Materials Transactions, Vol. 46, No. 8 (2005) pp. 1882 to 1889
#2005 The Japan Institute of Metals
Corrosion Behavior of Hastelloy-XR Alloy in O2 and SO2 Atmosphere
Rong Tu and Takashi Goto
Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan
As Hastelloy-XR alloy is a candidate structural material for the IS (Iodine-Sulfur) process in hydrogen production, oxidation and
sulfidation of Hastelloy-XR alloy in Ar–O2 and Ar–SO2 atmospheres were studied by thermogravimetry at temperatures from 1000 to 1300 K. In
Ar–O2 atmosphere, the mass change obeyed a linear-parabolic law at oxygen partial pressures (PO2 ) from 0.01 to 10 kPa. The oxidation scales
consisted of inner Cr2 O3 layer and outer Mn1:5 Cr1:5 O4 spinel layer. The surface morphology of the oxide scales changed from island-like to
buckled and to porous texture with decreasing PO2 . In Ar–SO2 atmosphere, the mass change obeyed a linear-parabolic law at SO2 partial
pressures (PSO2 ) from 0.05 to 5 kPa. The morphology of corrosion scales changed mainly with corrosion temperature. While oxidation was
dominant at 1073 and 1173 K forming double-layer scales of inner Cr2 O3 and outer Mn1:5 Cr1:5 O4 spinel, sulfidation was accompanied with
oxidation at 1273 K and PSO2 < 0:5 kPa with scales consisting of Fe3 O4 , FeCr2 O4 and Cr2 O3 layers and Ni3 S2 dispersed particles together with
CrS particles segregating at the grain boundary of Hastelloy-XR alloy.
(Received May 10, 2005; Accepted June 9, 2005; Published August 15, 2005)
Keywords: Hastelloy-XR alloy, SO2 gas, oxidation, sulfidation, corrosion, Iodine-Sulfur process
1.
Introduction
Since hydrogen can be a new clean energy source to reduce
CO2 emission and the green house warming, it is a key issue
to develop a mass production process of hydrogen. Hydrogen
can be produced by a thermochemical decomposition of
water, and recently IS (Iodine-Sulfur) process has succeeded
to produce hydrogen from water in a continuosly operated
closed cycle. IS process is fundamentally composed of three
processes, i.e., sulfuric acid (H2 SO4 ) decomposition around
1123 K, Bunsen reaction at 373 K and hydrogen iodide
decomposition at 473 to 573 K.1,2) In the process of H2 SO4
decomposition, SO3 and H2 O are produced, and then SO3
decomposed into SO2 and O2 . Such mixing atmosphere of
SO2 and O2 is significantly corrosive to structural metallic
materials.3) Thus, in bench-scaled IS process, the tubes and
vessels have been made of quartz glass and polymers (mainly
Teflon).4) However, in order to develop large-scaled industrial chemical plants, commercially available heat-resistant
alloys should be employed to construct the facility. The hightemperature mechanical properties and corrosion resistance
of many candidate alloys have been investigated in such
harsh environment,5,6) and then Hastelloy-X alloy (49Ni–
21Cr–18Fe–9Mo–0.6Mn–0.8Si, in mass %) has been expected as the structural material because of its high-temperature
strength and creep resistance.7) However, this alloy had not
enough corrosion resistance in IS process. On the other hand,
to improve the oxidation resistance of Hastelloy-X alloy in a
helium cooled very high temperature nuclear reactor
(VHTR), the contents of Mn and Si were optimized and
Hastelloy-XR alloy (49.47Ni–21.99Cr–17.80Fe–8.73Mo–
0.88Mn–0.33Si, in mass %) has been developed. The
oxidation behavior of Hastelloy-X alloy has been reported
by Wlodek et al.,8) Charlot9) and Shindo et al.7,10) However,
the oxidation behavior of Hastelloy-XR alloy has not been
investigated in wide-ranged temperatures and oxygen partial
pressures. Moreover, no corrosion resistance of HastelloyXR alloy in SO2 atmospheres at high temperatures has been
studied. In the present study, the corrosion behavior of
Hastelloy-XR alloy as a candidate structural material for IS
process was investigated in O2 and SO2 atmospheres at high
temperatures. Since SO2 would cause oxidation or sulfidation
depending on temperature and PSO2 , the oxidation in O2
atmosphere was separately examined to understand the
corrosion (oxidation and/or sulfidation) in SO2 atmosphere.
2.
Experimental
Hastelloy-XR alloy disks (10 mm in diameter by 1 mm
thickness) were polished with an alumina paste (1 mm) and
supersonically cleaned in acetone. The specimens were
exposed to Ar–O2 and Ar–SO2 atmospheres at oxygen partial
pressure (PO2 ) and SO2 partial pressure (PSO2 ) from 0.01 to
10 kPa at temperatures from 1000 to 1300 K for 43 ks. Ar–O2
and Ar–SO2 mixture gases were introduced from the bottom
of a reaction tube at a flow rate of 6:7 106 m3 s1 . The
total pressure in the reaction tube was fixed at 0.1 MPa. The
mass changes were continuously measured by thermogravimetry (A&T: HA202M, sensitivity: 10 mg). Crystal
structure, microstructure and composition of corrosion scales
were analyzed by X-ray diffraction (XRD, Rigaku: RAD-C),
scanning electron microscopy (SEM, Hitachi: S-3100H),
transmission electron microscopy (TEM, JEOL: EX-II) and
electron probe microanalysis (EPMA, JEOL: JXA-8621MX).
The potential diagrams of Fe–S–O, Ni–S–O and Cr–S–O
system were calculated by using the thermodynamic database
to estimate the corrosion behavior in Ar–SO2 atmosphere at
high temperatures.11,12)
3.
Results and Discussion
3.1 Oxidation behavior in Ar–O2 atmosphere
Figure 1 presents the XRD pattern of Hastelloy-XR alloy
after the oxidation for 43 ks at 1273 K and PO2 ¼ 1 kPa. The
oxide scales consisted of Cr2 O3 and Mn1:5 Cr1:5 O4 spinel
phases, and such two-phase oxide scales were observed
independent of oxidation conditions.
Figure 2 depicts the cross-sectional back-scattering SEM
image and EPMA analyses for the scales formed at 1273 K
and PO2 ¼ 10 kPa. The oxide scale of 4 mm in thickness
consisted of double layers with inner Cr2 O3 and outer
Mn1:5 Cr1:5 O4 spinel.
Corrosion Behavior of Hastelloy-XR Alloy in O2 and SO2 Atmosphere
Intensity (a. u.)
Mn1.5Cr1.5O4
Cr2O3
Hastelloy-XR
10°
20°
30°
40°
50°
60°
1883
(a)
70°
2θ / (Cu kα)
100 µm
Fig. 1 X-ray diffraction pattern of the scale on the Hastelloy-XR alloy after
the oxidation for 43 ks at PO2 ¼ 1 kPa and at 1273 K.
(b)
10 µm
Hastelloy-XR
20 µm
Mn1.5Cr1.5O4
(c)
Cr2O3
Fe
Ni
Mn
Cr
O
Fig. 2 Cross-sectional back-scattering SEM image and EPMA analyses of
the scale on the Hastelloy-XR alloy after the oxidation for 43 ks at 1273 K
and PO2 ¼ 10 kPa.
Figures 3(a) to (c) demonstrate the surface morphology of
oxide scales on the Hastelloy-XR alloy after the oxidation at
1273 K and PO2 from 0.01 to 10 kPa. Although the scales
always consisted of Cr2 O3 and Mn1:5 Cr1:5 O4 spinel layers,
the surface morphology significantly changed with PO2 .
Island-like scales were observed at PO2 ¼ 10 kPa [Fig. 3(a)].
It was confirmed by XRD that the islands in Fig. 3(a) was
Mn1:5 Cr1:5 O4 spinel and the flat area was Cr2 O3 by
mechanically removing the islands. During the cooling
process, sudden mass drops were often observed, implying
the island-like scale was resulted from the partial delamination of spinel outer layer due to thermal expansion mismatch
between Cr2 O3 and spinel layers. Buckled oxide scales were
observed at PO2 ¼ 0:1 kPa [Fig. 3(b)], which could be also
10 µm
Fig. 3 Surface SEM images of scales on the Hastelloy-XR alloy after the
oxidation for 43 ks at 1273 K, PO2 ¼ 10 (a), 0.1 (b) and 0.01 kPa (c).
caused of the thermal expansion mismatch between two
layers with the spinel layer remaining on the Cr2 O3 layer.
Well-adhered uniform scales with a large amount of micropores were formed at PO2 ¼ 0:01 kPa [Fig. 3(c)]. The micropores could relax the thermal stress between two layers.
Ecer et al. conducted a marker test for the oxidation of Ni–
Cr alloy at 1073 to 1373 K.13) The platinum marker stayed at
the interface between alloy and oxide scales. This implies
that the outward diffusion of cations could dominate the
oxidation of Ni–Cr alloys. Since Ni and Cr are major
components in Hastelloy-XR alloy, the similar oxidation
mechanism (i.e., outward diffusion of cations) can be
dominant in the present study.
Figure 4 presents the cross-sectional back-scattering SEM
image after the oxidation for 432 ks at 1273 K and PO2 ¼
10 kPa. The thickness of Cr2 O3 inner layer increased
1884
R. Tu and T. Goto
10-2
2
10 µm
Mass change, ∆M / kg m-2
1
10-3
10
-4
PO2 / kPa
0.01
0.1
1
10
1
1
Hastelloy-XR
Cr2O3
Mn1.5Cr1.5O4
10
-5
10
Mn1.5Cr1.5O4
Cr2O3
Fig. 4 Cross-sectional back-scattering SEM image of the scale on
Hastelloy-XR alloy after the oxidation for 432 ks at 1273 K and PO2 ¼
10 kPa.
2
10
3
10
4
10
5
Time, t / s
Fig. 6 Relationship between mass change and time for the oxidation of
Hastelloy-XR at 1273 K in Ar–O2 atmosphere.
Temperature, T / K
200 nm
Fig. 5 Cross-sectional TEM image of the scale on the Hastelloy-XR alloy
after the oxidation for 43 ks at 1273 K and PO2 ¼ 0:01 kPa.
significantly with time, and the slight increase in the spinel
outer layer was observed. Douglass et al.14) and Shindo et
al.7) have obtained the similar results for the oxidation of Ni–
Cr–Mn alloys. Douglass et al. reported that the double-layer
scales of inner Cr2 O3 and outer MnCr2 O4 spinel were formed
on a Ni–20Cr–1Mn alloy after the oxidation at 1373 K in
air.14) Shindo et al. reported that the increase in Cr2 O3 inner
layer thickness was much faster than MnCr2 O4 outer layer in
the oxidation of Hastelloy-X alloys and suggested that the
oxidation reaction should proceed mainly at the interface between inner and outer layers.7) Present study clearly indicated
that the growth rate of Cr2 O3 inner layer was significantly
greater than that of Mn1:5 Cr1:5 O4 outer layer, implying that
the oxidation of Hastelloy-XR alloy in the later stage was
mainly dominated by outward diffusion of Cr3þ ion.
Figure 5 presents the TEM micrograph for the crosssection of the oxide scale formed at 1273 K and PO2 ¼
0:01 kPa after 43 ks. The double-layer structure with inner
Cr2 O3 and outer Mn1:5 Cr1:5 O4 spinel was observed. The
diffusion rates of metallic ions in oxides including Cr2 O3
were reported as Fe3þ , Mn2þ > Fe2þ > Ti3þ > Co2þ >
Ni2þ > Mn3þ > Cr3þ .15) It is assumed that the spinel layer
formed at the outside of Cr2 O3 due to faster diffusion of
Mn2þ ion than that of Cr3þ ion particularly in an early stage
Log(Linear rate constant, kl / kg m-2 s-1)
Hastelloy-XR
-5
1500
1400
1300
1200
1100
1000
PO2 / kPa
0.01
0.1
1
10
20 (air)8)
-6
-7
-8
7
8
9
10
Reciprocal of temperature, T-1 / 10-4K-1
Fig. 7 Arrhenius plots of linear rate constants for the oxidation of
Hastelloy-XR alloys.
of oxidation.15,16) The thickness of outer layer was thicker
than inner layer.
Figure 6 shows the mass change of Hastelloy-XR alloy as
a function of oxidation time at 1273 K. The mass change
increased with increasing time and PO2 . The time dependence
of mass change obeyed a linear-parabolic law at PO2 from
0.01 to 10 kPa. The linear (kl ) and parabolic rate constants
(kp ) for the oxidation of Hastelloy-XR alloy were calculated
separately from the time dependence of mass change.
Figure 7 depicts the Arrhenius plots of linear rate constants
(kl ) for the oxidation of Hastelloy-XR alloy. The kl increased
with increasing the oxidation temperature and PO2 . Wlodek
studied the oxidation of Hastelloy-X alloy in air from 1140 to
1470 K and reported the linear-parabolic mass change
behavior at less than 1255 K.8) The transition from linear to
parabolic behavior occurred between 12 and 36 ks, which
Corrosion Behavior of Hastelloy-XR Alloy in O2 and SO2 Atmosphere
Log(Parabolic rate constant, kp / kg2 m-4 s-1)
Temperature, T / K
-7
1500
1400
1300
1200
1000
PO2 / kPa
0.01
0.1
1
10
-8
-9
(a)
1100
20 kPa(air) 8)
1885
PO2 / kPa
3.3
0.44
0.013
0.005
1393 K 9)
-10
-11
50 µm
PO2=10-17~10-20 Pa 7)
(b)
-12
7
8
9
10
Reciprocal of temperature, T-1 / 10-4K-1
Fig. 8 Arrhenius plots of parabolic rate constants for the oxidation of
Hastelloy-XR alloys.
was longer than that in the present study (2 to 10 ks), and the
kl obtained by Wlodek (PO2 ¼ 20 kPa) were close to that of
PO2 ¼ 0:1 kPa in the present study. In the present study, the
time for starting the parabolic behavior increased with
decreasing PO2 , which coincides the general trend of linear to
parabolic transition.17) These suggest that Hastelloy-XR alloy
can form protective oxide scale more easily than Hastelloy-X
alloy.
Figure 8 depicts the Arrhenius plots of parabolic rate
constants (kp ) for the oxidation of Hastelloy-XR alloy. The kp
values increased with increasing temperature and PO2 .
Wlodek obtained the kp values of Hastelloy-X alloy in air8)
which were almost the same as those of Hastelloy-XR alloy
in PO2 ¼ 10 kPa. Charlot et al. reported that the kp values of
Hastelloy-X alloy increased with increasing PO2 from 5 Pa to
3.3 kPa at 1393 K.9) Their values are almost in agreement
with the present values extrapolated to higher temperatures.
Shindo et al. obtained slightly lower kp values of Hastelloy-X
alloy than the present results.7) They studied in a helium
atmosphere containing a small amount of impurity (water
vapor and carbon dioxide), in which PO2 was estimated as
1017 to 1020 Pa. Charlot et al.9) and Shindo et al.7) reported
that the activation energy (Ea ) for the parabolic oxidation of
Hastelloy-X alloy was 234 kJ mol1 , which is almost in
agreement to that for the diffusion of Cr3þ in Cr2 O3 reported
by Giggins et al. (255 kJ mol1 ).18) Charlot et al. and Shindo
et al. implied that the rate-controlling step for the parabolic
oxidation of Hastelloy-X alloy could be the diffusion of Cr3þ
in Cr2 O3 layer. In the present study, the Ea for the parabolic
oxidation of Hastelloy-XR alloy was 220 kJ mol1 , which
almost agreed with those of Hastelloy-X alloy reported by
Shindo et al.7) and Wlodek (238 kJ mol1 ).8) Therefore, the
diffusion of Cr3þ in Cr2 O3 could be the rate-controlling step
for the parabolic oxidation of Hastelloy-XR alloy.
3.2 Corrosion behavior in Ar–SO2 atmosphere
Oxide scales consisting of Cr2 O3 and Mn1:5 Cr1:5 O4 spinel
2 µm
Fe
Ni
Mn
Cr
O
S
Fig. 9 Surface (a) and cross-sectional (b) SEM images of the scale on the
Hastelloy-XR alloy after the corrosion for 43 ks at 1073 K and PSO2 ¼
1 kPa.
were formed on the Hastelloy-XR alloy after the corrosion in
Ar–SO2 atmosphere between 1073 and 1273 K, as observed
in Ar–O2 atmosphere. Figure 9 depicts the surface, crosssectional SEM images and EPMA analyses of the scale on the
Hastelloy-XR alloy after the corrosion for 43 ks at 1073 K
and PSO2 ¼ 1 kPa. Buckled scales were observed on the
surface, being almost the same as that formed at 1273 K and
PO2 ¼ 0:1 kPa in Ar–O2 atmosphere as shown in Fig. 3(b).
The EPMA analysis showed that the buckled scale consisted
of inner Cr2 O3 layer and outer Mn1:5 Cr1:5 O4 layer, and a
1886
R. Tu and T. Goto
2
(a)
Mass change, ∆M / kg m-2
10-2
10-3
1073 K, PSO2=0.1 kPa
1073 K, PSO2=1 kPa
1273 K, PSO2=0.1 kPa
1273 K, PSO2=1 kPa
1
1
1
10-4
100 µm
102
(b)
103
104
105
Time, t / s
Fig. 11 Relationship between mass change and time for the corrosion of
Hastelloy-XR in Ar–SO2 atmosphere.
20 µm
(c)
Hastelloy
-XR
CrS
Ni3S2
CrS
Ni3S2
.
E
D
C
B A
Fig. 10 Surface SEM (a), cross-sectional back-scattering SEM image (b)
and a schematic of cross-section of the scale on the Hastelloy-XR alloy (c)
after the corrosion for 43 ks at 1273 K and PSO2 ¼ 0:1 kPa. (A:
Fe3 O4 (Ni3 S2 ), B: FeCr2 O4 (Ni3 S2 ), C: Cr2 O3 , D: Ni–Fe–Mo(CrS), E:
Hastelloy-XR)
small amount of sulfur was identified near the surface of
Hastelloy-XR alloy. No CrS or Ni3 S2 was detected by XRD.
At 1273 K, the corrosion scales showed two kinds of
microstructures depending on PSO2 . At PSO2 > 0:5 kPa, the
scale was almost the same as that shown at Fig. 9. At
PSO2 < 0:5 kPa, on the other hand, the scale had a multi-layer
microstructures with dispersions of sulfides as shown in
Fig. 10, where the surface SEM (a), cross-sectional back-
scattering SEM image (b) and a schematic of cross section (c)
of the scale on the Hastelloy-XR alloy after the corrosion for
43 ks at 1273 K and PSO2 ¼ 0:1 kPa were demonstrated.
Partially delaminated scale with many bumps was observed
[Fig. 10(a)]. The corrosion scale consisted of Fe3 O4 with
Ni3 S2 particles (layer A), FeCr2 O4 with Ni3 S2 particles
(layer B), Cr2 O3 (layer C), Ni–Fe–Mo metal with CrS
particles (layer D) and Hastelloy-XR substrate (layer E) as
depicted in Fig. 10(c).
It was reported that continuous Cr2 O3 layers were formed
in SO2 atmosphere at high temperatures for Ni–Cr alloys
containing high-content Cr.19,20) Hancock et al. studied the
corrosion of Ni–20Cr alloy in SO2 atmosphere at 1173 K, and
reported that the Cr2 O3 scale contained Ni3 S2 particles, and
the internal sulfidation caused the segregation of CrS at the
grain boundary in the alloy.19) Zurek et al. reported that the
mass change of Ni–22Cr–10Al–1Y alloy almost obeyed a
parabolic law, and scales consisted of Cr2 O3 , Al2 O3 , NiO,
NiCr2 O4 and a small amount of AlS at 1173 to 1273 K in SO2
atmosphere.20) In the present study, continuous Cr2 O3 layers
were formed on Hastelloy-XR alloy because Hastelloy-XR
alloy contained enough amount of Cr. The dispersion of
Ni3 S2 particles in the oxide scales and the segregation of CrS
at the grain boundary in the present study were similar to the
results by Hancock19) and Zurek.20) The corrosion of
Hastelloy-XR alloy became more significant at high temperature and low sulfur potential (at 1273 K and PSO2 <
0:5 kPa), and CrS was observed at the grain boundary of
Hastelloy-XR alloy due to the internal sulfidation of Cr. This
behavior was similar to that of Ni–Cr, Fe–Cr and Co–Cr
alloys in H2 S–H2 atmosphere studied by Narita et al.21) They
reported that Cr-containing alloys were hardly sulfidized at
high sulfur potentials, however sulfides were easily formed at
low sulfur potentials. At low sulfur potentials, copious
internal sulfidation (CrSx layer) was formed for Ni–Cr alloys,
whereas sulfidation was confined to grain boundary for Fe–Cr
alloys. As Hastelloy-XR alloy is a Ni–Cr–Fe based alloy, the
internal sulfidation mainly occurred at the grain boundary
and no CrSx layer was formed.
Figure 11 demonstrates the relationship between the mass
change and time for the corrosion of Hastelloy-XR alloy at
Corrosion Behavior of Hastelloy-XR Alloy in O2 and SO2 Atmosphere
Log(Parabolic rate constant, kp / kg2 m-4 s-1)
Log(Linear rate constant, kl / kg m-2 s-1)
-5
T/K
1073
1173
1273
-6
-7
-8
1
2
3
-8
-9
-10
T/K
1073
1173
1273
-11
1
4
2
Log(Parabolic rate constant, kp / kg2 m-4 s-1)
Log(Linear rate constant, kl / kg m-2 s-1)
-5
T/K
1073
1173
-6
-7
-8
T/K
1073
1173
-9
-10
-11
-8
-7
4
(a)
(a)
-8
3
Log(Partial pressure of SO2, PSO2 / Pa)
Log(Partial pressure of SO2, PSO2 / Pa)
-8
1887
-6
-7
-6
Log(Partial pressure of O2, PO2 / Pa)
Log(Partial pressure of O2, PO2 / Pa)
(b)
(b)
Fig. 12 (a) Relationship between linear rate constant (kl ) of Hastelloy-XR
alloy and PSO2 from 1073 to 1273 K. (b) Relationship between linear rate
constant (kl ) of Hastelloy-XR alloy and PO2 decomposed by SO2 from
1073 to 1173 K.
Fig. 13 (a) Relationship between parabolic rate constant (kp ) of HastelloyXR alloy and PSO2 from 1073 to 1273 K. (b) Relationship between
parabolic rate constant (kp ) of Hastelloy-XR alloy and PO2 decomposed by
SO2 from 1073 to 1173 K.
1
GT ¼ RT lnðPS2 2 PO2 =PSO2 Þ
1073 to 1273 K and PSO2 ¼ 0:1 to 1 kPa. The mass change
obeyed a linear-parabolic law at the whole conditions, and
increased with increasing temperature. The linear to parabolic transition occurred at about 20 ks, which was longer
than that of oxidation (2 to 10 ks in Fig. 6) implying that it is
more difficult to form protective scales in Ar–SO2 rather than
in Ar–O2 atmosphere. At 1073 K, the mass change at
PSO2 ¼ 1 kPa was higher than that at PSO2 ¼ 0:1 kPa, but
the trend was opposite at 1273 K. Since the oxidation was
dominant at 1073 K as above-mentioned, the mass change
increased with increasing PO2 calculated from the decomposition of SO2 as given by eqs. (1) to (4). When a
temperature increases from 1073 to 1273 K at PSO2 ¼
0:1 kPa, the equilibrium PS2 increases from 4:3 1012 to
2:8 1010 kPa and PO2 increases from 8:6 1012 to 5:6 1010 kPa.
SO2 ¼
1
S2 þ O2
2
ð1Þ
PO2 ¼ 2PS2
ð2Þ
ð3Þ
therefore,
PO 2
2
GT 3
¼ 1:414PSO2 exp
RT
ð4Þ
Figure 12(a) shows the relationship between linear rate
constant (kl ) and PSO2 for the corrosion of Hastelloy-XR alloy
in Ar–SO2 atmosphere. At 1073 and 1173 K, the kl increased
with increasing PSO2 , but decreased with increasing the PSO2
at 1273 K [Fig. 12(a)]. As the corrosion of Hastelloy-XR
alloy at 1073 and 1173 K was almost the same as the behavior
of oxidation as described in 3.1, the relationship between
kl and PO2 calculated from eq. (4) was demonstrated in
Fig. 12(b). The kl slightly increased with increasing PO2 .
Figure 13(a) shows the relationship between parabolic rate
constant (kp ) and PSO2 for the corrosion of Hastelloy-XR
alloy. The trend of kp was almost the same as that of kl . The
relationship between kp and PO2 calculated from eq. (4) was
1888
R. Tu and T. Goto
depicted in Fig. 13(b). The linear relation was observed at
1=n
1073 and 1173 K where kp / PO
(n ¼ 5 to 6). It is generally
2
understood that Cr2 O3 was a p-type semiconductor, and the
defect formation reaction could be expressed as eq. (5).
3
000
O2 ¼ 3OO þ 2VCr
þ 6h
2
ð5Þ
The equlibrium constant for reaction (5) is given by eq. (6)
because the concentration of vacancyes formed is proportional to the concentration of electron holes.
Keq ¼
000 2 6
000 6
ðVCr
Þ ðh Þ
ðV 000 Þ2 ð3VCr
Þ
¼ Cr
3=2
3=2
ðPO2 Þ
ðPO2 Þ
ð6Þ
Therefore,
3=16
000
VCr
/ const. PO
2
ð7Þ
The 3/16-power relationship can be predicted in the
oxidation where the Cr2 O3 formation is the rate controlling
1=n
step. Charlot et al. reported the relationship of kp / PO
2
9)
(n ¼ 5). Present results of n ¼ 5 to 6 can be close to that of
Charlot et al.9) and above-mentioned calculated results.
The potential diagram can be useful to understand the
corrosion behavior and the formation of oxide or sulfide
scales in SO2 atmosphere. The corrosion in SO2 atmosphere
may occur by dissociated O2 or S2 , thus the oxygen and sulfur
potentials may determine the corrosion process. We have
calculated the potential diagrams by using thermodynamic
database to compare with the experimental results. Figure 14
represents the potential diagrams for Ni–S–O, Fe–S–O and
Cr–S–O systems at 1273 K, where the three diagrams were
superposed in one diagram. The PS2 and PO2 at PSO2 ¼
0:1 kPa are 1:3 1010 and 2:6 1010 kPa, respectively.
These values correspond to point P in Fig. 14 and located in
the stable region of Cr2 O3 , Fe3 O4 and Ni. The outermost
layer of the corrosion scale consisted of Fe3 O4 with Ni3 S2
particles and no NiO [layer A in Fig. 10(c)]. In the layer A,
the PO2 can be in the stable area of Fe3 O4 and Ni, and the PS2
can be in the stable area of Ni3 S2 . Therefore, the PO2 and PS2
in the outermost layer (layer A) can be in the area A in
Fig. 14. The second layer of the corrosion scale consisted of
FeCr2 O4 and Ni3 S2 particles [layer B in Fig. 10(c)]. In the
Cr
Fe
Cr2(SO4)3(s)
Ni
NiS(s)
Cr3S4(s)
FeSO4(s)
layer B, the PO2 can be in the stable area of FeO and Cr2 O3
because FeCr2 O4 may be formed by the reaction of FeO and
Cr2 O3 . The PS2 can be in the stable area of Ni3 S2 . Therefore,
the PO2 and PS2 in layer B can be in the stable area of FeO,
Cr2 O3 and Ni3 S2 (area B in Fig. 14). The PO2 and PS2 in the
third layer [layer C in Fig. 10(c)] can be in the stable area of
Ni, Fe and Cr2 O3 (area C in Fig. 14) because the layer
consisted of Ni, Fe and Cr2 O3 . Since the layer D in Fig. 10(c)
consisted of Ni and Fe with CrS particles, the PO2 and PS2 can
be in the stable area of Ni, Fe and CrS (area D in Fig. 14).
The PO2 and PS2 in the layer E of Fig. 10(c) can be in the
stable area of Cr, Ni and Fe (area E in Fig. 14) because the
layer was non-corroded Hastelloy-XR alloy. Since the PO2
and PS2 in the corrosion scale may continuously decrease
from the outermost to inside, the PO2 and PS2 can be possibly
changed along the broken line from point A to E in Fig. 14.
The PS2 at the outermost of corrosion scale can be higher than
the initial S2 partial pressure. Since O2 may be consumed
significantly at the outermost layer, PS2 may be relatively
increased more than the initial PS2 . Therefore, the calculated
potential diagrams could be applicable to understand the
corrosion behavior in Ar–SO2 for Hastelloy-XR alloy.
4.
Conclusions
The corrosion behavior of Hastelloy-XR alloy in O2 and
SO2 atmospheres were investigated in the temperature range
between 1073 and 1273 K. In Ar–O2 atmosphere, the mass
changes mainly obeyed a linear-parabolic law. The corrosion
scales had a double-layer structure of inner Cr2 O3 and outer
Mn1:5 Cr1:5 O4 spinel layers. The surface morphology of scales
changed from island-like to buckled and to smooth porous
layers with decreasing PO2 from 10 to 0.01 kPa. In Ar–SO2
atmosphere, the mass changes obeyed a linear-parabolic law.
The corrosion scales formed at 1073 and 1173 K were similar
to the scales formed in Ar–O2 atmosphere at 1073 K, because
the oxidation was dominant at these low temperatures. At
1273 K and PSO2 < 0:5 kPa, the corrosion scales were
consisted of multi-layers changing from Fe3 O4 with Ni3 S2
particles (outmost), FeCr2 O4 , Cr2 O3 , and CrS particles
(innermost). The sulfidation became more significant in less
PSO2 at 1273 K. The thermodynamic potential diagrams may
be useful to understand the corrosion behavior of HastelloyXR alloy in Ar–SO2 atmosphere.
Fe2(SO4)3(s)
2
Acknowledgments
Ni3S2(l)
log(PS2 / kPa)
-3
-8
D
C
P
E
Fe3O4(s)
Fe(s)
Cr(s)
FeO(s)
REFERENCES
Ni(s) NiO(s)
-18
-23
-33
A part of this study was supported by Tokyu Foundation
for Inbound Student. The authors thank Mr. Y. Murakami of
Laboratory for Advanced Materials, Institute for Materials
Research for EPMA analyses.
B
CrS(s)
Cr2O3(s)
-13
NiSO4(s)
A
FeS(s)
Fe2O3(s)
-28
-23
-18
-13
-8
-3
2
log(PO2 / kPa)
Fig. 14 Potential diagrams for Ni–S–O, Fe–S–O and Cr–S–O systems at
1273 K.
1) M. Sakurai, H. Nakajima, K. Onuki, K. Ikenoya and S. Shimizu: Int. J.
Hydrogen Energy 24 (1999) 603–612.
2) A. Hammache and E. Bilgen: J. Energ. Resources Tech. 114 (1992)
227–234.
3) Y. Kurata, K. Tachibana and T. Suzuki: J. Jpn. Inst. Met. 65 (2001)
262–265.
4) T. Goto: Mater. Jpn. 40 (2001) 368–371.
Corrosion Behavior of Hastelloy-XR Alloy in O2 and SO2 Atmosphere
5) Y. Imai, S. Mizuta and H. Nakauchi: Boshoku Gijutsu 35 (1986) 230–
240.
6) F. Coen Porisini: Int. J. Hydrogen Energy 14 (1989) 267–273.
7) M. Shindo and T. Kondo: Tetsu to Hagane 68 (1982) 1628–1637.
8) S. T. Wlodek: Trans. Metall. AIME 230 (1964) 177–185.
9) L. A. Charlot and R. E. Westerman: Corrosion 23 (1967) 50–56.
10) M. Shindo and T. Kondo: Tetsu to Hagane 68 (1976) 1540–1549.
11) JANAF Thermochemical Tables, American Chemical Society and
American, Institute of Physics for National Bureau of Standards,
(1985).
12) I. Barin: Thermochemical Data of Pure Substances, VCH, Weinheim,
(1989).
13) G. M. Ecer and G. H. Meier: Oxid. Met. 13 (1979) 119–158.
1889
14) D. L. Douglass and J. S. Armijo: Oxid. Met. 2 (1970) 207–231.
15) M. G. C. Cox, B. Mcenaney and V. D. Scott: Philos. Mag. 26 (1972)
839–841.
16) R. K. Wild: Corros. Sci. 17 (1977) 87–93.
17) N. Birks and G. H. Meier: Introduction to High Temperature Oxidation
of Metals, Edward Arnold (publishers) Ltd., London, 1983, pp. 59–60.
18) C. S. Giggins and F. S. Pettit: Trans. Metall. AIME 245 (1969) 2495–
2507.
19) P. Hancock: 1st Int. Congr. Metall. Corros., Butterworths, London
(1961) pp. 193–198.
20) Z. Zurek, J. Jedlinski, K. Kowalski, V. Kolarik, W. Engel and J. Musil:
J. Mater. Sci. 35 (2000) 685–692.
21) T. Narita and T. Ishikawa: Mater. Sci. Eng. 87 (1987) 51–61.