8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
1/24
INSTITUTE OF PHYSICS PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS
J. Phys. D: Appl. Phys. 39 (2006) R101R124 doi:10.1088/0022-3727/39/6/R01
TOPICAL REVIEW
Elevated temperature erosive wear ofmetallic materials
Manish Roy
Defence Metallurgical Research Laboratory, PO Kanchanbagh, Hyderabad 500 258, India
Received 6 July 2005, in final form 9 December 2005Published 3 March 2006Online at stacks.iop.org/JPhysD/39/R101
AbstractSolid particle erosion of metals and alloys at elevated temperature isgoverned by the nature of the interaction between erosion and oxidation,which, in turn, is determined by the thickness, pliability, morphology,adhesion characteristics and toughness of the oxide scale. The mainobjective of this paper is to critically review the present state ofunderstanding of the elevated temperature erosion behaviour of metals andalloys. First of all, the erosion testing at elevated temperature is reviewed.This is followed by discussion of the essential features of elevatedtemperature erosion with special emphasis on microscopic observation,giving details of the erosionoxidation (EO) interaction mechanisms. TheEO interaction has been elaborated in the subsequent section. The EOinteraction includes EO maps, analysis of transition criteria from one
erosion mechanism to another mechanism and quantification of enhancedoxidation kinetics during erosion. Finally, the relevant areas for futurestudies are indicated.
Nomenclature
E Erosion rate
En Erosion rate at nth exposure
Mn Mass loss suffered due to erosion on nth exposure
Mn Mass gain experienced due to oxidation on
nth exposuret Time of exposure to eroding conditions
mn Cumulative mass of erodent for nth exposure
V Impact velocity
K1 Constant
p Velocity exponent of erosion rate
Kop Parabolic rate constant
m Incremental mass gain due to oxidation
Ao Arrhenius constant
Q Activation energy for oxidation
R Universal gas constant
T Absolute temperature
Z Thickness of the oxide scale
C ConstantKp Scaling constant
o Density of the oxide
F Particle flux rate
L Depth to which the deformed zone extends
W Indentation diameter
Constant
H Hardness
U Crater volume
mp Mass of single erodentr Radius of the erodent
N Number of erodents
Semi included angle of conical erodent
tb Time between impacts
tim Time of impact
Zb Thickness of the oxide scale grown between two
successive impacts
1. Introduction
Erosive wear or solid particle erosion is sometimes known
as impact wear. Solid particle erosion is defined as material
degradation due to the impact of particles travelling withsome significant velocity. It is mechanistically different from
other forms of erosion such as liquid impact erosion, slurry
0022-3727/06/060101+24$30.00 2006 IOP Publishing Ltd Printed in the UK R101
http://stacks.iop.org/JPhysD/39/R101http://stacks.iop.org/JPhysD/39/R101http://dx.doi.org/10.1088/0022-3727/39/6/R018/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
2/24
Topical Review
Table 1. Examples of systems subjected to elevated temperature erosion.
System Components References
Combustion systems Burner nozzles, reheater, super heater, [5,6]economizer tube banks, boiler heat exchanger,in bed tubes, tube banks, etc
Coal gasification systems Turbine, lock Hopper valves [7]Coal liquefaction system Valve to throttle the flow of product stream [7]Gas turbines Blades [8]
erosion and cavitation erosion, etc. A tribosystem suffering
from erosive wear can be characterized as an open system.
In such a system, the counterbody is continuously replaced.
Another important feature of solid particle erosion is that
the wear of the counterbody is completely uninteresting. In
solid particle erosion, the contact time between the erodent
and the target material is only momentary. In this respect,
erosion is different from other related processes such as sliding
wear, abrasive wear, grinding and machining, etc, in which thecontact between the tool/abrasive and the target/work piece is
continuous.
Room temperature erosion is an important problem
in several engineering applications. Rocket motor trail
nozzles [1], the engine of a helicopter operating in dusty
terrain [2], equipment in oil and mining industries [3, 4],
etc are subjected to solid particle erosion at ambient
temperature. Several engineering components are degraded
dueto solid particle erosion at elevated temperature. A number
of industrial systems, which undergo erosion at elevated
temperature, are summarized in table 1 [58]. At the same
time, it should be mentioned that erosion could also be used
constructively, as in the case of shot peening, sand blasting,mining, rock drilling and cutting applications, etc [911].
Erosion behaviour of metallic materials at room
and elevated temperatures has been reviewed in several
publications [1220]. The information contained in these
reviews will be repeated only to the extent of impressing upon
the readers the essential features of the erosion processes. It
is not the purpose of this work to provide a comprehensive
review of the status of research in the field of solid particle
erosion at elevated temperature; rather the aim is to discuss
and review some of the recent results which have enhanced
our understanding in the areas of elevated temperature erosion
of metallic materials. This review will be confined only to
metals and alloys. Elevated temperature erosion of ceramics,glasses and composites will not be considered here.
For the purpose of convenience, the review is divided into
several sections. Section 2 consists of elevated temperature
erosion tests. The salient features of elevated temperature
erosion constitute section 3. Examination of the eroded
surfaces is consideredin section 4. Section 5 dealswitherosion
oxidation interaction. Erosion enhanced oxidation kinetics
is analysed in section 6. Future research areas for elevated
temperature erosion arediscussed in section 7. Thisis followed
by concluding remarks in section 8.
2. Elevated temperature erosion test
Over the last few decades several test techniques or
methodologies have been developed for studying mechanisms
and assessing the extent of erosion. These tests can broadly
be divided into two categories, namely (1) those simulative
tests that are designed to simulate a specific type of erosion
and (2) those that are intended to be used for fundamental
studies. The main problem of the simulative test is that it is
very expensiveandit is difficult toconduct fundamental studies
for material development or for understanding the mechanisms
of erosion. In order to avoidtheseproblems, various laboratory
tests have been developed.The most common laboratory test involves blasting a
stream of airborne particles against the target as standardized
by G 76-83 [21]. In this type of test, a known quantity
of erodent is fed into an air stream, accelerated through a
converging nozzle and directed towards the test specimen. A
cleaned andweighed sampleis exposed to theparticle-laden air
streamfora predetermined time andweighed after interrupting
the test. The ratio of the weight loss suffered by the sample to
the weight of erodent gives the dimensionless erosion rate.
The air jet-type elevated temperature erosion rigs can be
classified into two broad groups. The first group comprises
erosion rigs, which are designed in such a fashion that both
the fluid stream carrying the particles and the eroding target
are heated to the same test temperature. These are called
isothermal erosion rigs. On the other hand, the second group
of erosion rigs, called non-isothermal rigs, have the facility to
heat the target material alone, while the fluid stream with the
particles is not preheated before allowing it to enter the erosion
chamber. In these types of rigs, the colder fluid stream cools
the target materials to some extent on impact. Nevertheless,
the target material still attains a steady-state temperature. The
above classification of the erosion rig as isothermal and non-
isothermal is largely artificial for reasons stated below.
1. The test specimen attains a constant temperature and
remains at that temperature throughout the test in both
types of rigs.2. The erodents are not heated and the cold particles impinge
thetarget material. However, unlikein thecase ofabrasion
or sliding wear, the contact between the particles and the
target materials in the case of solid particle erosion is
momentary, and hence negligible transfer of heat takes
place between the particles and the substrate. Thus, it is
immaterial as to whether theparticlesarepreheatedor not.
The non-isothermal erosion rigs are easy to fabricate but
they fail to simulate the erosion conditions. Hence, such rigs
are not popular at present. In contrast, isothermal type erosion
rigs can simulate erosion conditions but they are difficult to
fabricate. The schematic diagram of one such erosion rig,
fabricated by the author at DMRL, is shown in figure 1. Theunique feature of this rig is its ability to alter the particle
feed rate by over 100 times. Its particle feeding system is
R102
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
3/24
Topical Review
Air Heating
System
Temperature
Controller
For Air Heating
System
Air Pressure Indicator
Air Flow Rate IndicatorFluidized
Chamber
Test
Chamber
Sample
HoldingDevices
Particle Feeding System
Temperature
Indicator of
Hot Air
Temperature
Indicator of
The Sample
Outlet
Inlet
Air Heating
System
Temperature
Controller
For Air Heating
System
Air Pressure Indicator
Air Flow Rate IndicatorFluidized
Chamber
Test
Chamber
Sample
HoldingDevices
Particle Feeding System
Temperature
Indicator of
Hot Air
Temperature
Indicator of
The Sample
Outlet
Inlet
Temperature
Controller of the
Test Chamber
Figure 1. Schematic representation of jet type elevated temperature erosion rig [67].
a miniaturized conveyer belt system and particle feed rate is
controlled by controlling the speed of the motor of the system.
Further description of this rig is available elsewhere [22].
The test procedure involves heating the compressed air
to the required temperature and then heating and soakingthe test sample to that temperature. The heated samples are
then exposed to compressed, heated, fluidized and accelerated
air streams carrying the particles. The elevated temperature
erosion test is a multiple specimen test procedure. Cleaned,
dried and weighed samples are exposed to erodents for various
time intervals (say t1, t2, t3, . . . , where t1 < t2 < t3)
corresponding to various mass of erodents (m1, m2, . . . , mn).
A similar number of samples are again exposed to an air stream
without carrying particles. IfM1, M2, . . . , M n represent the
mass loss suffered by n samples when exposed to erodents for
time intervals oft1, t2, . . . , t n and ifM
1, M
2, . . . , M
n are the
massgainsexperienced by n samples when exposed to theplain
air stream without erodents for time intervals of(t1, t2, . . . , t n)then the incremental erosion rate E1, E2, . . . , En can be
computed as
En =(Mn M
n) (Mn1 M
n1)
mn mn1. (1)
This procedure is repeated until En1 is equal to En2 and
this E is considered to be the incremental erosion rate.
The main problem of the jet type of erosion rig is the
measurement of impact velocity. There are three different
types of velocity measurement techniques available. In the
first method, known as the photographic method, a high-
speed camera is used to photograph the successive positionsof a single particle as a function of time and thus compute
the velocity. The second method, known as the rotating
disc method, was developed by Ruff and Ives [23]. In this
method velocity is determined by estimating the time of flight
of particles between two discs fixed on a common shaft
rotating at some specified velocity. A modified version of
the rotating disc is the paddle wheel technique [24]. Thismethod gives a more reliable and statistically more accurate
velocity.
The third method uses the laser Doppler velocitimeter
(LDV). This technique is an accurate, non-interactive and
on-line velocity measuring device. The LDV uses the well-
known Doppler effect to measure the velocity of the particles.
When light is scattered from a moving object, the stationary
observer will see a change in the frequency of scattered light
proportional to the velocity of the object. A laser is used as a
light source because it is easily focused and is coherent.
In many applications, for example, pipe bends in a slurry
transportation system, the impact is primarily by big particles
with very low impact velocity. To simulate such a system,samples are impacted by dropping particles under gravity.
Such a system was first introduced by Bitter [25], a schematic
presentationof which is shown in figure 2. The system consists
of a ball dispenser unit, the velocity measuring system, ball
counting unit and the sample holder. The ball dispenser and
the sample holder can be moved up and down so as to alter the
height over which the eroding particle falls. Before the steel
ball impacts thesampleit passesthrough a multiple photodiode
unit, which measures the velocity of the passing ball and,
in addition, keeps track of the total number of balls passing
through.
Whirling arm rigs were developed to enable tests to be
carried out at precisely controlled velocity over a range ofimpact conditions. The target specimens are attached to the
tips of the rotor arms and whirled through a certain or a
R103
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
4/24
Topical Review
Ball
Feeder
Timer
Sample
Holder
Velocity MeasuringSystem
Figure 2. Schematic diagram of low velocity erosion rig.
Figure 3. Schematic diagram of whirling arm erosion rig.
narrow band of erosive particles. These rigs are very noisy
and consume considerable power. A whirling arm erosion
rig is schematically shown in figure 3. These kinds of rigs
simulate the degradation conditions prevalent in the fluidized
bed combustor. The whirling arm erosion rig was initially
developed by Tilly and Sage [26]. One of the advantages of
this type of rig is that the erodent velocity can be controlled
precisely as the velocity is governed by the rotating speed of
the arms. It also permits testing to be carried out over a widerange of impact velocities. It makes efficientuse of theerodent
as theentire amountof erodent is delivered to thetarget sample.
3. Salient features of elevated temperature erosionof metallic material
Over the last decade or so, a substantial amount of work on
elevated temperature erosion of metals and alloys has been
carried out [2762]. A compilation of these investigations,providingdetails of materialssubjectedto elevatedtemperature
erosion and the test conditions, is made in table 2. These
tests cover a wide range of test conditions and test materials.
Based on the work of the investigators compiled in table 2, the
important factors which influence the solid particle erosion
behaviour of metallic materials at elevated temperature are
described below.
3.1. Effect of temperature
The variation of erosion rate with temperature for a number
of metals and alloys is depicted in figures 4 and 5 [27, 28].
The erosion data presented in these figures pertain to highimpact velocities and mostly with angular particles. The
observed temperature dependence of the erosion rate can be
classified into three groups. In the first group, the erosion
rate initially decreases with the increase of temperature,
reaches a minimum and then starts increasing with increasing
temperature. Materials such as 5.0Cr0.5Mo, 17-4 PHSS,
410 SS, Alloy 800, Ti-6Al-4V, and tungsten belong to this
group. The second group comprises metals such as Ta and
lead (for oblique impact) andalloyssuch as 310SS (for oblique
impact), 1018 steel and 1100 aluminium (for normal impact)
which exhibit a temperature independent erosion rate up to a
critical temperature followed by an increase of the erosion rate
with increasing temperature. Finally, group three materialsshow a monotonically increasing erosion rate with increasing
temperature. Inco 600, carbon steel, 12Cr1MoV steel, and
2.25Cr1.0Mo steel, lead and 20245 Al are some of the typical
examples in this group.
3.2. Effect of impact velocity
The velocity dependence of erosion rate (E) is characterized
by the velocity exponent, p, given by
E = K1Vp, (2)
where K1 is a constant and V is the impact velocity. The
velocity exponents (p) obtained by various investigatorsare plotted in the velocitytemperature regime in figure 6.
The velocity exponent decreases with an increase in erosion
test temperature for 304 SS to values as low as 0.9 at low
impact velocities. At relatively higher impact velocities p
appears to lie in the range 23. Levy and Man [29, 30]
reported the influence of erodent size on the velocity exponent
for erosion of 9Cr1Mo steel at 923K using angular SiC
particles.
3.3. Effect of impact angle
The available data related to erosion rate and impact angle at
different temperatures are presented in this section. However,most of the metallic materials, irrespective of temperature of
erosion, exhibit a ductile behaviour, i.e. a maximum erosion
R104
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
5/24
Topical Review
Table 2. Compilation of the work of various investigators on elevated temperature erosion of metallic material.
Materials Name of the investigators Test conditions
310 SS, 304 SS, 1018, 2.25Cr1.0Mo, Levy et al [27] Nitrogen gas, 240 m SiC, V = 30ms1,5.0Cr0.5Mo, 410 SS, 17-4 PHSS T = RT to 1173K, = 30 and 90
1100 Al, 310 SS Finnie et al [32] Nitrogen gas, 250 m SiC, V = 3060 m s1,
T = RT to 0.8 Tm, = 30
9Cr1.0Mo, 2.25Cr1.0Mo, 5.0Cr10.5Mo, Levy and Man [29, 30, 47] Air, 130 m SiC and SiO2, V = 35ms1,
304 SS, 410 SS T = 9231123K, = 90 and 30
410 SS Levy et al [45] Air, 5 m fly ash, 130 m Al2O3,T = 1223K, V = 5 m s1, = 45
5.0Cr0.5MoSi Levy and Wang [40,48] Air, 90 m alumina, V = 30, 40 and 70m s1,T = 1123K, = 9
Inco 600 Tabakoff and Vittal [34] Air, 100800 m quartz, V = 60250ms1,T = 544, 666 and 846K, = 1070
Ti6Al-4V, 2074 Al, 410 SS, W, Ta, Pb Gat and Tabakoff [28] Air, 86 m quartz, V = 180300m s1,and 304 SS T = 3001023K, = 1590 also 164 m quartz,
V = 120ms1, T = 300483K, = 20, 60 and 90
Ni, Co Kang et al [50] Nitrogen, air, 20 m alumina, V = 90 and 140m s1,T = 9231073K, = 90
Ni, Co Chang et al [38, 47] Air, 20 m alumina, V = 140, 123 and 70m s1,
T = 8731053K, = 90
, 60
, 30
and 20
AISI 303 Shayler and Yee [31] Air, 4766 m fly ash, V = 150300 m s1,T = 300773K, = 35
304 SS, alloy 800, C-steel, 2.25Cr1.0Mo, Shida and Fujikawa [42] Argon gas, 120 m quartz, V = 40120m s1,12.0Cr1.0MoV T = 673923K, = 2090
304 SS, 316 SS and 410 SS Singh and Sundararajan [43] Air, 160 m SiC, V = 55110ms1,T = 300773K, = 30, 60 and 90
In 738, X 40, MA 754 and HA 188 Barklow et al [51] Burner rig, 20 m alumina, V = 200275 m s1,T = 1148 K
Stellite 6B and 1, In-100, MA-754, HA 8077, Wright et al [52] Argon gas, air, 12 m alumina, V = 43ms1,AISI 446 SS, FeCrAlY, AISI 446 T = 1033K, = 30
9.0Cr1.0Mo, 304 SS Sethi and Wright [53] Air, 1 m alumina, V = 2.7 and 4.3 m s1,T = 7331100K, = 30
304 SS, 416 SS, 430 SS and 17-4 PH SS Zhou and Bahadur [35] Air, 120 grit SiC, V = 65ms1,T = 9231073K, = 30
2.25Cr1.0Mo Sethi and Carey [54] Air, 1 m alumina, V = 2.7 m s1,T = 763863K, = 30
Alloy 800 HT, 310 SS Stott et al [55] Air, 4766 m fly ash, V = 150300 m s1,T = 300773K, = 35
Haynes 188, Waspaloy Chinadurai and Bahadur [56] Air, 150 m SiC, V = 50ms1,T = 3001073K, = 30
Ti6Al-4V Zhou and Bahadur [57] Air, 4766 m fly ash, V = 50ms1,T = 3001073K, = 1090
Carbon steel Xie and Walsh [58] Air, 43 m coal ash, V = 530ms1,T = 423673K
FeCrC cast steel Drotlew et al [59] V = 65ms1, T = 723K, = 1560
Ni, Ni20Cr Manish Roy et al [60] Air, 200 m SiO2, V = 35, 65 and 105m s1,
T = 3001073K, = 30, 45, 60 and 90
Fe3Al base alloy, Ni, Co Yu et al [61] SO2, air, 2050 m SiO2, V = 160190 m s1,
T = 8731073K, = 090
FeCrC Hayashi et al [62] Air, 23 m cold silica, V = 80ms1, T = 973K
Note: T = test temperature, V = impact velocity and = impact angle.
rate at oblique impact angles (1030). The universality
of such an observation is shown in figures 7 and 8,
wherein data from a large number of investigators have been
compiled [3136].
Levy [37] obtained a higher erosion rate at normal impact
than at oblique impact for 9Cr1Mo steel at 1123K using
rounded Al2O3 (130 m) erodent for a range of impact
velocities (3070 m s1), as shown in figure 9. But at a low
impact velocity of 20 m s1, a maximum in the erosion rate
occurred at oblique impact angle. Observations by Chang et al
[33, 38] shown in figure 10 indicate that the peak erosion rateof Co at a test temperature of 1053K occurs at an impact angle
of 60 when impacted with 20 m angular alumina particles at
impactvelocities of 70140m s1. However, when the erosion
test is carried out at 873 K and at impact velocity of 140m s1,
the erosion rate peaks at an impact angle of 30. Thus, there
exist apparently conflicting observations regarding the erosion
rateimpact angle behaviour.
3.4. Effect of particle size
Tabakoff and Vittal [34] carried out erosion tests on Inco 600
alloy using quartz particles having sizes between 70 and
800 m. Their work indicates that the erosion rate increasesmarginally with the increase of particle size, as shown in
figure 11. Zhou and Bahadur [35] investigated the effect
R105
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
6/24
Topical Review
of the particle size of SiC on the erosion rate of 304 SS
at 923 K (impact angle: 30 and impact velocity: 65 m s1).
These results indicate that erosion rate increases with the
increase in particle size up to 40 m and thereafter it becomes
independent of the particle size (figure 12). Levy et al
[29, 30, 37, 39], however, noted only the increase of erosion
rate with the increase of particle size for 9Cr1Mo steel eroded
at a temperature of 923K and that for 1018 steel eroded
at 723 K.
0
20
40
60
80
100
120
0 200 400 600 800 1,000 1,200
Temperature (K)
ErosionRatex
106(
Kg/Kg)
310 SS
304 SS
410 SS
2.25Cr-0.5Mo
17-4 PHSS
1018 Steel
5.0Cr-0.5Mo
Impact Velocity : 30 m/s
Impact Angle : 30o
Erodent : SiC (250 m)Gas : Nitrogen
Figure 4. Variation of erosion rate with temperature for a number ofalloys [27].
20
30
40
50
60
70
80
90
100
110
120
283 303 323 343 363 383 403 423
Temperature (K)
ErosionRatex104(
Kg/Kg)
Lead
Impact Angle : 20o
Impact velocity : 120 m/s
Erodent : Quartz ( 138 - 164 m)
0
20
40
60
80
100
120
140
283 303 323 343 363 383 403 423
Temperature (K)
ErosionRatex104(
Kg/Kg)
Impact Angle : 90o
Impact velocity : 120 m/s
Erodent : Quartz ( 130 m)
Lead
Tungsten
0
1
2
3
4
5
6
7
8
9
10
283 333 383 433 483
Temperature (K)
Erosionratex104
(Kg/Kg)
2024 Al
Ta
Ti-6Al-4V
410 SS
Impact Angle : 20o
Impact velocity : 120 m/s
Erodent : Quartz ( 138 - 164 m)2
2
3
3
4
4
5
283 333 383 433 483
Temperature (K)
ErosionRatex104(
Kg/Kg) Ta
2024 Al
410 SS
Ti-Al-4V
Impact Angle : 90o
Impact velocity : 120 m/s
Erodent : Quartz ( 138 - 164 m)
Figure 5. Effect of test temperature on the erosion rate of a number of metals and alloys [28].
3.5. Effect of particle shape
Levy et al [29, 30, 37, 39, 40] investigated the erosion rate of
a number of Cr containing steels at 1123 K with angular SiC
and spherical Al2O3 as erodent particles. A typical result for
9Cr1Mo steel is presented in figure 13. The erosion rate is
significantly higher when SiC is used as the erodent.
3.6. Effect of particle feed rate
Zhou and Bahadur [41] have investigated the influence of the
particle feed rate on the erosion rate of 304 and 430 SS over
a large temperature range. The reported results as illustrated
in figure 14 suggest that up to a temperature of about 773 K
increasing the feed rate by a factor of 16 has no influence on
the erosion rate. However, beyond 773 K, a lower feed rate
results in a substantially higher erosion rate. Roy et al [60]
also noted a decrease of erosion rate with the increase of the
particle feed rate, especially at low impact velocity.
3.7. Effect of eroded material characteristics
The interpretation of the available data on the effect of eroded
material characteristics on its erosion rate is complicated by
the fact that the behaviour of the oxide scale under erosion
conditions needs to be considered in addition to the behaviour
of the metallic material per se. The oxidation characteristic
of the eroded material plays a more important role than the
mechanical properties of the eroded materials at elevated
temperature.
R106
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
7/24
Topical Review
P values given alongside data points
Figure 6. Velocity exponent obtained by various investigators,plotted in velocitytemperature regime.
0
5
10
15
20
25
30
35
40
45
0 10 20 30 40 50 60 70 80
Impact Angle (Degrees)
ErosionRatex104
(Kg/K
g)
850 K
644 K
766 K
Material : INCO 600
Impact Velocity : 183 m/s
0
1
2
3
4
5
6
10 20 30 40 50 60 70 80 90
Impact Angle (Degrees)
ErosionRatex104(
Kg/Kg)
Material : 303 SS
Impact Velocity : 130 m/s
291 K
1023 K
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
0.050
0 20 40 60 80 100
Impact Angle (Degrees)
MaximumThicknessLoss(m/s)
Test Temperature = 573K
Impact Velocity = 120 m/sErodent = quartz (120 m)
304 SS
Cr-Mo-V
C Steel
0
1
2
3
4
5
6
0 20 40 60 80 100
Impact Angle (Degrees)
ErosionRatex104(
Kg/Kg)
1255 K
293 K
293K
1093 K
558 K372 K
743 K
Impact Velocity = 61 m/sImpact Velocity = 31 m/s
Erodent = SiC (250 m)Fluidizing Gas = Nitrogen
Al
310 SS
ab
cd
Figure 7. Variation of erosion rate with impact angle at various test temperatures: (a) INCO 600 [34], (b) 303 SS [31], (c) CrMoV, carbonand 304 SS [42] and (d) aluminium and 310 SS [32].
Even at elevated temperature, if one considers the erosion
behaviour of metallic material at high impact velocities and
feed rates, the oxidation plays an insignificant role and the
erosion behaviour is essentially metal erosion behaviour. The
experimental data of Shida and Fujikawa [42] pertaining to
1.25 Cr1MoV, 2.25CrMo, 12Cr1MoV and plain carbonsteel (up to 923 K), that of Singh and Sundararajan [36, 43]
pertainingto 304, 316, 410 stainless steel (up to773 K)and that
of Levy et al [44, 45] pertaining to 2.25Cr1.0Mo steel, 5Cr
0.5Mo steel, 1018steel, 304 SS, 310 SS, 410 SSand17-4PHSS
can be considered elevated temperature erosion of metals with
minimum or negligible oxidation. Under such conditions the
dependence of the strength of the material on temperature is a
reasonable indicator of the temperature dependence of erosion
resistance of the materials. Erosion data further indicate that
austenitic stainless steels have superior resistance to elevated
temperature erosion than ferritic steels. The 410 stainless
steel having a tempered martensitic matrix exhibits an erosion
resistance comparable to that of austenitic stainless steels.The oxidation effect is reported to be important for
elevated temperature erosion tests conducted at low impact
velocities and using rounded Al2O3 as an erodent. The
influence of Cr content on the erosion rate of steel is
illustrated in figure 15. It is indicated that the erosion rate
decreases to a very low value when Cr content in the steel
exceeds 1012% [35]. In the case of steel having Cr less than
10%, thick Fe2O3 scale was formed during erosion leading to
R107
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
8/24
Topical Review
0
5
10
15
20
25
30
35
40
45
10 20 30 40 50 60 70 80 90
Impact Angle (Degrees)
ErosionRatex104 (Kg/Kg)
Material : 304 SS
Impact Velocity : 305 m/s
1173 K
873 K
0
2
4
68
10
12
14
16
10 20 30 40 50 60 70 80 90
Impact Angle (Degrees)
ErosionRate
x104(
Kg/Kg) Material : 304 SS
Impact Velocity : 183 m/s1173 K
873 K
0
1
1
2
2
3
3
0 20 40 60 80 100
Impact Angle (Degrees)
ErosionRatex104(
Kg/Kg)
773 K
297 K
Material : 304 SS
Impact Velocity : 70 m/s
4
5
6
7
8
9
10
11
12
13
14
20 30 40 50 60 70 80 90 100
Impact Angle (Degrees)
ErosionRatex104(
Kg/Kg)
316 SS
410 SS
Impact Velocity : 129 m/s
763 K
RT
548 K
a b
c
d
Figure 8. Influence of impact angle and test temperature on erosion rate of (a) 304SS, (b) 304SS, (c) 304SS [35] and (d) 410SS and316SS [43].
high erosion rates [46]. Theexperimental results of Chang etal
[47], presented in figure 16, also indicate the importance of
the nature of the scale that forms during erosion at elevatedtemperature. It is also noted that materials with high scaling
rate such as nickel and cobalt exhibit the highest erosion
rates, while the alumina-forming alloys such as CoCrAlY and
NiCrAlY exhibit intermediate erosion rates. The superiority
of the Al2O3 forming alloy stems from the fact that the Al2O3forming scale forms much more slowly than the Cr2O3 scale.
Extensive work by Levy andco-workers [37,39,44,48,49]
implies that the morphology of the oxide scale that forms
during erosion is important. Segmented scales have a better
erosion resistance than thick, continuous and dense scale since
the spalled area is confined to oxide crystalline only in the case
of erosion of the segmented scale. A striking illustration of the
abovefact is obtained when Si is added to steels. Addition of Sito low chromium steel results in the formation of a segmented
scale even at high impact velocity and thereby reduces the
erosion rate substantially as compared with the same steel
without Si [49].
The above discussion clearly brings out certain features
of elevated temperature erosion of metallic materials. Almost
all metallic materials exhibit ductile erosion response at room
temperature, whereas at elevated temperature both brittle and
ductile erosion responses are noted [31, 36, 37]. The velocity
exponents for metallic materials are 2.5 during ambient
temperature erosion. At elevated temperature the velocity
exponent of metals and alloys varies over a wide range from
0.9 to even more than 3.0 [3133]. It is established that theerosion rate at room temperature increases with the increase
of the particle size up to 50 m and beyond such magnitude
0
5
10
15
20
25
30
35
40
45
0 20 40 60 8 00
Impact Angle (Degrees)
MetalThicknessLoss(m) 70 m/s
45 m/s
35 m/s
25 m/s
Figure 9. Variation of metal thickness loss with the impact angle for
9.0Cr1.0Mo steel at 1123 K [37].
the particle size has no effect on the erosion rate. Reported
literature indicates that the erosion rate increases with the
increase of particle size at high temperature [29,30,34,39]. At
ambient temperature, changing theparticle shape from angular
to spherical results in altering the erosion response from brittle
to ductile [63, 64]. At elevated temperature, brittle to ductile
response is noted irrespective of the particle shape [29,30,37].
Particle feed rate has a negligible effect on room temperature
erosion rate [6567]. A remarkable effect of the particle
feed rate has been noticed at elevated temperature [29]. The
characteristics of mechanical properties of eroding material
particles have nominal influence on room temperature erosionbehaviour [56, 6875]. In contrast, this aspect at elevated
temperature is hardly explored.
R108
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
9/24
Topical Review
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 20 40 60 8 00
Impact Angle (Degrees)
ErosionRate(x10
-1,Kg/m2sec) 1053 K, 140 m/s
873 K, 140 m/s
873 K, 70 m/s
Figure 10. Variation of erosion rate of Co with the impact angle at1053 K [38].
0
1
2
3
4
5
6
7
0 100 200 300 400 500 600 700 800 900
Particle Size (m)
ErosionRate(m3/Kg)x10
Figure 11. Influence of particle size on the erosion rate of Inco 600alloy [34].
0
50
100
150
200
250
300
350
400
0 20 40 60 80 100 120 140 160
Particle Size (m)
ErosionRate(Kg/Kg)
Figure 12. Effect of particle size on the erosion rate of 304 SS at923 K [35].
4. Examination of eroded surfaces
At elevated temperature, the material removal is governed
by the synergistic effect of erosion and oxidation (EO).
Detailed examination of the eroded target material using a
scanning electron microscope (SEM), a transmission electronmicroscope (TEM) and an optical microscope (OM) has been
carried out. The advent of single particle experiment is
0
50
100
150
200
250
300
0 10 20 30 40 50 60 7 0
Impact Velocity (m/s)
WeightLossx10-6 (Kg/Kg)
9Cr 1 Mo Steel
Nozzle Tester Corrosion Erosion Air
Particle Size: 130 m
Test Temperature : 1123K
Test Time: 2 hr
SiC
Al2O3
30o
30o
90o
90o
Figure 13. Variation of weight loss with impact velocity for9.0 Cr1.0Mo steel at 1123K showing the effect of particle feedrate on erosion behaviour [37].
0
50
100
150
200
250
300
350
400
450
0 200 400 600 800 1,00 ,200
Test Temperature (K)
ErosiveMassLossRate(Kg/Kg) Erodent Concentration
2.6x10-6
Kg/m2s
42.0x10-6
Kg/m2.s
Material:
304 SS
430 SS
Figure 14. Variation of erosion rate of 304 SS and 430SS with testtemperature showing the effect of feed rate on erosion rate [41].
particularly fruitful since such experiments allow a detailed
examination of large, individual craters. On the basis of
the extensive literature [57, 67, 7678], four different types of
EO mechanisms can be envisaged: in the first case, at low
temperatures, at high impact velocities and feed rates, there is
no oxide scale. Even if there is any oxide scale, it will be very
thin and it will be able to deform in the same manner as that
of the substrate target. Under such circumstances, erosion
takes place from the metallic surface and this mechanism
of erosion is called metal erosion. The erosion behaviourin this regime is similar to the ambient temperature erosion
behaviour of metallic materials. The erosion response in the
metal erosion regime is ductile, the velocity exponent of the
erosion behaviour is between 2 and 3 and the erosion rate
is independent of the particle feed rate. The metal erosion
mechanism is schematically shown in figure 17(a). In the
metal erosion regime, there are two modes by which materials
can be removed. These modes are ploughing and cutting. In
general, when a particle is in contact with a target at positive
rake angle, the cutting mode operates. On the other hand, the
ploughing mechanism operates at negativerake angle. Cutting
mechanisms result in generation of new surfaces while the
ploughing mechanism involvesthe displacement and extrusionof the material with no new surface generation. In addition,
the cutting mode is more efficient than the ploughing mode
R109
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
10/24
Topical Review
0
20
40
60
80
100
120
0 5 10 15 20 25 30
Chromium Content (wt %)
MetalThicknessLoss
(m)
Normal Impact
Oblique Impact
Test Temperature : 1123 K
Impact velocity : 35 m/sErodent: Al2O3, 130 mFluidizing Medium: Air
310 SS304 SS
410SS
9Cr
5Cr
2.25Cr
Figure 15. Influence of Cr content on the erosion rate of steel [35].
when considered in terms of energy consumed per unit volume
removal of the target material. The work carried out by a large
number of investigators [7982] has revealed that almost all
metals and alloys lose material by the formation of a lip and
its subsequent fracture.
Inmetalsand alloys, duringerosion,once thelip is formed,
it is fractured by several modes. In the case of ductile metals
like copper [83], brass [83], aluminium [84] and iron [85]
the lip fracture occurs by necking, and the resulting fracture
is ductile, as exemplified by the dimpled fracture surface.
This mode of lip fracture is shown in figure 18(a). In the
case of a high strength alloy such as CuBe in age hardenedcondition [83], 301 SS [85] and TD nickel [85] the lip removal
is greatly aided by the formation of adiabatic shear bands at
the base of the lip as shown in figure 18(b) and subsequent
easy separation/fracture across this band. In this case also
the fracture is ductile, as indicated by the presence of shear
dimples on the fracture surface. These two modes of material
removal considered above involve the fracture of pre-existing
lips. On the other hand, a new mode termed adiabatic
shear induced spalling involves the formation of intersecting
adiabatic shear bands at the base of the crater and subsequent
removal of chunks of material as illustrated in figure 18(c).
This mode of weight loss, which is highly efficient in terms of
energyexpendedper unit volumeof targetmaterial removed, isimportant only at normal impact where maximum resistance
is offered to curtail spreading of deformation. The erosion
response under such circumstances will be similar to that
observedforceramic materials. But theunderlyingmechanism
is entirely different. In the case of ceramic materials, material
removal occurs with the formation of intersecting cone or
radial cracks, which nucleate from pre-existing flaws once a
critical tensile stress is exceeded. Further, these cracks are
essentially brittle in nature. On the other hand, formation of
anadiabaticshearbandrequires critical strain[86]. In addition,
the fracture surface resulting from adiabatic shear induced
spalling exhibits sheardimples, implying an essentiallyductile
fracture.On the other extreme, at very high temperatures and low
velocities and particle feed rates, erosion takes place from
the oxide scale only, as shown in figure 17(b). Under such
conditions, a thick oxide scale is formed on the target material
during erosion and the deformed zone formed due to impact is
confined within the oxide scale. The erosion behaviour from
the oxide scale is characterized by a brittle erosion response,
strong velocity dependence and particle feed rate independentof the erosion rate. This erosion mechanism is termed oxide
erosion. In oxide erosion, material removal occurs with
the formation of intersecting cones and radial cracks, which
nucleate from pre-existing flaws once a critical tensile stress
is exceeded. At an intermediate temperature, impact velocity
and particle feed rate, an oxide scale of intermediate thickness
is formed. However, the depth of the deformed zone extends to
the metallic substrate beyond the oxide scale. Consequently,
the oxide scale beneath the eroding particle tends to crack, gets
pushed down into much softer base material and in the process
the softer base material gets squeezed out onto the top surface
through the cracks in the oxide scale. Over a period of time,
the repetition of such a process during each impact causes the
formation of a composite layer comprising the bulk metal and
broken pieces of oxide scale. Erosion takes place from this
composite layer. This mechanism, presented schematically
in figure 17(c), is termed oxidation affected erosion. The
interesting aspect of oxidation affected erosion is that the
volumefractionof theoxidein thecomposite layer isa function
of erosion conditions such as temperature, impact velocity
and particle feed rate. As a result, the erosion behaviour
in the oxidation affected erosion regime can vary from a
ductile to a brittle response depending on the amount of oxide
scale present in the composite layer. Further, unlike in the
case of metal erosion or oxide erosion, the oxidation affectederosion rate depends strongly on the test temperature and
particle feed rate. The final erosion mechanism is oxidation
controlled erosion and this is illustrated in figure 17(d). At
relatively higher temperaturesandlowerparticle feed rates and
impact velocities, the oxide scale that forms during erosion
is brittle and non-adherent. In such cases, the oxide scale
gets removed after it attains a critical thickness. The erosion
behaviour in this regime exhibits a brittle erosion response,
weak velocity dependence and particle feed rate dependent
erosion rate. Figures 19 and 20 show SEM images of the
morphologies of the eroded surfaces and transverse sections
of the eroded surfaces, obtained after exposing commercially
pure Ni to elevated temperature erosion. All the four
mechanisms described schematically in figure 17 can be seen
under SEM.
5. Erosion oxidation interaction
As mentioned previously, the erosion behaviour of metallic
materials at elevated temperature is governed by the nature
of interaction between erosion and oxidation. The nature of
interaction between erosion and oxidation in turn depends on
the thickness, morphology, adherence and the toughness of
the oxide scales that form in these materials. Before goinginto the details of mechanisms it is important to deal with the
theoretical aspects.
R110
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
11/24
Topical Review
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
1 2 3 4 5 6
ErosionRate(Kg/m2s)x106
Co
MA-754Ni-30Cr
CoCrAlY
Ni-20Al
Test Temperature : 1053 K
Impact Velocity : 140 m/s
Impact Angle : 30o
Erodent : Al2O3 (20 m)
0.0
0.5
1.0
1.5
2.0
2.5
0
1 2 3 4 5 6
ErosionRate(Kg/m2s)x106
Ni
Co
MA-754 Ni-30Cr
CoCrAlY
Ni-20Al
Test Temperature : 873 K
Impact Velocity : 140 m/s
Impact Angle : 30o
Erodent : Al2O3 (20 m)
Ni
(a)
3.(b)
Figure 16. Bar diagram showing the erosion rates of various alloys forming different types of scale under oblique impact at (a) 1073 K,(b) 873K [48].
5.1. Theoretical aspects
The details of the interrelationship between erosion conditions
and erosion mechanisms has been presented earlier. Furtherdiscussion on this interrelationship will be carried outsubsequently. Thus, only the salient theoretical aspects willbe considered in the following section.
5.1.1. Steady state oxide scale thickness. If it is assumed
that the oxide scale which forms on the eroding materialduring erosion is adherent and sufficiently ductile to withstandrepeated impacts without developing cracks, steadystateoxidescale thickness can be defined. It can be assumed that theoxidationof theeroding material follows the parabolickineticsgiven by equation (3)
m2 = Kop t, (3)
where m is the mass gain experienced by the metal per unitarea due to intake of oxygen to form oxide scale, Kop is the
parabolic rate constant and t is the time of exposure. Theparabolic rate constant is usually expressed in the form:
Kop = Ao expQRT
, (4)
where Ao is the Arrhenius constant, Q is the activation energy
for oxidation, R is the gas constant and T is the absolute
temperature.
In order to represent EO interaction in mathematical
terms one needs the rate of growth of the oxide scale thicknesswith time rather than the weight gain given by equation (3).
As noted by Lim and Ashby [87], once the composition of the
oxide scale is known, equation (1) can be transformed to give
Z2 = 2Kpt, (5)
Kp = 0.5 C2Ko
p
, (6)
where C is a constant for a given oxide composition (unit ism3 kg1). Kp is usually referred to as the scaling constant.
R111
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
12/24
Topical Review
Metal
Deformed
Region
Erodent
Lip
Lm
Metal Erosion with No Oxide Scale
Metal
Deformed
Region
Erodent
Lip
Lm
Metal Erosion with No Oxide Scale
Metal
Deformed
Region
Erodent
Lip
Lm
Metal Erosion with Thin Adherent Ductile Oxide Scale
V
Oxide
Scale
Metal
Deformed
Region
Erodent
Lip
Lm
Metal Erosion with Thin Adherent Ductile Oxide Scale
V
Oxide
Scale
a (Metal Erosion)
Metal
OxideZ Lo
Damaged
Zone
V
Metal
OxideZ Lo
Damaged
Zone
V
b (Oxide Erosion)
Metal
Deformed
Region
ErodentV
Oxide
Scale
Metal
Deformed
Region
ErodentV
Oxide
Scale
Metal
Erodent
V
Oxide
Scale
Lip
Metal
Erodent
V
Oxide
Scale
Lip
Metal
Composite LayerV
Metal
Composite LayerV
c (Oxidation Affected Erosion)
Metal
Oxide
Scale
Metal
Oxide
Scale
Metal
Newly Formed
Scale
Metal
Newly Formed
ScaleOxide
Debris
Oxide
Scale
Metal
Oxide
Debris
Oxide
Scale
Metal
d (Oxidation Controlled Erosion)
Figure 17. Schematic presentation of various erosion mechanisms at elevated temperature.
A value appropriate to the erosion conditions should be chosen
for Kp since Levy etal [88] have clearly demonstrated that the
oxide scales grow much more rapidly under erosion conditions
as compared with static conditions. From equation (5),
the rate of increase of oxide scale thickness with time is
given as
dZ
dt=
Kp
Z. (7)
IfEo is assumed to be the erosion rate of oxide scale and
F be the particle flux rate given by the ratio of particle feed rate
(f ) to the eroded area then the rate of decrease of the oxide
scale thickness due to its erosion is obtained as
dZ
dt=
EoF
o, (8)
where o is the density of the oxide.
Finally, a situation will arise when the oxide growth byoxidation (equation (7)) will be equal to the oxide removal by
erosion (equation (8)). Under such conditions, the steady state
R112
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
13/24
Topical Review
a
Figure 18. Schematic diagrams of various metal removal mechanisms during metal erosion.
Figure 19. SEM images showing different morphologies of eroded surfaces of Ni exposed at elevated temperature: (a) metal erosion,(b) oxidation affected erosion, (c) oxidation controlled erosion and (d) oxide erosion [67].
oxide thickness (Zss) can be obtained as
Zss =Kpo
EoF. (9)
Hence, the steady state oxide thickness increases with
increasing temperature(through Kp), decreasing oxide erosion
rate and decreasing particle flux rate. It should be mentioned
that Eo in equation (9) represents theerosion rate of pure oxide.
5.1.2. Depth of deformed zone in oxide and metal. A largenumber of experiments have consistently shown that the depth
to which the deformed zone extends is usually of the order of
R113
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
14/24
Topical Review
Figure 20. SEM images of the transverse section of the eroded surfaces of Ni shown in figure 19. (a) Metal erosion, (b) oxidation affectederosion, (c) oxidation controlled erosion and (d) oxide erosion [67].
indentation diameter, or in other words
L = W , (10)
where W is the indentation diameter and is a constant of
the order of unity. The magnitude ofW depends on whether
the erodent is spherical or angular. When a conical particle
of mass mp as shown in figure 21 impacts an eroding material
of hardness H with an impact velocity V, then the incident
energy of the impacting particle is given as 0.5 mpV2 and the
energy consumed in forming the crater of volume U is H U
where H is the hardness of the target material. Equating these
two energies, we get
H U = 0.5 mpV2. (11)
For a spherical erodent having radius r , U is given by
U = W4
64r(12)
since crater depth can be considered to be considerablysmallerthan the diameter of the erodent. Substituting equations (12)
in equation (11) and solving for W one obtains
W = 2.56 rV1/2
H
1/4, (13)
where is the density of the erodent. Hence, the depth of
deformation for a spherical erodent can be obtained by putting
equation (13) in equation (10) as
L(Sphe) = 2.56r V1/2
H
1/4. (14)
For conical particles having a half angle of = 30, U is
given asU =
24
W3tan
. (15)
Substituting equation (15) in equation (11) and solving for
W one obtains
W =2r1/3V2/3
H1/3. (16)
Thus the depth of deformation for a conical particle can
be obtained by putting the value ofW in equation (10) as
L(con) = 2.0r1/3V2/3
H1/3. (17)
Equations (14) and (17) are valid irrespective of whether the
target material is metallic material or oxide scale. However,
it should be mentioned that the hardness H in equations (14)
and (17) represents the hardness of the oxide scale Ho in the
case of thick oxide scale as target material and the hardness of
the base metal in cases where the thickness of the oxide scale
is very small.
5.1.3. Critical oxide thickness (Zc). An important factor
that should be considered to understand the EO interactionis that the oxide scale usually exhibits a ductile to brittle
transition as a function of both thickness and temperature.
Stephenson et al [89] have demonstrated this phenomenon
under impact conditions. Saunders and Nicholls [90] have
also noted similar ductile to brittle transition for chromia and
alumina coatings. Importantpointsrelated to this phenomenon
are given as follows.
(1) At all temperatures, the scale becomes brittle beyond a
critical thickness (Zc) and thus can be removed easily
by spalling or by cracking and chipping due to particle
impacts.
(2) This value of Zc changes discontinuously over narrow
temperature ranges usually in the range of 700800
C.(3) Belowandabovethis temperature range Zc is independent
of temperature.
R114
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
15/24
Topical Review
ts
/
Figure 21. Variation of the calculated ratio of the time between impacts to the time of impact with hardness.
5.1.4. Time between the impacts and time of impact. If a
spherical erodent of radius r and mass m is considered, then
the number of particles impacting the unit area of the eroding
material every second equals F /m where F is the particle feed
rate. If it is assumed that each impact event causes damage
over an area A, where A is the impact crater area and is a
constant of the order of unity, then the number of particles (N )
impacting an area ofA every second is given by
N=
F
m
A (18)
or
N=
F
m
0.25 W2, (19)
where A = 0.25 W2, W is the impact crater diameter. The
time between two impacts can now be obtained as
tb =4mp
F W2. (20)
Substituting the value ofW obtained from equation (13)
in equation (20) and noting that mp=
4/3 r3
, tb is obtainedas
tb(sph) =0.82r1/2H1/2
F V. (21)
In the case of conical particles with a hemispherical top
of size 2r, mass m and half angle , the time between impacts
can be obtained from equation (20) if is 30. This is because
the mass of such a particle is equal to that of a sphere of
diameter 2r . For conical particles W is given by equation (16)
and substituting equation (16) in equation (20) and putting
m = 4/3 r3, tb is obtained as
tb(con) =4r1/3H2/3
3F V4/3. (22)
In the above expression it is assumed that the rebounding
particle does not interfere with the incident particle. This
assumption is reasonable because the erosion oxidation
interaction becomes prominent only at a low particle feed rate.
It is also important to consider the contact duration
between the particle and the eroding material during each
impact. This duration known as time of impact (tim) is given
as [91, 92]
tim(sph) =1.28r
H1/2, (23)
where r is the radius of the spherical particle having hardness
H and density . Thus for spherical particles the ratio of
time between impacts and time of impact can be obtained by
dividing equation (21) by equation (23) as
tb
tim(sph) =
0.64H
V F. (24)
In the case of conical particles the time of impact can be
given to a good approximation as
tim(con) =2.8r1/3
H1/3V1/3. (25)
Thus the ratio oftim and tb is obtained from equations (22)
and (25) as
tb
tim(con) =
0.48H1/3
V F. (26)
In the above equations H represents the hardness of the
eroding material. The calculated values of the ratios tb/tim in
the case of spherical (continuous lines) and conical erodents
(broken lines) are illustrated in figure 21 as a function of
hardness of the eroding material and for four values of the
product of impact velocity and feed rate. It can be noted from
figure 21 that the time between impacts is several orders of
magnitude higher than the time of impact. Hence for theerosion process at elevated temperature, time of impact has
negligible influence compared with time between impacts.
R115
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
16/24
Topical Review
5.1.5. Oxide scale growth between impacts. On the
assumption that oxidation is governed by a parabolic law,
the growth of oxide scale thickness (Z) is described by
equation (5). Thus the thickness of the oxide scale (Zb) grown
between two successive impacts is given by
Z2b = 2Kptb. (27)
Substituting the value oftb from equations (21) and (22)
in equation (27), one obtains the following equation for Zb for
spherical and conical particles.
Zb(sph) = 1.26
Kpr
V F
1/2(H)1/4 (28)
and
Zb(con) = 1.63Kpr
F
1/2
1/6H1/3
V2/3 . (29)
While deriving the above two equations it is assumed
that the growth of the oxide scale is controlled by the metal
and/or the oxygen ion diffusion through the scale. Thus, it
is assumed that the oxide scale does not crack as it grows,
thereby providing an easy path for the oxygen diffusion to
the metal oxide interface. This assumption is reasonable
because it concerns the growth of the oxide scale at a very local
region surrounding the impact point wherein prior impact has
removed the scale. Thus, the scale needs to be adherent and
uncracked only in this local region for the parabolic kinetics
to be valid. It does not matter that the scale is heavily cracked
or spalled at the macroscopic level.ThemagnitudeofZb incomparisonwithsteadystateoxide
scale thickness (Zss) has an important bearing on the erosion
oxidation interaction. Therefore, the variation ofZss and Zbwith impact velocity is given in figure 22 for two different
temperatures and two different feed rates. The magnitude of
Zb is indicated for both spherical and conical erodents. The
values of various parameters used for calculating Zb and Zssare given in table 3. Since most of the work on elevated
temperature erosion is done with steel, the values appropriate
to steel are chosen for various parameters. A value of 105 for
Ao and 210kJ mol1 for Q are chosen on the basis of the data
reported by Quinn [93]. The value ofC is calculated based
on the formation of Fe2O3 scale as observed by Levy and co-
workers [29, 45, 59] for a variety of ferritic steel undergoing
erosion oxidation degradation. The hardness of the oxide
scale and the metal is assumed to be 3.0 GPa and 2.0GPa,
respectively. Zc iskeptwithintherangeof110 m, consistent
with values reported by various investigators [89, 90]. A
perusal of the erosion literature indicates that the particle feed
rate mostly lies in the range 0.110 kg m2 s1. The erosion
rate of the oxide scale is assumed to be proportional to V3. The
test temperature is chosen in the range of 8731173 K since the
oxidationeffect becomes importantat these temperaturesin the
case of steels. It is clear from figure 22 that Zss decreases much
more rapidly with impact velocity than Zb. Thus a transitionvelocity beyond which Zss < Zb can be obtained for a given
temperature and feed rate.
Table 3. Assumed values of various important parameters.
Constant/variable Symbol Values Units
Hardness of oxide Ho 3.0 GPaHardness of metal Hm 2.0 GPaParabolic rate constant Kop kg
2 m4 s1
Arrhenius constant Ao 105, 104, 103 kg2 m4 s1
Activation energy Q 210 kJ mol1
Erosion rate of oxide Eo kgkg1
Erosion rate constant Eoo 105107
Reference velocity Vo 10 m s1
Velocity exponent n 3Density of oxide o 5400 kg m
3
Density of particles p 3200 kg m3
Erodent radius r 100(10) mCoversion factors for C 1.3 104 m3 kg1
transforming Kop to KpParticle flux rate F 0.1, 1.0, 10.0 kgm2 s1
Particle shapeCritical oxide thickness Zc 1, 10 mImpact velocity V 5100 m s1
Test temperature T 8731173 K
5.2. Conditions for prevalence of various EO mechanisms
If there is no oxide scale on the metal surface or if the
thickness of the oxide scale is very small compared with the
depth to which the deformation extends in the metal surface,
the dominant erosion mechanism will be metal erosion. On
the other extreme, if the steady state thickness (Zss) of the
oxide scale is less than the critical thickness of spalling (Zc)
and the depth of deformation is lower than the steady state
thickness, oxide erosion is the prevalent erosion mechanism.
If, however, Zss is lower than Zc and the depth of the
deformed zone is bigger than Zss, the oxidation affectederosion mechanism is observed. Finally, if the steady state
thicknessof theoxidescaleis greater than thecriticalthickness,
the oxide scale will never attain steady state thickness. The
erosion will take place by spalling of the scale and oxidation
controlled erosion will be operative.
Roy et al [94], employing a new methodology, examined
thetransitioncriteria from themetalerosionregimeto theoxide
erosion regime. In their work, they eroded pre-oxidized Ni
samples having varying thicknesses of oxide scale at ambient
temperature. The ratio of the erosion rate of the oxide scale
(Eo) to the erosion rate of the substrate (E) is plotted against
the ratio of the initial thickness of the oxide scale (t) to the
depth of deformation (L) due to impact. Their observationis portrayed in figure 23. This figure reveals three distinct
regimes. In region one, the value oft/L is higher than 4.0.
The erosion rate of the oxide scale assumes a relatively high
but constant value. In region two, t /L is higher than 0.5 but
lower than 4.0. In this regime, there is a smooth change of the
relative erosion rate from a low value to a high value. Finally,
in regime three, t/L is lower than 0.5. In this regime, the rate
of change of the relative erosion rate is slow. In addition, there
appears tobea peak in therelative erosion rate atapproximately
t /L = 6.
Regime one can be deduced as an oxide erosion regime
on the basis of the erosion response, which is brittle, i.e.
higher erosion rate at normal impact and on the basis of thevelocity exponent, which is 3.0, of the erosion rate. Further,
the material in this regime was removed by brittle chipping.
R116
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
17/24
Topical Review
Figure 22. Variation of the calculated values of the steady state thickness and the thickness of the oxide scale grown between two successiveimpacts.
Figure 23. Variation of the ratio of the erosion rate of the oxide scale to the erosion rate of the substrate with the ratio of the thickness of theoxide scale to the depth of deformation [94].
In contrast, regime three is characterized by pure metallic
erosion because the thickness of the oxide scale wasnegligibly
low. In addition, the erosion response, the material removal
mechanisms and the velocity exponent of the erosion rate in
this regime are consistent with those of the erosion of metals
and alloys at ambient temperature. Regime two represents the
transition from the oxide erosion regime to the metal erosion
regime. It also shows that the transition is not sharp but rather
smooth. Thus as long as the depth of deformation is confined
within the oxide scale, erosion takes place from the oxidescale only and erosion behaviour similar to oxide erosion is
prevalent. When there is no oxide scale, metal erosion is
dominant. In the intermediate regime, erosion may take place
from the oxide layer, but the deformation will extend to the
substrate also.
5.3. Models for erosion oxidation interaction mechanisms
Several investigators [57, 61, 95103] have tried to identify
various possible mechanisms of interaction between erosion
and oxidation over the last few years. For example,
Hogmarketal [95] have identified six differentmechanismsof
interaction ranging from pure oxidation to pure erosion. Suchmechanisms are compiled in the report of Stack et al [96]
and Sundararajan and Roy [16]. Wellman and Nicholls [97]
R117
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
18/24
Topical Review
presented an excellent review on this aspect. Barklow and
Petit [98] have identified four different mechanisms such as
metallic erosion, oxidation modified erosion, erosion modified
oxidation and oxidation on the basis of the combined effect of
kinetic energy of the impacting particles, oxide growth rate
and particle flux rate. According to them the effect of thesubstrate becomes negligible when the particle impact energy
and particle flux rate are less. The influence of the substrate
becomes important at high particle flux rate and high impact
energy. Wright et al [52] classified the interaction between
corrosion and erosion into two regimes on the basis of kinetic
energies of the impacting particles. When the kinetic energy
is low, erosion behaviour exhibits brittle erosion mechanisms
characterized by a maximum erosion rate under normal impact
and a velocity exponent in excess of 3.0. When the kinetic
energy is high, erosion behaviour is ductile, maximum erosion
rate occurs at oblique impact angle and the velocity exponent
lies in the range 22.5. This classification was subsequently
modelled by Natesan and Liu [99] and later modified byStacket al [100].
Rishel et al [101] extended the idea originally proposed
by Kang et al [50] and Chang et al [102]. They defined
various mechanisms of erosionoxidation interaction using four
parameters: instantaneous scale thickness ( ), parabolic rate
constant under combined attack of erosion and oxidation Kce,
erosion rate constant of oxidation product K and parabolic
rate constant for oxidation (Kp) only. According to these
investigatorstherateof change of instantaneous scalethickness
is given byd
dt=
K(ce)
K . (30)
If d /dt is positive erosion of the oxide scale takes place.Erosion under such conditions will be characterized by a brittle
erosion response. If d/dt is negative oxidation affected
erosion becomes dominant, i.e. erosion takes place from a
composite layer consisting of oxide scale, metallic substrate
and erodent. If, however, d /dt is equal to zero erosion
enhanced oxidation plays a dominant role. Within this regime
three different modes are possible; Type I erosion enhanced
oxidation will be operative if K(ce) is equal to Kp. Type II
erosion enhanced oxidation can be seen ifK(ce) is greater than
Kp and Type III erosion enhanced oxidation occurs when the
oxide scale spalls. Finally, if there is no corrosion, metal
erosion will be prevalent. This model, in principle, canexplain
the effect of feed rate, particle size, etc on the observed erosionrate. However, the predictive capability of this model is
limited.
Sethi and Corey [54], with the help of oxide scale growth
kinetics, have demonstrated that the temperature dependence
of erosion rate shows three different regimes. These regimes
are (1) a low temperature regime where erosion rate is
independent of temperature, (2) an intermediate temperature
regime where erosion rate increases with the increase of
temperature and (3) a high temperature regime.
Stephenson and Nicholls [103] have defined three
different regimes on the basis of the ratio of contact radius (a)
and scale thickness (z). If this ratio is less than 0.1 a substrate
dominated regime can be seen. If this ratio is between 0.1and 1.0, oxide modified behaviour will be prevalent. If this
ratio is more than 1.0 oxidation dominated erosion will be the
operatingmechanism. Theyalso proposed the presenceof pure
oxidation in the case of the negligible presence of the erodent.
Another classification due to Stacket al [104] proposes
the presence of three regimes, namely, erosion dominated,
erosioncorrosiondominated and corrosion dominated. These
regimes are defined based on kinetic energy, temperature andcritical oxide thickness. The transition to corrosion dominated
behaviour can be attributed to the formation of critical oxide
thickness. Above the temperature at which this occurs, the
oxide formed in a given time interval cannot be removed from
the scale metal interface by erosion. Below the temperature at
which the critical scale thickness is attained, the oxide formed
can be removed during impact. The corrosion dominated
regime can be further divided into two more sub-regimes
depending on the velocity dependence of these regimes. The
corrosion dominated regimes can be corrosion dominated-1
and corrosion dominated-2 regimes. In the processes of
identifying these mechanisms, Stack assumed that oxidation
behaviour of the target material follows a parabolic oxidationbehaviour and the oxidation that occurs during each impact is
negligible.
Sundararajan [78] has broadly proposed two differ-
ent regimes of EO interaction, namely, erosion controlled
regimes and oxidation controlled regimes. In erosion con-
trolled regimes three different mechanisms can be envisaged.
These mechanisms are(a)metal erosion, (b)oxidationaffected
erosion and (c) oxide erosion. In oxidation controlled regimes
there are two different mechanisms, namely (a) oxidation con-
trolled erosion continuous and(b) oxidation controlled erosion
spalling. These zones can be identified on the basis of sev-
eral erosion related parameters. The most important feature of
this model is the ability of the model to predict the prevalentmechanisms of erosion once the conditions of erosion and the
thermo-physical properties of the erodent and target material
are known.
5.4. Erosion oxidation interaction map
While establishing the operative mechanisms for erosion
oxidation, it is noted that erosion conditions significantly
influence the oxidation. The factors which influence the
operating mechanisms are impact velocity, impact angle, feed
rate and temperature. Similarly, the particle size and shape
also have a profound effect on this map. The combined
influence of all these factors on erosion rate remainedqualitative for a considerably long time. Specific erosion
mechanismsfor variousmetallic materialshave beendescribed
by Hogmark et al [95], Wright et al [52], Kang et al
[50] and Sundararajan [78]. Barkalow and Petit [98] for
the first time have tried to organize such information in the
form of an erosion oxidation map where the prevalence of
various mechanisms is shown in the domain of the particle
kinetic energy and the scale growth rate. Sundararajan [78]
has attempted to organize erosion oxidation maps in the
domain of impact velocity and temperature. Stephenson
and Nicholls [77, 105] have plotted particle velocity versus
oxide thickness for a specific particle size and temperature
to construct such maps while Stack and Pena [106] plottedvelocity versus temperature. But the attempt has remained
limited with the consideration of theoretically postulated
R118
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
19/24
Topical Review
values of erosion rates. Numerous experiments conducted
by the author can be considered sufficient to attempt making
erosion oxidation interactionmapsfrom the experimental data.
Because of the significance of various erosion conditions
on erosion rate and erosion mechanisms, an erosion oxidation
interaction map can be represented in a multidimensionalconfiguration. A simplified representation of the erosion
oxidationmap canbe made in a two-dimensionalplot of impact
velocityand temperature fora givenimpact angle andfeed rate.
The influence of feed rate on the EO map at normal impact
for commercially pure Ni is shown in figure 24 whereas the
influence of impact angle on the EO map of Ni is illustrated
in figure 25. Similarly the effect of Cr addition on the EO
map can be noted in figure 26. In general all these maps
describe the transitional boundary between regimes of various
operative mechanisms of erosion at elevated temperature.
These maps are able to clearly delineate the metal erosion,
oxidation affected erosion, oxidation controlled erosion and
oxide erosion. The extent of each of these regimes depends ontemperature, feed rate, impact angle and impact velocity. An
examination of all these maps shows that:
(a) a low temperature imparts the metal erosion regime. With
an increase in temperature the metal erosion regime shifts
to the oxide erosion regime via the oxidation affected
erosion and oxidation controlled erosion regimes,
(b) a higher feed rate extends the metal erosion regime and in
turn alters theexistence andtheextentof theother regimes
and
(c) oblique impacts tend to reduce the extent of the metal
erosion regime and promote other regimes.
Theexistence andextent of a particular regime is governedby the oxide scale growth and the depth of deformation due to
impact. These two factors are opposing in nature with respect
to erosion conditions. Thus for a given condition, keeping
the impact velocity constant, as the temperature of erosion is
increased, the influence of oxide scale becomes dominant. As
a result, the oxide scale starts growing at a rate faster than
erosion. This leads to a situation where the thickness of the
oxide scale becomes significant compared with the depth of
deformation andat still highertemperaturewhere thethickness
of the scale attains critical thickness of spalling or attains a
thickness where the rate of growth of oxide scale is equal to
therate of erosion. This results in transition from metal erosion
to oxidation affected erosion or oxidation controlled erosionor even to oxide erosion. This can be seen in figure 25 at
impact angle of 30 and at impact velocity of 35 m s1 and at
feed rate of 0.2 g min1. Similarly keeping the temperature
of erosion constant as the impact velocity is increased, it not
only becomes more difficult for the oxide scale to grow but the
thickness of the oxide scale also becomes less compared with
the depth of deformation. This situation causes prevalence of
metal erosion in preference to oxidation affected erosion and
oxidationaffectederosion in preferenceto oxidationcontrolled
erosion, as shown in figure 24 (at impact angle of 90, feed rate
of 0.2 g min1 and temperature of 673 K).
The effect of the higher particle feed rate is analogous
to that of lower temperature. At a specific condition, higherparticle feed rate does not permit the oxide scale to grow as the
time intervals between impinging particles are shortened. This
(a)
(b)
Figure 24. Erosion oxidation interaction map for Ni, (a) for feedrate of 3.3 106 kg s1 and (b) for feed rate of 3.3 104 kg s1.
condition results in an expansion of the metal erosion regime
and the shifting of other erosion regimes to higher temperature
or lower impact velocities. The effect of impact angle can
be considered in terms of impact velocity. As the impact
angle isdecreasedthe normalcomponent of theimpact velocitydecreases. Thus, oblique impact gives rise to constricted metal
erosion regimes. Consequently, the oxidation affected erosion
regime and the oxidation controlled erosion regimes appear at
higher impact velocity and lower test temperature. At oblique
impact, it is possible to seethe presence ofoxideerosion, which
is not prevalent at normal impact in many cases.
It is noted in figures 2426 that the erosion rate tends
to increase and the erosion oxidation interaction mechanism
shifts from metal erosion to oxide erosion via oxidation
affected erosion and oxidation controlled erosion with the
increase of temperature. It can also be inferred that the erosion
rate is higher (1) in oxidation affected erosion than in metal
erosion, (2) in oxidation controlled erosion than in oxidationaffected erosion and (3) in the oxide erosion regime than in the
oxidation controlled erosion regime [107]. It is important to
R119
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
20/24
Topical Review
(b)
(a)
Figure 25. Erosion oxidation interaction map for Ni, (a) for impactangle of 30 and (b) for impact angle of 90.
mention that the erosion oxidation map has been constructed
experimentally for the first time. The influence of various
erosion conditions on such a map can be explained on the basis
of oxidation characteristics and mechanical properties of theeroding materials. Stacket al [108] constructed a similar map
previously. However, such a map was available only in the
low velocity regime. It also failed to highlight the influence of
other parameters such as particle feed rate, impact angle, alloy
composition, etc.
6. An analysis of enhanced oxidation kineticsdue to erosion induced roughness
The oxidation kinetics of metals and alloys are found to
alter during erosion or wear. The observation on enhanced
oxidation kinetics is primarily centred on iron base alloys. The
influence of sliding wear on oxidation was initially proposedby Quinn [109]. According to Quinn, the activation energy
for the parabolic rate constant (KP) remains the same during
(a)
(b)
Figure 26. Erosion oxidation interaction map for (a) Ni and(b) Ni20 Cr alloy.
static or wear induced oxidation. However, the magnitude
of KP of wear induced oxidation is higher than that for
static oxidation. A similar contention is also made by Lim
and Ashby [87] for wear induced oxidation and by Levyand co-workers [29, 45, 59] for erosion induced oxidation.
Roy et al [76] for the first time attempted to estimate the
altered oxidation kinetics during erosion and subsequently
tried to formulate a phenomenological model to explain such a
result.
The observations of Roy et al [76] are presented in
figures 27 and 28. The variations of mass gain per unit area
as a function of time at 1073 K for as-received and eroded Ni
are given in figure 27 whereas figure 28 depicts similar data at
1173 K for Ni20Cr alloy. It is clear that the oxidation rate of
Ni increases significantly for eroded samples when compared
with as-received samples. The oxidation rate increases with
the increase of impact velocity. The increase of oxidation ratewith increase of impactvelocityis higherat normalimpactthan
at oblique impact. With regard to the Ni20Cr alloy it is noted
R120
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
21/24
Topical Review
0.000
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0 10 000 20 000 30 000 40 000 50 000 60 000 70 000 80 000
Time (sec)
MassGain/UnitAreax103 (Kg/m
2)
As Received
V = 35 m/s, = 30oV = 65 m/s, = 30oV = 35 m/s, = 90oV = 65 m/s, = 90o
Material: Ni
V= Impact Velocity
= Impact AngleTemperature : 1073 K
Figure 27. Variation of mass gain per unit area of Ni with time at 1073 K [76].
0
2
4
6
8
10
12
14
16
18
0 10 000 20 000 30 000 40 000 50 000 60 000 70 000 80 000
Time (sec)
MassGainperU
nitAreax103
(Kg/
m2)
Material: Ni-20Cr
V= Impact Velocity
= Impact AngleTemperature : 1173 K
As ReceivedV = 35 m/s, = 30 oV = 65 m/s, = 30 oV = 65 m/s, = 90 oV = 35 m/s, = 90 o
Figure 28. Variation of mass gain per unit area of Ni20Cr alloy at 1173 K [76].
that the oxidation rate decreases with the increase of impact
velocity. Further, the influence of impact angle on oxidation
kinetics is negligible.
Theabove observation forNi is modelled by assuming that
the parabolic rate constant is related to the surface roughness
(Ra) or impact velocity as
K = KO1(Ra)x1 exp
Q
RT
, (31)
K = KO2 (V sin )x2 exp
Q
RT , (32)
where KO1 and KO2 are constants, V is the impact velocity and
is the impact angle, Q is activation energy, R is the universal
gas constant, T is test temperature andx1and x2 areroughness
exponent and velocity exponent, respectively. In order to
establish the relation between the parabolic rate constant
and impacting condition, the average activation energy for
the parabolic rate constant is determined by plotting ln K
(parabolic rate constant) against (1000 T1) i n K1. Using the
average activation energy, the average parabolic rate constants
at different eroding conditions are estimated. The natural
logarithm of the average parabolic rate constant is then plotted
against ln Ra and ln V sin . The slopes of best-fit straight
lines are then computed as constants KO1 and KO2 whereasthe intercepts of the lines with the ordinate are calculated as
exponents x1 and x2. The expressions for rate constants are
R121
8/3/2019 Elevated Temperature Erosive Wear of Metallic Materials 2006 ROY
22/24
Topical Review
y
Figure 29. Variation of the parabolic rate constant with (V sin )2.08
exp(Q/RT ) [76].
then obtained as
K = 4.67 109(V sin )2.08 exp
22 400
RT
, (33)
K = 9.62 107(V sin )4.26 exp
22 400
RT
. (34)
In order to assess the ability of the expression to predict
the parabolic rate constant, the estimated rate constants are
plotted against (V sin )2.08
exp(22 400/RT ) in figure 29.A straight line with a good fit suggests the suitability of the
expression.
A similar exercise cannot be carried out for Ni20Cr
alloy because of two factors. First, the nature of the oxide
scale that forms on Ni20Cr alloy can have a dual grain size
distribution [110]. Second, higher roughness in these alloys
results in lower sizes of the globular oxide grains. Hence even
though the roughness changes the amount of grain boundary
also gets altered and this in turn provides different extents of
short circuit paths. Thus the pre-exponential factors in the
Arrhenius type equation should be a function of the roughness,
grain size and the nature of the oxides that are present in the
oxide scales of Ni20Cr alloy. Hence a similar modelling forNi20Cr alloy cannot be achieved.
7. Areas of future research
In spite of significant progress in elevated temperature erosion
of metals and alloys, certain areas still need to be addressed.
The material flow behaviour during solid particle erosion is at
high strain, high strain rate, under adiabatic deformation and
undermultiaxial stress condition. The material flow behaviour
under such conditions, particularly at elevated temperature is
not well understood. The estimation of mechanical properties
of materials under such conditions is required for effective
modelling of the elevated temperature erosion behaviour ofmaterials. Thus there is a clear need to develop a simplified
test technique which will simulate the erosion conditions in a
test sample and evaluate the mechanical behaviour under such
conditions. For example the depth of deformation estimated
using equation (14) would be more accurate if dynamic
hardness is used instead of static hardness.
The deformation and fracture behaviour of the oxide scale
under erosion conditions is poorly understood. Further, thehigh strainrate flow behaviour of a layered structurecontaining
the oxide scale and the substrate material also needs to be
addressed. Such an understanding is required to model the
spalling behaviour of theoxide scale once it reaches thecritical
thickness of spalling. The dramatically different spalling
behaviour of the segmented scale and the compacted scale can
be explained more accurately only when concepts related to
such deformation behaviour get crystallized.
The kinetics of oxidation during erosion are an order of
magnitude higher than that under static conditions. An effort
to estimate the increased oxidation rate quantitatively either
through careful experiment or by rigorous theoretical analysis
is hitherto unexplored. All the reasons for such acceleratedoxidation are not yet revealed. Thus, there is a clear need
to carry out organized experiments to evaluate the oxidation
behaviour of materials under erosion conditions. Attempts
should be made to separate out the influence of possible factors
responsible for enhanc