A New Alloy Designed for Superheater Tubing in Coal-Fired Ultra Supercritical Boilers B. A. Baker

Transcription

1 Superalloys 718, 625, 706 and Derivatives 2005 Edited by E.A. Loria TMS (The Minerals, Metals & Materials Society), 2005 A New Alloy Designed for Superheater Tubing in Coal-Fired Ultra Supercritical Boilers B. A. Baker Special Metals Corporation 3200 Riverside Drive Huntington, WV Keywords: Nickel-base, superalloy, coal ash corrosion, ultra supercritical ABSTRACT Future increases in the demand for clean and efficient electrical power production will be unrelenting. Meeting the coming challenges will require power generating systems capable of operating with increased boiler pressures and temperatures. In turn, meeting the demands of the application will also require materials which can withstand these extreme conditions. This paper describes a new nickel-base tubing alloy, INCONEL alloy 740, developed for the purpose of meeting this challenge. This material possesses a unique combination of elevated temperature strength properties and resistance to coal ash corrosion required by the application. The material was developed to fulfill the minimum stress rupture requirement of 100,000 hour rupture life at a stress of 100 MPa and at a temperature of 750 C. In addition, metal loss of less than 2mm in 200,000 hours was defined as the target corrosion resistance for this material. The mechanical properties of this new material will be described, in addition to its coal ash and steam corrosion resistance and weldability. Thermal stability will also be discussed, focusing upon microstructural features of long-term exposed samples. Early methodology utilized to arrive at the current chemical composition will also be presented. INTRODUCTION When ultra supercritical steam conditions employing greater than 300 bar steam and 600 C steam temperature are adopted, it has been demonstrated that the efficiency of pulverized coalfired boilers can be increased to over 50% LHV (lower heating value). At this steam temperature, the superheater/reheater midwall temperature may be at 660 C or higher. At these metal temperatures, the conventionally used 9-12% Cr steels must be replaced by austenitic stainless grades. Projects like the European AD700 project and the German MARCKO DE2 project are planning for steam pressures up to 375 bar and steam temperatures to 700 C. This will lead to superheater/reheater mid-wall temperatures at over 740 C. Under these conditions, austenitic stainless steels cannot meet the stringent requirements of 100,000 hour rupture life at 750 C and 100 MPa and corrosion resistance defined as metal loss not exceeding 2mm after 200,000 hours. ALLOY DEVELOPMENT METHODOLOGY With no alloy in existence that could meet both the mechanical and corrosion demands of the described applications, an alloy development program was initiated. It was determined that the existing superalloy NIMONIC alloy 263 possessed the required rupture strength but lacked the needed corrosion properties. This alloy was used as the basis for the development, utilizing additions of Cr and Nb for enhancement of corrosion resistance while maintaining the required volume percentage of gamma prime for precipitation hardening. INCONEL and NIMONIC are trademarks of the Special Metals Corporation group of companies. 601

2 A rapid method of assessing rupture strength properties was utilized, involving testing samples and fixed stress of 100 MPa over a broad temperature range. Figure 1 shows a plot of collected data. The solid line represents the strength capability of NIMONIC alloy 263. Materials falling below that line were rejected from the program. Data shown for developmental alloys 1 and 2 fall above the alloy 263 data while data for alloy 3 lie well below. 1.0E E+04 1 Rupture Life, Hours 1.0E E E E Temperature, C Figure 1. Plot showing rupture life in hours versus temperature in degrees Celcius. The solid line is the isocline for NIMONIC alloy 263. The data points shown represent experimental materials evaluated in the alloy development program. Figure 2 shows the results of predictions made using Thermo-Calc of the atomic percentage of gamma prime as a function of Al and Ti content. The chart was constructed by mapping interpolated data. A minimum percentage of 15% was deemed acceptable. Assessment of preliminary data acquired to evaluate mechanical and corrosion properties of developmental heats of INCONEL alloy 740 resulted in the nominal composition shown in Table 1 being established. Nominal composition for other alloys mentioned in this study are included as well. 602

3 Table 1. Nominal Composition of the Alloys of This Study. Alloy C Ni Cr Mo Co Al Ti Nb Mn Fe Si Bal Bal Bal Bal Bal Bal MECHANICAL PROPERTIES OF INCONEL ALLOY 740 Stress rupture properties of INCONEL alloy 740 as determined by the AD700 consortium are shown in Figure 3. Table 2 shows tensile properties for hot-rolled 16mm diameter bar of INCONEL alloy 740. Table 3 shows room-temperature tensile data and Rockwell hardness values for 6.4mm thick hotrolled and solution annealed plate after receiving aging treatments at either 760 C or 800 C with aging timed varied between 4 hours and 16 hours. Strength levels were slightly higher for samples aged at 760 C versus 800 C; longer aging times resulted in slight strength increases at both temperatures. Elongation and reduction of area did exhibit a slight downward trend as aging time increased at each temperature; ductility values were in a similar range at both temperatures. Table 4 shows the effect of the same aging temperatures and times upon impact properties. Samples were half-size; a favorable impact strength of 75 J/cm 2 was retained after aging for 16 hours at both 760 C and 800 C Aluminum, Weight % Titanium, Weight % Figure 2. Predictions made using Thermo-Calc of the atomic percentage of gamma prime as a function of Al and Ti at 750 C content for Ni- 24Cr-20Co-1Fe-0.5Mo-0.03C. 603

4 ºC Stress MPa 750ºC 725ºC 775ºC 800ºC Rupture Life - hours Figure 3. Stress rupture properties of INCONEL alloy 740 as determined by the AD700 consortium. Table 2. Tensile Property Data for INCONEL alloy 740 Hot-Rolled Bar Test Temperature 0.2% Yield Strength Tensile Strength Elongation Reduction of Material F C ksi MPa ksi MPa % Area, % Condition Annealed a Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b Annealed and Aged b a 2100 F/30 minutes/water quench. b 2100 F/30 minutes/water quench F/16 hours/air cool. 604

5 Table 3. Room Temperature Tensile and Hardness Data for 6.4mm Thick Hot- Rolled and Solution-Annealed INCONEL alloy 740 Plate Rockwell C Hardness 0.2% YS, ksi UTS, ksi Elong., % R of A, % Aging Treatment C/4hr/ac C/8hr/ac C16hr/ac C4hr/ac C/8hr/ac C/16hr/ac Table 4. Room Temperature Charpy Impact Data for 6.4mm Thick Hot-Rolled and Solution-Annealed INCONEL alloy 740 Plate Impact Energy, Ft-Lb Impact Energy, J/cm 2 Aging Treatment C/4hr/ac C/8hr/ac C/16hr/ac C/4hr/ac C/8hr/ac C/16hr/ac Note: Sample cross section = 0.4 cm 2 at notch; average of triplicate tests. MICROSTRUCTURAL CHARACTERIZATION AND STABILITY OF INCONEL ALLOY 740 Detailed studies of the microstructural stability of INCONEL alloy 740 have been carried out by Dr. Xie and his colleagues at the University of Beijing, and also by Evans, et al., at Oak Ridge National Laboratory [1,2]. Xie, et al., examined the microstructure of solution annealed INCONEL alloy 740 hot-rolled and solution annealed bar after aging at 704 C (1300 F) and 725 C (1337 F) for 500, 1,000, 1,500 and 2,000 hours [1]. The SEM images are shown in Figures 4 The precipitate morphologies are fairly similar among the various exposure conditions. Very fine precipitates (16.6 wt. % after 2,000 hours) were distributed throughout the grains, which contribute the main strengthening effect. Larger (Nb,Ti)C particles (0.15% after 2,000 hours) were also observed throughout the grains, while the grain boundary precipitate after long term aging at 704 C (1300 F) was Cr 23 C 6 (0.15% after 2,000 hours). After aging at 725 C (1337 F) for 4,000 hours, a small amount of needle-like (η) and blocky precipitates (G-phase) had also formed in the grain boundaries. Figure 5 shows the coarsening rate of the precipitates at 704 C (1300 F), 725 C (1337 F) and 760 C (1400 F) for times to 4,000 hours. Gamma prime increases with time and temperature in the temperature range studied. At the initial stage of aging, the particles were small and spherical and eventually became cuboidal with time and temperature, confirming the work of Evans, et al [2], who examined a sample which had been solution-annealed and aged, followed by creep testing at 816 C (1500 F) and 138 Mpa (20 ksi). The relationship between the radius of and time is linear and follows the kinetics of diffusion controlled particle growth [1]. Both investigations confirmed the presence of the following precipitates: γ, η, G-phase, M 23 C 6 carbides and MC carbides. The composition for these respective precipitates is shown in Table 5 as reported by the Evans study, found in a creep sample tested at 816 C (1500 F) and 138 MPa 605

6 Figure 4. SEM images after aging for 500h (a), 1000h (b), 2000h (c) and 4000h (d) at 704 C (left) and 725 C (right). (20 ksi) for 2500 hours. While the η phase in this sample was not determined to have had an embrittling effect upon the alloy, it was noted that the laths extending across grains did appear to have grown at the expense of γ precipitates. In addition, at this high exposure temperature, the gamma prime had coarsened from ~60 nm to 240 nm in diameter and lost coherency with the matrix. However, the precipitates were still presumed to have enhanced matrix strength via dislocation pinning. 606

7 Figure 5. Coarsening of Precipitates in INCONEL alloy 740 (1). Table 5. Measured Composition (atom %) of the Phases Present After Creep Testing of INCONEL alloy 740 at 815 C (1500 F)/138 MPa (20 ksi) for 2,500 hours (2) Phase Ni Cr Co Fe Ti Al Nb Mo Si γ γ η M 23 C MC G CORROSION RESISTANCE OF INCONEL ALLOY 740 Corrosion testing was carried out in the interest of simulating coal ash corrosion conditions. The ash mixtures utilized consisted of 5% Na 2 SO 4-5% K 2 SO 4-90% (Fe 2 O 3 -Al 2 O 3 -SiO 2 in 1:1:1 ratio by weight). The ash mixture was prepared by grinding with mortar and pestle. The synthetic flue gas was mixed using electronic flow meters from component gases (including a pre-mixed cylinder of N 2-1.8% SO 2 ) and consisted of N 2-15% CO 2-3.5% O % SO 2. Test samples consisted of 7.6 mm diameter X 19 mm long cylindrical pins machined to a 32 microinch finish. Prior to testing, the synthetic coating was applied to the samples surfaces by first diluting with acetone and then using a bristle brush. The approximate mass gain resulting from application of the coating was 40 mg/cm 2. Samples were allowed to dry completely before testing. The coated samples were placed in a cordierite ceramic boat and exposed in a sealed horizontal, electrically heated muffle furnace having a 100 mm OD mullite tube. Samples were placed into the heated section of the furnace after sufficient purging and introduction of the test gas, using a sealed pushrod mechanism. Samples were cycled at approximately 500 h, 1000 h, 2000 h and 4000 h. Exfoliated surface products were removed and the coating was re-applied after each of the given time intervals. Metal loss and average depth of attack were assessed from sample cross sections. Average values for metal loss and depth of attack were determined from a total of six measurements made around the circumference of each pin. In the case of the depth of attack measurement, the maximum depth of intrusion was determined for each of the six fields examined at approximately 60 degree intervals around the circumference of the pin. Figure 6 shows depth of attack measurements for alloys tested at 700 C with the simulated flue gas and applied salt as described above. Alloys 617, A1 and A2 (alloy 740 variants with 6% 607

8 Mo) suffered significant thickness loss, exhibiting linear kinetics. These results could indicate that molybdenum at 6-9% is undesirable for coal ash corrosion resistance. Perhaps incorporation of molybdenum into the trisulfate ash forming at the material surface leads to an increased corrosion rate by affecting the solubility of protective chromium oxides, as observed in Mocontaining alloys which undergo high temperature Type I hot corrosion at about C [3,4]. Depth of Attack, Microns Expsoure Time, Hours 617 A-2 A Figure 6. Depth of attack after exposure at 700 C in N 2-15% CO 2-3.5% O % SO 2 with samples having a salt consisting of 5% Na 2 SO 4-5% K 2 SO 4-90% (Fe 2 O 3 -Al 2 O 3 -SiO 2 in 1:1:1 ratio by weight) applied to the surface (re-coated at the intervals shown). Figure 7 shows results from three different data sources after testing under fairly similar conditions. These data represent a number of different alloys ranging from Fe-Ni-Cr types to nickel-base superalloys. These data show that, in general, once the chromium content exceeds 25-30%, corrosion activity stabilizes. Figure 8 shows a comparison of both linear and parabolic extrapolations of data acquired for INCONEL alloy 740 and INCONEL alloy 617 under the same conditions as shown in Figure 6. Based upon 4000 hour data, these extrapolations show that INCONEL alloy 617 is not projected to meet the design criterion of less than 2mm of corrosion attack in 200,000 hours whereas INCONEL alloy 740 is, even if an eventual linear progression rate is assumed. FABRICATION CONSIDERATIONS Fabricability is a key consideration in the design of a material to be used for assembly of power boilers components. Weldability was demonstrated via assembly of a beveled butt joint using 6.4mm plate and matching filler wire with the manual gas tungsten arc welding (GTAW) technique. The plates were joined in the solution annealed and aged condition to simulate field construction of repair operations. A post-weld aging treatment was also applied. The weldment was completed using a current setting of 185A, with voltage at 14V using 100% Ar at 30 CFH as the shielding gas and a diameter, 2% thoriated tungsten electrode with a travel steep of ~2-4 inches per minute. Table 6 shows room temperature tensile results for 608

9 transverse samples from the assembled joint. Specimens passed a 2T face bend in the as-welded condition; specimens tested in the post-weld aged condition failed the 2T bend but passed a 4T bend. Average Thickness Loss (Microns/1000 Hours) Blough and Stanko [5] Castello, et al [6] /Baker and Smith [7] Mass % Cr Figure 7. Cross section loss (depth of attack) for high temperature alloys exposed at 700 C in laboratory flue gases (0.25% SO 2 ) with alkali sulfate-oxide coatings applied to the sample surface, as a function of chromium content. [5]: N 2-14%CO 2-10%H 2 O-3.6%O %SO 2,, 10% alkali sulfates [6]: N 2-15%CO 2-3.5%O %SO 2, 10% alkali sulfates [7]: N 2-15%CO 2-3.5%O %SO 2, 10% alkali sulfates 609

10 Maximum Attack, Microns INCONEL alloy 617 INCONEL alloy Linear Projection 740 Linear Projection 740 Parabolic Projection Time, Hours Figure 8. Depth of attack results for INCONEL alloy 617 and INCONEL alloy 740 samples tested in a simulated flue gas mixture comprised of N 2-15% CO 2-3.5% O % SO 2 at 700 C with a salt coating applied which was comprised of 5% Na 2 SO4-5% K 2 SO 4-90%(Fe 2 O 3 -Al 2 O 3 -SiO 2 in 1:1:1 ratio). Table 6. Room Temperature Tensile Results for Manual GTAW Weldments Made Joining 6.4 mm (0.25 ) Commercial Plate Using Diameter Experimental Wire. The As-Produced Base Material (Hot-Rolled and Solution Annealed) Was First Aged at 800 C (1472 F)/4 Hours, Welded and Then Aged after Welding at 800 C (1472 F)/4 Hours. Sample 0.2% Yield Strength MPa (ksi) Ultimate Tensile Strength MPa (ksi) Elongation % (107) 1100 (159.5) (109.8) 1079 (156.5) 14.8 CONCLUSIONS 1.) INCONEL alloy 740 successfully surpasses the established strength design target defined as aim stress rupture life exceeding 100,000 hours at 750 C and 100Mpa. 2.) INCONEL alloy 740 as defined for the purposes of this study exhibits standard gamma prime growth kinetics. At higher temperatures and longer times, η-phase formation occurs with a Widmanstatten morphology. 3.) INCONEL alloy 740 exhibits favorable resistance to coal ash corrosion in laboratory tests designed to simulate boiler conditions. 4.) INCONEL alloy 740 can be readily welded using GTAW and argon shielding. In the aged, welded and aged condition, the room temperature tensile properties approached those of the base material. 610

11 REFERENCES 1. S. Zhao, J. Dong, X. Xie, G. D. Smith and S. J. Patel, Thermal Stability Study on a New Ni- Cr-Co-Mo-Nb-Ti-Al Superalloy, presented at the 10 th International Symposium on superalloys, Seven Springs, PA, September 19-23, N. D. Evans, P. J. Masiasz, R. W. Swindeman and G. D. Smith, Microstructure and Phase Stability in INCONEL alloy 740 during Creep, Scripta Materiala, 51 (2004). 3. R. A. Rapp and Y. S. Zhang, Hot Corrosion of Materials: Fundamental Studies, JOM, Vol. 46. No. 12, 1994, p J. A. Goebel, et al., Mechanisms for Hot Corrosion of Nickel-Base Alloys, Metallurgical Transactions, Vol. 4, 1973, pp J. L. Blough and G. J. Stanko, Fireside Corrosion Testing of Candidate Superheater Tube Alloys, Coatings and Claddings - Phase II, CORROSION/97, Paper No. 140, NACE International, Houston,TX, Castello, P., Guttmann, V., Farr, N.*, Smith, G., Laboratory Simulated Fuel-Ash Corrosion of Superheater Tubes in Coal-Fired Ultra Supercritical-Boilers, Proceedings of the EUROMAT Conference, September 1999, München (D), Wiley-VCH Verlag - ORA/PRO B. A. Baker and G. D. Smith, Corrosion Resistance of Alloy 740 as Superheater Tubing in Coal-Fired Ultra-Supercritical Boilers, CORROSION/2004, Paper No , NACE International, Houston, TX,