See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/278111330

Feasibility of MARBN Steel for Application to Thick Section Boiler

Components in USC Power Plant at 650 degrees C

Conference Paper · January 2008

DOI: 10.1361/cp2007epri0092

CITATIONS READS

32 1,548

7 authors, including:

Fujio Abe Nobuyoshi Komai

National Institute for Materials Science Mitsubishi Heavy Industries
315 PUBLICATIONS 8,569 CITATIONS 38 PUBLICATIONS 382 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Structural integrity of fusion reactor components View project

All content following this page was uploaded by Fujio Abe on 22 February 2020.

The user has requested enhancement of the downloaded file.

 5th EPRI International Conference, October 3-5, 2007
Marco Island, Florida, USA

FEASIBILITY OF MARBN STEEL FOR APPLICATION

TO THICK SECTION BOILER COMPONENTS
IN USC POWER PLANT AT 650oC
F. Abe*
M. Tabuchi*
H. Semba**
M. Igarashi+
M. Yoshizawa+
N. Komai++
A. Fujita+++

* National Institute for Materials Science (NIMS),

1-2-1 Sengen, Tsukuba 305-0047, Japan
** National Institute for Materials Science, now Sumitomo Metal Industries, Ltd.
now 1-8 Fuso-cho, Amagasaki 660-0891, Japan
+
Sumitomo Metal Industries
1-8 Fuso-cho, Amagasaki, Hyogo 660-0891, Japan
++
Mitsubishi Heavy Industries
5-717-1, Fukahori-machi, Nagasaki 851-0392, Japan
+++ Mitsubishi Heavy Industries
2-1-1, Arai-cho, Shinhama, Takasago 676-8686, Japan

Abstract

A 9Cr-3W-3Co-VNbBN steel, designated MARBN (MARtensitic 9Cr steel strengthened by

Boron and Nitrides), has been alloy-designed and subjected to long-term creep and oxidation
tests for application to thick section boiler components in USC power plant at 650oC. The
stabilization of lath martensitic microstructure in the vicinity of prior austenite grain
boundaries (PAGBs) is essential for the improvement of long-term creep strength. This can
be achieved by the combined addition of 140ppm boron and 80ppm nitrogen without any
formation of boron nitrides during normalizing at high temperature. The addition of small
amount of boron reduces the rate of Ostwald ripening of M23C6 carbides in the vicinity of
PAGBs during creep, resulting in stabilization of martensitic microstructure. The
stabilization of martensitic microstructure retards the onset of acceleration creep, resulting in
a decrease in minimum creep rate and an increase in creep life. The addition of small
amount of nitrogen causes the precipitation of fine MX, which further decreases the creep
rates in the transient region. The addition of boron also suppresses the Type IV
creep-fracture in welded joints by suppressing grain refinement in heat affected zone. The
formation of protective Cr2O3 scale is achieved on the surface of 9Cr steel by several
methods, such as pre-oxidation treatment in Ar gas, Cr shot-peening and coating of thin layer
of Ni-Cr alloy, which significantly improves the oxidation resistance of 9Cr steel in steam at
650oC. Production of a large diameter and thick section pipe and also fabrication of welds
of the pipe have successfully been performed from a 3 ton ingot of MARBN.

Introduction

Energy security combined with lower carbon dioxide emissions is increasingly quoted to
protect global environment in the 21st century. Coal is an abundant, low cost resource used
1
for electric power generation. However, traditional coal-fired power plants have been
emitting CO2 at high levels relative to other electric power generation options. Adoption of
ultra supercritical (USC) power plants with increased steam parameters significantly
improves efficiency, which reduces fuel consumption and the emissions of CO2.

Materials development projects for advanced ultra supercritical (A-USC) power plants with
steam conditions of 700oC and above have been already initiated to gain net efficiency higher
than 50% at Thermie AD700 project aiming at 700oC in Europe (1) and at DOE Vision 21
project aiming initially at 760oC in the US (2). These projects involve the replacement of
9-12%Cr martensitic steels by nickel base superalloys for the highest temperature
components. However, it should be noted that nickel base superalloys are much more
expensive than austenitic and ferritic/martensitic steels. In Japan, we have been discussing
the feasibility of nickel base superalloys, austenitic steels and 9-12Cr martensitic steels for
components with steam temperatures above 700oC, at around 700oC and up to 650oC,
respectively. The combination of nickel base superalloys, austenitic steels and 9-12Cr
martensitic steels makes it possible to construct highly efficient and economically viable
A-USC power plant. The increase in steam parameters exceeding 600oC requires extensive
R&D of advanced ferritic/martensitic steels with sufficient long-term creep rupture strength
higher than conventional ones.

This paper describes feasibilities of MARBN steel (3), which is a MARtensitic 9Cr steel
strengthened by Boron and Nitrides and was developed by NIMS in corporation with private
companies in Japan, for application to thick section boiler components in USC power plant
with steam temperature of 650oC.

Critical issues for development of ferritic steels

Critical issues for the development of ferritic steels for thick section boiler components, such
as main steam pipe and header, with steam condition of 650oC are the improvement of
following properties.
(1) long-term creep strength (resistance to degradation in long-term creep strength)
(2) oxidation resistance in high-temperature steam
(3) resistance to Type IV cracking strength loss in welded Joints
(4) thermal cycling capabilities (resistance to creep-fatigue damage)
Usually, the creep rupture strength of 100MPa at operating temperature and 105 h is a target
value for the development of new steels used under creep conditions. At present, we have
no deciding criterion for oxidation resistance. We think that candidate steels for boiler
components operating at 650oC should exhibit oxidation resistance in steam at 650oC better
than that of Gr.91 in steam at 600oC, because Gr.91 are being now used for long duration in
power plants operating at 600oC. In terms of creep strength of welded joints, the weld
strength reduction factor (WSRF) of 105 h creep rupture strength should be larger than 0.75
for new steels at 650oC, because the WSRF for Gr.91 is 0.75 at 600oC (4).

Experimental

9Cr-3W-3Co-0.2V-0.05Nb steels with 140 ppm boron but different nitrogen contents were
used to investigate the effect of nitrogen on creep strength of 9Cr boron steel base metal.
The chemical compositions of the steels are given in Table 1 (5). The steels were basically
prepared by vacuum induction melting to 50 kg ingots. Hot forging and hot rolling were
carried out to produce plates of 20 mm in thickness. Creep tests were carried out at 650oC

2
for up to about 3 x104 h under constant load condition, using specimens of 10 mm in gauge
diameter and 50 mm in gauge length.

Table 1 Chemical compositions and heat treatment conditions of 9Cr-3W-3Co-VNb steels

with 0.014% (140 ppm) boron but different nitrogen contents.
(mass %)
C Si Mn Cr W Co V Nb N B normalizing tempering
o o
0.0015N 0.076 0.30 0.51 9.00 3.02 3.02 0.19 0.053 0.0015 0.0132 1150 C x 1 h 770 C x 4 h
0.0034N 0.078 0.30 0.51 8.99 2.91 3.01 0.19 0.049 0.0034 0.0139 1080oC x 1 h 800oC x 1 h
o o
0.0079N 0.078 0.31 0.49 8.88 2.85 3.00 0.20 0.051 0.0079 0.0135 1150 C x 1 h 770 C x 4 h
o o
0.030N 0.078 0.30 0.51 9.08 3.05 3.03 0.20 0.055 0.0300 0.0150 1150 C x 1 h 770 C x 4 h
o o
0.065N 0.081 0.31 0.51 8.90 3.07 3.00 0.20 0.054 0.0650 0.0144 1150 C x 1 h 770 C x 4 h

Another series of 9Cr-3W-3Co-0.2V-0.05Nb steels with different contents of boron and

nitrogen and conventional steel P92, whose chemical compositions are given in Table 2, were
used to investigate the effect of boron and nitrogen on creep strength of welded joints.
Plates of about 30 mm thickness were subjected to multi-layers Gas Tungsten Arc (GTA)
welding. The post weld heat treatment (PWHT) was carried out for each specimen
including the base metal at 740oC for 4.7 h.
The concentration of boron in M23C6 precipitates was analyzed by field emission scanning
Auger electron spectroscopy. The details of FE-AES analysis were described elsewhere (6).

Table 2 Chemical compositions and heat treatment conditions of 9Cr-3W-3Co-VNb steels

with different boron and nitrogen contents and conventional steel P92.
(mass %)
C Si Mn Cr W Mo Co V Nb N B Normalizing Tempering
o o
47B-17N 0.079 0.30 0.48 8.77 2.93 - 2.91 0.18 0.046 0.0017 0.0047 1080 C x 1 h 800 C x 1 h
90B-14N 0.074 0.30 0.48 8.93 3.13 - 2.92 0.18 0.046 0.0014 0.009 1080oC x 1 h 800oC x 1 h
o o
130B-15N 0.077 0.30 0.49 8.97 2.87 - 2.91 0.18 0.046 0.0015 0.013 1080 C x 1 h 800 C x 1 h
o o
180B-11N 0.078 0.30 0.49 8.91 2.85 - 2.91 0.18 0.047 0.0011 0.018 1080 C x 1 h 800 C x 1 h
160B-85N 0.079 0.31 0.51 8.81 3.05 - 3.10 0.20 0.055 0.0085 0.016 1150oC x 1 h 770oC x 4 h
o o
P92 0.09 0.16 0.47 8.72 1.87 0.45 - 0.21 0.06 0.050 0.002 1070 C x 1 h 780 C x 1 h

Results and discussion

Long-term stability of base metal

Combination of boron-strengthening and MX nitride-strengthening. The effect

of nitrogen addition on the creep rupture strength of 9Cr-3W-3Co-0.2V-0.05Nb steel
containing 140 ppm boron (Table 1) are shown in Fig. 1 (3, 5). The creep rupture strength
significantly increases with increasing nitrogen content from 0.0015 to 0.0079 %, but the
excess addition of nitrogen as high as 0.030 and 0.065% (300 and 650 ppm) decreases the
creep rupture strength. The excess addition of nitrogen of 0.030 and 0.065% causes the
formation of boron nitrides during normalizing heat treatment. The solubility product for
boron nitrides in 9 to 12Cr steels at normalizing temperatures of 1050 to 1150oC is given by

log [%B] = - 2.45 log [%N] - 6.81, (1)

where [%B] and [%N] are the concentration of boron and nitrogen in mass% (7). At a
boron concentration of 140 ppm, 95 ppm nitrogen can dissolve in the matrix without any
formation of boron nitrides during normalizing. The present results indicate that boron and

3
nitrogen in solid solution after normalizing is very effective for the improvement of creep
rupture strength. But the formation of boron nitrides reduces the strengthening effects.

200
9Cr-3W-3Co-VNb steel
650oC
Stress ( MPa )

0.0015 N
0.0034 N
100 0.0079 N
0.030 N
90 0.065 N
80 P92
T91
70 1
10 102 103 104 105
Time to rupture ( h )
Figure 1 Effect of nitrogen on creep rupture strength of 9Cr-3W-3Co-VNb steel
with 140 ppm boron at 650oC.

Enough creep ductility is also required for thick section components in USC plant, because
the thermal cycling capabilities of thick section components would be severely restricted by
fatigue damage. Yamaguchi and co-workers (8) have carried out creep-fatigue tests at
650oC for various kinds of ferritic steels and found the boron addition is also advantageous
for the improvement of creep-fatigue life. The creep-fatigue life is proportional to the
reduction of area in creep rupture test, namely, proportional to the creep ductility but not
proportional to the creep strength.

Figures 2(a) and 2(b) show the effect of nitrogen addition on the creep deformation behavior
of the steel at 650oC and 120 MPa. The addition of nitrogen causes a rapid decrease in
creep rate with time and also with strain in the transient region. It should be noted that the
creep rate in the transient region is the same among the three steels containing different
nitrogen of 79, 300 and 650 ppm (Fig.2(a)). The onset of acceleration creep is retarded up
to longer time by the addition of 79 ppm nitrogen but it shifts to earlier times with increasing
nitrogen content above 79 ppm. In Fig. 2(b), the onset of acceleration creep takes place at a
high strain of 0.045 in the very low nitrogen 0.0015N steel as shown by the arrow, while it
takes place at a low strain of 0.007 in the 0.0079N, 0.030N and 0.065N steels. This
suggests that the addition of nitrogen promotes the heterogeneity in creep deformation. The
acceleration of creep rate by strain dln  /d  after reaching a minimum creep rate is
evaluated from the slope of creep rate versus strain curves to be 32.2, 121.1, 113.0 and 90.7
for the 0.0015N, 0.0079N, 0.030N and 0.065N steels, respectively.

The minimum creep rate steeply decreases with increasing nitrogen content up to 79 ppm but
it turns to increase slowly above 79 ppm nitrogen, as shown in Fig.3. This reflects the
nitrogen content dependence of creep rupture strength shown in Fig.1.

In order to make clear the reason why the creep rates in the transient region are the same
among the three steels containing nitrogen of 79, 300 and 650 ppm (Fig. 2(a)), Fig.4 shows

4
 0 -1
10 10
650oC, 120MPa 650oC, 120MPa
10-1 10
-2

Creep rate ( 1 / h )
Creep rate ( 1 / h )

-2
10
10-3
-3
10
-4
10
-4
10
-5 10-5
10 0.0015N
0.0015N
0.0079N 0.0079N
10
-6
0.030N 10-6 . 0.030N
0.065N dlog / d 0.065N
-7
10-7 -1 0 1 2 3 4 10
10 10 10 10 10 10 0 0.05 0.1
Time ( h ) True strain

Figure 2. Effect of nitrogen on creep rate versus time curves and

creep rate versus strain curves of the steels at 650oC and 120 MPa.

10-3
Minimum creep rate ( 1 / h )

9Cr-3W-3Co-0.2V-0.05Nb-0.014B

-4
650oC, 120 MPa
10

10-5

-6
10

10-7 -3
10 10-2 10-1
Nitrogen content ( wt % )
Figure 3. Effect of nitrogen content on minimum creep rate of the steel
with 140 ppm boron at 650oC and 120 MPa.

the amount of dissolved nitrogen and precipitated nitrogen in the 0.0015N, 0.0079N and
0.065N steels after tempering. In the low nitrogen 0.0015N and 0.0079N steels, most of
nitrogen is in solution after tempering. But in the high nitrogen 0.065N steel, most of
nitrogen has already precipitated during tempering. Dissolved nitrogen can precipitate as
fine MX carbonitrides during creep at 650oC. Indeed, very fine vanadium-rich MX
carbonitrides were observed to have precipitated in the 0.0079N steel during aging for 1000 h
at 650oC (5). Fine MX carbonitrides precipitated during creep are responsible for the
significant decrease in creep rate in the transient region shown in Fig. 2. It should be noted
that the dissolved nitrogen concentration is roughly the same between the 0.079N and 0.065N
steels after tempering. This results in the same the creep rates in the transient region among
the steels containing nitrogen above 79 ppm.

5
 0.08

Concentration of nitrogen (mass %)

9Cr-3W-3Co-0.2V-0.05Nb-0.014B
After tempering

0.06 Precipitated nitrogen

Dissolved nitrogen

0.04

0.02

0
0.0015N 0.0079N 0.065N
Figure 4. Dissolved and precipitated nitrogen concentration
in the 0.0015N, 0.0079N and 0.065N steels after tempering.

Next, let us consider the reason why the onset of acceleration creep takes place at earlier
times in the high nitrogen steels (Fig. 2(a)). We have revealed that in the 0.0034N steel with
140 ppm boron, the fine distribution of M23C6 carbides along prior grain boundaries (PAGBs)
is still maintained during exposure at 650oC and that boron is enriched in M23C6 carbides near
PAGBs (5). The fine distribution of M23C6 carbides is also observed in the steel containing
no boron after tempering, but extensive coarsening takes place during exposure at 650oC.
This indicates that boron reduces the rate of Ostwald ripening of M23C6 carbides near PAGBs
during creep at 650oC. The fine distribution of M23C6 carbides near PAGBs retards the
onset of acceleration creep, resulting in lower minimum creep rate and longer time to rupture.
Figure 5 compares the enrichment of boron in M23C6 carbides between the 0.0015N and
0.065N steels. In the 0.0015N steel, no boron nitride formed during normalizing and hence
most of boron can contribute to the enrichment in M23C6 carbides near PAGBs. On the
other hand, in the 0.065N steel, large amount of boron nitrides formed during normalizing.
Therefore, most of soluble boron is consumed to form boron nitrides and only a slight amount
of boron contributes to the enrichment in M23C6 carbides. This is a reason why the onset of
acceleration creep takes place at earlier time in the high nitrogen steel.

10.0 10.0
Boron in M23C6 ( mol% )
Boron in M23C6 ( mol% )

(a) 132ppm B, 15ppm N (b) 144ppm B, 650ppm N

8.0 8.0
650o C, tr = 3304h, Grip 650o C, tr = 3688h, Grip
6.0 6.0

4.0 4.0

2.0 2.0

0.0 0.0
0 10 20 30 40 0 10 20 30 40
Distance from PAGB (m) Distance from PAGB (m)
Figure 5. Boron content in M23C6 carbides in the steels after creep rupture testing
at 650oC for 3000-4000 h, as a function of distance from prior austenite grain boundary.

6
Formation process of M23(CB)6. In iron-carbon-boron ternary phase diagram at 900oC,
Fe23(CB)6 appears at alloy compositions containing several % boron (9). Because the boron
content in the present steels is only 140ppm in maximum, Fe23(CB)6 is difficult to form in the
matrix of present steels. Therefore, we estimate segregation of boron at grain boundaries.
Using a binding energy of 62.7kJ/mol (10) reported for Type 316 stainless steel, Fig.6 shows
the concentration of boron at grain boundaries as a function of temperature. At a
normalizing temperature of 1100oC, segregation of several % boron can be achieved at grain
boundaries. During tempering at 800oC and subsequent creep at 650oC, the segregation is
more significant because of lower temperatures.
6
10
Concentration of boron ( ppm )

CGB = C0 exp (B/RT)

105 CGB CGB CGB

Concentration of boron
 = 0.3nm
4 d = 50m
10 Ci = 100ppm 
B = 62.7kJ/mol

103 Within grain Ci

Grain boundary
C0
2
10 C0
Distance
1
10
600 800 1000 1200
o
Temperature ( C )
Figure 6 Estimation of grain boundary segregation of boron.

Based on the estimation of grain boundary segregation of boron described above, we think
that at first during normalizing heat treatment at 1100oC, only grain boundary segregation of
boron takes place but no precipitation of M23C6 carbide because of high temperature, as
shown in Fig. 7. During tempering at 800oC, the precipitation of M23C6 carbide can take

(a) Normalizing temp. (b) Tempering temp (c) Creep test temp.
1100oC 800oC 650oC
. . .
.
. . .
.
. . . ... .
..
.
. .. Enrichment of
.. . Enrichment of
. . B in M23C6 . . B in M23C6
. .
.. .. . ..
. . . ..
. . . .. .. .. .. . . .
.
. . . . . .. . . .
.. .
.
. . .
. . . .
.
.. ..
. .
..
. .. ... . ..
.
..
Segregation of B Segregation of B
. . . . Segregation of B .
around G.B. . around G.B.
.
around G.B.

..
. . Boron BN M23C6 MX
Figure 7. Enrichment process of boron into M23C6 during heat treatment.

7
place preferentially at grain boundaries and lath-block boundaries. Because the grain
boundary segregation of boron is achieved, M23(CB)6 can form by enrichment of boron into
the M23C6 carbides. During creep at 650oC, the enrichment of boron in M23(CB)6 is more
significant because of lower temperature.

When the steel contains high nitrogen contents, boron nitrides can form during normalizing,
consuming most of soluble boron. This significantly reduces the segregation of boron at
grain boundaries and also reduces the enrichment of boron in M23C6, as shown in Fig.5.

Reduction of coarsening rate of M23(CB)6. According to the mechanism of Ostwald

ripening, the main factors affecting the coarsening rate of precipitate particles are diffusion
coefficient, solid solubility and interfacial energy (11). However, we have no experimental
result showing a significant decrease in diffusion coefficient and solid solubility by the
addition of small amount of boron. Furthermore, we have no evidence showing the
enrichment of boron at the interface between M23C6 carbide particle and alloy matrix.
Therefore, these parameters could be excluded from the main explanation of the reduction of
coarsening rate of M23C6 carbides by boron. It should be noted that Ostwald ripening in
solid matrix requires accommodation of local volume change around a growing particle,
because specific volume of carbide is larger than that of the alloy matrix. As a small carbide
goes into solution, carbon atoms take up interstitial sites in the matrix and vacancies are
created at the carbide interface, as shown in Fig.8. Then vacancies migrate through the
matrix and arrive at a growing carbide interface, which accommodates local volume change.
If boron atoms occupy vacancies around growing carbide particle, local volume change
cannot be accommodated. This causes the reduction of coarsening rate. We think the
main effect due to boron is to occupy vacancies around growing M23C6 carbide particle.

Dissolving Growing
M23C6 M23C6
diffusive flux
(C, Cr, W, Fe, - - )

vacancies

vacancies boron
Figure 8 Accommodation of local volume change around a growing M23C6 particle.

Long-term stability of welded joints and HAZ microstructure

Figure 9 compares the creep rupture data between the base metals and welded joints for
9Cr-3W-3Co-VNb steels containing different boron and nitrogen contents and P92 (Table 2)
at 650oC (12). Although P92 contains 20 ppm boron, the degradation of creep rupture
strength in welded joints is quite large at long times due to Type IV fracture. This suggests
that even if the steel contains boron, it does not necessarily suppress the Type IV fracture.
The degradation in creep rupture strength of welded joints is negligibly small in the
9Cr-3W-3Co-VNb steel even if it contains only 47ppm boron but no addition of nitrogen
(Fig.9(a)) and also even if it contains 160 ppm boron and 85 ppm nitrogen (Fig. 9(e)).

We have revealed that multi-axial stress condition in the fine-grained HAZ with lower creep

8
strength, resulting from mechanical constrain effect by the surrounding weld metal and base
metal with higher creep strength, is essential for the brittle Type IV fracture. Therefore, we
have systematically examined the effect of boron and nitrogen contents on grain refinement
behavior for the 9Cr steel during heating at temperatures between 900 and 1100oC. The
results are summarized in Fig. 10, overlapping on the composition diagram for the formation
of BN during normalizing heat treatment. No addition of boron causes the grain refinement
at around Ac3 temperature, irrespective of nitrogen content. The addition of boron and
nitrogen without any formation of BN during normalizing causes no grain refinement, while
the steels with the formation of BN during normalizing produce the fine-grained
microstructure. The present results suggest that soluble boron is essential for the
suppression of grain refinement during heating.

150 150
650oC o
650 C
(a) (b)

Stress ( MPa )
Stress ( MPa )

100 100

80 80
9Cr-3W-3Co-VNb-0.08C steel 9Cr-3W-3Co-VNb-0.08C steel
47 ppm B - 17 ppm N 90 ppm B - 14 ppm N
60 60
Base metal Base metal
Welded joints Welded joints

40 1 2 3 4 5 40 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10
Time to rupture ( h ) Time to rupture ( h )
150 150
650oC o
650 C
(c) (d)
Stress ( MPa )

Stress ( MPa )

100 100

80 80
9Cr-3W-3Co-VNb-0.08C steel 9Cr-3W-3Co-VNb-0.08C steel
130 ppm B - 15 ppm N 180 ppm B - 11 ppm N
60 60
Base metal Base metal
Welded joints Welded joints

40 1 2 3 4 5 40 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10
Time to rupture ( h ) Time to rupture ( h )
200 150
o
650 C 650oC
(e) (f)
Stress ( MPa )

Stress ( MPa )

100

80 P92
9Cr-3W-3Co-VNb-0.08C steel
100 (9Cr-0.5Mo-1.8W-VNb steel)
160 ppm B - 85 ppm N 20 ppm B - 500 ppm N
60
80 Base metal Base metal
Welded joints Welded joints

60 1 40 1 2 3 4 5
2 3 4 5 10 10 10 10 10
10 10 10 10 10
Time to rupture ( h ) Time to rupture ( h )

Figure 9 Creep rupture data for base metals and welded joints of the steels at 650oC.

9
 Nitrogen ( ppm )
10 100 1000

BN
-2
10 100
Boron ( mass % )

Boron ( ppm )
Solid solution P122 No grain refinement
P122
-3
10 1050 - 1150oC 10
P92 Grain refinement
P92

Large BN
-4 Small BN
10 1
No BN
-3 -2 -1
10 10 10
Nitrogen ( mass % )
Figure 10 Grain refinement and no grain refinement in 9 to 12Cr steels
during heating at 900 to 1100oC.

Oxidation resistance in steam

Figure 11 compares the weight gain due to oxidation in steam at 650oC among various 9 to
12Cr steels (13). The sheet specimens having a size of 10x20x2 mm were cut from bulk
materials, which were already heat treated, ground on a SiC paper of 320 grit, rinsed in
acetone and then supplied to the oxidation test in steam at 650oC. The weight gain of P92
and P122 in steam at 650oC is much larger than that of T91 at 600oC. This suggests that
existing steels, even in 12Cr steel P122, cannot satisfy the present criterion for oxidation
resistance that candidate steels for boiler components operating at 650oC should exhibit
oxidation resistance in steam at 650oC better than that of Gr.91 in steam at 600oC. It should

9Cr-3WVNb
25 9Cr-1.5MoVNb
8.5Cr-2WVNb + 0.8Si o
9Cr-3WVNb + 3Pd 650 C
Weight gain ( mg / cm2 )

20 T91
P92
P122

15

10

5
T91 (600oC)

0
0 200 400 600 800 1000
Time ( h )
Figure 11 Weight gain of the steels in steam at 650oC,
comparing with that of T91 at 600oC.

10
be noted that the weight gain of 9Cr-3WVNb steel with 3% Pd is significantly lower at 650oC
than that of T91 at 600oC (14). Thin scale of Cr-rich oxide, presumably Cr2O3, forms on the
specimen surface of 9Cr-3WVNb steel with 3%Pd in steam at 650oC, while thick scale
consisting of magnetite in the outer layer and Fe-Cr spinel oxides in the inner layer forms on
the surface of the other steels including T91. No evidence is found for the formation of any
oxide containing Pd on the surface of 9Cr-3WVNb steel with 3%Pd. Based on the results in
Fig.11, the formation of protective Cr2O3 scale is essential for the development of
oxidation-resistant ferritic steels for USC boilers at 650oC.

The formation of protective Cr-rich oxide scale is achieved by the combination of Si addition
and pre-oxidation treatment in argon gas (15), by the combination of shot-peening of Cr and
pre-oxidation treatment in air at 700oC (16), and by coating of Ni-20Cr and Ni-50Cr thin
layers (17). The oxidation test was carried out for 9Cr-3WVNb steel containing different Si
contents of 0 to 0.8% in steam at 650oC after pre-oxidation treatment in Ar gas. In the
condition of no pre-oxidation treatment, the addition of Si decreases the weight gain of the
steel in steam but the effect of Si is not large. The pre-oxidation treatment in Ar gas further
decreases the weight gain during subsequent oxidation in steam. This is more significant
with increasing Si concentration and with increasing pre-oxidation time, as shown in Fig.12.
For example, the combination of 0.3%Si addition and pre-oxidation treatment in Ar gas at
650oC for 100h satisfies the present criterion for oxidation resistance in steam at 650oC.

30
9Cr-3W-0.2V-0.05Nb 0%Si
Weight gain ( mg / cm2 )

0.1%Si
650oC, 1000h in steam after 0.3%Si
pre-oxidation in Ar at 650oC 0.5%Si
20
0.8%Si
o
T91 (600 C,1000h)
without pre-oxidation
10

0
0 200 400 600 800 1000 1200
Pre-oxidation time ( h )
Figure 12 Weight gain of 9Cr-3W-0.2V-0.05Nb steel containing different Si contents by
oxidation in steam at 650oC for 1000h, as a function of pre-oxidation time in Ar gas at 650oC.

The thin scale of Cr-rich oxides formed during pre-oxidation treatment is stable during
subsequent oxidation in steam at 650oC, as shown in Fig.13. The breakaway in the weight
gain curve is not observed for the specimen with pre-oxidation treatment.

The resistance to exfoliation of Cr2O3 scale is another requirement. At first we have tried to
examine the resistance to exfoliation of Cr2O3 scale by thermal cycling test in steam after
pre-oxidation treatment (16). The increase in weight gain in steam was substantially the
same between thermal cycling test and continuous oxidation test. This indicates that the
Cr2O3 scale formed during pre-oxidation is stable and highly resistant to exfoliation. Next,

11
 we have tried to cross-cut the Cr2O3 scale by a sharp edge after pre-oxidation and then
supplied to oxidation test in steam at 650oC for 1000h. After cutting the Cr2O3 scale, the
steel surface was directly exposed to steam. But no evidence was found for any exfoliation
or breakaway of surrounding Cr2O3 scale during subsequent steam oxidation. Again this
indicates that the Cr2O3 scale formed during pre-oxidation is very stable and highly resistant
to exfoliation. It should be noted that at present there is no standard test method for the
assessment of resistant to exfoliation.

(a) (b)
O
9Cr-3W-3Co-0.2V-0.05Nb-0.7Si-0.02B
Weight gain ( mg / cm2 )

6
Cr Fe
o
4 Oxidation in steam at 650 C
After pre-oxidation in Ar at 700oC, 50 h
2 Without pre-oxidation
Mn Si

0
0 1000 2000 3000 4000 5000
Time ( h )

Figure 13 (a) Weight gain of 9Cr-3W-3Co-VNbSi steel (MARB2) in steam at 650oC and
(b) formation of protective Cr2O3 scale during pre-oxidation in argon at 700oC for 50 h.

Production of pipe and fabrication of welds

The production of a large diameter and thick section pipe and subsequent fabrication of
circumference welds of the pipe have successfully been performed from a 3 ton ingot of
MARBN. The chemical compositions of MARBN are given in Table 3. From the
viewpoints of enhancing creep rupture strength for base metal (Fig.1) and welded joints
(Fig.8) and of no formation of BN during normalizing heat treatment (Fig.9), the boron and
nitrogen contents were set to be 0.013% (130ppm) boron and 0.007% (70ppm) nitrogen.
The residual impurity Al was minimized, because it consumes soluble nitrogen to form AlN
during creep, which degrades the creep strength (18). A 3 ton ingot, having a size of
785mm in diameter and 580mm in height, was prepared by vacuum melting. The ingot was
successfully hot-forged to a pipe, which was heat treated and machined to a size of 470 mm
in outer diameter, 65 mm in thickness and 1300 mm in length. The pipe was normalized at
1100oC for 3h followed by air cooling and then tempered at 780oC for 4 h followed by air
cooling. Figure 14 shows the appearance of pipe after machining. The creep specimens
were taken from the outside, centre and inside parts in the wall thickness and subjected to the
creep rupture testing at 650oC. The results are shown in Fig.15, together with those for the
small melts of 0.0034N and 0.0079N steels in Table 1. Although the creep rupture data for
the MARBN pipe are located just below those for the 0.0079N steel because of a little lower
nitrogen content, the MARBN pipe material exhibits substantially the same creep rupture
strength as that of the small melt of 0.0079N steel in Table 1. We are now continuing
long-term creep rupture tests.

12
 Table 3 Chemical compositions of MARBN pipe.
(mass %)
C Si Mn Cr W Co V Nb N B sol-Al
MARBN Pipe 0.08 0.33 0.51 9.03 2.79 3.01 0.19 0.056 0.0071 0.013 0.001

Figure 14 Appearance of MARBN pipe,

470 mm in outer diameter, 65 mm in thickness and 1300 mm in length.

Circumference welding of the MARBN pipe has successfully been carried out using gas
tungsten-arc welding (GTAW) process and using two kinds of filler wires; Ni-base AWS ER
NICr-3 (Alloy 82) and matching filler wires. The preheating temperature was kept to be
150oC or more to avoid low-temperature cracking, while the interpass temperature was below
200oC to avoid high-temperature cracking. After welding, the welded joints were subjected
to a postweld heat treatment (PWHT) at 740oC for 4 h. Figure 16 shows circumference TIG
welding of the MARBN pipe and appearance of welded joints after PWHT, with Ni-base
AWS ER NICr-3 and matching filler wires. The creep rupture testing of welded joints is
being now carried out.

300
650oC
9Cr-3W-3Co-VNb steel
200
Stress ( MPa )

140ppmB
79ppmN

140ppmB
34ppmN
100 Pipe
Outside
80 Center P92
Inside T91
60 1 2 3 4 5
10 10 10 10 10
Time to rupture ( h )
Figure 15 Creep rupture data for MARBN pipe,
together with 0.0034N and 0.0079N steels, T91 and P92 at 650oC.

13
 Figure 16 Circumference TIG welding of MARBN pipe and appearance of
welded pipes after PWHT, with Ni-base AWS ER NICr-3 and matching filler wires.

Conclusions

(1) A highly creep-resistant martensitic 9Cr steel (MARBN) has been alloy-designed on the
base of microstructure stabilization, especially in the vicinity of PAGB.
Creep-strengthening by nitrides as well as boron and no formation of boron nitrides during
normalizing heat treatment cause a significant improvement of creep rupture strength. The
creep rupture strength at 650oC and 105 h is estimated to be 80 to 90 MPa, which is much
higher than that of existing high strength steels P92 and P122.

(2) The addition of boron without any formation of boron nitrides during normalizing heat
treatment suppresses grain refinement in HAZ of welded joints and hence suppresses the
Type IV fracture in welded joints. This satisfies the criterion that the weld strength
reduction factor (WSRF) of 105 h creep rupture strength should be larger than 0.75 at 650oC.
It should be noted that the optimization of weld metal will be required in future.

(3) The formation of protective Cr-rich scale is achieved on the surface of 9Cr steel by the
pre-treatments, such as pre-oxidation treatment and coating. This significantly improves the
oxidation resistance of 9Cr steel in steam.

(4) The production of a large diameter and thick section pipe having a size of 470 mm in
outer diameter, 65 mm in thickness and 1300 mm in length and also fabrication of
circumference welds of the pipe have successfully been performed from a 3 ton ingot of
MARBN.

Acknowledgement

The authors would like to thank all members participating with Ultra-Steel Project at NIMS.
They are also grateful to members of the Creep Group at NIMS for their sincere efforts.

References

1. R. Blum and R. W. Vanstone, “Materials Development for Boilers and Steam Turbines
Operating at 700oC”, Proc. of 8th Liege Conference on Materials for Advanced Power
Engineering 2006, Liege, Belgium, (2006) pp.41-60.

14
 2. R. Viswanathan, J. F. Henry, J. Tanzosh, G. Stanko, J. Shingledecker and B. Vitalis, “U.S.
Program on Materials Technology for Ultrasupercritical Steam Coal-fired Power Plants”,
as in (1) pp.893-915.
3. F. Abe, “Metallurgy for Long-term Stabilization of Ferritic Steels for Thick Section Boiler
Components in USC Power Plant at 650oC”, as in (1) pp.965-980.
4. M. Tabuchi and Y. Takahashi, “Evaluation of Creep Strength Reduction Factors for
Welded Joints of Modified 9Cr-1Mo steel (P91)”, Proc. of 2006 ASME PVP Conference,
Vancouver, Canada, PVP2006-ICPVT11-93350 (2006).
5. H. Semba and F. Abe, “Alloy design and creep strength of advanced 9%Cr USC boiler
steels containing high concentration of boron”, Energy Materials, 1 (2006) 238-244.
6. T. Horiuchi, M. Igarashi, F. Abe, “Improved Utilization of Added B in 9Cr Heat-Resistant
Steels Containing W”, ISIJ International, 42 (2002) S67-S71.
7. K. Sakuraya, H. Okada and F. Abe, “BN type inclusions formed in high Cr ferritic heat
resistant steel” Energy Materials, 1 (2006) 158-166.
8. M. Kimura, K. Kobayashi and K. Yamaguchi, “Creep and fatigue properties of newly
developed ferritic heat-resisting steels for ultra super critical (USC) power plants“,
Materials Science Research Intern., 9 (2003) 50-54.
9. M. Lucco Borlera and G. Pradelli, “Iron-carbon-boron ternary phase diagram”, Met. Ital.,
59 (1967) 907.
10. L. Karsson and H. Norden, “Non-equilibrium grain boundary segregation of boron in
austenitic stainless steels-II. Fine segregation behaviour” Acta Metall., 36 (1988) 13-24.
11. M. Y. Wey, T. Sakuma and T. Nishizawa, ‘Growth of alloy carbide particles in austenite’,
Trans. JIM, 22 (1981) 733-742.
12. F. Abe, M. Tabuchi, M. Kondo and H. Okada, “Suppression of Type IV fracture in
welded joints of advanced ferritic power plant steels - effect of boron and nitrogen”,
Materials at High Temperatures 23 (2006) 145-154.
13.H. Kutsumi, T. Itagaki and F. Abe, “Improvement of Steam Oxidation Resistance for
Ferritic Heat Resistant Steels”, Proceedings of 7th Liege Conference on Materials for
Advanced Power Engineering 2002, edited by J. Lecomte-Beckers, M. Carton, F.
Schubert and P. J. Ennis, Liege, Belgium (2002) pp.1629-1638.
14. T. Itagaki, H. Kutsumi, H. Haruyama, M. Igarashi and F. Abe, “Steam Oxidation of
High-Chromium Ferritic Steels Containing Palladium”, Corrosion, 61 (2005) 307-316.
15. H. Kutsumi, H. Haruyama and F. Abe, “Application of the Pre-Oxidation Treatment in Ar
Gas to the NIMS High-Strength Steels”, Proceedings of 4th International Conference on
Advances in Materials Technology for Fossil Power Plants, Hilton Head Island, SC, USA
(2004) pp.463-471.
16. H. Haruyama, H. Kutsumi, S. Kuroda and F. Abe, “Effect of Shot Peening and Pre-
Oxidation Treatment in Air on Steam Oxidation Resistance of Mod.9Cr-1Mo Steel”, as in
(15) pp.412-419.
17. T. Sundararajan, S. Kuroda and F. Abe, “Steam Oxidation of 80Ni-20Cr High-Velocity
Oxyfuel Coating on 9Cr-1Mo steel: Diffusion-Induced Phase Transformation in the
Substrate Adjacent to the Coating”, Metallurgical and Materials Transactions A, 36A
(2005) 2165-2174.
18. F. Abe, H. Tanaka and M. Murata, “Impurity Effects on Heat-to-Heat Variation in Creep
Life for Some Heat Resistant Steels”, Proceedings of BALTICA VII International
Conference on Life Management and Maintenance for Power Plants, Helsinki, 12-14 June
2007 (2007) pp.171-184.

15

View publication stats