Microstructural aspects of superior creep resistance of a 10%Cr martensitic steel

Author’s Accepted Manuscript Microstructural Aspects of Superior Resistance of a 10%Cr Martensitic Steel

Creep

R. Mishnev, N. Dudova, A. Fedoseeva, R. Kaibyshev www.elsevier.com/locate/msea

PII: DOI: Reference:

S0921-5093(16)31177-7 http://dx.doi.org/10.1016/j.msea.2016.09.096 MSA34181

To appear in: Materials Science & Engineering A Received date: 18 August 2016 Revised date: 22 September 2016 Accepted date: 23 September 2016 Cite this article as: R. Mishnev, N. Dudova, A. Fedoseeva and R. Kaibyshev, Microstructural Aspects of Superior Creep Resistance of a 10%Cr Martensitic S t e e l , Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2016.09.096 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Microstructural Aspects of Superior Creep Resistance of a 10%Cr Martensitic Steel

R. Mishnev, N. Dudova, A. Fedoseeva, R. Kaibyshev

Belgorod State University, Belgorod 308015, Russia 

Corresponding author. Tel.: +7 4722 585417; fax: +7 4722 585417 E-mail address: [email protected] (N. Dudova)

Abstract

The microstructural evolution and the dispersion of secondary phases were studied in a lownitrogen 10%Cr martensitic steel with 3% Co and 0.008% B additives at 650C under an applied stress of 140 MPa. It was demonstrated that the superior creep strength of this steel can be attributed to the high resistance of M23C6 – type carbides and Nb-rich MX carbonitrides against coarsening, resulting in a stable of the tempered martensite lath structure (TMLS) under short-term creep conditions. The TMLS remains slightly changed under creep: lath coarsening occurs with a twofold decrease in the lattice dislocation density. M23C6–type carbides were found to give the main contribution in hindering the transformation of interlath boundaries to subgrain boundaries, impeding the migration of low-angle boundaries by exerting a large pinning pressure. A high Zener drag pressure is maintained up to rupture. The precipitation of a Laves phase under creep conditions results in a minor contribution to the overall pinning pressure. V-rich MX carbonitrides tend to dissolve with increasing time. No formation of a Z-phase was detected. M23C6 carbides retain their orientation relationship with ferritic matrix up to rupture. No significant strain-induced coarsening of M23C6 carbides, Laves phase, or MX carbonitrides was observed.

Keywords: martensite; steel; electron microscopy; phase transformation; precipitation; coarsening. 1. Introduction

1

High-chromium creep-resistant martensitic steels are widely used for critical components of fossil power plants that operate at temperatures up to 620°C due to their excellent creep resistance [1,2]. The heat treatment of these steels consists of normalizing from T≥1050C and tempering at temperatures ranging from 750 to 770C for 2–4 h [1-3]. The tempered martensite lath structure (TMLS) evolved from this treatment is composed of a hierarchical sequence of structural elements, i.e. prior austenite grains (PAG), packets, blocks, and laths with a high density of free dislocations [1,2,4]. Two types of carbides precipitate during tempering [1,2,5-7]. The major portion of M23C6 carbides with an average size of ~100 nm is located at high-angle boundaries of PAGs, packets, and blocks, while a minor portion precipitates at interlath boundaries of low-angle origin [1,3,5,8]. Nanoscale MX carbonitrides (where M is Nb or V and X is C or N) are distributed homogeneously throughout the ferritic matrix [1-3,5-8]. The superior creep strength of 9-11% Cr steels is attributed to the stability of TMLS under creep conditions [1,2]. Interlath boundaries play a key role in the creep strength of these steels [9,10]. The concept of alloying high-Cr heat-resistant steels aims to provide stability of the lath structure under creep conditions [1,2] by suppressing two sequential processes of transformation of TMLS. First, the knitting reactions between lattice dislocations and dislocations composing interlath boundaries have to be slowed down by MX carbonitrides and such substitutional atoms as W, Mo, Co and Cr, which hinder dislocation slip within the ferritic matrix and the rearrangement of lattice dislocations by climb [11,12]. As a result, interlath boundaries may be retained as irregular dislocation networks exerting long-range stress fields [13]. The transformation of interlath boundaries to subboundaries relieves internal elastic stresses and promotes their migration: this is the main reason for creep strength breakdown in the rupture timedependence of applied stress [10-19]. The secondary-phase distribution can evolve in two ways that facilitate the knitting reaction [12,20,21]. First, under long-term creep conditions, the nanoscale V-rich MX dispersoids are replaced by micron-scale Z-phase particles, which are, in fact, a thermodynamically stable nitride [22-24]. MX carbonitrides are not susceptible to significant coarsening due to the two-phase

2

separation into Nb- and C-rich MX and V- and N-rich MX particles [1,5,7,8,12,13,20,21,24,25]. The formation of the Z-phase reduces the efficiency of MX carbonitrides as pinning agents for dislocation glide and climb. Second, the depletion of W and Mo from the solid solution and their precipitation in the form of the Laves phase [1,3,12,26-28] mainly accelerates the diffusion process in the ferritic matrix, thus promoting dislocation climb and the knitting reaction. The creep strength breakdown correlates with the depletion of W from the solid solution [12] and/or is associated with the formation of coarse Z-phase particles at the expense of nanoscale MX carbonitrides [22-24]. The knitting reaction leads to the transformation of interlath boundaries to subgrain boundaries, followed by their migration and full replacement of TMLS by a well-defined subgrain structure [1,8,10,12,20]. This is second process of evolution of TMLS. Suppression of subboundary migration occurs as a result of Zener pinning pressure exerted by a dispersion of secondary-phase particles [8,12,20,21,29]. Boundary M23C6-type carbides are the major contributors to Zener pinning pressure [12,20,21,29]. In addition, precipitation of the Laves phase at boundaries during creep may cause an increase in Zener drag pressure [26-30]. MX carbonitrides make an insignificant contribution to the Zener drag force [1,2,12,20,21]. The drop of pinning pressure due to the coarsening of M23C6 carbides and Laves phase particles below 0.12 MPa leads to the complete transformation of TMLS to the subgrain structure and the onset of creep strength breakdown [8,9,12,20,21,27-30]. K. Maruyama et al. [10,15,16,31] attribute creep strength breakdown to the loss of particle pinning, mainly due to the coarsening of boundary M23C6 carbides. Recently, a new approach to microstructural alloying was suggested by the National Institute for Materials Science (NIMS), Japan. They proposed that the creep strength breakdown could be suppressed or shifted to higher rupture times by lowering the N content and increasing the B content in 9-11%Cr steels [5,32-35]. This approach allows an increase of the creep lifetime by a factor ranging from 10 to 40 compared with P92-type steel at the same applied stress. This finding is due to the strong effect of the N and B contents on the dispersion of secondary phases [32,33,35].

3

Carbon cannot replace nitrogen in the Z-phase and therefore, the lowering of the N content suppresses the formation of this phase. In addition, the high boron content provides a replacement of carbon by boron in M23C6 carbides, and the two-phase separation to B-free M23C6 carbides and the M23(BC)6 phase takes place [36,37]. The M23(B·C)6 phase is highly resistant to coarsening as compared to the B-free M23C6 carbides, due to the decreased energy of Fe(110) || M23(C,B)6(111) interfaces [36], in accordance with the well-known Gibbs–Thomson scheme [38]. However, the evolution of TMLS and the dispersion of secondary phases have not been reported to date in these types of high-Cr steels. Lowering of the N content leads to a significant decrease in the volume fraction of MX carbonitrides. However, the factors responsible for the suppression of the knitting reaction are unclear. The aim of the present work is to investigate the microstructural aspects of the superior creep resistance of a 10%Cr steel with 3% Co, 0.008% B, and 0.003% N additives under short-term creep at 650C. The creep behavior and microstructural evolution of this steel are compared with those of a 3%Co modified P92 steel [12,20].

2. Experimental procedure A steel with the following chemical composition (in mass%) was examined: 0.1% C; 0.06% Si; 0.1% Mn; 10.0% Cr; 0.17% Ni; 0.7% Mo; 0.05% Nb; 0.2% V; 0.003% N; 0.008% B; 2.0% W; 3.0% Co; 0.002% Ti; 0.006% Cu; 0.01% Al and Fe balance. This material is herein denoted 10% Cr steel. The 10%Cr steel was solution treated at 1060C for 0.5 h, cooled in air, and subsequently tempered at 770C for 3 h. Flat specimens with a gauge length of 25 mm and cross-sectional dimensions of 7 mm × 3 mm were subjected to creep tests to various strains and until rupture under an applied stress of 140 MPa at a temperature of 650C to examine the microstructural evolution under long-term aging in grip sections and under creep in gauge lengths. In addition, the creep rupture tests were carried out under applied stresses of 120, 160 and 180 MPa at the same temperature.

4

Structural characterization was performed using a Jeol JEM-2100 transmission electron microscope (TEM) with an INCA energy dispersive X-ray spectrometer and using the Z-contrast technique with a Quanta 600FEG scanning electron microscope (SEM) equipped with an electron backscatter diffraction (EBSD) pattern analyzer incorporating an orientation imaging microscopy (OIM) system. Low and high angle boundaries (LABs and HABs) were defined to have a misorientation of 2º<θ15º and >15º and depicted in misorientation maps using white and black lines, respectively. The transverse lath/subgrain sizes were measured on the TEM micrographs by the linear intercept method, counting all the clear visible (sub)boundaries. The dislocation densities were estimated by counting the individual dislocations in the (sub)grain/lath interiors per unit area on at least six arbitrarily selected typical TEM images for each data point. The density of particles located at (sub)grain/interlath boundaries was determined as the number of particles per unit boundary length. Other details of structural characterization were reported in previous works [5,12,20]. The equilibrium volume and mole fractions of phases and their chemical compositions were calculated by the version 5 of the Thermo-Calc software using the TCFE7 database by entering the BCC A2, FCC A1, M23C6 and LAVES PHASE C14 as equilibrium phases for the actual steel composition. To estimate the coarsening of the M23C6 and Laves phase particles the calculation was performed by the TC-Prisma software using the kinetic MOBFE1 and thermodynamic TCFE6 databases for the composition of (in mass%): 0.1% C; 10% Cr; 3.1% Co; 0.7%; Mo; 2.0% W and Fe balance.

3. Results 3.1 Tempered martensite lath structure Characteristics of microstructure and distribution of secondary-phase particles in the 10% Cr steel in the tempered condition were considered in previous works [6,35,37] in some details and are summarized in Table 1. The misorientation distribution of interlath boundaries showed two peaks at

5

approximately 2 and 6 (Fig. 1(a)). “Bain circles” appeared on a 001 pole figure (Fig. 1(a)) reconstructed from the separate PAG, indicating that an austenitemartensite transformation followed the Bain, Kurdjumov – Sachs and Nishiyama – Wassermann orientation relationships took place in the steel [39]. Nanoscale M23C6–type carbides with an average size of about 60 nm locate on the interlath boundaries and the coarse M23C6 carbides (about 100 nm) locate on high-angle boundaries of PAGs/packets/blocks/laths (Figs. 1(b),(c)). These carbides comprise almost continuous chains along the boundaries of PAGs, packet, blocks and laths. High-dense interlath carbides is a distinct feature of the 10%Cr steel despite the fact that some interlath boundaries are almost free of these carbides (Fig. 1(b)).The B-free M23C6 and B-rich M23(BC)6 phases were detected (Fig. 1(d)) [6,37]. Content of W in the B-free M23C6 carbides is about 25% higher than in the B-rich M23(BC)6 particles (Fig. 1(d)). The Thermo-Calc calculations predict the mole fraction of 2.1% and 0.28% of these phases, respectively. Therefore, the increased B content from 0.005% to 0.008% provided +50% increase in the mole fraction of B-rich M23(BC)6 phase predicted by Thermo-Calc calculation in comparison with the 3%Co modified P92 steel [20]. Nb- and V-enriched MX-type carbonitrides with sizes of 30 nm and 40 nm are uniformly distributed within the martensite laths (Fig. 1(e)). The mole fraction of the V-rich MX phase (0.0036%) is ten-fold smaller than that of the Nb-rich one (0.063%) due to a very low N content, as calculated by Thermo-Calc at 770C and confirmed by TEM analysis of replicas. Lowering of the N content from 0.05 mass.% to 0.003 mass.% leads to a decrease in the mole fraction of V-rich MX carbonitrides by a factor of ~120 compared with the 3%Co modified P92 steel [12,20], in which only one V-rich MX phase is predicted by Thermo-Calc. The total mole fraction of the MX phase in the 10%Cr steel is approximately 6.5 times smaller than that in the 3%Co modified P92 steel. Vrich MX particles contain approximately up to 70 mass.% V and 30 mass.% Nb, and Nb-rich MX particles contain approximately up to 80 mass.% Nb and 10 mass.% V (Fig. 1(f)). Therefore, the two-phase separation of MX carbonitrides to the Nb-rich and V-rich phases occurs. However, the lowered N content highly decreases the volume fraction of V-rich carbonitrides and increases the 6

proportion of V in Nb-rich MX dispersoids compared with steel with essentially the same chemical composition but distinguished by a higher (0.05%) N content [12]. Scarce, fine W-rich M6C carbides with an average dimension of 25 nm were found (Fig. 1(e), (f)) whereas no Laves phase precipitates were observed. 3.2 Creep behavior Figure 2(a) shows the creep rupture data of the 10%Cr steel compared with a P92-type steel and the 3%Co modified P92 steel [12]. The 10% Cr steel exhibits no creep strength breakdown at a temperature of 650C up to an applied stress of 120 MPa and a corresponding rupture time of 39 437 h, in contrast to 9%Cr steels containing 0.05%N. At 120 MPa, the creep lifetime of the 10%Cr-modified steel is higher by a factor of 8 compared with the 3%Co modified P92 steel, due to the lack of creep strength breakdown. At 140 MPa, the creep lives of the 10%Cr steel and the 3%Co modified P92 steel are nearly the same [12]. Therefore, creep under an applied stress of 140 MPa occurs in the short-term region for the 10%Cr steel, as for the 3%Co modified P92 steel [12]. The  vs time and  vs strain curves of the 10%Cr steel are shown in Fig. 2 (b) and (c), respectively. The contribution to creep deformation incurred by the transient creep strain, T, creep strain accumulation at the minimum creep rate ( min ×tr, where tr is rupture time), and tertiary creep strain, 3, [40-42] is shown schematically in Fig. 2(c). The transient creep behavior is described by the empirical relationship suggested by Garofalo [40,41]:

   0   T [1  exp( r '  t )]  s  t ,

(1)

where  is the strain,  0 is the instantaneous strain on loading, r’ is the rate of exhaustion of transient creep,  s is the steady-state creep rate that can be replaced by the min value [41], and t is time. The concept of the rearrangement of dislocations into a stable configuration over the relaxation time, , during creep deformation was invoked to provide a physical understanding of Eq.1 [41]. In this concept, the first-order reaction-rate kinetics is described as follows: d  (  S )  dt  7

(2)

For Eqs. 1 and 2, the following dependence is fulfilled: 1/= r’=K×  s , where K is a constant [41]. In the 10%Cr steel, the decrease in strain rate with time under transient creep occurs at a low rate compared with the 3%Co modified P92 steel [12,20] and requires significantly higher strain (Fig. 2(b) and (c)). The transient stage (2.3%) is more extended in the 10%Cr steel than in the 3%Co modified P92 steel [12,20], by a factor of 3 (Fig. 2(c), Table 1). The rate of exhaustion of transient creep, r’, evaluated graphically as the slope of the plot of ln(1   /  T ) vs. time, where  is the transient creep component (      0  min  tr ) [41], is 210-2 h-1. This value is -50% lower than that of the 3%Co modified P92 steel (310-2 h-1). On the other hand, the creep rate is described by  ~ A exp(b ) [18] and the transient stage of the 10%Cr steel is characterized by the parameter

d ln  / d 145, which is 3 times smaller than that in the 3%Co modified P92 steel (449) [12,20] (Fig. 2(c)). This finding can be attributed to the low rate of rearrangement of dislocations. The 10%Cr steel exhibits a minimum creep rate, min , instead of steady state that is typical for martensitic steels. The apparent region of steady-state creep, min ×tr, can be identified for convenience since it is difficult to accurately determine the time to the onset of the tertiary stage [42,43] (Fig. 2(b), (c)). It should be noted that the 3%Co modified P92 steel shows more defined steady-state creep stage than the 10%Cr steel [20] (Fig. 2(b),(c)). The offset strain, min~3.5%, at which the minimum creep rate is attained, is higher by a factor of ~3 than that in the 3%Co modified P92 steel [12,20] (Table 1), whereas the initial creep rate and the lowest creep rate are almost the same. Thus, the transient creep behavior of the 10%Cr steel shows a three-fold increase in the transient creep strain and the offset strain compared with the 3%Co modified P92 steel [12,20]. This finding is attributed to the three-fold reduction in the parameter, d ln  / d . Under tertiary creep, the variation of the creep strain with time is described by the following relationship [42,43]:

   0   T [1  exp(r '  t )]  s  t   3 exp[ p(t  t r )] ,

8

(3)

where p is the rate of acceleration of tertiary creep and tr is the time to rupture. In the 10%Cr steel, the transition from steady-state creep to tertiary creep occurs at a strain of ~5.1%, which is similar to that of the 3%Co modified P92 steel (~4.6%) [12,20] (Table 1). Tertiary creep strain, 3, of 10.8% is +50% larger than that in the 3%Co modified P92 steel [12,20] (7.1%). The rate of acceleration of tertiary creep, p, evaluated graphically as the slope of the plot of ln( /  3 ) vs. (t-tr), where  is the tertiary creep component (      0   T  min  tr ) [43], is 310-2 and 1.510-2 h-1 for the 10%Cr steel and the 3%Co modified P92 steel, respectively. The absolute value of the parameter d ln  / d 90 in the tertiary stage decreases compared with that in the transient stage (145) (Fig. 2(c)). This finding is indicative of the significantly slow kinetic processes for tertiary creep compared with transient creep. Thus, the tertiary creep behavior of the 10%Cr steel is typical of the short-term creep of high-Cr steels and is characterized by a low rate of recovery processes. Lowering the N and increasing the B contents significantly change the transient and steady-state creep behavior and slightly affect tertiary creep behavior.

3.3 Fractography The fracture surface of the crept specimen is shown in Fig. 3. Dimple rupture in transgranular manner was observed. A number of small dimples with diameter ranging from 1 to 2 m is indicative for the activation of numerous nucleating sites. Larger dimples with an average dimension of 10 m were nucleated by particles. Markings of serpentine glide were observed on their walls (Fig. 3) that is indicative for considerable plastic deformation during fracture. Therefore, the formation of coarse particles may promote premature fracture.

3.4 Stability of TMLS under long-term aging and short-term creep Microstructural evolution in the 10%Cr steel was investigated in the creep-interrupted specimens A, B, and C (Fig. 2(b), (c)), corresponding to the onset of apparent steady-state creep (278 h), the offset creep rate (780 h), and the onset of tertiary creep (1286 h), respectively, as well 9

as after rupture (1426 h) under long-term aging (grip sections) and short-term creep (gauge sections) conditions. Long-term aging at 650C leads to a +24% increase in lath thickness and a 30% decrease in the lattice dislocation density after 278 h (Fig. 4(a) and (b), Fig. 7 (a),(b), Table 2). Further aging up to 1426 h leads to insignificant changes in the structural parameters (Fig. 7 (a), (b), Table 2). However, the lath structure is completely retained (Fig. 4(c),(d)). The OIM picture (Fig. 5(a)) shows that the lath structure is mostly retained in the gauge section of the ruptured specimen, which is in contrast to the 3%Co modified P92 steel under the same creep conditions [12,20]. However, the partial destruction of “Bain circles” on the 001 pole figures is indicative for the re-orientation of the crystal lattice under creep [39]; although no changes in the misorientation of LABs occur under creep (Fig. 5(a)), as in the P92-type steel [13,14]. At the onset of apparent steady-state creep, there is no strain-induced lath coarsening and dynamic recovery. The lath thickness and density of lattice dislocations are the same in the gauge and grip sections of specimen A (Table 2, Figs. 4(b) and 6(a)). Upon further strain, well-defined strain-induced lath coarsening occurs that leads to the disappearance of particle-free interlath boundaries; while the lattice dislocation density remains unchanged (Figs. 6(b)-(d)), 7 (a) - (b), Table 2). The 20% difference in lath thickness between the gauge and grip sections at the offset strain in specimen B increases up to 30% after rupture (Figs. 6 (b)-(d), 7(a)). Transformation of some dislocation interlath boundaries into subgrain boundaries can be detected after rupture (Fig. 6(d)). The migration of these boundaries leads to the formation of subgrains of elongated shape. TEM observations distinctly show that boundary M23C6 carbides and Laves phase particles effectively pin the interlath boundaries (Figs. 5 (b)-(e) and 6). Detachment of interlath boundaries from these particles is rarely observed only at rupture (Fig. 6(d)). Lath thickening occurs due to the migration of particle-free interlath boundaries. Thus, lowering the N content and increasing the B content strongly hinder the replacement of TMLS by the subgrain structure under creep conditions [20].

10

3.5 Precipitates The evolution of a dispersion of secondary-phase particles is nearly the same in the grip and gauge sections. Strain-induced precipitation of Laves phase particles and coarsening of M23C6 carbides and MX carbonitrides are insignificant (Table 2, Figs. 4-6, 7(d)). Long-term aging leads to increased density of interlath particles due to additional precipitation of the Laves phase (Fig. 4(c)), whereas under creep, the density of these particles is lower and particle-free interlath boundaries can be detected (Fig. 5(b)-(e)). The W-rich Laves phase particles can be revealed on the SEM images obtained by Z-contrast technique as bright white particles whereas M23C6 carbides show up as grey particles (Figs. 4(c), 5(b)-(e)). The main features of the 10%Cr steel compared to other steels containing 0.05%N and 0.005%B [12,20,27] are the retention of the chains of M23C6 carbides and Laves phase particles at the interlath boundaries (Figs. 4(c), 5(b)-(e)), and the superior resistance of these phases to coarsening despite the difference between the dimensions of these particles located at high-angle and interlath boundaries. It is worth noting that the density of M23C6 carbides and the Laves phase particles located at interlath boundaries is higher than that at the boundaries of PAGs and packets. M23C6 carbides. No significant coarsening of M23C6 carbides occurs under transient creep (Fig. 7(d), Table 2). The onset of coarsening of these carbides was found at the offset strain (Fig. 8(a)). Increases of +27% and +40% were observed in the grip and gauge lengths of ruptured specimens, respectively (Fig. 7(d), Table 2). M23C6 carbides located at interlath boundaries seem to be more stable against coarsening than these carbides located at high-angle boundaries. No coarsening of interlath carbides is observed at the offset strain (Fig. 8(a)). The M23C6 carbides located on HABs grow from 100 to 150 nm (Figs. 5(e), 8(b)). The average size of M23C6 carbides located at interlath boundaries (Figs. 5(e), 8(c)) increases insignificantly up to 70 nm at rupture and no remarkable dissolution of these particles is found. As a result, interlath boundary M23C6 carbides remain as a major portion of these carbides even at rupture (Fig. 5(e)). It is worth noting that no difference in coarsening behavior of B-free M23C6 carbides and B-rich M23(BC)6 phases was found.

11

In the 10%Cr steel, the mean size of M23C6 carbides is almost half that in the 3%Co modified P92 steel [12,20] under the same conditions. Creep has an insignificant effect on the chemical composition of M23C6 carbides (Fig. 8(d)). The V content is 1.2 wt.% or less (Fig. 8(d)). Laves phase. Precipitation of Laves phase particles occurs under transient creep conditions (Fig. 8(a), Table 2). Fine W-rich M6C carbides are unstable and are replaced by the Laves phase [37] under transient creep (Table 2). The M6C carbides play the role of nucleation sites for the formation of the Laves phase. At the onset of apparent steady-state creep (specimen A, 278 h), the mean size of the Laves phase particles is slightly larger (152 nm) in the gauge section than that (132 nm) in the grip section (Table 2). However, at the offset strain (specimen B, 780 h), the Laves phase particles attain a steady size of approximately 160 nm in the gauge and grip sections, and no further coarsening occurs (Fig. 7(d), 8(a)-(c), Table 2). This size is insignificantly lower than that in the 3%Co modified P92 steel [20]. Particles of the Laves phase located on PAG boundaries are three times larger (~180 nm) (Fig. 8(b)) than those located on interlath boundaries (~ 60 nm) (Fig. 8(c)). Thus, the average sizes of M23C6 and Laves phase particles located on interlath boundaries are nearly the same. It is worth noting that a high portion of 1.7 wt.%W retains in the ferritic matrix after creep with a rupture time of 1426 h (not shown here) [44]. Depletion of W from the solid solution down to a thermodynamically equilibrium content of 0.9 wt.% appears after a rupture time of 4×104 h [44]. MX carbonitrides. At transient creep and the offset strain (specimens A and B), distinct straininduced growth of Nb-rich MX carbonitrides is found, whereas under tertiary creep (specimen C), the sizes of these dispersoids in the grip and gauge sections are the same and remain essentially unchanged up to rupture (Fig. 7(d), Table 2). Tungsten is completely depleted from the Nb-rich MX carbonitrides (Fig. 8(d)). TEM analysis reveals sparse V-rich carbonitrides in the gauge section of specimen A and grip section of specimen C (Table 2). Therefore, dissolution of V-rich MX carbonitrides occurs under long-term aging, and strain accelerates this process significantly. After

12

creep rupture, no V-rich carbonitrides were revealed in either the gauge or grip sections. These carbonitrides dissolve under creep conditions at 650C. Nevertheless, no formation of a Z-phase (CrVN) was detected, which is in contrast with other high-chromium martensitic steels [12,21]. Vanadium depleted from the MX phase dissolves in the ferritic matrix since no significant portion of this element was found in M23C6 carbides (Fig. 8(d)) [45]. In general, the average size of MX carbonitrides in the 10%Cr steel is insignificantly lower than that in the 3%Co modified P92 steel under creep conditions [20].

3.6 Microstructure after creep at other applied stresses There is no significant difference between microstructure of the 10% steel after rupture at an applied stress of 140 MPa and higher stresses. However, at an applied stress of 120 MPa, the microstructural evolution is distinctly different [44,46,47]. In the 10%Cr steel, as in other 10%Cr steels [45], the substitutional V element diffuses into M23C6 carbides. This effect leads to a +600% increase in the V content in these carbides, as the time increases from 1426 to 39 437 h. Distinct strain-induced coarsening leads to an average size of this carbides of 133 nm [44,46]. However, chains of M23C6 carbides retain at interlath boundaries, partially. Size of the Laves phase particles after long-term aging and long-term creep (39 437 h) is the same (~330 nm) [44,46]. In addition, creep induces the formation of Z-phase [44,46,47]. These changes in a dispersion of secondary phases lead to the lath coarsening up to 935 nm, the transformation of some laths to subgrains with round shape and a five-fold decrease in the lattice dislocation density.

4. Discussion Inspection of the experimental data shows that lowering the N content and increasing the B content highly increases the stability of TMLS and hinders strain-induced coarsening of secondary phase particles, which leads to the extension of the transient creep stage. Since no significant difference is revealed between the distribution of the Laves phase and MX carbonitrides in the 10%Cr steel and

13

the 3%Co modified P92 steel [20], the increased stability of the lath structure is attributed to the refinement of M23C6 carbides and their diminishing susceptibility to coarsening. It is obvious that the effect of N and B contents on these carbides plays a key role in enhancing the creep resistance of steels with a high B content [32-35].

4.1 Interfaces of M23C6 carbides and Laves phase Figure 9(a) depicts the calculation of the coarsening kinetics of a M23C6 particle at 650C in the model compositions of the 10%Cr and the 3%Co modified P92 (0.1C-9Cr-3.0Co-0.6 Mo-2.0 W-bal.Fe) [20] steels. The symbols indicate the experimental datum points for creep at 650C under a stress of 140 MPa. The calculated curves and the experimental data are in good agreement, verifying the given values of interphase energy. The lower coarsening rate in the 10%Cr steel is attributed to the strong decrease in interfacial energy from 0.36 to 0.1 J/m2. Since no difference was revealed in the coarsening behavior of B-free M23C6 carbides and B-rich M23(BC)6 phases, the decreased interface energy of Fe(110) || M23C6(111) or of Fe(110) || M23(C,B)6(111) could be attributed to the enhanced coherency [36,48] of both these phases. It should be noted that the content of (Ni+Mn) also plays an important role in the stability of M23C6 carbides. It was recently revealed [35] that if a steel contained ≥0.008 wt.% B, then the effect of (Ni+Mn) content on the growth of boundary particles was more pronounced than that of the B content. A decrease in the content of (Ni+Mn) in the 10%Cr steel enhances the creep strength compared to steels with even higher B contents but containing a higher content of Ni and Mn substitutional solutes [35]. The increased coherency and decreased interfacial energy of the M23(BC)6 phases is attributed to the replacement of C with B in the compound, while decreased interfacial energy of B-free M23C6 carbides may be associated with effect of Ni and Mn on the coherency of Fe(110) || M23C6(111) interfaces and the enrichment of these carbides by W. The interface between M23C6 carbides and the matrix in martensitic steels was shown to be coherent or semi-coherent, with orientation relationships, according to the TEM observations 14

reported

in

[48,49].

The

carbides

((011)(111)M23C6[ 1 1 1 ][ 01 1 ]M23C6),

obeyed

the

Kurdjumov-Sachs Nishiyama-Wasserman

((011)(111)M23C6[ 100 ][ 01 1 ]M23C6), Pitsch ((110)(100)M23C6[ 1 1 1 ][ 0 1 1 ]M23C6), as well as some newer orientation relationships [48-50]. The high coarsening resistance of M23C6 carbides in the 10%Cr steel is correlated with the retention of the orientation relationship with the ferritic matrix under creep conditions. The SAD patterns in Fig. 10 show that the beam direction for the M23C6 carbides located on interlath boundaries is [ 01 1 ], whereas that for ferrite is [ 1 1 1 ]. Therefore, this orientation relationship corresponds to the well-known Kurdjumov-Sachs orientation relationship [48-50]: Fe (110)  M23C6 (111) Fe [ 1 1 1 ]M23C6 [ 01 1 ]

(4)

In contrast to the 3%Co modified P911-type steel [48], the M23C6 particles in the 10%Cr steel studied do not lose their orientation relationships with the ferritic matrix. Therefore, carbides retain coherency of their boundaries and demonstrate enhanced stability under creep at 650C. It is worth noting that the coarsening rate of the Laves phase is described by theoretical calculations (Fig. 9(b)) based on the interfacial energy of 0.65 J/m2, which is lower than that in other 9-11%Cr steels [26,28]. Taking into account that M23C6 carbides at the interlath boundary as well as the Laves phase particles retain small sizes less than 100 nm at rupture, the Ostwald ripening [51,52] of these particles comprising the dissolution of small particles located on interlath boundaries and the growth of large particles located at boundaries of PAGs, packets and blocks, is retarded. It is worth noting that W solutes highly decrease diffusivity [12] and hinder the coarsening of M23C6 carbides and the Laves phase particles.

4.2. Evolution of lath structure during creep The main feature of TMLS evolution in the 10%Cr steel is the insignificant decrease in lattice dislocation density up to rupture, which is close to the accuracy to that obtained from dislocation 15

density calculations by TEM. Additionally, only a two-fold increase in lath thickness is observed under creep up to rupture. Moreover, these microstructural changes occur mainly under transient creep. It is obvious that the low r’ value in Eq.1 and low d ln  / d parameter in the transient stage are attributable to the low rate of the knitting reaction, which occurs mostly under transient creep. The knitting reaction providing mutual annihilation of lattice dislocations and the dislocations comprising interlath boundaries requires dislocation climb [4,53] and plays the role of the kinetic process [40-43] responsible for transient creep in high-chromium steels. Fine M23C6 carbides located at interlath boundaries strongly impede the knitting reaction in addition to exerting a high Zener drag force. The positive effect of the boundary carbides on the retardation of the knitting reaction is attributed to the hindering of dislocation motion by climb and glide. In addition, the Nbrich MX carbonitrides densely distributed in matrix, effectively pin lattice dislocations. W solutes also hinder the climb of lattice dislocations [12]. The synergetic effect of the two types of dispersoids and a high W content in the ferritic matrix on the dislocation motion results in the slow rate of the knitting reaction, which correlates with the pronounced transient creep stage. Thus, the precipitation of M23C6 carbides at interlath boundaries modifies the transient creep behavior of the 10%Cr steel in short-term conditions by slowing down the knitting reaction. Under apparent steadystate and tertiary creep, the knitting reaction additionally decelerates, and only a small fraction of interlath boundaries transform to subgrain boundaries, even in the ruptured specimen, that is in contrast to 9-11%Cr steels containing ~0.05%N [13,14,20,21,54,55].

4.3. Evolution of pinning pressures during creep The superior stability of TMLS under creep conditions in the 10%Cr steel can also be attributed to the Zener drag force. To estimate the role of different precipitates in the stabilization of TMLS under creep conditions, the pinning of (sub)grain/interlath boundaries by various secondaryphase particles was analyzed. The pinning pressures on the grain/subgrain boundaries from different precipitates were estimated using a technique described in [8,20,21].

16

Homogeneously distributed precipitates exert Zener pinning pressure on a boundary, which can be estimated as follows [12,20,21,29]: PZ 

3 FV d

(5)

where  is the boundary surface energy per unit area, and Fv and d are the volume fraction and size of dispersed particles, respectively. When particles are located at boundaries, the pinning pressure on a boundary depends on the size ratio of particles and subgrains/laths [8,12,20,21,56], and can be evaluated as follows:

PB 

FvB D d2

(6)

where FB is the volume fraction of the particles located at the boundaries, and D is the size of structural elements, i.e. the subgrain size or lath thickness. FB was estimated as presented in [20,21]. Figure 11 shows the densities of boundary M23C6 and Laves phase particles, which were calculated under the assumption that all M23C6 and Laves phase particles are located at interlath and (sub)grain boundaries. This change in the density of boundary precipitates during the test was taken into account calculating Zener drag exerted by M23C6 and Laves phase particles. Figure 12 presents the effect of the time of creep testing on pinning pressures. The pinning pressure exerted by homogeneously distributed MX carbonitrides was calculated using Eq. 5, and the pinning pressures exerted by M23C6-type particles and the Laves phase, which are situated on the boundaries, were calculated using Eq. 6. It is clearly observed that the pinning pressure from the boundary M23C6 carbides is extremely high whereas Zener pressure from MX carbonitrides is negligibly small due to the low density of MX carbonitrides. It should be noted that despite its continuous decrease with increasing time, the pinning pressure from the M23C6 phase is maintained at a high level up to rupture. Under apparent steady-state creep, the pinning pressure exerted by M23C6 carbides remains almost unchanged (approximately 0.26 MPa). At the onset of tertiary creep, this pinning pressure decreases progressively to 0.15 MPa at 1286 h, and then remains unchanged

17

up to rupture. Laves phase particles precipitated under transient creep exert a pinning pressure that is smaller by a factor ranging from 3 to 4 than that exerted by M23C6 carbides. The total pinning pressure exerted by both M23C6 carbide and Laves phase particles in the tertiary creep stage does not decrease below approximately 0.19 MPa, which is significantly larger than the critical value of 0.12 MPa estimated for pinning pressure PB(M23C6+Laves) in the 3%Co modified P92 steel [12]. Below this critical value, the transformation of TMLS to a subgrain structure occurs, leading to creep strength breakdown [12]. In the 10%Cr steel, the pinning pressure exerted by boundary particles is approximately 3 times larger than that in the 3%Co modified P92 steel (Fig. 12), despite the similar creep rupture times ( 2000 h). As a result, no formation of a well-defined subgrain structure at rupture was observed. Therefore, the high stability of TMLS in the 10%Cr steel is attributed to the low susceptibility of boundary M23C6 carbides to coarsening, which results in the low rate of the knitting reaction under transient creep and a pinning pressure that is high enough to hinder the migration of interlath boundaries or subgrain boundaries under apparent steady-state and tertiary creep. It is obvious that M23C6 carbides are the most effective pinning agent for stabilizing the lath structure and achieving extraordinarily high creep strength.

5. Conclusions The microstructures of a 10% Cr steel with 3%Co, 0.008% B, and 0.003% N, tempered and crept to different creep stages at 650С under a stress of 140 MPa, were studied. The main results can be summarized as follows: 1.

The steel tempered at 770C for 3 h has a tempered martensite lath structure with a lath

width of 380 nm. Fine M6C carbides are located on the boundaries in addition to nanoscale M23C6 – type carbides; and MX carbonitrides, represented mainly by Nb-rich MX and a fraction of V-rich MX precipitates that are an order of magnitude smaller, are distributed uniformly within the ferritic matrix. 18

2.

The steel demonstrates distinctly different short-term creep behavior characterized by a

highly extended transient creep stage up to ~2.3%, a low rate of exhaustion of transient creep, r’, of 2.010-2 h-1, a low d ln  / d parameter of ~145 in the transient stage, minimum creep rate of 1.410-5 h-1, and the onset of tertiary creep occurring at ~5.13%. The rate of acceleration of tertiary creep is also low (p 3.010-2 h-1), and the parameter d ln  / d in the tertiary stage is ~90. No creep strength breakdown appears up to an extremely long creep rupture time of approximately 40 000 h at a temperature of 650°C and under an applied stress of 120 MPa. 3.

The tempered martensite lath structure of the 10%Cr steel remains stable during creep

testing. An insignificant decrease in lattice dislocation density and a two-fold increase in lath thickness occur under transient creep, and correlate with the low rate of exhaustion of transient creep as well as the small value of the parameter d ln  / d in the transient stage. No significant changes occur in TMLS under apparent steady-state and tertiary creep conditions. 4.

Interlath boundary M23C6–type carbides and Laves phase particles are highly resistant to

coarsening. As a result, the Zener drag force exerted by boundary particles remains at a high level of 0.19 MPa, even in the tertiary creep stage. This suppresses the migration of interlath boundaries. Lath coarsening occurs due to the migration of M23C6 carbide free interlath boundaries. 5.

The high coarsening resistance of M23C6–type carbides is attributed to the coherency of their

Fe(110) || M23(C,B)6(111) interfaces and the two-phase separation to the B-free M23C6 carbides and B–rich M23(BC)6 phases and retention of a high portion of W in the ferritic matrix. No straininduced coarsening of these carbides occurs. Interlath boundary Laves phase particles are not susceptible to coarsening under creep conditions. Creep leads to dissolution of V–rich MX carbonitrides and insignificant coarsening of Nb-rich MX particles up to 55 nm. No formation of the Z phase was detected.

Acknowledgments

19

The study was financial supported by the Russian Science Foundation, under grant No. 14-2900173. The authors are grateful to the staff of the Joint Research Center, Belgorod State University, for their assistance with instrumental analysis.

20

References [1]

F. Abe, T.-U. Kern, R. Viswanathan, Creep-Resistant Steels, Woodhead Publishing,

Cambridge, 2008. [2]

R. O. Kaybyshev, V. N. Skorobogatykh, I. A. Shchenkova, New martensitic steels for fossil

power plant: Creep resistance, Phys. Metall. Met. 109 (2010) 186–200. [3]

M. I. Isik, A. Kostka, V. A. Yardley, K. G. Pradeep, M. J. Duarte, P. P. Choi, D. Raabe, G.

Eggeler, The nucleation of Mo-rich Laves phase particles adjacent to M23C6 micrograin boundary carbides in 12% Cr tempered martensite ferritic steels, Acta Mater. 90 (2015) 94–104. [4]

H. Kitahara, R. Ueji, N. Tsuji, Y. Minamino, Crystallographic features of lath martensite in

low-carbon steel, Acta Mater. 54 (2006) 1279–1288. [5]

I. Fedorova, A. Kostka, E. Tkachev, A. Belyakov, R. Kaibyshev, Tempering behavior of a

low nitrogen boron-added 9%Cr steel, Mater. Sci. Eng. A 662 (2016) 443–455. [6]

N. Dudova, R. Mishnev, R. Kaibyshev, Effect of Tempering on Microstructure and

Mechanical Property of Boron Containing 10%Cr Steel, ISIJ Int. 51 (2011) 1912–1918. [7]

A. Kipelova, A. Belyakov, V. Skorobogatykh, I. Shchenkova, R. Kaibyshev, Tempering-

induced structural changes in steel 10Kh9K3V1M1FBR and their effect on the mechanical properties, Metal Sci. Heat Treat. 52 (2010) 100-110. [8]

A. Kipelova, R. Kaibyshev, A. Belyakov, D. Molodov, Microstructure evolution in a 3%Co

modified P911 heat resistant steel under tempering and creep conditions, Mater. Sci. Eng. A 528 (2011) 1280–1286. [9]

F. Abe, Analysis of creep rates of tempered martensitic 9%Cr steel based on microstructure

evolution, Mater. Sci. Eng. A 510–511 (2009) 64–69. [10] H. G. Armaki, R. Chen, K. Maruyama, M. Igarashi, Contribution of recovery mechanisms of microstructure during long-term creep of Gr.91 steels, J. Nucl. Mater. 433 (2013) 23-29. [11] A. Kostka, K.-G. Tak, R. J. Hellmig, Y. Estrin, G. Eggeler, On the contribution of carbides and micrograin boundaries to the creep strength of tempered martensite ferritic steels, Acta Mater. 55 (2007) 539–550.

21

[12] A. Fedoseeva, N. Dudova, R. Kaibyshev, Creep strength breakdown and microstructure evolution in a 3%Co modified P92 steel, Mater. Sci. Eng. A 654 (2016) 1–12. [13] V. Dudko, A. Belyakov, R. Kaibyshev, Origin of Threshold Stresses in a P92-type Steel, Trans. Indian Inst. Metals 69 (2016) 223-227. [14] V. A. Dudko, A. N. Belyakov, R. O. Kaibyshev, Sources of High Creep Resistance of Modern High-Chromium Martensitic Steels, Doklady Phys. Chem. 464 (2015) 191–193. [15] M. Yoshizawa, M. Igarashi, K. Moriguchi, A. Iseda, H. Gh. Armaki, K. Maruyama, Effect of precipitates on long-term creep deformation properties of P92 and P122 type advanced ferritic steels for USC power plants, Mater. Sci. Eng. A 510–511 (2009) 162–168. [16] H. G. Armaki, R. Chen, K. Maruyama, M. Igarashi, Creep behavior and degradation of subgrain structures pinned by nanoscale precipitates in strength- enhanced 5 to 12 pct Cr ferritic steels, Metall. Mater. Trans. A 42 (2011) 3084-3094. [17] A. Aghajani, Ch. Somsen, G. Eggeler, On the effect of long-term creep on the microstructure of a 12% chromium tempered martensite ferritic steel, Acta Mater. 57 (2009) 5093–5106. [18] F. Abe, Creep behavior, deformation mechanisms and creep life of mod.9Cr-1Mo steel, Metall. Mater. Trans. A 46 (2015) 5610-5625. [19] K. Kimura, H. Kushima, K. Sawada, Long-term creep deformation property of modified 9Cr– 1Mo steel, Mater. Sci. Eng. A 510–511 (2009) 58–63. [20] N. Dudova, A. Plotnikova, D. Molodov, A. Belyakov, R. Kaibyshev, Structural changes of tempered martensitic 9%Cr–2%W–3%Co steel during creep at 650 °C, Mater. Sci. Eng. A 534 (2012) 632-639. [21] V. Dudko, A. Belyakov, D. Molodov, R. Kaibyshev, Microstructure evolution and pinning of boundaries by precipitates in a 9pctCr heat resistant steel during creep, Metall. Mater. Trans. A 44 (2013) 162–172. [22] H. K. Danielsen, J. Hald, On the nucleation and dissolution process of Z-phase Cr(V,Nb)N in martensitic 12%Cr steels, Mater. Sci. Eng. A 505 (2009) 169–177.

22

[23] L. Cipolla, H. K. Danielsen, D. Venditti, P. E. Di Nunzio, J. Hald, M. A. J. Somers, Conversion of MX nitrides to Z-phase in a martensitic 12% Cr steel, Acta Mater. 58 (2010) 669– 679. [24] J. Hald, Prospects for Martensitic 12 % Cr Steels for Advanced Steam Power Plants, Trans. Indian Inst. Metals 69 (2016) 183-188. [25] K. Suzuki, S. Kumai, Y. Toda, H. Kushima, K. Kimura, Two-phase separation of primary MX carbonitride during tempering in creep resistant 9Cr1MoVNb steel, ISIJ Int. 43 (2003) 1089 – 1094. [26] I. Fedorova, A. Belyakov, P. Kozlov, V. Skorobogatykh, I. Shenkova, R. Kaibyshev, Laves phase precipitates in a low-carbon 9%Cr martensitic steel during aging and creep at 923K, Mater. Sci. Eng. A 615 (2014) 153–163. [27] A. Kipelova, A. Belyakov, R. Kaibyshev, Laves phase evolution in a modified P911 heat resistant steel during creep at 923K, Mater. Sci. Eng. A 532 (2012) 71–77. [28] Q. Lu, W. Xu, S. van der Zwaag, Computational design of W and Co-containing creep resistant steels with barely coarsening Laves phase and M23C6 as the strengthening precipitates, Metall. Mater. Trans. A 45 (2014) 6067-6074. [29] F.J. Humphreys, M. Hatherly, Recrystallization, Related Annealing Phenomena, second ed., Elsevier, Oxford, 2004. [30] Y. Xu, M. Wang, Y. Wang, T. Gu, L. Chen, X. Zhou, Q. Ma, Y. Liu, J. Huang, Study on the nucleation and growth of Laves phase in a 10% Cr martensite ferritic steel after long-term aging, J. Alloys and Compounds 621 (2015) 93–98. [31] H. Gh. Armaki, R. Chen, K. Maruyama, M. Igarashi, Premature creep failure in strength enhanced high Cr ferritic steels caused by static recovery of tempered martensite lath structures, Mater. Sci. Eng. A 527 (2010) 6581–6588. [32] H. Semba, F. Abe, Alloy design and creep strength of advanced 9%Cr USC boiler steels containing high concentration of boron, Energy Mater. 1 (2006) 238-244. [33] M. Tabuchi, H. Hongo, F. Abe, Creep strength of dissimilar welded joints using high B-9Cr steel for advanced USC boiler, Metal. Mater. Trans. A 45 (2014) 5068–5075. 23

[34] Y. Liu, S. Tsukamoto, K. Sawada, M. Tabuchi, F. Abe, Precipitation behavior in the heataffected zone of boron-added 9Cr-3W-3Co steel during post-weld heat treatment and creep deformation, Metal. Mater. Trans. A 46 (2015) 1843–1854. [35] R. Kaibyshev, R. Mishnev, E. Tkachev, N. Dudova, Effect of Ni and Mn on the creep behavior of 9-10%Cr steels with low N and high B, Trans. Indian Inst. Met. 69 (2016) 203-210. [36] R. Sahara, T. Matsunaga, H. Hongo, M. Tabuchi, Theoretical investigation of stabilizing mechanism by boron in body-centered cubic iron through (Fe,Cr)23(C,B)6 precipitates, Metall. Mater. Trans. A 47 (2016) 1-11. [37] N. Dudova, R. Kaibyshev, On the precipitation sequence in a 10%Cr Steel under tempering, ISIJ Int. 51 (2011) 826-831. [38] D. A. Porter, K. E. Esterling, M. Sherif, Phase Transformation in Metals and Alloys, third ed., CRS Press, 2009. [39] R. K. Ray, J. J. Jonas, Transformation textures in steels, Int. Mater. Rev. 35 (1) (1990) 1-36. [40] F. Garofalo, Fundamentals of creep and creep rupture in metals, MacMillan, New York, NY, 1965. [41] J. Vanaja, K. Laha, M. D. Mathew, Effect of tungsten on primary creep deformation and minimum creep rate of reduced activation ferritic-martensitic steel, Metall. Mater. Trans. A 45 (2014) 5076-5084. [42] F. Dobes, J. Cadek, Contribution to the analysis of temporal interdependencies of creep, Kov. Mater. 19 (1981) 31-40 [43] J. Vanaja, K. Laha, Assessment of tungsten content on tertiary creep deformation behavior of reduced activation ferritic-martensitic steel, Metall. Mater. Trans. A 46 (2015) 4669-4679. [44] R. Kaibyshev, R. Mishnev, A. Fedoseeva, N Dudova, The role of microstructure in creep strength of 9-12%Cr steels, Mater. Sci. Forum, (2017) (accepted for publication). [45] Y. Xu, X. Zhang, Y. Tian, C. Chen, Y. Nan, H. He, M. Wang, Study on the nucleation and growth of M23C6 carbides in a 10% Cr martensite ferritic steel after long-term aging, Mater. Char. 111 (2016) 122–127.

24

[46] N. Dudova, R. Mishnev, R. Kaibyshev, The microstructural criterion for creep strength breakdown in a 10%Cr martensitic steel, Mater. Sci. Forum (2017) (accepted for publication). [47] R. Mishnev, N. Dudova, R. Kaibyshev, Long-term microstructural evolution in a 10%Cr high creep resistant martensitic steel at 650C, 8th International Advances in Materials Technology for Fossil Power Plants, ASM Inter., (2017) (accepted for publication). [48] A. Kipelova, A. Belyakov, R. Kaibyshev. The crystallography of M23C6 carbides in martensitic 9% Cr steel after tempering, aging and creep, Philos. Mag. 93 (2013) 2259-2268. [49] K. H. Kuo, C.L. Jia, Crystallography of M23C6 and M6C precipitated in a low alloy steel, Acta Metall. 33 (1985) 991-996. [50] P. R. Howell, J. V. Bee, R.W.K. Honeycombe, The crystallography of the austeniteferrite/carbide transformation in Fe-Cr-C alloys, Metall. Trans. A 10 (1979) 1213-1222. [51] A. Umantsev and G.B. Olson, Ostwald ripening in multicomponent alloys, Scripta Metall. Mater. 29 (1993) 1135-1140. [52] F. Liu, Dan H.R. Fors, A. Golpayegani, H.-O. Andren, G. Wahnstrom, Effect of boron on carbide coarsening at 873 K (600C) in 9 to 12 pct chromium steels, Metall. Mater. Trans. A 43 (2012) 4053-4062. [53] M. Mitsuhara, Sh. Yamasaki, M. Miake, H. Nakashima, M. Nishida, J. Kusumoto, A. Kanaya, Creep strengthening by lath boundaries in 9Cr ferritic heat resistant steel, Phil. Mag. Lett. 96 (2016) 76–83. [54] J. Pesicka, R. Kuzel, A. Dronhofer, G. Eggeler, The evolution of dislocation density during heat treatment and creep of tempered martensite ferritic steels, Acta Mater. 51 (2003) 4847–4862. [55] A. Kipelova, M. Odnobokova, A. Belyakov, R. Kaibyshev, Effect of Co on creep behavior of a P911 Steel, Metall. Mater. Trans. A 44 (2013) 577-583. [56] W. B. Hutchinson, B. J. Duggan, Influence of precipitation on recrystallization and texture development in an iron-1.2% copper alloy, Met. Sci. 12 (1978) 372-380.

25

List of figure and table captions Fig. 1. Microstructure of the 10%Cr steel after heat treatment: (a) OIM image, 001 pole figure from the one PAG indicated by a frame in the panel, and distribution of LAB misorientation (F – fraction, (%), MA – misorientation angle, (degree)); (b) SEM micrograph; (c-e) TEM micrographs show precipitated phases on foil (c) and replicas (d-e); (f) mass fraction of elements in different phases measured from replicas by EDX with TEM. Fig. 2. Plot of time to rupture vs. stress (a) and creep rate vs. time (b) and strain (c) curves at 140 MPa (b) of the 10%Cr steel at 650C, in comparison with data for P92 steel and the 3%Co modified P92 steel [12]. Fig. 3. SEM micrograph of the rupture surface of the 10%Cr steel specimen tested at 650C under a stress of 140 MPa. Fig. 4. Microstructure of the 10%Cr steel after long-term aging for: (a) 278 h, OIM image, 001 pole figure from the one PAG indicated by a frame in the panel, and distribution of LAB misorientation (F – fraction, (%), MA – misorientation angle, (degree)); (b) 278 h, TEM micrograph; (c) 1426 h, SEM micrograph; (d) 1426 h, TEM micrograph. Fig. 5. Microstructure of the 10%Cr steel after creep tests at 650C under a stress of 140 MPa: (a) creep-ruptured (1426 h), OIM image, 001 pole figure from the one PAG indicated by a frame in the panel, and distribution of LAB misorientation (F – fraction, (%), MA – misorientation angle, (degree)); (b-e) SEM micrographs: (b) specimen A (278 h); (c) specimen B (780 h); (d) specimen C (1286 h); (e) ruptured (1426 h). Fig. 6. TEM micrographs of the 10%Cr steel in the gauge section of specimens after creep tests at 650C under a stress of 140 MPa: (a) specimen A (278 h); (b) specimen B (780 h); (c) specimen C (1286 h); (d) ruptured (1426 h).

26

Fig. 7. Changes in the lath width (a), subgrain size (b), dislocation density (c), and mean size of precipitates (d) with time in the gauge (short-term creep condition) and the grip (long-term aging) sections of the 10%Cr steel specimens during creep testing at 650C under a stress of 140 MPa estimated by TEM. Fig. 8. TEM micrographs (replica) show precipitations in the gauge sections of specimens after creep tests at 650C under a stress of 140 MPa: (a) specimen B (780 h); (b) ruptured (1426 h), at grain boundaries; (c) ruptured (1426 h), at lath boundaries. Mass fraction of elements in different phases in creep-ruptured steel measured from replicas by EDX with TEM (d). Fig. 9. Calculated and experimental results of the coarsening of M23C6 (a) and Laves phase (b) particles at 650C in the 10%Cr steel and the 3%Co modified P92 steel [20]. Lines indicate the calculated temporal dependence of particle size in the model compositions. Symbols indicate the experimental results of the mean particle size during creep testing at 650C under a stress of 140 MPa. Fig. 10. Microstructure and SAD patterns from the M23C6 carbide and ferritic matrix of the 10% Cr steel creep-ruptured at 650C under a stress of 140 MPa. Fig. 11. Particle density at (sub)grain boundaries determined by SEM, as a function of time in the 10%Cr steel during creep testing at 650C under a stress of 140 MPa. Densities of M23C6 and Laves phase particles at (sub)grain boundaries were calculated using a technique from [20]. Fig. 12. Change in pinning pressures from different particles on the (sub)grain and lath boundaries in the gauge section of the 10%Cr steel, during creep testing at 650C under a stress of 140 MPa. The gray lines show comparative data for pinning pressures in the 3%Co modified P92 steel [20].

27

Table 1. Parameters of creep behavior of the 10% Cr steel during creep testing at 650C under a stress of 140 MPa, in comparison with data for the 3%Co modified P92 steel.

min , -1

10%Cr steel 3%Co modified P92 steel

h 1.410-5 1.210-5

tr, h 1426 1828

 min , % 3.5 1.0

T, % 2.3 0.77

min ×tr, % 2.83 3.84

3 , % 10.75 7.1

r’, h-1 0.02 0.03

p, h-1 0.03 0.015

Table 2. Effect of long-term aging (grip sections) and short-term creep (gauge sections) on structural parameters of the 10% Cr steel during creep testing at 650C under a stress of 140 MPa. Structural parameters

Time, h Lath width, nm Dislocation density, ×1014 mMean size of particles, nm

2

M23C6 Laves phase Nb-rich MX V-rich MX M6 C

0 380

A gauge / grip 278 490 / 470

B gauge / grip 780 560 / 470

C gauge / grip 1286 600 / 490

Creep ruptured neck / grip 1426 660 / 510

1.7

1.2 / 1.2

1.2 / 1.2

1.1 /1.2

1.0 / 1.1

70  30 40 25

80 / 77 152 / 132 47 / 35 55 / 40  / 25

88 / 79 157 / 149 48 / 36  / 40 /

96 / 85 160 / 159 55 / 51  / 50 /

100 / 89 168 / 162 55 / 52 / /

Initial state

28

29

30

31

32

33

34

35

36