1 edited by J. Lecomte-Bec ~:OE024913649* dB. Kuhn. HIGH

gth Liege Conference : Ma1 edited by J. Lecomte-Bec

1\111~11\~11~1

\t~\\\\t\1 ~:OE024913649*

ineering 2010 dB. Kuhn.

HIGH-TEMPERATURE PERFORMANCE OF A NEW NICKEL-BASED FILLER METAL FOR POWER GENERATION APPLICATIONS

J. Shingledecker1, K. Coleman1

, J. Siefetf, J. Tanzosh2, W. Newele

1. Electric Power Research Institute, Charlotte, NC USA 2. Babcock & Wilcox Research Center, Barberton, OH USA

3. Euroweld, Mooresville, NC USA

Abstract

A new nickel-based weld filler metal, EPRI P87, has been developed as a superior alternative to ERNiCr-3 for use in dissimilar metal welds (DMW) between ferritic and austenitic materials. EPRI P87 has a low coefficient of thermal expansion more closely matching alloys such as Grade 91 and 92 than other available filler metals. Additionally, the size of the carbon denuded region adjacent to the weld in the heat-affected-zone is minimized/eliminated by proper control of weld metal composition. In this work, the high-temperature mechanical behavior of DMWs utilizing EPRI P87 (GTA W and GMA W processes) was characterized through tensile and long-term creep-rupture testing. Microstructure analysis was also conducted on tested specimens to evaluate the HAZ regions and failure modes. Performance of the weld metal and welded joints is discussed and compared with ERNiCr-3 and typical 9%CrMo V filler metals.

Keywords: Dissimilar metal welds, DMW, nickel-based filler metal, lnco 82, ERNiCr-3, EPRI P87

1. Introduction

The metallurgical sources of failure in dissimilar metal welds (DMW) between ferritic and austenitic materials, specifically in high-temperature boiler tubing, include: the mismatch in coefficient of thermal expansion (CTE) between the weldment constituents, and the formation of a 'weak' carbon-free interface or carbon-denuded zone between the stainless or nickelbased filler metal and the ferritic steel during high-temperature exposure [1,2]. ERNiCr-3 (N06082), also referred to as Inco 82 or Inconel 82, is generally the preferred filler metal for DMWs today, but it is not immune to these metallurgical phenomenon. EPRI P87 filler metal was developed to address these issues as an alternative to ERNiCr-3 [3, 4]. Figure 1 compares the mean CTE as a function of temperature for EPRI P87 and ERNiCr-3 to various alloys. Over a broad range of temperatures, EPRI P87 has a CTE below that ofERNiCr-3 and closer to Grade 91 making it an attractive choice for DMWs involving creep strength enhanced ferritic (CSEF) steels. Figure 2 shows the carbon denuded region of a CrMo steel weldment interface after normalization for 2 hours at 1038°C (1900°F). The samples welded with nickel-base and stainless filler metals both show large carbide free regions, greater than l OOJlm in thickness, whereas the interface of the weldment produced with EPRI P87 is clean with no indication of carbon migration.

338

gth Liege Conference : Materials for Advanced Power Engineering 2010 edited by J. Lecomte-Beckers, Q. Contrepois, T. Beck and B. Kuhn.

All research, development, and initial stress-rupture testing on EPRI P87 was conducted using shielded-metal-arc welding (SMA W). A heat of EPRI P87 wire has been produced and

subjected to extensive weldability trials as reported in [5]. In this work, the behavior ofEPRI

P87 filler metal and DMWs from gas-tungsten arc and gas-metal arc welding (GTA Wand

GMA W) are presented and discussed.

13.00 -e-P22 -+-P91 -+-TP304H

u: --o- Super 304H -o- 347HFG -o-NF709

~ 12.00 -e- ERNiCr-3 -+- P87 ----~

! 18-20%Cr8-10%Ni Stainless Steel c 11.00 0 1i ~ c Ill 10.00 1:1.

~ ii 9.00 ~ Gl

.r::. 8.00 1-

0 .. c 7.00 Gl

EPRI P87

u 'i

6.00 0 CSEF Steel (Gr. 91)

(..)

c ftl

5.00 CD ~

Ferritic Steel (Gr. 22)

4.00 0 100 200 300 400 500 600 700 BOO

Temperature ("C)

Figure 1. Mean Coefficient of Thermal Expansion for EPRI P87 compared to other filler metals and alloys, data from ref [5]

Figure 2. Micrographs of the weldment interface of a CrMo steel using a nickel-based

(ERNiCrMo-3), stainless (ER309L) and EPRI P87 filler metals after normalizing and

tempering (scale bar= lOOf.II'J) showing EPRI P87 has virtually no carbon-denuded zone {3]

339

gth Liege Conference : Materials for Advanced Power Engineering 2010 edited by J. Lecomte-Beckers, Q. Contrepois, T . Beck and B. Kuhn.

2. Experimental Procedure

Two weldment configurations were used to produce the material for this study. The composition of the EPRI P87 filler metal and the 347H and Grade 91 plates are provided in Table I. Chemical composition ofthe EPRI P87 wire conformed to the SMAW consumable target composition. 50.4 mm (2") thick plates of347H and grade 91, approximately 915mm (36") in length, were welded with approximately 6.4mm (0.25") of cold-wire GTAW in the root, and the remaining --44.5 mm (1.75") of the joint was filled using hot-wire GTAW. 25.4mm (1'') plates of347H and grade 91, approximately 361mm (15'') in length, were welded using the GMA W process. Figures 3 and 4 show the joint designs for these thick section welds. Following welding, all plates were given a post-weld heat-treatment (PWHT) at 746°C (1375°F) for 4 hours.

Table 1. Chemical Composition (wt%) for Filler Metal and Alloys Used in this Study (EPRI P87 composition target based on ref./6])

c Mn Si s p Cr Ni Mo Nb Fe AI N Ti Co Cu V Zr Nb-+Ta EPRI PI7Wire 0.11 1.45 0.14 0 .00 1 <0.005 8.92 44.8 2.05 1.26 40.98 0.16 0.003 0.11 0.002 0.008 0.002 <0.001 SMAWERPI

P87 Tareet 0.1 1.5 0.3 0.008 0.008 9 bal 2 I 38 347H 0.054 1.3 0.51 0.0003 0.029 17.41 9.88 0 .5 1 0.73 Bal. 0 .065 0.25 0.4 0.74 Gr. 91 0.12 Q.415 0.24 0.0026 0.011 8.901 0.222 0.856 0 .07 Bal. 0.02 0.045 0.003 0 .031 0.1 94 0 .006

After PWHT, the weldments were subjected to destructive metallurgical analysis. No microfissuring or cracking was observed in any weldment except in the cold-wire GTAW root passes. These cracks were found in the weld metal and did not extend into the hot-wire GTAW passes. Extensive testing did not indicate impaired weld metal ductility or a high residual condition, thus the source of cracking appears to be joint geometry/backing bar material related [5]. All mechanical test specimens for the GTAW welds were taken from the hot-wire region to avoid any cracks which could influence test results.

Creep testing and elevated temperature tensile testing was performed in accordance with ASTM E21 and El39. Cross-weld (CW) tensile and stress-rupture tests were conducted on specimens, with a gauge length (GL) = 57.2mm (2.25") and a diameter (dia.) of 12.8mm (0.505"), from both the GTA W and GMA W plates with the weld centerline in the middle of the gauge. All weld metal (A WM) tensile tests were conducted on specimens, GL=31.8mm (1.25") and diameter of6.4mm (0.25"), from the GTAW plate. A WM creep and creeprupture tests were also conducted on specimens, GL=35.6mm (1.4") and dia=8.9mm (0.35"), from both the GTA W and GMA W plates.

3. Results

Cross-weld (CW) tensile tests were conducted on both weld joints from 427 to 760°C (800 to 1400°F) and on all-weld-metal (A WM) GTAW specimens from 482 to 704°C (900 to 1300°F). The weld metal elevated temperature yield strength was -300 MPa (43.5ksi) in this temperature range. The tensile strength for the weld metal and weldments is plotted in figure

340


5 and compared to other EPRI P87 GTAW room temperature tensile tests which failed in the EPRI P87 filler metal [5,7]. The current tensile strength for 347H plate and grade 91 in Figure 5 is taken from the ASME Boiler and Pressure Vessel Code, Section 11 Table U [8]; these data are based on a trend curve tied to the minimum room temperature strength of the alloy. Failure location is also indicated in the figure.

s· ~ V ,_i _20_.3_-_

.-------{ G<*91

EPRI Pl7 Fl6,

lloo-wl .. liG

9l5nun ln

left&tb,1 ......

Figure 3. Joint Design for the 50mm (2") Thick GTAW Joint

3&1mm \n

347tl Gliide91

EPRIN7FIII. G ..... W

»>I.Badcina 8or

5mm

Figure 4. Joint Design for the 25mm (1 '} Thick GMA W Joint

Examination of the plot shows the room temperature EPRI P87 filler metal tensile strength is above the room temperature minimum strength of Grade 91 (586MPa- 85ksi), and above ~538°C (1000°F) EPRI P87 exceeds the strength of Grade 91. The cross-weld failure

locations confirm this trend with tests at 427-482°C (800-900°F) failing in the 347H. At

temperatures above 900°F, all CW failures were in the grade 91.

341

800

700

600

'ii' D.. !. 500 .c .. Dl

~ 400 ;;; Gl

'ii 300 c Gl 1-

200

100

0

0

gth Liege Conference : Materials for Advanced Power Engineering 2010 edited by J. Lecomte-Beckers, Q. Contrepois, T . Beck and B. Kuhn .

.. ... 347H Plate (SA-240) Table U

---A-GTAW 91-P87-347H Cross-Weld UTS

.. . . . Gr. 91 (SA-387) Table U ::

-+-GMAW 91-P87-347H Cross-Weld UTS

0 GTAWTub8s[5,7] --o-GTAW P87 Weld Metal

s-- Cross-Weld Failure in EPRI P87

ij./.................. ....... Cross-Weld Failures in Stainless Steel

·-· -·--·~ •·- "• ..... ... ..

100 200 300 400 500 600 700 800

Temperature {"C)

Figure 5. Tensile Strength as a function of temperature for GTAW and GMAW CW, GTAW

A WM, and GTAW CW Tubes (from ref [5, 7}) compared to Gr. 91 and 347H minimum specified properties [8]

The test conditions and results for CW testing of the GTA Wand GMA W joints is provided in Table 2. These data are plotted as a function ofLarson-Miller-Parameter (LMP) with a

constant ofC=28 against the Grade 91 average and minimum (-20% on stress) rupture behavior in Figure 6. Cross-weld rupture data from subcritically PWHT'ed grade 91 joints made with EPRI P87 by SMA W process [3] and using nominally matching 9%CrMoV filler

metals [9] are plotted for comparison. High-stress ruptures were within the grade 91

scatterband and failures were in the base metal (BM). In the lowest stress tests, failures were

at or below the minimum scatterband for Grade 91 and failures were confirmed, by optical

microscopy, to be in the heat-affected-zone (HAZ) of the weldment which is consistent with

subcritically heat-treated Gr. 91 and other CSEF steels.

Table 2. Cross-Weld Stress-Rupture Test Conditions and Results Conditions Time to Rupture (hrs) Failure Location Temperature Stress CJF) MPa (ksi) GTAW GMAW GTAW GMAW

593 (1100) 172.4 (25.0) 326.1 208.2 91BM 91 BM 593 (1100) 151.7 (22.0) 507.0 674.0 91 BM 91 BM 625 (1157) 100 (14.5) 899.2 551.5 91HAZ 91 HAZ

342


Creep-rupture tests were conducted on A WM samples from 593 to 760°C (1100 to 1400°F) between 77 and 172 MPa (11.2 to 25ksi). Test times ranged from a few hours to a few

thousand hours with some tests continuing beyond 8800 hours. Minimum creep rate ( i.mn)

data were evaluated using a simple Norton-Bailey approach with a standard Arrhenius equation [10] where n is the stress exponent, A is a constant, Q is the activation energy, R is

the gas constant, and T is the temperature in Kelvin as follows:

imin = Aa"e-QIRT Equation I

where A=l.1398668xl0"4/hr, n=8, and Q=452.84916 KJ/mol. The test data for both the

GTA Wand GMA W weld metal and the Equation I fits are shown in Figure 7. No significant differences were observed between GTA Wand GMA W test results.

li 11.

1000

~ 11) 100 11)

I!! -Cl)

10 24000

--Gr. 91 Average -- Gr. 91 -20%

D Gr. 91 P87 SMAW (Subcritical PWHT) 0 Gr.91/347H P87 GTAW (Subcritical PWHT) t:. Gr.91/347H P87 GMAW (Subcritical PWHT) • Gr. 91 CW 9Cr Filler (Ref. [9])

25000 26000 27000 28000 29000

LMP {K,C=28)

30000

Figure 6. Stress-rupture results for the GTA Wand GMA W CWs plotted using the Larson

Miller Parameter (LMP, constant C=28) and compared to EPRI P87 CW SMAW process [3]

and typical grade 91 cross-weld data using matching 9%CrMo V filler metals from [9]

343


I .!! e 0: a. . ! u c !i

1.00E.+OO

1.00E-01

1.00E-02

1.00E-03

1.00E-04

1.00E-05

1.00E-06

1.00E-07

• 593C (1100F) GTAW

• 700C (1292F) GTAW • 760C (1400F) GTAW

• 760C (1400F) GMAW

- •· 593C Eqn. 1

t-----------~~"'-- --700CEqn. 1 -760CEqn. 1

I

10.0 100.0

Stress (MPa)

1000.0

Figure 7. Minimum creep-rate as a function of temperature and stress for EPRI P87

Rupture results for the EPRI P87 GTA Wand GMA W weld metal are shown in Figures 8, 9, and 10. Figure 8 shows the rupture ductility for the completed tests. In all cases, rupture elongation exceeded 20% and reduction of area exceeded 50%. Figure 9 is a comparison of the GTA Wand GMAW weld metal and reported behavior ofERNiCr-3 [11] at 760°C (1400°F). For tests greater than 100 hours, both the GTAW and GMA W EPRI P87 exceed the average reported time to rupture for ERNiCr-3 and the slope of the time to rupture curve fit to the EPRI P87 GT A W suggests this will persist at longer test times. GMA W rupture strength appears to be slightly less than the GT A W. Figure 10 is a LMP plot of all the rupture data compared to the EPRI SMA W development work [3] and the reported 100, 1000, and 10,000 hour data for ERNiCr-3. A dashed line is fit through the ERNiCr-3 data for comparison. At high stresses, all of the EPRI P87 product forms appear equivalent in terms of rupture strength. The lower-stress GT A W data appears to exceed the strength of the ERNiCr-3.

100 P87 All Weld Metal D GTAW- Elongation o GTAW- Red. of Area • GMAW- Elo!1g;abon • GMAW- Red. of Area

90 • 80 0 ~ Q l • 70 6 ~ • ~ 60

0 i

n D • .--• D

I!

I 0 0

1.0 10.0 100.0 1000.0 10000.0

Rupture Life (hours)

Figure B.Post-test creep-rupture ductility for EPRI P87 GTAW and GMAW weld metal

344


1000.0

Ci Q,

!. 100.0 ., ..

!: "'

10.0 1.0

!T60"C (1400"F) I [P87 GTAWFillerMetal 0 P87 GMAW Filler Metal

....... _ERNiCr-3 (ref. 11)

10.0 100.0

Rupture Life (hours)

1000.0 10000.0

Figure 9. Time to rupture/or EPRI P87 GTAW and GMAWweld metal at 7600C (1400°F) compared to reported ERNiCr-3 rupture strength {11]

1000.0

: !. ., 100.0

i

10.0

·~

• P87 GTAW Weld Metal

• P87 GMAW Weld Metal

o Development Heats, P87 SMAW, Weld Metal

-""'X-ecr ...... x ERNiCr-3 (ref. 11)

-*"- ..... 593•c L ~......... 1oo•c -B800h1S "' ..... r I -760Qhrs

I--- "' .. [~ ..... g

' .... x .... ~ 700"C ...._ .._ -7800hrs

18000 19000 20000 21000 22000 23000 24000 25000 26000

LMP (K,C=20)

Figure 10. LMP plotofEPRI P87 GTAW and GMAWrupture data (arrows indicate ongoing tests) compared to EPRI P87 SMAW [3] and reported ERNiCr-3 data [11]

4. Discussion

Tensile tests on EPRI P87 DMWs show the filler metal exceeds the minimum strength of both

Grade 91 and 347H stainless steel at room temperature. At elevated temperature, cross-weld tensile tests failed in either the 347H (lower temperatures) or in the grade 91 (higher temperatures). However, at room temperature, the strength of the EPRI P87, although above

the grade 91 minimum, was inferior to the base metals and failure occurred in the weld metal.

For higher strength alloys such as CSEF steel P92 or advanced stainless steels such as HR3C

(31 OHCbN}, the low ambient temperature strength of EPRI P87 may result in failure of a qualification test. However, as demonstrated by the elevated temperature data in Figure 5, the filler metal will have more than adequate strength at service temperatures, even for higher

345


tensile strength materials. Thus for EPRI P87 to be used in higher-strength joints, relief from current welding qualification rules by supplemental elevated temperature testing requirements may be needed.

Cross-weld creep-rupture results show that the behavior of subcritically heat-treated grade 91 joints made with EPRI P87 filler metal (GTA W, GMA W, and SMA W) are similar to crossweld tests of joints made with typical 9%CrMo V filler metals. At high-stresses, creep failure is observed in the base metal but with decreasing stress failure occurs in the fme-grained HAZ region of the grade 91. The data in Figure 6 suggest little difference between the various filler metals. However, the creep strength of the EPRI P87 far exceeds that of grade 91. In the case of a normalized and tempered (N&T) grade 91 weld made with EPRI P87, the large mismatch in creep strength could affect rupture results. Creep-rupture testing on N&T grade 91-EPRI P87 joint produced by SMA W exhibited fusion line failures almost independent of test condition [3]. Therefore, more cross-weld rupture testing is advisable prior to utilizing EPRI P87 in N&T CSEF steels.

Fusion Line

Figure 11. Micrograph of Grade 91-EPRI P87 FM Fusion Line region after completed stressrupture testing at 625°C-JOOMPa (rupture life 899.2 hours)

All weld metal tests on EPRI P87 GT A W and GMA W show excellent rupture ductility and consistent creep rates irrespective of welding process. A power-law creep exponent of n=8, which is in the typical range of many engineering alloys, appears to fit the data over the range of stresses and temperatures in this study. Creep-strength was as good or better compared to ERNiCr-3 with longer-term tests suggesting a small creep strength advantage for the EPRI P87 GTA W welds. Additionally, unlike the high Cr content in ERNiCr-3 which contributes to the formation of a carbon denuded region during service, the EPRI P87' s low Cr content, essentially matching that of Grade 91, results in a uniform structure near the weld fusion line. Figure 11 is a post stress-rupture test micrograph of the Gr. 91 and EPRI P87 (GT A W) weld after almost 900 hours at 625°C and 1 OOMPa. Clearly, if carbon migration has occurred, it is not at a level distinguishable in this optical micrograph (as opposed to the carbon free regions in the micrographs in Figure 2).

346


5. Conclusions

GT A W and GMA W cross-weld and weld metal tensile and creep-rupture tests were conducted on dissimilar metal welds between grade 91 and 347H using EPRI P87 filler metal. Evaluation of the test results show the elevated tensile strength of the weld metal exceeds that of either base metal at service conditions. For joining higher-strength alloys, welding qualification modifications may be necessary due to the low room temperature strength of the EPRI P87 filler metal, but the elevated temperature strength is adequate. Creep-rupture results show EPRI P87 has excellent creep ductility. The test data from this study were used to provide a simple representation of minimum creep-rate behavior over the range of temperatures and stresses tested. Rupture life was compared to ERNiCr-3 on both an isothermal and a time-temperature parameter basis. At high-stress, the EPRI P87 had equivalent rupture strength to ERNiCr-3. At lower-stresses, the data suggest the EPRI P87 produced by GTA W is slightly stronger than ERNiCr-3. Microstructural analysis confirmed little to no carbon denuded region was present in the grade 91 after stress-rupture testing. Overall, the tensile strength, creep behavior, rupture strength, lack of carbon migration, and physical properties show EPRI P87 filler metal is an excellent alternative to ERNiCr-3 for dissimilar metal welds.

Acknowledgements

Special thanks to D. Glanton and M.K. Havens for their assistance with the metallography.

References

[1] Dissimilar-Weld Failure Analysis and Development Program. EPRI, Palo Alto, CA: 1985-1989. CS-4252. [2] D.l. Roberts, R.H. Ryder, R. Viswanathan, "Performance of Dissimilar Metal Welds in Service," Journal of Pressure Vessel Technology, Vol. 107 August 1985. [3] Development of a New Nickel Filler for Dissimilar Metal Welds and Repair. EPRI, Palo Alto, CA: 2009. 1018991. [4] D. Gandy and K. Coleman. "Alternative Filler Materials for DMWs Involving P9l Materials," Proceedings to the EPRI Fifth International Conference on Advances in Materials Technology for Fossil Power Plants (Marco Island, FL, Oct. 3-5, 2007) [5] J.A. Siefert, J.M. Sanders, J.M. Tanzosh, W.F. Newell, J.P. Shingledecker. "Development ofEPRl P87 Solid Wire." Proceedings to WELDS2009 (Fort Meyers, FL, June 24-26, 2009). [6] "EPRI P87 Product Data Sheet," Metrode Data Sheet D87, 2006 [7] "Welding Qualifications for Advanced Alloys." Unpublished EPRI research, 2009. [8] 2007 ASME Boiler and Pressure Vessel Code. Section ll- Materials Cl 2007 The American Society of Mechanical Engineers. [9] D. Jandova, J. Kasl, V. Kanta. "Influence of Substructure on Creep Failure of P91 Steel Weld Joints." Proceedings: Creep & Fracture in High Temperature Components,:!"' ECCC Creep Conference, Apri/2I-23, 2009, Zurich, Switzerland.© 2009 DEStech Publication, Inc.l77-188. [10] T.H. Courtney. Mechanical Behavior of Materials. © 2000, McGraw-Hill Companies, Inc. [11] Incoloy® alloy 800H & 800HT®, SMC-047 ©Special Metals Corporations, 2004 (Sept 04).

347

gth Liege Conference : Materi~ edited by J. Lecomte-Beckers,

ng 2010 Kuhn.

IMPROVED CREEP AND OXIDATION BEHA VIOR OF A MARTENSITIC 9CR STEEL BY THE CONTROLLED ADDITION OF

BORON AND NITROGEN

Peter Mayr1.2, Ivan Holzer, Francisca Mendez-Martin2, Mihaela Albu3

, Stefan Mitsche3,

Vanessa Gonzalez4, Alina Agiiero4

1 Department of Materials Science and Engineering, Massachusetts Institute of Technology, 77 Massachusetts Av, MA 02139, USA

2 Institute of Materials Science and Welding, Graz University of Technology, Kopernikusgasse 24, 8010 Graz, Austria

3 Institute for Electron Microscopy, Graz University of Technology, Steyrergasse 17, A-80 I 0 Graz, Austria

4 Instituto Nacional de Tecnica Aeroespacial, Ctra. de Ajalvir Km 4, 28850 Torrej6n de Ardoz, Spain

Abstract

This manuscript gives an overview on recent developments of a martensitic steel grade based on 9Cr3W3CoVNb with controlled additions of boron and nitrogen. Alloy design by thermodynamic equilibrium calculations and calculation of boron-nitrogen solubility is discussed. Out of this alloy design process, two melts of a 9Cr3W3CoVNbBN steel were produced. The investigation focused on microstructural evolution during high temperature exposure, creep properties and oxidation resistance in steam at 650°C. Microstructural characterization of "as-received" and creep exposed material was carried out using conventional optical as well as advanced electron microscopic methods. Creep data at 650°C was obtained at various stress levels. Longest-running specimens have reached more than 20,000 hours of testing time. In parallel, long-term oxidation resistance has been studied at 650°C in steam atmosphere up to 5,000 hours. Preliminary results of the extensive testing program on a 9Cr3W3CoVNbBN steel show significant improvement in respect to creep strength and oxidation resistance compared to the state-of-the-art 9 wt. % Cr martensitic steel grades. Up to current testing times, the creep strength is significantly beyond the +20% scatterband of standard grade P92 material. Despite the chromium content of 9 wt. % the material exhibits excellent oxidation resistance. Steam exposed plain base material shows comparable oxidation behavior to coated material, and the corrosion rate of the boron-nitrogen controlled steel is much lower compared to standard 9 wt.% Cr steel grades, P91 and P92.

Keywords: 9-12% chromium steels, microstructural evolution, creep testing, oxidation testing

348


1. Introduction

In Japan and Europe, research and development works in the field of advanced martensitic creep resistant steels has a long history [ l-3 ]. The applications for such materials are in particular large diameter and thick section boiler components, such as main steam pipes and headers of ultra-supercritically (USC) operated power plants. The current target of ferritic steel development is set to a creep rupture strength of I 00 MP a after I 00,000 hours at 650°C and oxidation rates lower than for grade P9l steel.

Within the last years, the alloy design philosophy for martensitic steels strongly targets the enhancement of long-term microstructural stability [ 4 ]. Migration of lath and block boundaries, causing the coarsening of martensite laths and blocks by annihilation of excess dislocations, closely correlates with the onset of accelerated creep. A fine dispersion of M23C6 and MX particles, which exert a pinning force on migrating boundaries and dislocations, has been identified as main contributor in suppressing recovery of the martensitic microstructure. The work presented in this manuscript is following an extensive research program published by the National Institute for Materials Science (NIMS), Japan [4]. In detail, the NIMS alloy design concept follows stabilization of M23C6 carbides and grain boundaries by the addition of boron in combination with finely dispersed MX particles as a possible solution to stabilize the martensitic substructure. As basis for the systematically study of the influence of boron and nitrogen on long-term creep behavior, a Fe-0.08C-9Cr-3W-3Co-0.2V-0.05Nb (wt. %) steel with different concentrations ofboron and nitrogen was investigated [5,6].

The influence of boron on 9Cr3W3CoVNb steel was studied in the range of 0 to 140 ppm (parts per million) while nitrogen level was kept constant at a minimum between I 0 to 30 ppm [5]. The influence of boron additions on creep properties was studied by uniaxial-creep testing at 650°C at different stress levels. At stress levels higher than 100 MPa, no big difference in creep strength between OB, 0.0048B and 0.0092B steel was found. Only 0.0139B steel showed improved creep strength. While at lower stress levels, creep strength of the lower boron steels decreased significantly, 0.0139B steel showed a stable creep behavior. It was concluded that controlled addition of boron can delay the drop of creep strength in 9Cr3W3CoVNb steel. As a result of the boron series creep tests, a boron content of 140 ppm was identified as optimum for improved creep behavior. The combined effect of boron plus additional MX precipitation was studied by 9Cr3W3CoVNb test melts with a fixed boron content of 140 ppm and varying nitrogen levels of 15, 79 and 650 ppm [6]. Increasing the nitrogen level to 79 ppm resulted in further improvement of the creep strength of the base 9Cr3W3CoVNb-O.Ol40B steel. However, further increase of nitrogen to 650 ppm showed no beneficial effects and in turn reduced creep strength of the alloy. The main idea behind the NIMS alloying concept is to combine boron-strengthening with nitride-strengthening. However, excess addition of nitrogen in combination with boron causes the formation of boron nitrides (BN) during processing and heat treatments at elevated temperatures. Precipitation of boron nitrides offsets the beneficial effect of boron and nitrogen addition. Sakuraya et. al [7] showed in their work on BN formation in high Cr heat resistant steels that, a particular relationship between boron and nitrogen has to be kept to avoid

349


formation of coarse BN particles. From their experimental observations the authors derived a mathematic formulation of the solubility product ofBN.

Within the present work, the NIMS alloy design strategy is expanded by thermodynamic studies of the boron nitride solubility product. Slightly modified alloy chemistry is proposed and first results of creep and oxidation tests as well as the microstructural evolution during creep are presented.

1. Experimental

1.1. Alloy Design

Thermodynamic equilibrium calculations were performed using the software Matcalc [8-11] utilizing the thermodynamic database Fe-data6 [12]. The description of BN in the database was slightly modified, based on the work of Fountain et al. [13]. Alloy design was supported through a calculated equilibrium phase diagram for the 9Cr3W3CoVNbBN steel and a boron nitride and boride solubility diagram.

1.2. Materials

A 20 kg heat (Heat 1) of 9Cr3W3CoVNb steel with controlled additions of 120 ppm boron and 130 ppm nitrogen was produced by vacuum induction melting. The addition of boron and nitrogen was set to avoid the formation of large boron nitrides but still allow the precipitation of strengthening MX particles. For homogenization of Heat 1, the ingot (110 mm square) was forged to fmal dimensions of 50 mm square. A second heat (Heat 2) of a 9Cr3W3CoVNb steel with slightly reduced boron (110 ppm) and nitrogen (100 ppm) content was cast (100 kg) and rolled to 20 mm thick plates. The exact chemical composition in weight% (wt. %) for Heat 1 and Heat 2 is given in Table 1. The final quality heat treatment for both heats consisted of normalizing at 1150 oc for 1 hour followed by tempering at 770 oc for 4 hours.

Table 1: Chemical composition of9Cr3W3CoVNbBN test melts in wt.% (balanced Fe . Analysis c Si Mn Cr w Co V Nb 8 N

Heat 1 0.074 0.29 0.44 9.26 2.84 2.95 0.21 0.056 0.012 0.013 Heat2 0.090 0.30 0.51 9.26 2.92 2.88 0.20 0.050 0.011 0.010

In both heats, precipitation of BN particles was investigated. Material of Heat 1 was used for microstructural characterization and creep testing. Fractured creep samples were investigated by means of optical and electron microscopy to determine the microstructural evolution during creep. Heat 1 material was also used for oxidation testing in steam atmosphere. Heat 2 material was used for basic metallographic investigations and for additional creep tests.

1.3. Steam Oxidation Laboratory Testing

The schematic of the closed loop laboratory rig employed at INT A is shown in Figure 1. Prior to testing, laboratory air is displaced from the chamber by means of N2 which is kept flowing while heating up to the test temperature (approximately 600°C/h). Once the test

350


temperature is reached, the N2 flow is closed and pure steam is introduced at a linear velocity of 8 cm/s. To carry out weight measurements or to remove samples, the furnace is cooled to about 300°C under N2 atmosphere and the specimens are subsequently removed. The reheat cycle is also carried out under N2 atmosphere.

-PUII.GI:

Figure 1: Sketch of the closed loop laboratory test rig for the oxidation testing of samples in steam atmosphere at 650°C.

1.4. Microstructure Characterization

Base material microstructure was characterized by light optical microscopy, scanning electron microscopy (SEM) and analytical transmission electron microscopy (TEM). Precipitation of BN was studied by SEM and energy dispersive X-ray spectroscopy (EDS) on fracture surfaces of base material specimens. The smallest BN particle size being able to identify is about 0.3 J.Lm. Details are reported in [14]. The oxidized specimens were characterized by optical and field emission scanning electron microscopy (FESEM) using a JEOL JSM 840 system equipped with an energy dispersive X-ray spectrometer (EDS) KEVEX MICROANALYST 8000 with a RONTEC signal processor.

2. Modified Alloy Design Concept

The investigated alloys are based on a 9Cr3W3Co VNbBN steel. The calculated equilibrium phase diagram for the chemical composition equal to that of Heat l (see Table l) indicates that the steel should be free of delta ferrite at room temperature. The nominal chromium content of 9 wt.% enables a complete austenitization on cooling in the temperature range from 1183 oc to 881 °C. Normalizing at ll50°C will therefore dissolve most of the precipitates

351


formed during solidification and allow a homogenization of the elemental distribution through diffusion. On cooling, a uniform martensitic microstructure will form which then is tempered at 770°C.

Figure 2: Calculated equilibrium phase diagram with varying chromium content for a 9Cr3W3CoVNbBN steel with a chemical composition equal to that of Heat I .

To take advantage of boron and nitrogen addition, the formation of large boron nitrides (BN) and borides (M2B) has to be avoided. Therefore, in the present work the allowable boron and nitrogen limits for high chromium martensitic steels were determined by thermodynamic equilibrium calculations. The BN solubility product for a 9Cr3W3CoVNbBN steel, based on the chemical composition of Heat 1, was calculated for various BIN ratios and different normalizing temperatures, including also possible formation of borides (M2B). Similar to boron nitrides, boride particles would also reduce the amount of boron dissolved in the metal matrix or incorporated in chromium carbides (M23C6) and therefore diminish the beneficial effect of boron addition. Results of the BN solubility calculation are shown in Figure 3. The solubility products (solid lines) are shown for different normalizing temperatures ranging from 900°C to l250°C. The dashed grey lines represent the solubility product for the M2B phase. For reason of comparison, the BN solubility lines reported by Fountain et al. [13] at ll50°C and Sakuraya et al. [15] are included in the figure. The grey shaded area in Figure 3 indicates critical BIN ratios e.g. at l250°C where BN formation is to be expected. No formation of BN is expected below the respective BN solubility line. Figure 3 clearly shows that the experimental fmdings reported by Sakuraya et al. [15] are well described by the thermodynamic equilibrium considerations in the temperature range between 900°C and 1250°C. Sakuraya et al. [7] reported no BN formation for temperatw:es higher than 1250°C. Due to the temperature dependence of the BN solubility product it can be assumed that coarse BN (> IJ!m) exist for high boron and nitrogen contents and low normalizing temperatures.

352

101

10°

eft. 10'1

~ - 10'2

c: 0 ~

0 CO 10-3

10"'

10-5


---- ··-----------------<>- BN solubility line ~ Heat 1 ® no BN (<0.21Jrn) • coarse BN (>11Jm)

-a. M,B solubility line ~ Heat 2 () small BN (<1~m)

10'3 10'2

Nitrogen I wt.%

Figure 3: BN solubility product for a 9Cr3W3CoVNbBN steel as a function of normalizing temperature in the range of 900°C to 1250°C. The grey shaded areas indicate BN formation. The calculation is compared to the solubility products reported by Sakuraya et al. [15], and

Fountain et al. (13], and the two heats investigated within this work.

3. Microstructure of9Cr3W3CoVNbBN steel

Optical micrographs of 9Cr3W3CoVNbBN base material (Heat I) show a tempered martensitic microstructure (see Figure 4). The average prior austenite grain size is 250 J.llll in diameter, and precipitates are finely distributed mainly along prior austenite grain and martensite lath boundaries. Via electron microscopy, MX (V,Nb)(N,C) carbonitrides and M23C6 (Cr,Fe,Mo,W)2JC6

carbides have been identified at grain and lath boundaries and/or in the matrix of the base material. MX {V,Nb){N,C) particles have a representative diameter smaller than lOO nm, and Cr-rich carbides have a diameter between 150 and 250 nm. By SEM and EDS only very few and small {<1~m) BN particles for Heat 1 and no BN particles for Heat 2 have been detected. A detailed description of the method of detecting BN particles is given in [15]. These experimental findings support the results of the thermodynamic equilibrium calculations in the alloy design stage.

353


Figure 4: Optical micrograph of normalized and tempered 9Cr3W3CoVNbBN base material (left) and small BN particle (right) identified by EDS- both images for Heat 1.

After creep exposure for 8,971 hours at 650 oc and 130 MPa, light-optical and electronmicroscopic investigations have been performed to study the microstructural evolution during creep exposure. The prior austenite grain and martensite lath structure, after almost 10,000 h of creep, show no significant changes. The precipitate size and number density of carbonitrides and carbides remain almost the same as in the base material. The electron microscopy investigation reveals the precipitation of two new phases: laves phase (intermetallic phase A2B, (Fe,Cr)2(W,Mo) with approximately 38 atom % of tungsten and more than 45 atom % of iron, distributed mainly at the grain boundaries, and modified Zphase (complex nitride [(Cr,V,Nb)N]) with a very low phase fraction. Details on the microstructural evolution of 9Cr3W3Co VNbBN steel during creep exposure can be found in [ 16 ]. Based on this study, an enhanced microstructural stability can be attributed to 9Cr3W3CoVNbBN steel.

4. Oxidation behavior

Specimens of9Cr3W3CoVNbBN (Heat 1) were exposed to pure, flowing steam at 650°C for 4,908 h. Figure 5 shows the mass variation of the new steel as function of time as well as that of9 wt.% ferritic steels P91 and P92 for comparison purposes. 9Cr3W3CoVNbBN exhibits a mass gain considerably lower than that of the other two 9 wt. % Cr steels. This behavior is surprising, as work done by several research groups have indicated the steam oxidation resistance of steels strongly correlates with the Cr contents, and results so far have indicated that 9 wt.% Cr steels oxidized very fast when exposed to steam at T > 600°C [17-20]. For instance, Quadakkers and collaborators have concluded that a chromium content of 11 wt. % or more is required to form a protective scale [17].

354

40


....... Substrate P92

N 35 1 ...... Substrate P91

E 30 ....... Substrata NPM ..!:! .r 25

& 20 c I'G -5 15

= 10 ::E

5 0~-+~~~+---~~--~~~~--~~--

0 1000 2000 3000 4000 5000

Time (h)

Figure 5: Mass gain of conventional (P91, P92) 9 wt.% steel and 9Cr3W3CoVNbBN (NPMl) substrates exposed to flowing steam (1 bar) at 650° C

Figure 6 shows an image of a specimen exposed to 650°C hot steam for 4,908 h. Several nodular features can be observed, especially at the samples edges but most of the surface still appears to be smooth. On the cross section of the exposed surface, a very thin(< 100 nm) Cr and 0 rich layer, most likely Cr20 3 can be observed with difficulty. This type of oxide is typical of steels containing ll wt. % of Cr or more. Quadak:kers found that Co tends to increase the oxidation resistance of 9-12 Cr wt. % ferritic steels [ 17], but not to the extent observed for 9Cr3W3CoVNbBN. On the other hand, other groups have observed that the higher the W content, the lower the oxidation resistance of these types of steels [ 18,21,22]. It is therefore not clear yet why this 9 wt. % steel forms a protective oxide layer and exhibits such high steam oxidation resistance. Further studies are required.

Figure 6: SEM image of a 9Cr3W3Co VNbBN specimen exposed to steam at 650°C for 4,908 h

355


W rich precipitates

Figure 7: FESEM image ofthe cross section of9Cr3W3CoVNbBN exposed to steam at 650°C at a nodule site: (a) for 1,260 h, (b) for 4,908 h

The cross section of representative nodules formed after 1,260 and 4,908 h are shown in Figure 7a and b respectively. The microstructure corresponds to a dual oxide layer similar to that observed on ferritic steels containing 9 wt.% in Cr, with a top outwards growth of a Fe304 layer and an inside growing mixed oxide layer containing (Fe, Cr)J04, FeO and Cr20 3 [23] which also contains W, Co and V. Interestingly, W rich nanoscale precipitates which are also present in the substrate can be observed in this inner oxide layer. The nodules in the 1,260 h specimen exhibit lower thickness and width than those observed in the specimen exposed for 4,908 h, indicating that these oxides are growing and therefore are not protective. Moreover, the number of nodules on the 4,908 h specimen is higher than that observed on the 1 ,260 h specimen.

Figure 8: Cross section of the edge of a 9Cr3W3CoVNbBN specimen exposed to steam at 650° C for 4,908 h

At one specimen's edge (Figure 8), the corresponding EDS composition map reveals Cr-rich layers within the inner oxide and the original specimen surfaces. The average Cr content (19 wt. %) in this inner oxide zone is significantly higher than that of the substrate, but because it is not uniformly distributed it is not of a protective nature. According to Quaddakers and coworkers, this pattern results from the repetition of cycles in which Cr rich spinels form due to rapid Cr diffusion from the bulk alloy until the spinel cannot be sustained due to continuous

356


Cr depletion, resulting in the formation of Fe rich oxides [I7]. Formation of oxidation nodules has been observed on oxidation resistant 12 wt. % steels and has been attributed to compositional inhomogeneity [ 17,24]. For this 9Cr steel the presence of nodules can be rather attributed to local cracks, imperfections and/or rupture of the protective scale on a Cr depleted substrate. The fact that the samples edges are fully covered with nodules support this hypothesis as due the higher triaxiality of stresses, corners and edges are more prone to develop cracks.

5. Creep Strength

Uniaxial creep tests of 9Cr3W3CoVNbBN base material specimens have been carried out at 650°C. Both Heat 1 specimens tested at I30 MPa and I 00 MPa have already fractured, as shown in Figure 9. The base material specimen tested at a higher stress level of 130 MPa fractured after 8,97I h. This corresponds to an improvement of rupture stress at approximately 9,000 hours of +60 MPa and +40 MPa compared to grades P9l and P92 respectively. After 23,997 hours of creep exposure, creep rupture strength of 9Cr3W3CoVNbBN steel is still beyond the +20% average creep strength of grade 92 material. Heat 2 material tests are still running but up to almost I 0,000 hours of testing, creep strength is superior to that of standard grade 92 material by more than +20%.

(ij' a.. ~ Cll Cll

~ U5

200

150

100

50

···· ························ ....

·······.. o--··- .. __ _

-----·-

-·--- ........

- Heat-1

o.-.• ------ I a- • ·-. __ _ •

···---- ... __ _

~

9Cr-3W-3CoVNbBN - 650°C

--mean P92 ECCC ········· P92 +20% -·-·-·mean P91 VdT-V • 9Cr3VV3CoVNbBN (Heat 1)

·---.

o 9Cr3VV3CoVNbBN (Heat 2) all tests running 20~==~~~~~~==~~~~~==~==~~~~

100 1,000 10,000 100,000

Time [h]

Figure 9: Creep testing data of9Cr3W3CoVNbBN base material of Heat I and Heat 2 at 650°C compared to creep rupture data of well-established grades P91 and P92.

357


6. Summary and Conclusions

Alloy design for a 9 wt. % Cr steel has been successfully improved by thermodynamic studies of the boron-nitrogen solubility product. The boron and nitrogen content of a 9Cr3W3CoVNbBN steel was adjusted to have either only very small boron nitrides (Heat 1) or no precipitation of BN (Heat 2). This confirms that in these two heats most of the boron stays either in solution or is incorporated in M23C6 carbides. Therefore, the added boron can develop its maximum strengthening contribution.

The boron strengthened 9Cr3W3CoVNbBN steels exhibits an unexpected high resistance to steam oxidation at 650°C. Despite its low Cr content it develops a thin Cr rich protective oxide scale. However, several oxidation nodules form, very likely due to imperfections in the protective oxide layer. These nodules grow in thickness, width and number as a function of steam exposure time. Nevertheless, the increased oxidation resistance of the new steels shows high potential and is currently studied in detail.

The creep behavior up to testing times of almost 25,000 hours at 650°C is very promising. The 9Cr3W3CoVNbBN steel shows an improved creep rupture strength more than +20% beyond the average creep strength of standard grade 92 steel. This higher creep rupture strength is attributed to an enhanced microstructural stability due to the controlled addition of boron and nitrogen. Soluble boron stabilizes the martensitic microstructure and delays recovery whereas boron incorporated in M23(C,B)6 carbides reduces their tendency to coarsening.

Acknowledgements

PM would like to thank the Austrian Academy of Sciences and the Max Kade Foundation for granting a one-year research fellowship. The authors would like to acknowledge the support by the European Cooperation in Science and Technology (COST) through Action 536 "Alloy development for Critical Components of Environmental friendly Power planT (ACCEPT)" and the Austrian Research Promotion Agency - FFG (#818512). AA and VG wish to acknowledge the Spanish Ministry of Science and Innovation for the financial support (ENE2008-06755-C02-0 1 ).

References

[I] Kern TU, Wieghardt K and Kirchner H, in 4th int conf Advances in Materials Technology for Fossil Power Plants, Hilton Head Island, ASM International, 2005. [2] Masuyama F, in 4th int conf Advances in Materials Technology for Fossil Power Plants, Hilton Head Island, ASM International, 2005. [3] Cerjak, H.-H.; Dimmler, G.; Holzer, 1.; Kozeschnik, E.; Mayr, P.; Pein, C.; Sonderegger, B.: Materials science forum 539-543 (2007) , S. 2954 - 2959 [4] Abe F, in 4th int conf Advances in Materials Technology for Fossil Power Plants, Hilton Head Island, SC, EPRI, 2004. [5] Abe F, Horiuchi T and Sawada K, Materials Science Forum, 2003,426-432, 1393-1398. [6] Semba Hand Abe F, "in 4th int conf Advances in Materials Technology for Fossil Power Plants, Hilton Head Island, SC, EPRI, 2004.

358


[7]Sakuraya K., Okada H., Abe F., ISIJ Int., vol. 46 (2006), 1712. [8]Kozeschnik E., Buchmayr B., Mathematical Modelling of Weid Phenomena 5, Institute of Materials, London, 2001, 349. [9]Svoboda J., Fischer F. D., Fratzl P., Kozeschnik E., Mater. Sci. Eng. A, vol 385 (2004) 166. [IO]Kozeschnik E., Svoboda J., Fratzl P., Fischer F. D., Mater. Sci. Eng. A, vol 385 (2004) 157. [11]Kozeschnik E., Svoboda J., Fischer F. D., CALPHAD, vol28 (2005) 379. [12]Saunders M., Fe-data6.tdb. [13] R.W. Fountain, Chipman J., Transactions of the Metallurgical Society of AIME, vol. 224 (1962) 599. [14]Mayr P., PhD-Thesis, Graz University of Technology, 2007 [15]Sakuraya K., Okada H., Abe F., in: Advances in Materials Technology for Fossil Power Plants: Proceedings of the Fourth International Conference, (2004) 1270. [16] Mayr, P.; Mendez Martin, F.; Albu, M.; Cerjak:, H.-H.: in: Creep and Fracture in High Temperature Components- Design & Life Assessment Issues (2009), S. 1029- 1037 [17]Quadakkers W.J. and Ennis P.J., Materials for Advanced Power Engineering 1998, Part I (1998) p. 123 [18]Nava-Paz J.C. and Knodler R., Materials for Advanced Power Engineering 1998, Part I (1998) p. 451 [19]Fukuda Y., Tamura K. and Sato T., Materials for Advanced Power Engineering 1998, Part I (1998) p. 461 [20]Ennis P.J. and Quadakkers W.J., Materials for Advanced Power Engineering 2002, Part 11 (2002) p. 1131 [2l]Knodler R. and Scarlin B., Materials for Advanced Power Engineering 2002, Part Ill (2002) p. 1601 [22]Agiiero A. and Muelas R., Materials Science Forum, 461-464 (2004) p. 957 [23]Zurek J., Nieto-Hierro L., Piron-Abellan J., Niewolak: L., Singheiser L. and Quadakkers W.J., Materials Sceince Forum, 461-464 (2004) p. 791 [24]Lepingle V., Louis G., Pete1ot D., Lefevre B. and Vandenberghe B., Materials Science Forum, 461-464 (2004) p. 1039

359

Documents

1 edited by J. Lecomte-Bec ~:OE024913649* dB. Kuhn. HIGH