Assessing the Printability of Rene 65 Powder for Repairing Degraded GTD 111 Gas Turbine Blades Using L-DED and L-PBF

Abstract

Restoring components in the hot gas path of turbine engines after service-induced degradation is crucial for economic efficiency. This study investigates the printability of Rene 65 powder on a degraded first-stage turbine blade using two additive manufacturing techniques: Laser Powder Bed Fusion (L-PBF) and Laser Powder Directed Energy Deposition (L-DED). Deposited material was evaluated using optical microscopy (OM), scanning electron microscopy (SEM), and Electron Backscatter Diffraction (EBSD) to characterize its crystallographic texture, while microhardness testing provided insight into its mechanical properties. Our results show that L-PBF excels at replicating intricate features, such as small cooling holes, and produces a highly texturized microstructure oriented parallel to <001> under optimal parameters (80 W, 400 mm/s, unidirectional scanning), although at a slower pace. In contrast, L-DED offers a versatile, rapid, and cost-effective method for repairing medium to large parts, yielding an equiaxed microstructure and higher as-printed hardness—approaching GTD 111 values due to an aging effect from high heat input. Both processes effectively restored the dimensional integrity of degraded blade tips, paving the way for more sustainable and economical maintenance strategies in the aerospace industry.

1. Introduction

Nickel-based alloys are widely used in the hot path components of gas turbines. GTD 111, in particular, is a commonly used material for manufacturing turbine blades in thermoelectric power plants [1]. Due to the combination of nickel alloys and complex geometries, turbine blades are highly expensive. The replacement cost of a single first-stage gas turbine blade can exceed USD 15,000 [2]. As a result, remanufacturing degraded components is becoming increasingly popular, as it significantly reduces the total cost of replacement, as well as energy consumption and downtime [3,4].

The turbine serves as the core component of a thermoelectric plant, comprising various mechanical elements engineered to endure extreme operational conditions. These include elevated temperatures (~1000 °C), high corrosion pressure (~130 psi), gas outlet velocities reaching 2.5 Mach, and a rotational speed of 3600 rpm. Gas turbines are essential in converting thermal energy from combustion into mechanical power, which is subsequently transformed into electrical energy via the Brayton cycle [1]. This thermodynamic cycle consists of compression, combustion, and expansion stages, with the expansion phase further divided into three sub-stages. Specifically, the initial expansion sub-stage of a GE7FA gas turbine features 92 blades fabricated from the nickel-based alloy GTD 111.

Figure 1 presents typical defects observed in deteriorated turbine blades. Contact surfaces, which experience both oscillatory vibrations and centrifugal forces, are highly vulnerable to erosive degradation. Furthermore, exposure to a corrosive environment significantly contributes to oxidation, corrosion, and sulfidation, which can result in cracking and material spalling. Among the various wear mechanisms, fretting wear has been identified as the predominant factor affecting turbine blade longevity [5]. To counteract these wear mechanisms, several studies have demonstrated successful outcomes in repairing turbine blade tips using additive manufacturing techniques [6,7].

The reliability of joined components depends on the absence of defects such as cracks or fissures, making it essential that manufacturing and repair processes ensure a defect-free outcome. However, nickel-based alloys present significant weldability challenges, including solidification cracking, liquation cracking, and strain-age cracking. Traditional repair techniques for turbine blades, such as Plasma Arc Welding (PAW), Gas Tungsten Arc Welding (GTAW), and manual brazing, required highly skilled labor and often resulted in metallurgical compromises due to the use of dissimilar filler alloys, such as Inconel 625 (IN 625) [8].

Recent advancements in additive manufacturing (AM) have accelerated its use in energy, aerospace, and medical sectors by enabling precise layer-by-layer deposition of complex geometries without costly tooling, thus, extending component lifespan and reducing repair costs and environmental impact [9]. Laser-based AM technologies such as L-PBF and L-DED also offer lower heat input and diminished susceptibility to metallurgical issues compared to traditional welding, which is particularly beneficial for complex, lightweight designs. However, the iterative melting and cooling inherent in AM can generate significant residual stresses, leading to microcracking, warpage, and dimensional inaccuracies. Furthermore, phenomena like “balling”—where molten metal coalesces into spherical droplets rather than smooth tracks—can undermine surface quality and mechanical integrity.

As explained in [10], meticulous parameters optimization (e.g., laser power, scan speed, layer thickness) and careful consideration of support structures and part orientation are essential for controlling residual stresses, distortions, and crystallographic orientation. To address these challenges, researchers have refined process variables (e.g., energy input, hatch spacing, preheating) to lower thermal gradients and residual stresses; post-processing measures like heat treatment, hot isostatic pressing, and surface finishing also help mitigate defects and improve mechanical properties. In addition, as reviewed in [11], part designs must account for potential warping and stress concentrations by using optimized supports and strategic orientations, thereby guiding both practitioners and researchers toward more reliable and repeatable AM processes.

This study investigates the feasibility of repairing GTD 111 turbine blades using L-PBF and L-DED with Rene 65 powder, without preheating, focusing on dimensional accuracy, deposit integrity, and the effect of key process parameters on the microstructure of LP-PBF and L-DED deposits. The central thesis is that advanced additive manufacturing methods can be leveraged to restore degraded turbine blades with minimal distortion, robust microstructures, and competitive mechanical properties, ultimately reducing costs and extending component service life. Accordingly, the primary aim of this research is to (1) evaluate how specific process parameters (e.g., laser power, scan speed) influence the dimensional accuracy and metallurgical quality of as-deposited materials, (2) characterize the resulting microstructures and microhardness profiles, and (3) compare the microhardness and suitability of L-PBF versus L-DED for repairing critical turbine components.

2. Materials and Methods

Rene 65 powder (15–45 μm PSD), produced via gas atomization, was used in both AM processes. This precipitation-hardenable nickel-based superalloy is strengthened by gamma prime [12]. Experiments were conducted on degraded GTD 111 blades, a cast, precipitation-hardenable alloy derived from Rene 80, commonly used in gas turbines for its directionally solidified microstructure [13,14]. The powder and substrate’s chemical compositions were analyzed using optical emission spectroscopy (OES) with an Ametex-Spectromax device, as shown in

Table 1. Chemical composition of GTD 111 substrate and Rene 65 powder (wt%).

Alloy	Ni	C	Al	B	Co	Cr	Fe	Mo	Nb	Ti	W	Zr	Ta
Rene 65	Bal.	---	1.8	0.01	12.3	15.6	0.8	3.2	0.7	3.5	3.7	0.03	---
GTD 111	Bal.	0.1	3.4	---	10.8	14.1	---	1.6	---	5.1	4.4	---	3.2

. The high Ti+Al content promotes gamma prime formation but increases cracking susceptibility during fusion processes [15]. René 65 was selected over René 80 due to its superior mechanical properties, higher temperature capability, and improved oxidation resistance, making it more suitable for the intended application.

Figure 2 presents details of the materials used in the experiments. Figure 2A shows a micrograph of Rene 65 powder cross-sections, confirming the absence of porosity. Figure 2B illustrates the powder morphology, predominantly spherical, with some elongated particles (yellow arrows), satellite formations (white arrows), and occasional agglomerations (orange arrows), which can impact flowability, particularly in the L-DED process.

The base material was extracted from two first-stage GE7FA gas turbine blades, as shown in Figure 2C. Coupons were extracted via wire Electrical Discharge Machining (EDM), perpendicular to the longest blade axis, approximately in the <100> direction. Blade 1 operated for 30,000 h (about half its lifespan), while Blade 2 accumulated 22,000 h. All samples underwent heat treatment at 1120 °C for two hours in an argon atmosphere for stress relief and to homogenize the microstructure, preventing cracking during additive manufacturing (AM). Figure 2D depicts a 3D microstructural reconstruction of GTD 111 after heat treatment; metallographic preparation was carried out using Marble’s etchant, revealing columnar dendritic growth in the <001> direction, which enhances creep and fatigue resistance compared to equiaxed solidification [15]. Microhardness testing was performed using a Leco LM 100AT (LECO Corporation, St. Joseph, MI, USA) device with a load of 0.3 kg according to ASTM E384, measuring the hardness variation across the substrate and the L-PBF and L-DED deposits.

2.1. L-DED

L-DED experiments were conducted using a TrumpF TruLaser Cell 3000 (TRUMPF Inc., Farmington, CT, USA), a five-axis CNC system with a solid-state laser (1030 nm wavelength, 1 mm spot size). The powder was delivered coaxially through a three-jet nozzle at a 12 mm working distance and 12% amplitude, with argon used as both carrier and shielding gas. The nozzle remained perpendicular to the sample surface throughout all trials.

Figure 3 outlines the turbine blade reconstruction process using L-DED and L-PBF, structured into four stages: previous characterization, exploratory phase, experimental phase, and rebuild. The initial characterization included a microstructural analysis of all AM samples. The exploratory phase involved single-bead deposition, followed by geometric and visual inspections before advancing to the main experimental stage.

In L-DED, a 0.7 mm hatch spacing was used, with power and scanning speed combinations optimized by analyzing deposition characteristics such as discontinuities, dilution, layer slope, height, and width. Multilayer depositions varied power, speed, and layer rotation. The best parameters were then applied to rebuild a GTD 111 sample wall using each AM process. A full factorial 2³ design with replication was used, resulting in 16 randomly tested conditions.

Table 2. Parameters used in the L-DED experimental stage.

Power (W)	Scanning Speed (mm/s)	Rotation (°)
400	7.5, 9.2	0, 180
300	7.5, 9.2	0, 180

presents the developed parameters, and Figure 4 illustrates the schematic paths used in the experimental stage.

2.2. L-PBF

Laser Powder Bed Fusion (L-PBF) experiments were conducted using a Concept Laser Mlab (GE Additive company, Lichtenfels, Germany) machine with a 100 W fiber laser, a maximum scanning speed of 7 m/s, and a fixed 25 μm layer thickness in an argon-controlled atmosphere. The build volume of the machine is 90 mm × 90 mm × 80 mm. The methodology follows Figure 3, and the experimental parameters are listed in

Table 3. Parameters used in the L-PBF experimental stage.

Power (W)	Scanning Speed (mm/s)	Rotation (°)
50	200, 400	0, 90
80	200, 400	0, 90

. Layer rotation was alternated between 0° and 90°.

A complete factorial 2³ design with replication was used for data analysis, varying power, scanning speed, and layer rotation. The layer thickness and hatch spacing were set at 25 μm and 80 μm, respectively. Table 3 details the parameters used in the L-PBF experimental stage. Each combination was tested twice on the GTD 111 substrate.

3. Results and Discussion

3.1. Characterization of Degraded, Post-Service Blades

Microstructural characterization was conducted at the center of the blade tips on degraded samples from both blades, revealing similar features across the components.

Figure 5A presents an unetched, polished GTD111 blade sample after heat treatment, where typical casting defects such as carbides (C) and pores (P) are visible. In Figure 5B, a high-magnification view highlights a blocky primary carbide alongside noticeable microstructural defects, including compositional inhomogeneities within the eutectic γ/γ’ phase and additional matrix pores.

Figure 5C illustrates the coexistence of carbides, eutectic, and eta phases. The eta phase (Ni₃Ti, D024 structure) appears along grain boundaries as platelets formed from MC decomposition into M₂₃C₆, usually in proximity to pores and carbides. These features predominantly reside in the interdendritic regions.

Finally, Figure 5D shows a mix of blocky and script-like carbide morphologies, a transformation attributed to increased carbon content and a higher cooling rate. Hardness measurements for both blades post-service were consistently around 400 HV.

Degradation alters the microstructure from its original state of cuboidal and spherical γ’ precipitates [16]. Figure 6A,B reveals eutectic islands and carbides at grain boundaries, creating regions of compromised structural integrity. Figure 6C shows a depletion of γ’ at the γ/γ’ interface, while Figure 6D highlights significant changes in the size, shape, and volume of γ’ precipitates, with primary and secondary γ’ coexisting.

Figure 7 presents the EDS analysis results. In Figure 7A, a chemical mapping reveals an eutectic phase enriched in Ni and Ti but depleted in Cr—consistent with a Ni₃Ti composition—along with MC-type carbides predominantly containing Ti and W. Figure 7B shows script carbides (“Chinese script”) with elongated arms along grain boundaries, enriched in Ti and Ta, which may facilitate crack nucleation and impair mechanical properties [17,18]. Blade 1 exhibits an average area fraction of 33.93% ± 1.56%, while Blade 2 shows 30.91% ± 1.13%. The average γ’ precipitate size is similar between the two—0.32 μm for Blade 1 and 0.35 μm for Blade 2. The lower area fraction and slightly larger precipitate size in Blade 2 are attributed to its longer service duration.

3.2. L-DED Manufacturing

3.2.1. Experimental Stage

Figure 8 shows 15 mm × 3 mm × 3 mm deposits printed onto a blade coupon manufactured in GTD 111 superalloy according to the parameters described in Table 2. In general, no detachments or surface discontinuities were evidenced by visual inspection. The cross-section at the central part of each deposit was analyzed to determine the porosity of the samples by image analysis. Porosity measurements indicate that samples with a 180° rotation between layers exhibit lower porosity, with values generally below 0.35%.

The deposits were made up of four layers built with four beads as seen in the micrographs included in Figure 9A,B, which correspond to deposits with different layer orientations applied with 400 W and 9.2 mm/s. Figure 9C shows the 300 W-7.5 mm/s-0° deposit, where no significant changes in the microstructure are evident.

Figure 10 summarizes the microhardness results, which reveal a consistent trend among the specimens. Hardness increases markedly at the Heat Affected Zone (HAZ) before stabilizing in the as-printed area, with L-DED samples reaching around 400 HV—comparable to the GTD 111 substrate (Figure 10A). A hardness map for the 300 W-7.5 mm/s-180° specimen (Figure 10B) shows a uniform distribution between the substrate and the printed metal. Additionally, subsequent layer deposition introduces an aging effect, incrementally increasing hardness and homogenizing properties throughout the deposit in situ. These hardness values are consistent with those reported in previous studies [14,19] following heat treatment.

Figure 11 shows the EBSD inverse pole figure (IPF) map from the cross-section of an as-printed sample aligned with the build direction. Using 400 W and 7.5 mm/s with printing orientations from 0° to 180°, the grains display equiaxed dendritic growth with no specific crystallographic alignment. The 2 mm × 1.5 mm area reveals a random texture with similar grain morphologies and rotation angles across samples, with no discernible differences.

A Design of Experiment (DOE) analysis using analysis of variance (ANOVA) evaluated the impact of power, scanning speed, and layer rotation on porosity and hardness at a 95% confidence level (α = 0.05). Although individual factors did not yield statistically significant changes in hardness (p > 0.05), a clear trend was observed: higher power and scanning speed increased hardness, and a 180° interlayer rotation further elevated hardness values. The model’s reliability is confirmed by an adjusted R² of 97.6%.

3.2.2. Wall-Tip Repair

After finalizing the DOE, a setting of 300 W power and 7.5 mm/s scanning speed was selected to restore a 3 mm high wall, with the restoration dimensions mimicking those of a squealer tip. These tests confirmed that the L-DED process, when controlled numerically, is a viable method for repairing GTD 111 blades (see Figure 12).

3.3. L-PBF Manufacturing

3.3.1. Experimental Stage

Figure 13 shows the experimental results, where sixteen 5 mm × 5 mm × 2 mm parallelepiped deposits were made following the parameters in Table 3 (see Figure 13A). All deposits adhered well to the substrate except for sample 50 W-400 mm/s-0° (indicated by the white arrow), which nearly detached due to the lowest energy density. Each sample was then sectioned transversely, prepared metallographically, and analyzed for porosity using ImageJ software 1.54p (Figure 13B).

The porosity analysis in the L-PBF parts demonstrated that optimizing process parameters plays a crucial role in minimizing porosity. In particular, the parameter combinations 50 W-400 mm/s-90° and 80 W-400 mm/s-0° produced the densest as-printed parts, highlighting the importance of fine-tuning these settings to enhance material quality.

Optical microscopy focused on low-porosity printed samples is critical since porosity can affect strength and fatigue performance [20]. The microstructure of sample printing with 50 W-400 mm/s-90° of layer rotation, examined at multiple magnifications, demonstrates how processing parameters shape AM parts (Figure 14). Specifically, a columnar dendritic structure develops along the build direction, causing anisotropy on the mechanical properties such as tension or creep [21], while temperature gradients and growth rates yield varied solidification patterns, including both columnar and equiaxed grains [22].

Hardness measurements along the build direction were performed only on dense samples (Figure 15), excluding those with significant porosity. The results show that lower energy densities decrease the hardness of deposits produced with Rene 65 powder, whereas samples processed at 80 W exhibit higher hardness values due to an aging effect from successive layers. Additionally, regions containing defects demonstrate a significant reduction in hardness. Extended results are described in detail in [23].

Figure 16 presents the crystallographic texture of as-printed samples using EBSD inverse pole figure (IPF) maps. In Figure 16A, a SEM image of the sample is shown, while Figure 16B displays the corresponding IPF grain map from a transverse section. The grain texture shows alignment with <001>, a trend attributed to the vertical heat flux during processing. Additionally, Figure 16B highlights the interface between the GTD 111 substrate and the Rene 65 deposit, with the directionally solidified microstructure of GTD 111 visible at the top.

The DOE analysis using ANOVA evaluated the effects of power, scanning speed, and layer rotation on porosity and hardness at a 95% confidence level (p = 0.05). Results showed that both power and scanning speed, along with their interaction, significantly affected porosity, with the model achieving an R² of 86%. However, due to high porosity, the hardness variable was unstable—only three samples exhibited consistent hardness (see Figure 15)—preventing further statistical analysis.

3.3.2. Wall-Tip Repair

After extensive parameter optimization and analysis of the microstructure, texture, microhardness, and deposit integrity, a combination of 80 W, 400 mm/s, and 0°-layer rotation was selected. These settings were used to 3D print a wall simulating the repair of a turbine blade squealer tip. The wall, shown in Figure 17, measures 3 mm in height and 1.8 mm in thickness, replicating actual repair conditions.

4. Conclusions

This study demonstrates that additive manufacturing can effectively restore the dimensions of first-stage tip gas turbine blades by depositing precipitation-hardenable powders. The research successfully rebuilt the squealer tip using Rene 65 through L-DED and L-PBF processes without pre-heating. While L-PBF offers superior resolution and dimensional stability—yielding a highly texturized microstructure oriented parallel to <001> with optimal parameters of 80 W, 400 mm/s, and unidirectional scanning—it operates at a slower pace but reduces the need for extensive post-machining.

In contrast, L-DED is celebrated for its versatility, speed, and cost-effectiveness, making it ideal for repairing medium to large parts with minimal heat impact on the substrate, despite requiring parameter optimization for directional solidification. Moreover, L-DED tends to produce an equiaxed microstructure and higher as-printed hardness, approaching GTD 111 values due to an aging effect from high heat input. Both processes proved effective for repairing squealer tips, with L-PBF excelling in capturing fine details such as cooling holes.

The outcomes of this study provide a clear repair pathway for hot-path turbine components using additive manufacturing processes. Future work could focus on optimizing the L-DED process parameters, particularly controlling solidification rates to achieve consistent crystallographic orientations and ideal columnar grain structures. Such optimization should be performed cautiously due to the susceptibility to solidification cracking. Additionally, modern L-PBF machines equipped with large build chambers offer the opportunity to fabricate entire blades, enabling large-scale studies.