J. of Supercritical Fluids 29 (2004) 129
Corrosion in high-temperature and supercritical water and
aqueous solutions: a reviewPeter KritzerFreudenberg Vliesstoffe KG,
D-69 465 Weinheim, Germany Received 22 April 2002; received in
revised form 6 January 2003; accepted 14 February 2003
Abstract The aim of the present article is to review some of the
common corrosion phenomena and describe the predominant corrosion
mechanisms in high-temperature and supercritical water. Corrosion
in aqueous systems up to supercritical temperatures is determined
by several solution-dependent and material-dependent factors.
Solution-depending factors are the density, the temperature, the pH
value, and the electrochemical potential of the solution, and the
aggressiveness of the attacking anions. Material-dependent
parameters include alloy composition, surface condition, material
purity, and heat treatment. Corrosion phenomena that are observed
include intergranular corrosion, pitting, general corrosion, and
stress corrosion cracking. The solubility and dissociation of both
attacking species and corrosion products play the most important
role for corrosion in high-temperature water. Both solubility and
dissociation processes are strongly inuenced by the density, or the
ionic product, respectively, of the solvent. High values of both
parameters favor ionic reactions, and thus, accelerate
electrochemical forms of corrosion. At low densities, water behaves
like a non-polar solvent, and thus, ions associate. In these cases,
the concentation of e.g. aggressive H+ drops down and thus,
solutions containing species such as HCl become neutral and thus
less aggressive. Further, corrosion products plug the surface and
material loss stops. Materials parameters have inuence especially
on the initiation of corrosion. In the present article, these
factors are linked with the physical and chemical properties of
high-temperature and supercritical water. An outlook is also given
for future research needs. 2003 Published by Elsevier B.V.Keywords:
Corrosion; High-temperature water; Supercritical water;
Supercritical water oxidation; Density; Ionic product
1. Physical and chemical properties of high-temperature and
supercritical water The physical and chemical properties of
hightemperature and supercritical water (T > 374 C; p > 22.05
MPa) have been investigated and reviewed in detail by E.U. Franck
and co-workers [15]. A
Tel.: +49-6201-80-4003; fax: +49-6201-88-4003. E-mail address:
[email protected] (P. Kritzer).
computer program is available to calculate these properties [6].
Fig. 1 shows schematically the course of density when crossing the
critical values of pressure or temperature. At a certain
temperature, a drop can be observed, at which density and ionic
product of high-temperature water fall down drastically. With
increasing pressure, this drop shifts toward higher temperatures
and its step height decreases (Fig. 2). It will be shown in the
following sections that this drop is responsible for the
uncommon
0896-8446/$ see front matter 2003 Published by Elsevier B.V.
doi:10.1016/S0896-8446(03)00031-7
2
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
Fig. 1. Phase diagram of water and schematic course of density
versus pressure and temperature.
Fig. 3. Ionic product of high-temperature water and steam at
different temperatures versus pressure. At low pressures, water
behaves as a non-polar solvent with low self-dissociation. High
pressures can increase the ionic product to values above those
found for water at ambient conditions.
corrosion behavior of most materials in high-temperature water.
Supercritical water has both liquid-like and gas-like
characteristics like density between both states, high diffusivity
and good heat-transporting properties. From this point of view,
supercritical water can be seen as a dense gas. Thus, supercritical
water is a medium with excellent transport properties and possesses
a complete solvency for most gases [7,8], and organic compounds
[911]. The ionic product of high-temperature water and steam versus
temperature is illustrated in Fig. 3. Steam and low-density
supercritical water behave like non-polar solvents
with low solvency for ionic compounds. Typical solubility values
for sodium chloride and sodium sulfate in low-density supercritical
water are 100 and 1 ppm, respectively. The solubility of NaCl
versus pressure and temperature is illustrated in Fig. 4 (after
Ref. [12,13]). These characteristics can be tuned by a change of
pressure. In the vicinity of the drop of density and ionic product
mentioned above, a slight modication of temperature and pressure
has a huge effect on the physical and chemical parameters of the
solvent. While chemical reactions occurring in water of high
density are dominated by ionic pathways, low-density water favors
radical reactions [1426]. High-density supercritical water is still
a good solvent for organic compounds, but also for gases and salts.
All these special features lead to interesting applications of
supercritical water. In the present article, the following
expressions are used as follows: Supercritical water: T > Tc
(374 C); p > pc (22.05 MPa) Subcritical water: T < Tc ; p
> pSaturation (p may also be above pc ) Steam: T < Tc (T may
also be above Tc ); p < pSaturation 2. Applications of high
temperature and supercritical water In the last two decades,
high-temperature and supercritical water have become interesting
mediums for
Fig. 2. Drop of physical properties of high-temperature water at
different pressures. Note that the drop shifts towards higher
temperatures at higher pressures. Data after Steamtables, see
Kritzer et al. [171]; reprinted with kindly permission of the
National Association of Corrosion Engineers (NACE).
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
3
Fig. 4. Solubility of NaCl in supercritical water at different
pressures versus temperature. Salt solubility follows the density
drop mentioned above (e.g. Fig. 2). Data after Martynova [13].
Table 1 Applications of supercritical water Application Chemical
reactions Hydrothermal syntheses Waste oxidation Radioactive waste
reduction Biomass conversion Plastic degradation Synthesis of
nano-particles Properties exploited High solvency for organics,
tunable conc. of H+ and OH Solubilities High solvency for organics
and oxygen High solvency for organics and oxygen; solubilities High
solvency for organics High solubility of the monomers Low
solubility of salts Refs. [1826] [27] [20,2833] [3436] [3741] [42]
[4346]
chemistry [1846]. A number of reviews dealing with the general
characteristics of supercritical water have been published recently
[4,5,1417]. Here, only a brief overview is given of the enormously
growing number of different applications so of which are listed in
Table 1.
3. Corrosion in high-temperature and supercritical water 3.1.
Short glossary of the typical forms of corrosion In this section,
the principle corrosion mechanisms observed in high-temperature
water will be described. An evaluation is given, which corrosion
phenomena are most relevant for different applications.
3.1.1. Pitting corrosion Pitting corrosion is a localized form
of corrosion occurring in the passive state of the metal. An
example is shown in Fig. 5. In this case, the oxide lm in principle
has a protective nature. However, aggressive anions such as
chloride or bromide can penetrate into the oxide lm and destroy the
lm locally. Typical initiation points are inclusions or grain
boundaries. Small pits are formed in the rst step. The oxidation
and dissolution of metal components such as nickel and/or chromium
ions followed by their reaction as Lewis acids with the water leads
to a strong acidication of the solution inside the pits. Due to
migration processes from the bulk solution, the concentration of
aggressive anions increases. Thus, the solution becomes more and
more corrosive inside the pits and corrosion progresses. Pit growth
generally occurs in
4
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
Fig. 5. Typical form of pitting corrosion. The phenomenon is
stochastic and due to complicate and changing local chemistry
inside the pits, penetration rates are not linear. Note that the
neighbored surface remains unattacked, underlining the principle
stability of the oxide lm. Corrosion rates are in the range of 1000
m in 100 h (nickel-base alloy 625; conditions: [HCl] = 0.05 mol/kg;
[O2 ] = 0.48 mol/kg; T = 160 C; p = 24 MPa; t = 124 h; see Ref.
[173]).
Fig. 6. Typical wave-like form of general corrosion (shallow
pitting). Since this form of corrosion is diffusion-controlled,
corrosion rates are linear. Note that due to the general
instability of the oxide lm, the complete surface is attacked.
Grain boundaries at the surface might be etched (not visible in
this gure due to too low magnication). Corrosion rates are in the
range of 500 m in 100 h (nickel-base alloy 625; conditions: [HCl] =
0.05 mol/kg; [O2 ] = 0.48 mol/kg; T = 350 C; p = 24 MPa; t = 124 h;
see Ref. [173]).
high rates. Increasing temperature additionally weakens the
oxide lm, and thus, pitting occurs much easier at higher
temperatureswhich is indicated, e.g. by the decrease of the lower
limit of the electochemical potential, at which pitting is observed
with increasing temperature [4767]. The reason for this behavior is
(a) the higher number of locally limited defects in the lm [47,48]
and (b) the increased tendency of the oxide lms to incorporate
anions at higher temperatures [5053]. Note that still large areas
of the surface remain unattacked, while in neighbored areas,
corrosion rates can exceed some 10 m/h. Due to its stochastic and
non-predictive nature, pitting is a dangerous form of corrosion.
Review articles dealing with pitting corrosion have been published
by several researchers [6879]. 3.1.2. General corrosion In contrast
to pitting corrosion, the reason for general corrosion is a general
instability of the oxide lm, and thus, corrosion attacks the entire
surface. The typical morphology is a shallow, wave-like pitting
(Fig. 6). General corrosion happens in such cases, where none of
the alloy components could form a protective layer. This is the
case during the active and transpassive dissolution of materials.
Both processes
are described below. The starting points of general corrosion
are the same like these found for pitting, the weakest points of
the surface. It has been shown for different alloys and stainless
steels that pitting switches to general corrosion above a certain
temperature called the inversion temperature [60,62]. At this
temperature, which is typically in the range of 200250 C, also the
pitting potential no longer decreases, but shifts into that of the
transpassive dissolution. In case of chromium-containing alloys,
chromate is released under these conditions. It was shown that
concentrations of oxygen even below some 100 ppm are sufcient for
chromate formation [60]. The absolute material loss caused by
general corrosion might be high, but due to the
diffusion-controlled character, the corrosion rates are linear and
thus, can be predicted. General corrosion is a common phenomenon in
oxidizing high-temperature water. In cases, where chromate is
released, grain boundaries on the surface might be attacked since
chromate act as a grain-boundary etching agent [60,62]. It must be
mentioned that chromium(III) oxides are still found under these
conditions, although the chromate formation is thermodynamically
favored, which underlines the low conversion rates of solid Cr(III)
towards Cr(VI).
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
5
Fig. 7. Intergranular corrosion. Due to the difference between
grain boundaries and grains, corrosion occurs preferentially in
this regions. Corrosion at this picture occurred at supercritical
temperature; signicant material loss could not be observed.
Corrosion rates are in the range of 10 m in 100 h (nickel-base
alloy 625; conditions: [HCl] = 0.05 mol/kg; [O2 ]=0.48 mol/kg; T =
450 C; p = 24 MPa; t = 150 h, see Ref. [173]).
Fig. 8. Stress corrosion cracking starting at the bottom of a
shallow pit and passing through the whole wall of a tubular reactor
in the temperature transition region between passive and
transpassive states. In this example, failure occurred in less than
50 h in a solution containing HCl and oxygen (nickel-base alloy
625; conditions: [HCl] = 0.10 mol/kg; [O2 ] = 0.48 mol/kg; T = 220
C; p = 38 MPa; t = 24.5;Von Mises stress values are rd. 100 N/mm,
see Ref. [173]).
3.1.3. Intergranular corrosion (intercrystalline corrosion, IC)
The phenomenon of IC has found much attention in literature [8085].
Grain boundaries and the neighbored areas of the grains normally
are chemically different compared to the bulk grains themselves.
Additionally, new phases can be formed at the grain boundaries such
as metal carbides or nitrides. Further, an enrichment, or
segregation, respectively of trace elements at the grain boundaries
lead to detrimental conditions. During IC, either the grain
boundaries or neighboring grain areas might be attacked (Fig. 7).
Local electrochemical elements might be formed. Different corrosion
mechanisms are observed, so IC can be observed under nearly all
conditions. Both, penetration depth and amount of dissolved
material regularly are low. Therefore, IC is not as critical as the
other forms of corrosion. However, whole grains may be dissolved at
longer times and by the inuence of mechanic stress, IC may lead to
the dangerous stress corrosion cracking (SCC). 3.1.4. Stress
corrosion cracking SCC is an extremely dangerous form of corrosion,
since its nature and its occurrence are stochastic (Fig. 8). Thus,
failures can be catastrophic. Therefore, publications dealing with
this form of corrosion
are numerous [86103]. SCC is observed along the grain boundaries
(inter-granular) or through the grains (trans-granular). SCC is
commonly present in the transition ranges between the active and
the passive, or the passive and the transpassive potential,
respectively. Thus, SCC was observed in high-temperature water in
the presence of either hydrogen (active region) or oxygen
(transpassive region). Most detrimental anions are chloride,
bromide, and sulde. SCC needs both a chemical and a mechanical
component. At higher mechanical stress of the material, the
chemical aggressiveness of the solution does not need to be so
high. On the other hand, in highly aggressive environments,
relatively low values of stress can cause SCC. SCC commonly leads
to a failure of the entire reactor due to leakage. If further
burnable gases like hydrogen are released from such a leakage in
the heating section of a reactor, res and even explosions cannot be
excluded. 3.2. Brief historywet air oxidation and supercritical
water oxidation The most widespread use of high-temperature water
is as a heat transporting and heat transmission medium in power
station cycles. Water in this environment
6
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
contains low concentrations of corrosive species such as salts
and oxygen. Additionally, suitable corrosion inhibitors are added
to the water in such systems as noted later. Thus, in practice,
corrosion phenomena that are dominant, are those, in which the
mechanical component plays the most important role. Corrosion
occurring under these conditions has been extensively investigated
[104113]. In nuclear power generation cycles, also the effect of
radioactivity has to be taken into account. Some 30 years ago,
concentrations of inorganic compounds and oxygen above the ppm
level in high-temperature water were only present in the industrial
application of wet air oxidation (WAO) [114118]. WAO is the
oxidative treatment of organic wastewaters at temperatures up to
320 C and pressures up to 20 MPa. Reaction times are 3060 min. The
reactor material mostly is titanium grade 2, which has been proven
to be resistant under these conditions. The purpose of this kind of
waste treatment is to crack stable organic molecules and form
harmless compounds like acetic acid, which can be easily destroyed
subsequently with biological treatment. Currently, over 100 WAO
plants are run successfully world-wide [114,118]. The typical
oxidizer in such systems is oxygen from air, but also nitric acid
has been used. The process of supercritical water oxidation
(SCWO)the most widely investigated supercritical water application
so far, can be seen as an extension of WAO to higher temperatures
and pressures. Typical process parameters of SCWO are reaction
temperatures of 400700 C and pressures of 2450 MPa. During the
process, organic compounds react completely with the oxidantmostly
oxygenin a single-phase reaction to form CO2 and H2 O. Hetero-atoms
like halogens, sulfur, or phosphorus present in the organic wastes
are transformed into their mineral acids HF, HCl, HBr, H2 SO4 or H3
PO4 , respectively. HI is not formed since iodides are rapidly
converted to elemental iodine in the presence of oxygen.
Organically bound nitrogen forms N2 predominately and small amounts
of N2 O. Thereby, N(III) nitrogen is oxidized by the oxidant, while
N(+V), being a powerful oxidant itself, is reduced. Undesirable
by-products common in incineration like dioxins, or higher NOx are
normally not formed, since reaction temperatures not favorable for
their formation. SO2 is transferred to SO3 and removed with the
aqueous
efuent as sulfate. The variety of employed wastes include
warfare agents, rocket propellants, radioactive wastes, and wastes
from the paper and chemical industries [2836]. In contrast to WAO,
SCWO is able to achieve complete oxidation of most organic
materials at short reaction times that range from seconds to
minutes. Common reactor materials like stainless steels at service
times of some 1000 h used in SCWO of aqueous waste streams of
organic compounds containing no hetero-atoms do not exhibit
remarkable corrosion rates. Unfortunately, the degradation of such
compounds is economically uninteresting, since they can be treated
cheaper with alternative methods such as incineration. It has been
found in different studies that SCWO could become economically
interesting for the treatment of highly toxic material like
chlorinated dioxines or S- and P-containing warfare agents.
However, the highly corrosive efuents resulting from the treatment
of these waste streams, together with the precipitation of salts
and a subsequent reactor plugging, has hindered a wider application
of the process so far [33]. The number of investigations of the
corrosion resistance of different reactor materials under SCWO
conditions is enormous [119201]. However, a material that can
withstand every conceivable chemical conditions has not yet been
found. Most reviews dealing with the SCWO process are overly
optimistic, and critical publications dealing with the process
problems are few [28,33]. To solve these existing problems, the
following routes for future research have been suggested [33]:
Determination the corrosion mechanisms occurring under
high-temperature conditions. Understanding the inuence and
interaction of corrosion-determining parameters of the solution
such as temperature, pressure, solution components, and pH.
Understanding the inuence of the reactor material parameters, the
material itself, alloy composition, heat treatment and the level of
impurities. Determination of conditions, where corrosion is low. It
is probable that no single reactor material can withstand every
conceivable reaction conditions. Evaluation, of whether a
particular corrosive species can be excluded, compensated or
avoided so that a
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
7
corrosion-resistant material for the desired application can be
chosen. Construction of special reactor designs to limit contact of
the corrosive species with the reactor wall or temperaturepressure
regimes, in which highly corrosive species are formed, are avoided
[201]. Several SCWO pilot plants are under operation world-wide. A
recent overview over these activities is given elsewhere [28]. In
recent years, also the application of high-temperature and
especially supercritical water as medium for chemical reactions,
synthesis and crystallization processes (see Table 1) has become a
eld of wide interest. In these applications, the presence of salts,
acids, or gases might be necessary for the chemical reaction, since
they might be educts and/or products. 3.3. Solution parameters
inuencing corrosion 3.3.1. Ionic reactions and oxide lm stability
As mentioned above, high-temperature water is a medium of tunable
density, ionic product and dielectric constant (see Figs. 13).
Hence, its solvent character can vary from highly polar at high
densities to nearly non-polar at low densities. High polarity
favors the solubility and/or the dissociation of ionic species like
salts, acids, and bases, and thus favors ionic reactions.
Howeverand this stands in contrast to water at ambient
conditionsalso non-polar gases and organic compounds might be
completely soluble at sufciently high temperatures and pressures
[7,8]. Low-density water suppresses ionic reactions and favors
radical reaction pathways, especially at high temperatures.
Generally, the high temperatures drastically accelerate all
chemical reactions, and thus, kinetic effects become less and less
important. Aqueous corrosion processes commonly are seen as ionic
reactions. The rate-determining step generally is the dissolution
of a protective surface saltmostly an oxide. In high-temperature
water, this oxide might be of different nature or composition than
the surface oxides found under ambient conditions. In general,
oxides are preferentially formed compared to hydroxides at high
temperatures, and the content of crystal water is lower. At ambient
conditions, thermodynamically unstable compounds might form a
protective layer in such
Fig. 9. Stability island of a protective oxide and the principle
mechanisms for its dissolution. Horizontal direction means chemical
dissolution (variation of the pH without changing the potential),
while vertical direction means electrochemical dissolution
(variation of the electrochemical potential without changing the
pH). The stability island correlates with the passivity region of
the metal.
cases, when their kinetic stability is high, and thus, their
dissolution rate is very low. On the other hand, in
high-temperature water of high density, thermodynamic stability is
a necessary condition for a protective lm, since all chemical
dissolution rates are accelerated. The principle reactions to
dissolve such an oxide lm are similar to those found at room
temperature and are illustrated in Fig. 9. A detailed description
of the different dissolution processes is given in the following
sections. 3.3.2. Temperature Corrosion rates in water generally
increase with temperature. Indeed, most corrosion processes have a
minimum temperature, below which corrosion is limited or does not
occur. For example, stainless steels are attacked by
chloride-induced pitting corrosion above some 80100 C, while the
materials are resistant at lower temperatures. However, these
ndings are only true as long as the other solution properties
remain unchanged over the investigated temperature range. The rst
experimental corrosion experiments in supercritical aqueous
solutions gave surprising results: Corrosion was lower by orders of
magnitude at the highest supercritical experimental temperatures of
500 C compared
8
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
with corrosion at subcritical 300 C [119121,130 132,135,136].
This behavior was traced to the change of physical and chemical
properties of water. 3.3.3. pH-value: chemical dissolution The pH
value normally is based on the equilibrium reaction of the
self-dissociation of water H2 O = H+ + OH This reaction has an
endothermic character and thus, the equilibrium shifts towards the
right side with increasing temperature. For example, at a pressure
of 25 MPa, it runs through a maximum at about 300 C (Fig. 2). At
these temperatures, the concentrations of both H+ and OH are about
three orders of magnitude above the values in ambient water, so
that water can be considered to be both acidic and alkaline. A
technical application for this phenomenon are acidor base-catalyzed
reactions such as estherications or hydrolysis reactions in
high-density subcritical water [26]. Under these conditions, strong
mineral acids like H2 SO4 , which are necessary for catalyzing such
reactions at room temperature, are not used, and thus, a separation
of the acidic catalyst in a second step is unnecessary. Higher
temperatures at constant pressure lead to a reduction of the number
of hydrogen bonds and thus favor the non-polar character of water
[15,16]. Consequently, self-dissociation drastically drops down
above a certain temperature. The dissociation of mineral acids
follows the course of density of the solvent [170,202,203]. Since
the dissociation of strong acids is exothermic, the decrease is
strongly monotonous. In Fig. 10, the dissociation of HCl versus
temperature is shown (data after Ref. [203]). Other strong acids
behave more or less similar. Again, the large pressure inuence at
high temperatures is remarkable and, e.g. HCl dissociation
correlates well with the solution density [203]. The pH value, or
dissociation rate, respectively, is one of the factors, which has
most inuence on corrosion in high-temperature water. High or low
values of pH lead to a chemical dissolution, which is described as
dissolution of the protecting oxide at constant electrochemical
potential, and indicated as the horizontal direction in Fig. 9.
This form of dissolution is caused by the amphoteric character of
most oxides; they can be dissolved either in acidic or in alkaline
solutions.
Fig. 10. Dissociation of HCl in water versus temperature after
Frantz and Marshall [203]. Dissociation follows the density of the
solvent. For comparison, the pH of 1 M acetic acid at room
temperature is indicated. It can bee seen that HCl at high
temperatures still acts as acid, if the pressure is high enough.
The behavior of other acids is similar (see, e.g. Ref.[202]).
Typical examples are the oxides of iron, nickel, and chromium.
The pH region of highest stability depends on the isoelectrical
point of the oxide and can change with temperature. Theoretical
work of Adschiri et al. determined the solubility of the protecting
oxides of iron, nickel and chromium in supercritical water versus
pH value at non-oxidizing conditions [188]. For all three metal
oxides, the solubility at neutral or slightly alkaline conditions
is lowest. Chromium oxide show the lowest solubility of the three
oxides, nickel oxide the highest. In a later publication, the same
authors investigated the effect of oxygen on the corrosion of pure
iron, nickel, chromium and different alloys versus pressure at
slightly supercritical temperatures [194,195]. In all cases, an
increased pressure increases the corrosion rate for the pure metals
[194]. Chromium shows the highest stability compared to iron and
nickel, in coincidence with the solubility data. This behavior also
explains the observations, why chromium-containing alloys show a
better corrosion resistance than chromium-free ones [195]. In a
extensive study, Schroer et al. investigated the inuence the
chromium content in binary NiCr alloys and the effect of an
addition of aluminum and tungsten as alloying elements [198]. The
authors also found a strong pressure dependency of the corrosion.
At comparable pressuretemperature conditions 410 C/40 MPa, the
corrosion rates for all these alloys were comparably high, and lie
within one order of magnitude. The
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
9
different corrosion rates were claimed on the different
morphologies of the Cr(III) oxide/hydroxide layer. It must be
mentioned that in special cases, no corrosion might occur, even
although the oxide layer is removed. This is the case for
solutions, in which no oxidative species are present that could
further oxidize the pure metal after initial oxide removal. This
state normally is called immunity of the metal or the alloy. The
value of pH also inuences the solubility of the primary corrosion
products. These are formed intermediately at the rapidly corroding
surface, when local super-saturation occurs leading to a
precipitation of these salts. For example, it has been shown that
corrosion rates of nickel-base alloys in strongly oxidizing
high-temperature solutions of different acids are inuenced by the
solubility of the nickel salts of the acid anion [167,172,173]. On
the other hand, the substance with the lowest solubility, which for
stainless steels and nickel-base alloys in acidic solution are Cr
(III) oxides or hydroxides, are present as solid corrosion products
in form of thick scales on the corroded surface, while nickel is
dissolved selectively. Typical concentrations of metal ions
released from a nickel-base alloy 625 reactor (composition: 62 wt.%
Ni, 22 wt.% Cr, 9.0 wt.% Mo) under oxidizing alkaline, neutral, and
acidic conditions are shown in Table 2 [166]. In the same table,
enrichment factors (EF) for Cr and Mo, normalized to the nickel
content in the alloy, and solution, respectively, are listed.
Acidic solution leads to high concentrations of chemically
dissolved Ni in the efuent. Cr and Mo show low EF values, which
means that Ni is dissolved preferentially. Increased pH values
reduce the Ni concentrations by
orders of magnitude according to the thermodynamic stability of
NiO under these conditions (see Section 3.3.7), while Cr and Mo are
not inuenced that much. Chromium and molybdenum behave similar to
each other. Both metals are dissolved chemically, mainly in acidic
solutions, and electrochemically as thermodynamically favored CrO2
or MoO2 , mainly in al4 4 kaline solutions. The low concentrations
of dissolved Cr under acidic conditions underlines the low
solubility of the chromium(III) oxides. Under neutral, and
especially alkaline conditions, Cr and Mo show a high enrichment
compared to Ni. This, together with the decrease in concentration
of all metals at alkaline conditions can be attributed to the
protecting NiO, which suppresses a further dissolution of the alloy
[165,166,172]. 3.3.4. The electrochemical potential and the
solubility of gases: electrochemical dissolution At room
temperature, non-polar gases like oxygen, hydrogen, or nitrogen are
water-soluble only in low ppm concentrations according to their
Henrys constants. Increasing the pressure at room temperature leads
to only slightly higher gas solubility. Increasing the temperature
at these high pressures results in a slight decrease of the gas
solubility up to around 100 C. Above that temperature, gas
solubility increases drastically [204208]. At supercritical
parameters of the solvent, gases are completely miscible with the
supercritical solvent [7,8]. Since corrosion reactions are
generally oxidation reactions, they are very sensitive to changes
in the electrochemical potential. Under these conditions, the
Table 2 Concentrations of metals released from a tube reactor of
nickel-base alloy 625 corroded by acidic, neutral and alkaline
chloride solutions (Tmax = 350 C; p = 24 MPa; [O2 ] = 0.5 mol/kg;
[HCl] = [NaCl] = [NaOH] = 0.05 mol/kg; further experimental details
are described in Ref. [166]). [Nidissolved ] ppm HCl + O2 NaCl + O2
NaCl/NaOH + O2 1560 10 0.1a [Crdissolved ] ppm 17 24 1.8 EF (Cr)
0.03 6.9 51.6 [Modissolved ] ppm 12 6 0.7 EF (Mo) 0.05 4.2 48.8
The enrichment factor (EF) of an element X compared to nickel is
calculated after the following equation by taking into account the
concentrations of dissolved metals in solution and the alloy
composition (Ni: 62.8 wt.%, Cr: 21.9 wt.% Mo 9.0 wt.%). EFNi (X)
=a
[X]/[Ni] (dissolved) [X]/[Ni] (alloy)
Value is indeed below 0.1 ppm (the detection limit of the
analytical technique), so the EF values for Cr and Mo are even
higher.
10
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
so-called electrochemical dissolution, which is a dissolution at
constant pH value as indicated by the vertical direction of Fig. 9,
occurs. In general, when a metal is protected by an oxide layer, it
can be considered to be in its passive state. For the formation of
the oxide lm, a minimum electrochemical potential is necessary.
Below that potential, the metal might undergo a so-called active
dissolution. In most stainless steels and nickel-base alloys,
chromium is the oxide-forming element. In acidic solutions, the
pure metal and its alloys are passivated by thermodynamically
stable, solid Cr(III) compounds (Cr(OH)3 ; CrOOH; Cr2 O3 ). The
so-formed oxides perfectly cover the underlying metal and protect
it from further degradation. With increasing temperature, the water
content of the oxide decreases and the cristallinity increases. At
some 100 C, amorphous CrOOH is the favored species, while at
temperatures above some 400 C, crystalline CrOOH (Grimaldiite), and
especially Cr2 O3 predominate [166,173,198]. Below the potential,
at which Cr(III) is stable, Cr(II) compounds might be formed, but
these do not form a protective layer. Consequently, the underlying
metal is dissolved rapidly by active dissolution (Cr0 Cr(II)).
Increasing the electrochemical potential above that of the passive
range might lead again to a high dissolution of the metal or alloy.
Chromium then forms the soluble hexavalent species chromate (CrO2
), hydrogen chromate (HCrO4 ) or chromic 4 acid (H2 CrO4 ), which
cannot protect the metal anymore [60,62,166,170,172]. The former
insulating effect of the oxide lm is lost. This process of high
dissolution is called transpassive dissolution and is observed for
many metals. However, titanium, niobium, or tantalum, respectively,
do not undergo a transpassive dissolution at room temperature up to
anodic potentials of some 100 V. In this case, the oxides are
perfect insulators, and also oxygen release caused by water
oxidation is suppressed. This leads to their outstanding corrosion
performance in highly oxidative environments. It should be noted
that due to the changing properties of oxide lms with temperature,
observations made for room-temperature corrosion may not be valid
at higher temperature in most cases. For example, the number of
surface effects in the oxide increases with temperature. An
increased number of defects leads to reduced corrosion resistance.
Thus, metals with high corrosion resistance at low
Fig. 11. Schematical course of the electrochemical potential for
the formation of soluble chromate in acidic and alkaline solution
(after Chen et al. [209]) and the solubility of oxygen (after Zoss
et al. [208]). Figure after Kritzer et al. [170].
temperatures may dissolve rapidly at increased temperatures.
With an increased solubility of oxygen in water, the oxidizing
power of the solution increases, and electrochemical processes
become more and more important (Fig. 11). As an example, the
process of supercritical water oxidation is given and compared with
an oxidative treatment in air at the same temperatures, but ambient
pressure. Taking into account SCWO-typical oxygen concentrations of
around 5 mol/kg and a reaction pressure of 25 MPa, this results in
an oxygen partial pressure of 2.5 MPa. This value is approximately
two orders of magnitude above the corresponding value for
applications in air at ambient pressures. Nevertheless, nickel-base
alloys have shown their corrosion resistance against these strong
oxidizing conditions due to formation of protecting NiO [168].
However, if ionic chlorides are present under these harsh
conditions, and thus, the oxide lm does not protect perfectly any
more, corrosion initiation occurs rapidly and the following
corrosion rates are high [166]. Additionally, the oxidation of the
metal occurs easier at higher temperatures due to the decrease in
the protective nature of the oxide layer. The electrochemical
potential for the transpassive transformation of chromium and
Cr(III) by forming soluble Cr(VI) compounds decreases with
increasing temperature [209] (Fig. 12; [169,170,172,173]). This
means that chromium can be much more easily oxidized in
high-temperature water. Additionally, higher pH-values favor the
Cr(III)Cr (VI) transformation. As mentioned above, already low ppm
values of
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
11
3.3.5. Inuence of anions Anions can promote or inhibit the rates
of different corrosion processes, according to how they interact
with the protective oxide lm layer of the metal. Several possible
inuences are: Oxide-lm destruction: this locally restricted form of
corrosion is commonly known for the halides chloride, bromide, and
iodide, but not uoride, in high-temperature water [166,167]. Under
harsher conditions, also other anions like sulde or sulte may lead
to a localized destruction of the oxide layer. Such localized
corrosion (e.g. pitting corrosion, SCC) is extremely dangerous for
technical applications, since its occurrence is stochastic and
corrosion rates are high and not linear. Corrosion product
dissolution rate: as mentioned before, the rate of
diffusion-controlled corrosion reactions is determined by the
solubility of the primary corrosion products. For example, the
general corrosion of nickel-base alloy 625 at high subcritical
temperatures in oxidizing solutions of HCl and HNO3 occurs in the
same temperature range, but the corrosion rates differ by an order
of magnitude (12; [174]). This can be explained by the higher
solubility of Ni(NO3 )2 compared with that of Ni(Cl)2 and the
diffusion-controlled character of the general corrosion. Anions as
oxidizing agents: especially nitrate is known to be a powerful
oxidizing agent in high-temperature water and thus can corrode
metals also in the absence of oxygen. Furthermore, ions like
sulfate, which are not seen as oxidizing agent at room temperature,
may act as strong oxidizers in high-temperature water [210216].
Under these conditions, the thermodynamically favored suldes,
sultes, or elemental sulfur can be formed. Thereby, the metals
might undergo a fast active dissolution [165]. This dissolution is
even worsen by the formation of thin sulfur lms on the dissolving
metal surface, observed by Marcus and Protopopoff during the
corrosion of chromium and nickel in high-temperature sulfate
solutions [217,218]. Oxide-lm supporting by incorporation: several
anions are incorporated in the oxide lm of nickel-base alloys and
lead to their increased stability. Such modied layers then might
possess a drastically reduced solubility leading to secondary
passivation.
Fig. 12. Corrosion rates of general corrosion caused by
solutions of different acids under similar conditions (Tmax = 500
C; p = 24 MPa; [O2 ] = 0.5 mol/kg; [HCl] = [HNO3 ] = 0.1 mol/kg).
The high corrosion rates caused by nitric acid can be explained
with the different solubilities of the main corrosion products
Ni(NO3 )2 and NiCl2 . The nitrate is by far more soluble. Data
after Ref. [174]; reprinted with kindly permission of Kluwer
Academic Publishers.
oxygen are sufcient for chromate formation in neutral
high-temperature solutions [60,62]. Both, the increased oxidative
aggressiveness of oxygen-containing solutions and the reduced
resistance against oxidation of chromium result in a high
dissolution of the pure metal. Chromium-containing alloys will
corrode in similar high corrosion rates, if no other alloy
component can form an alternative protective lm. This behavior is
the reason, why nickel-base alloys and stainless steels are rapidly
corroded in oxidizing acidic high-temperature water of high
density. Under these conditions, chromium is dissolved oxidatively,
while the oxides of the other two main components, iron and nickel,
are dissolved chemically. Molybdenum and tungsten behave chemically
similar to chromium and are also dissolved oxidatively by formation
of hexavalent species [165,166,172,173]. Due to the lower
solubility of chromium and iron oxides compared to that of nickel
oxides [188,194], iron, and especially chromium enriched solid
corrosion products can be found under such acidic conditions, while
Ni is missing [60,62,165,166]. Further, nickel-base alloys and
stainless steels behave similar under these conditions their high
dissolution rates vary by below one order of magnitude (see, e.g.
Ref. [141,163,172,198,199]).
12
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
Table 3 Inuence of inorganic ions on the corrosion of
nickel-base alloys and stainless steels in high-temperature water
Ion F Cl ; Br SO2 ; SO2 ; S2 O2 3 4 3 S2 NO 3 CO2 ; PO3 3 4 OH + Ha
b
Mode of action Weak complex former Penetrate into & destroy
protecting oxide-lm Oxidative in high-temperature water by forming
S2 and S0b Reductive in high-temperature waterb Strongly oxidizing;
main corrosion products well soluble Low-soluble salts Low-soluble
salts Enhanced solubility of protecting oxides
Result Homogeneous corrosion possiblea ; passivating inuence?
Strong localized corrosion: pitting and SCCa Strong homogeneous
degradation possible Release of H2 possible; SCC possible Strong
general corrosion possible Corrosion-inhibition possible Strongly
passivating; corrosion-inhibition possible Strong general corrosion
possiblea
In the presence of oxidizing compounds. See Refs.
[166,210216].
This positive behavior was proven experimentally for carbonate,
phosphate, uoride, and hydroxide in high-temperature solutions
[56,60,167,175]. It must be the subject to future research, to what
extent these anions could inhibit the detrimental character of,
e.g. other halides. An overview over the inuence of different
inorganic ions on the corrosion in high-temperature water is given
in Table 3. 3.3.6. The connection of the corrosion-determining
factors The interdependencies between the solution parameters and
corrosion are illustrated in Fig. 13 (data after Ref. [170]).
Generally, corrosion rates can be directly correlated with the
density of the solution. Since density and ionic product show a
similar drop in the vicinity of the critical point, a correlation
between corrosion
rates and ionic product would lead to a similar result. This
attempt rst was made by Kriksunov and Macdonald, who created a
suitable model for predicting corrosion of stainless steels and
nickel-base alloys in oxidizing HCl solutions as function of
density [120]. It has been proven experimentally for many chemical
solutions later [161175,186,194]. The dependency of the upper limit
of general corrosion from the density drop of water is shown in
Fig. 14. As mentioned above, high-temperature water of high density
(above 200300 kg/m3 ) still has an ionic character and thus shows
high solvency for ionic species. Thus, density inuences directly
the dissociation of acids, salts, and bases and the solubility of
salts.
Fig. 13. Connection between solution parameters and corrosion.
Thickness of arrows indicate importance of pathways. Figure taken
from Ref. [170].
Fig. 14. Experimentally found corrosion rates of general
corrosion caused by HCl solutions at different pressures (Tmax =
500 C; p = 24/38 MPa; [O2 ] = 0.5 mol/kg; [HCl] 0.05 mol/kg). Note
that the lower temperature limit of general corrosion is
independent of pressure, while the upper temperature limit strongly
increases with pressure (and correlates with the density drop).
Data Ref. [172]; reprinted with kindly permission of the National
Association of Corrosion Engineers (NACE).
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
13
The dissociation of acids, salts, and bases correlate directly
with the density (see, e.g. Figs. 2 and 10) or the ionic product of
water. For example, HCl is completely dissociated by forming H+ and
Cl at high densities, while at low densities, exclusively the
non-dissociated HCl is present (and the solution becomes neutral).
In this case, HCl is completely miscible with the supercritical
water. As mentioned above, both H+ and Cl are corrosive species,
while HCl itself is relatively harmless at temperatures up to about
600 C. High densities thus create a highly corrosive environment in
the presence of acids. The solubility of salts is directly inuenced
by solution density. Low-density water can only slowly remove salts
in their ionic form. Under these conditions, the energies necessary
to break the salt crystal lattice cannot be compensated by
hydration energies. Consequently, salts, which are highly soluble
at ambient conditions, are nearly insoluble in low-density
high-temperature water. Typical values for the solubility of NaCl
at different temperatures and pressures (densities) are shown above
in Fig. 4. The solubility of Na2 SO4 is even lower by around two
orders of magnitude. Both, the protecting oxides on metal and
alloys and the primary corrosion products are salts. Enhancing
their solubility automatically means enhancing corrosion. In
low-density water, a removal of formed ionic compounds by the bulk
solution in an open equilibrium process does not occur in
remarkable rates. Thus, the formed primary corrosion products can
plug the surface and cause a new kind of passivity. In principle,
the complete miscibility of oxygen with low-density supercritical
water and thus high partial pressure of oxygen should accelerate
the cathodic corrosion reaction. On the other hand, the low
solubility of ions hinders corrosion processes in the following
way: The cathodic reaction at the metal | solution interface
follows the equation O2 + 2H2 O + 4e 4OH
Fig. 15. Density range of high corrosion at different
temperatures. General corrosion is low at densities below about
200300 kg/m3 . At 300 C, a pressure of already 10 MPa leads to high
corrosion rates, while at 500 C, pressures above around 50 MPa are
necessary for high corrosion. Increasing the pressure at constant
temperature increases the rate of electrochemical corrosion. Note
that the shift from no corrosion towards strong corrosion is only
sharp at lower temperatures, while there is no clear dividing line
at higher temperatures.
However, this reaction leads to an enrichment of negative
charged hydroxyl ions at the metal surface, which cannot be
dissolved by the non-polar solvent. Thus, a further cathodic
reaction cannot occur. As conclusion, corrosion in high-temperature
and supercritical water strongly correlates with density or ionic
product. Corrosion is high at high densities and
negligibly low at lower densities (density below about 200300
kg/m3 ). Consequently, the generalization that supercritical water
is not corrosive at all is true for supercritical water at low, but
not for such at high densities. Highly pressurized supercritical
water causes comparably high corrosion rates like dense subcritical
water (Fig. 14). Fig. 15 illustrates the pressuretemperature regime
where high ionic corrosion is expected. For nickel-base alloys, at
a temperature of 300 C and pressures above only around 10 MPa (i.e.
above the saturation pressure), high rates of corrosion are
expected. On the other hand, temperatures of 500 C need minimum
pressures of approximately 50 MPa for high rates of ionic
corrosion. The density of steam is much lower compared with the
density of water above the saturation pressure (see Fig. 15).
Consequently, electrochemical corrosion rates caused by steam are
by orders of magnitude below those found for high-density water at
the same temperature. The mechanism of high-temperature corrosion
is explained in Section 3.5. It must be noted that the correlation
between the dielectric constant and corrosion does not lead to a
similar exact result. This could be explained by the less inuence
of the dielectric constant on the solubility of salts in
high-temperature water.
14
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
Stress corrosion cracking (SCC) has been observed in
chloride-containing media at high densities [146,166]. It has been
shown that the effect of bromide, which is also known for its
SCC-inducing behavior, is at least not much worse than that of
chloride [167]. Iodide causes pitting at room temperature, but is
oxidized rapidly to non-corrosive iodine in the presence of oxygen
at higher temperatures [167]. As mentioned above, the most
dangerous areas for SCC are the potential transitions ranges
between active and passive, or passive and transpassive region,
respectively. SCC in the activepassive transition is well known
from high-temperature water in power applications, where traces of
hydrogen gas might be formed and described in detail elsewhere
[86,91,101]. In special cases, SCC is also observed at other
potentials. For example, SCC can occur, if pitting corrosion has
thinned the reactor walls leading to an increased mechanical
stress. However, no observation of SCC in reactors at low-density
supercritical solutions has been reported in literature thus far.
Recently, Fournier et al. investigated the SCC behavior of high
nickel alloys in de-aerated and aerated low-density supercritical
water [178]. The authors found a dependency of SCC and oxygen
content, but this behavior was also observed in a pure oxygen
atmosphere. The main reason for SCC susceptibility are niobium
carbides. Nevertheless, the investigations do not cover the
high-density region, where SCC occurred predominantly as showed in
several screening tests. There are several reasons for the current
lack of SCC in low-density water. First, the susceptibility of the
materials against corrosion is much higher at other temperatures
and higher densities, so if the material fails, it will not fail
under low-density supercritical conditions. Second, for SCC to
begin, the protective oxide layer has to be destroyed locally,
which can occur through the penetration of chloride. In low-density
solutions, the concentration of free ionic chlorides and other
potentially detrimental anions is extremely low. Consequently, the
oxide lm is either attacked generally by chemical dissolution
processes or is not attacked at all. Finally, like other ionic
corrosion phenomena, SCC requires the removal of dissolved metal
ions out of the small crevices. However, the capacity of
low-density water to dissolve salts and thus to remove the
corrosion products is too low for crack propagation.
Nevertheless, the results of Founier et al. are important for
the long-time stability of reactor materials under supercritical
water conditions. It must be also the aim of future work to
evaluate, whether SCC is a problem in the hydrogen-generating
process of supercritical biomass conversion. While hydrogen in
trace concentrations extremely accelerates SCC, there are
indications that high nickel alloys are much less susceptible for
SCC at high hydrogen concentrations [219,220]. Nevertheless, most
critical for initiation of hydrogen-induced SCC are the presence of
impurities, which are enriched at the grain boundaries of the alloy
[86,101,221]. It must be again noted that such an enrichment might
occur in-situ during long-time operations at temperatures above 600
C, so that the initial SCC resistance might be much better than the
resistance during operation. 3.3.7. Low corrosion pHpotential
regions This section should give an overview how to reach
pHpotential-regions, in which low corrosion is expected. As
mentioned above, at least one of the alloying componentschromium or
nickelmust be able to form a protecting oxide. In Fig. 16, the
stability islands of chromium and nickel are shown schematically.
At an acidic pH and moderate electrochemical potential (state A),
chromium protects the alloy. Note that pure nickel would undergo a
chemical dissolution at these conditions. An increase of potential
at constant pH leads to an instability of both chromium and nickel
(state B). Chromium thereby is dissolved oxidatively as chromate,
while nickel is dissolved chemically. Experimental investigations
performed with pure nickel and
Fig. 16. Stability islands of chromium and nickel. Chromium is
more stable against acidic solution, while nickel better tolerates
oxidizing conditions. For the reactions leading to
instability/stability, see text. E = electrochemical potential.
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
15
chromium under such acidic, oxidizing conditions showed a fast
dissolution of nickel [171]. Chromium was also completely unstable,
but dissolution rates were much slower [171]. The mechanism leading
to this behavior is described in Section 3.3.4 and illustrated in
Fig. 11. By increasing the pH to neutral values at these high
electrochemical potentials, a state C can be reached, where nickel
is able to protect the alloy. Under these conditions, protecting
NiO is formed, which has been found experimentally protecting
nickel-base alloys in low-density supercritical water
[166,167,172,197]. Note that pure chromium would undergo an
electrochemical dissolution at these conditions by forming the
well-soluble chromic acid H2 CrO4 . The pH increase can be obtained
as follows: an addition of small amounts of a base like NaOH leads
to a pH shift into the alkaline direction. Experimental results
have shown that under these conditions, nickel-base alloys are not
corroded also by oxygen- and chloride-containing high-density
solutions [166,177].
A similar pH shift is reached when the subcritical supercritical
transition is crossed, e.g. by increasing the temperature, and the
density of the water and the dissociation of acids become low. This
explains the low corrosion of low-density supercritical water also
if acids are present [120,170,172]. Alternatively, the reduction of
the electrochemical potential can bring the process back to state
A, in which the material is protected by chromium oxides. This
could be obtained by the exclusion of oxidizing compounds, or
possibly by an electrochemical protection of the reactor wall.
Whether such an protection works in practice would have to be the
subject of further investigations. 3.3.8. The corrosion mechanism
of nickel-base alloys and stainless steels in oxidizing acidic
solutions The corrosion mechanism of nickel-base alloys in
oxidizing high-temperature HCl solution has been published earlier
(see Ref. [166]). This mechanism, which is a summary of the above
sections, is shown in Fig. 17. Corrosion in oxidizing sulfuric acid
solutions
Fig. 17. Mechanism of the corrosion of nickel-base alloys in
oxidizing HCl solution (slightly modied after Ref. [166]).
Corrosion in the passive state is pitting corrosion (see Fig. 5),
while in the transpassive region, general corrosion (see Fig. 6)
occurs. Note that the rates of pitting can be much higher than
these of general corrosion, while the whole material loss can be
less. Stress corrosion cracking (SCC; see Fig. 8) is observed
primarily in the transition region between passive and transpassive
region, but can also start at the bottom of pits in the passive
region (increased mechanical stress!). In the low-density region,
NiO passivates the alloy. Salts or corrosion products dissolved at
lower temperatures might precipitate here (esp. NiCl2 ).
16
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
Table 4 Typical corrosion temperatures and corrosion rates of
intergranular corrosion (IC), pitting corrosion, general
dissolution and stress corrosion cracking (SCC) observed for alloy
625 in oxidizing aqueous solutions of various acids ([acid] =
0.050.2 mol/kg; [O2 ] = 0.5 mol/kg; p = 24 MPa) Temperature of
occurrence IC Pitting > 100 > 150200 Corrosion rate ( m/100
h) 1020 5001000a Occurs in All solutions HCl/O2 NaCl/O2 HBr/O2
HCl/O2 NaCl/O2 HBr/O2 H2 SO4 /O2 HNO3 /O2 HCl/O2 HBr/O2 Inhibitors
No OHc CO2 c 3 PO3 c 4 Fd OHc CO2 c 3 PO3 c 4 Fd OHc CO2 d 3 PO3 c
4 Fd
General dissolution
> 250300
500b
SCC
250300
1000a
Data for HCl, H2 SO4 , H3 PO4 or HNO3 , respectively, were taken
from Kritzer et al. [see text]. a Corrosion rates not linear. b
Rates in sulfuric acid and hydrochloric acid. Values in nitric acid
are signicantly higher ( > 1500 c Effect proven. d Effect has to
be proven.
m in 100 h).
follows a similar behavior, but in this case, no pitting
corrosion in the passive state is observed [165]. Mechanisms
explaining the corrosion in oxidizing aqueous solutions of HBr
[167], HF [167], HNO3 [174], and H3 PO4 [175] solutions are
published elsewhere. The temperatures and corrosion rates of the
various forms of corrosion observed for nickel-base alloy 625 is
shown in Table 4. 3.3.9. Chemical water treatment strategies in
power stations Since heat-transport applications in power station
cycles are the most important application of high-temperature and
supercritical water to date, in this section, a short overview will
be given on the various types of
conventional water treatment to minimize corrosion. These
different treatments are summarized in Table 5. Until the early
seventies, phosphate treatment was commonly used for subcritical
circulation boilers. In this kind of boiler, steam is continuously
removed from the boiler system. Since salts are more soluble in the
aqueous phase, an enrichment can occur in this phase. This has led
to high corrosion rates and sudden breakdown of some reactors in
the past. It was found that the concentrated phosphate solutions
were able to dissolve the protecting iron oxides of the stainless
steels, followed by a rapid oxidation and dissolution of the
underlying alloy. Due to the low solubilities of phosphate in
low-density water, this strategy cannot be applied for
supercritical reactors.
Table 5 Typical concentrations of additions in different
water-treatment strategies in power stations (after Ref. [228])
Phosphate treatment Feed water Boiler water pH value [O2 ] [N2 H4 ]
N 2 H4 Na2 HPO4 9.5 < 7 ppb 1030 ppb Volatile treatment N2 H4
/NH3 N/A 9.5 < 7 ppb 1030 ppb Oxygen treatment (combined water
treatment) O2 /NH3 N/A 89 100 ppb 0
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
17
To avoid precipitation, the so-called all-volatile treatment was
introduced. The key issue of this kind of treatment are the
reducing, alkaline conditions caused by an ammonia/hydrazine
containing solutions. Unfortunately, it was shown that such a
treatment could cause elevated corrosion of stainless steels and
nickel-base alloys. The oxide lm on stainless steels in alkaline
high-temperature solutions contains of a double-layer with porous
Fe3 O4 in the outer and protecting Fe2 O3 in the inner layer [222].
However, under reducing conditions, the formation of the protecting
inner lm might be incomplete and the alloy shows elevated corrosion
due to active dissolution. In case of nuclear applications, the
materials of construction are weakened by high doses of radioactive
rays. Further, any metal ions released from these materials carry
off the radioactivity and cause radioactive waste. Problems in
supercritical boilers arise from the possibility of salt plugging
if corrosion products, which are released at lower, subcritical
temperatures, deposit at higher temperatures. This again could lead
to a plugging of the lines. In the combined water treatment method,
the solution is oxidizing through an addition of around 100 ppb
oxygen. By addition of ammonia, the pH value is shifted to the
weakly alkaline region. Under these conditions, the formation of
the protecting iron oxide lm is supported. Consequently, corrosion
rates observed in power stations using this kind of treatment, are
signicantly lower. However, chromium and molybdenum, both are
frequent alloying elements of stainless steels, are not stable
under these conditions and transformed oxidatively to soluble
chromate and molybdate, respectively. Both metals are transformed
to their corresponding soluble acid H2 MeO4 (Me = Cr, Mo) in
low-density supercritical water and thus can pass through the
reactor without causing plugging problems. Consequently, a small
amount of both compounds is released during the operationespecially
during warm-up procedures. In subcritical once-through reactors,
the presence of both NH3 /O2 does not cause problems, since
oxidation of ammonia is slow. In contrast, the oxidation rate of
ammonia is strongly accelerated under supercritical conditions. The
main oxidation product is nitrogen gas, but up to several percent
of N2 O might be formed. Both reactions lead to ammonia and oxygen
consumption.
Do to the consumption of ammonia, the pH is lowered to less
alkaline values. Since iron oxides have their highest stability at
a pH around 10, this could lead to an enhanced dissolution of the
oxide layer. However, as ammonia is present in clearly higher
concentrations than oxygen, this effect should be of minor
relevance. Nevertheless, the under-stoichiometric reaction could
lead to favored formation of N2 O. Since oxygen is present clearly
below the stoechiometric amount for ammonia oxidation, it can be
assumed that the consumption of oxygen is close to 100%. It must be
kept in mind that oxygen was introduced to support the protective
character of the iron oxide lm, so after its consumption, corrosion
is enhanced. It should be clear that none of the present chemical
treatment strategies can eliminate all corrosion problems. An
alternative treatment could be the addition of small amounts of
NaOH, or KOH instead of NH3 as PH stabilizer, respectively, which
cannot be oxidized by the solution and thus might lead to
acceptable corrosion rates inside the reactor. 3.4. Material
parameters 3.4.1. Stainless steels and nickel-base alloys: alloy
composition It is obvious that the composition of an alloy has a
large inuence on its corrosion resistance. Chromium as alloying
element improves the resistance against acidic and oxidizing media
and reduces pitting corrosion, nickel improves the corrosion
behavior in alkaline environments. Molybdenum causes a passivating
effect at low, reducing potentials, at which other alloys show
active dissolution. Consequently, the rst choice of an alloy used
in highly oxidizing, acidic solutions at moderate high temperatures
is one with high chromium content. However, as mentioned above, the
tendency towards chromate formation increases with temperature and
thus, chromium loses its protective effect. As shown in different
studies, the conventional high-chromium nickel- and iron-based
alloys show a similarly high corrosion rate under these conditions.
To found a stable alloy, the alloying elements have to form an
oxide layer which completely covers the alloy. The oxides of
niobium, tantalum, aluminum, zirconium and yttrium are known to be
chemically
18
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
stable in oxidizing high-temperature water. It could be an item
of future research to evaluate, whether alloys containing these
elements in concentrations, which are high enough for oxide lm
formation, show the desired corrosion resistance. Additionally,
solubility screening tests with further oxides, e.g. these of the
rare earth elements, or with the metals themselves, and nally with
model alloys containing these metals, could also lead to an
improved further strategy for the development of a stable alloy. In
neutral or weakly alkaline oxidizing high temperature solution, as
well as in low-density supercritical solution, nickel forms the
protective layer, while chromium is unstable. It must be the item
of future research to prove, if chromium is need as alloying
component for reactor materials under these conditions at all. The
presence of chromium always bears the risk of a release of
chromate, which is crucible in the efuent solution. In reducing
high-temperature water, e.g. during the reductive conversion of
biomass, a further increase of molybdenum in the alloy composition
might improve corrosion resistance, if chromium fails. 3.4.2.
Stainless steels and nickel-base alloys: heat treatment and surface
condition Most commonly used reactor materials are stainless steels
and nickel base alloys. For these alloys, material parameters have
a high inuence on general corrosion, and especially pitting
corrosion and SCC. The presence of non-metallic inclusions in the
metal matrix were found to be starting points of pitting and
general corrosion [68,171]. Most detrimental inclusions are
manganese sulde and titanium nitride. These inclusions lead to weak
points of the oxide layer, and thus to preferential starting points
of corrosion (see Fig. 18). Also, the composition of grain
boundaries of the alloy has a remarkable effect on corrosion (see
sections above). Depending on the kind heat-treatment of the alloy,
the grain boundaries may be depleted of one or more of the alloy
components. Most detrimental from corrosion side is a chromium
depletion in areas of some hundreds of nanometers around the grain
boundaries, at which chromium carbides are formed. In the
neighbored regions of these carbides, chromium concentrations can
decrease to values of 50% of the concentration in the bulk alloy
[81,82,84,86]. This
Fig. 18. Micro-pit with a diameter of approximately 2 m starting
at an inclusion under transpassive conditions (nickel-base alloy
625; polished surface, [HCl] = 0.05 mol/kg; [O2 ] = 0.48 mol/kg; T
= 350 C; p = 24 MPa, t = 0.75 h; see Ref. [171]. Corrosion spreads
out over the surface rapidly leading to the shallow pits shown in
Fig. 6). Reprinted with kindly permission of the National
Association of Corrosion Engineers (NACE).
leads to a less noble alloy in small restricted areas.
Additionally, impurities like phosphorus or sulfur may become
concentrated in the neighboring areas [85]. Non-metallic
inclusions, grain boundary depletion, and strong enrichment of
non-metallic contamination can form and disappear in certain
temperature zones. At temperatures above some 900 C, depending on
the alloy, the solubility of inclusions and non-metallic forms of
contamination in the bulk metal matrix is high enough, so the
formation of separated phases does not occur. At temperatures below
some 600 C, the migration velocity of the impurities towards the
grain boundaries and the formation of these phases is too slow, and
thus, the time frame, during they will occur, exceed some 10,000 h.
As a result, the most detrimental temperature region is between
about 600 and 900 C, where the solubility of the impurities is
relatively low, but their migration to the grain boundaries and
precipitation is fast enough to happen within certain times. These
impurities are known to be responsible for intergranular attack and
stress corrosion cracking (see section above). With the shifting of
the process temperatures of either SCWO and supercritical biomass
conversion into these temperature ranges, also the risk for
impurity-induced SCC increases. It should be an object of future
research to evaluate the effect
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
19
The surface condition also has an inuence on corrosion behavior.
Polished surfaces are known to be more resistant to corrosion than
non-polished ones, since the number of starting points of corrosion
is lower [171]. However, after corrosion once has started, the
corrosion rate of polished surfaces might be signicantly higher,
since the total anodic current available for metal dissolution is
distributed among a lower number of parallel growing pits resulting
in an accelerated growth rate of the individual pit [171]. 3.4.3.
Other materials Different other reactor materials other than
stainless steels or nickel-base alloys have been suggested for
supercritical water applications. For an literature overview over
the mechanisms of these materials, see (Table 6). Such materials
generally have lower mechanical strength and thus, need special
reactor constructions for their use in supercritical water
applications. Among them were noble metals like gold and platinum,
the valve metals titanium, niobium and tantalum and different oxide
ceramic materials. Gold, platinum, and titanium have been
successfully used as liners inside a stainless steel pressure tube
[141,185]. Ceramics can be applied in form of a swimming tube
inside an outer high-grade alloy pressure tube [155]. In such
reactors, the outer tube is only in contact with
pressure-transmitting de-ionized water, while corrosive reactions
can be performed at the same pressure inside the ceramic tube.
Fig. 19. A series of Micro-pits starting at grain boundaries
also under transpassive conditions (nickel-base alloy 625; polished
surface, [HCl] = 0.05 mol/kg; [O2 ] = 0.48 mol/kg; T = 350 C; p =
24 MPa, t = 0.75 h; see Ref. [171]. Reprinted with kindly
permission of the National Association of Corrosion Engineers
(NACE).
of the purity of the alloy on long-term stability of a
SCWO/supercritical biomass conversion reactor. Also, the very rst
step of general corrosion starts at inclusions, or grain
boundaries, respectively, and leads to micro-pits, which is shown
in Fig. 19. It must be the subject of future work, whether such
precipitations can be formed or worsen under high-temperature
supercritical water applications leading to the occurrence IC or
SCC at longer operation times, or if alloys with higher purity are
required under these conditions.
Table 6 Literature data for the corrosion and corrosion
mechanisms of different reactor materials in oxidizing
supercirtical solutions of different species Material Stainless
Steels Nickel base alloys Corroding agent and Refs. HCl
[119,126,127,133,139,151,163,196]; H2 SO4 [185,194]; NaCl [195] HCl
[127,129,131,139,141,150,151,163,166,173,179,180,183,184,192,196,197];
H2 SO4 [165,173,185,194]; H3 PO4 [173,175]; HNO3 [174]; HF [167];
HBr [167]; NaCl [166,173,183,184]; Na2 SO4 [165,173]; NaOH
[169,173] HCl [182,187,193] HCl [171,173,183,187,193] HCl
[171,173,183,187,193] HCl [171,173,183] HCl
[131,142,157,159,183,190,200]; H2 SO4 [159,186]; H3 PO4 [159,186];
NaCl [159] HCl [176]; H2 SO4 [176] HCl [156,190]; H2 SO4 [156]; H3
PO4 [156] HCl [140,141,190] HCl [153] HCl [153,155]; H2 SO4 [155]
H3 PO4 [155]; HF [167]; HBr [167]
Iron Chromium Nickel Molybdenum Titanium Niobium Tantalum Gold
and platinum non-Oxide ceramics Oxide ceramics
20
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
3.5. Exceptions and cautions In the following section,
exceptions and cautions are described, which cannot be classied by
the previous sections. These show that corrosion in
high-temperature aqueous solutions can still be unpredictable. The
presence of high concentrations of NaOH in low-density
supercritical solutions leads to the formation of a liquid NaOH
phase. This phase heavily attacks, e.g. alumina ceramics [155],
which are chemically dissolved. In presence of oxygen, most
metallic materials are attacked [169,170,172]. Under these
circumstances, the formed metal oxides like NiO are dissolved
rapidly in the liquid melt and thus, passivation cannot occur
[172]. However, in the absence of oxygen, nickel-base alloys are
immune [173]. Niobium and tantalum, which show an excellent
corrosion resistance in high-density subcritical solutions,
surprisingly show a fast corrosion at higher temperature, although
their oxides are chemically stable [156,176]. The reason for this
unexpected behavior was found in a phase transformation of the
protective oxide layer: At lower temperatures, an amorphous,
protective oxide layer is present, which is converted to a
non-protective crystalline one at higher temperatures. The metals
are oxidized completely in short times and only the stable oxides
remain. In special cases, the unexpected occurrence of new phases
creates a highly corrosive environment. While, e.g. an oxidizing
solution ([O2 ] = 0.5 mol/kg; T = 430470 C; p = 24 MPa) containing
0.1 mol/kg phosphoric acid causes almost no corrosion of
nickel-base alloys, an increase of the acid concentration to 0.2
mol/kg leads to deadly high corrosion rates of half a millimeter
per hour at the upper side of the reactor. The corrosion mechanism
describes this behavior with the immediate presence of nickel(III)
phosphates, which possess a melting point lower than the process
temperature and thus are removed rapidly [175]. At temperatures
above about 600700 C, another mechanism of corrosion, the so-called
high-temperature corrosion may occur. At these temperatures, the
common reactor materials iron, nickel or chromium begin to form
volatile corrosion products in the presence of acids and salts such
as NiCl2 or NiBr2 . The corrosion products can be removed quickly
or even melt on
the surface leading to high general corrosion rates. Although
these reactions contain oxidation processes of the metals, they are
not seen as electrochemical reactions. Reviews dealing with the
process of high-temperature corrosion are published [223228].
4. When and where is the corrosion low? The text above might
lead to the impression that corrosion is always high in
high-density, hightemperature water. However, it has been shown
that corrosion of nickel-base alloys and stainless steels is low
under the following conditions: At low concentrations of salts and
especially certain acids, corrosion rates are low (also in the
presence of high concentrations of oxygen). All strong acids (e.g.
HCl, H2 SO4 , HNO3 , HBr)also in tracesshould be avoided. The
presence of bases increases the stability of protective iron- and
nickel oxides and thus has a corrosion-inhibitive effect. In the
presence of phosphate, carbonate or uoride, these ions are
incorporated in the oxide lm leading to a protective layer. This
mechanism might also work in the presence of aggressive ions like
chloride, but this assumption has to be proven in future
experimental work. Tolerance of the metal surface against chloride
or bromide in the presence of a certain passivating species has to
be evaluated. It has been shown that some materials might be
corrosion resistant in cases, where high corrosion is observed for
other materials. For example, titanium is resistant against
oxidizing high-temperature solutions of HCl also at high pressures
[159,187], where nickel-base alloys fail. On the other hand,
sulfuric acid attacks titanium at conditions, where nickel-base
alloys are immune [187]. Therefore, it is absolutely necessary to
know the composition of the attacking solution. For each solution,
an optimized reactor material can be chosen. It must be critically
mentioned that all corrosion tests that have been performed for
SCWO applications have lasted not more than several hundreds hours.
This is mainly because of the highly corrosive environment that
lead to early failure. Long-time tests are necessary to obtain
information about the corrosion
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129
21
resistance for power generating systems at typical operation
times of months to years. It must also be mentioned that
salt-doping (esp. sodium phosphates) of high-temperature water,
which is common for subcritical power systems to reduce its
aggressiveness, will lead to severe salt precipitation problems.
The use of a combination of oxygen, to passivate the material, and
ammonia, to adjust the pH, in the so-called oxygen treatment might
cause problems since both components are able to react with each
other. In every case, most problematic in corrosion considerations
is the high-temperature, high-density regime.
5. Outlook With the increased number of applications of
high-temperature and supercritical water, the need for a
corrosion-resistant reactor material increases. Unfortunately,
there is still a lack in basic physical and chemical data. Without
these data, supercritical water may be a dangerous medium in some
cases. Especially lacking are: Research on the solubility of oxides
and salts: the solubility of inorganic compounds (e.g. oxides) in
high-temperature water under various pH and electrochemical
potential conditions. A lot of work in this eld has been done by
geochemical researchers, but these investigations mainly are
focused on pressures above some 100 MPa, which is out of scope for
most supercritical water applications. Nevertheless, a
collaboration between geologists, corrosion scientists and
supercritical water users should be very fruitful for a better
understanding of dissolution behavior and thus corrosion occurring
in high-temperature, high-pressure water. Research on the phase
behavior of supercritical aqueous systems. Few investigations exist
for systems containing more than two compounds. But such
multi-phase systems are present if supercritical water acts as a
reaction medium. Since not all physical phenomena are understood in
detail, extrapolations and predictions obtained from experiments at
lower temperatures are sometimes inaccurate. Research on the
possibility of inhibit corrosion by an addition of certain
compounds (see Section 4). Such corrosion inhibitors could offer
the
possibility to tolerate higher concentrations of corrosive
anions inside the reactors as a preferable technical alternative
over a dilution of the solutions. Research on the inuence of metal
purity on stress corrosion cracking. This item is important for
applications, where the temperature of the reactor materials are in
a range above the sensibilization temperature of some 600 C, and
especially, if hydrogen is present. So far, the corrosion item in
the process of biomass conversion, where hydrogen is released, has
not reached much attention. Corrosion tests under realistic
conditions. Such tests are necessary and cannot be replaced by
theoretical work. It has been shown in the past that calculations
dealing with the transition from subcritical to supercritical water
do not always account for the drastic change of physical and
chemical properties. Tests should also consider the mechanical
inuence. Lots of corrosion tests done in the past have been
performed with coupon samples and thus do not take into account
stress corrosion phenomenaalthough SCC is one of the most dangerous
forms of corrosion for technical applications. Embrittlement can
occur suddenly after long operation times. Until extensive
long-time studies have not been performed, a too optimistic
attitude rather hinders supercritical water applications.
Acknowledgements The experimental work was carried out during my
PhD at the Institut fr Technische Chemie at the Forschungszentrum
Karlsruhe, between 1995 and 1998. I would like to thank especially
Gnter Franz, Claus Friedrich, Wilhelm Habicht, Michael Schacht and
Basil Kanellakopoulos for the help and fruitful discussions.
Further, I would like to thank Nikolaos Boukis and Eckhard Dinjus
for the opportunity of performing this work in their
group/institute and the support they gave. Finally, I want to
acknowledge the National Association of Corrosion Engineers (NACE),
Houston, TX, USA, for honouring my work with the A.B. Campbell
Award 2000. Parts of this work were presented at the First
International Symposium on Supercritical-cooled Reactors
(SCR-2000), Tokyo, 2000.
22
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129 E. Kiran,
P.G. Debenedetti, C.J. Peters (Eds.), Supercritical
uidsfundamentals and applications, NATO Science Series E, Vol. 366,
Kluwer Publishers, Dodrecht, 2000, p. 345. T. Mizuno, Properties
and application of supercritical water, Zairyo-to-Kankyo 47 (1998)
298. B. Kuhlmann, E.M. Arnett, M. Siskin, Classical organic
reactions in pure superheated water, J. Org. Chem. 59 (1994) 3098.
P.E. Savage, S. Gopalan, T.I. Mizan, C.J. Martino, E.E. Brock,
Reactions at supercritical conditions: applications and
fundamentals, AICE J. 41 (1995) 1723. H. Schmieder, N. Dahmen, J.
Schn, G. Wiegand, Industrial and environmental applications of
supercritical uids, in: R. van Eldik, C.D. Hubbard (Eds.),
Chemistry under Extreme or Non-Classical Conditions, Spektrum,
Heidelberg, 1996, p. 273. T. Adschiri, Material synthesis in
supercritical water, Zairyo-to-Kankyo 49 (2000) 126. P.E. Savage,
Organic chemical reactions in supercritical water, Chem. Rev. 99
(1999) 603. P.G. Jessop, T. Ikariya, R. Noyori, Homogeneous
catalysis in supercritical uids, Chem. Rev. 99 (1999) 475. A.
Baiker, Supercritical uids in heterogeneous catalysis, Chem. Rev.
99 (1999) 453. J.A. Darr, M. Poliakoff, New directions in inorganic
and metalorganic coordination chemistry in supercritical uids,
Chem. Rev. 99 (1999) 495. D. Broll, C. Kaul, A. Kramer, P. Krammer,
T. Richter, M. Jung, H. Vogel, P. Zehner, Chemistry in
supercritical water, Angew. Chem. Int. Ed. Engl. 38 (1999) 2999. A.
Rabenau, Die Rolle der Hydrothermalsynthese in der prperativen
Chemie, Angew. Chem. 97 (1985) 1017. R.W. Shaw, N. Dahmen,
Destruction of toxic organic materials using supercritical water
oxidation: current state of the technology, in: E. Kiran, P.G.
Debenedetti, C.J. Peters (Eds.), Supercritical FluidsFundamentals
and Applications, NATO Science Series E, Vol. 366, Kluwer
Publishers, Dodrecht, 2000, p. 425. H. Schmieder, E. Dinjus,
Supercritical water oxidation: state of the art, Chem. Eng.
Technol. 22 (1999) 903. J.W. Tester, J.A. Cline, Hydrolysis and
oxidation in subcritical and supercritical water: connecting
process engineering science to molecular interactions, Corrosion 55
(1999) 1088. H.J. Bleyl, J. Abeln, N. Boukis, H. Goldacker, M.
Kluth, A. Kruse, G. Petrich, H. Schmieder, G. Wiegand, Hazardous
waste disposal by supercritical uids, Sep. Sci. Technol. 32 (1997)
459. S.F. Rice, R.R. Steeper, Oxidation rates of common organic
compounds in supercritical water, J. Hazardous Mater. 59 (1998)
261. P. Kritzer, E. Dinjus, An assessment of supercritical water
oxidation (SCWO)existing problems, possible solutions and new
reactor concepts, Chem. Eng. J. 83 (2000) 207. G. Limousin, C.
Joussot-Dubien, Y. Garrabos, S. Sarrade, Modelling of dodecane
hydrothermal oxidation in a tubular
References[1] E.U. Franck, berkritisches Wasser als
elektrolytisches Lsungsmittel, Angew. Chem. 73 (1961) 309. [2] E.U.
Franck, Physicochemical properties of supercritical solvents, Ber.
Bunsenges Phys. Chem. 88 (1984) 820. [3] E.U. Franck, Fluids at
high pressures and temperatures, J. Chem. Thermodynamics 19 (1987)
225. [4] R.W. Shaw, T.B. Brill, A.A. Clifford, C.A. Eckert, E.U.
Franck, Supercritical watera medium for chemistry, Chem. Eng. News
69 (51) (1991) 26. [5] E.U. Franck, Supercritical water and other
uidsa historical perspective, in: E. Kiran, P.G. Debenedetti, C.J.
Peters (Eds.), Supercritical FluidsFundamentals and Applications,
NATO Science Series E, Vol. 366, Kluwer Publishers, Dodrecht, 2000,
p. 307. [6] ASME Steam Tables, 6th ed., The American Society of
Mechanical Engineers, New York, 1992 (computer program). [7] T.M.
Seward, E.U. Franck, The system hydrogenwater up to 440 C and 2500
bar pressure, Ber. Bunsenges Phys. Chem. 85 (1981) 2. [8] M.L.
Japas, E.U. Franck, High pressure phase equilibria and PVT-data of
the wateroxygen system including waterair to 673 K and 250 MPa,
Ber. Bunsenges Phys. Chem. 89 (1985) 1268. [9] K. Brllos, K. Peter,
G.M. Schneider, Fluide Mischsysteme unter hohem Druck.
Phasengleichgewichte und kritische Erscheinungen in den binren
systemen cyclohexanH2 O, n-heptanH2 O, biphenylH2 O und benzol-D2 O
bis 420 C und 3000 bar, Ber. Bunsenges Phys. Chem. 74 (1970) 682.
[10] Z. Alwani, G. Schneider, Druckeinu auf die Entmischung ssiger
Systeme. VI. Phasengleichgewichte und kritische Erscheinungen im
System BenzolH2 O zwischen 250 und 368 C bis 3700 bar, Ber.
Bunsenges Phys. Chem. 71 (1967) 633. [11] J.F. Connolly, Solubility
of hydrocarbons in water near the critical temperature, J. Chem.
Eng. Data 11 (1966) 13. [12] G.W. Morey, The solubility of solids
in gases, Econ. Geol. 62 (1957) 225. [13] O.I. Martynova,
Solubility of inorganic compounds in subcritical and supercritical
water, in: D. Jones, G. de, J. Slater, R.W. Staehle (Eds.), NACE-4:
High Temperature High Pressure Electrochemistry in Aqueous
Solutions, NACE National Association of Corrosion Engineers,
Houston, TX, 1976, p. 131. [14] K. Arai, T. Adschiri, Importance of
phase equilibria for understanding supercritical uid environments,
Fluid Phase Equilibria 158160 (1999) 673. [15] K.P. Johnston, P.J.
Rossky, Solution chemistry in supercritical water: spectroscopy and
simulation, in: E. Kiran, P.G. Debenedetti, C.J. Peters (Eds.),
Supercritical FluidsFundamentals and Applications, NATO Science
Series E, Vol. 366, Kluwer Publishers, Dodrecht, 2000, p. 323. [16]
A.A. Chialvo, P.T. Cummings, Molecular simulation and modeling of
supercritical water and aqueous solutions, in: [17] [18]
[19]
[20]
[21] [22] [23] [24] [25]
[26]
[27] [28]
[29] [30]
[31]
[32]
[33]
[34]
P. Kritzer / J. of Supercritical Fluids 29 (2004) 129 coiled
reactor, in: Proceedings of the 6th Meeting on Supercritical
Fluids, Nottingham, UK, Institute National Polytechnique De
Lorraine (INPL), Vandoeuvre Cedex, France, 1999, p. 471. A. Gidner,
L.B. Stenmark, K.M. Carlsson, Treatment of different wastes by
supercritical water oxidation, Proceedings of the 20th IT3
Conference, 2001, Philadelphia, PA. R.L. Smith Jr, P. Atmaji, Y.
Hakuta, M. Kawaguchi, T. Adschiri, K. Arai, Recovery of metals from
simulated high-level liquid waste with hydrothermal
crystallization, J. Supercritical Fluids 11 (1997) 103. H.
Schmieder, J. Abeln, N. Boukis, N. Dinjus, E. Kruse, A. Kluth, M.
Petrich, G. Sadri, M. Schacht, Hydrothermal gasication of biomass
and organic wastes, J. Supercritical Fluids 17 (2000) 145. H.R.
Appell, Y.C. Fu, S. Friedman, P.M. Yavorsky, I. Wender, Converting
Organic Wastes to Oil. A Replenishable Energy Source, Report of
Investigations 7560, Bureau of Mines, Washington, DC, 1971, 20 pp.
X. Xu, Y. Matsumura, J. Stenberg, M.J. Antal Jr, Carboncatalyzed
gasication of organic feedstocks in supercritical water, Ind. Eng.
Chem. Res. 35 (1996) 2522. T. Minowa, F. Zhen, T. Ogi, Cellulose
decomposition in hot-compressed water with alkali or nickel
catalyst, J. Supercritical Fluids 13 (1998) 253. M. Sasaki, B.
Kabyemela, R. Malaluan, S. Hirose, N. Takeda, T. Adschiri, K. Arai,
Cellulose hydrolysis in subcritical and supercritical water, J.
Supercritical Fluids 13 (1998) 261. S. Hayashi, K. Nomaguchi, T.
Okusawa, O. Yokomizo, Y. Ishagaki, H. Ishimaru, An energy balance
study of a newly developed recycling system for waste plastics,
Papers of Ship Res. Inst. 35 (1998) 133. M. Poliakoff, J.A. Darr,
T. Ikehans, A. Cabanas, The synthesis of crystalline metal oxide
particles in sub- and supercritical water mixtures, in: Proceedings
of the 6th Meeting on Supercritical Fluids, Nottingham, UK,
Institute National Polytechnique De Lorraine (INPL), Vandoeuvre
Cedex, France, 1999, p. 483. M.M. Hoffmann, J.S. Young, J.L.
Fulton, Unusual dysprosium ceramic nano-ber growth in a
supercritical aqueous solution, J. Mater. Sci. 35 (2000) 4177. E.
Tani, M. Yoshimura, S. Somiya, Formation of ultrane tetragonal ZrO2
powder under hydrothermal conditions, J. Am. Ceram. Soc. 65 (1983)
11. D.W. Matson, R.D. Smith, Supercritical uid technologies for
ceramic-processing applications, J. Am. Ceram. Soc. 72 (1989) 871.
G. Okamoto, Passive lm of 18-8 stainless steel structure and its
function, Corrosion Sci. 13 (1973) 471. G. Bellanger, J.J. Rameau,
Behaviour of hastelloy C22 steel in sulphate solutions at pH 3 and
low temperatures, J. Mater. Sci. 31 (1996) 2097. P.E. Manning, D.J.
Duquette, The effect of temperature (25289 C) on pit initiation in
single phase and Duplex 304L stainless steels in 100 ppm Cl
Solution, Corrosion Sci. 20 (1980) 597.
23
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47] [48]
[49]
[50] J.H. Wang, C.C. Su, Z. Szklarska-Smialowska, Effects of Cl
concentration and temperature on pitting of AISI stainless steel,
Corrosion 44 (1988) 732. [51] R.M. Carranza, M.G. Alvarez, The
effect of temperature on the passive lm properties and pitting
behaviour of a FeCrNi alloy, Corrosion Sci. 38 (1996) 909. [52] T.
Fujii, Electrochemical study on the corrosion behavior of metals
and alloys in aqueous solutions at high temperature and high
pressure, Trans. Nat. Res. Inst. Metals 18 (1976) 101. [53] A.M.
Olmedo, M. Villegas, M.G. Alvarez, Corrosion behaviour of alloy 800
in high-temperature aqueous solutions: electrochemical studies, J.
Nucl. Mater. 229 (1996) 102. [54] J. Hickling, N. Wieling,
Electrochemical investigations of the resistance of Inconel 600,
Incoloy 800, and type 347 stainless steel to pitting corrosion in
faulted PWR secondary water at 150250 C, Corrosion 37 (1981) 147.
[55] L.F. Lin, G. Cragnolino, Z. Szklarska-Smialowska, D.D.
Macdonald, Stress corrosion cracking of sensitized type 304
stainless steel in high temperature chloride solutions, Corrosion
37 (1981) 616. [56] W.F. Bogaerts, A.A. Van Haute, M.J. Brabers,
Relative critical potentials for pitting corrosion of 304 stainless
steel, Incoloy 800 and Inconel 600 in alkaline high-temperature
aqueous solution