Manufacturing turbine components for aero-engines remains challenging as these parts must sustain extreme thermomechanical loads over long service durations.^1–3 To ensure efficient heat dissipation at high temperatures, many critical components incorporate intricate internal cooling structures, such as those in vented discs and hollow blades.^4,5 Traditional welding methods struggle to satisfy the stringent requirements of precision, reliability, and high performance in these complex geometries. Low-pressure diffusion bonding (DB), which achieves metallurgical bonding with minimal macroscopic deformation through precise temperature-pressure-time control, has therefore emerged as a key technology for manufacturing and repairing turbine disks and blades.^6–8 However, conventional direct DB faces a persistent strength–ductility trade-off. Interfacial defects such as kiss-bonds, microvoids, and segregated carbides/borides impede atomic diffusion and weaken the joint, particularly in superalloys with high deformation resistance at elevated temperatures.^9–11 Although metallic interlayers or commercial BNi-type fillers can eliminate the initial interfaces, their low yield strength (YS) and limited alloying capability often lead to insufficient joint strengthening pronounced strain incompatibility with base metals (BMs),^12–14 resulting in premature instability failure under thermomechanical loading.^15,16

Recent efforts have focused on modifying interlayer compositions and architectures to alleviate these limitations.^17–19 For example, although high-entropy alloy (HEA) interlayers can suppress brittle intermetallics, most are single-phase solid-solution alloys with insufficient key strengthening elements (e.g., Al, Ti, and Ta), limiting the L1₂ precipitation and strengthening capability for the joint. In our previous studies, we developed a novel multi-interlayer composite bonding (MICB) strategy by combining BNi fillers with L1₂-strengthened HEA foils into a sandwich-structured interlayer, achieving a significant improvement by promoting nanoprecipitation inside the joint.^18,20 Nevertheless, the precipitate density (<30 vol%) and size (<80 nm) remained far below those of BMs due to insufficient elemental supply,^21–23 leading to limited work-hardening capability and unsatisfactory ductility (typically <60% of BM). This mechanical mismatch still amplifies interfacial stress concentrations under cyclic loading, compromising components’ service life (e.g., fatigue, creep).

In this study, a strong-yet-ductile chemically complex intermetallic alloy (CCIMA) was developed and it was integrated into a “BNi-2/CCIMA/BNi-2” architecture for joining powder metallurgy (PM) superalloys via the MICB strategy. The L1₂-structured CCIMA (Co–Ni–Al–Ti–Ta–Nb–B) contains a high concentration (>23 at%) of L1₂-stabilizing elements, enabling intensive out-diffusion into adjacent BNi-2 layers and driving the formation of ultrahigh-density (>57 vol%) and large (>350 nm) cuboidal L1₂ precipitates within the joint. This markedly enhances precipitation strengthening while preserving ductility, due to the CCIMA's exceptional work-hardening capability.²⁴ The outer BNi-2 interlayers ensure defect-free bonding via transient liquid phase (TLP) mechanisms, whereas the CCIMA core provides a robust feedstock for strengthening-phase kinetics. This CCIMA-modified MICB approach simultaneously achieves strong bonding quality, superior strength–ductility synergy, and improved resistance to deformation-induced damage, offering a promising pathway for next-generation manufacturing and repair of high-value turbine components.

2. Experimental

2.1. Material preparation

Self-prepared CCIMA foils with a nominal composition of 38.5Co–38Ni–13.5Al–3Ti–3Ta–3.5Nb–0.5B (at%) were fabricated using arc-melting first. All ingots were repeatedly flipped over and remelted at least eight times to ensure chemical homogeneity, and then they were cast in a Cu mold to produce 3.5 mm-thick sheets. Subsequent cyclic rolling (∼20% thickness reduction per step with immediate annealing at 1000 °C for 15 min) and a recrystallization treatment at 1000 °C for 4 h followed by air cooling were applied until the sheet thickness reached approximately 1.3 mm (Fig. 1a). After that, CCIMA foils with thicknesses below ∼100 µm were achieved (Fig. 1b and c) by mechanical thinning treatments including wire-cutting and grinding to 2000 # grit finish. Commercial BNi-2 filler metals of approximately 50 µm in thickness (Fig. 1d and e) were combined with CCIMA foils to establish a sandwich-structured “BNi-2/CCIMA/BNi-2” interlayer for joining the BM pieces. Fig. 1f illustrates that the CCIMA exhibits fully recrystallized single-phase L1₂ grains with an average size of ∼13 µm. The BM, a third-generation PM superalloy typically used for turbine disk applications, consists of a γ matrix and three types of γ′ precipitates: micro-scale primary γ′ phases along the grain boundaries (GBs) of the γ matrix; dispersed submicron secondary γ′ phases within the γ grains; and smaller tertiary γ′ phases distributed among the phases (Fig. 1g). Prior to bonding, these as-hot isostatic pressed (as-HIPed) BM bars were heated by a sub solid-solution treatment of 1130 °C for 2 h followed by oil cooling. The chemical compositions of the BMs and BNi-2 interlayers are given in Table 1.

Fig. 1 Schematic illustrations of the CCIMA foil fabrication, MICB procedure, and microstructural characterization of the CCIMA interlayer and BMs. (a) Processing steps for CCIMA thin foil fabrication including arc-melting, casting, cyclic rolling with intermediate annealing, and final recrystallization. Surface morphology and cross-sectional thickness of (b) and (c) the developed CCIMA interlayer and (d) and (e) a commercial BNi-2 filler metal. (f) Fully recrystallized L1₂ grains in the CCIMA foils. (g) BM microstructure showing three distinct γ′ precipitate types. (h) Configuration of a custom vacuum DB furnace consisting of vacuum chamber, pressurization system, and control system. (i) Close-up view of the uniform temperature zone within the furnace chamber. (j) MICB joint configuration with the sandwich-structured “BNi-2/CCIMA/BNi-2” interlayer.

Table 1 Chemical compositions of the BM and the BNi-2 interlayer (at%)

Alloy	Ni	Co	Fe	Ti	Al	Cr	Mo	W	Ta	Si	Nb	B	C	Zr
BM	Bal.	20.1	—	4.5	7.5	13.6	2.3	0.7	0.8	—	0.6	0.1	0.24	0.03
BNi-2	Bal.	—	2.7	—	—	6.8	—	—	—	7.8	—	14.4	0.25	—

---

2.2. Multi-interlayer composite bonding

Before the MICB process, BM bars (∼14 mm diameter) were wire-cut into ∼25 mm-high cylinders and were mechanically ground and polished to achieve surface roughness (R_a) <0.2 µm and parallelism <0.1 mm. All interlayers need to be immersed in alcohol and ultrasonically cleaned for 5 min to remove surface contaminants. Sandwich-structured interlayers were assembled between BM pieces, and the assemblies were loaded into a custom vacuum DB furnace (ZTF2-150, jointly developed by the Shaanxi Zhituo Solid-state Additive Manufacturing Technology Co., Ltd and the Northwestern Polytechnical University) under a high vacuum of 3 × 10⁻³ Pa (Fig. 1h, i, and j). To ensure the melting of liquid BNi-2 filler metals, the subsequent complete isothermal solidification (eliminating residual eutectic products), and minimization of the microstructural degradation in BMs during thermal cycling, the bonding temperatures were set as 1090 °C, 1120 °C, 1150 °C, and 1180 °C with a constant holding time of 2 hours.

2.3. Microstructural characterization and tensile testing

After bonding, the bonded joints were sectioned into metallographic specimens for microstructural characterization. Samples were ground to 2000# grit and subsequently electro-polished in a mixed solution of C₂H₅OH : HNO₃ = 4 : 1 under a direct voltage of 20 V at −30 °C for 20 s. Joint microstructure and crystallographic characteristics were analyzed using a field emission scanning electron microscope (FEI-SEM, Helios G4 CX, 10 kV) equipped with electron backscattered diffraction (EBSD, EDAX). Elemental distributions and compositions in the joints, particularly for light elements (e.g., C, B), were quantified via an electron probe microanalyzer (EPMA, JXA-8230). A scanning transmission electron microscope (STEM, Talos F200X, 200 kV) with energy-dispersive spectroscopy (EDS) was employed to identify additional phases formed by interdiffusion between interlayers and BMs. For nanoprecipitate analysis, chemical compositions were characterized via three-dimensional atom probe tomography (3D-APT, CAMECA LEAP 5000 XR) in a voltage mode at 70 K, with a pulse repetition rate of 200 kHz, a pulse fraction of 20%, and an evaporation detection rate of 0.2% atom per pulse. Then the 3D distributions of distinct atoms were reconstructed using the AP Suite 6.1 workstation. Needle-shaped specimens for 3D-APT analysis were prepared using a focused ion beam/scanning electron microscope (FIB/SEM, FEI Scios).

Dog-bone-shaped tensile specimens with a gauge section of 12.5 × 3.2 × 1.5 mm³ were wire-cut from bonded joints. All surfaces, including fillet radii, were ground to 2000# grit. Room temperature (RT) tensile testing was performed using a universal testing machine (Instron 3382) at a strain rate of 1 × 10⁻³ s⁻¹. Three specimens per bonding condition were prepared and tested to ensure data repeatability. Hardness distributions across the joints were conducted by a Vickers micro-hardness tester (Wilson, VH1102) under a load of 4.9 N for 15 s. In situ tensile testing was conducted to analyze the macro-deformation mechanisms within interlayer regions and BMs during stretching. After tensile testing, joint fracture morphologies and deformation substructures were characterized by SEM and TEM to elucidate the underlying micro-deformation mechanisms of joints.

3. Results

3.1. Joint microstructure

Fig. 2 shows the microstructures and EPMA elemental maps of the MICB joint bonded at 1150 °C for 2 h. The defect-free joint exhibits a hetero-structured architecture comprising five distinct zones (Fig. 2a): two isothermally solidified zones (ISZs) originating from the outer BNi-2 interlayers (Fig. 2b); a central CCIMA region (Fig. 2c), and two diffusion-affected zones (DAZs) within the BMs, influenced by boron diffusion from BNi-2 (Fig. 2d). The corresponding chemical compositions of these zones are summarized in Table 2. In the DAZs, bright Cr–Mo–W-rich M₃B₂-type borides precipitate predominantly along GBs (Fig. 2d and EPMA maps), consistent with observations from our previous study.²⁰ The outer ISZs were dominated by coarsened γ-phase grains formed during isothermal solidification. In the core CCIMA region, the microstructure retained its initial recrystallized L1₂ grain structure (Fig. 2b). High-magnification SEM image and EPMA maps (Fig. 2c and e) reveal a limited number of secondary phases, including dark grey Co–Cr-rich phases and bright Nb–Ta-rich borides, which likely formed due to interdiffusion between the CCIMA and the BNi-2 interlayers/BMs.

Fig. 2 Microstructural and elemental analysis of the MICB joint bonded at 1150 °C for 2 h. (a) Hetero-structured joint comprising two DAZs, two ISZs, and a central CCIMA region. (b) High-magnification SEM image detailing joint morphology. (c) CCIMA region microstructure showing L1₂ grains with intergranular Co–Cr-rich phases and Nb–Ta-rich borides. (d) DAZ morphology exhibiting the BM's γ/γ′ microstructure and intergranular Cr–Mo–W-rich M₃B₂-type borides. (e) Joint morphology with the corresponding EPMA elemental maps.

Table 2 Chemical compositions (at%) of different zones measured by EPMA in the 1150 °C joint

Location	B	Al	Si	Ti	Cr	Fe	Co	Ni	Nb	Mo	Ta	W
CCIMA	0.99	5.59	0.74	3.89	3.87	0.88	25.36	54.97	1.22	0.27	2.05	0.17
ISZ	1.06	4.08	0.78	2.84	8.90	1.18	29.95	48.28	0.88	0.56	1.28	0.22
DAZ	1.83	3.52	0.23	4.61	15.89	0.49	21.48	48.95	0.39	1.28	0.92	0.40

---

To identify the two distinct phases formed in the CCIMA region, STEM analyses were performed (Fig. 3 and 4). STEM-bright field (STEM-BF) imaging reveals the original L1₂ grains alongside newly formed slatted/blocky borides (corresponds to the Nb–Ta-rich borides in Fig. 2c), with a small number of dislocation cell structures also present (Fig. 3a). Selected area electron diffraction (SAED) analysis (Fig. 3c) confirms that the matrix retained the L1₂ structure. The Nb–Ta-rich borides were found to share the same M₃B₂ crystal structure with the Cr–Mo–W-rich borides in the DAZs (Fig. 2d), as evidenced by the corresponding SAED pattern (Fig. 3d), the STEM-high angle annular dark-field (STEM-HAADF) image and elemental mapping (Fig. 3e). Additionally, dislocation cell structures associated with the Co–Cr-rich phases (Fig. 2c) were also observed (Fig. 3a). STEM-BF and dark-field (DF) images (Fig. 4a and b) demonstrate that these regions are surrounded by dense dislocation networks, which can be attributed to the mismatch stresses resulting from the new phase formation. Corresponding SAED patterns indicate that these areas exhibit a face-centered cubic (FCC) matrix, distinct from the surrounding L1₂ matrix (Fig. 4c and d). Notably, within them, cuboidal L1₂ particles enriched in Ni, Al, Ti, Ta, and Nb were precipitated, as depicted in the STEM-HAADF analysis (Fig. 4e). These findings indicate that during the bonding thermal cycle, atomic interdiffusion triggered two distinct phase transformation pathways within the CCIMA region: (1) formation of Nb–Ta-rich M₃B₂-type borides and (2) precipitation of dual-phase regions comprising an FCC matrix and L1₂ nanoparticles. Their resultant volume fractions were strongly influenced by the bonding parameters, which govern the elemental diffusion kinetics and, in turn, dictate the phase evolution pathways.

Fig. 3 STEM characterization of the CCIMA region microstructure. (a) STEM-BF image showing the L1₂ matrix, borides and dislocation cell structures. (b) High-magnification STEM-BF image and corresponding SAED patterns from (c) L1₂ matrix and (d) M₃B₂ borides. (e) STEM-HAADF elemental maps exhibiting M₃B₂ borides enriched in Nb and Ta.

Fig. 4 STEM characterization of dislocation cell structures in the CCIMA region. (a) STEM-BF image showing the FCC matrix region surrounded by dislocation cells. (b) Corresponding STEM- HAADF pattern of the FCC matrix region containing an FCC matrix and L1₂ nanoparticles. (c) SAED pattern from the L1₂-structured CCIMA region, and (d) the FCC matrix region. (e) Elemental maps revealing the L1₂ matrix and the dual-phase (FCC + L1₂) region.

3.2. Microstructural evolution

Fig. 5a–l illustrate the microstructural evolution of joints bonded at temperatures ranging from 1090 °C to 1180 °C for 2 hours, while Fig. 5m–o present the quantified volume fractions of secondary phases as a function of bonding temperature. Across all bonding conditions, the joints consistently exhibit five characteristic zones identified earlier: two DAZs, two ISZs, and one central CCIMA region (Fig. 5a–d). As detailed in previous analyses, the DAZs within the BMs retained the original γ/γ′ dual-phase microstructure and intergranular Cr–Mo–W-rich borides, which formed due to boron diffusion from the BNi-2 interlayers (Fig. 2d).²⁰ Within the CCIMA region, as illustrated in Fig. 5e–h, the volume fraction of the Nb–Ta-rich borides progressively decreases as the bonding temperature increases, with 14.7 vol% at 1090 °C to nearly 0 vol% at 1180 °C (Fig. 5m). When bonding at 1150 °C, no obvious bright borides were detected (Fig. 5g). Instead, a limited number of FCC + L1₂ dual-phase structures (shown in Fig. 4) appeared in the CCIMA region.

Fig. 5 Microstructural evolution and quantitative analysis of MICB joints bonded for 2 h at varying bonding temperatures. Joint morphology of the CCIMA region consisting of large Nb–Ta-rich borides, and the ISZ with numerous L1₂ nanoparticles in the joints bonded at (a), (e) and (i) 1090 °C, (b), (f) and (j) 1120 °C, (c), (g) and (k) 1150 °C, and (d), (h) and (l) 1180 °C. Quantitative dependencies on bonding temperature: (m) boride volume fraction in the CCIMA region, (n) average L1₂ nanoparticle size and (o) L1₂ nanoparticle volume fraction in ISZ.

In the ISZs, coarsened γ-phase grains predominate, and within them, numerous dispersed L1₂ nanoparticles precipitate out (Fig. 5i–l). Quantitative analysis using Image-Pro Plus software shows that at lower temperatures (e.g., 1090 °C), these newly formed nanoparticles exhibited relatively lower volume fractions (∼33.2 vol%) and small average sizes (∼104 nm diameter), as demonstrated in Fig. 5n and o. As the temperature increases, both the volume fraction and the particle size increase accordingly, peaking at 1150 °C with values of approximately 57 vol% and 367 nm, respectively. However, when the temperature reached 1180 °C, slight reductions in both the volume fraction (∼55 vol%) and the size (∼346 nm) were detected (Fig. 5l, n, and o). Notably, as the temperature increased from 1090 °C to 1150 °C, the nanoparticle morphology evolved from near-spherical to more cuboidal forms. At 1150 °C, high-density cuboidal L1₂ nanoparticles were widely distributed throughout the ISZ (Fig. 5k). However, after bonding at 1180 °C, the nanoparticles regained a more rounded morphology (Fig. 5l), primarily due to the lattice misfit with the FCC matrix resulting from compositional tuning, a mechanism that will be discussed in detail later.

3.3. Mechanical properties and fracture behavior

Fig. 6a presents the engineering stress–strain curves of the joints tested at RT, and the corresponding mechanical properties as a function of bonding temperature, including YS, ultimate tensile strength (UTS), and total elongation (TE), are shown in Fig. 6b and c. As the temperature increases from 1090 °C to 1120 °C, the properties of the joints exhibit minimal enhancement: YS from 985.7 ± 15 MPa to 995.2 ± 20 MPa, UTS from 1155.1 ± 20 MPa to 1179 ± 18 MPa (Fig. 6b), and TE remained low, from 5.6 ± 1% to 6.5 ± 0.5% (Fig. 6c). Fracture profile of the 1090 °C-bonded joint (the inset of Fig. 6a) indicates the limited ductility is attributed to the presence of aggregated Nb–Ta-rich borides in the CCIMA region, which initiated brittle fracture early during plastic deformation. Corresponding fractography reveals a predominantly quasi-cleavage morphology, characterized by a small number of dimples and extensive cleavage planes (Fig. 6e and f).

Fig. 6 Mechanical properties and fracture behavior of MICB joints under varied bonding conditions. (a) Engineering stress–strain curves for joints bonded at different temperatures. (b) YS/UTS and (c) TE as a function of bonding temperature. (d) Ashby plot comparing UTS-TE combinations (red stars: this work) with literature data for superalloy joints.^{13,16,25–39} (e) Cleavage fracture surface of the 1090 °C-bonded joint with the (f) high-magnification SEM image showing cleavage facets. (g) Ductile fracture morphology in the 1150 °C-bonded joint revealing (h) fine dimples at high magnification.

At a bonding temperature of 1150 °C, the brittle borides in the CCIMA region are significantly reduced, resulting in a significant improvement in mechanical performance. The maximum UTS reached 1325.7 ± 15 MPa, exceeding 93% of the BM strength after the same thermal cycle, and the TE of 27.7 ± 3% was comparable to the ductility of the BM (Fig. 6b and c). Regarding the fracture surfaces, a typical ductile feature with uniformly distributed fine dimples was observed, indicating a substantial enhancement in ductility (Fig. 6g and h). However, when the bonding temperature was further increased to 1180 °C, mechanical performance declined: YS, UTS, and TE dropped to 877.8 ± 10 MPa, 1219.6 ± 17 MPa, and 15.7 ± 2%, respectively. This degradation is likely due to microstructural deterioration in the BMs under such high temperatures, such as grain coarsening or precipitate dissolution, caused by excessive thermal exposure during bonding. Statistical analysis via Ashby plots^{13,16,25–39} reveals that the MICB-fabricated joints in this study achieve a remarkable combination of high tensile strength and significantly improved elongation (Fig. 6d). Compared to TLP joints bonded with a BNi-2 interlayer of identical thickness under the same bonding parameters (1150 °C for 2 h), the MICB joints exhibit a moderate increase in UTS (from ∼1292.5 MPa to ∼1325.7 MPa) and, more notably, a substantial improvement in ductility, enhancing from ∼15.4% to ∼27.7%. This represents a substantial advancement over conventional DB techniques, which have long struggled to simultaneously achieve high strength and ductility. This kind of balance in mechanical performance is critical for superalloy joints operating under complex loading conditions for extended durations.

4. Discussion

Building on the established microstructural evolution and mechanical performance of the MICB joints, this section explores the underlying deformation and strengthening mechanisms through multiscale characterization. By integrating crystallographic data, elemental diffusion kinetics, in situ tensile observations, and TEM analyses of deformation substructures, this study systematically elucidates how the CCIMA-modified multi-interlayers influence the interfacial bonding, as well as the macro- and microscale deformation mechanisms in the hetero-structured joints.

4.1. Bonding mechanisms and atomic diffusion behaviors

Fig. 7a–d show the crystallographic characteristics of the joint bonded at 1150 °C, and the corresponding chemical profiles across joints bonded at different temperatures are demonstrated in Fig. 7e–g. Based on the grain size differences shown in the band contrast map and inverse pole figure (IPF) (Fig. 7a and b), five distinct regions are identified: a central CCIMA region with fine grains, flanked by two ISZs composed of coarsened γ-phase grains, and two DAZs within the BMs containing relatively finer grains. This results in a hetero-structured grain size distribution, alternating between fine-grained regions (DAZ/CCIMA, ∼63.4 µm) and coarser regions (ISZs, ∼115.4 µm), as illustrated in Fig. 7d. The recrystallization map (Fig. 7c) reveals that twin boundaries (light green lines) are exclusively found in the DAZs and the CCIMA region, suggesting that melting during bonding occurred only within the ISZs. Fig. 7h–k schematically illustrate the bonding behaviors governed by multi-interlayer design. During the MICB process employing a “BNi-2/CCIMA/BNi-2” sandwich interlayer (Fig. 7h), the outer BNi-2 melts upon reaching its melting point (T_m), forming a transient liquid film. This liquid promotes rapid atomic interdiffusion, particularly of melting point depressants (MPDs) such as Si and B (Fig. 7i). As these MPDs diffuse toward the BMs or CCIMA region, the melting temperature of the remaining liquid increases. Once the bonding temperature (T_b) equals this new local melting point, isothermal solidification initiates and progresses during the holding stage (Fig. 7j), giving rise to the ISZs.^40–42 Due to the absence of supercooling at a constant bonding temperature, nucleation is suppressed during the isothermal solidification, while the grain growth is promoted. As a result, coarse γ-phase grains, exceeding 110 µm, are formed within the ISZs (Fig. 7b).

Fig. 7 Crystallographic, compositional, and bonding behavior analysis of the MICB joints. (a) Band contrast map, (b) corresponding IPF, (c) recrystallization distribution map, and (d) grain size distribution with fine-grained DAZ, CCIMA region, and coarsened-grained ISZ in the joint bonded at 1150 °C for 2 h. Elemental profiles across the joints bonded at (e) 1090 °C, (f) 1150 °C and (g) 1180 °C showing atomic fractions of Ni, Co, Al, Ta, Nb, Ti, and Cr. (h) Schematic MICB bonding mechanisms including the (h) multi-interlayer configuration, (i) atomic diffusion, (j) isothermal solidification, and (k) precipitation process.

During the subsequent homogenization, concentration gradients established between the BNi-2 interlayers, CCIMA region, and BMs drive elemental interdiffusion, leading to a more uniform chemical distribution across the joint. Fig. 7e–g show the chemical profiles of joints bonded at 1090 °C, 1150 °C, and 1180 °C, indicating the temperature-dependent diffusion kinetics. At 1090 °C, atomic diffusion is limited, leading to the pronounced elemental partitioning. Fig. 7e illustrates that the CCIMA region maintains a CoNi-based composition (38.5Co–38Ni at%), while the ISZs and BMs remain Ni-rich systems. Moreover, it is notable that the presence of Nb–Ta-rich borides in the CCIMA region causes significant local compositional fluctuations. As the bonding temperature increases, accelerated diffusion reduces the boron accumulation (from the BNi-2) in the CCIMA region, facilitating the elimination of brittle borides and thereby promoting the compositional homogeneity.²⁰ Concurrently, Ni/Cr diffuses from the ISZs into the CCIMA region, while Co migrates reversely, promoting more balanced elemental distributions (Fig. 7f). At 1180 °C, Fig. 7g reveals a nearly homogeneous chemical profile throughout the hetero-structured joint.

Beyond the diffusion of major FCC-forming elements, the diffusion of L1₂-stabilizing elements (e.g., Al, Ti, Ta, and Nb) critically governed the precipitation of L1₂ particles and the resultant strengthening mechanisms. As described in Section 3.2, a significant number of L1₂-γ′ nanoparticles precipitated within the γ-phase grains of the ISZs. To quantitatively assess their composition, 3D-APT was employed. The elemental distribution maps and corresponding compositional profiles across the FCC matrix and L1₂ nanoparticles are depicted in Fig. 8. The results reveal pronounced partitioning behavior among the alloying elements: Co, Cr, Fe, Mo, and Si preferentially partition to the FCC matrix, while Ni, Al, Ti, Ta, Nb, Ta, and B are enriched in the L1₂ nanoparticles (Fig. 8a and b). It is notable that the original BNi-2 interlayer does not contain any L1₂-forming elements (see Table 1). Therefore, the formation of L1₂ precipitates in the ISZs is primarily attributed to the diffusion of Al, Ti, Ta, and Nb from the CCIMA region (Fig. 8c), which initially contains approximately 13.5% Al, 3.0% Ti, 3.0% Ta, and 3.5% Nb (at%). As the bonding temperature increases (under a fixed holding time), enhanced diffusion coefficients promote the migration and accumulation of these elements into the ISZs.⁶ During the subsequent cooling stage, their supersaturation in the FCC matrix drives extensive precipitation of L1₂ nanoparticles, increasing their volume fraction accordingly (Fig. 5n).^1,22,43,44 However, when the bonding temperature reaches 1180 °C, the concentration of these L1₂-forming solutes begins to decline slightly due to the continued outward diffusion beyond the CCIMA region, leading to a minor reduction in the L1₂ volume fraction, from ∼57 vol% at 1150 °C to ∼55 vol% at 1180 °C.

Fig. 8 3D-APT analysis of L1₂ precipitation in the ISZ. (a) Atom maps showing elemental distributions within L1₂ precipitates and the FCC matrix. (b) Compositional profile across the FCC/L1₂ interface. (c) Schematic illustrating L1₂ precipitation in the ISZ induced by diffusion of Al, Ti, Ta, and Nb from the CCIMA region.

Concurrently, Fig. 5i–l exhibit that the morphology of L1₂ nanoparticles evolves from spherical to cuboidal with increasing bonding temperature. This transition is attributed to variations in the lattice misfit between the FCC matrix and L1₂ precipitates driven by compositional fluctuations, particularly involving Ta. Zhang et al.⁴⁵ found that Ta additions in Ni-based superalloys significantly affect L1₂ precipitate morphology by altering the lattice misfit from positive to negative values and increasing its absolute magnitude, thereby promoting the transition from spherical to cuboidal precipitates. In the present study, elevated bonding temperatures enhance Ta diffusion from the CCIMA region into the ISZs, raising local concentrations of it and facilitating the observed cuboidal transformation of the L1₂ nanoparticles.

4.2. Macro-deformation mechanism in the hetero-structured joint

Through partial melting, isothermal solidification, atomic interdiffusion, and subsequent phase precipitation, a hetero-structured joint consisting of alternating DAZs, ISZs, and a central CCIMA region formed in the MICB process. To evaluate the mechanical implications of this structural heterogeneity, microhardness profiling was conducted across the joint. The microhardness distributions for joints bonded at various temperatures exhibit distinct zone-dependent deformation characteristics (Fig. 9a). At 1090 °C, the presence of brittle Nb–Ta-rich borides results in a pronounced hardness peak (∼530 HV) within the CCIMA region (Fig. 9b). In contrast, adjacent ISZs, where L1₂ nanoparticle precipitation is insufficient, exhibit hardness valleys (∼384 HV and ∼395 HV). These sharp peak-valley transitions in hardness can induce local stress concentrations and strain mismatches between adjacent zones. This mismatch triggers plastic instability and causes the premature failure, particularly within the CCIMA region during the early stages of deformation,^14,15,46,47 as illustrated in the engineering stress–strain response shown in Fig. 6a. As the bonding temperature increases, brittle borides significantly reduced while abundant L1₂ nanoparticles precipitate out (Fig. 9c), thereby a significantly more uniform hardness profile was achieved in the hetero-structured joint. The improved mechanical homogeneity across the interface effectively enhances the deformation compatibility of the joint, enabling a synergistic balance between strength and ductility.

Fig. 9 Microhardness distributions across MICB joints bonded at different temperatures. (a) Microhardness profiles for joints bonded at 1090 °C and 1150 °C. Joint microstructure with microhardness indentations: (b) peak hardness region in the 1090 °C-bonded joint associated with brittle Nb–Ta-rich borides in the CCIMA region and (c) homogeneous hardness distribution in the 1150 °C-bonded joint.

However, despite macroscale hardness uniformity promoting global deformation compatibility, localized microscale heterogeneity persists. Specifically, the ISZs still display slightly lower hardness values (∼379 HV and ∼385 HV) compared to the CCIMA region (∼418 HV) and the DAZs (∼392 HV), indicating minor variations in local mechanical properties. To further investigate the deformation behavior of this hetero-structured joint, in situ tensile testing combined with digital image correlation (DIC) analysis was employed. This method enables direct quantification of strain gradients across the joint, revealing how zone-specific mechanical responses contribute to the overall deformation and govern the failure mechanisms. By mapping the spatial evolution of strain during tensile loading, this approach provides critical insights into the strain partitioning and plastic instability initiation across the hetero-structured joint.

In situ tensile testing was conducted using a dedicated tensile stage (PicoFemto-SEP, SEM5000N, China) integrated within a SEM (SEM5000, CIQTEK, China) (Fig. 10a). Prior to testing, dog-bone-shaped tensile specimens were carefully polished to eliminate surface defects and subsequently sprayed with metallic particles (∼250 nm in diameter) to create random speckle patterns for DIC analysis (Fig. 10b and c). The tensile test was conducted under displacement control at a loading rate of 2 µm s⁻¹, with periodic pauses at predefined strain intervals to acquire high-resolution SEM images from a fixed region of interest (ROI) (Fig. 10d). Based on these speckled SEM images (Fig. 10e), full-field strain maps were computed via DIC using the VIC-2D software, enabling spatially resolved qualification of strain evolution across the hetero-structured joint (Fig. 10f–i). In addition, geometrically necessary dislocation (GND) density maps were extracted from strain gradient tensors through a custom MATLAB algorithm. Calculation methodological details are provided in our previous studies.⁴⁸ GND density evolution was statistically analyzed for each zone as a function of strain, correlating the local dislocation activity with key features such as GBs, bonded interfaces, and second-phase particles (Fig. 11).^49,50

Fig. 10 In situ tensile analysis of the MICB joint bonded at 1150 °C for 2 h. (a) Tensile testing stage configuration. (b) Dimensions of the in situ tensile specimen. (c) Specimen with fiducial marks for DIC. (d) The in situ tensile curve exhibiting paused strain levels. (e) SEM image of a pristine specimen with DIC markers. Strain distribution maps at strains of approximately (f) 2%, (g) 10%, (h) 22%, and (i) fracture. (j) Initial joint microstructure (unmarked region). (k) Scattered fracture profile showing the failure zone in the DAZ and (l) high-magnification SEM image exhibiting the intergranular crack behavior. (m) Initial DAZ morphology with numerous intergranular Cr–Mo–W-rich borides.

Fig. 11 Evolution of GND density during in situ tensile testing. GND density maps at strains of (a) yield point, (b) 2.0%, (c) 6.8%, (d) 8.7%, (e) 10.0%, (f) 12.4%, (g) 14.7%, (h) 17.0%, (i) 22.0%, and (j) fracture. (k) Average GND density evolution in the DAZ, ISZ, and CCIMA region as a function of strain. (l) Post-fracture GND density profile across the hetero-structured joint exhibiting peak concentration at the bonded interface. (m) Local strain histogram in different zones at fracture.

At the onset of plastic deformation (∼2% global strain), strain localization was observed primarily within the ISZs (Fig. 10f). These zones, characterized by coarsened γ-phase grains and relatively lower YS, initiated plastic strain earlier than the CCIMA region and DAZs. As the applied strain increased to ∼10%, sustained strain concentration in the ISZs led to significant dislocation accumulation and strain hardening (Fig. 10g). Consequently, higher applied stress was required to continue deformation, consistent with the stress–strain behavior shown in Fig. 10d. Through load transfer across interfaces, the rising stress in the ISZs progressively activated plastic deformation in adjacent DAZs and the CCIMA region (∼22% global strain, Fig. 10h). This interfacial stress mediation facilitated strain redistribution, enabling the coordinated macro-deformation across the joint's hetero-structured architecture. Upon fracture (∼28% global strain, Fig. 10i), severe strain concentration, exceeding 35%, was concentrated in the ISZs, whereas the average strain in the CCIMA region and DAZs remained below 20%. This strain gradient highlights the ISZ's role as “soft” buffer zones, effectively delaying catastrophic failure by accommodating localized plasticity and dissipating strain energy.^51,52 Post-fracture analysis comparing the initial joint microstructure (Fig. 10j) and the speckled fracture profile (Fig. 10k) indicated that fracture initiation occurred approximately 200 µm away from the ISZs, within the DAZ, specifically in the BM region. This fracture location suggests that the MICB joint performance is nearly comparable to that of the BM, confirming the effectiveness of the multi-interlayer “sandwich” architecture in enabling uniform deformation while simultaneously strengthening the joint. High-magnification fracture profile in the DAZ revealed intergranular cracking (Fig. 10l), which is attributed to local stress concentrations induced by residual Cr–Mo–W-rich borides along the GBs, as observed in the initial DAZ microstructure (Fig. 10m).

Fig. 11a–j systematically track the GND density evolution across the hetero-structured joint under incremental strains, with Fig. 11k quantitatively comparing the average GND density evolution in each zone. During the initial plastic deformation (<9% strain), GNDs predominantly accumulated at interfacial regions, including GBs within all three zones and the bonded interfaces. Intragranular GND accumulation was also observed along two primary {111} slip planes, manifested as parallel and intersecting bright blue bands (Fig. 11a–d). Due to the earlier yielding of the ISZs, this region consistently exhibited higher GND densities than the DAZs and the CCIMA region at equivalent strain levels (Fig. 11k). As strain increased to 17%, interfacial GND accumulation intensified significantly, particularly at the bonded interfaces (Fig. 11e–h). Simultaneously, slip bands within the ISZ grains impinged upon GBs, amplifying the localized GND concentrations (Fig. 11i). Throughout the 0–22% strain range, all zones displayed parallel, gradual increases in dislocation density without abrupt transitions (Fig. 11k), indicating stable plastic flow and global deformation compatibility.

The bonded interfaces with irregularly serrated morphology, originating from partial melting and isothermal solidification of the BNi-2 interlayer, play a key role in accommodating deformation. These geometrically irregular interfaces effectively dispersed strain gradients and suppressed crack nucleation, directly contributing to the joint's exceptional ductility.^43,53 Furthermore, post-fracture analysis reveals a peak GND density at the bonded interface (∼5 × 10¹⁴ m⁻², Fig. 11l), approximately twice that of adjacent DAZ and CCIMA regions, arising from mismatches in mechanical response and strain partitioning. Average strain in each zone upon failure confirms the significant interzonal disparities: the ISZ experienced ∼23.1% strain versus ∼15.5% in the DAZ and ∼14.6% in the CCIMA region (Fig. 11m). Crucially, despite these strain gradients, fracture initiated within the DAZ (specifically in the BM region, Fig. 10k), not at the bonded interfaces, indicating the superior interfacial cohesion and deformation compatibility achieved via the MICB process. The BNi-2-derived serrated interfaces effectively alleviate stress concentrations, retard crack propagation, and reduce brittle failure susceptibility. This mechanism parallels toughening strategies employing serrated GBs in high-strength steels,^54,55 Ni-based superalloys,^53,56 and HEAs.^43,57

4.3. Hierarchical strengthening and toughening mechanisms

The exceptional strength–ductility synergy achieved in the MICB joint stems from a hierarchical design strategy integrating joint structural heterogeneity, GND/precipitation strengthening, and atomic-scale diffusion control. At the macroscale, the hetero-structured joint, comprising alternating DAZ, ISZ, CCIMA, ISZ, and DAZ, facilitates coordinated deformation through effective load transfer and strain partitioning. Acting as a “soft buffer zone”, the ISZ accommodates early-stage plastic strain to delay strain localization. As deformation progresses, the CCIMA region and DAZs sequentially engage via load transfer mechanisms, resulting in a more uniform strain distribution across the joint. Crucially, the serrated bonded interfaces, formed during isothermal solidification of the BNi-2 interlayers, exhibit outstanding cohesion and deformation compatibility. Their irregular geometry effectively dissipates stress concentrations, slows crack propagation, and mitigates interfacial fracture, contributing significantly to the joint toughness. At the microscale, the mechanical response of the joints is enhanced through the synergistic effects of GNDs and precipitation hardening. During deformation, GNDs preferentially accumulate along GBs and bonded interfaces, inducing significant back stress that elevates the joint's global work-hardening capability.^50,51 In the ISZ, the coarse-grained microstructure provides an extended mean free path for dislocation glide, thereby promoting superior strain accommodation and facilitating the onset of plastic deformation. GNDs nucleate and propagate within the coarse grains of the ISZ. With increasing strain, dislocations within the adjacent fine-grained CCIMA/DAZ regions are activated and progressively hindered at the GBs and the bonded interfaces. The mutual interaction of the associated stress fields suppresses further dislocation multiplication, thereby strengthening these regions while concurrently enhancing the local strain-hardening capacity in the interfacial regions. Moreover, microstructural features intrinsic to each zone, particularly the precipitates, act as potent obstacles to dislocation nucleation and glide, further strengthening the joint. Fig. 12 illustrates dislocation structures and their interactions with precipitates in both the CCIMA region and the ISZ of the joint bonded at 1150 °C under ∼5% applied strain.

Fig. 12 Deformation substructures in the CCIMA region and the ISZ of the 1150 °C bonded joint at ∼5% strain. The CCIMA region: (a) BF-TEM image showing dense dislocations within L1₂ grains. (b) Corresponding STEM-BF image indicating APB-coupled superdislocations. (c) Their high-magnification morphology featuring dislocation-pairs. The ISZ: (d) STEM-BF image with FCC/L1₂ dual phase showing massive dislocation within the FCC channels and a limited number of APB superdislocations within the L1₂ particles. (e) High-magnification BF image exhibiting dislocation pile-ups within the FCC channel. (f) HRTEM image of an FCC/L1₂ interface, with FFT patterns of the L1₂ particle and SFs in the FCC matrix. (g) STEM-HAADF image and corresponding elemental maps.

Notably, in the CCIMA region, bright-field TEM (BF-TEM) imaging reveals high-density dislocations distributed within L1₂-ordered grains (Fig. 12a). The STEM-BF image (Fig. 12b) confirms these dislocations as anti-phase boundary (APB)-coupled superdislocations, exhibiting the characteristic dislocation-pair morphology illustrated in the high-magnification DF image (Fig. 12c). These superdislocations, resulting from the ordered nature of the L1₂ phase, require elevated shear stress for propagation due to the substantial energy penalty associated with disrupting atomic ordering. This constraint on dislocation mobility significantly enhances work-hardening capability by promoting dislocation entanglement and impeding slip transmission.^24,58,59 In the ISZs, STEM-BF and STEM-HAADF imaging coupled with elemental mapping confirm the presence of a high density of cuboidal L1₂ nanoparticles embedded within FCC matrix channels (Fig. 12d and g). These precipitates, enriched in Al, Ti, Ta, and Nb through diffusion from the CCIMA region, serve as potent barriers to dislocation motion via the Orowan mechanism (Fig. 12d).^21,23,43,60 Detailed TEM analysis reveals dislocations accumulating predominantly within the FCC matrix channels, while only limited observation of APB-coupled superdislocations inside the L1₂ precipitates themselves (Fig. 12e). During plastic deformation, dislocations progressively bow around these precipitates and form pile-ups at interfaces, substantially increasing the critical resolved shear stress and contributing to precipitation strengthening.^60–62 Furthermore, high-resolution TEM (HRTEM) imaging (Fig. 12f) demonstrates coherent L1₂/FCC interfaces and stacking faults (SFs) within the FCC channels. These features not only inhibit the interface decohesion but also help accommodate plastic strain, further enhancing the work-hardening response of the ISZs and, by extension, the entire joints.⁶³

At the atomic scale, multi-element interdiffusion in the MICB joint promotes both the coherency strengthening and order strengthening, which are critical to its superior mechanical performance. In the ISZs, nanoscale L1₂ precipitates maintain coherent interfaces with the FCC matrix. Solute partitioning of elements like Ta, Al, Ti, and Nb between these dual phases modulates interfacial lattice misfit (δ), generating long-range elastic strain fields around the precipitates, which imposes an additional atomic-scale barrier to dislocation motion and elevates the strength.^1,64 Crucially, the strong atomic bonding and lattice registry at these coherent interfaces facilitate dislocation transmission and local strain accommodation, significantly enhancing fracture toughness. Additionally, order strengthening, driven by high APB energy intrinsic to the L1₂-ordered structures, is a dominant mechanism in the CCIMA region and becomes increasingly significant in the ISZs. During interdiffusion between the CCIMA and the ISZs, partitioning of Ta/Nb/Ti into L1₂ nanoparticles substantially raises their APB energy,^24,59 thereby increasing the critical stress for dislocation shearing and improving joint strength and work-hardening capability.⁶⁵

Consequently, this hierarchical strengthening and toughening strategy, combining macroscale strain delocalization through a hetero-structured architecture, microscale dislocation pinning by GBs/bonded interfaces and L1₂ nanoprecipitates, and atomic-scale strengthening via compositional tuning, ensures sustained work-hardening throughout deformation. This approach effectively overcomes the traditional strength–ductility trade-off, establishing the MICB joint as a promising prototype for next-generation high-temperature structural applications that demand both mechanical robustness and damage tolerance.

5. Conclusions

A multi-interlayer composite bonding (MICB) strategy employing a “BNi-2/CCIMA/BNi-2” sandwich structure was developed to fabricate high-performance superalloy joints. The CCIMA interlayer, consisting of fully recrystallized L1₂ grains, synergizes with BNi-2 filler metals to suppress interfacial defects and promote atomic interdiffusion, enabling the robust metallurgical bonding of superalloys. Key conclusions can be drawn as follows:

(1) The incorporation of a “BNi-2/CCIMA/BNi-2” sandwich interlayer during the MICB process produced a hetero-structured joint comprising alternating DAZ/ISZ/CCIMA/ISZ/DAZ zones. Driven by the atomic interdiffusion between interlayers and BMs, each zone developed distinct microstructures: the DAZs within the BMs exhibited γ/γ′ dual-phase structures along with Cr–Mo–W-rich borides; the central CCIMA region, initially composed of L1₂ grains, incorporated limited FCC/L1₂ dual-phase regions and Nb–Ta-rich borides; the ISZs consisted of coarsened γ-phase grains containing uniformly distributed cuboidal L1₂ nanoparticles precipitated during cooling.

(2) As the bonding temperature increased, the volume fraction of Nb–Ta-rich borides in the CCIMA region progressively decreased, while both the size and the volume fraction of L1₂ precipitates in the ISZ increased. Optimal mechanical properties, UTS of ∼1325.7 MPa (>93% of the BM's strength) and elongation of ∼27.7% (comparable to the BM's ductility) were achieved in the joint bonded at 1150 °C for 2 hours, which is attributed to the elimination of brittle borides and the formation of ∼57 vol% cuboidal L1₂ nanoparticles with an average diameter of ∼367 nm.

(3) The joint's exceptional strength–ductility synergy stems from a hierarchical strengthening and toughening design consisting of macroscale hetero-structured architecture, microscale GND and precipitate hardening, and atomic-scale strengthening via compositional tuning. Specifically, strain redistribution enabled by the alternating soft/hard zone configuration, dislocation obstruction by GBs/serrated bonded interfaces and L1₂ nanoparticles, and enhanced atomic bonding energy and tailored coherent phase interfaces, collectively enhance the work-hardening capacity and suppress strain localization.

Author contributions

L. Yuan: conceptualization, methodology, visualization, software, resources, data curation, writing – review & editing. F. Y. Jiang: data curation, writing – review & editing. Y. J. Du: software, writing – review & editing. Y. Z. Yang: visualization, writing – review & editing. J. Gan: software, writing – review & editing. P. K. Liu: methodology, writing – review & editing. D. Hao: methodology, writing – review & editing. J. Y. Zhang: conceptualization, methodology. J. H. Luan: software, data curation. H. Zhang: methodology, writing – review & editing. J. L. Li: methodology, writing – review & editing. J. T. Xiong: methodology, writing – review & editing. T. Yang: conceptualization, methodology, resources, writing – review & editing.

Conflicts of interest

All the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The raw/processed data required to reproduce these findings are available from the corresponding author upon reasonable request.

Acknowledgements

The authors greatly acknowledge the financial support from the National Natural Science Foundation of China (Grant No. 52222112), and the Hong Kong Research Grant Council (RGC) (Grant No. 21205621). APT research was conducted at the Inter-University 3D APT Unit of City University of Hong Kong, which is supported by the CityU grant 9600011 and 9360161. This work was supported by Sinoma Institute of Materials Research (Guang Zhou) Co., Ltd (SIMR) to assist the TEM characterization.

References

1. R. C. Reed, The superalloys: fundamentals and applications, Cambridge University Press, 2008.
2. G. Hoppin III and W. Danesi, JOM, 1986, 38, 20–23.
3. G. Gudivada and A. K. Pandey, J. Alloys Compd., 2023, 963, 171128.
4. L. Xu, Z. Sun, Q. Ruan, L. Xi, J. Gao and Y. Li, Energies, 2023, 16, 668.
5. F. N. Nourin and R. S. Amano, J. Energy Resour. Technol., 2021, 143, 080801.
6. N. F. Kazakov, Diffusion bonding of materials, Elsevier, 2013.
7. Y. Peng, Z. Li, W. Guo, J. Xiong and J. Li, J. Mater. Process. Technol., 2024, 324, 118267.
8. Y. Wang, Z. Ding, P. Li, Z. Qin, X. Bi, L. Zhang, C. Li, H. Dong, H. Sun and Y. Cheng, Composites, Part B, 2025, 297, 112330.
9. J. Xiong, Y. Peng, M. Samiuddin, L. Yuan and J. Li, J. Mater. Eng. Perform., 2020, 29, 3277–3286.
10. R. H. AL-Nafeay, A. O. AL-Roubaiy and H. Omidvar, IOP Conf. Ser.:Mater. Sci. Eng., 2021, 1094, 012141.
11. A. I. Ghahferokhi, M. Kasiri-Asgarani, K. Amini, R. Ebrahimi-kahrizsangi and M. Rafiei, Weld. World, 2021, 65, 329–343.
12. M. Pouranvari, A. Ekrami and A. Kokabi, Mater. Des., 2013, 50, 694–701.
13. X. Zhang, K.-S. Li, N. Yao, J.-M. Shi, T.-W. Lu, W.-R. Nie, R.-Z. Wang, X.-C. Zhang and S.-T. Tu, Mater. Sci. Eng. A, 2025, 923, 147723.
14. S. Li, J. Li, J. Shi, Y. Peng, X. Peng, X. Sun, F. Jin, J. Xiong and F. Zhang, J. Mater. Sci. Technol., 2022, 96, 140–150.
15. S. Chen, G. Liu, H. Li, L. Yang, X. Zhang and J. Shen, Adv. Eng. Mater., 2024, 26, 2302241.
16. G. Liu, F. Lin, J. Yang, X. Wang and J. Zou, Vacuum, 2022, 206, 111501.
17. L. Yuan, J. Xiong, J. Ren, T. Yang, X. Xiao, Y. Ma, J. Luan and J. Li, Composites, Part B, 2022, 231, 109568.
18. L. Yuan, F. Jiang, D. Hao, Y. Yang, T. Chou, J. Zhang, J. Gan, J. Li, J. Xiong and T. Yang, J. Mater. Sci. Technol., 2025, 222, 152–163.
19. Y. Lin, X. Jiangtao, D. Yajie, R. Jin, S. Junmiao and L. Jinglong, J. Mater. Sci. Technol., 2021, 61, 176–185.
20. L. Yuan, Y. Yang, J. Gan, T. Chou, Y. Zhao, D. Hao, J. Zhang, J. Li, J. Xiong and T. Yang, Acta Mater., 2025, 121186.
21. T. Yang, Y. Zhao, Y. Tong, Z. Jiao, J. Wei, J. Cai, X. Han, D. Chen, A. Hu and J. Kai, Science, 2018, 362, 933–937.
22. T. Yang, Y. Zhao, W. Liu, J. Kai and C. Liu, J. Mater. Res., 2018, 33, 2983–2997.
23. Y. Zhao, T. Yang, Y. Tong, J. Wang, J. Luan, Z. Jiao, D. Chen, Y. Yang, A. Hu and C. Liu, Acta Mater., 2017, 138, 72–82.
24. Y. Zhao, W. Xiao, Z. Zhao, Q. Li, J. Cui, J. Luan, C. Liu, P. Liaw and T. Yang, Scr. Mater., 2023, 229, 115371.
25. J. Lin, D. Yu, P. Lin, D. Ma, L. Xing, P. He, T. Lin and W. Yang, J. Mater. Res. Technol., 2024, 28, 3670–3675.
26. J. Xiong, L. Yuan, Y. Zhu, H. Zhang and J. Li, J. Mater. Sci., 2019, 54, 6552–6564.
27. H. Chen, S. H. Kim and C. Jang, J. Mater. Sci., 2020, 55, 3652–3667.
28. Y. Peng, C. Li, H. Li and Y. Liu, J. Manuf. Process., 2021, 72, 252–261.
29. J. Y. Zhang, B. Xu, M. Y. Sun, D. Z. Li and Y. Y. Li, J. Mater. Sci. Technol., 2020, 40, 54–63.
30. S. Li, P. Liu, J. Song, W. Guo, J. Xiong and J. Li, J. Alloys Compd., 2023, 953, 169899.
31. S. Li, Y. Zu, Y. Du, L. Gao, P. Liu, Y. Peng, W. Guo, J. Xiong and J. Li, Mater. Charact., 2022, 194, 112359.
32. S. Li, X. Sun, Y. Du, Y. Peng, Y. Chen, Z. Li, J. Xiong and J. Li, Crystals, 2021, 11, 1127.
33. Z. Li, Z. Guo, J. Li, W. Guo and J. Xiong, J. Mater. Res. Technol., 2024, 29, 1971–1982.
34. Y. Peng, J. Li, S. Li, Y. Chen, Z. Li, W. Guo and J. Xiong, Mater. Charact., 2022, 189, 111999.
35. G. Zhang, R. Chandel and H. Seow, Sci. Technol. Weld. Joining, 2001, 6, 235–239.
36. B. Zhang, C. Chen, Z. Cai, H. Tian, J. Hou, L. Chai, Y. Lv and J. He, J. Alloys Compd., 2022, 890, 161681.
37. I. Sah, D. Kim, H. J. Lee and C. Jang, Mater. Des., 2013, 47, 581–589.
38. Y. Jiao, G. Sheng, Y. Zhang, C. Xu and X. Yuan, Mater. Sci. Eng. A, 2022, 831, 142204.
39. Y. Peng, C. Li, Q. Guo, H. Li and Y. Liu, Mater. Charact., 2023, 202, 113031.
40. W. MacDonald and T. Eagar, Annu. Rev. Mater. Res., 1992, 22, 23–46.
41. G. O. Cook III and C. D. Sorensen, J. Mater. Sci., 2011, 46, 5305–5323.
42. B. Binesh and A. J. Gharehbagh, J. Mater. Sci. Technol., 2016, 32, 1137–1151.
43. Y. Zhou, J. Zhang, J. Zhang, X. Yao, J. Luan, Q. Li, S. Liu, B. Xiao, J. Ju and S. Zhao, Acta Mater., 2024, 268, 119770.
44. K. Xie, P. Ma, Y. Fang, H. Yang, S. Wan, Z. Wu, J. Shi, K. G. Prashanth, P. Gargarella and L. Zhang, Composites, Part B, 2025, 306, 112786.
45. J. Zhang, T. Huang, K. Cao, J. Chen, H. Zong, D. Wang, J. Zhang, J. Zhang and L. Liu, J. Mater. Sci. Technol., 2021, 75, 68–77.
46. A. Sadeghian, S. E. Mirsalehi, F. Arhami, A. Malekan, N. Saito and K. Nakashima, Metall. Mater. Trans. A, 2021, 52, 1526–1539.
47. T. Zou, Z. Lang, H. Chen, Z. Ye, T. Wu, W. Wang, J. Yang and J. Huang, J. Manuf. Process., 2024, 109, 79–88.
48. Y. Du, J. Song, J. Xiong, S. Li, F. Jin, J. Li, G. Wen and W. Guo, Mater. Des., 2024, 237, 112581.
49. Z. Cheng, H. Zhou, Q. Lu, H. Gao and L. Lu, Science, 2018, 362, eaau1925.
50. J. Li, G. J. Weng, S. Chen and X. Wu, Heterostructured Materials, Jenny Stanford Publishing, 2021, pp. 323–360.
51. Y. Wang, Y. Zhu, Z. Yu, J. Zhao and Y. Wei, Acta Mater., 2022, 241, 118395.
52. X. Lu, Y. Xu, H. Ran, G. Fan, S. Gao, N. Tsuji, C. Huang, H. Gao and Y.-W. Zhang, Acta Mater., 2025, 289, 120928.
53. H. L. Danflou, M. Marty and A. Walder, Superalloys, 1992, 1992, 63–72.
54. M. Tanaka, H. Iizuka and F. Ashihara, J. Mater. Sci., 1988, 23, 3827–3832.
55. R. Misra, Mater. Technol., 2025, 40, 2485218.
56. Y. T. Tang, P. Karamched, J. Liu, J. C. Haley, R. C. Reed and A. J. Wilkinson, Acta Mater., 2019, 181, 352–366.
57. S. Wu, T. Yang, B. Cao, J. Luan, Y. Jia, L. Xu, Y. Mu, T. Zhang, H. Kong and X. Tong, Scr. Mater., 2021, 204, 114066.
58. W. Xiao, W. Xu, J. Zhang, J. Zhang, S. Liu, B. Xiao, Y. Zhou, J. Ju, Q. Li and Y. Zhao, Acta Mater., 2025, 121084.
59. B. Xiao, J. Zhang, S. Liu, Y. Zhou, J. Ju, J.-J. Kai, Y. Zhao, X. Yang, L. Xu and S. Zhao, Acta Mater., 2024, 262, 119459.
60. X. Xi, D. Lin, Z. He, R. Ma, H. Wei, Z. Shi, W. Zhao, B. Chen, C. Tan and Z. Dong, Composites, Part B, 2025, 291, 112038.
61. A. Goodfellow, Mater. Sci. Technol., 2018, 34, 1793–1808.
62. T. Pretorius and E. Nembach, Acta Mater., 2001, 49, 1971–1980.
63. B. Cao, W.-W. Xu, C. Yu, S. Wu, H. Kong, Z. Ding, T. Zhang, J. Luan, B. Xiao and Z. Jiao, Acta Mater., 2022, 229, 117763.
64. J. Li, J. Zhang, Z. Li, Q. Wang, C. Dong, F. Xu, L. Sun and P. K. Liaw, J. Mater. Sci. Technol., 2024, 186, 174–187.
65. L. Liu, Y. Zhang, J. Han, X. Wang, W. Jiang, C. T. Liu, Z. Zhang and P. K. Liaw, Adv. Sci., 2021, 8, 2100870.

---