Novel approach for dilution engineered CoCrFeMnNi in-situ weld overlay coating

Dilution trends and arc physics interpretation

Fe dilution and cladding compositions under the tested conditions (Table 1) were determined employing wide-area energy-dispersive spectroscopy (EDS). Average Fe dilution (at%), standard deviations, and bonding quality are presented in Figure 3. Fe dilution increased systematically with weld current and decreased with traverse speed, as evidenced by the contraction of the Fe compositional domain from experiment 12 to 1 (Figure 3). The weld current dependence originates from four interdependent arc physics mechanisms. Primarily, arc pressure varies proportionally to the square of the weld current, as formulated in Lancaster's electromagnetic force model ³⁰

P a r c = μ 0 I 2 4 π r 2

(5)

OPEN IN VIEWER

Figure 3.

Compositional variation and interfacial bonding at the experimented range of cladding parameters (n = 10).

Wherein μ₀ denotes magnetic permeability (4π × 10⁻⁷ H/m), ‘I’ represents current, and r is arc radius. Elevating current from 120 to 180 A augments pressure by (180/120)² = 2.25-fold, thereby intensifying substrate penetration and melt pool convection. Secondarily, elevated current raises plasma temperature, diminishing melt pool viscosity and promoting Fe diffusion from the substrate. ³⁰ Tertiarily, current amplifies Marangoni convection induced by surface tension gradients, enhancing filler-substrate intermixing. ³¹ Quaternary, arc spreading expands with current, increasing substrate melting. ¹⁰ The observed dilution increment of 2.47-fold (12.9 to 31.9 at% Fe at 120 mm/min) correlates closely with the 2.25-fold pressure escalation, corroborating arc pressure as the predominant current-driven mechanism. Standard deviations were consistently below 1 at% across all conditions, signifying uniform in-situ alloying attributable to vigorous Marangoni (dγ/dT < 0) and Lorentz (J × B) convection within the GTAW molten pool, which homogenizes the multicomponent melt prior to solidification.^25,32 It also should be noted that small EDS/WDS fluctuations only produce minor variations in D_i and do not alter the overall identification of the dilution window or the optimized process condition as the D_i is less sensitive to the compositional fluctuations.

Traverse speed exerts influence via residence time. Elevated speeds diminish heat input per unit length, thereby restricting Fe incorporation. At 160 A, decreasing speed from 120 to 80 mm/min elevated H.I. from 896 to 1344 J/mm, increasing Fe from 30 to 37 at%, in accordance with DuPont-Marder energy balance principles for single-pass arc welds. ¹¹

Anomalies and non-monotonic behavior

Two deviations from monotonic trends were observed: (i) reduced dilution at higher heat input (120 A, 100 mm/min, 806 J/mm: 23 at% Fe vs 140 A, 120 mm/min, 784 J/mm, 27 at%); (ii) disparate dilutions at equivalent heat input (120 A, 80 mm/min versus 180 A, 120 mm/min, both 1008 J/mm, 27.2 versus 31.9 at% Fe). Heat input, being temporally averaged, neglects instantaneous arc dynamics. Dilution emerges from the interplay of arc pressure-dominated melting (scaling with I², prevalent at high current) and conduction-limited melting (scaling with H.I., dominant at low current). At 120 A (∼14.4 kPa), subdued arc pressure constrains penetration irrespective of speed, at 180 A (∼32.4 kPa), pressure overrides residence time effects. This dual-mechanism paradigm extends conventional heat input frameworks, underscoring the necessity of current-resolved dilution evaluation for GTAW-based multi-principal element alloy in-situ processing.

Critical thresholds: critical heat input (H.I._c) and Critical dilution (D_c)

The parametric study shows a minimum substrate dilution that is necessary for reliable interfacial bonding, referred to as critical dilution (D_c), below which resulting a poor bonding. Understanding the dilution-bonding relationship is essential to define the critical heat input and dilution. An element-specific dilution index (D_i) (Eq.4) was employed for better compositional predictability in MPEAs. ²⁴ D_i = 0 indicates no substrate contribution (composition = filler), D_i = 1 indicates complete substrate replacement (composition = substrate), and D_i > 0 represents concentrating and Di < 0 diluting elements. This formulation enables precise tracking of each element's dilution/concentration behavior.

Critical heat input (H.I.c) Identification

Plotting D_i versus H.I. (Figure 4(a)) reveals distinct element specific dilution trajectories. At the lowest heat input (672 J/mm, Exp. 9), all elements cluster near the central reference line (D_i ≈ 0), and indicating composition matches filler with negligible substrate melting. However, macroscopic observation (Figure 4(b)) and SEM interfacial analysis (Figure 4(d)) show semi-molten un-bonded deposits (balling), analogous to lack of fusion defects in laser powder bed fusion. ³³ A sharp transition occurs at 784 J/mm (Exp.10). Fe dilution jumps from ∼12.9 to ∼27 at% (D_i = 0 → 0.18), Co/Cr/Mn/Ni dilute proportionally, and bonding quality shifts from balling to sound metallurgical fusion (Figure 4(c) & (e)). This 672–784 J/mm interval represents the steepest dilution sensitivity region, analogous to balling-to-bonding transitions in arc-welding critical heat input studies. ²⁴ The threshold H.I. ≈ 784 J/mm provides sufficient energy for three pre-requisite conditions. One is fulfill the thermal energy for complete filler melting which overcomes latent heat + sensible heat. Two is reducing viscosity of the filler liquid metal and enable the wetting (γ_LV cos θ > 0). ³⁴ Third is to ensure the good interfacial mixing by establishing concentration gradients for diffusion bonding ³⁵

OPEN IN VIEWER

Figure 4.

(a) Variation of the element specific dilution / concentration with heat input. (b & c) Macro images of the claddings at 672 and 784 J/mm respectively. (d & e) Cross sectional view at the interface of claddings at 672 and 784 J/mm respectively.

Critical dilution (D_c) as bonding threshold

The Fe dilution of 23 at% (Figure 3, Exp.5), representing a minimum substrate contribution for bond integrity. Below this threshold (Exp.9), insufficient substrate melting will prevents the mechanical interlocking and diffusion bonding. Above this threshold metallurgical bonding is maintained but composition deviates from equiatomic target (20 at% Fe), necessitating further refinement. Mechanistically, 23% Fe dilution ensures (1) adequate substrate penetration for interfacial Fe gradient establishment; (2) sufficient melt pool volume ratio (substrate:filler ≈ 1:3) to sustain convective mixing. ¹¹ (3) thermal compatibility between substrate (Fe-dominant BCC) and filler (FCC-stabilizing Co/Cr/Ni) for epitaxial solidification at interface. These results explicitly confirm critical thresholds H.I. (∼784 J/mm) and least possible dilution without suffering the bonding exist as critical dilution (D_c) threshold for good bonding. The H.I and dilution thresholds were calibrated for single pass and 8 mm thick LCS substrate only. This will serve as a reference to multi-pass, multi-layer depositions and multi-thickness LCS substrate cladding studies.

Powder volume effect on dilution

In Stage II, the narrow process window from Stage I (120 A, 100 mm/min) was adopted as baseline, with powder volume fraction (P_v) varied from 20% to 60% in 10% increments. Compositions were analyzed and summarized in Figure 5, wherein the nominal equiatomic composition is depicted at the bottom (X-axis: element at%; Y-axis: P_v). Three claddings were deposited for each P_v level ensured reproducibility, with standard deviations of 0–1.5 at% affirming compositional uniformity. P_v controls Fe dilution by repartitioning GTAW arc energy between metal-cored filler melting and LCS substrate melting. As Per Eq.6, ¹¹ total arc energy (E_arc) equals filler melting energy (E_fm), substrate melting energy (E_s), and losses (E_l). At fixed cladding conditions (120 A, 100 mm/min, 16 V), constant electrical input per length implies P_v-induced E_fm will be rising and necessitate reduced Es, limiting substrate melting and Fe pickup.

E arc = E fm + E s + E l

(6)

OPEN IN VIEWER

Figure 5.

(a) Efm Vs Pv graph (b) Variation of CoCrFeMnNi MPEA cladding composition with respect to P_v (c) D_E Vs D_A graph.

Thermo-Calc based calculations (Figure 5) illustrate P_v's impact. E_fm surges three orders of magnitude from 1637 to 5172 J/cm³ as P_v increases from 0.2 to 0.6 (Figure 5(a)), owing to greater solid mass and elevated heat capacity/latent heat of the Co–Cr–Mn–Ni mixture. Constant E_arc thus curtails substrate volume melted, reducing Fe pickup. EDS data (Figure 5(b)) corroborate this fact and Fe falls from ∼23 at% (20% P_v) to ∼11 at% (60% Pv; vertical dotted circle), achieving equiatomic composition at 30% Pv (horizontal dotted circle). Co and Mn rise proportionally with declining Fe, whereas Ni and Cr remain stable, reflecting their higher diffusion activation energies (>300 kJ/mol vs. ∼270 kJ/mol for Co/Mn). ³⁶ To verify dilution predictability over empiricism, estimated dilution (D_E) from energy partitioning (modified Banovic model, accommodating no-feed filler and porosity for P_v dependent E_fm) was compared to actual dilution (D_A) from wide area EDS across P_v levels. Figure 5(c) reveals near-perfect correlation (R² = 0.981), with 20–60% P_v data aligning tightly on the linear fit. This validates Fe dilution is possible to engineer via energy partitioning, predictable from thermo-physical data rather than trial-and-error. Thus, Stage II establishes P_v as a precise, quantitative lever for dilution control.

Structural characterization of the dilution engineered cladding

Figure 6 summarizes the solidification microstructure of the Pv optimized in-situ GTAW cladding. The macro and cross-sectional views (Figure 6(a)–(c)) show a sound single track bead with the classical sequence controlled by the temperature gradient (G) to growth rate (R) ratio G/R. ³⁷ Near the fusion boundary, high G and low R (typically G > 10⁵ K/m, R ∼10⁻² m/s for arc welding) produce planar growth (Figure 6(d)), which evolves into columnar dendrites as G decreases and R increases with radial heat flow (Figure 6(e)). ³⁷ When G/R falls below the critical threshold (∼2.5 × 10⁶ K·s/m² for FCC alloys), a columnar to equiaxed transition occurs, giving fine equiaxed grains at the surface (Figure 6(f)).^38,39 This G/R controlled progression agrees with reported CoCrFeMnNi GTAW welds and laser claddings, where columnar grains are typically 100–250 µm and equiaxed grains 20–80 µm.^40,41 Grain size in the present cladding refines from ∼190 ± 50 µm (columnar) to 62.5 ± 19.6 µm (CET) and 30.6 ± 9.1 µm (equiaxed) at the surface (Figure 6(d)′–(f)′), matching ranges for arc welded Cantor alloys.^42,43 This refinement reflects GTAW's intermediate cooling rates of 72–109 °C/s, higher than arc melting (∼10 °C/s, >500 µm grains) and lower than laser processing (10³–10⁵ °C/s). ⁴⁴ These results show that dilution engineered GTAW can deliver grain refinement together with weld typical zoning, satisfying the microstructure equivalence component of the central hypothesis. ⁴⁴ The reported G and R values are literature based contextual parameters used for interpretation, rather than directly measured quantities in the present work. The present study does not independently decouple dilution and cooling rate, because both evolve simultaneously during GTAW cladding. Dilution mainly controls the final chemistry through Fe pickup from the substrate, where as cooling rate governs the solidification morphology and grain refinement therefore, the observed microstructure is best interpreted as the coupled effect of composition and thermal history.

OPEN IN VIEWER

Figure 6.

(a) Macro image of the Pv optimized cladding (b) Cross-sectional view (c) Microscopic image at the interface (d-e) Microscopic SEM images of different regions marked in image b (d’-f’) Grain size distribution at d,e and f locations (g) XRD phase analysis (h) Elemental mapping in the cladding.

XRD analysis (Figure 6g) reveals a dominant single FCC phase with peak positions corresponding to (111), (200), (220) and (311) and lattice parameter a ≈ 0.3594 ± 0.0003 nm (Nelson-Riley extrapolation) indistinguishable from equiatomic CoCrFeMnNi produced by vacuum melting or laser processing. This phase selection is fully consistent with established MPEA phase forming rules^45,46 the CoCrFeMnNi in-situ cladding exhibits negative mixing enthalpy (ΔH_mix = −4.5 kJ/mol), small atomic size mismatch (δ ≈ 1.6%) and valence electron concentration ≈ 8, all favoring FCC solid solution formation over intermetallics across the measured dilution range. The absence of secondary peaks above the XRD detection limit confirms that dilution engineering has retained the alloy within the FCC stability field. However minor phases below the detection limit of XRD cannot be completely excluded and EBSD or TEM would be required to detect nano-scale precipitates or ordering. High-magnification SEM–EDS elemental maps (Figure 6 h) reveal uniform distribution of Co, Cr, Fe, Mn, and Ni across the cladding, with minor micro segregation typical of as welded MPEAs and no macro-scale coring or banding. Sluggish diffusion in high-entropy liquids (η ∼10⁻³ Pa·s) ⁴⁷ further suppresses coring during rapid 72–109 °C/s cooling, avoiding defects common in conventional alloys.

Corrosion resistance evaluation

Potentiodynamic polarization (PDP) curves (Figure 7(c)) for the in-situ CoCrFeMnNi cladding, VM alloy and LCS substrate show clear difference in the electrochemical parameters (TableS4, eq-5 & 6 supplementary data). ⁴⁸ Both MPEAs exhibit passivation between −0.38 and −0.1 V (SCE). The in-situ cladding shows a stable passive region (−0.340 to −0.180 V), indicating a robust oxide film, whereas the VM alloy passivates less effectively due to its coarse microstructure. Grain boundaries will promote local galvanic activity, while very coarse grains reduce re-passivation pathways. Optimal corrosion resistance typically occurs at 10–70 µm grain sizes.^49,50 The in-situ cladding's top surface grains (30.6 ± 9.1 µm) fall within this window, whereas the VM alloy (>500 µm) lies far outside, contributing to inferior performance. Cr/Mn oxide inclusions cover ∼1.5% area, Figure 6(d) and are isolated, not continuous. Though they are potential sites for localized attack, their limited extent and the stable passive behavior suggest minimal influence under the test conditions (3.5 wt% NaCl, RT, 30 min OCP). Grain refinement, rapid solidification, and minimized segregation enhance the cladding's corrosion resistance. Corrosion rates follow the trend of in-situ cladding (0.93 mpy) < VM alloy (2.17 mpy) << LCS substrate (44.77 mpy). The cladding also shows a nobler E_corr (93 mV) and 57% lower I_corr than the VM alloy (217 mV). The results show an indication of promising short-term passive behavior, while long term immersion and localized corrosion testing are necessary for service relevant validation.