Effect of the Laser Cladding Parameters on the Crack Formation and Microstructure during Nickel Superalloy Gas Turbine Engines Repair

Abstract

Cracking of nickel superalloys with a high content of $\gamma$’-phase remains an unresolved problem, including in technologies for repairing gas turbine engines blades. Laser cladding is a method of material deposition used to repair parts exposed to aggressive environment and surface wear. Cladding parameters have a high influence on cracking susceptibility nickel superalloys. Alloy ZhS32 has a high propensity for hot cracking when exposed to laser radiation. In this work, the study of the structural and phase features of ZhS32 alloy was carried out. A high tendency to form segregation of refractory elements and carbides in the intergranular areas was found. The features of the structure and phase composition of the material for different cladding parameters were studied. The main contribution of technological parameters to the formation of cracks is shown.

1. Introduction

Nickel superalloys combine properties such as high thermal and mechanical strength, creep resistance, and high hardness at high temperatures. Nickel-based heat resistant alloys are used to reduce life cycle costs and provide longer component life, mainly because they are less prone to stress cracking than other alloys. Because of these attractive properties, these alloys have a high demand and a wide range of applications, which have recently expanded considerably. Industries where they can be used include rocket engines, plant systems, land-based gas turbines, and equipment for the petroleum and chemical industries, and they include the most critical and highly stressed parts, such as blades for gas turbine engines (GTEs). For cast parts, such as gas turbine blades, directional solidification of alloys is used. In this case the crystals are oriented along the main stresses and the boundaries of weak grains remain unstressed [1,2]. Directional crystallization makes it possible to achieve the level of plasticity to the level of deformable alloys [3].

The structure of cast parts is characterized by significant chemical and structural heterogeneity [4]. Laser treatment of materials in the realization of promising technological processes of laser welding and laser cladding significantly reduces the segregation heterogeneity of nickel superalloys [5,6,7].

One of the important and demanded directions of work with GTE blades is carrying out their repair by various methods. Laser surfacing method during repair provides better formation of geometry and structure of the deposited area due to better controllability of thermal deposition and narrower zone of thermal influence in comparison with other methods [8,9]. However, laser repair of welds does not exclude weld metal cracking. The laser cladding process involves rapid heating and rapid cooling of the material, which to a greater extent can lead to the formation of microcracks. On the surface of the machined material, the energy of the powerful laser beam is absorbed by the metal, and then this energy is converted to thermal energy, where thermal stress is instantly formed. Furthermore, if the thermal stress exceeds the strength limit of the metal being processed, defects such as cracks form. Therefore, cracking remains one of the most basic problems in production for nickel superalloys, which is also due to the wide crystallization interval of multicomponent systems. However, laser cladding of nickel-based superalloys can produce minimal thermal stresses and minimal dilution factor with a narrow zone of thermal influence, which in turn can reduce the susceptibility to cracking.

The initiation and growth of cracks in nickel precipitation hardened alloys can occur for several reasons. Solidification cracks occur during alloy solidification and are associated with the presence of liquid films along the grain boundaries in the melting zone. Liquation cracks are observed in the heat-affected zone as a result of microsegregation of elements, or as a result of the formation of low-melting compounds during non-equilibrium heating [10,11,12]. The composition of the materials used has a great influence on the appearance of cracks and the nature of cracking. The susceptibility to cracking under strain aging is promoted by the high content of Ti and Al, since they contribute to the deposition of elements such as C, S, and B. The niobium in the composition has a negative effect on crack resistance due to the formation of niobium carbide. In addition, the accumulation of such elements as Nb, Ta, Mo, W, Ti in the intergranular area and the formation of their intermetallic compounds contribute to the embrittlement of products and the appearance of cracks. It is also worth noting that Nb additives will have a stronger effect on curing cracking than Ti additives [13,14,15,16].

The object of the research is a nickel superalloy ZhS32 of the Ni-Co-W-Al-Cr-Ta-Re alloying system. This alloy belongs to the non-weldable materials and has a high tendency to crack. The alloy under study consists of a $\gamma$ matrix based on a solid solution with a FCC structure, a hardening $\gamma$’ phase based on the intermetallic compound Ni${}_3$ (Al, Ti) with an FCC structure and carbides of the MC type (where M is Nb and Ta) and complex carbides of the M${}_{23}$C${}_6$ and M${}_6$C type (where M is W, Mo, Cr) [16,17,18].

Studies of the technological possibility of obtaining defect-free samples from Ni-Co-W-Al-Cr-Ta-Re alloys in works [19,20,21] showed the dependence of cracking on the parameters of the laser cladding process (laser power and scanning speed). In [21,22,23], the possibility of obtaining defect-free samples from nickel alloys of Ni-Co-Cr-W-Al alloying systems using induction heating and different scanning strategies was reported. This shows the fundamental possibility of selecting technological regimes for materials with low weldability.

The purpose of this study was to determine the causes of cracking during laser cladding on cast blade segments with directionally crystallized structure. This paper considers the influence of laser cladding parameters, on the formation of structure in the molten material and deposited layers, the nature of nucleation and development of defects such as cracks was investigated. Laser radiation power, scanning speed and powder feed rate are the variable parameters in this work. The effect of different scanning strategies on the formation of thin clad walls was also studied.

2. Materials and Methods

The studied samples were got by laser melting and laser cladding processes. The technological complex for laser cladding includes an industrial robot LRM-200iD_7L (Fanuc, Oshino-mura, Japan), an LS-3 Yb fiber laser (IPG Photonics, Oxford, MA, USA), with a maximum power of 3000 W, a welding head FLW D30 (IPG Photonics, Oxford, MA, USA), with focal lengths up to 500 mm, SO12 nozzle (SPbSMTU, St. Petersburg, Russia) and Metco Twin 150 powder feeder (Oerlikon, Freienbach, Switzerland), with a powder feed rate of 2 to 150 g/min.

Extensive experimental studies have been carried out including melting of work blade samples, single roll cladding and multi-layer single roll wall cladding. The research was carried out to study the processes of crystallization and defect formation at different parameters of the surfacing process, the shape of the melt bath and the crystallographic orientation of the grains.

Technological capabilities of the laser cladding machine allow experimental research to be carried out with a wide range of technological modes. However, it is known that the technological window for the processing of heat-resistant alloys by laser processing is in a narrow range. Based on works conducted with heat-resistant alloys and literature sources [19,20,21], the ranges of technological parameters of laser powder surfacing process and strategy of layer deposition which, according to the authors, are the most interesting to study have been determined.

Table 1. Cladding parameters.

Strategy Number	Type of Sample	Power, W	Scanning Speed, mm/s	Powder Feeding Speed, g/min
1	Single beads	200	3	2.4
	Thin wall	200	3	2.4
2	Single beads	200	9	7
	Thin wall	200	9	7
3	Single beads	600	9	7
	Thin wall	600	9	7

shows the parameters of laser cladding. Using the welding head FLW D30 at a focal distance of 200 mm, we obtain a laser beam of 0.3–0.5 mm diameter on the substrate. To achieve a beam diameter of 1 mm on the work surface, we lower the welding head to a focal distance of 150 mm and keep the laser beam diameter constant throughout the process for single bead and thin wall surfacing. High-purity argon was applied at a set flow rate of 20 L/min for local protection.

Single-wall surfacing was performed using three different strategies. Schematics of the strategies are shown in Figure 1. The first strategy was single-sided surfacing, i.e., with a fixed starting point for each layer. After each layer, a pause of 30 s was made to cool the weld pool. The second strategy was a one-sided build-up of layers, but without a pause for cooling. After applying the layer, the tool moved to the starting point and applied the next layer. The third strategy was reverse surfacing. After the build-up layer was applied, the tool began surfacing the next layer without moving to the starting point. There was no pause between layers. In addition, single rolls were cladded separately and laser melting was also considered to further analyze the influence of process parameters, such as scanning speed and laser power, on the crystallization process and on the occurrence of macro-defects. The parameters of laser cladding of thin walls were chosen according to the results of single roll formation and laser melting (Table 1).

The experiments were carried out on segments of gas-turbine engine (GTE) blades made by casting ZhS32 alloy. Powder of the same alloy is used for laser cladding, powder sizes ranging between 100 and 140 $\mathsf{\mu }$m. The morphology of the powder is shown in Figure 2, while the chemical compositions of the substrate and metal powders are shown in

Table 2. Chemical composition of powder and substrate material.

Element, wt.%	Ni	Ta	Re	Co	Nb	Mo	W	Cr	Al
Powder	59.37	4.49	3.76	9.48	1.64	1.09	9.34	5.12	5.71
Substrate	59.57	4.38	3.51	9.24	1.64	1.39	9.28	5.28	5.84

.

The structure of the samples was studied using an optical microscope Leica DMi8 (Leica Microsystems, Wetzlar, Germany). Metallographic etching was made in a mixture of acids HCl:HNO${}_3$ in ratio 3:1. Researches of microstructure and chemical elements distribution were made on scanning electron microscope Tescan Mira 3 (TESCAN, Brno, Czech Republic) using EDS analysis console Aztec Live Advanced Ultim Max 65 (Oxford Instruments NanoAnalysis, Wycombe, UK). Thin-wall samples were also studied using a system for analyzing back-scattered electron diffraction AZtecHKL Standard C-Nano (Oxford Instruments NanoAnalysis, Wycombe, UK), this made it possible to determine the crystallographic directions of grains in the samples structure and to analyze grains misorientation. X-ray diffraction measurements of thin-wall samples were carried out using a diffractometer D2 PHASER (Bruker AXS GmbH, Karlsruhe, Germany). The quantitative composition of segregations was determined by the HAADF-STEM and EDS method using a Thermo Scientific™ Talos F200X G2 S/TEM, 4-segment EDX (FEI Corporation, Hillsboro, OR, USA).

3. Results and Discussions

3.1. Influence of Laser Melting Parameters on the Appearance of Defects in a Sample

Cracks were found in samples from a ZhS32 alloy after laser melting and the characteristic macrostructure of the samples is shown in Figure 3. Changing the melting regimes has a noticeable effect on both the shape of the melt pool and the formation of cracks. Macrodefects were not detected during melting with low laser power and low scanning speed. An increase in the scanning speed does not lead to a change in the shape of the remelting zone, however, cracks are formed that pass along the axis of the melt pool from the line of fusion with the base metal to the surface of the pool (it is shown in Figure 3a,b). An increase in the radiation laser power leads to a transition from the thermally conductive processing mode to the mode with deep penetration; this is expressed by the formation of a mushroom-shaped melt pool. At the same time, the number and size of cracks increases. Structure consists of columnar dendrites and equiaxed crystals.

Microstructure studies show that melting actively occurs deep into the substrate along grain boundaries (Figure 4a,b). Carbide particles are present in the original structure of the metal (Figure 4c), they become stress concentrators and contribute to the appearance of cracks in the samples during melting. The results of studies by the EDS method are presented in Figure 5, the high content of niobium, tantalum and molybdenum in the composition of the particles can be seen. The size of detected carbide particles in the remelting zone is 2 to 8 µm, in the unmolten substrate the carbides size can be up to 15 µm.

The main reason for the formation of cracks is the presence of inclusions with a high content of niobium, tantalum and molybdenum. Inclusions based on carbides and carbonitrides have a negative effect and lead to cracking in nickel superalloys [24,25,26]. As a result of the studies, (Nb, Ta, Mo) C were found in the intergranular areas of the original blade structure. These inclusions become a source of internal stresses during crystallization and lead to the formation of cracks.

Furthermore, microcracks are formed because of the formation of the eutectic $\gamma$+$\gamma$’, which is clearly seen in Figure 6. The appearance of this type of microcracks depends on the melting parameters, since the higher the heat input, the more branched the microcrack becomes.

Therefore, the laser power and scanning speed affect both the crystallization process and the appearance of cracks after melting.

3.2. Influence of Laser Cladding Parameters on the Appearance of Defects during the Formation of Single Beads

Single beads are formed using a ZhS32 alloy metal powder alloy with a feeding rate stated in Table 1. The values of laser power and scanning speed correspond to the same values during melting. Two types of cracks were found in the samples: cracking along inclusions and pores, cracking in the volume of the deposited bead (Figure 7 and Figure 8). The nature of defects depends on the modes of laser cladding, as with melting samples. The mode in which crack formation does not occur during melting (200 W, 3 mm/s) leads to the formation of a defect-free macrostructure during bead cladding (Figure 8a). The formation of a small number of pores and cracks in the bead volume occurs when the scanning speed increases (Figure 7b and Figure 8b). The pores in the samples are solitary and up to 40 µm in size. Cause of cracks in the bead volume may be volumetric changes in the process of crystallization, which depend on the rate of crystallization of the alloy and on the selected parameters of cladding.

At high scanning speeds, an increase in laser power and heat input leads to the formation of numerous cracks and pores, as these parameters correlate with each other. Induced heat may also play a role in the formation of defects in the melt bath. A large root cavity can be caused by an unstable gas-vapour channel and its disintegration during machining (Figure 7c). Cracks are also treated after exposure to carbides in the structure of the original blade; it is also worth noting that in this case the sample has a large number of pores and inclusions compared to other strategies (Figure 8c).

Therefore, the laser power and scanning speed affect the crystallization process and the appearance of cracks in beads forming.

3.3. Influence of Laser Cladding Parameters on the Structure of Thin Walls

Thin walls are deposited using three scanning strategies. The parameters of the laser cladding process were limited by a lower power and scanning speed (Table 1), because the results of single beads cladding showed a high defectiveness of the samples when using high heat input. The resulting wall length is 50 mm, wall height 9–11 mm and the number of layers is 100–120.

Photos of the longitudinal sections of the walls are shown in Figure 9. Layer heights vary depending on the laser cladding modes.

Table 3. Layer height for different laser cladding strategies.

	Laser Cladding Strategies	Strategy 1	Strategy 2	Strategy 3
Power, W
200		120–145	140–165	185–220
600		330–370	390–440	480–520

lists the layer height ranges. The layer height tendency is to increase when moving from strategy 1 to strategy 3. This tendency is the same for both low and high heat inputs.

The greatest cracking is observed when using strategies 1 and 2 (cladding in one direction) at any laser power. At low laser power, cracking occurs from the bottom of the wall. The substrates were not preheated before growing, so this cracking occurs at a high temperature gradient. Using strategy 3 with a change in cladding direction reduces cracking; cracks are not detected at low laser power (200 W). Furthermore, the number of cracks is reduced at high laser power. The use of strategy 3 reduces the temperature gradient. Thus, the tendency to crack is reduced.

The microstructure of thin walls when using modes with low heat input is a fine-mesh structure with segregations of alloying elements along the cell boundaries (Figure 10). Segregations include tungsten, molybdenum, niobium, and rhenium, which are refractory metals and can precipitate along grain boundaries during crystallization (Figure 11). Carbides of composition NbC were also found. This distribution of elements is characteristic of the alloy under study, which has already been shown in the study of single beads.

Directional crystallization occurs and the structure of the samples is dendrites with high heat input (Figure 10). Elemental analysis of the structure sections (Figure 12) shows the precipitation of refractory metals (Ta, W, Mo, Re, Nb) in the interdendritic areas and inclusions of carbides. At low heat input, there was a markedly higher number of segregations than at high heat input. The composition of the segregations at low heat input increases the content of rhenium and tungsten. At high heat input, rhenium and tungsten dissolve into a solid solution ($\gamma$-phase).

XRD results for powder and thin walls are shown in Figure 13. The lattice parameters for the $\gamma$- and $\gamma$’- phases are close, so the diffraction peaks are superimposed on all the studied samples. However, there is a shift in the diffraction angle, which is associated with a change in the lattice parameter of the $\gamma$-phase depending on the heat input during laser cladding and the choice of cladding strategy. Strategy 1 results in the expending of the peaks; this is associated with the formation of smaller grains because of a higher temperature gradient than with other strategies. With a high heat input, strategy 3 allows directed dendritic growth. This is confirmed by the high intensity of the peak from the (200) plane in the direction in which dendrites grow.

The peaks of carbide phases of the MC type (where M is Nb, Ta, Mo) were also found in the diffractograms. Diffraction peaks of intermetallic compounds of refractory metals (Nb, Re, Mo, W, Ta) are at diffraction angles of 39–42 2$\theta$, and this confirms intermetallic compounds of refractory metals detected by the EDS method. Low heat input leads to the appearance of a significant number of carbides and intermetallic compounds of refractory metals, and high heat input leads to a decrease in the content of these phases.

The choice of strategy also affects the phase composition. This is because of the difference in temperature gradient, which leads to different rates of crystallization and phase formation.

Quantitative determination of the segregations composition was carried out using HAADF-STEM; the results are shown in Figure 14 and in

Table 4. Quantitative composition of segregations.

#	Al	Cr	Co	Ni	Nb	Mo	Ta	W	Re
1	2.98	3.05	4.77	18.58	0.98	2.26	24.68	21.87	20.84
2	3.42	1.16	2.50	18.98	18.07	1.48	52.67	1.26	0.45
3	2.24	1.34	2.25	13.51	30.09	2.43	42.90	4.67	0.58
4	2.63	4.84	6.50	30.56	6.47	6.56	19.14	17.35	5.94
matrix	4.49	4.99	9.12	57.74	1.41	1.21	12.46	5.17	3.40

.

Segregations of Nb, Mo, W, Ta, and Re elements (points 1 and 4) occur because of their incomplete dissolution during laser cladding and the precipitation of intermetallic compounds along the grain boundaries. The high heat input reduces the refractory metal content of the intermetallic segregations. Re and W dissolve into a $\gamma$-phase, as the element distribution map in Figure 12 shows. Segregations with a high content of Nb and Ta (points 2 and 3) may be mixed carbides, but the presence of carbon in these particles was not detected during the studies, which may be because of a low signal compared to the matrix elements. Carbide particle sizes up to 1.0 µm in samples with low heat input and up to 0.7 µm in samples with high heat input. The solidification rate is higher in high heat input conditions, so the growth of carbide particles stops earlier.

Segregations of refractory elements were also observed in [27,28], where the relationship between cracking and segregations was indicated. The formation of a significant number of segregations leads to embrittlement of the material and a decrease in crack resistance.

Grain misorientation maps are shown in Figure 15. Cracking occurs along the boundaries of differently oriented grains with high heat input. Powder particles are not completely melted; rounded grains are visible at low heat input in Figure 15. Cracking in this case occurs along the boundaries of the powder particles. Strategy 3, with a change in the cladding direction, reduces the tendency to cracking, however, cracking is observed along the boundaries of misoriented grains at high laser power.

The effect of scanning strategy on cracking propensity has been reported in many studies on nickel superalloys. It is possible to select a scanning strategy that reduces the cracking of samples [27,28]. The results of the study showed the possibility of further choice of scanning strategy to get defect-free samples for the alloy under study.

The main role in the formation of cracks is played by the parameters of the laser cladding process. Changing the scanning strategy of the cladding does not lead to a change in the nature of cracking, but may reduce the tendency to cracking. The presence of segregations is detected in any mode and scanning strategy. The results showed that the temperature gradient has an effect on cracking.

The microstructure of the samples (Figure 10) has the same patterns for different strategies. Carbides and segregations of refractory metal intermetallides are formed in the samples. The number of carbide particles and segregations is greater at low heat input (about 6%) than at high heat input (about 3%). Rhenium and tungsten in intermetallic composition are high at low heat input. These elements are dissolved in the solid solution of nickel ($\gamma$-phase) at high heat input.

Strategy 1 uses a pause of 30 s between laser cladding layers. As a result, crack formation starts from the substrate (Figure 9). EBSD maps (Figure 15) show that cracks propagate along fine grains at low heat input. As the SEM results show, at low heat input, a large number of segregations are formed, and cracks form along the segregations. At high temperature, cracks form along the grain boundaries with misorientation. In this case, crack formation was influenced by the stress state of the solidifying structure. This type of cracking is characteristic of nickel alloys [10,12,13].

The difference between strategy 2 and strategy 1 is the absence of a pause for cooling of the cladding roll during laser cladding. Therefore, the temperature gradient is lower, but crack formation occurred, and the cracking pattern is the same as in strategy 1. The SEM (Figure 10) and XRD images (Figure 13) show that the content of carbide phases and intermetallic refractory metal segregations is higher in strategy 2. Thus, the use of laser cladding without a pause does not eliminate cracking.

In strategy 3 the temperature gradient is lower than in strategies 1 and 2. The formation of cracks at low heat input does not occur. At the same time, the formation of a significant amount of segregations indicates the need to increase the heat input in order to achieve a more homogeneous composition. The tendency to reduce cracking remains at high heat input.

4. Conclusions

Analysis of the results allowed us to identify the causes of defects and the nature of their occurrence, depending on the technological parameters of laser processing:

• the appearance of cracks on the cast substrate occurs mainly by the liquation mechanism because of the melting of the fusible carbides MC at the grain boundaries;
• cracking in melted metal and single beads occurs during crystallization solidification and is associated with grain misorientation, which depends on the shape of the melt pool and the process parameters;
• cracking of thin walls occurs along the grain boundaries, increasing the heat input leads to increased cracking, the choice of scanning strategy can lead to reduced cracking.

The main cause of cracking in ZhS32 material is the high temperature gradient during laser cladding. A reduction in the tendency to crack is observed when scanning strategies that reduce the temperature gradient are used.

The use of low heat input leads to the formation of carbides and segregations of refractory metals. Segregation of elements causes embrittlement and crack development, but is not the source of crack formation. At high heat input, carbide growth is reduced, and the composition of the segregations changes as some of the elements dissolve in the $\gamma$-phase.