P R E P R I N T – ICPWS XV

Berlin, September 8–11, 2008

Corrosion of steels in H2O-CO2 atmospheres at temperatures between

500°C and 700°C

D. Huenert, W. Schulz and A. Kranzmann

Federal Institute of Materials Research and Testing

Unter den Eichen 87, 12205 Berlin, Germany

Email: daniela.huenert@bam.de,
wencke.schulz@bam.de,
axel.kranzmann@bam.de

The introduction of carbon capture technology into thermal power plants benefits from combustion of fuel and
pure oxygen due to high partial pressures of CO2 in the flue gas. Consequently IGCC plants with 80 bar
reactors and high temperature turbine equipment, boilers with 700°C superheater tubes will be the next
generation power plants. In this paper, the attack of a model gas composition on different typical power plant
construction steels with chromium contents in the range of 1 to 24% is discussed. The test conditions were an
corrosion time up to 1000 h, a pressure of 80 bar, a fast flowing gas and temperatures between 500 and 700°C
depending on the maximum working temperature of the individual steel. In particular, the growth of FeO at
temperatures lower than 570°C, which is the eutectic decomposition temperature in the oxygen-iron phase
diagram, is observed. Iron carbides are primarily formed in low alloyed steels near the oxide steel interface.
For the studied steels containing more then 9% chromium, a growth of Fe3O4 and Fe-Cr-spinel was typical.
Steels with chromium contents below 12% did not form a dense and therefore protective Cr2O3 layer. The
oxide scale thickness increases strongly with temperature and decreases with raising chromium content. Data
of oxide scale thickness up to corrosion time of 1000 h will be presented. In the test environment the
maximum working temperature was estimated to be 50K below the working temperatures in current power
plant environments.

Introduction In the Oxyfuel power plants, pure oxygen will

The discussion about the effect of rising CO2
concentration in air had caused, that all industrial
nations search for options to decrease the part of
CO2 in the atmosphere. The expression of this
collective aim is the Kyoto protocol, in which the
involving states are committed to reduce the part of
CO2, CH4, N2O, H-FKW/HCF, FKW/PFC and SF6
to an amount of 5% in a time period from 1990 to
2012. If all states fulfil their obligations, the part of
CO2 in the air could be stabilized below 350 ppm.
Industrial facilities and power plants are well
localised emission sources and produce
approximately one fourth of the emitted CO2. The
running research programs of the Federal Ministry
of Economics to develop CO2 Reduction
Technology (COORETEC), as well as EU-
programs, consider the complete process from CO2
accruement to CO2 suspension. Consequently the
application of power plants with CO2 separation is
an aim. The technologies with CO2 capture require
rising efficiencies, which cause higher working
temperatures, pressures and modified gas
compositions. In this account better knowledge
about the corrosion behaviour of the materials have
to be obtained.
be inserted for the combustion, resulting in a flue
gas composed of H2O and CO2. In addition with the
unconsumed oxygen and fuel specific impurities,
these gases compose the flue gas.
The metallic units have to resist surface
temperatures of 950°C in gas turbines and 650°C in
coal burning power plants. Future power plants are
supposed to work in the range of 700°C. Less is
known about the oxidation resistance of
conventional steels in advanced power plants
during service under future power plant conditions.
The research work of the BAM is focused on the
corrosion behaviour of power plant steels and
nickel based alloys with different chromium
contents in contact with combustion gases
consisting of H2O, CO2 and O2 [1] [2]. The
oxidation behaviour, oxidation rates and the formed
oxide scales with its phase composition of one low
alloyed steel, three 9-12% Cr-steels and one
austenitic steel in H2O-CO2 has been the subject of
our study.
Experimental
In the experiments commercial available steam
generation tubes were used as tubular samples,
which were streamed by pressurised combustion
gases from the inside. At the outside laboratory air
at ambient pressure was provided and caused an
oxidation. The flowing combustion gas consisted of
H2O and CO2 (30 mol %, 70 mol %) at a pressure
of 80 bar according to the main combustions
products in IGCC and Oxyfuel. Impurities were left
out to study the corrosion mechanisms of the major
components.
A thermal load was applied by heating the tubes
from the outside, allowing the experimental study
of different temperatures in one tube. The tubes
were exposed at temperatures between 500°C and
700°C for a period of 250 h and 1000 h. The gas
velocities varied with rising inner tube diameter
between 2,1 and 0,1 cm/s. At the end of the tubes, a
gaseous or fluid sample can be removed from the
flue gas to clarify, which alloying additions
evaporate during the corrosion process.
The exact compositions of the investigated steels
are shown in table 1. The ferritic and martensitic
materials differ mainly in chromium content. As a
substitute of the low alloyed steels, 13CrMo44 (1%
Cr), as high alloyed steels, the P92, X20 and an
experimental steel B (9-12% Cr) and as austenitic
steel, the DMV 310 N was investigated.
Figure 1: Sketch of facility for the corrosion at 80
bar
Table 1: Composition of the testing basic material
Preparation
After the corrosion tests, tubes were prepared
with a fixation spray first and subsequently filled
with epoxy resin in order to prevent a spalling off
of the oxide scales during machining. Exact
information was published elsewhere [3]. The
oxidation products were analysed with optical
microscopy, scanning electron microscopy
(SEM/EDX, LEO 1520 VP, JEOL microprobe
JXA-8900 RL) and transmission electron
microscopy (JEOL 4000).
2
Figure 2: Sample for analyses
Measurement of scale thickness
On each sample (fig. 2), 100 single measuring
points on eight different positions were measured
around the tube segment to give a good accuracy of
scale thicknesses and to evaluate irregularities on
the cross section.
Results
Microstructure of the base material
The low alloyed steel 13CrMo44 has a ferritic
microstructure. A martensitic microstructure is
characteristic for the steels X20CrMoV12-1, B and
P92. As an example of austenitic steel, DMV 310 N
containing 24% chromium was used.
Composition of oxide scales
In a flowing mixture of H2O-CO2 at a pressure
of 80 bar in a temperature range between 550°C
and 700°C, differences in phase compositions were
observed after an exposure time of 1000 h [fig. 3].
Figure 3: General phase sequence of steels
differing in chromium content
A typical phase sequence of a low alloyed steel
at 600°C is shown in fig. 4a. Above the wustite
point a heterogeneous layer in contact with the base
material has been formed consisting of 50% FeO
and 50% iron carbides. The carbides are arranged
like columns in a wustite matrix. According to
rising oxygen partial pressures from the metal-
oxide interface to the oxide-gas interface, the
following phases consist of wustite and maghemit.
Maghemit is a defect spinel with the structure of
magnetite but the overall formula of hematite [5].
Maghemit is a result of cooling, in which the iron
mobility is decreased and oxygen diffusivity is
increased [4]. Below the wustite point the scale
consists of an inner magnetite and an outer
hematite layer according to the thermodynamically
preferred phases below the wustite point. The
heterogeneous mixture of wustite and iron carbides
observable at higher temperatures is absent at Fe3O4
550°C.

Fe2O3 Fe3O4 + Spinel

FeO
chromium precipitates
FeO + Fe3C
Alloy (P92)
Alloy (13CrMo44)
c)
a)

On 9-12% Cr-steels, a continuous and

multilayered oxide scale has been formed. Figure
4b and c shows chromium precipitates in an inner
FeO matrix, followed by Fe3O4 with spinel stringers
and an outer thin layer of Fe2O3 (maghemit) at the
oxide-gas interface.

Fe2O3

d)

Fe3O4 + spinel stringers

iron oxide
Alloy (B) Cr2O3
b)

In the case of alloy B, the maghemit phase is

absent. Carbides are only found in the case of Co
and W free alloys at the metal-oxide interface.
Despite of the similar chromium content of steel
X20 and B, the structure of their oxide scales is
different. During the corrosion of X20, oxide
islands as well as areas with thin oxides were
formed (fig. 4d). In contrast to the islands with in
this manner described phase sequence, the thin
neighbouring oxide scale shows only iron rich
oxides with no chromium spinel. This oxide
morphology is well known for 9-12% chromium
steels exposed to air at high temperatures [6]. A
continuous oxide scale with no islands has been
formed on the steel B.
Alloy (DMV 310N)
e.)
Figure 4: Oxide scales of a) 13CrMo44 b) steel B
c) P92 d) X20CrMoV12-1 and e)
DMV310N in flowing H2O-CO2 at a
pressure of 80 bar at the maximum
working temperatures of the individual
steels after an exposure time of 1000 h
The microstructure of the oxide scale on an
austenitic steel at 730°C is shown in fig. 4e. A thin
and continuous chromium oxide layer was formed

3
in contact with the base material, which is followed
by an iron oxide (Fe2O3).
Scale thicknesses
Depending on the temperature, different oxide
scale thicknesses for the investigated steels were
observed. Figure 5 shows the scale thicknesses as a
function of temperature after an exposure time of
1000 h. Generally, the oxide scales of 13CrMo44,
P92 and alloy B becomes thicker with increasing
temperature. A non-arrhenius-typ scale growth is
ascertained for the steel X20CrMoV12-1
potentially indicating changes in reaction
mechanism. The oxide scale increases with
temperature until a temperature of 600°C is reached
and is reduced for higher temperatures.
10000
13CrMo44
P92
1000
X20CrMoV12-1
scale thickness, µm
B
DMV 310 N
100
10
1
9,8 10,8
1/ T *10 [1/K]
Figure 5: Logarithmic exposition of oxide scale
thickness data for the investigated steels
after an 1000 h exposure in a flowing
mixture of H2O-CO2 at a pressure of 80
bar and at different temperatures
At the steel 13CrMo44 with the lowest
chromium content, thick oxide scales has been
developed during the corrosion in a flowing and
pressure loaded mixture of H2O and CO2. This
ferritic steel shows the highest scale thicknesses
with rising temperature compared to steels higher
in chromium content. After a temperature of 550°C
a sharp increase in scale thickness is observable.
The high alloyed steels P92, alloy B and X20
present a relatively low scale thickness below
600°C (fig. 5). Above 600°C an increased oxide
scale growth for the steels P92 and steel B is
observable. In contrast, the steel X20 shows a
reduction of scale thickness with rising
temperatures above 600°C. Scale thickness
reducing factors could be spallation, preparation
effects, fast evaporation of alloying elements from
the oxide surface [7] or a change in corrosion
mechanism.
18
500
relative scale thickness, µm
xscale
xrel = 550
13
50µm 600
650
8
3
-2
0 5 10 15 20 25
Chromiumequivalent
Figure 6: Relative scale thickness versus
chromium equivalent at different
temperatures
At the austenitic steel (DMV310N) with the
highest chromium content of 24 %, thin scales of
approximately 5µm has been grown at all
investigated temperatures.
Figure 6 shows the relative scale thickness data of
the studied steels (table 1) during the exposure in
flowing H2O-CO2 (30-70 mol %) at 80 bar as a
function of chromium equivalent and different
temperatures. The scale thicknesses are given
relative to the scale thickness of alloy B at 500°C
11,8 12,8
over 1000 h at 80 bar in order to compare the scale
-4
thicknesses. A scale thickness of 50 µm was found
in this case. The dashed lines illustrate the assumed
dependence on chromium equivalent. The
temperature effect on scale thickness is pronounced
for low alloyed steels and marginal for steels with
higher alloying additions. With rising temperatures,
the curve maximum has been shifted to higher
values of the relative scale thickness. The reduction
of oxide scale thickness is not only due to an
increased level of chromium in the alloy.
Especially secondary alloying additions like
tungsten and cobalt seem to influence the corrosion
behaviour significantly and lower the relative oxide
scales. The relative scale thickness is also
dependent on pressure of the gas mixture. With
increasing pressure the oxide scale becomes thicker
[4].
Figure 7 shows the scale thickness of the
studied steels differing mainly in chromium content
during the exposure in H2O-CO2 (30-70 mols %) as
a function of time at temperatures of 500, 550 and
600°C. For the low Cr-steel 13CrMo44 an
incubation time for the exposure at a pressure of 80
bar is characteristic at all investigated temperatures.
At the beginning of oxidation, the scales do not
grow very much, whereas after an exposure of 1000
h substantially thicker scales are formed. Already,
small additions of chromium show a substantially
decreasing effect on scale thickness. These 9-12%
chromium steels grow probably parabolic. The
austenitic steel DMV 310N shows the thinnest
oxide scale of 5 µm after a 1000 h corrosion time.
4

500°C
550°C
600°C
Figure 7: Scale thickness as a function of
time at temperatures of 500,
550 and 600°C.
Discussion
The corrosion mechanisms and observed oxide
scale compositions in a flowing H2O-CO2-
atmosphere at different temperatures are changed
compared to the corrosion results in pure H2O, CO2
and O2 respectively.
Under the described conditions the stable oxide
phase on the low alloyed chromium steel
13CrMo44 is wustite, despite the oxygen partial
pressures in the mixture clearly favour the
formation of hematite at every examined
temperatures. Because of its defect rich structure,
wustite leads to a fast diffusion of iron to the
outside and the enhanced growing of oxide scales.
Above the wustite point in the Fe-O binary phase
diagram at 570°C, wustite is the
5
thermodynamically preferred phase and below
570°C magnetite is the stable phase. Magnetite
exhibits a spinel structure with limited defects,
resulting in the reduction of oxide scale growth. In
our experiments the wustite was observable below
the wustite point. Alloying additions like Mn have
the ability to reduce the wustite point temperature
[14]. The low alloyed steel 13CrMo44 has
sufficiently high amounts of Mn, which could
result in the stabilisation of wustite below 570°C
and cause the high oxide scale growth down to low
temperatures. Another explanation for existence of
wustite below 570°C is that the equilibrium is not
reached due to the fast diffusion and reaction of
iron to form oxides. An additional explanation for
the temperature of 600°C considers the formation
of a laminar grain boundary layer at test conditions,
where oxygen is provided depending on the oxygen
diffusion through the boundary layer. The deep
decreasing partial pressure profile in the laminar
media layer could meet the FeO/Fe3O4-equlilibrium
oxygen partial pressure and hence favour the
formation of wustite.
An addition of chromium leads to a significant
decrease in oxidation rates because of the formation
of protecting Fe-Cr-spinels [4] [6]. Depending on
chromium content, Fe-Cr-particles or a continuous
and protecting layer of Fe-Cr-spinel has been
formed at the martensitic 9-12% chromium steels.
The spinel is located in the inner corrosion zone
and show zones enriched in chromium depending
on the chromium content in the alloy. At the alloy
interface a chromium rich spinel, and close to the
gas side, an iron-rich spinel is exclusive observed.
According to the reported results, a dense Cr2O3-
layer could not be observed for any of the
examined 9-12% Cr-steels after a 1000h exposure
in flowing H2O-CO2 at different temperatures
(relevant for power plants). It is reported that in
pure steam, a thin protecting Cr2O3-layer is first
formed above a chromium content of 15%, which is
not stable due to interdiffusion by iron and
formation of iron rich oxides [6] [8] [9]. The
observed results in H2O-CO2 indicate that H2O has
a more lasting effect on the phase sequence for the
high chromium steels as CO2 has. The analysed
iron-rich spinel cause a loss of protective properties
of the oxide scale [10] [11] [12] [13]. At the steels
close to the gas side, maghemit was observable for
all investigated steels with the exception of alloy B.
During corrosion and during cooling a deep oxygen
profile from the metal surface to the oxide/gas
interface is formed. Accordingly, the iron vacancies
are growing in oxygen rich sections, near the
oxide/gas interface. In the same direction the
oxygen rich wustite becomes metastable below
temperatures of 570°C and decomposes in
magnetite and residual wustite [14]. Residual
wustite was also observed near the gas/oxide
interface in our experiments for the steel
13CrMo44, supporting this theoretical mechanism.
The decomposed magnetite could further transform
into maghemit, because of the high oxygen supply
at the oxide/gas interface [4].
In contrast to the 9-12% chromium steels, the
austenitic steel with the highest chromium content
of 24% shows a dense and continuous Cr2O3 layer
in contact with the base material. As a result of the
fast diffusion of iron, the Cr2O3-layer is covered by
iron oxides. This effect is well known as transient
oxidation [15].
A strongly developed carbide formation in the
base material and close to the oxide/alloy interface
is observed at a pressure of 80 bar for all cobalt and
tungsten free alloys up to 12% chromium. In the
case of the low alloyed steel 13CrMo44,
additionally strong carbide formation occurs in the
oxide to form a complete layer of 50% wustite and
50% iron carbides. This additional carbide-oxide
layer grows temperature and time dependent. The
carbide formation is obviously coupled with the
formation of wustite. Oxide scales composed of
magnetite shows no additional carbides in the
oxide. At a corrosion time of 125 h, no carbide-
oxide layer was observable. After 250 h, a thin
layer grows and becomes thicker with time. The
additional carbide-oxide layer is restrained for
steels with higher chromium levels. For all Co- and
W-containing steels, carbide formation is absent in
the base material.
Conclusion
H2O and CO2 are the main constituents of the
flue gas in power plants with oxygen feeded
combustion. Chromium containing protective
scales are not formed under conditions of power
plants for the next generation at steels with 9-12%
chromium. This results in unexpected lower
application temperatures for these classes of power
plant steels with consequences on potential
efficiency of electricity production using fossil
fuels. Finally it is assumed that the working
temperatures have to be reduced about 50 K
dependent on the material. The exposure after 1000
h suggests that the 1 % chromium steel is only
applicable up to 500°C. The 9-12% chromium
steels are able to be used below 600°C in an
atmosphere of H2O-CO2, whereas from the
viewpoint of corrosion resistance excluding
mechanical data, the austenitic steel DMV 310 N
may be serviceable at temperatures up to 700°C,
which is not finally concluded.
Acknowledgement
The authors would like to acklowledge Dipl.-
Ing. H. Nitschke, R. Saliwan-Neumann , G. Oder,
Dipl.-Ing. I. Urban and Dr.-Ing. Oesterle for their
contributions to this work.
6
The authors acknowledge the financial support
from the Ministry of Economics, Germany (Grant
ID: 0327727B) and the COORETEC program.
Literature
[1] Bundesministerium für Wirtschaft und Arbeit:
Research and Development Concept for Zero
Emission Fossil-Fuelled Power Plants (2003),
Documentation No. 527
[2] A. Kranzmann and D. Huenert: Technologien
für den Klimaschutz und ihre Anforderungen
an Werkstoffe. In: Chemie Ing. Technik, Vol
78, Issue 12, pp1809-1817, 2006
[3] H. Nitschke, D. Hünert, A. Kranzmann:
Präparation und Darstellung von HT-
Korrosionsschichten aus CO2-H2O
Atmosphären: In: Sonderbände der Praktischen
Metallographie 39. Fortschritte in der
Metallographie, 2007
[4] D. Huenert and A. Kranzmann: Influence of
pressure and chromium content on corrosion
reactions at 600°C in a CO2-H2O atmosphere.
In: Corrosion 2008, Conference March 2008,
New Orleans
[5] J. Huber : Zur Natur von γ-Fe2O3. Inaugural –
Dissertation, 2004, Gießen
[6] G.C. Wood, I.G. Wright, T.Hotgkiess, D.P.
Whittle: A Comparison of the oxidation of Fe-
Cr, Ni-Cr and Co-Cr Alloys in Oxygen and
Water Vapour. In: Werkstoffe und Korrosion,
Heft 11, 1970
[7] E. Opila: Votalitility of common protective
oxides in high-temperature water vapor:
current understanding and unanswered
questions. In: Materials Science Forum, Vo.
461-464, pp. 765-774, 2004
[8] J. Žurek, E. Wessel, L. Niewolak, F. Schmitz,
T.-U. Kern, L. Singheiser, W.J. Quadakkers:
Anomalous temperature dependence of
oxidation kinetics during steam oxidation of
ferritic steels in the temperature range 50-
650°C. In: Corrosion Science 46 (2004) 2301-
221
[9] W. Bendick, J. Gabrel, J.-C. Vaillant, B.
Vandenberghe: Development of a new 12%
Cr-Steel for tubes and pipes in power plants
with steam temperatures up to 650°C. In:
Materials for advanced power engineering
2002, Proceedings of the 7th Liège
Conference, Schriften des Forschungszentrums
Jülich, Energy Technology, Vol. 21, Part III,
1361 - 1370, 2002
[10] K. Segerdahl, J.-E. Svensson and L.-G.
Johansson: The high temperature oxidation of
11% chromium steel: Part I – Influence of
 pH2O. In: Materials and Corrosion 53, 247-
255, 2002
[11] R. Knödler: Oxidation of advanced
ferritic/martensitic steels and of coatings in
flowing steam at 650°C. In: Materials for
Advanced Power Engineering, Part III, 2002
[12] S. Osgerby and L. N. Mc Cartney: Steam
oxidation of 9-12 Cr martensitic steels:
characterisation and modelling the spalling of
oxide scale. In: Materials for Advanced Power
Engineering, Part III, 2002
[13] H. Kutsumi, T. Itagaki and F. Abe:
Improvement of steam oxidation resistance for
ferritic heat resistant steels. In: Materials for
Advanced Power Engineering, Part III, 2002
[14] Punge-Witteler: Wüstit in Oxidschichten auf
niedrig-legierten Stählen in Wasserdampf. II.
Wüstitzersetzung. In: Werkstoffe und
Korrosion 39, 204-207, 1988
[15] Birks: High Temperature Oxidation