A Mechanism of Argon Arc Remelting of LPBF 18Ni300 Steel Surfaces

Abstract

This study aims to reduce pores, cracks, and other defects on the surface of laser powder bed fusion (LPBF)-fabricated 18Ni300 steel and improve its surface quality. Remelting was carried out on the surface with an argon arc as the heat source. Then, the surface layer was characterized using SEM, EDS, XRD, EBSD, and hardness testing. The results showed the following: When the pulse current $I$ increased from 16 A to 20 A, the surface hardness of LPBF 18Ni300 increased due to a decrease in defects and an increase in the martensite phase. The driving forces of convection in the molten pool (such as buoyancy, Lorentz magnetic force, surface tension, and plasma flow force) rose with an increase in current. When the current $I$ exceeded 20 A, the convection became more intense, making it easier for gas to be entrained into the melt pool, forming pores and introducing new defects, resulting in a decrease in surface hardness. The primary factors affecting the hardness of LPBF 18Ni300 after surface argon arc remelting were pore (defect) weakening and phase transformation strengthening, while the secondary factors included grain refinement strengthening and texture strengthening. The solidification mode of the remelted layer was: L → A → M + A′. The phase transition mode of the heat-affected zone was: M + A′ → A_reverse → M_temper. Compared with the base material and heat-affected zone, the grains in the remelted layer formed a stronger <001> texture with a larger average size (2.51 μm) and a lower misorientation angle. The content of the residual austenite A′ was relatively high in the remelted layer. It was distributed in the form of strips along grain boundaries, and it always maintained a shear–coherent relationship with martensite.

1. Introduction

The 18Ni300 steel, as a maraging steel, has the advantages of ultra-high strength, good toughness and wear resistance and is widely used in aerospace, defense, mold, and other fields [1,2,3]. The laser powder bed fusion (LPBF) process has been one of the research hotspots in recent years. The parts manufactured by this method have the advantages of fine grain size, high forming efficiency, and excellent performance [4,5,6]. Huang Gao et al. [7] manufactured 18Ni300 steel using the laser powder bed fusion process, with a laser power up to 2000 W. The results showed that the manufacturing rate of this process was about 3 times higher than that of low-power laser powder bed fusion. LPBF 18Ni300 steel not treated by solution aging had a yield strength (YS) of up to 983 ± 10 MPa, an ultimate tensile strength (UTS) of 1100 ± 11 MPa, and an elongation (E) of 9.8 ± 0.7%, which were comparable to forged 18Ni300 steel. This indicated that the 18Ni300 produced by laser powder bed fusion could meet many service conditions. However, studies [8,9,10] showed that the parts manufactured by LPBF generally had some defects, such as pores, unmolten metal powder inclusions, and microcracks, due to uneven deformation, uneven molten pool temperature, and different cooling rates of materials in various passes during the forming process of laser powder bed fusion. The above-mentioned defects might also appear on the surfaces of the parts, reducing their surface hardness, wear resistance, corrosion resistance, etc. [11,12,13] and limiting the further use of LPBF 18Ni300 in wear-resistant and corrosion-resistant environments.

Surface remelting is one of the most effective ways to reduce such defects as pores, unmelted metal powder inclusions, and microcracks on the surface of LPBF parts and improve their surface quality. Currently, laser surface remelting (LSR) has been widely studied [14,15,16]. Zhaowei Xiang et al. [17] conducted numerical and experimental studies on laser surface remelting under different linear energy densities (LEDs). The results showed that LSR was a feasible method for improving surface quality. When a dual LSR with medium LED was adopted, the surface morphology was quite smooth and free of obvious defects. Jiantao Zhou et al. [18] adopted in situ laser surface remelting (LSR) to eliminate defects in LPBF AlSi10Mg alloy. The results showed that the average grain size decreased by 8.7%, the low-angle grain boundaries increased from 70.9% to 99.7%, and the surface smoothness and densification were significantly enhanced after LSR under nitrogen protection. In addition to LSR, argon arc surface remelting can also be used to improve surface quality, achieve uniform microstructure, and reduce cracks, pores, and other defects. Moreover, it has simple requirements for experimental equipment, low process costs, and high work efficiency [19]. Li Ya-long et al. [20] adopted tungsten inert gas welding to remelt NiCrBSi coatings. The results showed that remelting almost completely eliminated the structural non-uniformity in the sprayed coating and increased the elastic modulus by 84.4%, the fracture toughness by 18.5×, and the microhardness by 42.8%. Argon arc remelting is a feasible method for significantly improving the structural and mechanical properties of coatings. Currently, few researchers adopt argon arc remelting to reduce the above-mentioned defects on the surface of LPBF parts.

In this work, the argon arc generated by the high-precision cold welder will be used as a heat source to remelt, solidify, and crystallize the surface of LPBF 18Ni300 to achieve the goals of eliminating defects and improving surface quality. The research will be carried out from three aspects—molten pool behavior, solidification process, and performance impact—to clarify the mechanism and lay a theoretical foundation for the application of pseudo-laser argon arc remelting technology.

2. Materials and Methods

2.1. Materials

The vacuum aerosol powder of 18Ni300 steel is selected as the raw material, and its chemical composition (mass fraction, wt%) is shown in

Table 1. Chemical composition of 18Ni300 (mass fraction, wt%) [21].

Elements	Ni	Ti	Co	Al	Mo	Si	Cr	Mn	C	Fe
Content	17.70	0.72	9.05	0.077	4.70	0.025	0.031	0.022	0.007	Bal.

[21]. The powder particle size is 25–53 μm, and the loose density is 4.18 g/cm³. The powder is put into a vacuum drying furnace for 8 h at 40 °C to remove water vapor from the powder.

The above powder is put into the DMP Flex350 LPBF equipment (GF Company in Schaffhausen, Switzerland), and the Ar gas with a purity of 99.9% is introduced into it. The laser scan path and main parameters are set on the computer, such as a laser power of 230 W, a laser scanning speed of 1100 mm/s, a single-layer forming thickness of 30 μm, and a scanning spacing of 0.1 mm. After forming, the wire cutting method is adopted to process 18Ni300 steel into a shape of 55 × 10 × 2 mm.

The surfaces of LPBF 18Ni300 steel are polished with 100-mesh, 400-mesh, and 800-mesh sandpaper in sequence. Then, it is put into the JM-60ST ultrasonic cleaner (Shenzhen Jiemeng Technology Co., Ltd., Shenzhen, China), and anhydrous ethanol is added for cleaning to remove rust, oil stains, impurities, etc. The ultrasonic vibration frequency is 40 kHz, the power is 480 W, the cleaning time is 10 min, and the temperature is 40 °C.

2.2. Argon Arc Remelting

The GT-2500 pseudo-laser cold welder produced by Anhui Zhilang Electromechanical Equipment Co., Ltd. (Fuyang, China) is adopted, and its parts and the working principle are shown in Figure 1. This experimental equipment precisely controls the welding time through microcomputer (CPU) technology and instantly releases the electrical energy stored in the large capacity capacitor in the form of pulsed argon arc between the tungsten electrode and the workpiece. Characterized by short welding time, low heat generation, and high heat-source concentration coefficient, it can nearly achieve the effect of laser welding. Therefore, the manufacturer named it “pseudo-laser cold welder”.

Argon (Ar) gas is first introduced for protection. Based on practical experience, the parameters of the experimental equipment should be as follows: pulse current: 16–22 A, pulse time: 18 ms, and argon gas flow rate: 12 L/h. Then, manual processing is adopted to melt the surface of the processed LPBF 18Ni300 steel point by point. The welding method is shown in Figure 2, where each circle represents a melting point. After a melting point is completed, the next melting point is melted with the circular edge of the previous melting point as the center. When one pass in the X direction is melted and the next pass needs to be melted in the Y direction, the circular edge of the previous pass is also used as the melting center. Meanwhile, the principle of the next melting point is melted with the circular edge of the previous melting point as the center should be followed, ultimately achieving complete remelting of the entire target area.

2.3. Structural Observation and Performance Test

The microstructure of the cross section of the surface layer is observed with JSM-IT500 scanning electron microscope (SEM) produced by Japan Electron Optics Laboratory (Tokyo, Japan) and 4XB optical metallographic microscope (OM) produced by Shanghai Optical Instrument Factory (Shanghai, China). Elemental analysis is performed with the energy dispersive spectrometer (EDS) attached to the scanning electron microscope. Crystallographic data of the cross section of the surface layer are detected with (Zeiss) Sigma field emission scanning electron microscopy (FE-SEM) produced by Carl Zeiss AG (Jena, Germany) and an electron back-scattered diffraction (EBSD) system. The EBSD specimen is vibration-polished and analyzed with a step size of 0.25 μm and an 800 × 920 scan grid. Physical phase analysis is performed on the surface of the sample with SmartLab X-ray diffractometer, produced by Rigaku Corporation (Tokyo, Japan), with Cu Kα radiation (λ = 0.154056 nm), tube pressure 40 kV, tube current 40 mA, continuous scanning mode, scanning range 10°~90°, and scanning speed 8°/min. The HV-1000A automatic turret Vickers hardness tester, produced by Shandong Laizhou Huayin Testing Instrument Co., Ltd. (Laizhou, China), is used to test the microhardness of the surface with a load of 4.9 N and a holding time of 10 s.

3. Results and Discussion

3.1. Macro-Morphology

Figure 3 shows the macro-morphology of argon arc remelting under different pulse currents: (a) 16 A; (b) 18 A; (c) 20 A; (d) 22 A. The remelted surface appears as a fish scale. According to the measurement with a vernier caliper, the diameters of the molten pool with pulse currents $I$ of 16 A, 18 A, 20 A, and 22 A are 2.85 mm, 3.43 mm, 3.57 mm, and 3.85 mm, respectively. It indicates that the diameter of the molten pool increases gradually with the increase in the pulse current. Meanwhile, it can be seen that the roughness of the remelted surface and the depression depth at the center of the molten pool have increased as well.

3.2. Macrostructure

Figure 4 shows the microstructure of LPBF 18Ni300 steel. During the LPBF process, many defects can be seen in the sample, such as pores, microcracks, and unmelted metal particles, due to uneven deformation, solution splashing, gas ingress, and insufficient laser power during the solid–liquid–solid transformation of 18Ni300 in each pass. These defects may reduce the effective load-bearing area, leading to a decrease in the mechanical properties of the material [10,22].

Figure 5 shows the macrostructure of the remelted layer with different pulse currents $I$: (a) 16 A; (b) 18 A; (c) 20 A; (d) 22 A. Typically, a clear boundary between the remelted layer and the base material requires a notable difference in grain morphology and size across the two regions [23,24]. From the figure, it can be seen that the boundary between the remelted layer and the base material is not obvious at low magnification (the boundary depicted in the figure is in an enlarged state). The primary reason is that the equipment has high accuracy, a high energy concentration of argon arc heat source generated, and a small operating point, which can almost meet the heat source effect of laser powder bed melting. Both argon arc remelting and laser powder bed melting produce smaller melt volumes, higher undercooling during solidification, faster cooling rates, and similar crystallization conditions. Therefore, the shape and size of the grains are not significantly different under low magnification.

The heat flow of the argon arc not only acts on the surfaces of LPBF 18Ni300 but also the thickness direction. The changes in average thicknesses of remelted and heat-affected layers as a function of the current are shown in

Table 2. Changes in current intensity $I$ and average thickness H of remelted and heat-affected layers.

Current Intensity/A	16	18	20	22
Thickness of remelted layer/μm	308 ± 7	333 ± 4	352 ± 5	381 ± 16
Thickness of heat-affected layer/μm	87 ± 7	92 ± 12	98 ± 4	117 ± 13

.

The relationship between argon arc heat input $Q$ nd current $I$ is as follows:

$$Q=\mathsf{\eta }UIt$$

(1)

where $\mathsf{\eta }$ is thermal efficiency, $U$ is the argon arc voltage, and $t$ is action time. The thickness of the workpiece is 2 mm, and the speed of manual argon arc remelting is not fast. The form of the argon arc heat source is considered a point heat source. The energy of the argon arc heat source follows the Gaussian distribution. The heat input $Q$ and the depth and diameter of the molten pool increase (namely, increase in the thickness of the remelted layer) as a function of the pulse current $I$. Correspondingly, the thickness of the heat-affected layer increases as well.

When the argon arc with high energy density is applied to LPBF 18Ni300, the surfaces instantly form a molten metal flow in the pool. This flow occurs under some driving forces. The driving forces involved are illustrated in Figure 6 [25,26,27,28] and primarily include: (1) Buoyancy. The density of 18Ni300 liquid decreases with increasing temperature, resulting in a higher temperature and lower density at the center of the molten pool, and a lower temperature and higher density at the edges. Liquid metal sinks along the boundary of the molten pool and rises along the axis (Figure 6a). (2) Lorentz force. The current on 18Ni300 converges toward the tungsten electrode and the center of the molten pool surface. The converging current field and corresponding electromagnetic field jointly generate a downward and inward Lorentz force, pushing the metal liquid downward along the axis of the molten pool and upward along the boundary of the molten pool (Figure 6b). (3) Surface tension. The surface tension of metal liquid decreases as the temperature rises, and the surface tension at the center of the molten pool is greater than that at the boundary. An outward shear stress is generated on the surface of the molten pool along the surface tension gradient of the molten pool, causing the metal liquid to flow from the center of the surface to the edges and return below the surface of the molten pool. This convection is also known as “Marangoni convection” (Figure 6c). (4) Plasma flow force. The plasma moves outward at high speed along the surface of the molten pool and may apply an outward shear force on the surface of the molten pool, causing the liquid metal to flow from the center of the molten pool surface to its edges and then return below the surface of the molten pool (Figure 6d). The action of these driving forces causes the convection of liquid 18Ni300 in the molten pool and the material to refill the pores, making the remelted structure more uniform and dense. Compared with Figure 4, the defects, such as pores, microcracks, and unmelted metal particles in the remelted layer, are significantly reduced after argon arc remelting, as shown in Figure 5. Meanwhile, strong driving forces cause the metal liquid to flow to other places and then solidify very quickly, which results in uneven surfaces after remelting to a certain extent (Figure 3).

As the current $I$ gradually increases to 22 A, larger pores appear in the upper part of the remelted layer, and a small number of small shrinkage pores also appear at the top. The magnitude of driving forces, such as buoyancy, Lorentz force, surface tension, and plasma flow force, is related to the amount of heat input. Namely, it is directly related to the pulse current intensity $I$. The higher the current intensity, the greater the driving force, and the more intense the convection in the molten pool. Accordingly, the protective gas (Ar) out of the air or welding gun is more likely to be drawn into the melt. If not escaping in time, it may form pores (Figure 5d). Higher driving forces may cause more metal liquid to flow to other places. During solidification, it is difficult to timely supplement the voids caused by volume reduction, resulting in the formation of shrinkage pores at the top of the remelted layer (Figure 5d).

Among chemical elements in 18Ni300 steel, Ni, Mo, and other elements with minimum segregation energy and lower E_seg are more likely to diffuse to the grain boundaries, causing element segregation. To understand the overall distribution of chemical elements, EDS surface scanning analysis has been performed, as shown in Figure 5c, and the results are shown in Figure 7. The content of each element in the entire surface is shown in

Table 3. EDS surface scan analysis of element content (mass, wt%) in Figure 5c.

Elements	Fe	Ni	Co	C	Mo	O
Content	59.0 ± 1.1	16.1 ± 1.2	9.3 ± 0.2	7.4 ± 0.2	4.8 ± 0.2	2.9 ± 0.1

. Locations showing the uneven distribution of elements in Figure 7 have been marked with circles A and B, mainly in the base material. The shaded area in the figure indicates that the concentrations of major elements, such as Fe, Ni, and Co, are relatively low. Compared with Figure 4, it is found that the main element deficiency is caused by incomplete fusion pores at location A, and the O element concentration is relatively high at location B, which are possibly oxides of Fe, Ni, and Co. The argon arc remelts the surface layer, and the molten pool forms convection under the driving forces, making the distribution of elements in the remelted layer more uniform.

3.3. Grain Structure

Figure 8 shows the microstructure of the grains in the molten pool. Cellular crystals, cellular dendritic crystals, dendritic crystals, and equiaxed crystals can be seen from the bottom to the center of the molten pool. According to the solidification theory [29], the grain shape in the molten pool is related to the ratio of G/R where G is the temperature gradient at the solid–liquid interface and R is the crystallization rate. At the boundary of the molten pool, the melt nucleates with the base material at the bottom of the molten pool, and the temperature gradient G is great. Due to the Gaussian distribution of argon arc heat, the temperature at the center is extremely high, and the heat continuously conducts from the center of the molten pool, resulting in a slow crystallization rate R. Therefore, the G/R ratio is great, and the crystal particles are in cellular crystallization, grow preferentially in the opposite direction along the heat dissipation direction, and form a perpendicular angle of α = 90° with the tangent line to the bottom boundary of the molten pool. As the solid–liquid interface moves toward the center of the molten pool, the temperature gradient G gradually decreases, the overall heat of the molten pool decreases, the heat transferred to the solid–liquid interface decreases, the crystallization rate R increases, and the G/R ratio gradually decreases. This may cause constitutional supercooling, and some protruding parts of the cellular crystals first extend into the constitutional supercooling area, gradually grow branches, and transform into dendritic crystals. The liquid phase at the center of the molten pool may form a wide constitutional supercooling area with a negative temperature gradient G. Due to no obstruction around, these crystal nuclei may grow freely and form equiaxed grains.

The grain size gradually increases from the bottom to the center of the molten pool. According to the classical solidification principle [30], the relationship between the critical nucleus radius and the $r_k$ supercooling degree $\Delta T$ is as follows:

$$r_k={\displaystyle \frac{2\sigma }{\Delta G_r}}={\displaystyle \frac{2\sigma T_m}{L_m·\Delta T}}$$

(2)

where $\sigma$ is the interface energy between the crystal nucleus and the liquid phase, $\Delta G_r$ is the change of volume free energy, $L_m$ is the latent heat of crystallization, $T_m$ is the melting point, and $\Delta T$ is the supercooling degree. It can be seen that the greater the supercooling, the smaller the critical radius of crystal nucleation, the higher the nucleation rate, and the finer the grain size of the alloy. Crystallization first occurs at the bottom of the molten pool, with the base material as the nucleation core. Here, the supercooling degree $\Delta T$ is the highest, the nucleation rate is high, and the grain size is the smallest. As the solid–liquid interface moves toward the center of the molten pool, the supercooling degree $\Delta T$ gradually decreases and the grain size gradually increases.

EBSD is adopted to detect and analyze argon arc remelting samples of LPBF 18Ni300, with a remelting current of 20 A and a time of 18 ms. Figure 9 shows the grain morphology (Figure 9a) and misorientation angle distribution (Figure 9b) of the base material, heat-affected zone, and remelted layer in the sample. It can be seen that the grain morphology of the base material is similar to that of the remelted layer and is mostly in the long-rice-shaped cellular structure, while the morphology in the heat-affected zone has changed significantly, mostly growing into a lath shape. The misorientation angles are also different and decreasing in sequence from the heat-affected zone (maximum misorientation angle) to the base material and the remelted layer.

Figure 10 shows the orientation of martensitic grains in different areas. The grain orientation of the base material is not obvious, which may be related to the small number of samples taken; the grains in the heat-affected zone have a tendency toward <111> direction, but the texture is not serious; the texture in the <001> direction is clearly formed in the remelted layer.

Figure 11 shows the grain sizes in different areas. It can be seen that the grain size has grown sequentially from the base material through the heat-affected zone to the remelted layer. The average grain size of the base material is about 1.76 μm; the average grain size in the heat-affected zone is 2.39 μm with uneven distribution, some even growing up to 35.7 μm; the average grain size is 2.51 μm in the remelted layer.

The grain morphology, orientation angle difference, grain orientation, and grain size difference of the base material, heat-affected zone, and remelted layer are all related to the heat distribution and heat-source processing accuracy during the argon arc remelting process. The argon arc acts on LPBF 18Ni300, and the temperature is the highest at the center of the molten pool on the surface of the part. The temperature gradually decreases from top to bottom in the thickness direction of the part. According to the solidification theory [31], grains may preferentially grow in the opposite direction of heat dissipation. Because the remelted layer for sample testing is selected in middle and lower parts of the molten pool, the grains are mostly in the cellular structures (Figure 9a), and the grains in the remelted layer are mostly in a <001> direction with high directional consistency (Figure 10e,f), and the misorientation angle is also very small (Figure 9b). The heat continuously transfers outward, and the martensite undergoes a phase transition in the heat-affected zone so that the grain morphology, misorientation angle, grain orientation, and grain size have all undergone changes, which will be discussed in Section 3.4.

Although the heat sources of argon arc remelting and laser powder bed fusion are similar to a certain extent, some differences exist in processing accuracy. The minimum heating area is 10⁻⁸ cm² for the laser heat source and 10⁻³ cm² for the argon arc heat source. Therefore, the molten pool formed during argon arc heating is larger, and the time required for complete solidification is longer, so the grains may grow more easily in the remelted layer than those in the base material (Figure 11). Laser has a smaller minimum heating area and requires more processing passes. Because processing of the next pass may affect the size and direction of the grains in the previous pass, the consistency of grain orientation is not as high as that of the argon arc (Figure 10a,b,e,f).

3.4. Physical Phase Analysis

Figure 12 shows the XRD diffraction pattern of 18Ni300. The 18Ni300 steel is mainly composed of martensitic phase (M) with body-centered cubic (bcc) structure and a small amount of austenitic phase (A) with face-centered cubic (fcc) structure. After argon arc remelting, the diffraction peak intensity at the martensitic phase significantly increases, and a new diffraction peak of the austenitic phase (200) increases at a diffraction angle of 44.5°.

Figure 13 shows the physical phase distribution after argon arc remelting with a current of 20 A and a time of 18 ms (based on EBSD phase detection in Figure 9 and with the same detection location as Figure 9), where red represents martensite M and blue represents austenite A. The overall proportion of austenite is very small, about 0.2%. A small amount of the residual austenite phase appears as dots in the LPBF state and is distributed at some grain boundaries; there is almost no residual austenite in the heat-affected zone and all has been transformed into martensite; residual austenite is relatively high in the remelted layer and has grown into strips along the grain boundaries. Figure 14 shows the orientation of residual austenite A′ grains in the remelted layer, which are mainly growing along the <001> direction.

During LPBF forming and remelting, 18Ni300 undergoes rapid non-equilibrium solidification after being melted by two heat sources—laser and argon arc. Its solidification mode is consistent as: L → A → M + A′. Due to significant thermal and volume effects of the transformation between new and old phases in the process of austenite to martensite transformation in the base material and remelted layer, two phases are squeezed against each other in order to achieve the lowest energy state and some austenite at grain boundaries does not have enough space for martensitic transformation, resulting in residual austenite [32]. As shown in Figure 14, the strip-shaped residual austenite has a <001> direction, which is consistent with the direction of martensite (Figure 10e,f), and the new and parent phases always maintain a shear–coherent relationship. In combination with the study of Wu W.W. et al. [33,34], a classic Nishiyama–Wassermann (N–W) orientation relationship exists between residual austenite (A′) and martensite (M): (110)M || (111)A and [001]M || [−101]A. Because the molten pool formed by the argon arc heat source is larger than that formed by the laser heat source, the reduction rate of free energy of the system is slower, and the driving force for the austenite to martensite transformation is not as strong as the latter. Therefore, the content of residual austenite is higher than that with the laser heat source.

The martensitic phase transition in the heat-affected zone is reversible. When heated, it can be reversed to austenite A_reverse, and when rapidly cooled, austenite can be transformed back to martensite M_temper. Because the volume of the melting zone may shrink during solidification and is subjected to tensile stress, the compressive stress generated by volume increase due to martensitic phase transition can be released in the heat-affected zone. Therefore, the transformation of martensite is relatively thorough, almost without residual austenite (Figure 13). According to the above analysis, the heat-affected zone rapidly undergoes heating and cooling processes again, and its phase transition mode is M + A′ → A_reverse → M_temper. Most martensitic phase transitions nucleate at such positions as vacancies, dislocations, and grain boundaries, causing atoms to shift in a shear manner and grow at an extremely fast rate (equivalent to 10⁵ cm/s). When energy imbalance is caused by such factors as specific volume difference, body mismatch, and elastic modulus between old and new phases, martensite achieves the lowest energy state by adjusting its morphology and structure. Therefore, most grains grow into a lath shape and a few into a needle shape, with uneven shape and size (Figure 9a). A few grains even grow up to 35.7 μm (Figure 11b) and the misorientation angle of grains is also great (Figure 9b). The martensitic structure is a body-centered cubic (bcc) structure, with atoms densely packed in the <111> direction, and martensite mainly grows along this direction. However, due to the energy imbalance between new and old phases, the growth morphology and size are uneven, and the consistency of final growth direction is not high. Only some grains are growing in the <111> direction (Figure 10c,d).

3.5. Vicker’s Hardness

Figure 15 shows the Vickers hardness of LPBF 18Ni300 after surface argon arc remelting. As the heat input $Q$ increases (current $I$ increases), the surface hardness of LPBF 18Ni300 first increases and then decreases. When the current is 20 A, the hardness is 389.0 HV (the highest), 11.2% higher than that (349.8 HV) of the base material. When the current reaches 22 A, the surface hardness is 325.3 HV, 7% lower than that of the base material. After argon arc remelting, the hardness of the heat-affected zone is the highest (average 423.6 HV) in different areas of the same sample, about 21.1% higher than that of the base material. In the remelted layer, the hardness gradually increases from the center to the bottom of the molten pool.

Figure 16 shows the schematic diagram of the argon arc remelting results on the surface of LPBF 18Ni300. Blue color represents the remelting zone where pore defects are basically eliminated, and the grain size gradually decreases from the center of the molten pool to the bottom, forming a texture in the <001> direction; red color represents the heat-affected zone, and the microstructure is tempered with M; gray color represents the base material containing a large number of defects.

Main factors affecting the surface argon arc remelting hardness of LPBF 18Ni300 are analyzed and discussed as follows:

(1)

Weakening of pores (defects). In general, the strength and porosity of materials can be expressed as [31,35]:

$${\sigma }_p={\sigma }_0\exp (-\mathrm{n}\mathrm{p})$$

(3)

where ${\sigma }_0$ is the strength when the porosity is 0, $\mathrm{p}$ is the porosity, and $\mathrm{n}$ is a constant. As observed, there is an exponential function relationship between material strength and porosity, and the material strength may be greatly reduced with the increase in its porosity. After argon arc remelting, the defects, such as pores, microcracks, and unmelted metal particles in the remelted layer of LPBF 18Ni300, are significantly reduced (as shown in the blue part of Figure 16). When materials are subjected to a load, pore elimination may increase the cross-sectional area of the load on the one hand; and stress concentration is prone to occur near voids, avoiding the formation of crack sources and accordingly increasing material strength on the other hand. However, when the current increases to 22 A, the driving force in the molten pool increases, the convection becomes more intense, and larger pores and a few small shrinkage pores appear in the remelted layer, resulting in a significant decrease in hardness. Therefore, the weakening mechanism of pores (defects) is the primary factor affecting the hardness change of LPBF 18Ni300 after surface argon arc remelting.

(2)

Phase transition strengthening. As shown in Figure 13, the heat-affected zone is completely transformed into tempered martensite, which has high strength and hardness after the argon arc action. According to the mixing ratio principle [36,37], the strength of a material under load is as follows:

$${\sigma }_c={\sigma }_{\alpha }{v}_{\alpha }+{\sigma }_{\beta }{v}_{\beta }$$

(4)

where ${\sigma }_c$ is material strength, ${\sigma }_{\alpha }$ is the phase strength of $\alpha$, $v_{\alpha }$ is the phase volume fraction of $\alpha$, ${\sigma }_{\beta }$ is the phase strength of $\beta$, and $v_{\beta }$ is the phase volume fraction of $\beta$. According to Table 2, as the current $I$ increases, the thickness of the heat-affected layer gradually increases (as shown in the red part of Figure 16), and the volume content of tempered martensite increases as well. Then, the hardness increases accordingly. Due to a small amount of tempered martensite, the phase transition strengthening mechanism plays a secondary role in the surface argon arc remelting of LPBF 18Ni300.

(3)

Fine-grain strengthening. According to the Hall–Petch formula, the material strength and the grain size and content conform to the following relationship [38,39]:

$${\sigma }_s={\sigma }_0+k{d}^{-1/2}$$

(5)

where ${\sigma }_s$ is the yield strength, ${\sigma }_0$ and $k$ are constants, and $d$ is the grain size. The smaller the grain diameter $d$, the higher the yield strength. As shown in Figure 8 and Figure 15, the grain size gradually decreases and the hardness gradually increases from the center to the bottom of the molten pool, which conforms to the Hall–Petch formula. However, in Figure 11, average grain sizes of the base material, heat-affected zone, and remelted layer are approximately 1.76 μm, 2.39 μm and 2.51 μm, respectively. After the calculation, the grain sizes and hardness values of the base material, heat-affected zone, and remelted layer are not closely related to Formula (5), mainly due to the significant influence of the martensitic phase transition strengthening mechanism. Overall, the contribution of the fine-grain strengthening mechanism to the hardness of LPBF 18Ni300 after surface argon arc remelting is not as significant as the first two factors.

(4)

Texture strengthening. The formula for texture strengthening can be expressed as [40,41]:

$${\sigma }_{tex}={\displaystyle \frac{I_0}{m}}$$

(6)

where ${\sigma }_{tex}$ is the strength contributed by texture strengthening, $I_0$ is a constant and $m$ is the Schmidt factor. The strength ${\sigma }_{tex}$ contributed by texture strengthening is inversely proportional to the Schmidt factor $m$. According to Figure 10, the martensite Schmid factor is 0.26 for the LPBF 18Ni300 sample base material after surface argon arc remelting, 0.29 for the heat-affected zone, and 0.24 for the remelted layer. The difference in Schmidt factor is not significant, and grains with the texture in the <001> direction are mostly distributed at the bottom of the molten pool. After the calculation according to Formula (6), the correlation between hardness change and texture strengthening is not significant. So, the contribution of the texture strengthening mechanism to hardness is very small.

4. Conclusions

Surface remelting of LPBF 18Ni300 with an argon arc heat source can effectively reduce surface defects, such as pores and cracks, and improve hardness. The following conclusions can be drawn after testing and analysis:

(1)

The energy of the argon arc heat source follows a Gaussian distribution. The thicknesses of the remelted layer and the heat-affected layer increase as a function of the current $I$. The molten pool generated by argon arc action may undergo convection under such driving forces as buoyancy, Lorentz force, surface tension, and plasma flow force. The larger the pulse current $I$, the greater the driving forces, and the more intense the convection in the molten pool. Then, air is more likely to be drawn into the solution, forming pores. The convection of the molten pool does not cause macroscopic element segregation.

(2)

From the bottom of the molten pool to the center of the remelted layer, grain morphologies were in the order of cellular crystals, cellular dendritic crystals, dendritic crystals, and equiaxed crystals. So, the grain size gradually increases. The solidification mode of the remelted layer was as follows: L → A → M + A′. The phase transition mode of the heat-affected zone was as follows: M + A′ → A_reverse → M_temper. Compared with the base material and heat-affected zone, the grains in the remelted layer formed a texture of <001> direction with a larger average size of 2.51 μm and a lower misorientation angle. The content of the residual austenite A′ was relatively high in the remelted layer, was distributed in the form of strips along grain boundaries, and always maintained a shear–coherent relationship with martensite.

(3)

When the pulse current I increased from 16 A to 20 A, the surface hardness of LPBF 18Ni300 increased due to the reduction in defects and the increase in the martensite phase. When the current was higher than 20 A, the convection became intense, and gas was easily drawn into the melt to form pores, leading to an increase in defects and a decrease in surface hardness. When the current was 20 A, the surface hardness was the highest, 389.0 HV, which was 11.2% higher than that of the base material. The primary factors affecting the hardness change of LPBF 18Ni300 surface argon arc remelting were pore (defect) weakening and phase transition strengthening, while the secondary factors included fine grain strengthening and texture strengthening.