Wear 267 (2009) 244254

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Erosioncorrosion resistance of engineering materials in

various test conditions
S.S. Rajahram , T.J. Harvey, R.J.K. Wood
national Centre for Advanced Tribology at Southampton, School of Engineering Sciences, University of Southampton Higheld,
Southampton SO17 1BJ, UK

a r t i c l e i n f o a b s t r a c t

Article history: Erosioncorrosion is a complex phenomenon which involves the interaction between the mechanical
Received 16 September 2008 processes of solid particle erosion and the electrochemical processes of corrosion. A whole range of issues
Received in revised form 5 January 2009 is faced by a designer when trying to obtain relevant information on erosioncorrosion performance
Accepted 5 January 2009
of a material. Amongst the constraints are the dispersed test conditions and test rigs available in the
literature making comparisons and quantifying erosioncorrosion wear rates of different materials very
Keywords:
difcult. The aim of this work is to evaluate the repeatability of erosioncorrosion experiments and to
Slurry pot erosion tester
investigate the role of different parameters inuencing erosioncorrosion. The materials tested in this
Erosioncorrosion
Error analysis
work are stainless steel (SS316L/UNS S31603), carbon steel (AISI 1020/UNS G10200) and nickel-aluminium
Synergy bronze (NAB/UNS C63200). A slurry pot erosion tester was used as the test apparatus and test parameters
such as erodent size, erodent concentration, ow velocity and test solutions were varied to study their
effect on erosioncorrosion. SEM analysis showed that a similar erosioncorrosion mechanism is seen
for SS316L and NAB with formation of multiple extruded lips and platelets typically seen for erosion
dominated material. In contrast the surface of AISI 1020 revealed the formation of craters, pits and shallow
indentations which suggests that corrosion mechanism has a dominant inuence on the material. Error
rates in tests were found to have an average of 5.5% which are relatively low indicating good repeatability
of test measurements and data from the test rig. The erosioncorrosion resistance of AISI 1020, SS316L
and NAB were compared and it was found that SS316L showed the lowest erosioncorrosion mass loss
rates in all test conditions followed by NAB and then AISI 1020. However in terms of synergistic ranking,
NAB showed the best resistance to the combined action of erosion and corrosion with the highest negative
synergy value. Positive synergy was evident for AISI 1020 in 3.5% NaCl and SS316L in 0.3 M HCl. A wear map
is presented to evaluate erosioncorrosion trends of the materials. This work combines the assessment
of test repeatability, variation in test conditions and comparison of material performance which are key
stages in a material selection process.
2009 Elsevier B.V. All rights reserved.

1. Introduction applications in these environment cannot be overstated as compo-

nent wear can be accelerated by the aggressive conditions in these
Wear caused by solid particle erosioncorrosion has received harsh environments. Synergy is the additional wear rate experi-
signicant attention amongst researchers in recent years due to enced by a metal under the combined action of erosioncorrosion
the severity of the problems caused by this phenomenon to com- conditions which is higher than the sum of wear rate due to pure
ponents in service, resulting in material loss and subsequent erosion and ow corrosion [13]. This factor can reduce the ser-
equipment failure. Erosioncorrosion related problems occur in vice life of engineering material signicantly. The interactions of
power plants, oil and gas processing plants and chemical plants these two processes are complex in nature, therefore, careful mate-
where there is an interaction between solid particles, corrosive rial characterization in terms of its erosion resistance, corrosion
uid and a target material. The problem has been reported to affect resistance and synergy effect has to be understood for appropriate
static equipment for example pipelines, valves, heat exchangers, material selection in any engineering application.
pressure vessels and various rotating equipment namely compres- Synergism as described earlier is the difference between the
sors, turbines and pumps. The importance of material selection for combined erosioncorrosion wear rate and the sum of pure ero-
sion and pure corrosion acting separately and can be expressed by
Eq. (1).
Corresponding author. Tel.: +44 23 8052 7667; fax: +44 23 8059 3230.
E-mail address: ssr1y07@soton.ac.uk (S.S. Rajahram). S = T (E + C) (1)

0043-1648/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2009.01.052
 S.S. Rajahram et al. / Wear 267 (2009) 244254 245

where T is the total wear rate due to erosioncorrosion, E is the wear Table 2
Surface roughness of the test samples.
rate due to pure erosion, C is the wear rate due to ow corrosion
and S is the additional wear rate due to the synergistic interaction Surface roughness SS316L AISI 1020 NAB
between erosion and corrosion. Synergism consists of two main Ra 0.28 0.04 m 0.21 0.04 m 0.51 0.06 m
components which is the erosion enhanced corrosion rate (C) and
the corrosion enhanced erosion rate (E) shown in Eq. (2).
Table 3
Sand optical analysis [24].
S = E + C (2)
Coarse Medium Fine

The complexity of the erosioncorrosion phenomena is not Size (m) 3D 665 49 294 43 106 41
only limited to the interaction between the various parameters Section 553 114 167 60
affecting erosion and corrosion. A whole range of issues is faced Circularity 3D 0.78 0.05 0.79 0.05 0.76 0.06
by the designer when trying to obtain relevant information on Section 0.72 0.09 0.71 0.10
erosioncorrosion performance of a material. It would be ideal if Ratio 3D 0.75 0.12 0.77 0.10 0.72 0.13
the testing of material under erosioncorrosion conditions could Section 0.70 0.14 0.69 0.14
be done in the eld to obtain accurate data on material perfor- SF 3D 0.61 0.18 0.63 0.19 0.68 0.19
mance. However, this is economically and practically unviable. Section 0.68 0.18 0.70 0.20
Researchers and designers have to rely on information obtained
from laboratory-based erosioncorrosion experiments. The con-
straint is often the various laboratory experiments are very specic where  C ,  E,  T are errors due to corrosion, erosion and
in terms of tests conditions and environments. The need for stan- erosioncorrosion respectively. From the equation, it can be seen
dardization of erosioncorrosion testing is crucial. The rst step that, in order to obtain reliable data for synergy calculations, accu-
towards standardization is selecting a suitable rig which can pro- rate measurements of all three components are required. Even the
duce repeatable results. By embarking on this, test conditions could slightest skew in measurement can affect synergy calculations and
be reproduced every time, giving condence in the results when propagate to signicant error levels.
varying parameters in experiments. The materials selected for this work are stainless steel (SS316L),
In order to determine the total wear rate caused by the combined carbon steel (AISI 1020) and nickel-aluminium bronze (NAB).
effect of erosion and corrosion, various rigs have been designed to SS316L is widely used in the oil and gas industry, chemical plants
measure this effect. Amongst these rigs are slurry pot erosion tester and power plants and is known for its superior corrosion perfor-
[49], jet impingement rig [1014], Coriolis erosion tester [1517], mance owing to its passivity. The presence of chromium in SS316L
pipe ow loop [18,19] and rotating cylinder apparatus [2022]. Each allows the formation of a thin layer of chromium oxide lm that
rig has its advantages and disadvantages in terms of ease of usage, inhibits corrosion [23]. AISI 1020 however, is not as corrosion resis-
cost, ease of maintenance and control of test parameters. Labora- tant as stainless steel and corrodes in atmospheric conditions where
tory simulation of erosioncorrosion testing is often scaled down to the relative humidity exceeds 60%. However it is widely used in
predict actual eld conditions. The pipe ow loop provides a good many engineering applications due to economical reasons. NAB on
indication of actual erosioncorrosion rates in slurry pipelines but the other hand is widely used for propulsion and seawater han-
is seldom used due to the high cost of construction, maintenance dling systems in naval platforms and similar to SS316L also inhibits
and actual space limitation within laboratories. A slurry pot erosion corrosion through the formation of a thick adherent oxide lm.
tester was selected for this work as it allows variation of parameters These three materials were tested to determine synergy levels and
with good accuracy [49]. Erosioncorrosion testing is performed their resistance to the combined action of erosion and corrosion.
with this rig by changing the slurry medium to corrosive slurry. Wear mechanisms were investigated for these material to develop
The current work involves a comparison of engineering mate- a mechanistic understanding on erosioncorrosion.
rials to study their sensitivity to parameters like ow velocity,
sand size, sand concentration and test solutions. It is vital that 2. Experimental
these parameters are well understood and evaluated in a con-
trolled condition to determine the synergistic effect that arises from 2.1. Materials
erosioncorrosion. A set of experiments was performed to com-
pare erosioncorrosion resistance of a passive and active material The materials tested in this work are stainless steel (SS316L),
in a controlled environment. This work also aims to provide quan- carbon steel (AISI 1020) and NAB with a hardness of 220HV(0.1) ,
titative information on the applicability of the slurry pot erosion 254HV(0.1) , 235HV(0.1) respectively. The chemical compositions of
tester to obtain repeatable results. The multiple tests performed the materials tested are given in Table 1. The test samples were
provide useful information on the percentage error and con- cylindrical in shape with dimensions of 5 mm diameter and 70 mm
dence level of each test. Error analysis is important in evaluating length and were tested as received. The average surface rough-
erosioncorrosion data. For synergy, the error is calculated using ness of the test samples is given in Table 2. Erodents used in the
Eq. (3). erosion and erosioncorrosion experiments are natural uncrushed
 silica sand. Table 3 summarises the average size of the sand used
for testing. Optical analysis of the sand was performed by National
S = C2 + E2 + T2 (3)
Physical Laboratories (NPL) [24]. The hardness of the sand used was

Table 1
Chemical composition of materials used in the experiment (wt.%).

P S Mn Al Ni Fe Cr Mo Si C Cu

SS316L (UNS S31603) 0.034 0.027 1.39 10.5 Bal 16.7 2.04 0.35 0.015
AISI 1020 (UNS G10200) 0.05 0.05 0.6 Bal 0.35 0.25
NAB (UNS C63200) 1.10 9 5 4.5 0.01 Bal
246 S.S. Rajahram et al. / Wear 267 (2009) 244254

Fig. 3. Placement of samples on two nylon-coated arms, secured with o-rings to

prevent slurry ingress [4].

Fig. 1. Optical image of silica sand [24].

were conducted in different test solutions. For the pure erosion
measured as 1100HV(0.3) . An optical cross-section image of the sand experiments, 0.1 M NaOH was used as the test solution with the
is shown in Fig. 1. addition of silica sand. The pure corrosion experiments were per-
formed using 3.5% NaCl for AISI 1020 and NAB while 0.3 M HCl was
2.2. Experimental rig used for SS316L. Erosioncorrosion experiments were performed
using similar corrosive solutions as used in the pure corrosion
A slurry pot erosion tester was used to perform erosion experiments with the addition of silica sand. A variety of test solu-
corrosion experiments. Fig. 2 shows the diagram of the assembled tions were evaluated to obtain positive synergy which included a
slurry pot enclosed in the Faraday cage. Details of this experimental range of HCl solution with different molarity, from 0.1 M HCl to
rig have been described elsewhere [4]. The rig is driven by a 3.5 kW 0.5 M HCl. The rationale behind the choice of the test solutions for
motor which is connected to the slurry shaft through a toothed each material will be discussed in detail in Section 3.2. The tempera-
belt and two pulleys (on the shaft and on the motor). Cylindrical ture of the test solutions was maintained at 40 1 C throughout the
test samples are inserted between two nylon-coated arms at the duration of the experiment. Tests were conducted at different veloc-
end of the shaft as shown in Fig. 3 [4]. The speed of the motor is ities, sand sizes and sand concentrations to study the effect of the
controlled through a variable speed drive with maximum rotation respective parameters on erosioncorrosion. Tables 46 shows the
speed of up to 3500 rpm (18.7 m s1 ). The pot is made of uPVC with test matrix of the experiments conducted for the materials. To allow
a maximum capacity of 4 litres and has a cup type design copper a manageable test programme, a complete test matrix was con-
cooler which allows the temperature of the slurry to be controlled ducted for AISI 1020 while SS316L was tested at selected mid range
by the circulation of hot/cold water. The pot is designed with bafes values to compare its erosioncorrosion performance against AISI
in it, to allow mixing of solid particles in the slurry, preventing it 1020. NAB was tested at one test condition in order to evaluate syn-
from settling at the bottom of the pot. The rig assembly is enclosed ergy values for different materials. The weight of the test samples
within a Faraday cage which allows electrochemical measurements was measured before and after the experiments to determine the
to be made and also acts as a protective safety barrier when running mass loss for the different test conditions. Samples were cleaned in
experiments. an ultrasonic bath and dried in a dessicator before gravimetric mass
loss measurements were made. Gravimetric measurements were
2.3. Methodology performed using a precision balance with an accuracy of 0.01 mg.
Measurements were repeated six times for each sample and the
In order to evaluate the pure erosion (E), ow corrosion (C) and maximum and minimum values were discarded. The remaining 4
erosioncorrosion (T) rates at different test conditions, experiments readings were averaged to obtain an average mass loss for each test
sample. As the rig design allows two samples to be used in a test, the
average mass loss is taken between the two samples as the effective
mass loss rate in mg h1 .

3. Results and discussion

3.1. Error analysis

In order to determine the repeatability and to evaluate the

experimental error in the experiments, a variation of replicated
gravimetric measurements were performed on a single sample. Six
gravimetric measurements were performed for each sample and
the standard deviation was calculated for each set of measure-
ments. Using the standard deviation data, a population graph was
produced by sorting each standard deviation data in bins as shown
Fig. 2. Assembled slurry pot erosion tester enclosed in a Faraday cage (pot capacity in Fig. 4. The average of the measurements (0.046 mg) lies towards
4 l) [4]. the right hand side of the population peak indicating that the data
 S.S. Rajahram et al. / Wear 267 (2009) 244254 247

Table 4
Test matrix for AISI 1020.

Sand concentration Solution Sand size Velocity (m s1 )

3 5 7 10 12

Corrosion 0 wt.% 3.5% NaCl

0.5 wt.% Fine

Medium
Coarse

1 wt.% Fine

Medium

Coarse
Erosion 0.1 M NaOH
2 wt.% Fine

Medium
Coarse
5 wt.% Fine

Medium
Coarse

0.5 wt.% Fine

Medium
Coarse

1 wt.% Fine

Medium

Coarse
Erosioncorrosion 3.5% NaCl
2 wt.% Fine

Medium
Coarse
5 wt.% Fine

Medium
Coarse

Table 5
Test matrix for SS316L.

Sand concentration Solution Sand size Velocity (m s1 )

3 5 7 9 10 12

Corrosion 0 wt.% 0.3 M HCl

0.5 wt.% Fine

Medium
Coarse

1 wt.% Fine

Medium

Coarse
2 wt.% Fine

Erosion 0.1 M NaOH Medium
Coarse
3 wt.% Fine

Medium
Coarse
5 wt.% Fine

Medium
Coarse

1 wt.% Fine

Medium

Coarse
3 wt.% Fine

Erosioncorrosion 0.3 M HCl Medium
Coarse
5 wt.% Fine

Medium
Coarse

Table 6
Test matrix for NAB.

Sand concentration Solution Sand size Velocity (m s1 )

5 7 9

Corrosion 0 wt.% 3.5% NaCl

Erosion 1 wt.% 0.1 M NaOH Fine

Medium
Coarse

Erosioncorrosion 1 wt.% 3.5% NaCl Fine

Medium
Coarse
248 S.S. Rajahram et al. / Wear 267 (2009) 244254

Table 7
Compilation of mass loss rate, error and condence level (CL) obtained from the various test time durations for different erosioncorrosion tests.

Variable Level AISI 1020 in NaOH (mg h1 ) AISI 1020 in NaCl (mg h1 ) SS316L in NaOH (mg h1 )

Rate Error CL Rate Error CL Rate Error CL

0.5 1.5687 0.1335 0.99640 12.3025 0.1624 0.99991 0.9228 0.0413 0.99900
1 2.1274 0.2329 0.98823 16.1755 0.0943 0.99997 1.7611 0.0815 0.99787
Sand concentration (%)
2 4.8180 0.0897 0.99983 18.6633 0.0922 0.99999 3.3228 0.2270 0.99767
5 10.7205 0.0762 0.99997 26.3642 0.1091 0.99999 8.2325 0.2378 0.99958

Fine 1.0791 0.0751 0.99759 16.2400 0.7753 0.99886 1.2733 0.0044 0.99999
Sand size Medium 2.1274 0.2329 0.98823 16.1755 0.0943 0.99997 1.7611 0.0815 0.99787
Coarse 1.8713 0.0640 0.99942 16.7415 0.9461 0.99841 1.3346 0.0264 0.99980

3 0.3057 0.0594 0.98163 12.1420 0.1153 0.99995 0.4843 0.0738 0.98859

5 0.8785 0.0499 0.99839 12.4540 0.3917 0.99951 1.0525 0.1039 0.99516
Speed (m s1 ) 7 2.1274 0.2329 0.98823 16.1755 0.0943 0.99997 1.7611 0.0815 0.99787
10 6.5178 0.1955 0.99955 24.7228 0.7011 0.99960 4.3705 0.4558 0.99461
12 14.5988 1.3601 0.99569 30.8208 4.1761 0.99094 6.8204 0.5365 0.99692

Synergy measurements

NAB (mg h1 ) AISI 1020 (mg h1 ) SS316L (mg h1 )

Water 4.7618 0.3039 0.99797 2.8627 0.0608 0.99932 3.1745 0.1556 0.99761
Solution (7 m s1 , 1 wt.% medium sand)
NaCl 3.4645 0.3480 0.99499 16.1755 0.0943 0.99997 2.9244 0.0961 0.99892
NaOH 2.1274 0.2329 0.98823 1.7611 0.0815 0.99787
HCl 31.9258 0.9012 0.99920 3.0954 0.2685 0.99256

is skewed towards low error points. This shows that the gravimetric
techniques adopted for measurements in these experiments have
minimal high error values. To determine the mass loss rates, mul-
tiple tests were performed for a single test parameter at different
durations. The reason for this is to determine the average mass loss
rate in mg h1 and its relationship with increasing duration. In the
current work, a linear relationship is observed for the mass loss
with increasing duration (see Fig. 5). The slope of the linear t pro-
vides the mass loss rate value at a given duration while the error
is obtained from the slope. A condence level is produced to pro-
vide information on the linearity as well as the overall error. A value
closer to unity for the linear t indicates better linearity and lower
error between the mass loss with test duration. Table 7 tabulates
the mass loss rate, error and condence level obtained from the
various test durations.
The calculated condence level was extracted and sorted into a
population graph as shown in Fig. 6. The average condence level Fig. 5. Mass loss rate relationship with increasing time duration.
was determined to be 0.9972, and most of the data fell above 0.994,
providing condence in the linear relationship of mass loss rate.
This is also reinforced by the count of percentage error shown in Fig. 7 that demonstrates the experiments conducted have an
error rate of less than 10% with an average of 5.5%, only three val-
ues recorded errors of more than 10%. The low relative errors in
the experiments also provide condence on the reproducibility of

Fig. 4. Population graph of gravimetric measurements standard deviation with aver- Fig. 6. Population graph of condence level with average, maximum (max), mini-
age. mum (min) and overall number (N).
 S.S. Rajahram et al. / Wear 267 (2009) 244254 249

Fig. 7. Population graph of relative error of mass loss rate with average, maximum
(max), minimum (min) and overall number (N).
Fig. 9. Comparison of erosion and erosioncorrosion rates in different test solutions
for SS316L and AISI 1020 (7 m s1 , 1% sand concentration, medium sand).
the experimental data. The relative errors for the mass loss rate
obtained from the experiments were analyzed further to investi-
gate the error distribution and Fig. 8 shows the error variation with tions, attributed to the chromium oxide lm layer that minimizes
increasing mass loss rate. The majority of the high relative errors charge transfer to and from the material. In order to obtain positive
occurred during the measurement of low mass loss rates. This trend synergy and to assess the erosioncorrosion performance of SS316L,
is not surprising since at lower mass loss rates, the small differences the range of test solutions were extended by increasing the molarity
in the samples before and after testing would inuence the error in of HCl to 0.3 M and 0.5 M. It has been shown that the passive lm of
weighing the samples. SS316L becomes unstable in HCl due to the aggressive chloride ions
acting on the lm [25]. Fig. 10 shows the mass loss variation with
3.2. Test solutions increasing HCl concentration, it was observed that even an increase
to 0.3 M HCl, the passivity of the chromium oxide lm becomes
In order to ascertain the behaviour of the materials in an erosion unstable in erosioncorrosion conditions leading to an increased
dominated (passive) condition and a corrosion dominated (active) mass loss rate. The mass loss rate becomes more signicant at 0.5 M
condition, a range of test solutions were used on SS316L and AISI HCl. From these initial tests, an informed decision was made on the
1020 to decide the best test solution for this work. The solutions choice of test solution for erosion and erosioncorrosion conditions
tested were from a wide range of pH values including acidic pH for both materials. 0.1 M NaOH was chosen as the test solution for
1 (0.1 M HCl), neutral pH 7 (distilled water) and alkali pH 14 erosion experiments as it exhibited the most passive behaviour for
(0.1 M NaOH). In order to simulate seawater conditions 3.5% NaCl both materials. 3.5% NaCl was chosen as the test solution for AISI
was also used. Besides deciding on a suitable test solution for ero- 1020 in erosioncorrosion conditions while 0.3 M HCl was chosen
sion and erosioncorrosion conditions, this part of the work also for SS316L. The choice of test solutions was based on the criteria
aims to look at synergy encountered by these materials and identify for assessing positive synergy on both materials which will be dis-
a test solution that would produce positive synergy, which is of pri- cussed further in Section 3.6. The second criterion for the choice
mary interest to researchers in the eld of erosioncorrosion. Fig. 9 of the test solution was based on the aggresivity of the test solu-
shows the results of the erosion and erosioncorrosion experiments tion towards the test rig which eliminated 0.5 M HCl from further
conducted with the various test solutions. AISI 1020 exhibited rel- consideration in the current work.
atively low mass loss rates in 0.1 M NaOH and distilled water while The effect of increasing the sand concentration on erosion
it actively corroded in 3.5% NaCl and 0.1 M HCl. SS316L in contrast and erosioncorrosion is depicted in Fig. 11. A similar relation-
exhibited passivity in all test solutions with low mass loss rates. This ship is observed in all four test conditions where the increase
shows that SS316L has good corrosion resistance to all the test solu-

Fig. 10. Evaluation of mass loss for SS316L in various test solutions (7 m s1 , 1% sand
Fig. 8. Distribution of percentage error with increasing mass loss rate. concentration, medium sand (294 m)).
250 S.S. Rajahram et al. / Wear 267 (2009) 244254

Fig. 11. Mass loss rate with sand concentration variation (7 m s1 , medium sand).

in sand concentration results in an increase in the erosion and Fig. 12. Mass loss rate with linear velocity variation (1% sand concentration, medium
erosioncorrosion mass loss rate for SS316L and AISI 1020. For both sand (294 m)).
metals, the erosion mass loss rate increased with sand concen-
tration but was lower than the erosioncorrosion mass loss rate.
AISI 1020 exhibited a signicantly higher mass loss rate compared mass loss rates than SS316L in both conditions. It is known that
to SS316L during erosioncorrosion conditions which is expected the velocity exponent for the erosion rate increases to the power
for a less noble metal. The erosion rate is known to increase with of between 2 and 3 for ductile material [30]. In the current work,
sand concentration because of the increase in the number of par- the velocity exponent for AISI 1020 and SS316L in pure erosion
ticles impacting the surface causing material loss [26]. However, conditions was evaluated to be 2.7 and 2.1 respectively, values
this relationship is not that simple as it has also been seen that the consistent for ductile metals. This relationship is more complex
increase in sand concentration can lead to a decrease in erosion for erosioncorrosion due to the presence of the additional cor-
rate attributed to the rebounding of the particles near the surface rosion effect which also varies with velocity. For pure erosion, the
of the material, protecting the surface from subsequent impacts increase in erosion rate with velocity is associated with the increase
[27]. No such effect was observed in this work possibly due to in kinetic energy of the erodents causing more effective damage on
the low sand concentrations (up to 5%) used in this experiment the metal surface. For corrosion, the increase in velocity increases
where the screening effect is not observed. The higher increase the effective mass transfer between the metal and corrosive uid
in mass loss rate with sand concentration for both materials in hence accelerating the electrochemical process of corrosion. AISI
erosioncorrosion conditions on the other hand is believed to be 1020, which is susceptible to corrosion, suffers severe mass loss in
due to the contribution of corrosion in accelerating the wear rate. erosioncorrosion conditions due to the combined effect of veloc-
SS316L exhibits a lower mass loss rate as compared to AISI 1020 ity in both processes. For a passive metal like SS316L, the increase
due to the ability of SS316L to form a chromium oxide rich pas- in velocity accelerates the mechanical removal rate of the passive
sive lm which inhibits corrosion. Although the rate of removal lm thus accelerating the overall erosioncorrosion process [28]. At
of the passive lm for SS316L is dependent on the percentage of the same time the increase in ow velocity can also accelerate the
solids [28], the low erosioncorrosion rates show that the passive repassivation rate by increasing the supply of dissolved oxygen and
lm is fast reforming in this material. More recent studies by Hu chromium ions necessary to repair the chromium oxide lm that
and Neville [29] identied a critical solid loading above which the has been destroyed by solid particle impingement [31]. The rate of
depassivationrepassivation process becomes critical for a passive depassivation and repassivation then becomes critical in enhancing
metal as it results in a higher charge transfer from the material asso- or suppressing the total erosioncorrosion rate for SS316L.
ciated with increasing particle impacts. This nding is considered
important for passive metals as it proves that there is an upper and 3.4. Sand size variation
lower threshold for erosioncorrosion wear rate. When the sand
concentration is lower than the critical limit, the effective removal One of the many factors that affect erosion and erosion
of the passive lm to accelerate mass transfer is minimal, similarly corrosion rates is the particle size. Fig. 13 shows the variation of
above the critical sand concentration particle rebounding prevents mass loss rate with particle size. Tests conducted on both ero-
effective material removal to occur as it prevents incoming particles sion and erosioncorrosion for SS316L and AISI 1020 showed that
from impacting the surface. medium sand produced the highest mass loss rate followed by
coarse sand with the lowest rates seen for tests with ne sand con-
3.3. Velocity variation ditions. This nding is rather interesting as it implies that there
exists an optimal particle size range in a slurry pot erosion tester
As the velocity increases, the erosion and erosioncorrosion that causes maximum erosion and erosioncorrosion which lies
rates is found to increase for AISI 1020 and SS316L (see Fig. 12). between 100 m and 600 m. The concept of collision efciency
However at low speeds of 3 m s1 and 5 m s1 the mass loss dif- dened as the ratio of the number of particles striking the surface
ference between both velocities is minimal. This can be explained and the number of particles contained within the volume of sus-
by the ow eld effect in the slurry pot. At low velocities, the ero- pension swept by that area in unit time is a useful parameter in
dents are not completely suspended in the uid therefore effective examining the particle size effect in erosion process [32]. Smaller
erosion is unable to take place. A higher increase in mass loss rate particles are said to suffer from particle retardation prior to impact
with velocity variation for erosioncorrosion conditions is observed resulting lower collision efciency of the particles. This will lead to
compared to pure erosion conditions with AISI 1020 having higher a lower dissipation of kinetic energy at impact and decrease the ero-
 S.S. Rajahram et al. / Wear 267 (2009) 244254 251

become larger, the possibility of particles rebounding and impact-

ing incoming particles might also increase similar to the effect of
increasing the sand concentration which causes the screening
effect. This provides a possible explanation on the optimal sand
size for maximum erosion. The threshold for particle size effect is
also similar for erosioncorrosion where the increase in particle
size leads to an increase in the erosioncorrosion rate but above a
critical value, the erosioncorrosion rate is independent of particle
size [20].

3.5. Erosioncorrosion mechanisms

Fig. 14 shows the SEM micrographs taken for erosioncorrosion

experiments at 7 m s1 , 1% sand concentration and medium sand
size (294 m) for SS316L, AISI 1020 and NAB. The SEM micrographs
for SS316L (Fig. 14(a) and (b)) and NAB (Fig. 14(e) and (f)) shows
Fig. 13. Mass loss rate with sand size variation (7 m/s, 1% sand concentration). surface roughening with the formation of multiple superimposed
craters and platelets as the main mechanism for erosioncorrosion,
sion rate. It was also noted that for particles less than 100 m, the indicating these two materials undergo multiple impacts leading to
particles might fail to rebound from the surface due to the squeeze the formation of vulnerable lips that can be removed by subsequent
lm effect and at smaller sizes they might even fail to penetrate the impacts. The SEM micrographs show similar mechanisms to the
squeeze lm and impact the surface [7]. However, as the particles cutting mechanism proposed by Hutchings where the type of cut-

Fig. 14. SEM Images of a test samples subjected to erosioncorrosion taken at different magnications (a) SS316L 2000, (b) SS316L 3000, (c) AISI 1020 2000, (d) AISI
1020 3000, (e) NAB 2000, (f) NAB 3000 (test conditions: velocity 7 m s1 , sand concentration 1%, medium sand (294 m)).
252 S.S. Rajahram et al. / Wear 267 (2009) 244254

ting depends on whether the particle rolls forward or backwards to

remove the material from the surface [33]. It can be deduced that for
SS316L and NAB, erosion dominated the overall erosioncorrosion
process. For AISI 1020, shown in Fig. 14(c) and (d), the mechanism
that occurs for erosioncorrosion is slightly different to SS316L
and NAB. Although surface roughening due to particle impacts
are still observed, the formation of lips and platelets are minimal
and in contrast the surface shows formation of craters, indenta-
tions and pits. One possible explanation for this is that for AISI
1020 under erosioncorrosion conditions, electrochemical corro-
sion plays a dominant role in material removal by attacking the
weakened layer of the eroded section causing material removal,
exposing the indented area. Subsequent corrosion attack on this
indented area causes the formation of corrosion pits. The lips which
were formed could have also been removed by the initiation of
cracks at the root of the distressed lips due to corrosion attack which Fig. 16. Classication of materials tested by electrochemical loss (TE) versus
mechanical loss (E).
is vulnerable to material removal [1]. From the SEM micrographs
a clear distinction can be made on the erosioncorrosion mecha-
nisms for an erosion dominated material (SS316L and NAB) and a erosioncorrosion of cast NAB for velocities between 3.1 m s1 and
corrosion dominated material (AISI 1020). 6 m s1 , negative synergies of up to 31.4% were observed infer-
ring that NAB has good resistance under the combined action of
3.6. Synergy erosion and corrosion. NAB forms a thick adherent oxide layer of
cuprous oxide (Cu2 O) and alumina (Al2 O3 ), which remains intact
AISI 1020 exhibited negative synergy in 0.1 M NaOH and 0.1 M under solid particle impingement suppressing erosioncorrosion
HCl and only exhibited positive synergy in 3.5% NaCl as shown in particularly at low kinetic energies.
Fig. 15. Although these results are expected for a passive solution Comparing all three test materials under erosioncorrosion
like 0.1 M NaOH, the existence of negative synergy in 0.1 M HCl is conditions in 3.5% NaCl, SS316L is observed to have the lowest
rather surprising since it is expected that in an aggressive acidic material loss with a mass loss rate of 2.9 mg h1 followed by NAB
condition like 0.1 M HCl, positive synergy would occur. Examining (3.46 mg h1 ) and AISI 1020 (16.18 mg h1 ) (see Table 7). However,
the synergy trends in SS316L, negative synergy was observed on in terms of synergistic ranking, NAB is seen to have the highest
all test solutions except 0.3 M HCl where a substantial increase of negative synergy (63.9%) in erosioncorrosion conditions sug-
positive synergy was seen. In erosioncorrosion studies conducted gesting that this material has the best resistance to the combined
by Matsumura et al. [25] with 5% hydrochloric acid, it was found action of erosion and corrosion followed by SS316L (11.5%) while
that the chloride ions adsorbed on the surface for SS304 inhibited AISI 1020 exhibited positive synergy (5.5%) indicating that this
erosioncorrosion which explains the negative synergy behaviour materials suffers additional wear due to the interaction between
on both materials at low concentrations of HCl. The study further erosion and corrosion. Looking at the SEM surface for AISI 1020 it
shows that on passive metals such as stainless steel, the rate of could be deduced that corrosion attacks the distressed lips formed
synergy is increased by the act of solid particles damaging the sur- during erosion hence accelerating the total wear rate. The rank-
face which increases the corrosion activity hence the synergism. An ing provides useful information as it shows that a material that
important nding from their studies is that the erosioncorrosion is erosioncorrosion resistant does not necessarily suffer from
rates in a stable passive lm environment (1% sulphuric acid) are low synergy levels. Therefore careful judgement should be made
higher than those in an environment where the passive lm is when comparing material loss rate and synergy rates to make an
unstable (5% hydrochloric acid) which shows that the passive lm informed decision on the most suitable material for a given appli-
is not the only factor in determining the synergism of the material. cation.
NAB was only tested in 3.5% NaCl and a high negative synergy of Another method for classifying erosioncorrosion data is by
64% was observed. In the work conducted by Barik et al. [34] on the plot of electrochemical loss (TE) versus mechanical loss
(E) as shown in Fig. 16 for analyzing positive synergy [4]. In
this map erosioncorrosion is divided into four regimes namely
corrosion regime, erosion regime, erosioncorrosion regime
and corrosionerosion regime. The corrosion dominated region
(TE 10E) indicates that the material losses due to electrochemical
process is equal to or more than ten times the losses due to mechan-
ical process and vice versa for the erosion dominated region. The
effect of sand size variation is plotted on this graph. It can be seen
that for AISI 1020, the mass loss rates fall between the corrosion
and corrosionerosion regime indicating that this material is dom-
inated by active corrosion which is consistent for AISI 1020 which is
susceptible to corrosion. However for SS316L in 0.3 M HCl although
still classied under the corrosionerosion regime with one point
lying on the erosioncorrosion regime, the inuence on erosion and
corrosion is seen as almost proportionate to each other exhibit-
ing the passivity of the metal suffering corrosion when there are
solid particles impinging on it. The experimental data from the cur-
rent work was compared with data from published literature for
Fig. 15. Synergy measurements for the three material tested variation (7 m s1 , 1% erosioncorrosion of carbon steel [3] and stainless steel [28,35,36].
sand concentration, medium sand (294 m)). A common trend is indentied where the points for carbon steel in
 S.S. Rajahram et al. / Wear 267 (2009) 244254 253

all cases exhibit higher electrochemical loss at almost similar rates, 7. Comparing all three test materials under erosioncorrosion
varying only in mechanical loss as the test conditions are changed. conditions in 3.5% NaCl, SS316L is seen to have the lowest mate-
In contrast, points for SS316L vary vertically along the wear map rial loss with a mass loss rate of 2.9 mg h1 followed by NAB
indicating varying electrochemical loss with change in test condi- (3.46 mg h1 ) and AISI 1020 (16.18 mg h1 ). However in terms of
tions. A conclusion that can be made from this plot is that for a synergistic ranking, NAB is seen to have the highest negative syn-
material that is corrosion dominated such as carbon steel, the cor- ergy (63.9%) in erosioncorrosion conditions suggesting that
rosion rate is already high even without the presence of erosion this material has the best resistance to the combined action
therefore the electrochemical loss is not signicant. The electro- of erosion and corrosion followed by SS316L (11.5%) while
chemical loss of SS316L on the other hand varies signicantly in the AISI 1020 exhibited positive synergy (5.5%) indicating that this
presence of erosion because the removal of the chromium oxide material suffers additional wear due to the interaction between
layer by solid particle impingement accelerates the corrosion rate erosion and corrosion.
substantially.
Acknowledgements
4. Conclusions
This work was partly sponsored by National Physical Laborato-
ries (NPL). The authors would like to thank Mark Gee and Andrew
A set of erosion, corrosion and erosioncorrosion experiments
Gant (NPL) for supply of materials and technical help.
were carried out to evaluate the repeatability of erosioncorrosion
experiments and to investigate the role of different parameters
inuencing erosioncorrosion. Three test materials namely stain- References
less steel 316L, carbon steel (AISI 1020) and NAB were tested to
[1] Y. Li, G.T. Burstein, I.M. Hutchings, The inuence on the erosion of aluminium
determine synergy levels in these materials and were ranked rela- by aqueous silica slurries, Wear 186187 (1995) 515522.
tive to its erosioncorrosion resistance. The following conclusions [2] R.J.K. Wood, S.P. Hutton, The synergistic effect of erosioncorrosion: trends in
can be made from the current work: published results, Wear 140 (1990) 387394.
[3] R.J.K.Wood, Erosion-corrosion synergism for multi-phase owline materials,
Proceedings of the 3rd International Conference on Cavitation, Inst. Mech. Eng,
Cambridge, Paper C453/011, (1992), 151-158.
1. The low error rates and the high condence levels obtained from
[4] T.J. Harvey, J.A. Wharton, R.J.K. Wood, Development of synergy model for
the experiments showed that the measurements and slurry pot erosion-corrosion of carbon steel in a slurry pot, Tribol. Mater. Surf. Interfaces
erosion rig provide repeatable and reproducible test results. The 1 (1) (2007) 3347.
[5] H.M. Clark, K.K. Wong, Impact angle, particle energy and mass loss in erosion
error in calculation for erosioncorrosion rates was found to have
by dilute slurries, Wear 186187 (1995) 454464.
an average of 5.5% and only a few more than 10%. [6] H.M. Clark, Particle velocity and size effects in laboratory slurry erosion mea-
2. AISI 1020 exhibited passive behaviour in 0.1 M NaOH and dis- surements OR. . . Do you know what your particles are doing? Tribol. Int. (2002)
tilled water but was actively corroding in 3.5% NaCl and 0.1 M 617624.
[7] H.M. Clark, R.B. Hartwich, A re-examination of the particle size effect in slurry
HCl with mass loss increase of between 14 mg h1 and 30 mg h1 . erosion, Wear 248 (2001) 147161.
SS316L in contrast, exhibited passivity in all test solutions with [8] H.M. Clark, On the impact rate and impact energy of particles in a slurry pot
mass loss rates of less than 2 mg h1 . The test conditions were erosion tester, Wear 147 (1991) 165183.
[9] H.M. Clark, Specimen diameter, impact velocity, erosion rate and particle den-
extended by increasing the molarity of HCl to study the effect of sity in a slurry pot erosion tester, Wear 162164 (1993) 669678.
erosioncorrosion on SS316L. It was found that at 0.3 M HCl and [10] J.B. Zu, I.M. Hutchings, G.T. Burstein, Design of a slurry erosion test rig, Wear
0.5 M HCl, SS316L started producing positive synergy with mass 140 (1990) 331344.
[11] R.J.K. Wood, J.A. Wharton, A.J. Speyer, K.S. Tan, Investigation of
loss rates of between 4.8 mg h1 and 7.7 mg h1 . erosioncorrosion processes using electrochemical noise measurement,
3. Increasing the velocity and the sand concentration produced Tribol. Int. 35 (2002) 631641.
higher mass loss rates for AISI 1020 and SS316L. Mass loss [12] A.J. Speyer, Wear corrosion sensing in owing seawater, Ph.D. Thesis, University
of Southampton, 2002.
increase of between 8 mg h1 and 18 mg h1 is seen for AISI
[13] K.S. Tan, J.A. Wharton, R.J.K. Wood, Solid particle erosioncorrosion
1020 and SS316L in erosion-corrosion conditions. The increase behaviour of a novel HVOF nickel aluminium bronze coating for marine
in kinetic energy of the particles is suggested to be the reason for applicationscorrelation between mass loss and electrochemical measure-
ments, Wear 258 (2005) 629640.
the higher mass loss rates.
[14] A. Neville, T. Hodgkiess, J.T. Dallas, A study of the erosioncorrosion behaviour
4. Medium sand (294 m) showed the highest mass loss rate of engineering steels for marine pumping applications, Wear 186187 (1995)
followed by coarse sand and ne sand in erosioncorrosion con- 497507.
ditions for both SS316L and AISI 1020. The difference in mass loss [15] H.M. Clark, H.M. Hawthorne, Y. Xie, Wear rates and specic energies of some
ceramic, cermet and metallic coatings determined in the Coriolis erosion tester,
between medium sand and ne sand was found to be between Wear 233235 (1999) 319327.
3.6 mg h1 for SS316L and 0.6 mg h1 for AISI 1020. [16] H.M. Clark, J. Tuzson, K.K. Wong, Measurements of specic energies for erosive
5. SEM micrographs showed that a similar erosioncorrosion wear using a Coriolis erosion tester, Wear 241 (2000) 19.
[17] Y. Xie, H.M. Clark, H.M. Hawthorne, Modelling slurry particle dynamics in the
mechanism is seen for SS316L and NAB with the formation Coriolis erosion tester, Wear 225229 (1999) 405416.
of multiple extruded lips and platelets typically seen for ero- [18] R.J.K. Wood, T.F. Jones, Investigations of sand-water induced erosive wear of
sion dominated materials. In contrast AISI 1020 surface revealed AISI 304L stainless steel pipes by pilot-scale and laboratory-scale testing, Wear
255 (2003) 206218.
the formation of craters, pits and shallow indentations which [19] R.J.K. Wood, T.F. Jones, J. Ganeshalingam, N.J. Miles, Comparison of predicted
suggests that the material is corrosion dominated. The surface and experimental erosion estimates in slurry ducts, Wear 256 (2004) 937947.
features observed from SEM explains the reason why AISI 1020 [20] M.M. Stack, J.S. James, Q. Lu, Erosioncorrosion of chromium steel in a rotating
cylinder electrode system: some comments on particle size effects, Wear 256
suffered signicantly higher mass loss rates, as corrosion was
(2004) 557564.
accelerating wear by attacking the distressed lips formed during [21] M.M. Stack, H.W. Wang, Simplifying the erosioncorrosion mechanism map for
particle erosion. erosion of thin coatings in aqueous slurries, Wear 233235 (1999) 542551.
[22] M.M. Stack, H.W. Wang, W.D. Munz, Some thoughts on the construction of
6. The additional wear rate due to the interaction between erosion
erosioncorrosion maps for PVD coated steels in aqueous environment, Surf.
and corrosion (synergy) was evaluated for SS316L, NAB and AISI Coat. Technol. 113 (1999) 5262.
1020 in different test solutions and positive synergy of 5.5% and [23] K.R. Trethewey, J. Chamberlain, Corrosion for Science and Engineering, 2nd ed.,
57% was evident for AISI 1020 in 3.5% NaCl and SS316L in 0.3 M HCl Longman Scientic and Technical, 1995, pp. 341347.
[24] M.G. Gee, N.J. McCormick, L. Newton, R. Darling, Analysis of the size and shape
respectively. NAB exhibited negative synergy (63.9%) in 3.5% of abrasives, NPL Report No.: DEPC MPE 009, National Physical Laboratories,
NaCl. Teddington, UK, February 2005.
254 S.S. Rajahram et al. / Wear 267 (2009) 244254

[25] M. Matsumura, Y. Oka, H. Hiura, M. Yano, The role of passivating lm in pre- [31] P. Roberge, B.C. Syrett, ErosionCorrosion, NACE International, 2004.
venting slurry erosioncorrosion of austenitic stainless steel, ISU Int. 31 (1991) [32] R.S. Lynn, K.K. Wong, H.M. Clark, On the particle size effect in slurry erosion,
168176. Wear 149 (1991) 5571.
[26] W. Tsai, J.A.C. Humphrey, I. Cornet, A.V. Levy, Experimental measurement of [33] I.M. Hutchings, Tribology-Friction and Wear of Engineering Materials, Edward
accelerated erosion in a slurry pot tester, Wear 68 (1981) 289303. Arnold, 1992.
[27] S. Turrene, M. Fiset, J. Masounave, The effect of sand concentration on the [34] R.C. Barik, J.A. Wharton, R.J.K. Wood, K.S. Tan, K.R. Stokes, Erosion and
erosion of materials by a slurry jet, Wear 133 (1989) 95106. erosioncorrosion performance of cast and thermally sprayed nickel-
[28] B.W. Madsen, Measurement of erosioncorrosion synergism with a slurry wear aluminium bronze, Wear 259 (2005) 230242.
test apparatus, Wear 123 (1988) 127142. [35] Y.G. Zheng, Z.M. Yao, W. Ke, Erosioncorrosion resistant alloy development for
[29] X. Hu, A. Neville, The electrochemical response of stainless steels in liquid solid aggressive slurry ows, Mater. Lett. 46 (2000) 362368.
impingement, Wear 258 (2005) 641 648. [36] T. Hodgkiess, A. Neville, S. Shrestha, Electrochemical and mechanical inter-
[30] M.M. Salama, An alternative to API 14E erosional velocity limits for Sand-Laden actions during erosioncorrosion of a high velocity oxy-fuel coating and a
uids, J. Energy Resour. Technol. 122 (2000) 7177. stainless steel, Wear 233235 (1999) 623634.