HIGH-TEMPERATURE CL-INDUCED CORROSION

In this part of the course, the main focus is on Cl-induced corrosion, caused by
accumulative sulfation of metal chlorides deposited in the inner layer of deposits, as this
has proved to be a major corrosion mechanism in biomass- and waste-fired boilers
[Nielsen et al., 2000].

A substantial amount of work, has been conducted in order to address problems in

relation to corrosion in systems fired fully or partly (i.e. co-fired) with straw, both in the
laboratory, under rather well-controlled conditions, but also in full-scale boilers,
applying either cooled deposition/corrosion probes, or, cutting out sections of real tubes
or test sections of superheaters.

1. Introduction to High-Temperature Cl-Induced Corrosion

The basic causes of reduction in the boiler tube lifetime in waste and biomass combustors
are well known in general terms; being the presence of chlorine, alkali, and heavy metals
(Cu, Sn, Pb, and Zn) in the fuel, which combine to form low-melting, highly corrosive
deposits on the tubes, as well as the heterogeneous chemical and physical nature of waste
and biomass, which makes it difficult to handle, and causes combustion conditions that
are often very hard to control [Nielsen et al., 2000].

Tube failures typically occur on the upstream side of the final stage superheater, and near
the locations of the soot blowers. Maximum tube thinning is located on the crown of the
tube facing the gas flow. The deposit layers in these zones consist typically of mainly
alkali and alkaline-earth sulfates (very low [Cl]), a scale of iron oxide, and an innermost,
often laminated layer rich in metal chlorides [Wright and Krause, 1996].

The sequence of events, in the corrosion process involving such deposits, has been
postulated to be the following [Miller et al., 1972a,b];

• Potassium (and other alkali and alkaline-earth), chloride, and, sulphur dioxide
in the deposit, react with oxygen, producing K2SO4 and HCl or chlorine (Cl2).
• The liberated HCl/Cl2 reacts with the underlaying alloy to produce ferrous
chloride, which can further react with oxygen, to form iron oxide, and free
chlorine, serving to perpetuate the corrosion process [Wright and Krause,
1996].
[Otsuka and Kudo, 1992] used simulated deposits to investigate the relative effects of
chlorides and sulfates on the corrosion of austenitic steels and Ni-base alloys. They found
that high-Ni alloys such as 825, G3, 625, 690, 600, and pure Ni, exhibited good corrosion
resistance to Cl-based deposits at 400 °C, whereas Fe-based alloys such as 304 and 316L
were severely corroded. In contrast, predominantly for sulphate-rich deposits
(containing much smaller amounts of Cl), the corrosion behavior at 550 °C, appeared to
be a function of the alloy Cr-content, with the higher-Cr alloys corroding slowest. For
alloys of similar Cr-content, increasing Ni-content appeared to degrade the resistance to
corrosion. The best-performing alloy in all the environments tested, was Alloy 625
[Otsuka and Kudo, 1992].

Straw contain significant amounts of K and Cl, which is being present as KCl in parts of
the combustion system. Significant levels of KCl has been detected in fly ash, but
particularly in inner layers of deposits formed in straw-fired plants. The flue gas will
contain a certain level of SO2(g) originating from the combustion of straw. It has been
suggested that corrosion of superheater tubes is driven by sulfation of alkali chloride to
alkali sulfates, eg. KCl to K2SO4, causing a relatively high partial pressure of chlorine,
close to the metal (heat transfer) surface [Krause, 1989; Reid, 1971; Alexander, 1963]. Thus,
the initiation reaction in the presence of KCl-rich deposits, may be one of the below
reactions, depending on the presence of SO2:

2 KCl + O2 + SO2  K2SO4 + Cl2 (1)

2 KCl + SO2 + H2O  K2SO4 + 2 HCl (2)

KCl + Fe2O3 + ½ O2  K2Fe2O4 + Cl2 (3)

This role of SO2 was demonstrated in laboratory experiments [Miller et al., 1972a,b], in
which alloys were exposed to sulphate-chloride salt mixtures, in a synthetic flue gas with
and without SO2. Introduction of only 250 ppmv SO2, a level very common in waste-fired
boilers, caused large increases in the corrosion rates. Increasing the SO2–concentration to
2500 ppmv, a level not uncommon for coal-fired boilers, caused a significant increase on
the attack of C-steel, but had only minor effect on T11 [Miller et al., 1972a,b]. Type 321
stainless steel underwent slightly less corrosion, at the higher [SO2]-level. The effect of
increasing temperatures was also investigated, increasing the temperature from 427 to
538 °C caused a much higher increase in the corrosion rate, than an increase in the amount
of NaCl in the deposit did. These results indicate that only small amounts of alkali
chloride in a superheater deposit may cause serious corrosion. It is important to
emphasize that without a deposit on the metal, the gas phase attack by the synthetic flue
gas containing 250 ppmv SO2 was slight, showing that formation of HCl in the deposit, is
the critical factor [Miller et al., 1972a,b].

Chlorine is known to be able to penetrate, and destroy, the oxide scale, probably through
cracks and pores, and at the metal/scale interface react with Cr and Fe, thereby forming
volatile metal chlorides. A detailed study of the disruption of the oxide scale on steel by
HCl in flue gas, was reported by [Mayer et al., 1980]. Experiments were conducted at a
metal temperature of 538 °C, and a flue gas temperature of 1060 °C. Without HCl in the
flue gas, a continuous non-porous scale consisting of FeO, Fe2O3, and Fe3O4 was formed.
With 0.1 %(v/v) HCl present in the gas, the inner scale layers were continuous, but the
outer Fe2O3-layer became blistered and cracked. This Fe2O3-layer in turn became porous
and discontinuous when [HCl] in the gas was increased to 0.2 %(v/v) [Mayer et al., 1980].
A further increase of [HCl] to 0.8 %(v/v) caused complete disintegration of the Fe 2O3-
layer, and the Fe3O4-layer became irregular and porous. The maximum [HCl] that was
investigated was 2 %(v/v), a level which destroyed the continuity of both outer layers,
thereby exposing the underlying metal to attack by oxygen and HCl [Mayer et al., 1980].

A mechanism of scale disruption proposed by [Crolet, 1993a,b], involves direct reaction,

between the metal and the Cl, thereby producing a volatile oxychloride. Both eqn (1) and
eqn (3) above, causes release of Cl2-gas which may then migrate through the oxide, to
react with the metal at the metal-oxide scale interface, according to eqn (4);

Cr + Cl2  CrCl2 (4)

The CrCl2 formed migrates out of the metal due to its high volatility, to areas with a
higher oxygen partial pressure, where it is converted to Cr2O3 and Cl2 (eqn (5)).

2 CrCl2 + 1½ O2  Cr2O3 + 2 Cl2 (5)

The precipitated Cr2O3 is non-protective, and is therefore susceptible to spallation.

Chlorine then migrates into the corrosion front, to form metal chlorides again. A similar
reaction can occur for Fe, forming FeCl3, if there is no Cr available for reaction at the
metal-scale interface. Iron chloride, FeCl3, will migrate more readily from the corrosion
product, as it is more volatile than chromium chloride (CrCl2) and is expected to form an
oxide on top of the Cr2O3, at higher partial pressures of oxygen. Thus, at lower
temperatures which do not encourage evaporation of CrCl3, FeCl3 will still evaporate
from the corrosion front [Montgomery et al., 2000a,b].

Temperature also has an important effect on the corrosion in combustors. Early tests
showed that both the metal, and the gas temperature, influenced the corrosion rates. At
gas temperatures of 460 and 593 °C, the corrosion rates and the temperature-dependence
of corrosion were significantly lower than at the higher gas temperatures, an effect
especially pronounced at metal temperatures above 371 °C [Vaughan et al., 1974, 1977].

Laboratory studies of simulated deposits [Otsuka and Kudo, 1992] have shown that, for
both sulphate- and chlorine-based deposits, corrosion was only slight to negligible, at
temperatures below the melting point of the deposit. As soon as constituents of the
deposit melted, rapid corrosion was observed. Also, in the absence of a deposit, corrosion
from a simulated flue gas containing 1500 ppmv HCl was minimal, at temperatures up
to 550 °C [Otsuka and Kudo, 1992].

Several other corrosion mechanisms, may be responsible for a corrosive attack of the
nature observed in grate-fired units;
 [Spiegel and Grabke, 1991] suggests that chlorine may be liberated close to the
metal surface, by NaCl reacting with the oxide scale, forming Na2CrO4 and
Na2Fe2O4;
 [Fujikawa and Maruyama, 1986] and [Hiramutsu et al., 1989] proposed that NaCl
reacts with chromium carbide, preferably found at the grain boundaries, in order
to form Na2CrO4 and Cl2, and, finally;
 [Shinata et al., 1986, 1987] studied the effect of NaCl on chromium, Cr-Ni alloys,
and stainless steels. They suggested a mechanism where NaCl reacts with Cr 2O3
to form Na2CrO4 in an initiation step, followed by the formation of a eutectic melt
between NaCl and Na2CrO4.

2. Danish Corrosion Studies

Several field investigations of corrosion have been conducted at straw-fired CHP-boilers

in Denmark, over the past two decades, providing basic data on corrosion rates of
different materials at different temperatures, and, exposure times [Larsen and Henriksen,
1996; Henriksen and Larsen, 1997; Montgomery and Karlsson, 1999; Montgomery et al.,
2002a,b; Nielsen, 1998; Frandsen et al., 2006].

Below is an outline of Danish experiences with and investigations on corrosion, in

thermal fuel conversion systems fired solely, or partly with straw.

[Nielsen, 1998] showed in a lab-scale study, that metal samples covered by KCl-crystals,
and exposed, isothermally at 550 °C to an SO2-laden gas, after 1 month had a thin layer
(5 – 20 µm) of K2SO4 with iron threads included, formed close to the oxide scale. This
layer resembled the structure of the inner layer found in mature deposits from the
Masnedø CHP-boiler. For samples having been exposed only 1 week in the oven, the
formation of the characteristic K2SO4-layer had barely started, and at this point, small
KCl-inclusions were still present in the layer [Nielsen, 1998].

The following mechanism for the formation of a K2SO4-layer in the inner deposit was
proposed [Nielsen, 1998]:

1. Initially, KCl-rich deposits on the superheater surface, e.g. by inverse sublimation

of gaseous KCl, KCl(g)  KCl(s). The evidence of such a layer has been reported
e.g. at the Maribo-Sakskøbing straw-fired CHP-boiler.
2. KCl reacts with FexOy, FeCl2, or K2SO4, to form a molten phase. This reaction is
likely to happen, as KCl forms eutectic mixtures with FeCl2 (at 355 °C), with K2SO4
(at 690 °C), and possibly also with FexOy, even though the latter has not been
verified.
3. The molten KCl reacts with gaseous SO3, and forms K2SO4. Gaseous SO3 may be
present, since FexOy is known for catalysing the oxidation of SO2 to SO3 [Harlow,
1944, 1949; Reid, 1971].

Thus, potassium chloride, KCl, forms a melt with K2SO4, and different iron components
(FeXOY, FeCl2), adjacent to the oxide scale. The sulfation reaction occurs rapidly in the
melt, releasing HCl, or Cl2, close to the metal surface. The chlorine gas has the ability to
diffuse through the oxide scale, and react with Fe forming volatile iron chlorides (FeCl2,
FeCl3) [Nielsen, 1998]. The iron chlorides are thermodynamically stable, at the interface
between metal and metal oxide, where a very low partial pressure of oxygen exist. They
then diffuse out through the scale, and oxidize as they reach areas with higher partial
pressures of oxygen. Iron oxides will be formed either on top of the metal oxide layer or
in the dense mixed layer [Nielsen, 1998].

By this mechanism, optimal conditions for the sulfation of KCl are present, close to the
Fe-rich oxide scale, which may increase the rate of SO3-formation, catalytically. The
hypothesis is consistent with the compact structure of the K 2SO4-layer, as compared to
the porous KCl-layer, located just on top of it. It is also consistent with the presence of Fe-
oxides in the K2SO4-layer: The iron oxide threads reach to the very outer edge of the
K2SO4-layer, while no iron oxide threads are seen, traversing through the KCl, at any
point in the mature deposits. This strongly indicates that the K 2SO4-formation is related
to the presence of, and close contact with, FexOy [Nielsen, 1998].

This mechanism can explain the shift in corrosion behaviour with temperature which has
been observed in full-scale corrosion tests [Larsen & Henriksen, 1996, Montgomery and
Larsen, 2002].
 Total corrosion Internal corrosion Material loss
Corrosion rate mm/1000 hs. 0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 5 10 15 20 25 30 35
Chromium content in wt. %

Figure 1: Corrosion rate with various chromium contents in the alloys at approx. 580°C.
Source: [Montgomery and Larsen, 2002].

2.1. Effect of Alloy Composition (Cr-Content)

In conventional coal-firing, alloys with high Cr-content causes lower corrosion rates,
however the contrary is the case for straw-firing plants. Results from three separate
investigations at straw-fired plants [Larsen et al., 1996; Henriksen et al., 1995;
Montgomery et al., 2000a], testing a variety of steels, revealed that at high temperatures,
high-Cr steels were rapidly attacked. Figure 1 show data on total corrosion rate, as a
function of the Cr-content in the metal, for a range of alloys from a 2½% Cr, to a 30% Cr
steel [Montgomery et al., 2000a]. The low-Cr steels exhibit a high material loss whereas
the high-Cr steels are characterised by severe internal attack, in the form of chromium
depletion. These results indicate a minimum the in total corrosion rate of steels, at an
intermediate Cr-content of 12-18% Cr [Montgomery and Karlsson, 1999].

2.2. Effect of Metal Surface Temperature

Selective removal of Cr and Fe from an alloy, causes a porous metal structure, and often
also grain boundary attack. For austenitic steels, grain-boundary corrosion is a precursor
to corrosion within the grains. The corrosion products formed are Cr oxide [Cr 2O3] and
iron oxides [Fe2O3, Fe3O4]; being present at the surface of the specimen. In many cases, Ni
is also present in the outer corrosion product, as non-reacted nickel [Montgomery and
Larsen, 2002a,b].
 1.2
Corrosion rate in mm/1000h X3CrNiMoN17-13
TP347CG
1
TP347FG
12Cr1Mo1W
0.8
13Cr13Mn
Esshete 1250
0.6

0.4

0.2

0
450 500 550 600
Calculated metal temperature in C

Figure 2: Corrosion rate of specimens in the test superheater section of the Masnedø
straw-fired CHP-plant relating metal temperature to corrosion rate. Source:
[Montgomery and Larsen, 2002a,b].

In Figure 2, the corrosion rate of 12-18% Cr steels is plotted against temperature. It is

clearly visible that the rate of corrosion is increasing with increasing temperature of the
metal surface[Montgomery et al., 2002a,b].

For the austenitic alloys investigated, a gradual change in the corrosion morphology, as
a function of temperature, was observed. Corrosion rates were modest at 460°C, where a
protective Cr-rich Fe oxide scale, was formed. When increasing the temperatures of up to
525°C, the corrosion rates become more linear, with respect to temperature. At around
500-520C, both grain boundary attack and protective oxide scales has been observed for
the same specimen. Above this temperature, Cr-depletion and grain boundary attack are
always observed [Montgomery et al., 2002a,b].

2.3. Why Corrosion Rates are Lower in Co-Fired Systems

Co-firing of coal and straw, in a suspension-fired boiler, at straw shares of 10-20 % on a

thermal base, removes the danger of catastrophic Cl-induced corrosion, as was observed at
MKS1 [Andersen, 1998, Wieck-Hansen et al., 2000]. The corrosion rates observed in co-firing
systems are much lower than those having been observed in straw-fired units, and are closer
to the corrosion rates observed in coal-fired power plants. For comparative purposes, the
corrosion rates for TP347H FG have been plottet in an Arrhenius-type plot, see Figure 3,
assuming a parabolic corrosion kinetics [Montgomery and Larsen, 2002]:

x2  2 kp  t

where x is the thickness of the corroded material, and, t is the duration of time of exposure.

Figure 3: Arrhenius plot

showing that corrosion of
TP347H FG in co-firing
compared to pure coal- respec-
tively straw-firing. Source:
[Montgomery and Larsen,
2002].

Parabolic kinetics is assumed for all three types of corrosion, although it is apparent from
previous work, that linear kinetics is more applicable for corrosion in 100 % straw-fired
boilers, at high temperatures [Montgomery and Karlsson, 1999]. It is clear from Figure 3,
that the corrosion is significantly lower in co-fired plants as when compared to ‘pure’ straw-
firing [Montgomery and Larsen, 2002].

The most important outcome of the MKS1 corrosion investigation was that there was no
sign of Cl found in the deposits, within the corrosion products, or, at the metal-oxide scale
interface [Wieck-Hansen et al., 2000]. This shows that KCl released from the straw, reacts
with Al and Si in the coal ash, forming non-corrosive K-Al silicates and HCl in the gas phase
[Andersen, 1998]:

2 KCl + Al2O3 + 6 SiO2 + H2O  K2O.Al2O3.6SiO2 + 2 HCl (5)

2 KCl + SO2 + H2O + ½ O2  K2SO4 + 2 HCl (6)

Formation of HCl(g) does not pose nearly as great a corrosive potential as deposition of KCl,
followed by subsequent sulfation at the metal-oxide scale surface. An increase in the straw-
fraction fired from 10 % to 20 % on a thermal base, caused a higher local concentration of
KCl in the combustion chamber, which is converted to more K2SO4 (eqn. (6), above), causing
the HCl emissions from the stack to increase [Wieck-Hansen et al., 2000].

The intensive corrosion investigations conducted co-fired plants, show that the corrosion
caused by co-firing of 10 % straw is low in general, and that it is not, or only slightly,
increased compared to coal-firing.

The corrosion may, though, be dependent on the coal type. Especially, coal types with a
considerable content of water soluble alkali, may provide a higher risk of increased
corrosion rates, than observed in the tests at MKS1.

Based on the corrosion studies carried out at MKS1 and SSV4, it was concluded that there
are no corrosion related problems when co-firing 10 % straw (thermal base) at standard
steam temperature of 540 C. Furthermore, the results do not show any considerable
corrosion risks caused by introduction of co-firing of 10 % straw at plants with elevated
steam temperatures – up to 580 C [Wieck-Hansen et al., 2000; Frandsen et al., 2006].

3. Lab-Scale High-Temperature Corrosion Study

Deposit-induced Cl-corrosion was further studied under well-controlled laboratory

conditions, simulating the conditions in straw-fired boilers as well as boilers co-firing coal
and straw by [van Lith et al., 2009]. This was done by exposing pieces of superheater tube
(TP 347H FG) covered with synthetic deposits of known Cl-content to gas mixtures
simulating straw-firing and co-firing of coal and straw, at 560 °C, for 3 days.

The effect of the SO2 concentration in the flue gas is presented in Figure 4. For samples
exposed to the cofiring gas composition, the degree of sulfation in the deposits, was found
to be considerably higher than that observed for the samples exposed to the straw-firing
gas composition. For example, for specimens covered with a 25 wt % KCl deposit, the
KCl-particles in the sample exposed to straw-firing gas composition are only partly
sulfated, whereas the deposit is completely sulfated, in the case of the sample exposed to
the cofiring gas composition. The higher degree of sulfation can be explained by the
higher concentration of SO2 in the co-firing flue gas. The SEM study of the deposits after
exposure shows that the sulfation of the KCl proceeds according to the “unreacted
shrinking core model”, see Figure 5 [van Lith et al., 2009].

The corrosion attack was found to be quite different for samples exposed to the straw-
firing gas composition, compared to the co-firing gas composition.
 Exp. 1, 7, and 8, 100 % KCl Exp. 1, 7, and 8, 100 % KCl

Corrosion rate (mm/1000h)

Corrosion rate (mm/1000h)
60 0.8 60 0.8
50 50

Mixed layer (m m)
Oxide layer (m m)

0.6 0.6
40 40
30 0.4 30 0.4
20 20
0.2 0.2
10 10
0 0.0 0 0.0
straw-firing co-firing co-firing straw-firing co-firing co-firing
(60 ppm) low SO2 (500 ppm) (60 ppm) low SO2 (500 ppm)
(200 ppm) (200 ppm)

a) Oxide layer thickness b) Mixed layer thickness

Figure 4: Effect of the SO2 concentration in the gas flow (given in brackets) on the
corrosion attack, for specimens covered with a 100% KCl deposit, exposed at 560 C, for
3 days.

a) Gas composition: straw-firing b) Gas composition: cofiring

Large bright area: KCl; smaller grey particles: Small grey particles: SiO2; Light grey areas /
SiO2; small light grey spots: K2SO4 rims: K2SO4

Figure 5: SEM micrographs of cross sections of the top of initially 100% KCl deposits, after
exposure at 560 C, for 3 days. The metal surface is visible in the bottom of the pictures.

As can be seen for specimens covered with a 100% KCl deposit, the sample exposed to
the straw-firing gas composition shows an almost continuous layer of internal attack,
followed by a rather uniform oxide and mixed layer (see Figure 6a), whereas the sample
exposed to the co-firing gas composition, shows very irregular and localized internal
attack and non-uniform thicknesses of the oxide scale and mixed layer (see Figure 6b).
Furthermore, very little (≤1 wt %) or no Cl was detected at the corrosion front in the
samples exposed to the co-firing gas composition, whereas ~1-4 wt % Cl was detected at
the corrosion front, in the samples exposed to the straw-firing gas composition [van Lith
et al., 2009].

Samples exposed to the cofiring gas composition, with a reduced SO2 concentration,
clearly showed a lower degree of sulfation of KCl, in the deposit. For example, for the
specimens covered with a 100% KCl deposit, it can be seen that the top particles of the
deposit are completely sulfated, when exposed to the cofiring gas composition (gray
areas in Figure 7a), whereas only rims of the top particles are sulfated, when exposed to
the cofiring gas composition, with a reduced SO2 concentration (gray rims in Figure 7b),
resulting in an overall degree of sulfation of around 20% according to the ICP-OES
analysis. For the samples with a 25 wt % KCl deposit, complete sulfation was observed
when exposed to the co-firing gas composition, while some KCl was left in the deposits
of the samples, exposed to the co-firing gas composition with reduced SO2 concentration
(around 93% sulfation according to the ICP-OES analysis) [van Lith et al., 2009].

KCl particle
epox KCl particle
y

mixed
layer K2SO4
oxide
layer
epox epoxy
y
selective corrosion mixed layer

selective corrosion oxide layer

sound metal sound metal

a) Gas composition: straw-firing b) Gas composition: cofiring

Figure 6: SEM micrographs of cross sections of corrosion products of specimens

covered with a 100% KCl-deposit, after exposure at 560 C, for 3 days. Note: Pictures
have different scales.

The thickness of the oxide scale of the sample exposed to the co-firing gas composition,
with reduced SO2 concentration, was smaller than for the sample exposed to the straw-
firing gas compositions, but was - on average - higher than the sample exposed to co-
firing gas composition. However, the mixed layer was noticeably thinner than for the
samples exposed to the straw-firing and cofiring conditions (see Figure 6b), indicating
that there is no clear relationship between the thickness of the mixed layer and the SO 2
concentration in the flue gas, and that the concentration of HCl in the flue gas possibly
also plays a role [van Lith et al., 2009].

Besides the gas compositions, various parameters were studied systematically. Most
specimens suffered some internal attack, mostly by selective corrosion, but in some cases
also by grain boundary attack. In all experiments with KCl and KCl-SiO2 deposits, the
corrosion products consisted of an oxide scale, containing oxides of Cr and Fe, and on
top of that a characteristic mixed layer of iron oxide threads in a K2SO4-matrix. However,
the thickness and shape of this layer was found to be strongly dependent on the
experimental conditions. An increase of the percentage of KCl in the deposit resulted in
a more uniform and deeper internal corrosion attack. The presence of HCl in the flue gas
did not seem to be essential for Cl-induced corrosion to occur, when a deposit containing
KCl was present, but it enhanced the corrosion rate.

K2SO4

KCl
K2SO4

KCl

a) Gas composition: cofiring b) Gas composition: cofiring, low SO2

Figure 7: SEM micrographs of cross sections of the top of initially 100% KCl deposits,
after exposure at 560 C, for 3 days.

The degree of sulfation of KCl in the deposits after exposure was quantified by wet
chemical analysis and was shown to be dependent on many parameters, including the
SO2-concentration in the gas flow, the concentration of KCl in the deposit, and the SiO2
and KCl particle sizes in the deposit. Whereas SiO2-particles were found to be chemically
inert, with respect to the flue gas and the corrosion attack, CaO-particles reacted with
HCl from the flue gas, and the resulting CaCl2 played an important role in the corrosion
mechanism. As a result, the corrosion rate was strongly enhanced when CaO was present
in the deposit, instead of SiO2 [van Lith et al., 2009].

4. Corrosion Studies and Material Aspects in Biodust-Fired Systems

After years optimizing dedicated biomass-fired CHP-plants, typically based on grate-

firing technology, and coal-straw co-firing in suspension, it was decided to focus
attention of biodust-firing, e.g. firing of biomass pellets in burners in PF-fired units. This
called for red light, since thermal treatment of biomass, at the relatively high
temperatures utilized in burners, i.e. > 1500 C, indicated potential corrosion problems,
due to excessive release of K, Cl and S at such high temperatures.

In order to understand the mechanisms of corrosion attack in biodust-fired plants, a

comprehensive analysis of corrosion products was conducted by [Okoro, 2016]. Two
complementary methodologies based on analysis of cross sections and plan views were
applied to achieve comprehensive characterization of corrosion products. The suitability
of these methods for both lab- and full-scale corrosion investigations, was demonstrated
by the combined use of complementary information from microscopy, energy dispersive
X-ray spectroscopy and various X-ray diffraction characterization techniques [Okoro,
2016].

In light of the wide variation in operating conditions in biomass fired power plants,
systematic and well-controlled, but realistic lab-scale exposures were carried out in order
to understand the effect of process parameters, such as gas phase chemistry, exposure
time and temperature variations, on the corrosion process. By conducting corrosion
exposures under: oxidizing, oxidizing-chlorinating, oxidizing-sulphidizing and
oxidizing-chlorinating-sulphidizing gas mixtures, it was possible to observe the
contributions of chlorination, sulphation and sulphidation on the complex corrosion
mechanisms. Comparative exposure of KCl-coated and KCl-free samples, under each of
the gas atmospheres, revealed that the corrosion attack due to the presence of a deposit,
overrules the effect of the gas phase chemistry. Further, using the oxidizing-chlorinating-
sulphidizing gas mixture corresponding to straw-firing conditions, the evolution of
corrosion with time and temperature variation was studied [Okoro, 2016].

For the investigated austenitic stainless steels, FeCrAl alloys and Ni-based superalloy,
the formation of a protective oxide layer that suppressed corrosion attack was not
observed. However, it was established that among the alloying elements present in these
alloys, Ni exhibits a relatively greater resistance to corrosion attack. Surface modification
approaches with the aim to form resistant oxides or coatings on superheater surfaces,
were also evaluated. This included preoxidation to obtain Al2O3 and Ti-Cr-rich oxides, as
well as formation of SiOx coating, Ni electroplating and NiAl coatings on commercial
alloys. For most of the modified surfaces, in particular α-Al2O3 obtained by pre-oxidation,
poor mechanical integrity of the oxide layer nullifies its otherwise excellent chemical
integrity, towards corrosion attack. Nonetheless, corrosion attack on most of the modified
surfaces was substantially lower than attack on the virgin alloys. Thus, with further
optimization, these approaches may provide alternative corrosion resistant materials for
application in biomass fired power plants [Okoro, 2016].

The high-temperature corrosion experienced by superheaters is a very significant

challenges for biodust-firing in power plants, where the maximum steam temperature is
limited by the corrosion. Application of protective coatings on superheater tubes is a
possible solution to combat fireside corrosion. Therefore, [Duo, 2018] investigated the
corrosion performance of promising coatings in real boilers. The coatings were based on
Ni and Al, where Ni was electroplated followed by Al pack cementation. The pack
cementation was adapted so that it could be undertaken at 650ºC for the integrity of the
coated austenitic stainless steel tubes [Duo, 2018].

The resulting Ni2Al3 coating was first tested under laboratory conditions relevant to
biomass combustion, resulting in protective behavior in most areas, and, localized attack
with selective removal of Al. Porosity developed at the interface between the Ni-Al
coating and the Ni layer and this zone gradually expanded with time due to isothermal
interdiffusion, but did not lead to spallation [Duo, 2018].

The Ni2Al3- and Ni-coatings were exposed in two different power plants for one year: (a)
a straw-fired plant with an outlet steam temperature of 540 C (MSK) and (b) a wood-
fired plant with an outlet steam temperature of 520 C (Randers). The Ni coatings showed
good corrosion resistance at the Randers plant, while at the MSK plant the pure nickel
layer was no longer present, after exposure. At the Randers plant, the Ni2Al3-coatings
showed protective behavior in most areas and corrosion attack only occurred in local
areas, resembling the behavior observed in the laboratory exposures. At the MSK plant,
the Ni2Al3 coatings spalled off and were only found as remnants in the deposit. The
failure of the coatings was with high certainty linked to the frequent thermal cycling in
the MSK plant, as well as the higher temperature as compared to the Randers plant [Duo,
2018].

The boiler exposure was extended from 1 year (2015-2016) to 2 years (2015-2017) at
Randers, however the operation temperatures were slightly higher in the 2 nd year (2016-
2017). The Ni-coatings were more heavily attacked after the second year of exposure,
compared to the first year exposure. After the first year, the Ni2Al3 coatings were still
present in most areas with only localized attack. After the second year, no areas with
intact Ni2Al3-coating, were observed. The Ni2Al3-coatings were attacked chemically,
resulting in Al-depletion and ultimately in spallation of the completely corroded coating
[Duo, 2018].

Preoxidized and CeO2-dispersed Ni2Al3-coatings showed similar corrosion behavior as

the normal Ni2Al3-coatings after exposure at the Randers plant for one year (2016-2017).
A nitrided Fe-Al coating slowed down the corrosion attack in some areas and a Fe-rich
oxide formed on the outer surface. In other areas, grain boundary attack of the steel and
selective attack of Cr was observed, similar to corrosion observed on uncoated tubes.

5. Learning Objectives

The module on Deposit Build-Up, Consolidation and Shedding, have the following
overall learning objectives;

 Why should we worry about chemical attacks/corrosion in boilers ?

 How does corrosion happen in straw-fired plants ?
 Lab-scale corrosion study: What is the effect of major oxides like CaO and SiO 2,
[SO2(g)] and [HCl(g)] on corrosion underneath deposits ?
 How does corrosion happen in waste-fired plants ?
 How does corrosion happen in biodust-fired systems ?

References:

Alexander, P.A.; Laboratory Studies of the Effects of Sulphates and Chlorides on the Oxidation
of Superheater Alloys; In: The Mechanism of Corrosion by Fuel Impurities, Eds. R.
Johnson and D. J. Littler, Butterworths, 1963, pp 571-582.
Andersen, K.H.; Deposit Formation during Coal-Straw Co-Combustion in a Utility PF-Boiler;
Ph.D.-Thesis, Department of Chemical Engineering, Technical University of
Denmark, ISBN-87-90142-43-8, 1998.
Crolet, J.L.; Mechanisms of uniform corrosion under corrosion deposits; J. Mat. Sci., 28, pp.
2589-2606, 1993a.
Crolet, J.L.; The eltrochemistry of corrosion beneath corrosion deposits; J. Mat. Sci., 28, pp.
2577- 2588, 1993b.
Frandsen, R.B., Montgomery, M.; Larsen, O.H.; Field test corrosion experiences when co-firing
straw and coal: Ten years status within Elsam; 4th International Workshop "Materials Issues
Governing the Performance of Advanced 21st Century Energy Systems" held in
Wellington, New Zealand, February 2006.
Harlow, W.F.; Causes of High Dewpoint Temperatures in Boiler Flue Gases; Proc. Inst. Mech.
Engrs., 151: 293, 1944.
Harlow, W.F.; Causes of Flue Gas Deposits and Corrosion in Modern Power Plants; Proc. Inst.
Mech. Engrs., 160: 359, 1949.
Henriksen, N., Larsen, O.H.; Corrosion in ultra supercritical boilers for straw combustion;
Materials at High Temperatures, 14 (3) pp. 227-236, 1997.
Hiramutsu, N., Uematsu, Y., Tanaka, T., Kinugasa, M.; Effect of Alloying Elements on NaCl-
Induced Hot Corrosion of Stainless Steels; Materials Science and Engineering, A120, 319-
330, 1989.
Krause, H.H.; Corrosion by Chlorine in Waste-Fuelled Boilers; In: R.W. Bryers (Ed.),
Incinerating Municipal and Industrial Waste, Hemisphere Publishing Corporation,
1989.
Larsen, O.H., Henriksen, N.; Ash Deposition and High Temperature Corrosion at Combustion
of Aggressive Fuels; Proc. Power Plant Technology, Kolding 4-6 Sept. pp 7.1-7.18., 1996.
Larsen, O.H. , Henriksen, N., Inselmann, S., Blum, R.; The Influence of Boiler Design and
Process Conditions on Fouling and Corrosion in Straw and Coal/Straw-Fired Ultra Supercritical
Power Plants; Proc. 9th European Bioenergy Conference: Copenhagen, Denmark, 1996.
van Lith, S.C., Frandsen, F.J., Montgomery, M., Wilhelmsen, T., Jensen, S.A.; Lab-Scale
Investigation of Deposit-Induced Chlorine Corrosion of Superherater Materials under Simulated
Biomass-Firing Conditions. Part 1; Exposure at 560 °C; Energy and Fuels, 23, 3457-3468,
2009.
Mayer, P., Westwood, H.J., Manolescu, A.V.; Corrosion Related Problems in Fossil Fired
Boilers; J. Materials for Energy Systems, 2, pp. 55-64, 1980.
Miller, P.D., Krause, H.H., Zupan, J., Boyd, W.K.; Corrosive Effects of Various Salt Mixtures
Under Combustion Gas Atmospheres; Corrosion-NACE, 28(6), pp. 222-225, 1972.
Miller, P.D., Krause, H.H., Vaughan, D.A., Boyd, W.K.; The Mechanism in High
Temperature Corrosion in Municipal Incinerators; Corrosion-NACE, 28(7), pp. 274-281, 1972.
Montgomery, M., Karlsson, A.; In situ Corrosion Investigation at Masnedø CHP Plant;
Materials and Corrosion, 50 579-584, 1999.
Montgomery, M., Frandsen, F.J., Karlson, A., Larsen, O.H.; Corrosion Investigation in
Straw-fired Power Plants in Denmark; Proc. United Eng. Found. Conf. on Effects of Coal
Quality on Power Plant Performance: Ash Problems, Management and Solutions, Park
City, UT, USA, May 9-13, 2000a.
Montgomery, M., Frandsen, F.J., Larsen, O.H.; The Effect of Co-Firing with Straw and Coal
on High Temperature Corrosion; Proc. United Eng. Found. Int. Conf. on ’Effects of Coal
Quality on Power Plant Management: Ash Problems, Management and Solutions’,
Park City, Utah, May 8-11, 2000b.
Montgomery, M., Karlsson, V, Larsen, O.H.; Field test corrosion experiments in Denmark
with biomass fuels Part I: Straw firing; Materials and Corrosion 53 pp. 121-131, 2002a.
Montgomery, M., Biede, O., Larsen, O.H.; High Temperature Corrosion in straw fired power
plants: Influence of steam/metal temperature on corrosion rates for TP347H; Materials for
Advanced Power Engineering 2002 Part II pp 957-966, 2002b.
Montgomery, M., Larsen, O.H.; Field test corrosion experiments in Denmark with biomass
fuels. Part 2: Co-firing of straw and coal; Materials and Corrosion, 53(3), 185-194, 2002.
Nielsen, H.P.; Deposition and High-Temperature Corrosion in Biomass-Fired Boilers; Ph.D.-
Thesis, Department of Chemical Engineering, Technical University of Denmark,
ISBN-87-90142-47-0, 1998.
Nielsen, H.P., Frandsen, F.J., Dam-Johansen, K., Baxter, L.L.; The Implications of Chlorine-
Associated Corrosion on the Operation of Biomass-Fired Boilers; Prog. Energy Comb. Sci., 26,
283-298, 2000.
Okoro, S.C.; High-Temperature Corrosion in Biodust-Firing; PhD-Thesis, Technical
University of Denmark, Department of Mechanical Engineering, Section for Materials
and Surface Engineering, 2016.
Otsuka, N., Kudo, T.; High-Temperature Corrosion of Advanced Materials and Coatings; In: Y.
Saito, N. Onay, T. Maruyama (Eds.), Elsevier Science Publishers, B.V. 1992.
Reid, W.T.; External Corrosion and Deposits - Boilers and Gas Turbines; American Elsevier,
New York, 1971.
Shinata, Y., Nishi, Y.; NaCl-Induced Accelerated Oxidation of Chromium; Oxidation of
Metals, 26, pp 201-212, 1986.
Y. Shinata, Takahashi, F., Hashiura, K.; NaCl-Induced Hot Corrosion of Stainless Steels;
Materials Science and Engineering, 87, pp 399-405, 1987.
Spiegel, M., Grabke, H.J.; High Temperature Corrosion of Low and High Alloy Steels in
Simulated Waste Incineration Atmospheres; In: Incinerating Municipal and Industrial
Waste. Fireside Problems and Prospects for Improvement edited by R. W. Bryers,
Hemisphere Publishing Corporation, pp 758-764, 1991.
Vaughan, D.A., Krause, H.H., Boyd, W.K.; Batelle Interim Report to US EPA, June 1974.
Vaughan, D.A., Krause, H.H., Boyd, W.D.; Chloride Corrosion and Its Inhibition in Refuse
Firing; Proceedings of the International Conference on Ash Deposits and Corrosion
From Impurities in Combustion Gases, Henniker, New Hampshire, June 26-July 1, pp
473-493, 1977.
Wieck-Hansen, K., Overgaard, P., Larsen, O.H.; Co-firing coal and straw in a 150 MWe power
boiler experiences; Biomass & Bioenergy, 19, 395-409, 2000.
Wright, I., Krause, H.H.; Assessment of Factors Affecting Boiler Tube Lifetime in Waste-Fired
Steam Generators: New Opportunities for Research and Technology Development; National
Renewable Energy Laboratory, Golden, Colorado, Rep. No. NREL/TP-430-21480, 1996.
Wu, D.; Corrosion Resistant Coatings for Biomass Firing; PhD-Thesis, Technical University
of Denmark, Department of Mechanical Engineering, Section for Materials and
Surface Engineering, 2018.