Solidification Analysis of a High Cr-Ni Heat-Resistant Alloy: Microstructure and Hardness Evaluation

Abstract

Ni-Cr-based alloys are known for their high resistances to corrosive environments, high temperatures, and stresses, finding wide application in furnace systems in petrochemical plants. However, very few is known about their solidification features and how they can impact processing and properties. This study investigated the alloy solidification, obtaining a variety of as-cast microstructures in the high-Cr 28Cr48Ni5W3Co14Fe alloy. It presents results of optical, hardness, and CALPHAD analyses in a straightforward examination of solidification according to the application in the cast condition. The directional solidification setup allowed to assess a variety of samples solidified at different cooling rates, and showing different microstructural coarsening. The analyses determined secondary dendritic arm spacings (SDAS) ranging from 16 μm to 120 mm. It was determined the formation of the FCC+M₂₃C₆ quasi-eutectic constituent filling the interdendritic. An effective correlation was chosen to compute the corresponding cooling rates. Finally, an in-depth applied analysis of tube-support components for pyrolysis through cast simulation revealed the SDAS variation to be a critical feature in pre-programming casting designs. These findings play a fundamental role in advancing cast operations of heat-resistant alloy components, enhancing both quality and processing setups.

Keywords:
Heat resistant alloy; SDAS; microstructure; hardness

1. Introduction

Advancements in casting technologies for nickel-based heat-resistant (refractory) alloy as-cast products necessitate a deeper understanding of their solidification behavior. In the petrochemical sector, a range of products may be manufactured by foundry, mainly using conventional casting. Parts used in these environments may be exposed to extreme temperatures and heavy loads¹,². Therefore, alloys with high Cr and Ni content alloys are widely used as refractory materials in petrochemical plants under high temperatures (from 1010 °C to 1150 °C). Creep resistance and long service lifetimes until fracture are factors for which these alloys are chosen for a safe design.

Despite the development of numerous cast alloys in recent years³, components manufactured through conventional casting, particularly tube supports, large-sized parts, and those exhibiting significant thickness variation between sections, continue to pose critical challenges in the manufacturing process. According to Tillack and Guthrie⁴ fabricability is an important selection criterion applicable to the heat-resistant alloys for petrochemical plants. Avoiding defects, and controlling cooling rate and microstructure can be particularly important to reduce costs, offering projects in which solidification is well controlled. As such, products with higher quality may be attained. In this way, there is a significant opportunity to improve the quality control of the process and reduce defects in alloys manufactured with a better understanding of solidification and the determination of correlations involving solidification thermal parameters, microstructure coarsening parameters (SDAS), and mechanical properties. The present contribution focuses on this issue.

Early analysis of solidification of Ni-based alloys was highly concentrated on superalloys and low Cr Ni-based alloys. Therefore, mathematical analysis of solidification in multi-component refractory systems remains absent, which requires experimental validation of the solidification process parameters. Manufacture of components from these alloys can be difficult due to the low ductility of the heat-resistant alloys at room temperature. It can be difficult to remove and repair casting defects because cracks tend to propagate due to the brittleness of the alloy at room temperature. Therefore, it is important to control the scale of the dendritic microstructure in order to identify the critical spacing and minimize crack initiation and propagation during the solidification and repair stages.

Few works in this field comprise the problem of controlling solidification cooling rate based on SDAS. Tinoco and Fredriksson⁵ observed that SDAS tended to decrease as the cooling rate was increased for a modified IN625 alloy (22Cr63Ni9Mo3Nb). After measuring much higher cooling rates than Tinoco and Fredriksson⁵, Wu et al.⁶ established a relationship between average cooling rate and powder size (after converted to SDAS) in a Ni-Cr-Co alloy, having the following composition (wt.%): Cr 12.92, Co 20.83, Mo 2.64,W 3.85, Al 3.57, Ti 3.53, Nb 1.51, Ta 1.65, C 0.048, 0.027, Zr 0.043, Hf 0.2, Ni–Base. The cooling rate range of the powders was 1.8×10³–3.6×10⁴ K.s^-1.

This study aims to examine the influence of various process conditions on the SDAS of the 28Cr48Ni5W3Co14Fe alloy. This type of high-chromium, nickel-based alloy, largely employed in cast operations, requires thorough examination of its solidification to enhance design and control over its properties. A valuable approach has been employed to assess cooling rates. In addition to exploring microstructure and solidification kinetics, a practical challenge was faced through the casting simulation of tube supports to investigate potential microstructural drawbacks during manufacturing. Hardness profiles were utilized to strengthen the discussions.

2. Materials and Methods

The 28Cr48Ni5W3Co14Fe alloy was processed by directional solidification. The material used was the same as that used in parts of a pipe support component and was provided for this study. The following composition in wt. % was determined using a Q4 Tasman Bruker spectrometer, as shown in Table 1.

Table 1
Chemical composition determined in wt. %. of the 28Cr48Ni5W3Co14Fe alloy.

C	Cr	Ni	W	Co	Si	Mn	Fe
0.40	28.0	48.0	4.6	2.8	1.4	0.6	Bal.

The alloy was cut into small pieces that were melted in a Si carbide crucible using an induction furnace (Power-trak 50-30 R model, Inductotherm VIP, Rancocas, NJ, US) and poured into a preheated mold (at 880 ^oC) using a directional solidification (DS) device. Pre-heating is typically used in the DS system to prevent the immediate solidification of the molten alloy after pouring. This also facilitates the heating process of the molten alloy to reach the desired level of superheating. The 304 stainless steel mold dimensions were 60 mm inner diameter, 160 mm height, and 5 mm thickness. The DS equipment was manufactured by Fortelab Ind. de Fornos Elétricos, São Carlos, Brazil. An overheating temperature of approximately 1380 °C (5% above liquidus temperature) was employed, while the alloy was poured with the water-cooled system of the DS equipment already activated. The controlled water flow was implemented to trigger forced cooling from the bottom, persisting until the ingot achieved complete solidification. This technique produces a range of cooling rates whose values decrease as the distance from the base of the ingot in contact with the surface cooled by the water flow increases.

Samples at the heights 5 mm, 10 mm, 15 mm, 20 mm, 25 mm, 45 mm, 70 mm, and 90 mm were polished and etched with an Agua Regia solution (20 mL of HNO₃ + 60 mL of HCl) for 30 s to reveal the dendritic arrangements. In order to register the optical images of the samples, a microscope (BX14 M-LED, Olympus, Japan) was used. The intercept method was used to measure the microstructural spacing values along the referred sections⁷. For each tested sample, at least 40 SDAS measurements were taken. As macrograph was characterized by very refined equiaxed grains while both longitudinal and transverse sections were examined in order to determine the dendritic coarsening.

Figure 1 shows the division of the ingot with the parts identified for each type of microstructural analysis in the present investigation.

Figure 1
Experimental scheme for microstructural analysis in the 28Cr48Ni5W3Co14Fe alloy ingot. T is transverse and L is longitudinal.

For indentation experiments, specimens were machined and surfaces finished to a level comparable to sanding with 1200 grit to ensure flatness and prevent scratches or grooves. Brinell (HB) tests were carried out using a BK 300b model, Wolpert. A minimum of five measurements were tested for each section of interest. The load used was 3,000 N.

CALPHAD computations were performed using Scheil and equilibrium models with the TCHEA7 database. The software used for such computations has been the Thermo-Calc (Stockholm, Sweden). Cast simulations were developed using the Magmasoft® ((MAGMA Giessereitechnologie GmbH, Aachen, Germany)) software designed explicitly for casting process simulations. The processing map of a tube-support structure for pyrolysis was determined by employing the experimental SDAS and HB equations. In other words, the experimental equations for SDAS and HB were implemented in Magmasoft simulation so that both SDAS and HB evolutions could be estimated in a real component.

3. Results and Discussion

Several optical microstructures were recorded. However, only some representative ones are shown in Figures 2 and 3. Focus here is to analyze the massive variation in the scale of the dendritic microstructure at different heights in the casting: 5 mm, 10 mm, 15 mm, 20 mm, 25 mm, 45 mm, 70 mm, and 90 mm. Microstructures found at intermediate positions, in relation to those observed in Figures 2 and 3, depict intermediate coarsening. Moreover, extreme positions, P (5 mm and 90 mm), as those in Figure 2 already exhibit all dendritic-scale changes in the ingot. It can be seen that the SDAS of each section varied considerably from the bottom to the top of the casting. This means that the 28Cr48Ni5W3Co14Fe alloy is very sensitive to the solidification conditions during manufacturing. Micrographs like those allowed the coarsening evolution to be established for this alloy. Based on the alloy composition as well as on the solidification conditions, the interdendritic (dark) areas are probably constituted by M₇C₃ and M₂₃C₆ carbides⁸. The morphology of these carbides (occupying the dendrite boundaries) can be seen in the coarser particles in Figure 3.

Figure 2
Representative optical microstructures at different positions, P, along the directionally solidified body of the 28Cr48Ni5W3Co14Fe alloy.

Figure 3
High magnification optical images at different positions, P, along the directionally solidified body of the 28Cr48Ni5W3Co14Fe alloy.

The occurrence of primary carbides has been discussed in for many years⁹-¹³. Most of the works reported the dark interdendritic carbide as being the M₇C₃. M₂₃C₆ has been mostly reported to form during aging⁹-¹³. To preliminarily identify this phase, thermodynamic computations can be utilized, as will be discussed later.

The solidification sequences depicted in Figure 4, described by the Scheil curve, is a possible method for representing as-cast processes under non-equilibrium conditions. It also characterizes the directional solidification experiment by considering solute rejection at the solidification interface and the absence of diffusion back into the solid state. It can be observed that the 28Cr48Ni5W3Co14Fe alloy begins to solidify slightly above 1345 °C with the formation of the FCC dendrites, followed by the precipitation of M₂₃C₆-type carbides starting near 1330 °C. Therefore, the alloy of interest probably has the interdendritic regions filled with the M₂₃C₆. Comparing equilibrium and Scheil profiles gives notion of the extension of the solidification interval, attaining approximately 30 °C under non-equilibrium growth. The Sigma (s) phase was not predicted to form under Scheil modeling. However, it does form below 900 °C in the equilibrium state. This phase has not been identified in the high-magnification optical images in Figure 3.

Figure 4
Sequence of phase precipitation from the liquid considering the 28Cr48Ni5W3Co14Fe alloy under (a) non-equilibrium and (b) equilibrium solidification.

Although some authors proposed the nucleation of primary carbides⁹-¹⁴, in the present case, only secondary ones have formed, in an almost eutectic fashion. This is because the solidification interval is very narrow in the equilibrium diagram (i.e., ~10^oC), as shown in Figure 4b. Moreover, an alternate morphology of phases indicates some eutectic type reaction forming the FCC+M23C6 morphology, as seen in Figure 3b. The description of the M23C6-D84 in the Thermo-Calc databases for the elements employed remains: (Fe,Cr,Ni,Co,Mn)₂₀(Fe,Cr,Ni,W,Co,Mn)₃(C)₆. It can be inferred that most atoms of W remain in solution in Ni-rich FCC, stabilizing such phase and avoiding the formation of primary precipitates.

In order to determine cooling rate during solidification of the alloy, literature containing data extracted from similar compositions⁵,⁶ was explored. Despite both correlations can be considered similar as can be seen in Figure 5, that proposed by Tinoco and Fredriksson⁵ was preferred because it was developed under slow to intermediate cooling conditions, which are more aligned with the current ones¹⁵. Therefore, square points in Figure 5 represent a projection of the experimental SDAS determined here so that the cooling rates driven by the correlation in Tinoco and Fredriksson⁵ were assumed. Both alloys in Tinoco and Fredriksson⁵ and Wu et al.⁶ are Ni-Cr-based, which allows the application of SDAS scaling laws safely, given that the solidification mode is the same as the alloy of interest here.

Figure 5
SDAS as a function of solidification cooling rate in the 28Cr48Ni5W3Co14Fe alloy.

Brinell hardness shows higher values as SDAS decreases, as shown in Figure 6. This effect is related to the refinement of the microstructure; in other words, the finer the microstructure, the higher the hardness. Smaller SDAS means a better distribution of refined interdendritic second phases. These phases act as barriers to dislocation movement, and can effectively block the movement of dislocations, increasing hardness. Both experimental equations for SDAS and HB were implemented in a commercial cast simulator. The experimental correlation HB = 324 × (SDAS)^-0.11 aligns with several previous studies on as-cast alloys that establish a relationship between hardness and dendritic spacing¹⁶-¹⁸. Moreover, SDAS is considered a predominant factor in as-cast microstructures as demonstrated by Ghassemali et al.¹⁹.

Figure 6
Hardness Brinell as function of SDAS in the 28Cr48Ni5W3Co14Fe alloy.

Figure 7 shows a strong variation in SDAS along the tube-support, especially when considering the regions between the holes (smaller SDAS) up to the boundary between one row and the next row of holes (larger SDAS). This line of variation was indicated with an arrow in Figure 7b at right side. It is worth noting that the relationships in Figure 7 were consistent with those in Figure 5 and Figure 6. As such, SDAS and HB were inferred based on the cooling rate computed values.

Figure 7
Examples of process maps through Magmasoft® simulation of a tube-support with the 28Cr48Ni5W3Co14Fe alloy: (a) cooling rate, and (b) SDAS.

Different regions with varying thicknesses and varied cooling rates can be observed in Figure 7. It was possible to predict local hardness values (see left side images in Figure 7b). Suitable proportional correlations between the yield strength and the hardness for Ni-based alloys has been established in the literature²⁰. Higher yield strength, thus, may be associated with a reduction of the mean free path of dislocations, which leads to an effective enhancement of this property. It can be seen that the side flange regions of the part are critical areas as they have lower hardnesses (i.e., lower yield strength) than the rest of the structure. In addition, large variations in hardness near the holes, together with shrinkage during cooling can be a source for initiation of cracks and defects in the casting. Regions characterized by smaller hardnesses (i.e., higher SDAS) can experience plastic deformation and initiate cracks from the hole surface.

4. Conclusions

The main novelty of this work lay in the determination of fundamental equations to plan the casting process and avoid defects in the alloy. The equations were accurately determined and applied in a real case to enhance the quality of the cast part. The CALPHAD calculations revealed the solidification sequence of the 28Cr48Ni5W3Co14Fe alloy, showing an extended solidification interval in Scheil model, and the formation of M23C6-type carbides.

For the first time in this type of alloy, based on the CALPHAD method, the primary formation of M₂₃C₆ was not identified. Instead, it formed as a quasi-eutectic phase under the described solidification conditions. Literature data was helpful to determine the cooling rates, with preference for the correlation in Tinoco and Fredriksson⁵ due to the low cooling rate levels consistent with the present setup. Smaller SDAS resulted in enhanced hardness via microstructural refinement. A reduced SDAS indicates a more uniform distribution of the interdendritic constituent, which serves as reinforcement under loading conditions. The simulation analysis of SDAS variations highlights critical areas, such as side flanges and holes, prone to plastic deformation and crack initiation due to lower hardness. It was emphasized the importance of understanding microstructural coarsening of Ni-based heat-resistant alloys for ensuring structural integrity, especially during cast operations.

5. Acknowledgments

The authors acknowledge FAPESP (grant 2023/06107-3) and CNPq. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. We also thank MAGMA Inc. for the access the software, and ENGEMASA for the donation of the alloy.

6. References

• ¹ Schillmoller CM. HP modified furnace tubes for steam reformers and steam crackers. Toronto: Nickel Development Institute; 1991. (NIDI Technical Series; 10058).
• ² Elliott P. A practical guide to high-temperature alloys. Mater Des. 1991;12(6):299-307.
• ³ Razorenov SV, Savinykh AS, Garkushin GV, Atroshenko SA. Effect of heat treatment and test temperature on the strength properties of cast heat-resistant nickel base Inconel 718 superalloy under shock-wave loading. Metals. 2022;12(7):1098.
• ⁴ Tillack DJ, Guthrie JE. Wrought and cast heat-resistant stainless steels and nickel alloys for the refining and petrochemical industries. Toronto: Nickel Development Institute; 1998. (NIDI Technical Series; 10071).
• ⁵ Tinoco J, Fredriksson H. Solidification of a modified Inconel 625 alloy under different cooling rates. High-Temp Mater Process. 2004;23(1):13-24.
• ⁶ Wu K, Liu G, Hu B, Li F, Zhang Y, Tao Y, et al. Solidification characterization of a new rapidly solidified Ni–Cr–Co based superalloy. Mater Charact. 2012;73:68-76.
• ⁷ Kaya H, Çadırlı E, Gündüz M. Directional cellular growth of Al–2 wt%Li bulk samples. Appl Phys, A Mater Sci Process. 2009;94(1):155-65.
• ⁸ Dao VH, Yoon KB, Yang G, Oh JS. Determination of creep constitutive model for 28-48WCo alloy based on experimental creep tests at 817-982 °C. J Mech Sci Technol. 2018;32(9):4201-8.
• ⁹ Roach DB, VanEcho JA. ASTM STP 756: comparison of the properties of the HK-40 and HP-45 cast heat-resistant alloys: stainless steel castings. Philadelphia: ASTM; 1982. p. 275-312.
• ¹⁰ Almeida Soares GD, Almeida LH, Silveira TL, Le May I. Niobium additions in HP heat-resistant cast stainless steels. Mater Charact. 1992;29(3):387-96.
• ¹¹ Almeida LH, Ribeiro AF, Le May I. Microstructural characterization of modified 25Cr–35Ni centrifugally cast steel furnace tubes. Mater Charact. 2002;49(3):219-29.
• ¹² Sustaita-Torres IA, Haro-Rodríguez S, Guerrero-Mata MP, de la Garza M, Valdés E, Deschaux-Beaume F, et al. Aging of a cast 35Cr–45Ni heat resistant alloy. Mater Chem Phys. 2012;133(2-3):1018-23.
• ¹³ Shatynski SR. The thermochemistry of transition metal carbides. Oxid Met. 1979;13(2):105-18.
• ¹⁴ Feurer U, Wunderlin R, Fachber DGM. Deutsche Gesellschaft für Metallkunde, Oberursel. In: Kurz W, Fisher DJ, editors. Fundamentals of solidification. Switzerland: Trans Tech Publications; 1986. p. 38. Appendix 8.
• ¹⁵ Oliveira PHF, Mancilha PHS, Reyes RAV, Gouveia GL, Bolfarini C, Spinelli JE, et al. Influence of the cooling rate on the solidification path and microstructure of a AlCoCrFeNi_2.1 alloy. Mater Charact. 2023;203:113121.
• ¹⁶ Afonso CRM, Martinez-Orozco K, Amigó V, Della Rovere CA, Spinelli JE, Kiminami CS. Characterization, corrosion resistance and hardness of rapidly solidified Ni–Nb alloys. J Alloys Compd. 2020;829:154529.
• ¹⁷ Dias M, Bogno A, Spinelli JE, Oliveira R, Cheung N, Garcia A, et al. Microstructure and hardness of an Al-8-wt%Si-2.5-wt%Bi alloy subjected to solidification cooling rates from 0.1 to 800-K-s −1. Adv Eng Mater. 2023;25(3):2201060.
• ¹⁸ Sobral BS, Vieira PM, Lima TS, Spinelli JE, Cheung N, Garcia A, et al. Effects of Zn addition on dendritic/cellular growth, phase formation, and hardness of a Sn-3.5-wt% Ag solder alloy. Adv Eng Mater. 2023;25(6):2201270.
• ¹⁹ Ghassemali E, Riestra M, Bogdanoff T, Kumar BS, Seifeddine S. Hall-Petch equation in a hypoeutectic Al-Si cast alloy: grain size vs. secondary dendrite arm spacing. Procedia Eng. 2017;207:19-24.
• ²⁰ Şerban V, Codrean C, Vodă M, Chicot D, Decoopman X. Correlation between yield stress and hardness of nickel–silicon-boron-based alloys by nanoindentation. Mater Sci Eng A. 2014;605:294-300.

Publication Dates

• Publication in this collection
  16 Dec 2024
• Date of issue
  2024

History

• Received
  11 July 2024
• Reviewed
  02 Oct 2024
• Accepted
  30 Oct 2024