You are on page 1of 20

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

net/publication/233568408

High Temperature Oxidation of the Austenitic (35Fe27Cr31Ni) Alloy Sanicro 28


in O2 + H2O Environment

Article  in  Oxidation of Metals · August 2010


DOI: 10.1007/s11085-010-9199-1

CITATIONS READS

17 182

6 authors, including:

T. Jonsson Christian Proff


Chalmers University of Technology Paul Scherrer Institut
61 PUBLICATIONS   970 CITATIONS    18 PUBLICATIONS   503 CITATIONS   

SEE PROFILE SEE PROFILE

Mats Halvarsson Jan-Erik Svensson


Chalmers University of Technology Chalmers University of Technology
154 PUBLICATIONS   3,170 CITATIONS    202 PUBLICATIONS   6,556 CITATIONS   

SEE PROFILE SEE PROFILE

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

Li ion battery materials studied in operando in SEM View project

High temperature corrosion of additively manufactured materials View project

All content following this page was uploaded by T. Jonsson on 15 April 2015.

The user has requested enhancement of the downloaded file.


Oxid Met (2010) 74:93–111
DOI 10.1007/s11085-010-9199-1

ORIGINAL PAPER

High Temperature Oxidation of the Austenitic


(35Fe27Cr31Ni) Alloy Sanicro 28 in O2 + H2O
Environment

C. Pettersson • T. Jonsson • C. Proff •


M. Halvarsson • J.-E. Svensson • L.-G. Johansson

Received: 5 August 2009 / Revised: 15 March 2010 / Published online: 18 April 2010
Ó Springer Science+Business Media, LLC 2010

Abstract The present study investigates the high temperature oxidation of alloy
Sanicro 28 (35Fe27Cr31Ni) in 5% O2 and in 5% O2 ? 40% H2O. Polished steel
coupons were isothermally exposed in a tube furnace at 600, 700 and 800 °C for up
to 168 h. The samples were investigated by gravimetry, grazing angle X-ray dif-
fraction (XRD), Auger electron spectroscopy (AES), scanning electron microscopy
(SEM), transmission electron microscopy (TEM) and scanning transmission elec-
tron microscopy/energy dispersive X-rays (STEM/EDX). The results show that the
material forms a protective scale in both environments. The scale is duplex. The
inner part of the scale consists of corundum type chromium-rich (CrxFe1-x)2O3, and
the outer layer consists of spinel type oxide. Chromia is lost from the protective
oxide by vaporization of CrO2(OH)2 in O2 ? H2O environment. The capacity of
Sanicro 28 to suffer chromia vaporization without forming a rapidly growing iron-
rich oxide is attributed to its high Cr/Fe ratio. The spinel formed at the oxide/gas
interface could in addition be beneficial for oxidation behavior in wet oxygen
because it may slow down chromia evaporation.

Keywords Oxidation  Sanicro 28  Water vapor  Evaporation

Introduction

Fe–Cr–Ni steels are known to form a dense, slow-growing chromium-rich oxide


scale at high temperature which provides protection against oxidation and corrosion.

C. Pettersson (&)  J.-E. Svensson  L.-G. Johansson


Department of Chemical and Biological Engineering, High Temperature Corrosion Centre,
Chalmers University of Technology, 412 96 Göteborg, Sweden
e-mail: tj@chalmers.se

T. Jonsson  C. Proff  M. Halvarsson


Department of Applied Physics, Chalmers University of Technology, 412 96 Göteborg, Sweden

123
94 Oxid Met (2010) 74:93–111

The supply of chromium to the scale, by diffusion in the alloy, has to be sufficient
for such a protective scale to form. Hence, depending on the corrosive environment
and the alloy microstructure, different concentrations of Cr in the alloy are
necessary for corrosion protection. In many applications, e.g. in power plants, the
environment contains a mixture of oxygen and water vapor.
The presence of water vapor is known to increase the corrosivity of the
environment. The critical Cr content needed to form a Cr rich (Fe,Cr)2O3 slow
growing scale has for example been shown to be higher in the presence of water
vapour [1, 2]. Several explanations have been suggested in order to explain the
higher corrosivity of O2/H2O mixtures [2–11]. The material transport mechanisms
in the alloy have been reported to change, normally increase, in the presence of
water vapour [3]. Hydrogen has been reported to become incorporated into the
oxide scale, changing the defect-dependent properties [5, 6]. Exposure to O2/H2O
mixtures in the temperature range 500–900 °C, have been reported to result in
chromium evaporation [7–13]. According to Ebbinghaus the most probable
vaporising species is CrO2(OH)2 in this temperature range [14]. The vaporization
of chromia from the scale tends to convert the protective chromium-rich oxide to
less protective hematite. The ability of an alloy to endure Cr vaporisation without
losing the protective properties of the Cr rich oxide is expected to depend on the rate
of supply of Cr from the alloy to the oxide scale.
The present study investigates the corrosion behavior of the highly alloyed
austenitic Sanicro 28 (35Fe27Cr31Ni) in dry and wet oxygen. Due to the high
chromium content of the alloy, corrosion attack is expected to be slower and more
easy to study, compared to other alloys, e.g. 304L. The focus of this work is on the
initial stages of the corrosion process, especially before breakdown of the protective
properties of the oxide scale. As described above, different mechanisms have been
suggested to explain the higher corrosivity in O2/H2O mixtures. A detailed analysis
with high spatial resolution is therefore needed to further investigate the thin
complex oxide scales formed in O2 and O2/H2O mixtures in order to better
understand the good oxidation behavior of alloys such as Sanicro 28 at 600–800 °C.
In this work, the detailed microstructural analysis was performed using a focused
ion beam (FIB) workstation in combination with a transmission electron microscope
(TEM) equipped with an energy dispersive X-ray (EDX) system. This made it
possible to analyze the complex oxide scales and describe the microstructure in
detail. The objective of this investigation is to link the microstructure of the oxide
scales formed in dry oxygen and in O2/H2O mixtures, i.e. under Cr evaporation, to
the oxidation mechanisms in order to understand the oxidation resistance of highly
alloyed alloys such as Sanicro 28 in the presence of O2/H2O mixtures.

Experimental

Exposure Set-Up and Sample Preparation

Table 1 shows the composition of the alloy studied. The samples used were coupons
with the dimensions 15 9 15 9 2 mm3. A hole (/ = 1.5 mm) was drilled for

123
Oxid Met (2010) 74:93–111 95

Table 1 Chemical composition of alloy Sanicro 28


Element Fe Cr Ni Mn Si Mo C Cu Cr/Fe

Sanicro 28 (wt%) 34 27 31 2.0 0.7 3.5 0.02 1.0 0.77

better handling. The exposures were carried out in a horizontal furnace fitted with a
46 mm i.d. SiO2-glass tube. All exposures were isothermal, the temperature was
kept at 600, 700 and 800 °C ± 2 °C. The samples were mounted three at a time
using an alumina sample holder with dimensions 40 9 50 9 2 mm3, featuring three
slits for mounting. The distance between the slits was 12 mm. The samples were
positioned parallel to the direction of the flow. Two different gas mixtures were
used: 5% O2 (\10 ppm H2O) and 5% O2 ? 40% H2O (N2 in balance). In the water
containing exposures the reaction gas was led through a humidifier producing 40%
water vapor. The gas velocity was 3 cm/s. The dry flow rate was calibrated using a
Bios DC2 Flow Calibrator. Before exposure the samples were ground to 1000 grit
SiC and polished with 1 lm diamond spray until the surface appeared mirror-like.
The polished samples were degreased and cleaned in acetone and ethanol using an
ultrasonic bath. The samples were stored in a desiccator prior to exposure. The
gravimetric measurements were made using a six decimal SartoriusTM balance.

X-Ray Diffraction (XRD)

Crystalline corrosion products were analyzed by X-ray diffraction (XRD) using a


Siemens D5000 powder diffractometer, equipped with grazing incidence beam
attachment and a Göbel mirror. CuKa radiation was used and the angle of incidence
was 0.5°. The detector measured between 20° \ 2h \ 65°.

Scanning Electron Microscopy (SEM)

Analytical scanning electron microscopy was carried out using an FEI Quanta 200
ESEM in high vacuum, secondary electron mode was used. The microscope was
operated at 10–12 kV for imaging.

Auger Electron Spectroscopy (AES)

The Auger analyses were performed using a Scanning Auger Microprobe (PHI 660).
The electron beam voltage was 10 kV and the beam current was about 150 nA. The
depth profiles were obtained using a differentially pumped ion gun (Ar?) with
acceleration voltage 4 kV. The etch rates were calibrated on flat samples of Ta2O5
with a well defined oxide thickness of 1000 Å. The collected raw data were refined
by using MultiPak v.6.0 software. The software made it possible to distinguish
between the signals from oxidized and metallic iron and chromium, based on the
chemical shifts of these elements in oxidized form. Due to the small peak shift in
nickel, it was not possible to divide the Ni signal into a metal state and an oxidized
state.

123
96 Oxid Met (2010) 74:93–111

Focused Ion Beam Workstations (FIB)

A FEI Strata DB 235 M combined focused ion beam/scanning electron microscope


(FIB/SEM) workstation and a FEI FIB 200 THP workstation were used for TEM-
sample preparation. In order to do so the samples were first cut with a diamond saw
into thin slices (400 lm) and then polished down to a thickness below 150 lm. The
next step was to stamp out 3 mm half discs and then polish them to an approximate
thickness of 50 lm. The oxide morphology was first investigated with the SEM
column in the FIB/SEM and then a protective 20 lm long, 4 lm wide and 100 nm
thick Pt-strip was deposited with electrons in-situ to protect the oxide scale. A
thicker Pt-strip (2 lm) was then deposited with ions in order to better protect the
oxide during the ion milling. The material on each side of the Pt-strip was then
milled away with the ion beam. The aim was to produce 20 lm wide and 1–5 lm
deep electron transparent areas, approximately 100 nm thick, of the oxide scale and
the subjacent metal. The ion current was decreased during the final operations to
create a smooth electron transparent area. If the thickness was 10 lm or more, the
current used was 20 nA, if the thickness was less than 10 lm and more than 2–
4 lm, 3–7 nA was used, if the thickness was less than 2–4 lm and more than
0.5 lm, 1 nA was used and for the final milling down to 100 nm, 500–100 pA was
used. During the final part the sample was also tilted approximately 18 in order to be
able to get a smooth electron transparent area with parallel sides.

Transmission Electron Microscopy (TEM)

A Philips CM 200 FEG transmission electron microscope (TEM) was used in this
work. It has a field emission electron gun, which was operated at an accelerating
voltage of 200 kV. The TEM is equipped with a Link Isis energy dispersive X-ray
(EDX) system and the software can be used to create scanning transmission electron
microscope (STEM) images, EDX spot analyses, EDX linescans and EDX maps.
These techniques were used to analyse the electron transparent cross-sections
produced in the FIB. The TEM was operated with gun lens 5 (intermediate strength)
together with an extractor voltage of 4.02 kV. Due to the technical difficulties of
quantifying the light anion elements (e.g. oxygen), the linescans presented in this
paper show weight percentages of total cation content. It is important to emphasize
that, while AES gives a mean value of the surface, TEM analysis is very sensitive in
a specific point.

Results

Gravimetry

Figure 1 shows mass change as a function of time for Sanicro 28 in O2 and


O2 ? H2O at 600, 700 and 800 °C. The small mass gains indicate that the alloy
shows protective behavior in all cases. As expected, mass gain increases with
temperature. In dry O2 the mass gain curves are sub-parabolic. In wet O2 the mass

123
Oxid Met (2010) 74:93–111 97

Fig. 1 Mass change versus exposure time of Sanicro 28 exposed at 600, 700 and 800 °C in 5% O2 and in
5% O2 ? 40% H2O

gain curves peak at 24 h at 600 and 700 °C, indicating chromium volatilization. At
800 °C the slope of the mass gain curve is always positive while the mass gains are
considerably lower than in dry O2.

Crystalline Phases Detected by XRD

Table 2 presents the crystalline phases detected by XRD, corundum type (a-
(Fe,Cr)2O3) and spinel type (M3O4) were identified. The austenitic substrate was
detected when the oxide layer was thin enough, i.e. in all cases except after 168 h at
800 °C. At 600 °C corundum-type oxide was the dominant oxidation product in
both atmospheres. At 800 °C the diffraction patterns were dominated by spinel type
oxide.

Table 2 Crystalline phases detected by XRD


Temperature Atmosphere (time) a-(FeCr)2O3 M3O4 Substrate

600 °C O2 (24 h) m w s
O2 (168 h) m w s
O2 ? H2O (24 h) s w s
O2 ? H2O (168 h) s w s
800 °C O2 (24 h) w m s
O2 (168 h) w m –
O2 ? H2O (24 h) m s m
O2 ? H2O (168 h) w s –

s Strong, m medium, w weak intensity, and (–) = not detected

123
98 Oxid Met (2010) 74:93–111

Surface Morphology

Figure 2a shows a SEM/BSE image of the as-prepared alloy surface (before


exposure). Small areas with different contrast (dark grey) can be seen, that form a
network pattern. The network is considered to correspond to the grain boundaries in
the alloy. SEM/EDX analysis showed that the dark grey areas had the same
composition as the matrix. These observations imply that they consist of pores that
have accumulated at the alloy grain boundaries. Figure 2b, c show SEM/SE and
SEM/BSE plan view images after exposure for 168 h in O2 at 600 °C. Figure 2b
shows a relatively smooth base oxide with oxide nodules on top. The density of
nodules is less in the grain boundary region. The BSE image in Fig. 2c shows dark
lines at the alloy grain boundaries. The dark lines are attributed to an accumulation
of pores similar to that observed in the unexposed material. More information on
these pores is given below. Plan view SEM imaging after exposure to O2 ? H2O at
600 °C showed the same morphology as in dry O2 and is not shown. Figure 2d–g
shows plane view SEM/SE images of the surface after 168 h at 700 and 800 °C.
Exposure at 700 °C resulted in a morphology similar to that described for 600 °C.
However, at 700 °C the oxide morphology is similar all over the surface, making it
difficult to discern the grain boundaries. At 800 °C the oxide is rougher and ridges
of thicker oxide have formed above the alloy grain boundaries. On top of some of
the thicker oxide ridges cracks were observed in both atmospheres (as in Fig. 2f).
The presence of water vapor had no significant influence on surface morphology,
regardless of temperature.

Characterization of the Oxide Scale

Exposure in Dry O2 at 600 °C

Figure 3 shows a STEM cross section image of the oxide scale formed after 24 h in
dry O2 at 600 °C. The oxide is about 100 nm thick, which corresponds well with the
gravimetric results (Table 3). The STEM image suggests that the oxide scale is
duplex. Accordingly, Auger analysis and STEM/EDX (not shown) indicated that it
consists of an inner chromium-rich part and an outer iron-rich part. In addition,
XRD showed the presence of corundum type oxide and spinel type oxide (Table 2).
Hence, the results suggest the presence of iron-rich spinel oxide (M3O4) and Cr-rich
corundum type M2O3. A 200 nm diameter void can be seen at a grain boundary in
the alloy substrate. Several ion milled cross sections were prepared and investigated
in the TEM from samples with different exposure times showing grain boundary
regions and regions far away from grain boundaries. Voids were frequently detected
at the alloy grain boundaries, immediately below the scale. It is considered that
these voids correspond to the network of dark lines seen in the BSE image after
exposure at 600 °C. Voids in the alloy were observed after exposure in both
atmospheres and also after exposure at 800 °C (see below). The evidence does not
suggest that the size and shape of the voids changes significantly upon exposure,
regardless of exposure time and temperature (600–800 °C). As expected, increasing
the exposure time from 24 to 168 h resulted in a thicker oxide, reaching about

123
Oxid Met (2010) 74:93–111 99

Fig. 2 a shows a plan view SEM/BSE image of an unexposed sample. b and c show SE and BSE plan
view images of the sample exposed for 168 h in dry O2. d and e show plan view SE images of samples
exposed in dry and wet O2 after 168 h at 700 °C and f and g show plan view SE images of samples
exposed in dry and wet O2 after 168 h at 800 °C

123
100 Oxid Met (2010) 74:93–111

Fig. 3 STEM image from the sample exposed for 24 h in O2 at 600 °C showing the oxide scale and the
subjacent metal

Table 3 Oxide scale thicknesses of Sanicro 28 after exposure to 5% O2 and 5% O2 ? 40% H2O at
600 °C
Environment 600 °C Exposure Oxide scale thickness Oxide scale thickness Calculated oxide
time (h) (nm) [TEM] (nm) [AES] thickness (nm)
[Cr2O3]

5% O2 24 100 90 80
5% O2 168 200–400 220 120
5% O2 ? 40% H2O 24 100 130 80a
5% O2 ? 40% H2O 168 150–200 150 20a
a
Underestimated due to vaporization

Fig. 4 AES depth profile of Sanicro 28 exposed at 600 °C for 168 h in dry O2

200 nm (Table 3; Figs. 4, 5a). AES depth profiling was performed on the middle of
a grain. An ion-milled cross-section was prepared that did not include a grain
boundary (see Fig. 5a). The morphology and composition of the oxide were in line

123
Oxid Met (2010) 74:93–111 101

Fig. 5 a TEM image showing a cross section of the oxide scale formed at 600 °C after 168 h in dry O2. c
The position of the line indicates where a linescan was acquired. The results from the linescan are shown
in (b). b STEM/EDX linescan from the sample exposed at 600 °C for 168 h in dry O2. The linescan was
acquired from the metal (A) through the oxide scale (B), see (a). c CBED patterns of the oxides formed in
dry O2 after 168 h. The inner oxide was indexed as a corundum structure with the zone axis [2 1 1 0] (I).
The outer oxide was indexed as a spinel type structure with the zone axis [0 0 1] (II)

with the observations made after 24 h (Figs. 4, 5a, b). Thus, the AES depth profile
reveals that the inner part of the oxide scale is enriched in chromium while the outer
part is iron-rich. The outer part of the scale also contains relatively high levels of
nickel. The results from a STEM/EDX linescan shown in Fig. 5b is in accordance
with the AES results, showing a duplex scale with an inner chromium-rich layer and
an outer layer which is enriched in iron and nickel. The composition of the inner layer
was typically 70–80% Cr, 10–15% Fe, 5% Mn and 5% Ni (cationic %) while the outer
part of the scale contained 15% Cr, 60% Fe, 25% Ni and 5% Mn (cationic %). Silica-
rich areas were detected at the metal/oxide interface. The inner oxide layer was
indexed as corundum type M2O3 by convergent beam electron diffraction (CBED)
(Fig. 5c), the cation composition being 85% Cr, 10% Fe and 5% Ni. The outer oxide
layer was indexed as spinel type M3O4 (Fig. 5c II) with a cation composition of 20%
Cr, 50% Fe, 25% Ni and 5% Mn corresponding to the general formula
(Ni,Fe)(Fe,Cr,Mn)2O4. The identification of oxides of spinel and corundum type is
in accordance with XRD (see Table 2). The results from the linescan (Fig. 5b) show
that the alloy immediately below the scale is depleted in chromium and enriched in
nickel. The Cr depleted zone extends approximately 100 nm into the alloy.

Exposure in O2 ? H2O at 600 °C

Exposure in O2 ? H2O at 600 °C resulted in a scale similar to that formed in dry


O2. Thus, according to TEM analysis of the oxidized surface after 24 h (not shown)
the scale was about 100 nm thick and appeared to be duplex. The results of Auger
depth profiling and XRD after 24 h exposure was also similar to the dry case,
suggesting the presence of iron-rich spinel oxide (M3O4) and Cr-rich corundum type
M2O3. After 168 h in O2 ? H2O, the oxide scale is somewhat thinner than in the
corresponding exposure in dry O2 (about 150 nm and 200 nm, respectively, see
Figs. 4, 6 and Table 3). Because of chromium volatilization (see below), mass gain
is not a good measure of oxidation rate in O2 ? H2O environment, see Table 3. The
TEM image (Fig. 7a) shows that the scale has a thickness in the range 150–200 nm.
As in dry O2, the inner oxide layer was indexed as corundum type M2O3 (Fig. 7b)
by CBED. The composition was 85% Cr with 5% each of Fe, Ni and Mn (cationic
%) while the outer layer was indexed as spinel type M3O4 oxide (Fig. 7c) with a
cationic composition of 20% Cr, 65% Fe, 10% Ni and 5% Mn. This is supported by
AES depth profiling showing that the inner part of the oxide scale is enriched in
chromium, while the outer part is more iron-rich, see Fig. 6. According to AES
depth profiling (see Figs. 4, 6) the outer part of the oxide scale formed in O2 ? H2O
environment is lower in chromium compared to dry O2. In addition, the outer part of
the oxide scale formed in O2 ? H2O environment contains less nickel than in the
dry case.

123
102 Oxid Met (2010) 74:93–111

123
Oxid Met (2010) 74:93–111 103

Fig. 6 AES depth profile of Sanicro 28 exposed at 600 °C for 168 h in 5% O2 ? 40% H2O

Fig. 7 a TEM image showing a cross section of the oxide formed in O2 ? H2O after 168 h at 600 °C. b
The inner oxide was indexed as a corundum structure with the zone axis [2 1 3 2]. c The outer oxide was
indexed as a spinel type structure with the zone axis [1 1 3]

123
104 Oxid Met (2010) 74:93–111

Exposure in Dry O2 at 800 °C

Figure 8a shows a STEM image of the oxide scale formed after 24 h in dry O2 at
800 °C. The scale is duplex, consisting of a fine-grained inner layer and an outer
layer with larger columnar grains. The total scale thickness was about 500 nm. The
duplex nature of the scale is also seen in the linescans (Fig. 8b). Thus, the inner
oxide layer is dominated by Cr (80 cationic %) with small amounts of Fe, Mn and
Ni while the outer oxide layer contains about 50% Cr together with Mn (30%) and
small amounts of Fe and Ni. Figure 9a shows a corresponding STEM cross section
image after 168 h, showing a scale thickness in the range 500 nm. The STEM image
shows a ridge of thicker oxide which has formed above an alloy grain boundary
(compare the plane view image in Fig. 2f). The STEM/EDX elemental maps reveal
that the oxide scale is two layered and chromium-rich. Point analyses show a
composition similar to that observed after 24 h. Thus, the inner layer contains[80%
chromium together with 10% manganese, 5% iron and 5% nickel (cationic %) while
the outer layer contains 60% chromium and 30% manganese together with small
amounts of iron and nickel. Based on the combined evidence from XRD (Table 2)
and EDX analysis, the upper layer of the scale is considered to consist of spinel-type
oxide (M3O4) with the approximate composition MnCr2O4 while the inner layer
consists of Cr-rich corundum-type M2O3. The duplex scale structure was also
verified by STEM/EDX linescans of cross sections (not shown). The alloy beneath
the oxide scale was depleted in chromium and enriched in nickel. After 24 h in dry
O2, the Cr depleted zone had reached a width of about 600 nm at the scale metal
interface. Also, the depleted zone extended further into the substrate along the alloy
grain boundaries. A discontinuous silica layer was present at the scale/alloy
interface, the silica containing significant amounts of Cr, Mn and Fe. The STEM/
EDX maps show chromium-rich regions in the alloy substrate. These regions are
also enriched in Mo and are attributed to carbide precipitates. Similar to the
unexposed samples and to the samples exposed at 600 °C, voids are often observed
where alloy grain boundaries reach the scale/metal interface, see Fig. 9a.

Exposure in O2 ? H2O at 800 °C

The STEM cross section image in Fig. 10a was made after 168 h in O2 ? H2O at
800 °C, showing an oxide thickness of about 600 nm. Again, chromia vaporization
causes mass gain to be an unreliable measure for oxidation rate in this environment,
explaining the anomalously low calculated scale thickness in Table 4. As at 600 °C,
similar scale structures were observed after exposure in O2 ? H2O and in dry O2.
Thus, Fig. 10a shows a two-layered scale including unconnected silica-rich regions
at the scale/metal interface. Also in this case, the evidence from XRD (Table 2) and
STEM/EDX allows us to conclude that the fine-grained inner part of the scale
consists of Cr-rich corundum type (M2O3), while the larger columnar grains in the
outer part consist of spinel type oxide (M3O4). However, the outer (spinel) oxide
layer formed in O2 ? H2O at 800 °C was significantly poorer in chromium
compared to that formed in dry O2 (see STEM/EDX maps in Fig. 10b). According
to linescans through the scale (not shown), the outer oxide layer contained 30%

123
Oxid Met (2010) 74:93–111 105

Fig. 8 a STEM image showing a cross section of the oxide formed in O2 after 24 h at 800 °C. The
position of the line indicates where a linescan was acquired. The results from the linescan are shown in
Fig. 9b. b The results from the STEM/EDX linescan acquired from the metal (A) through the oxide scale
(B), see (a)

123
106 Oxid Met (2010) 74:93–111

Fig. 9 a STEM image showing a cross section of the oxide scale formed in O2 after 168 h at 800 °C. b
Quantitative STEM EDX elemental maps of the oxide scale formed in O2 after 168 h at 800 °C, see
STEM image in (a)

(cationic) Cr after exposure in O2 ? H2O environment whereas it contained 60%


(cationic) Cr after exposure in dry O2. Hence, the outer part of the scale had the
approximate composition Mn2CrO4 after exposure in O2 ? H2O at 800 °C while it
can be represented as MnCr2O4 in dry O2. STEM/EDX point analyses showed that

123
Oxid Met (2010) 74:93–111 107

Fig. 10 a STEM image showing a cross section of the oxide scale formed in O2 ? H2O after 168 h at
800 °C. b STEM/EDX elemental maps of the oxide scale formed in O2 ? H2O after 168 h at 800 °C, see
STEM image in (a)

123
108 Oxid Met (2010) 74:93–111

Table 4 Oxide scale thicknesses of Sanicro 28 after exposure to 5% O2 and 5% O2 ? 40% H2O at
800 °C
Environment Exposure Oxide scale Calculated oxide
800 °C time (h) thickness (nm) [TEM] thickness (nm) [Cr2O3]

5% O2 24 400 350
5% O2 168 500 850
5% O2 ? 40% H2O 24 300 220a
5% O2 ? 40% H2O 168 600 370a
a
Underestimated due to vaporization

the silica formed at the metal/oxide interface typically contained 75 at.% Si together
with about 10% Cr, 10% Fe and 5% Ni (cationic %).

Discussion

The results show that alloy Sanicro 28 forms a protective, chromium-rich oxide in
O2 at 600, 700 and 800 °C and that the presence of water vapor does not
compromise the alloy’s ability to resist high temperature corrosion (Fig. 1). It is
well-known that chromium is lost from the protective oxide formed on Cr and FeCr
alloys in atmospheres containing water vapor and oxygen at 600–800 °C [15]. The
vaporizing species has been shown to be chromic acid (CrO2(OH)2) [14]:
Cr2 O3 ðsÞ þ 2H2 OðgÞ þ 1:5O2 ðgÞ ! 2CrO2 ðOHÞ2 ðgÞ ð1Þ
A study of the vaporization of chromic acid from chromia was recently published
by Opila et al. [16]. In good agreement with that paper, the equilibrium partial
pressure of CrO2(OH)2 over Cr2O3 in a mixture of 0.60 atm O2 and 0.40 atm H2O
was recently reported to be 0.15 ppm at 600 °C and 0.30 ppm at 700 °C [17].
Recalculating the data from Opila et al. to our experimental conditions gives the
equilibrium pressure of 0.1 ppm at 600 °C and 0.7 ppm at 800 °C for CrO2(OH)2.
The present study confirms that vaporization of chromic acid occurs when alloy
Sanicro 28 is exposed to a mixture of water vapor and oxygen at 600–800 °C. Thus,
TG curves registered in O2 ? H2O environment show mass losses at 600 and
700 °C after exposure times of more than 24 h. The effect of chromia vaporization
is also evident from the TG curves in O2 ? 40% H2O at 800 °C. However, since the
oxidation rate is higher at this temperature the slope of the mass change curve does
not become negative.
The influence of Cr vaporization on scale composition is noticeable especially at
800 °C. Thus, exposure to O2 ? H2O environment at 800 °C resulted in the
formation of an outer (spinel type) oxide that can be represented by Mn2CrO4 while
in dry O2 the corresponding part of the scale had the approximate composition
MnCr2O4. Similar results were obtained on an austenitic 38Fe25Cr35Ni steel at
900 °C in the presence of O2 ? H2O [18]. The effect of chromium vaporization on
the MnCr spinel composition can be represented in the following way:

123
Oxid Met (2010) 74:93–111 109

2MnCr2 O4 ðsÞ þ 5x=ð1 þ xÞO2 ðgÞ þ 5x=ð1 þ xÞH2 OðgÞ


! 2x=ð1 þ xÞMnII MnIII x Crð2xÞ O4 ðsÞ þ 6x=ð1 þ xÞCrO2 ðOHÞ2 ðgÞ ð2aÞ
In the present case the reaction can be expressed in a simpler manner:
2MnCr2 O4 ðsÞ þ 5=2O2 ðgÞ þ 3H2 OðgÞ ! Mn2 CrO4 ðsÞ þ 3CrO2 ðOHÞ2 ðgÞ ð2bÞ
[Comment: In the interest of clarity, Fe and Ni that are present in significant
amounts in the spinel are omitted from (2a, 2b)]
In contrast, the inner part of the scale, consisting of Cr rich M2O3, was unaffected
by the presence of water vapor. This is interpreted to imply that the supply of
chromium to the oxide by diffusion in the alloy (primarily by grain boundary
diffusion) is sufficiently rapid to maintain a high chromium concentration in the
inner part of the scale, thereby maintaining its protective properties. It can thereby
be concluded that the higher Cr evaporation rate at the higher exposure temperature
is compensated by a higher diffusivity of Cr to the oxide scale. According to our
calculations above the equilibrium pressure of CrO2(OH)2 changed from 0.1 ppm at
600 °C to 0.7 ppm at 800 °C. This may be compared to the diffusivity of Cr in the
alloy where the depletion zone was measured to be about 100 nm at 600 °C while it
was about 600 nm at 800 °C. The spinel formed at the oxide/gas interface could in
addition be beneficial for oxidation behavior in wet oxygen because it may slow
down chromia evaporation.
The behavior of Sanicro 28 may be contrasted to the behavior of austenitic
FeNiCr-base alloys with lower chromium content, e.g. alloy 304L that tends to suffer
a partial breakdown of the protective properties in O2 ? H2O environment at 600 °C
[19]. It is suggested that the superior oxidation resistance of Sanicro 28 is mainly due
to its high Cr/Fe ratio, producing a faster supply of chromium from the alloy,
compared to e.g. 304L. In addition, it is argued that the formation of a layer of spinel
oxide at the scale/gas interface contributes to the oxidation behavior of Sanicro 28 in
wet oxygen. Thus, the loss of chromium from the corundum-type oxide is expected to
be slower compared to the cases (e.g. 304L) where this cap layer is absent.
Reports on the accumulation of silica at the scale alloy interface of high
temperature steels are frequent in the literature [20–22]. It has been suggested that
the formation of a silica sub-scale can contribute significantly to the oxidation
resistance [23]. However, in the present case, the silica at the interface can hardly
act as a diffusion barrier since it does not form a continuous film, and is thus not
expected to contribute much to oxidation properties. The present results may be
compared to the oxidation behavior of the austenitic 38Fe25Cr35Ni steel 353MA,
which is very similar in composition except that it has higher Si content and Ce
additions. Despite the high Si content (1.3 wt%) in 353MA no diffusional barrier of
silica was reported to form under similar exposure conditions (O2 and O2 ? H2O
after 168 h in the temperature range 700–900 °C) [12, 18]. This implies that it
would not be enough to increase the Si content in order to create a silica diffusional
barrier during initial oxidation of Sanicro 28 at these temperatures. It may be noted
that Si can have other beneficial effects on oxidation behavior (see below).
The voids at the alloy grain boundaries observed before exposure, and their
development upon exposure, is an interesting phenomenon. The fact that the voids

123
110 Oxid Met (2010) 74:93–111

were observed where alloy grain boundaries reach the scale/metal interface implies
that they may play a role in oxidation, e.g. by decreasing scale adherence which could
be detrimental to oxidation properties. Evans et.al has reported on the influence of Si
content on void formation in 20Cr–25Ni stainless steels after long exposure times
(6000 h) at 900 °C [24]. A series of 20Cr–25Ni stainless steels with Si content
ranging from 0.05 to 2.35 wt% was investigated and voids were reported to form in
the alloy substrate in all cases. In contrast, void formation at the metal/oxide interface,
as in the present case, was reported only on the alloy with the lowest Si content. The
latter type of voids was suggested to be caused by the Kirkendall effect in connection
to Cr depletion of the alloy by preferential oxidation. The other type of void reported
by Evans was suggested to be associated to the dissociation of M23X6 type precipitates
and was correlated to scale thickness. In this context it is interesting to compare void
formation on alloy Sanicro 28 (35Fe27Cr31Ni) with alloy 353MA (38Fe25Cr35Ni),
which has a higher Si content but is otherwise similar in composition. While the
present study shows the formation of scale/metal voids in Sanicro 28, such voids did
not form in the case of alloy 353MA (Exposure conditions and exposure time are
comparable for the two alloys) [12, 18]. Instead of voids, alloy 353MA formed
globular silica aggregates at the scale metal interface. The observations are in
accordance with Evans’s report regarding the influence of Si on void formation. It is
argued that in this case, the apparent influence of Si on void formation is not due to the
formation of a continuous silica film that slows down oxide growth (as suggested by
Evans) because neither 353MA nor Sanicro 28 form such a film. Instead it is suggested
that in the case of the high Si alloy 353MA, the voids become filled by silica,
explaining the silica globules. It may be noted that there is a large difference in
exposure time between this study and the work of Evans et al. and that comparison
between the two studies should be made with some caution.

Conclusions

Alloy Sanicro 28 forms a protective scale in O2 and O2 ? H2O environment at 600–


800 °C. The scale is duplex and the inner part consists of corundum type chromium-
rich (CrxFe1-x)2O3, while the outer layer consists of spinel type oxide. In O2 ? H2O
environment chromia is lost from the protective oxide by vaporization of
CrO2(OH)2. As a result, the spinel oxide in the outer part of the scale becomes
depleted in chromium. However, the inner layer is virtually unaffected and the scale
retains its protective properties. The capacity of alloy Sanicro 28 to suffer chromia
vaporization without forming a rapidly growing iron-rich oxide is attributed to its
high Cr/Fe ratio. Hence, the loss by vaporization is compensated for by a rapid
supply of chromium to the protective scale by diffusion in the alloy. The spinel
formed at the oxide/gas interface could in addition be beneficial for oxidation
behavior in wet oxygen because it may slow down chromia evaporation.

Acknowledgments This work was carried out within the High Temperature Corrosion Centre (HTC) at
Chalmers University of Technology. A grant from the Knut and Alice Wallenberg Foundation for
acquiring the FEG-SEM instruments is gratefully acknowledged.

123
Oxid Met (2010) 74:93–111 111

References

1. P. Kofstad, High Temperature Corrosion (Elsevier Applied Science Publishers Ltd, London and New
York, 1988).
2. C. T. Fujii and R. A. Meussner, Journal of the Electrochemical Society 111, 1215 (1964).
3. M. Thiele, H. Teichmann, W. Schwarz, and W. J. Quadakkers, VGB Kraftwerkstechnik 77, 129
(1997).
4. W. J. Quadakkers and P. J. Ennis, Materials for Advanced Power Engineering 1998 (Liege, Belgium,
1998).
5. G. Hultquist, B. Tveten, E. Hörnlund, M. Limbäck, and R. Haugsrud, Oxidation of Metals 56, 313
(2001).
6. G. Hultquist, B. Tveten, and E. Hörnlund, Oxidation of Metals 54, 1 (2000).
7. H. Asteman, Water Vapour Induced Active Oxidation of Stainless Steel (Department of Chemistry,
Göteborg University, Göteborg, 2002).
8. H. Asteman, K. Segerdahl, J.-E. Svensson, and L.-G. Johansson, Materials Science Forum 369–372,
277 (2001).
9. H. Asteman, J.-E. Svensson, and L.-G. Johansson, Oxidation of Metals 57, 193 (2002).
10. H. Asteman, J.-E. Svensson, and L.-G. Johansson, Journal of the Electrochemical Society 151, B141
(2004).
11. H. Asteman, J.-E. Svensson, M. Norell, and L.-G. Johansson, Oxidation of Metals 54, 11 (2000).
12. T. Jonsson, S. Canovic, F. Liu, H. Asteman, J. E. Svensson, L. G. Johansson, and M. Halvarsson,
Materials at High Temperature 22, 231 (2005).
13. F. Liu, J. E. Tang, T. Jonsson, S. Canovic, K. Segerdahl, J.-E. Svensson, and M. Halvarsson,
Oxidation of Metals 66, 295 (2006).
14. B. B. Ebbinghaus, Combustion and Flame 93, 119 (1993).
15. H. Asteman, J.-E. Svensson, and L.-G. Johansson, Journal of the Electrochemical Society 151, B141
(2004).
16. E. J. Opila, D. L. Myers, N. S. Jacobson, I. M. B. Nielsen, D. F. Johnson, J. K. Olminsky, and M. D.
Allendorf, Journal of Physical Chemistry 111, 1971 (2007).
17. B. Pujilaksono, T. Jonsson, M. Halvarsson, I. Panas, J.-E. Svensson, and L.-G. Johansson, Oxidation
of Metals 70, 163 (2008).
18. T. Jonsson, F. Liu, S. Canivic, H. Asteman, J. E. Svensson, L. G. Johansson, and M. Halvarsson,
Journal of the Electrochemical Society 154, C603 (2007).
19. H. Asteman, J.-E. Svensson, M. Norell, and L.-G. Johansson, Oxidation of Metals 54, 11 (2000).
20. S. N. Basu and G. J. Yurek, Oxidation of Metals 36, 281 (1991).
21. L. Mikkelsen, S. Linderoth, and J. B. Bilde-Sorensen, 6th International Symposium On High-Tem-
perature Corrosion and Protection Of Materials (Les Embiez (Var) France, 2004).
22. Y. Wouters, G. Bamba, A. Galerie, M. Mermoux, and J.-P. Petit, 6th International Symposium on
High-Temperature Corrosion and Protection of Materials (Les EMBIEZ (Var), France, 2004).
23. F. H. Stott, G. C. Wood, and J. Stringer, Oxidation of Metals 44, 113 (1995).
24. H. E. Evans, D. A. Hilton, R. A. Holm, and S. J. Webster, Oxidation of Metals 19, 1 (1983).

123
View publication stats

You might also like