As a researcher focused on advanced materials for energy applications, I have been deeply involved in the development of heat-resistant steels for ultra-supercritical (USC) power plants. The demand for higher efficiency and reduced emissions in thermal power generation has driven the need for steels that can withstand extreme temperatures and pressures. In this context, steel casting plays a pivotal role in producing components like turbine rotors and casings, where microstructure control is critical. This article presents my comprehensive investigation into the high-temperature phases of an ultra-high nitrogen 9Cr martensitic heat-resistant cast steel, aiming to optimize its microstructure through detailed phase analysis and heat treatment design. The study leverages computational thermodynamics, experimental techniques, and microstructural characterization to unravel the behavior of phases during casting and heat treatment, ensuring the alloy’s reliability in service.
The significance of this work stems from the reliance on imported materials for 600°C-class USC components in China, highlighting an urgent need for localization. The performance of heat-resistant steels under prolonged exposure to high temperatures hinges on a homogeneous matrix free of detrimental secondary phases. In steel casting, the as-cast structure often contains inhomogeneities and undesirable phases that can degrade mechanical properties. Therefore, my research targets the identification and control of high-temperature phases, such as δ-ferrite and carbides/nitrides, which form during solidification and subsequent processing. By integrating simulation and experimental tools, I seek to establish heat treatment protocols that minimize these phases, thereby enhancing the steel’s creep resistance and toughness.
My approach begins with the material composition, as listed in Table 1. This ultra-high nitrogen steel features a balanced array of alloying elements, including chromium, vanadium, tungsten, and cobalt, designed to promote martensitic transformation and precipitation strengthening. The high nitrogen content (0.3 wt%) is particularly noteworthy, as it influences phase stability and nucleation kinetics. In steel casting, such compositions are melted and poured into molds, leading to complex solidification patterns that affect phase distribution. To understand the equilibrium phases, I employed Thermo-Calc software with the TCFE9 database to simulate the phase diagram from 400°C to 1600°C, as shown in Figure 2. The simulation predicts the presence of austenite (γ), ferrite (δ), and intermetallic phases like Laves and carbides (M23C6, M6C) across temperature ranges, providing a baseline for experimental validation.
| C | N | Si | S | P | V | Cr | Mn | Co | Ni | Nb | Mo | W | Fe |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.027 | 0.3 | 0.06 | 0.006 | 0.013 | 0.55 | 9.13 | 0.083 | 2.745 | 0.139 | 0.017 | 1.09 | 6.684 | Bal. |
The as-cast microstructure, examined via optical microscopy after etching with picric acid, reveals a martensitic matrix with dispersed high-temperature phases. These phases appear as blocky and rod-like particles, ranging from 4 to 20 μm in size, often located within grains or at grain boundaries. Some particles contain dark precipitates in their cores, suggesting complex compositional gradients. In steel casting, such heterogeneity arises from non-equilibrium solidification, where elements segregate and form secondary phases. To quantify this, I used image analysis to measure phase fractions, finding that these high-temperature phases constitute approximately 3-5% of the area. Their presence is concerning because they can act as stress concentrators, reducing ductility and fatigue life. Thus, my subsequent experiments aim to characterize these phases and devise heat treatments to refine or eliminate them.
Phase transformation temperatures were determined using dilatometry. A cylindrical sample (φ5 mm × 10 mm) was heated from 25°C to 1300°C at 5°C/min in a vacuum dilatometer. The relative length change (ΔL/L0) and its first derivative (d(ΔL/L0)/dT) are plotted in Figure 3. The derivative curve shows four inflection points, which I correlated with phase transitions: (1) around 700°C, likely corresponding to the dissolution of low-temperature ferrite; (2) near 755°C, possibly associated with carbide precipitation or further ferrite changes; (3) at approximately 880°C, indicating austenite formation (Ac3); and (4) above 1045°C, signaling the onset of δ-ferrite and Laves phase formation. These experimental points align well with the Thermo-Calc simulation, confirming the phase stability ranges. For instance, the austenite start temperature (Ac1) is simulated at 830°C, while the dilatometry shows a broad transformation between 869°C and 895°C, reflecting kinetic effects. The data can be modeled using the Koistinen-Marburger equation for martensite formation during cooling, but for heating, I applied an Arrhenius-type relationship to describe phase transformation kinetics:
$$ \frac{dX}{dt} = k(T) \cdot (1 – X)^n $$
where \( X \) is the transformed fraction, \( k(T) = A \exp\left(-\frac{Q}{RT}\right) \) is the rate constant, \( Q \) is the activation energy, \( R \) is the gas constant, \( T \) is temperature, and \( n \) is the Avrami exponent. From the dilatometry curves, I estimated \( Q \) values for austenitization and δ-ferrite formation, which guide heat treatment design.
In-situ observation of phase transformations was conducted using a laser scanning confocal microscope (LSCM). Samples were heated from room temperature to 1200°C at 40°C/min, held at key temperatures for 2 minutes, then cooled at 100°C/min. This allowed real-time monitoring of surface relief changes associated with phase transitions. At 700°C, I observed the emergence and disappearance of a phase, identified as low-temperature ferrite based on its morphology and temperature range. At 880°C, the austenite transformation was subtle but evident in video recordings as a gradual smoothing of surface contours. Most strikingly, at 1090°C, δ-ferrite nucleated and grew rapidly, covering much of the surface, as shown in Figure 6. Upon cooling, this δ-ferrite transformed back to austenite and then to martensite, confirming its reversibility. These observations are crucial for steel casting processes, as they reveal the dynamic nature of phase evolution during thermal cycles, informing cooling rate selection to avoid excessive δ-ferrite retention.
To systematically evaluate heat treatment effects, I designed 24 different processing routes, varying homogenization temperature (1050°C to 1150°C), holding time (2 to 10 hours), cooling rate (air cooling vs. furnace cooling), normalizing parameters (temperature and time), and tempering conditions. Each route aimed to minimize high-temperature phases while achieving a fully martensitic structure. Specimens were characterized using scanning electron microscopy (SEM) equipped with energy-dispersive X-ray spectroscopy (EDS) and X-ray diffraction (XRD). Table 2 summarizes the key findings from selected treatments, highlighting trends in phase size and distribution.
| Treatment Route | Homogenization | Normalizing | Tempering | High-Temperature Phase Size (μm) | δ-Ferrite Content (%) | Microhardness (HV) |
|---|---|---|---|---|---|---|
| Route 1 | 1080°C/10 h, AC | 1100°C/2 h, AC | 780°C/2 h, AC | 4.2 ± 1.5 | 0 | 320 ± 15 |
| Route 2 | 1080°C/10 h, FC | 1100°C/2 h, AC | 780°C/2 h, AC | 6.8 ± 2.1 | 2.5 | 290 ± 20 |
| Route 3 | 1120°C/10 h, AC | 1150°C/4 h, AC | 750°C/4 h, AC | 5.5 ± 1.8 | 1.8 | 310 ± 18 |
| Route 4 | 1080°C/10 h, AC | 1100°C/6 h, AC | 820°C/2 h, AC | 7.3 ± 2.4 | 0 | 280 ± 22 |
From this data, I derived empirical relationships between processing variables and phase attributes. For high-temperature phases (mainly carbides/nitrides), the size increases with slower post-homogenization cooling, longer normalizing times, and higher tempering temperatures. This can be expressed as:
$$ d = k_1 \cdot \left( \frac{1}{v_c} \right)^{m_1} + k_2 \cdot t_n^{m_2} + k_3 \cdot T_t^{m_3} $$
where \( d \) is the average phase diameter, \( v_c \) is the cooling rate after homogenization, \( t_n \) is normalizing time, \( T_t \) is tempering temperature, and \( k_i, m_i \) are material constants. For δ-ferrite, the content is primarily controlled by cooling rate after homogenization; slower cooling promotes higher δ-ferrite volume fractions, as described by:
$$ f_{\delta} = C \cdot \exp\left(-\frac{v_c}{v_0}\right) $$
with \( f_{\delta} \) as the δ-ferrite fraction, \( C \) and \( v_0 \) as constants. These equations aid in optimizing steel casting post-processing to achieve desired microstructures.
EDS analysis on SEM samples provided compositional insights into the high-temperature phases. As shown in Figure 8, blocky particles are rich in chromium, vanadium, nitrogen, and carbon, while core precipitates contain additional tungsten and silicon. This indicates that these phases are complex carbonitrides of Cr, V, W, and Si, likely of the MX or M23C6 type. The nitrogen enrichment is expected given the high nitrogen content, which stabilizes nitrides and influences precipitation kinetics. In steel casting, such phases inevitably form due to element segregation, but their size and dispersion can be managed. XRD patterns confirmed the presence of martensite, retained austenite, and peaks corresponding to Cr2N and VN, aligning with EDS results.
Further, I conducted thermodynamic calculations to predict phase fractions using the CALPHAD method. The equilibrium phase amounts at different temperatures are listed in Table 3, derived from Thermo-Calc. This table helps identify temperature windows for heat treatment where undesirable phases are minimal.
| Temperature (°C) | Ferrite (δ) | Austenite (γ) | Martensite (α’) | Laves Phase | M23C6 | MX Carbonitrides |
|---|---|---|---|---|---|---|
| 800 | 0 | 95.2 | 0 | 0.5 | 3.1 | 1.2 |
| 1000 | 0 | 98.7 | 0 | 0 | 0.8 | 0.5 |
| 1100 | 12.3 | 87.1 | 0 | 0.6 | 0 | 0 |
| 1200 | 35.4 | 64.2 | 0 | 0.4 | 0 | 0 |
Based on these findings, I optimized the heat treatment process for this ultra-high nitrogen martensitic heat-resistant cast steel. The recommended route involves homogenization at 1080°C for 10 hours followed by air cooling to limit δ-ferrite, normalizing at 1100°C for 2 hours to fully austenitize and dissolve Laves phases, and tempering at 780°C for 2 hours to stabilize the martensite and control carbide growth. This yields a fine martensitic lath structure with high-temperature phases reduced to an average size of 4 μm and no residual δ-ferrite, as verified by metallography. The success of this optimization underscores the importance of integrating simulation and experiment in steel casting research, enabling tailored microstructures for demanding applications.
In conclusion, my investigation demonstrates that high-temperature phases in ultra-high nitrogen 9Cr martensitic heat-resistant cast steel are predominantly chromium, vanadium, tungsten, and silicon carbonitrides, which cannot be entirely eliminated but can be refined through controlled heat treatment. δ-Ferrite, however, can be fully suppressed by fast cooling after homogenization. The phase transformation sequence during heating includes low-temperature ferrite dissolution near 700°C, austenitization around 880°C, and δ-ferrite formation above 1045°C, as confirmed by dilatometry and in-situ microscopy. These insights provide a foundation for improving the manufacturing of cast steel components for USC power plants, where microstructure homogeneity is paramount. Future work will focus on creep testing to correlate phase characteristics with long-term performance, further advancing the field of high-performance steel casting.
To facilitate reproducibility, I include the detailed experimental protocols. For dilatometry, samples were machined to precise dimensions and tested under argon atmosphere. LSCM specimens were polished to a mirror finish and observed using a 1500°C furnace stage. SEM/EDS was performed at 20 kV accelerating voltage, with phase compositions averaged over 10 measurements. XRD used Cu-Kα radiation, scanning from 20° to 100° 2θ. All heat treatments were conducted in tube furnaces with temperature control within ±5°C. The data analysis employed software like ImageJ for microscopy and Origin for curve fitting.
Lastly, I reflect on the broader implications. In steel casting, the challenge of managing high-temperature phases is universal, especially for nitrogen-alloyed grades. My approach of combining Thermo-Calc, dilatometry, and LSCM offers a robust framework for phase analysis, applicable to other alloy systems. By sharing these methodologies, I hope to contribute to the development of next-generation heat-resistant steels, supporting the global transition to efficient, clean energy. The integration of computational tools with advanced characterization will continue to drive innovation in materials science, ensuring that cast steels meet the ever-increasing demands of high-temperature service.