Вы находитесь на странице: 1из 23

1.

20 Gas Turbine Oxidation and Corrosion


J. R. Nicholls and N. J. Simms
Cranfield University, Cranfield, Bedfordshire MK43 0AL, UK

ß 2010 Elsevier B.V. All rights reserved.

1.20.1 Introduction 518


1.20.2 Gas Turbine Operating Conditions 520
1.20.3 Gas Turbine Hot Gas Path Components and Materials 522
1.20.4 Oxidation and Thermal Cycling 524
1.20.4.1 Oxide Formation and Spallation 524
1.20.4.1.1 Oxidation processes 524
1.20.4.1.2 Modeling oxide growth and spallation 526
1.20.4.2 Thermal Barrier Coating Systems 527
1.20.4.2.1 Damage mechanisms 527
1.20.4.2.2 Modeling TBC spallation 529
1.20.5 Hot Corrosion 529
1.20.5.1 Hot Corrosion Mechanisms 530
1.20.5.2 Deposit Formation and Fuel Specifications 532
1.20.5.3 Damage Locations 535
1.20.5.4 Rates of Hot Corrosion Attack and Modeling Hot Corrosion Processes 535
1.20.6 Current/Future Issues 538
1.20.7 Summary 538
References 539

Abbreviations Symbols
APS Air plasma spray y Reciprocal value of the C to H ratio in mol
CVD Chemical vapor deposition A/F Air to fuel mass flow ratio
EBPVD Electron beam physical vapor MO Metal oxide
deposition S/F Steam (water) to fuel mass flow ratio
EFC European Federation of Corrosion XA Contaminant concentration in inlet air (ppm) by
EPSRC Engineering and Physical Sciences weight
Research Council (United Kingdom) XF Contaminant concentration in fuel (ppm) by weight
EU European Union XS Contaminant concentration in injected steam or
FOD Foreign object damage water (ppm) by weight
HP High pressure XT Total contamination in fuel equivalent (ppm) by
HRSG Heat recovery steam generator weight
HVOF High velocity oxy-fired l Amount of excess air: actual mass flow divided by
IGCC Integrated gasification combined cycle the air mass flow for stoichiometric
ISO International Organization for combustion
Standardisation
LPPS Low pressure plasma spray
NGV Nozzle guide vane
1.20.1 Introduction
NMAB National Materials Advisory
Board (USA)
Gas turbines have been developed rapidly over the
NRC National Research Council (USA)
last 60 years, with applications currently ranging
PLPS Photoluminescence piezospectroscopy
from aeroengines to marine propulsion to industrial
TBC Thermal barrier coating
mechanical power to utility scale electrical power
TGO Thermally grown oxide
generation.1,2 Whatever be the intended application

518
Gas Turbine Oxidation and Corrosion 519

Combustor

Air inlet Compressor Exhaust


Power
turbine

Figure 1 Schematic diagram of simple cycle gas turbine showing fundamental components.

of the gas turbine, the basic features and principles of Thus, within these constraints, there are several
operation are common. Figure 1 illustrates the fun- routes to improve the efficiency of gas turbines:
damental parts of a gas turbine.  increasing the power turbine inlet temperature; this
Details of the operation of gas turbines (which use
is the traditional route with limitations being set by
the Brayton cycle) are available from numerous
the materials of construction for the hot gas path,
sources1,2 but can be summarized as follows for the
but now also being set by cooling technologies and
purposes of this chapter:
thermal barrier coating (TBC) performance;
 air enters the gas turbine through the compressor;  lower the exit temperature;
 the air is compressed (and heated adiabatically) as  using the heat in the exit gases in another way; for
it passes through a series of vane/rotor stages of example, by passing the gases through a heat
the compressor; recovery steam generator (HRSG) to generate
 this air is passed into combustion chamber(s) steam which is then passed through steam turbine
where fuel is introduced and burnt raising the to generate power (i.e., using a conventional Ran-
temperature of the gas stream; kine cycle); the overall system is known as a com-
 this hot, high pressure gas stream is then passed bined cycle, and overall efficiencies for such
through the power turbine where its temperature systems have been reported to exceed 60%.
and pressure are reduced as it passes through a series In addition, there are several other variations of
of vane/rotor stages which drive both the compressor gas turbine cycles that continue to be investigated:
and mechanical/electrical power conversion unit.
 Indirect firing; where the compressed air is heated
The Carnot efficiency of the gas turbine (as with by an external heat source via heat exchangers
any heat engine) is related to the difference between before passing through the power turbine;
the maximum and minimum absolute temperatures  Oxy-firing; where the fuel is burnt with oxygen to
of the working fluid (in which case, the inlet and generate a combusted gas stream of mostly CO2 and
outlet temperatures of the power turbine are divided steam, from which the steam can be condensed to
by the inlet temperature). For a current gas turbine produce a CO2 rich gas stream for recycling into the
with an inlet temperature of 1400  C (1673 K) and an compressor or for storage (after further cleanup).
exit temperature of 600  C (873 K), this would give a In terms of the sizes of gas turbines:
Carnot efficiency of 48%. Real cycle efficiencies
 the largest utility power turbines are rated at
are always lower than Carnot efficiencies, and in this
case, for a simple cycle gas turbine, the efficiency 340 MWe in simple cycle mode and 530 MWe
would be closer to 38–40%.3 If the inlet tempera- in combined cycle mode4;
 jet engines are rated in terms of their thrust, with
ture is to be raised to 2100  C or 2373 K (just
below stoichiometric combustion for maximum the largest generating up to 100 000 lbs.1
flame temperature1 – a near impossible achievement) From the foregoing, it is evident that materials in
and the exit temperature remains at 600  C (873 K), gas turbines need to be capable of operating at high
then the maximum thermal efficiency will be 63%, bulk temperatures with high static and fluctuating
without dropping the exit temperature to extract stresses, while by also withstanding the surrounding
energy from this lower grade heat source. environments. The environments generated within
520 High Temperature Corrosion Environments

the hot gas paths of gas turbines can be both physi- Excess air is always used in the process to ensure
cally and chemically aggressive, with particles pro- that the fuel is completely consumed, to reduce the
ducing erosion or deposition while gaseous and temperature of the combusted gas stream, and to
vapor-phase species may produce different forms of increase the mass flow through the power turbine.
deposition, as well as oxidation and hot corrosion. The This produces a high temperature combusted gas
balance between these various degradation mecha- stream that is oxidizing. For some industrial gas tur-
nisms is component and turbine dependent (as are bines, water/steam is also added to increase the mass
the temperatures and stresses involved). flow through the power turbine and/or reduce NOX
Thus, from a generic materials point of view, emissions5–9 (in some cases, steam-cooled compo-
issues of concern for gas turbine hot gas path compo- nents also exhaust into the main combusted gas flow).
nents include: However, unwanted products of combustion pro-
cesses can be produced:
 creep,
 fatigue (thermal, low cycle, and high cycle),  CO from incomplete combustion of the fuel;
 thermal cycling,  NOX from thermal deposition of air or fuel-
 oxidation, derived nitrogen species;
 hot corrosion,  SOX from oxidation of fuel borne impurities.
 erosion/foreign object damage (FOD),
Various approaches have been developed to con-
 synergistic oxidation/corrosion/mechanical prop-
trol these unwanted combustion products, such as
erty interactions.
improved combustor design, staged combustion,
This chapter focuses on the oxidation and hot water injection, etc. Not all are appropriate (or feasi-
corrosion performance of materials in a range of ble) for every gas turbine application, so their use
different gas turbines and on how thermal cycl- varies. However, increasingly stringent environmental
ing, erosion, and mechanical loading affect this regulations are being developed and applied in various
performance. countries; especially to utility scale gas turbines.
Unfortunately, the combusted fuel gas stream also
includes species derived from minor and trace ele-
1.20.2 Gas Turbine Operating ments in the fuels and oxidant streams: these elements
Conditions commonly include sulfur, alkali metals, chlorine,
magnesium, calcium, lead, vanadium, etc. However,
Gas turbines have conventionally been fired using when gasification processes are considered, even
hydrocarbon fuels (such as diesels, jet fuels, and more elements can be involved, including cadmium
natural gases) burnt using air as the oxidant. More and arsenic.10 These various elements can combine to
recent fuels have included syngases produced by coal cause hot corrosion damage to the hot gas path of the
or biomass gasification processes. In addition, use of gas turbine. The reactions involved, their effects, and
oxygen as the oxidant is now being actively investi- methods to limit them (including fuel specifications)
gated for utility gas turbines as one potential route to are described in detail in Section 1.20.5.
enable CO2 capture technologies to be introduced. During the last 60 years, there has been a continu-
In all cases, the basic reaction of the fuels with the ous drive to increase the efficiencies of all types of gas
oxidant can be summarized as: turbine, but they still fall short of the Carnot theoreti-
cal optimum, which would be achieved at stoichiomet-
Fuel þ air ! Combusted gases
ric combustion combined with no mechanical losses.
or As a result, the firing temperatures and pressures
of gas turbines have increased steadily (Figure 2).
C þ yH þ lð1 þ y=4ÞðO2 þ 3:719N2 þ 0:043ArÞ
Current industrial gas turbines for power generation
! CO2 þ ðy=2ÞH2 O þ lð1 þ y=4Þð3:719N2 applications have firing temperatures of up to
1430  C and pressures of up to 34 bar, while for
þ 0:043ArÞ þ ðl  1Þð1 þ y=4ÞO2
jet engine applications, firing temperatures are up to
where y is the reciprocal value of the C to H ratio in 1550  C and pressures up to 45 bar.1
mol, l is the amount of excess air actual mass flow Figure 3 illustrates the temperature and pres-
divided by the air mass flow for stoichiometric sure variations expected through a generic gas tur-
combustion. bine. There are wide variations in the actual
Gas Turbine Oxidation and Corrosion 521

1300 24
SGT800
1200 SGT400 21
SGT700
Firing temperature (⬚C)
1100 18

Pressure ratio
SGT100
1000 15
SGT200
900 12

TB5000
800 9

700 6
1975 1980 1985 1990 1995 2000 2005
Year
Figure 2 Increase in turbine entry temperatures and pressure ratios over time for Siemens industrial gas turbines
(4–50 MW). Reproduced from Hannis, J. M. In Power Generation in an Era of Climate Change; Proceedings of the 7th
International Charles Parsons Turbine Conference; Strang, A. et al. Eds.; IOM Communications, 2007; pp 37–49.

20 1400
Maximum gas pressure
18 Minimum gas pressure 1300
Tempest-gas pressure in/out
16 Cyclone-gas pressure in/out 1200

Metal temperature (°C)


TB5000-gas pressure in/out
14 1100
Gas pressure (bar)

Typhoon-gas pressure in/out


12 Other GTs-gas pressure in/out 1000
Maximum metal temperature
10 Minimum metal temperature 900
8 800

6 700
4 600

2 500
0 400
400 500 600 700 800 900 1000 1100 1200 1300 1400
Gas temperature (°C)
Figure 3 Variation in temperature and pressure of the gas stream passing through typical industrial gas turbines.
Reproduced from Simms, N. J.; Nicholls, J. R.; Oakey, J. E. In Lifetime Modelling of High Temperature Corrosion Processes;
Schütze, M. et al. Eds.; Maney, 2001; pp 379–397, EFC No. 34.

temperatures and pressures that are produced by (with flight times of up to 12 h), with peak operating
different gas turbines. This results from a number of temperatures for short periods during takeoff and
factors such as: landing. In contrast, industrial gas turbines designed
for utility scale electrical power generation should be
 power requirement of the turbine,
ramped slowly to their optimum operating tempera-
 design approach,
ture and then ideally remain there for hundreds if not
 duty and cycle required,
thousands of hours. Other differences include weight
 materials/lifting restrictions.
limitations and the various contaminants that can
The gas turbines used for the different applications enter the gas turbines during their operation. Thus
outlined in the introduction, while obviously related the gas turbine usage defines the maximum gas tem-
by their basic principles of operation, have different perature and pressure, duty cycle, operating time at
detailed requirements. For example (Figure 4), civil temperature, number of shut-down cycles, and mate-
aeroengines go through a routine cycle on every flight rial operating environment.
522 High Temperature Corrosion Environments

1300

1200

1100

Temperature (°C) 1000

900

800

700

600

500

400
0 20 40 60 80 100
(a) Time (min)

1300

1200

1100

1000
Temperature (°C)

900

800

700

600

500

400
0 10 20 30 40 50 60 70 80 90 100
(b) Time (h)
Figure 4 Examples of gas temperature cycles for an aeroengine flight and an industrial gas turbine.12 (a) Example of short
duration aeroengine cycle (Note: 4000–7500 cycles between overhauls, depending on cycle duration). (b) Example of
industrial gas turbine cycle (Note: desired component lives of >24 000 h).

1.20.3 Gas Turbine Hot Gas Path improved mechanical properties and thermal stabili-
Components and Materials ties. As these compositions have been progressively
optimized, new materials have been developed with
As efficiencies of all types of gas turbines increase improved high temperature mechanical properties
with firing temperature and pressure, there has (Figure 5). Until the 1970s, such improvements in
always been a drive for the hot gas paths to use high temperature mechanical properties could be
materials close to their mechanical and environmen- coupled to improved oxidation and hot corrosion
tal operating limits.1,2 However, viable component resistance,2,14 with alloys being designed to have suf-
lives for the hot gas path of the gas turbines have ficient chromium and aluminum contents to form
only been achieved as a result of significant materials self-protective oxide scales. However, more recent
developments coupled with improved manufacturing alloy developments, targeted at improved mechanical
methods and engine designs (specifically efficient properties, could only be achieved to the detriment
cooling). Originally steel parts were used, but as the of oxidation and hot corrosion resistance of these
severity of operating conditions was increased, it alloys. Thus there is now a requirement for many
was necessary to invent and develop whole series hot gas path components to be coated to enable
of nickel-based and cobalt-based superalloys with them to resist the surrounding environments and
Gas Turbine Oxidation and Corrosion 523

Uncooled Cooled Coated


2100 Demonstrator
blades blades blades
technology

Turbine entry temperature (K)


1900
Trent
Production
1700 technology

1500
Conway RB211
1300 Spey SX cast
W1 Avon
Dart DS cast
1100
Cast alloys
Wrought alloys
900
1940 1960 1980 2000 2020
Year of introduction
Figure 5 Improvements in materials strength with time and new alloy development. Reproduced from Rickerby, D. S.;
Hicks, M. A. In Power Generation in an Era of Climate Change, Proceedings of the 7th International Charles Parsons Turbine
Conference; Strang, A. et al. Eds.; IOM Communications, 2007; pp 229–240.

to permit economically viable component lives. Higher TBCs means that first stage blades/vanes can now be
temperature alloy mechanical performance, as well as operated at gas temperatures in excess of the melting
associated engine capability, has also been enhanced points of the alloys used for these components.3,4
through improvements in casting technologies. These Critical components within the hot gas path of gas
have permitted a change in the microstructure of turbines include:
nickel-based superalloys from equiaxed to direction-
ally solidified and single-crystal blade materials  combustor can/tiles/transitions ducts,
(eliminating the grain boundaries that partly dictate  vanes and blades (both in multiple stages),
high temperature creep damage). Originally these  seals,
casting technologies were developed for small aero-  discs.
engine blades, but now larger industrial gas turbine Examples of alloys used for these components are
blades/vanes can also be processed.3 given in Table 1.
The development of a range of increasingly sophis- A wide range of coatings have been developed and
ticated component cooling technologies (including air used for protection along the gas turbines hot gas path.
impingement, film cooling, etc. Figure 6) has enabled The different routes for coating application that have
gas stream temperatures to be decoupled from metal developed are favored by different gas turbine manu-
temperatures along the gas turbine hot gas path. Most facturers. Thus, the situation of using a ‘standard’
of the cooling systems rely on air diverted from coating for protection against hot corrosion damage
selected locations in the air compressor, but there are has not developed. Instead, there are several coatings
some land-based gas turbines that use steam cooling. from the different classes of coating materials available
The use of such cooling technologies has enabled gas that are currently being used (Table 2):
temperatures to be progressively increased, much fur-
ther than would be expected from alloy developments.  Overlay coatings:
The introduction of cooled components has  typically Ni–Co–15–23% Cr–8–12% Al–0.1–
allowed the adoption and development of TBCs as 0.9% Y,
part of the turbine thermal management system.  applied by high velocity oxy-fuel (HVOF)
These low conductivity ceramic coatings protect the or low pressure plasma spray (LPPS) coating
underlying metallic materials from the surrounding processes,
high temperature gas streams and allow a balance  some newer coatings also include Re, Hf, Ta, or
between component life, component operating tem- Si in some combination.
perature, and cooling air requirements to be addressed.
The principal of TBCs is illustrated in Figure 7. The  Diffusion coatings:
use of advanced air cooling technologies coupled with  Aluminizing, chromizing, or chrome-aluminizing;
524 High Temperature Corrosion Environments

Front chamber double end Air passes


feed to leading edge holes through holes
in impingement
plate cooling
the aerofoil

Baffle plate

Air exits leading


edge holes to Air exits through
cool NGV trailing edge
pedestal bank

Film cooling

HP NGV cooling
flows, showing
Front chamber Single end feed impingement and
double end feed baffle plates

Figure 6 Example of air cooling systems in advanced gas turbine aerofoil.1 Courtesy of Rolls Royce.

Thermal  Platinum-aluminizing;
Combustion barrier Bond
gases coating coat Substrate Cooling air  Electroplate of platinum followed by CVD
aluminizing process.
 Slurry coatings;
Tgas
 Silicon-aluminizing.
 Coating performance varies with base alloy
Temperature

compositions, which have a notable effect in


terms of hot corrosion resistance, with such
coatings on lower Cr-based alloys performing
significantly worse than on higher Cr substrates.
Tcoolant

Distance
1.20.4 Oxidation and Thermal Cycling
Figure 7 Principal of operation of thermal barrier coatings
(TBCs). 1.20.4.1 Oxide Formation and Spallation
1.20.4.1.1 Oxidation processes
 Chemical vapor deposition (CVD) processes Gas turbine materials will oxidize in the combustion
applied by in-pack, above pack or gaseous gases produced in all gas turbine systems. The earlier
routes; superalloys were relatively rich in chromium and
 Gaseous routes enable coating of internal so formed slow growing protective Cr2O3 scales at
cooling passages in advanced aero-foil designs; their temperatures of use. However, as component
Gas Turbine Oxidation and Corrosion 525

Table 1 Examples of materials used in gas turbine hot gas path components (in wt%)

Ni Co Fe Cr Al Ti Mo W Others

CMSX-4 Bal 9.0 6.5 5 1 0.6 0.1 Hf


IN738LC Bal 8.5 16 3.4 3.4 1.7 2.6 1.7 Ta, 0.9 Nb
GTD111 Bal 9.5 14 3 4.9 1.5 3.8 2.8 Ta
IN939 Bal 19 22.5 1.9 3.7 2.0 1.4 Ta, 1.0 Nb
IN6203 Bal 19 22 2.3 3.5 2.0 1.1 Ta, 0.8 Nb, 0.8 Hf
GTD222 Bal 14 22.5 1.2 2.3 2.0 1.0 Ta, 0.8 Nb
IN792 Bal 9.0 12.4 3.1 4.5 1.9 3.8 3.9 Ta
FSX-414 10 Bal 1 29 7.5
MarM 247 Bal 8.3 5.5 1 0.7 10 1.5 Hf
Hastelloy X Bal
Haynes 230 Bal <3 22 0.3 2 14
RR1000 Bal 18.5 15 3 3.6 5 2 Ta, 0.5 Hf
IN718 Bal 18.5 19 0.5 0.9 3 5.1 Nb
Astroloy Bal 15 15 4.4 3.5 5.2
Waspalloy Bal 13.5 <2 19.5 1.4 3 4.3
U500 Bal 19 4 19 3 3 4
U720Li Bal 15 16 2.5 5 3 1.2
Haynes 214 Bal 3 16 4.5 0.1 Zr, 0.01 Y
PM2000 Bal 20 5.5 0.5 0.5 Y as Y2O3
Aluchrom YHf 0.2 Bal 20 5.8 0.5 Y, 0.05 Hf, 0.05 Zr

Table 2 Examples of gas turbine coating compositions The development of superalloys for high metal tem-
(wt%) peratures resulted in materials with progressively less
Overlays Ni Co Cr Al Y Other
chromium and more aluminum. This resulted in
alloys that formed slower growing Al2O3 scales at
Amdry 963 Bal – 25 6 0.4 high temperatures that were not volatile in the sur-
Amdry 995 32 Bal 21 8 0.5 rounding gas streams.2
Amdry 997 Bal 23 20 8.5 0.6 4 Ta
LCO22 32 Bal 21 8 0.5
At temperatures higher than 950–1000  C,
CT102 32 Bal 21 8 0.6 a-Al2O3 scales are the preferred scale to be formed.
CT103 Bal 22 17 12 0.6 Considerable research targeted at a range of base
CT104 Bal 23 22 11 0.4 4 Hf, 0.4 Si alloys and different applications has been carried
GT29 – Bal. 29.0 6.0 0.3 out on the growth of these scales over the year.2,16
SV20 Bal – 25 5.5 0.7 1 Ta, 2.7 Si
SV30 Bal. – 25.0 5.0 0.7 0.1 Ta,
However, at lower metal temperatures, more com-
0.35 Ti plex oxides may form on the surfaces of superalloys:
SC2231 30 Bal 28 8 0.5 0.5 Si various transition aluminas (e.g., y- or g-Al2O3) as
SC2453 Bal 10 28 12 0.6 3 Re well as mixed (Cr, Al)2O3 oxides, spinel oxides of Cr
and/or Al with base metal oxides, and oxides of other
Diffusion Description
minor alloying elements that are present in these
Aluminized Applied in-pack, above pack or by gas complex superalloy materials may form.
phase CVD In general, the combination of metal temperature
Chromized Applied in-pack, above pack or by gas
and a ‘clean’ oxidizing environment (i.e., a combus-
phase CVD
RT22 Variations on platinum electro-plate followed tion gas stream with trace metal contaminants at too
CN91 by aluminizing low a level to cause any hot corrosion) will result in
MDC150 the formation of slow-growing stable oxide scales for
SermaLoy Slurry application the appropriate blade and vane materials, plus any
1515
coating system, in the gas turbine hot gas path (but it
should be noted that such materials will vary between
temperatures rose, so did the rate of Cr2O3 formation, gas turbine manufacturers, and with the intended use
and other oxidation mechanisms, such as formation of of gas turbine).
volatile oxide and oxyhydroxide species in the fast There will, therefore, be a difference between
flowing combustion gases, also became signifcant.2,15 gas turbines: the higher peak temperatures of
526 High Temperature Corrosion Environments

aeroengines will encourage a-Al2O3 scale formation, At high temperatures, coatings will degrade by
compared to industrial machines operating close to interdiffusion with the substrate as well as surface
oxide transition temperature. Another important dif- oxidation; this can be particularly important for alu-
ference in the exposure conditions will be number minum which tends to diffuse inwards as well as
and severity of thermal cycles. Aeroengine tempera- being consumed due to oxidation. The development
tures are cycled much more frequently and rapidly of diffusion barriers to reduce or prevent such
than industrial gas turbines. As a result, the slow- inwards diffusion is an active research topic.
growing scales may spall from the component sur-
faces and a new oxide will form during the next 1.20.4.1.2 Modeling oxide growth and
period of operation: whether this is the same oxide spallation
or not will depend on the composition of the under- Oxide growth and spallation models have been devel-
lying alloy/coating. oped for alumina type oxide scales by several research
The coating of superalloys with most overlay and groups around the world, most notably: NASA,18
diffusion coatings (Table 2) results in a surface rela- Jülich–Cranfield19,20 and Toulouse.21 These models
tively rich in aluminum, which assists in the forma- are essentially all empirical, but are based on the math-
tion, retention, and reformation of a slow-growing ematical understanding of oxidation processes:
Al2O3-based (or at least alumina rich) oxide scale.
 The NASA cyclic oxidation and spallation (COSP)
Overlay coatings traditionally include ‘reactive ele-
model is based on rate constants determined from
ments’ such as Y or Hf that are well known to enhance
mass change data and uses these combined with
the adhesion of alumina scales (see Chapter 1.23,
idealized parameters that describe the spallation
High Temperature Corrosion of Chromia-forming
process to predict the damage to a metal in terms
Iron, Nickel and Cobalt-base Alloys). Scales can
of mass change or metal loss for different time
develop through transient oxides (i.e., various
periods, cycle durations, etc.
Al2O3) and other oxides (e.g., NiO, spinels) and mul-
 The Toulouse p–kp model uses a spalling proabil-
tiple oxide layer developments (e.g., undergrowths of
ity (p) and parabolic rate constant (kp) to predict
Cr2O3 and a-Al2O3) before reaching a slow growing
the mass change of materials as a function of time
stable structure (e.g., Figure 8). This development
and temperature.
depends on the exposure temperature and coating–
 The Jülich–Cranfield model uses the concept of an
base alloy composition.
aluminum reservoir to calculate the metal loss due
Diffusion coatings have also progressively devel-
to alumina growth and spallation. Spallation pro-
oped to assist in the growth and retention of alumina
cesses are assumed to commence after some criti-
scales, in this case, using metals such as Pt and Pd.
cal mass gain (or oxide thickness) which is
These coatings will also develop scales through the
influenced by thermal cycling.
formation of transient oxides (i.e., various Al2O3) and
other oxides (e.g., NiO) and multiple oxide layer Continuous oxide growth and coating interdif-
developments (i.e., undergrowth of a-Al2O3) before fusion models are being developed by several
reaching a slow growing stable structure. As for over- research groups, most notably NASA,22 DTU,23 and
lay coatings, this development depends on the expo- Toulouse.24,25 These use various finite difference
sure temperature and coating–base alloy composition. or finite element methods, in some cases combined

α-Al2O3 Cr2O3 NiAl2O4 Nio

2 μm
(a) Back-scattered image (b) Gray scale image
17
Figure 8 Scale grown on Ni–27Cr–8Al at 1353 K. (a) Back-scattered image and (b) grey scale image.
Gas Turbine Oxidation and Corrosion 527

with thermodynamic phase predictions, to calculate as well as to provide a transition zone between the
elemental profiles throughout the oxide–coating– metal and the ceramic layer. Thus bond coat materi-
substrate system as a function of time and tempera- als are often classically established environmental
ture, again using the ideal a-Al2O3 scale formation. protection coatings, for example, a platinum alumi-
nide diffusion coating system or an MCrAlY overlay
coating. However, bond coatings for APS and EB
1.20.4.2 Thermal Barrier Coating Systems
PVD TBCs require different initial characteristics
1.20.4.2.1 Damage mechanisms to enable the coatings to be successfully applied.
Two types of TBCs are now frequently found in gas APS TBCs generally need to be applied to deliber-
turbines. These are generally classified by their appli- ately roughened bond coat surfaces to provide
cation routes: air plasma sprayed (APS) and electron mechanical keying, whereas EBPVD TBCs are usu-
beam physical vapor deposition (EB PVD) (Figure 9). ally applied to smoother, polished, or media finished
However, currently their compositions are generally bond coat surfaces.
the same: 7 wt% Y2O3 – balance ZrO2. New thermal The spallation of TBCs after exposure at high tem-
barrier ceramics are also being researched,27 but a perature is now well recognized as the life-limiting
review of these is beyond the scope of this chapter. factor dictating TBC performance and has been fre-
Both types of TBCs are applied to gas turbine quently reported.28 However, the causes of this spall-
components after they have had a bond coating ation remain controversial (e.g., Figure 10). Features
applied. The purpose of this bond coating is to reported include28:
encourage and maintain the formation of a stable a-
 spallation is usually observed on samples/compo-
Al2O3 scale at the interface between metal and ceramic,
nents after the growth of 5–10 mm of oxide scale;
 cracks grow roughly parallel to the bond coat/
thermally grown oxide/TBC interfaces, before
large areas of TBC spall;
 cracks may grow:
 in TBC near TGO interface (between bond-
coat peaks, or not),
 within the TGO,
 along bond coat/TGO interface, depending on
the system;
 bondcoat/TGO/interfaces may roughen with
exposure (for EBPVD TBCs on Pt–Al bond coats);
 mixed oxides (Cr,Al)2O3 or spinels (e.g., (Ni,Co)
100 μm
(a) Cr2O4) may form above or below the stable alu-
mina scales in areas close to crack growth.
Both academic and industrial research is actively
continuing in this area, but has moved from charac-
terization towards methods of finding damaged TBCs
nondestructively and also developing predictive
models for TBC spallation (see below). But it is
worth noting that there are many significant differ-
ences between the preparation and processes used for
applying bond coatings, initiating the growth of ther-
mally grown oxide and applying the TBC layer that
are used by different coating companies, all of which
100 μm may interact and therefore influence susceptibility to
(b) TBC spallation.
Particular interest has been taken in finding ways
Figure 9 Two main types of TBCs.26 (a) Air plasma
sprayed thermal barrier coating (APS TBC) and (b) electron
of detecting the cracks in/close to TGO using non-
beam physical vapor deposition thermal barrier coating destructive inspection methods, before the cracks
(EB PVD TBC). have a chance to coalesce and cause TBC spallation
528 High Temperature Corrosion Environments

TBC
Transient oxides
and detached areas TBC

Oxide protrusions TGO


and inclusions TGO

NiCoCrAIY bond coat NiCoCrAIY bond coat

(a) (b)

Defects
TBC TBC
Defects

TGO TGO
Crack
Crack
NiCoCrAIY bond coat NiCoCrAIY bond coat

(c) (d)
(A) Possible causes of TBC spallation [70]

Thermochemical

α-Al2O3 TBC TBC


α-Al2O3 Impacting
particle
Spinel

(aAI<10–17) Delamination
BC BC
IAI

(a) AI depletion (c) Undulating interface Tg


Compacted Erosion
area
TBC TBC
TBC

Spinel α-Al2O3 Hot


BC
α-Al2O3 area
IN Crack of BC
BC inclusion (e) Foreign object damage

(b) Outward Ni transport (d) Local oxide thickening

(B) Thermomechanical
Figure 10 Schematic diagrams of alternative views of the causes of TBC spallation. (A) Possible causes of TBC spallation29
and (B) possible causes of TBC spallation. Reproduced from Simms, N. J.; Kilgallon, P. J.; Roach, C.; Oakey, J. E.
In Microscopy of Oxidation, Proceedings of 5th International Conference, Newcomb, S. and Tatlock, G. Eds.; Science
Reviews, 2003; pp 273–280.

to occur. Promising nondestructive techniques cur-  photoluminescence piezospectroscopy (PLPS) –


rently include30: this uses lasers to excite the Cr3+ contaminant ions
that are always present in Al2O3 oxide scales in these
 images produced by differential heat conduction material systems and detects the shift in the charac-
(using lasers or other heat sources) – essentially teristic peaks produced: the greater the shift the more
relying on cracked regions of the coating system stressed the TGO. This technique can be used to
not allowing heat to pass through as readily as fully examine a coated TBC component and to detect
adherent coated regions; variations in stresses in the TGO (cracked areas
Gas Turbine Oxidation and Corrosion 529

Top coat valley

TGO centre

Top coat peak

Figure 11 Finite element approach to TBC spallation modeling as a function of TGO growth. Reproduced from Hermosilla, U.
PhD Thesis, Nottingham University 2008; UK EPSRC Supergen Plant Life Extension Consortium.

showing up as low stressed TGO). This technique 6.6 6.8 7 7.2 7.4 7.6 7.8 8 8.2
0
works better on EB-PVD TBCs than APS TBCs due –0.5 Activation energy from
NiCoCrAIY bond coats kp α-Al2O3 growth on
to the columnar microstructure of the former that –1 Pt.Al Alloys = 85 kcal/mole

allows light signals to pass through more easily. –1.5 Same slopes Activation energy
–2 = 85.1 kcal/mole
log 1st

–2.5
1.20.4.2.2 Modeling TBC spallation –3
Platinum aluminide Pt overlayer on NiCoCrAIY
Several industrial and academic research groups are –3.5 bond coats
14791
working on the development of models to predict –4
spallation of TBCs.31–37 There are essentially two –4.5
–5
distinct approaches:
104/T
 Finite element modeling of a selected area of an
Figure 12 Observations of TBC spallation that form a
idealized TBC system to predict development of basis for analytical approaches to life modeling.
stresses around TGO as it grows (e.g., Figure 11); Reproduced from Yanar, N. M.; Kim, G.; Hamano, S.;
these models are being developed by several EU Petit, F. S.; Meier, G. H. In Microscopy of Oxidation,
and US research groups.32–34,36,37 Proceedings of 5th International Conference, Newcomb, S.
 Phenomenological approaches based on observa- and Tatlock, G., Eds.; Science Reviews, 2003; pp 249–260.

tion of the performance of many TBC-coated


more rapidly and can be life limiting for all compo-
samples under a range of conditions to identify
nents. As a result, hot corrosion of gas turbine hot gas
the most important variables and their effects as a
path materials has been the subject of active research
function of exposure time and temperature (e.g.,
and industrial experience for more than 50 years.2
Figure 12); these types of models are being dev-
The potential problems that may be encountered in
eloped particularly by Dechema and Cranfield in
gas turbines fired on conventional natural gas, diesel,
Europe31,35 and Pittsburg in USA.29
and jet fuels have been well characterized,2,15,38–45 and
fuel standards derived (Table 3). In recent years, the
1.20.5 Hot Corrosion ranges of fuels that need to be considered have wid-
ened in response to changes in availability and price. In
Although oxidation of gas turbine hot gas path addition, investigations have also been carried out into
materials is rarely life-limiting, unless localized over- the potential effects on gas turbines from using fuel
heating has occurred (e.g., following TBC spallation), gases derived from solid fuels (e.g., gasification of coal
the hot corrosion of these materials can occur much and/or biomass).47
530 High Temperature Corrosion Environments

Table 3 Fuel specifications46

Fuel type True distillates Ash-forming fuels

Kerosene No.2 Distillate Blended residuals and crude Heavy residuals

Sulfur (%) 0.01/0.1 0.1/0.8 0.2/3 0.5/4


Ash (as delivered) (ppm) 1/5 2/50 25/200 100/1000
Ash (inhibited) (ppm) – – 25/250 100/7000
Trace metal contaminants (untreated) (ppm)
Sodium + potassium 0/0.5 0/1 1/100 1/350
Vanadium 0/0.1 0/0.1 0.1/80 5/400
Lead 0/0.5 0/1 0/1 0/25
Calcium 0/1 0/2 0/10 0/50

100 μm
(a) 600 μm (b)

Type I hot corrosion Type II hot corrosion


Figure 13 Typical appearance of type I and type II hot corrosion during their propagation stages.50 (a) Type I hot corrosion
and (b) type II hot corrosion.

These studies have successfully identified the two widely recognized forms of hot corrosion, distin-
causes of hot corrosion and some methods to limit guished by their propagation modes in gas turbines
the damage caused. A considerable amount of data on (Figure 13)2,15,38–45:
hot corrosion damage for a wide range of materials,
 Type I hot corrosion has traditionally been
under different exposure conditions, has been gen-
observed in a temperature band between the melt-
erated during the course of research projects and
ing point of the surface deposit and vapor deposi-
industrial usage.2,42 This dataset has permitted,
tion dew point for the deposit (above this dew
experimentally determined, safe operating conditions
point deposit is unstable and will evaporate). An
to be produced. However, the development of quan-
example of this could be the formation of sodium
titative predictive models for hot corrosion damage
sulfate deposits. This form of corrosion damage is
to materials of current interest in gas turbines is an
characterized by the appearance of internal sulfide
on-going activity,48,49 with some embryonic models
phases in a zone of selective alloy depletion beneath
produced for these complex corrosion phenomena.
a fairly smooth scale–metal interface. The typical
temperature ranges quoted for this type of attack
1.20.5.1 Hot Corrosion Mechanisms
are typically 750–900  C. This form of attack has
Hot corrosion of materials in a power turbine occurs been called high temperature hot corrosion.
when a liquid deposit forms on the surface of a com-  Type II hot corrosion occurs at a lower range of
ponent. After a period of incubation as the deposit temperatures (typically quoted as 600–750  C).
forms on the surface and reacts with the initially This form of attack relies on the formation of a
protective oxide, hot corrosion reactions move into complex mixed metal sulfate deposit: an example
their more rapid propagation stages. There are of this could be the formation of a mixed nickel
Gas Turbine Oxidation and Corrosion 531

sulfate–sodium sulfate deposit. Such deposits are 4

(mole metal ions/mole Na2SO4)


not very stable and require the presence of signifi- Cr2O3

Log concentration ppm


cant levels of SO3 to stabilize them (hence the term
Fe2O3
‘gas phase induced acidic fluxing’). The corrosion 3
occurs over the stability range of the liquid deposit NiO
from its melting point to its dissociation tempera-
Al2O3
ture. The attack is characterized by the formation of 2
pitted or uneven surfaces with layers of corrosion Co3O4
products/deposits on the surface and no subsurface
depletion zones or internal sulfide formations 1 SiO2(900 ⬚C)
(although a layer of sulfides may be observed at 5 7 9 11 13 15 17
the metal–scale interface). This form of attack has –Log aNa2O
been called low temperature hot corrosion.
Figure 14 Oxide solubility in molten Na2SO4 at 927  C.
Reproduced from Sims, C. T.; Stoloff, N. S.; Hagel, W. C.
The mechanisms of hot corrosion are generally
Eds. Superalloys; Wiley: New York, USA, 1987; Vol. II.
believed to be a result of the interaction of the surface
oxide scales at first and then of the underlying alloys
with molten surface deposits.2,15,38,39,42 These are KCl-NaCl-K2SO4-Na2SO4
commonly described as fluxing mechanisms. Both Na2SO4 K2SO4
acidic and basic fluxing reactions are possible depend- 884⬚ ~932⬚ 1069⬚
ing on the deposit composition and the composition of 95
0
oxide/underlying coating/alloy. P108⬚
P696⬚ 85
In basic fluxing, oxide ions are produced in a P686⬚ P736⬚ 0
deposit of Na2SO4 by removal of oxygen and sulfur [II]
75
0
[I] [III]
from the deposit by reaction with underlying coating/ Na 2S
O4
65 K
628⬚ 0
alloy. Then, the oxide scales (e.g., MO) that form can ⬚ P567⬚ 2S
O
P576P ⬚
react with the oxide ions via reactions such as2,15,38: 536⬚ 534 P550⬚
4
55 P
N

0
aC

MO þ O2 ¼ MO2
l

2 0 KCl 690⬚
65 60
In acidic fluxing, the oxide scale dissolves to 0

donate oxide ions to the deposit melt2,15,38:


0
70

MO ¼ M2þ þ O2
75
0

Acidic conditions can be developed in a molten


deposit by at least two different processes, leading 800⬚ ~658⬚ 774⬚
to the terminology gas phase acidic fluxing and alloy- Na2Cl2 K2Cl2
Mol %
induced acidic fluxing.
As a result of these different possible hot corrosion Figure 15 Na2SO4–K2SO4–NaCl–KCl phase diagram.
Reproduced from Phase Diagrams for Ceramists; American
mechanisms, the solubility of commonly formed oxi-
Ceramic Society, 1964.
des in molten deposits by acidic and basic reactions has
been the subject of several studies (e.g., Figure 14).
In practice, the temperature ranges of type I and combustion gas stream (e.g., calcium, magnesium,
type II hot corrosion are affected by the composition lead, or chlorine) or elements from the components
of deposits formed. For example, the presence of being corroded (e.g., molybdenum or tungsten).
potassium in a deposit can significantly lower the Some of these elements may also participate in the
melting point of the deposit and lower the tempera- hot corrosion reactions between the deposit and the
ture range for type I when compared to sodium component materials.
sulfate alone (Figure 15). Similarly, NaCl may also Along with the distinctive propagation morpholo-
lower deposit melting points51 and also assist the gies of types I and II hot corrosion, it is also possible
breakdown of any protective oxide scales. for an intermediate form of attack to develop which
Other elements can also affect the melting ranges has some of the features of both. This has been
of these deposits: these may be contaminants from described as mixed mode attack or a transition-type
532 High Temperature Corrosion Environments

attack. It is characterized by the formation of layered combustor, these contaminants can join those exit-
corrosion products/deposit on the component surface ing in the compressor and react as they pass through
and intermittent areas of internal sulfides/depletion the combustor and power turbine.
zones and an uneven scale/metal interface. This form  Salts entering the compressor can be diverted from
of attack has been observed at temperatures between the main air flow into smaller flows that are used to
traditional type I and type II corrosion regimes. cool components in the power turbine; these salts
An alternative mode of hot corrosion attack is seen do not get the chance to react with the fuel derived
in turbines in which the combustion gas stream con- contaminants as they pass through the combustor,
tains vanadium (usually a contaminant that arrives so different compositions of deposits can form
via a fuel oil). This form of hot corrosion also from this contaminant routing.
depends on the formation of liquid deposits on the
For gas turbines that operate under steady condi-
surfaces of components, but these low-melting-point
tions (i.e., some larger land-based gas turbines), it is
deposits do not require the presence of high levels of
possible to use knowledge of these conditions to
SO3. In the presence of sodium-containing deposits,
calculate the total contaminants entering the system
mixed deposits, for example, sodium vanadyl vanadate
(whether from fuel or oxidant or coolant) giving rise
may form with melting points as low as 535  C.51 These
to an equivalent fuel contaminant level49,54,55:
sodium vanadate compounds can flux the protective
oxide from component surfaces and then continue XT ¼ ðA=F ÞXA þ ðS=F ÞXS þ XF
with a fluxing reaction with the metallic substrates
where XT is the total contamination in fuel equiva-
below. It has been found that this form of hot corrosion
lent (ppm) by weight; A/F is the air to fuel mass flow
can be suppressed by deliberately adding dopants (such
ratio; S/F is the steam (water) to fuel mass flow ratio;
as magnesium or chromium compounds) to the fuel;
XA is the contaminant concentration in inlet air
these are designed to react with the vanadium species as
(ppm) by weight; XS is the contaminant concentra-
they pass through the combustion process and deposit
tion in injected steam or water (ppm) by weight; XF is
onto component surfaces as solid deposits with high
the contaminant concentration in fuel (ppm) by
melting points that remain stable and unreactive.
weight.
This approach provides a method of assessing
1.20.5.2 Deposit Formation and Fuel
potential corrosiveness of gas streams within the gas
Specifications
turbine and is one route to setting specifications for
For hot corrosion to occur, a liquid or partially mol- safe operation of such turbines.
ten deposit must be created on the surface of compo- It is also possible to consider how the various
nents. Such deposits can be formed by a number of contaminant species that enter the power turbine
different mechanisms,52,53 but all require the com- can deposit out onto components. There are several
busted gas streams passing through the gas turbine well-known deposition processes (and associated
hot gas path to contain both trace metal species (such models52,53,56–62) that can occur within a gas turbine
as alkali metals) and reactive gas species (such as hot gas path (illustrated schematically in Figure 16
SO2/SO3). The source of the contaminants may be for an aerofoil):
either the fuel or the oxidant (usually air) or compo-
 Direct inertial impact ion; larger particles
nent coolant (usually air, but sometimes steam in
(>5 mm) from the gas stream impact on compo-
large land based gas turbines) or a mixture of all
nents instead of following the gas flows around
three. For example:
them; depending on their properties and the sur-
 Sea salt contains a mixture of metal chlorides and face condition of the components, these particles
lower levels of sulfates: these contaminants may may stick to the component surface, bounce off (or
enter a turbine through its air inlet, pass through partially stick).
the compressor and the combustor before enter-  Eddy diffusion; smaller particles may become
ing the power turbine. The salts may build up as entrained in the turbulent eddies that can form in
deposits on the compressor vanes/blades before certain locations in the gas paths (e.g., towards the
shedding and passing further into the turbine as trailing edge of aerofoil suction surfaces, or around
particles, or may pass through as vapor or aerosols. blade platforms, disc edges, seals, etc.); these par-
 Gas turbine fuels usually contain sulfur compounds ticles can then impact on component surfaces, and
together with a number of trace metal species; in the again depending on their properties and the surface
Gas Turbine Oxidation and Corrosion 533

Diffusion
High velocity
thin boundary Suction
layer side

Film cooling Air cooling


provides barrier increases DT
to deposition and thermophoresis Alkali deposition when
surface temperature
below dewpoint

Inertial
Low-velocity deposition
thicker boundary layer

Pressure
side

Figure 16 Potential deposition around a gas turbine aerofoil surface. Reproduced from Grimethorpe Topping Cycle
Project, UK DTI Clean Coal Programme, Summary Report 1995.

condition of the components, these particles may temperature gradient) are well known but are
stick to the component surface, bounce off (or par- believed to be less probable in the environments
tially stick). It is quite likely that the composition of present within a gas turbine.
the particles will vary with size, so this may cause It is clear from the brief descriptions of the differ-
the behavior of larger and smaller particles to differ, ent types of deposition processes above that the
as well as the differences in impact conditions. deposition mechanisms that will operate will depend
 Vapor deposition; this transfers trace elements on a large number of potential variables that include
from the bulk gas stream to the surface of compo- ones resulting from (a) the composition of the com-
nents; however, there are several variations of this busted gas flow and (b) gas turbine component opera-
type of process which include: tions/geometry.
 condensation of a vapor of a particular species Combusted gas stream factors (local to area of
from a hot gas stream onto a cooler metal sur- deposit formation):
face (e.g., gaseous sodium sulfate condensing
 trace metal contaminant levels,
out as solid or liquid sodium sulfate);
 condensation of a vapor of a particular species  SO2/SO3 levels,
 particle sizes,
from a hot gas stream into a solid or liquid of
 particle composition.
another species on a cooler surface (e.g., gaseous
sodium hydroxide or sodium chloride condens- Gas turbine component operations and geometri-
ing out as solid or liquid sodium sulfate); reac- cal factors (local to area of deposit formation):
tion with SO2/SO3 could happen during the
condensation process or on the component sur-  bulk gas temperatures,
face after condensation;  component temperatures,
 these can be complex processes that depend on  difference between bulk gas and metal temperatures,
the activities and relative stabilities of all the  gas pressure,
contaminants involved in the gas stream and at  component shape and size, affecting:
the surface of the components, as well as the  gas flows around component,
temperature differences between the bulk gas  development of eddy flows,
stream and component surface, system pressure,  development and thickness surface boundary
boundary layer thickness, and other flow related layers.
parameters.
The source of the contaminant can alter the form
Other deposition processes such as Brownian diffu- in which it enters the gas turbine, passes through it
sion or thermophoresis (i.e., diffusion along a and deposits onto components in the power turbine.
534 High Temperature Corrosion Environments

For example, sodium chloride can enter the system  Pressurized combustion systems (such as pressur-
from the fuel or combustion air, pass through the ized fluidized bed combustion) in which solid fuels
combustor as a vapor or as particles (or both), and are burnt, and then the hot pressurized gas stream
then deposit via direct inertial impaction, eddy produced cleaned and passed through the power
impaction, and/or vapor condensation on different turbine part of the gas turbine (in some cases, the
areas of an aerofoil. compression part could be used to produce the
For a given gas turbine design, only factors related pressurized air stream needed for the combustion
to the combusted gas stream composition may be open process);
to an operator to control (and so minimize deposition  Use of natural gases with increasingly high
and hot corrosion). Gas turbine manufacturers, some H2S levels (beyond the range of previously accept-
gas turbine operators, and standards bodies have all able gas compositions);
considered the potential damage that can be caused to  Natural gases or syngases are processed to reduce
gas turbines by fuel contaminants. As a result, a num- their carbon contents and enhance their hydrogen
ber of fuel specifications have been produced that contents;
include limitations on elements related to hot corro-  Bio-oil fuels;
sion (the fuel standards also include limitations on  Gas turbines fired using oxygen and any of the
many other fuel properties for other purposes). Exam- other potential fuels to produce a combusted gas
ples of such fuel specifications are given in Table 3. stream of mostly CO2 and steam, from which the
All these specifications have the aim of reducing the steam can be condensed to produce a CO2 rich gas
damage caused by contaminants to levels acceptable to stream for recycling into the compressor or for
gas turbine manufacturers or users. storage (after further cleanup).
Along with controlling the contaminants entering
the gas turbine hot gas path via the fuel, for many gas All these systems can be expected to have similar
turbines (land and marine based systems), it is also contaminants (e.g., S, Cl, alkali metals, Ca, Pb, V) to
possible to limit the contaminants entering the system those traditionally found in the various oil or natural
via the combustion air (using air filters to remove gas systems fired on air. However, both the absolute
particles and aerosols). For aeroengines, the contami- and relative levels of these contaminants will be
nants entering the system via the air intakes depend on different in each of the different systems, as a result
the types of flight path, its height, and location in the of different fuels being used in them. In addition,
world: for example, for civil jet engines, island hopping other trace metals (e.g., Cd, Sb, Zn), which may give
with short flights at low altitudes over the sea versus rise to vapor phase species when different fuels
long haul flights with long periods at high altitudes.63 are gasified, may affect the deposition and hot cor-
In addition, it should also be recognized that con- rosion processes in the gas turbines that use the fuel
taminants can be accidentally put onto gas turbine gases produced.49,50,60,64 Coal, biomass, and waste
components via cleaning fluids, greases, and other products have significantly different compositions
lubricants used during component manufacturing, to each other, and these variations persist to differ-
inspection and/or assembly. However, every effort ing degrees for each element/fuel utilization process
should obviously be made to avoid this deposit for- by the time the fuel gas streams reach the gas
mation route wherever possible. turbine.
Fuel specifications have been derived mainly from Thus, new gas turbine fuel specifications will
experience with traditional gas turbine fuels such as be needed for new types of fuels that are being con-
diesels, jet fuels, heavy fuel oils and natural gas. How- sidered. These are currently under active consi-
ever, in recent years, economic and environmental/ deration and in some cases research.14,49,50,60,62
regulatory pressures have encouraged the develop- However, in considering the use of gas turbines for
ment of power systems that use a wider range of new applications, it is necessary to consider both the
fuels, as well as new types of power systems3,14,60,64–66: composition of the combusted gas stream expected in
the power turbine and detailed design of gas turbine as
 Integrated gasification combined cycles (IGCC) both should be optimized for the new gas turbine
in which solid fuels such as coal, biomass, and applications, to minimize deposition and subsequent
waste products are gasified, and then the fuel gases hot corrosion (as well as allowing for new bulk gas
(or syngases) produced cleaned up and passed into a compositions in some of the novel power generating
gas turbine for combustion and power generation; cycles).
Gas Turbine Oxidation and Corrosion 535

1.20.5.3 Damage Locations corrosion in different areas of components. It should


be noted that as operating temperatures are slowly
The potential locations for hot corrosion along the
being raised, potential areas for damage are not
gas turbine hot gas path are controlled by deposition
restricted to aerofoil surfaces traditionally consid-
processes (outlined in Section 1.20.4) and the heat
ered. In addition, platforms, shrouds, roots, and
transfer processes between bulk gas stream and var-
shanks may also develop hot corrosion damage as
ious components that enclose it. Heat transfer pro-
may internal cooling air passages. In future advanced
cesses (via radiation, convection, and conduction
gas turbines, one may expect that even outer parts of
mechanisms) cause unique distributions of metal
discs and seals may also suffer from oxidation or
temperatures to be set up on each type of compo-
corrosion damage as the gas temperatures and com-
nent; for example, the vanes and blades at each stage
ponent operating temperatures are raised further.
of any turbine will have different temperature dis-
tributions. The air-cooled vanes/blades in the
higher pressure stages may have wider temperature 1.20.5.4 Rates of Hot Corrosion Attack and
distributions than the ‘uncooled’ vanes/blades in Modeling Hot Corrosion Processes
lower pressure stages. However, the temperature
The rates of hot corrosion attack on a component
profiles on cooled vanes/blades depend on the cool-
depend on the alloy and coating used as well as on
ing methods used (e.g., convection cooling, film
exposure conditions (including local environment).
cooling), details of the blade designs, and whether
As a result of the potential for hot corrosion to be
TBCs are used. Thus, the surface temperature dis-
a life-limiting failure mechanism for some hot gas
tributions on all components along a gas turbine hot
path components within gas turbines, there have been
gas path are unique (and are often commercially
many studies of the mechanisms of hot corro-
sensitive). Figure 17 gives examples of predicted
sion damage in gas turbines over the years.2,15,38–45
temperature distributions on a cooled and an
Failed or damaged components from operating gas
uncooled turbine aerofoil surface.
turbines have confirmed that the various potential
It is quite possible for cooled blades or vanes in the
hot corrosion mechanisms are all possible in practice
higher pressure stages to have the combination of
and have enabled these potential problems to be well
surface temperature and deposition conditions to
characterized. Thus, operating experience and
cause oxidation as well as type I and type II hot
research studies have identified the main factors
that will influence the rates of hot corrosion; in no
particular order, these are2,15,38–45:
 metal temperature,
 gas temperature,
 gas pressure,
 rate of deposit formation (or deposition flux),
 deposit composition,
 gas composition,
 component geometry,
 alloy composition,
 coating composition,
 exposure time (hot corrosion damage goes through
an incubation stage before reaching the character-
istic propagation stage).
Some of these factors relate more to deposition
processes than hot corrosion processes, but are nec-
essary to set up conditions needed for hot corrosion
to proceed.
(a) Cooled aero-foil (b) Uncooled aero-foil Unfortunately, each type of component in every
Figure 17 Examples of temperature distributions on (a)
gas turbine operating with different fuels provides a
cooled and (b) uncooled gas turbine aerofoil surfaces. different combination of exposure conditions. Thus
Courtesy of Siemens Industrial Turbo-Machinery Ltd. carrying out systematic studies on all the different
536 High Temperature Corrosion Environments

potential variables is not a viable option. However, Such studies have confirmed that base alloys con-
information is available from: taining higher chromium (e.g., IN939, IN6203DS,
IN738LC, and FSX-414) tend to perform well under
 manufacturers and operators of fleets of gas tur-
hot corrosion conditions. Such materials are still used
bines, that use fuels from different sources and
for components in many turbines. However, gas turbines
turbines sited or operated in different locations;
that use high component operating temperatures have
 rainbow trials of components in operating gas
moved to using single crystal superalloys for blades and
turbines;
vanes in the higher pressure stages. These lower chro-
 burner rigs:
mium materials (e.g., CMSX-4) do not have good hot
 high velocity,
corrosion resistance, so they require coating to pro-
 low velocity,
vide them with the required surface protection.14,49,50
 laboratory tests:
Figure 19 gives a qualitative indication of the
 deposit recoat methods,
relative performance of different coating types (dat-
 dean rig,
ing from the late 1980s).69,70 The introduction of
 buried in ash methods.
single crystal materials with poor hot corrosion per-
Moving down the above list, control over some of formance and the differences between type I and
the exposure variables generally increases and their type II hot corrosion mechanisms, that are now
cost decreases, but their relevance to gas turbine known, have complicated this picture.
service also decreases unless care is taken to match Quantitative data on hot corrosion of gas turbine
exposure conditions to particular in-service condi- substrate materials and coatings under well-
tions. In addition it is necessary to ensure that mate- characterized conditions are not frequently reported
rials performance data is produced in a form that is due to the difficulties of carrying out such
useful for comparison with component performance experimental programs and the commercial sensitiv-
determined from use in real environments (i.e., metal ities of the results if they are targeted at particular
loss data rather than weight change data). operating conditions or materials systems.14
Figure 18 illustrates hot corrosion damage found However, over the last 15 years, a series of experi-
for two example gas turbine materials in burner rig mental programs have been carried out at Cranfield
studies carried out to compare the performance of University with the aim of generating systematic data
base alloys and coatings under well-characterized sets of the performance of generic gas turbine mate-
operating conditions. rials (e.g., IN738LC and CMSX-4) under ranges of

1200
1200
1000
Corrosion (μm)

Corrosion (μm)

1000
800
800
600
600 1200 1200
1000 400 1000
)
400 (h
) (h
800 e 800 e
m
Ti m Ti
200 600 200 600
400 400
700 750 200 200
800 850 700 750
800 850
900 Tempera 900 Tempera
tu re (⬚C) ture (⬚C
(a) IN792 )
(b) IN738
Figure 18 Examples of hot corrosion performance of two example materials in burner rig tests as a function of metal
temperature and exposure time.67 (a) IN792 and (b) IN738.
Gas Turbine Oxidation and Corrosion 537

exposure conditions that reflect hot corrosion condi- standards. The datasets generated illustrate some of
tions in industrial and aeroengines that have been the features of hot corrosion damage in gas turbine
accelerated to differing degrees. These datasets are environments (Figures 20 and 21).
in the form of metal losses determined at a number of Figure 20 illustrates three different regimes for
locations to give a cumulative probability distribution hot corrosion damage11,47,48,50:
for the damage observed under specific exposure
conditions. The methods being used for this work  At low deposition fluxes; accelerated oxidation pro-
have been included in EU guidelines for hot corro- cesses are observed and the hot corrosion is still in
sion testing that now form parts of draft ISO an incubation phase after this exposure time (500 h);

Ni
Co NiCrAIY
C
rA
IY
Oxidation resistance

CoNiCrAIY
Pl
at
inu
m
alu
m

Aluminides
ini

CoCrAIY
de
s

High chromium

Corrosion resistance - chromium content


Figure 19 Qualitative indication of the relative performance of different coating types.27,68–70

600
100 vppm SOx, 100 vppm HCl
‘Maximum’ corrosion damage (μm)

500 3000 vppm SOx, 300 vppm HCl

‘Buried in ash’
400

300 Flux limited


corrosion
‘Protective’
200 scale formation
(incubation)

100

0
0.1 1 10 100 1000
Alkali sulphate deposition flux (μg cm–2 h–1)

Figure 20 Dependency of type II hot corrosion damage as a function of deposition flux for IN738LC. Reproduced from
Simms, N. J.; Nicholls, J. R.; Oakey, J. E. In Lifetime Modelling of High Temperature Corrosion Processes; Schütze, M. et al.
Eds.; Maney, 2001; pp 379–397, EFC No. 34.
538 High Temperature Corrosion Environments

100
μg/cm2/h
Flux 1.5 ug/cm2/h
90 μg/cm2/h
Flux 5.0 ug/cm2/h
CMSX-4 μg/cm2/h
Flux 15.0 ug/cm2/h
(1.5 mg cm–2 h–1) 80 μg/cm2/h
Flux 1.5 ug/cm2/h
μg/cm2/h
Flux 5.0 ug/cm2/h
70 μg/cm2/h
Flux 15.0 ug/cm2/h

Probability [ % ]
60
50
40
30
20
10
0
0 200 400 600 800 1000
Incubation time (h)
Figure 21 Dependence of type II hot corrosion incubation times on deposition flux for CMSX-4 and IN738LC.
Reproduced from Nicholls, J. R.; Simms, N. J.; Encinas-Oropesa, A. In Materials Issues Governing the Performance of
Advanced 21st Century Energy Systems; Levi, T. et al. Eds.; Science Reviews, 2008; pp 35–48.

 At intermediate deposition fluxes; corrosion rates process. In terms of hot corrosion, there are several
are proportional to deposition fluxes; issues that are believed to be of increasing interest in
 At high deposition fluxes; corrosion rates are not the near future3,14,60,64–66:
sensitive to deposition flux as it is no longer the  Corrosion of internal blade cooling passages (from
rate limiting process in hot corrosion reactions
contaminants in the cooling air);
(this is equivalent to previous and older ‘buried  Deposit-induced corrosion damage below blade
in ash’ test methods).
platforms/on disc rims (due to increasing compo-
Figure 21 illustrates the much shorter incubation nent temperatures);
times found for the single crystal material CMSX-4  Much higher SOx levels than traditionally consid-
when compared to that of the older alloy IN738LC.48 ered (from dirtier/cheaper fuels) modifying type
In addition, the effect of increasing deposition flux on the I and type II hot corrosion reactions;
reduction of incubation times for the IN738LC is shown.  Gasifier fuel gas derived deposits with calcium,
In such laboratory tests, the exposure conditions magnesia, alumina, silica, and iron compounds;
have been adjusted to accelerate the hot corrosion  Effects of novel environments (e.g., high steam
damage relative to that expected in operating gas levels from H2 firing, or high CO2/steam mixtures
turbines. However, by determining the sensitivity of from oxy-fired systems) combined with traditional
the damage to changes in exposure conditions, it is gas turbine degradation issues;
possible to determine the expected degree of accel-  Interaction between mechanical properties (e.g.,
eration linked to specific operating conditions and fatigue) and hot corrosion;
then compare the data generated to that observed in  Development of single crystal alloys with composi-
practice under service conditions. tions that produce higher hot corrosion resistance.
The overall data set being generated by this work
is being used to produce empirical-mechanistic based
models of hot corrosion damage to assist with the
1.20.7 Summary
assessment and prediction of component lives within
the hot gas paths of gas turbines.
Gas turbines are currently used in a wide variety of
applications: from jet engines to large-scale power
1.20.6 Current/Future Issues generation. Each of these applications presents its
own particular set of challenges for the use of specific
The development of gas turbine technologies for gas turbine designs. However, many of these systems
many different potential applications is a continuing experience damage related to oxidation and hot
Gas Turbine Oxidation and Corrosion 539

corrosion in the hot gas paths of gas turbines. These 14. Simms, N. J.; Bale, D. W.; Baxter, D.; Oakey, J. E. In
Materials for Advanced Power Engineering 2002,
various forms of damage are related to a combination Lecomte-Beckers, J. et al. Eds.; Jülich, 2002; p 73, 675–690.
of operating conditions of the particular gas turbine 15. Birks, N.; Meier, G. H.; Petit, F. S. Introduction to the High
and its components together with contaminants Temperature Oxidation of Metals; Cambridge University
Press: Cambridge, 2006.
entering hot gas paths via air or fuel streams. The 16. Schütze, M. et al. Eds. Lifetime Modelling of High
various forms of oxidation and hot corrosion that are Temperature Corrosion Processes; Maney, 2001, EFC
expected under such exposure conditions are well No. 34.
17. Nijdam, T. J.; van der Pers, N. M.; Sloof, W. G. In Novel
known and are characterized with qualitative descrip- Approaches to Improving High Temperature Corrosion
tion of the potential mechanisms developed mostly Resistance; Schütze, M. and Quadakkers, W. J., Eds.;
during the 1970s. However, quantitative data on the Woodhead, 2008; pp 582–598, EFC No. 47.
18. Smialek, J. L.; Nesbitt, J. A.; Barrett, C. A.; Lowell, C. E. In
various damaging reactions are less readily available, Cyclic Oxidation of High Temperature Materials;
and models to enable the prediction of such damage Schütze, M. et al. Eds.; IOM Communicvations, 1999,
are subjects of current research activity. The use of Chapter 9, EFC No. 27.
19. Nicholls, J. R.; Newton, R.; Bennett, M. J.; Evans, H. E.;
new materials designed for strength rather than cor- Al-Badairy, H.; Tatlock, G. J.; Naumenko, D.;
rosion resistance is proving to be problematic for Quadakkers, W. J.; Strehl, G.; Borchardt, G. In Lifetime
some gas turbine applications. The on-going wish to Modelling of High Temperature Corrosion Processes;
Schütze, M. et al. Eds.; Maney, 2001; pp 83–106, EFC No. 34.
increase gas turbine efficiencies and at the same time 20. Nicholls, J. R. In Life Assessment of Hot Section Gas
widening the range of potential fuels and including Turbine Components; Townsend, R. et al. Eds.; IOM
new types of power generation system, to meet envi- Communications, 2000; pp 135–154.
21. Poquillon, D.; Monceau, D. Oxid. Met. 2003, 59, 409.
ronmental and regulatory wishes, is presenting a new 22. Nesbitt, J. A. In Lifetime Modelling of High Temperature
series of challenges in terms of the oxidation and hot Corrosion Processes; Schütze, M. et al. Eds.; Maney,
corrosion performance of components in various gas 2001; pp 359–378, EFC No. 34.
23. Dahl, K. V.; Hald, J. In Materials for Advanced Power
turbine hot gas paths. Engineering 2006; Lecomte-Beckers, J. et al. Eds.; Jülich,
2006; pp 675–690.
24. Bertrand, N.; Desgranges, C.; Nastar, M.; Girardin, G;
References Poquillon, D.; Monceau, D. Mater. Sci. Forum 2008,
595–598, 463–472.
25. Bacos, M. P.; Josso, P.; Vialas, N.; Poquillon, D.;
1. Rolls Royce: The jet engine. Rolls Royce, 2005. Pieraggi, B.; Monceau, D.; Nicholls, J. R.; Simms, N.;
2. Sims, C. T.; Stoloff, N. S.; Hagel, W. C. Eds. Superalloys; Encinas-Oropesa, A.; Ericsson, T.; et al. Appl. Therm. Eng.
Wiley: New York, USA, 1987; Vol. II. 2004, 24, 1745–1753.
3. Hannis, J. M. In Power Generation in an Era of Climate 26. Simms, N. J.; Kilgallon, P. J.; Roach, C.; Oakey, J. E. In
Change; Proceedings of the 7th International Charles Microscopy of Oxidation, Proceedings of 5th International
Parsons Turbine Conference; Strang, A. et al. Eds.; IOM Conference, Newcomb, S. and Tatlock, G. Eds.; Science
Communications, 2007; pp 37–49. Reviews, 2003; pp 273–280.
4. Data sheet for SGT5-8000H; Siemens AG, 2007. 27. Nicholls, J. R. Mater. Res. Soc. Bull. 2003, 28(9),
5. Pavri, R.; Moore, G. D. General Electric Reference 659–670.
Document, GER-4211, 2001. 28. Newcomb, S. and Tatlock, G. Eds.; Microscopy of
6. De Backer, W.; Bopp, C. In Proceedings of ASME Oxidation; Proceedings of 5th International Conference
TURBOEXPO 2000, Munich, Germany 2000; Paper Science Reviews, 2003.
2000-GT-0307. 29. Yanar, N. M.; Kim, G.; Hamano, S.; Petit, F. S.; Meier, G. H.
7. Ingistov, S. Proceedings of ASME TURBOEXPO 2000, In Microscopy of Oxidation, Proceedings of 5th
Munich, Germany, 2000; Paper 2000-GT-305. International Conference, Newcomb, S. and Tatlock, G.,
8. Horlock, J. H. Proceedings of ASME TURBOEXPO 2001, Eds.; Science Reviews, 2003; pp 249–260.
New Orleans, Louisiana, 2001; Paper 2001-GT-0343. 30. Osgerby, S.; Rinaldi, C.; De Maria, L. In Materials for
9. Bhargava, R.; Meher-Homji, C. B. In Proceedings of ASME Advanced Power Engineering 2006; Lecomte-Beckers, J.
TURBOEXPO 2001, Amsterdam, The Netherlands 2002; et al. Eds.; Jülich, 2006; pp 217–232.
Paper 2002–30560. 31. Renusch, D.; Schütze, M. Mater. Sci. Forum 2008,
10. Bradshaw, A.; Simms, N. J.; Nicholls, J. R. Fuel 2008, 87, 595–598, 151–158.
3529–3536. 32. Busso, E. P.; Lin, J.; Sakurai, S.; Nakayama, M. Acta
11. Simms, N. J.; Nicholls, J. R.; Oakey, J. E. In Lifetime Mater. 2001, 49, 1515.
Modelling of High Temperature Corrosion Processes; 33. Busso, E. P.; Evans, H. E.; Wright, L.; McCartney, L. N.;
Schütze, M. et al. Eds.; Maney, 2001; pp 379–397, EFC Nunn, J. E.; Osgerby, S. Mater. Corros. 2008, 59(7),
No. 34. 556–565.
12. ADSEALS – Investigation in advanced high temperature 34. Hermosilla, U. PhD Thesis, Nottingham University 2008;
seals, European Union FP5 project, G4RD-CT-2000- UK EPSRC Supergen Plant Life Extension Consortium.
00185, 2004 35. Nicholls, J. R. DARP: Advanced Aero-engine Materials
13. Rickerby, D. S.; Hicks, M. A. In Power Generation in an Era (ADAM) – Next Generation Single Crystal and Lifing
of Climate Change, Proceedings of the 7th International Methodologies; GR/S26149/01, Final Report, EPSRC, 2007.
Charles Parsons Turbine Conference; Strang, A. et al. 36. Padture, N. P.; Gell, M.; Jordan, E. H. Science 2002, 296,
Eds.; IOM Communications, 2007; pp 229–240. 280–284.
540 High Temperature Corrosion Environments

37. He, M. Y.; Hutchinson, J. W.; Evans, A. G. Mater. Sci. Eng. 57. Fackrell, J. E.; Tabberer, R. J.; Young, J. B. In The Impact
A 2003, 345, 172–178. of Ash Deposition in Coal Fired Plants; Williamson, J. et al.
38. Kofstad, P. High Temperature Corrosion; Springer, 1988. Eds.; Taylor and Francis, 1994; pp 123–134.
39. Giggins, C. S.; Pettit, F. S. Hot Corrosion of Metals and 58. Hepworth, J. K.; Fackrell, J. E.; Pinder, L. W.; Wilson, J. D.
Alloys – A unified Theory, PWA Report No. FR-11545, 1979. In Life Assessment of Hot Section Gas Turbine
40. Corrosion, 9th ed.; ASM International: Ohio, USA, 1987; Components, Townsend, R. et al. Eds.; IOM
Vol. 13, ISBN 0-87170-007-7. Communications, 2000; pp 281–292.
41. Stringer, J.; Wright, I. G. Oxidation Metals 1995; 44(1/2), 59. Simms, N. J.; Oakey, J. E.; Stephenson, D. J.; Smith, P. J.;
265–308. Nicholls, J. R. In Proceedings 8th International Conference
42. Special Issue on Hot-salt corrosion standards, test on Erosion by Liquid and Solid Impact; Wear 1995;
procedures and performance, High Temp. Technol. 186–187(Part 1), 247–255.
1989, 7(4). 60. DeCorso, M.; Newby, R.; Anson, D.; Wenglarz, R.;
43. Rapp, R. A. Corros. Sci. 2002, 44, 209–221. Wright, I. Coal/biomass fuels and the gas turbine:
44. Rapp, R. A.; Goto, K. S. In Symposium on Fused Salts; Utilization of solid fuels and their derivatives; ASME,
Braunstein, J., Selman, J. R., Eds.; The Electrochemical Paper 96-AT-76, 1996.
Society, 1979; p 159. 61. Fackrell, J. E.; Tabberer, R. J.; Young, J. B.; Wu,
45. Shores, D. A. In High Temperature Corrosion; Rapp, R. A., Z. Gas turbines fired on coal-derived gases –
Ed.; NACE, 1983; p 493. Modelling of particulate and vapour deposition, UK
46. Foster, A. D.; von Doering, H. E.; Hilt, M. B. Fuels Flexibility DTI Report No. COAL R280, DTI/Pub URN 05/661,
in Heavy Duty Gas Turbines; GE: Schenectady, NY, 1983; March 2005.
GER3428a. 62. Grimethorpe Topping Cycle Project, UK DTI Clean Coal
47. Simms, N. J.; Nicholls, J. R.; Oakey, J. E. Mater. Sci. Programme, Summary Report 1995.
Forum 2001, 369–372, 833–840. 63. Leggett, A. J.; Simms, N. J.; Rickerby, D. S. In Power
48. Nicholls, J. R.; Simms, N. J.; Encinas-Oropesa, A. In Generation in an Era of Climate Change, Proceedings of
Materials Issues Governing the Performance of Advanced the 7th International Charles Parsons Turbine Conference;
21st Century Energy Systems; Levi, T. et al. Eds.; Science Strang, A. et al. Eds.; IOM Communications, 2007; pp
Reviews, 2008; pp 35–48. 241–251.
49. Bordenet, B.; Bossman, H. P. Mater. Sci. Forum 2004, 64. Wenglarz, R.; Oakey, J. E. In Materials for Advanced
461–464, 907–916. Power Engineering 2006, Lecomte-Beckers, J. et al. Eds.;
50. Simms, N. J.; Encinas-Oropesa, A.; Nicholls, J. R. Mater. Jülich, 2006; pp 233–244.
Sci. Forum 2004, 461–464, 941–948. 65. Ruth, L. A. In Life Cycle Issues in Advanced Energy
51. Phase Diagrams for Ceramists; American Ceramic Systems, Norton, J. F. et al. Eds.; Science Reviews, 2003;
Society, 1964. pp 7–14.
52. Hamed, A.; Tabakoff, W.; Wenglarz, R. J. Propul. Power 66. Paisley, M. A.; Welch, M. J. In ASME Turbo Expo 2003;
2006, 22(2), 350–360. ASME, 2003; GT2003-38294.
53. Young, J. B.; Wu, Z. Particulate and vapour deposition in 67. High Temperature Corrosion of Aerospace Alloys,
gas turbines fired on coal-derived gases, Final report DTI AGARD-AG-200, 1975.
Cleaner Coal R&D Programme. Project 142, 2004. 68. Coatings for High Temperature Structural Materials,
54. Foster, A. D.; et al. Fuels flexibility in heavy-duty gas NMAB report of the NRC; National Academic Press:
turbines, General Electric Reference Document, Washington DC, 1996.
GER-3428a; GE: Schenectady, NY, 1983. 69. Mom, J. A. NLR Report MP 81003U, Amsterdam, 1981.
55. Wolf, J.; Werning, H. J. Proceedings of VGB Conference 70. Novak, R. C. Coating development and use: Case studies.
on ‘Corrosion and Corrosion Protection in Power Plant Presentation to the Committee on Coating for High-
Technology 1995 (VGB, 1995), paper 27. Temperature Structural Materials; National Materials
56. Fackrell, J. E.; Tabberer, R. J.; Young, J. B.; Fantom, I. R. Adisory Board; National Research Council: Irvine;
ASME International Gas Turbine & Aeroengine Congress, California, April 18–19, 1944. Cited in references [27]
The Hague, 1994; Paper 94-GT-177. and [68].

Похожие интересы