Print Main Menu Zoom In Zoom Out Previous View Next Page

CLARKE’S SOLUTION CLEANING USED

FOR CORROSION PRODUCT REMOVAL:
EFFECTS ON CARBON STEEL SUBSTRATE
S.A. Wade and Y. Lizama
Faculty of Science, Engineering and Technology, Swinburne University of
Technology, Hawthorn, Victoria, Australia

SUMMARY: Immersion of a test sample in Clarke’s solution (a combination of hydrochloric acid and
inhibitors) is one of the standard techniques used for removal of corrosion products from iron and steel.
If, however, samples are left for too long in the solution there is a possibility that in addition to removing
the corrosion product that the underlying material may also start to be attacked. In this work we report the
results of a range of tests investigating the effect of immersion in Clarke’s solution on the attack of five
different grades of carbon steel. A combination of weight loss measurements, high resolution microscopy
and chemical analysis has been used to look at the effects of immersion time, steel microstructure and
presence of inclusions on the rate and morphology of attack of the different steels.

Keywords: Steel, Corrosion, Clarke’s solution, Inclusions

1. INTRODUCTION
In many corrosion studies it is important to be able to remove the corrosion product that is formed on the surface of a metal
coupon during testing. This is required, for example, in order to be able to accurately determine the rate of loss of metal and
also to be able to determine the extent and morphology of corrosion attack. One of the standard techniques used for removal of
corrosion products from iron and steel is a chemical cleaning procedure using the so-called Clarke’s solution (a combination of
hydrochloric acid and inhibitors) [1, 2]. If samples are left for too long in the solution however there is a possibility that in
addition to removing the corrosion product that the underlying material may also start to be attacked.
The development of Clarke’s solution dates back to the 1930’s, when S.G. Clarke published a paper on the use of inhibitors
with hydrochloric acid for rust removal while protecting the underlying metal [1]. It was shown that strong hydrochloric acid
combined with antimony chloride and stannous chloride provided a suitable cleaning combination. The antimony oxide plays
the main role of inhibiting attack, while the stannous chloride is used to reduce the ferric chloride produced by the rust solution
to ferrous chloride. Ferric chloride was suggested to increase corrosion whereas ferrous chloride appeared not to affect the
steel. Other key points that Clarke noted were that:
 “Vigorous agitation of the liquid, preferably by mechanical means, increases the speed of rust removal, and is essential in
order to bring up supplies of stannous chloride from the bulk of the solution to reduce ferric chloride as fast as it is formed,
thus preventing local accumulation at the surface of the specimens.”
 “The tin-antimony solution should be rejected when it becomes permanently colored yellow by ferric chloride, indicating
oxidation of all the stannous chloride.”
These points have been reiterated here as despite the widespread use of Clarke’s solution, the original paper by Clarke has
received very few references. In addition, the information provided in the ASTM standard [2] is reasonably vague about some
of the details that are important to the best practice use of Clarke’s solution.
Somewhat surprisingly there have been very few reports investigating the effect of Clarke’s solution on rust removal since the
original paper. In 2003 Singh and Kumar published a paper that looked at trying to reduce what they noted to be a variation in
the attack of different grades of steel when cleaned with Clarke’s solution [3]. The difference in the corrosion of the various
steel types was discussed in relation to the presence of copper in steels. It was shown that the addition of copper chloride to the
cleaning solution reduced both the corrosion rate and also the variation in corrosion rate of different steels immersed in
Clarke’s solution. More recently Chaves and co-workers have performed studies of various rust cleaning methods, including
Clarke’s solution [4, 5]. In relation to Clarke’s solution high resolution microscopy revealed that “cleaned” steel samples were
typically found to be covered in a doughnut-shaped surface product of unknown origin.
The work presented in the current paper was motivated by a combination of the results of the aforementioned recent studies of
Clarke’s solution and the authors >10 years use of the solution to clean various grades of carbon steel. The aims were to better
understand if there was any dependence on steel type on the level of attack and also to provide more information on the source
of the doughnut-shaped features reported by Chaves et al. [4, 5].
Corrosion & Prevention 2015 Paper 050 - Page 1
Print Main Menu Zoom In Zoom Out Previous View Next Page

2. EXPERIMENTAL PROCEDURE
A variety of steel grades were used in the work, chosen to cover a range of alloy types including those used in marine (S-
Marine) and land (S-Land) based structural applications. Details of the elemental compositions of the steels tested, as
determined by atomic emission spectroscopy, are provided in Table 1. The coupons from the structural grade steels (S-Marine
and S-Land) were cut from supplied plate to dimensions of approximately 25 mm × 25 mm × 6 mm. The 1020, 1045 and 1214
test coupons were cut from supplied rod and had an average diameter of 30 mm and a thickness of approximately 3.5 mm.
Prior to testing the surface of each of the coupons was ground using silicon carbide papers, finishing with 1200 grit size.
Table 1 Elemental composition in weight percentage of various carbon steels used in this work.

Element (Wt %)
Steel grade
Fe C Mn Si S P Cu Al
1020 bal. 0.18 0.66 0.18 0.02 0.02 0.02 -
1045 bal. 0.50 0.84 0.15 0.02 0.04 0.15 -
1214 bal. 0.06 1.23 0.02 0.37 0.05 0.05 -
S-Marine bal. 0.13 1.22 0.26 0.02 0.03 0.01 0.03
S-Land bal. 0.14 0.65 0.14 0.01 0.01 0.01 0.04

The Clarke’s solution used in tests was prepared as recommended by ASTM G1 [2], with the ratio of 1000 mL hydrochloric
acid (specific gravity 1.19), 20 g antinomy trioxide (Sb2O3) and 50 g stannous chloride (SnCl2). Individual clean steel coupons
were tested by placing them in a beaker containing 75 mL of Clarke’s solution, which was agitated using an ultrasonic bath for
the required duration. Unless otherwise stated, each steel coupon was tested using fresh (i.e. previously unused) Clarke’s
solution. After cleaning the coupons were rinsed thoroughly in a large volume (~1 L) of deionized water followed by an
ethanol rinse and drying with a nitrogen gas flow. Figure 1 shows an example of steel coupons being cleaned in the ultrasonic
bath.

Steel
coupons

Figure 1 Photo of steel coupons being cleaned with Clarke’s solution in an ultrasonic bath.
The main techniques used to determine the effect of Clarke’s solution on the cleaned metal coupons were weight loss and
scanning electron microscopy (SEM, typical voltage 5 kV). Coupons were weighed before and after cleaning using a high
accuracy mass balance. It is worth noting that due to size differences the initial masses of the steel coupons used in this work
varied from approximately 15 to 30 mg. Therefore simply providing absolute mass losses as a base of comparison would be
misleading. To account for this and the typical localised nature of corrosion attack observed, the mass losses due to Clarke’s
solution cleaning are reported as the mass change per the surface area of individual coupons. Multiple measurements (×25)
with the mass balance of the same steel coupon (mass ~28 mg) gave a standard measurement deviation of approximately
0.13 mg. To help further improve the accuracy of mass measurements, coupons were weighed multiple times resulting in
standard errors of <0.1 mg. Each test was normally performed on 3 coupons of each steel type. For the SEM investigation
images were taken at multiple locations across the coupon using magnifications between 100× and 20,000×. Standard digital
optical photographs were also taken of the coupons after cleaning to provide a general indication of the surface.
For the inclusion analysis, coupons of each steel type were first ground using silicon carbide papers (320, 500 and 1200 grit
size). The coupons were then polished through a sequence of fine polishing with 3 and 1 m diamond suspensions, ending with
a 0.025 m finish using colloidal silica suspension. The surface of individual polished coupons was imaged at 6 different
locations using an optical microscope (100× objective). A commercial software package was then used to obtain a quantitative
analysis of the size and number of inclusions present in the images recorded.
To observe the microstructure of the different steel types, coupons polished to a 0.025 m finish (using the method discussed
above) were etched using 2% Nital, a mix of 98 ml absolute ethanol (100%) and 2 ml nitric acid (70%). An optical microscope

Corrosion & Prevention 2015 Paper 050 - Page 2

Print Main Menu Zoom In Zoom Out Previous View Next Page

was subsequently used to record the microstructure with a 20× objective lens. Figure 2 shows examples of the microstructure
of the different grades of steel studied.

1020 1045 1214 S-Marine S-Land

Figure 2 Microstructure of the steel grades used in this study. Areas of pearlite phase are indicated by the black grains, while
the white grains are the ferrite phase. Both the 1020 and 1214 steels also have numerous grey/black spots that denote
inclusions.

3. RESULTS
As is somewhat typical of experimental work, the data reported in this paper is the result of a series of repeat tests with
modifications made to the testing procedure as new insights were obtained. For example in the first sets of tests, measurements
were taken using the same single batch of Clarke’s solution, without changing it from coupon to coupon. Gradually however it
became apparent that the amount of previous use of a batch of Clarke’s solution had a significant effect on the level of attack
sustained by a particular steel type when immersed for the same duration. This is illustrated in Figure 3, which shows SEM
images of the surfaces of coupons of marine grade structural steel (S-Marine) which have been exposed to either fresh or
previously used Clarke’s solution. The exact amount of previous use of the “old” Clarke’s solution was not recorded at the
time of testing but is estimated to be in the range of several hours of cumulative cleaning. It is quite clear that the coupons
immersed in previously used Clarke’s solution had greater levels of attack.

(a) (b) (c)

2 min, fresh Clarke’s solution 5 min, fresh Clarke’s solution 10 min, old Clarke’s solution
(d) (e) (f)

15 min, fresh Clarke’s solution 30 min, fresh Clarke’s solution 30 min, old Clarke’s solution
Figure 3 SEM images of S-Marine steel coupons cleaned with Clarke’s solution for various durations with either freshly
prepared or previously used Clarke’s solution (all images ×5000 magnification).
Figure 3 also shows the effect of immersion time in Clarke’s solution for the marine grade structural steel. The tests performed
in fresh Clarke’s solution are provided in Figures 3a, b, d and e. For the shorter immersion durations (2 and 5 min) the attack
on the steel for fresh Clarke’s solution was quite minimal, with some very sparse pitting that was believed to correspond to
sites of inclusions. As the exposure duration increased (15 and 30 min, Figures 3d and e) some areas of more general attack

Corrosion & Prevention 2015 Paper 050 - Page 3

Print Main Menu Zoom In Zoom Out Previous View Next Page

started to occur in addition to the pitting. Even for the 30 min immersion however the steel surface is not overly affected when
fresh Clarke’s solution is used.
The average changes in the mass of steel coupons after immersion in Clarke’s solution are provided in Table 2. As previously
discussed providing absolute mass changes can be misleading as the magnitude of change is relative to the size of the test
coupon used, i.e. a coupon with larger surface area will have a larger mass loss than a smaller coupon for the same immersion
time. For the sake of comparison however a steel coupon of similar size to those used in the current test subjected to a
corrosion rate of 25 m/yr (top of low corrosivity category in AS4312-2008 standard [6]), exposed for 30 days would have a
mass loss of 30 mg. Likewise the same coupon subjected to a corrosion rate of 80 m/yr (top of high corrosivity category in
AS4312-2008 standard), exposed for 1 year would have a mass loss of 1.4 g. The values shown in Table 2 confirm that the age
of the Clarke’s solution can have a significant effect on the amount of attack of the metal and that if fresh Clarke’s solution is
used the attack is relatively small.
Table 2 Mass losses of steel coupons immersed in Clarke’s solution, the errors in mass loss were
calculated from the spread of coupon data.

Immersion Mass Loss (mg/cm2)

Steel grade
time Old Clarke’s Fresh Clarke’s
S-Marine 2 0.021  0.002 (0.4 mg) 0.028  0.026 (0.5 mg)
S-Marine 10 0.180  0.147 (6.3 mg) 0.076  0.034 (1.4 mg)
S-Marine 15 0.487  0.187 (8.8mg) 0.031  0.011 (0.6 mg)
S-Marine 30 3.950  1.134 (72.8 mg) 0.166  0.022 (3.0 mg)
1020 15 - 0.498  0.254 (8.3 mg)
1045 15 - 0.433  0.159 (7.5 mg)
1214 15 - 0.272  0.029 (4.7 mg)
S-Land 15 - 0.409  0.161 (7.4 mg)

To further examine how the age of Clarke’s solution affected the level of attack on steel a test was performed using the same
(i.e. unchanged) volume of solution to clean a series of coupons of marine grade structural steel. This involved immersing a
single coupon in fresh Clarke’s solution for 5 min, removing it and measuring mass loss, immersing another coupon in the
same sample of Clarke’s solution for 5 min, removing it and measuring mass loss, and repeating this process. This meant that
for the last test coupon the Clarke’s solution had already been used for a total of 25 min beforehand. The results of this test are
shown in Figure 4 which indicates that the mass loss of the steel coupons definitely appears to increase despite being immersed
for the same time as the amount of previous use of the Clarke’s solution goes up.
4.0
Mass change (mg/cm2 100)

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0
0 5 10 15 20 25 30 35
Cumulative usage time of Clarke's solution (min)

Figure 4 Effect of age of Clarke’s solution on subsequent mass loss per surface area of S-Marine grade steel coupons,
error bars are spread of data points.
In addition to looking at how exposure time and the age of the Clarke’s solution effects steel, tests were undertaken to look at
whether the steel grade used affected the amount or type of attack. One of the initial observations made was that the 1214 steel

Corrosion & Prevention 2015 Paper 050 - Page 4

Print Main Menu Zoom In Zoom Out Previous View Next Page

type appeared to have a higher pit density compared to the other steel types tested, when exposed for the same duration. This is
shown in Figure 5 using relatively low magnification SEM images to provide a general overview of the surfaces of two of the
different steel types (1045 and 1214). It is clear from the images shown in Figure 4 that the 1214 steel has many more areas of
localised attack than the 1045 steel. The other steel grades studied (i.e. 1020, S-Land and S-Marine) showed similar or even
lower levels of pitting attack as the 1045 steel sample shown.

1045 1214
Figure 5 Relatively low magnification SEM images of steel coupons cleaned for 15 min with Clarke’s solution,
the black areas are the localised pitting (images ×1000 magnification).
To clarify the cause of the higher level of pitting in the 1214 steel when immersed in Clarke’s solution tests were performed to
determine if this related to the presence of inclusions in the steel. As shown in Table 1 the 1214 steel has higher levels of
manganese, sulphur and phosphorus which are designed to increase the number of MnS inclusions in this steel to help improve
its machinability [7]. To confirm this SEM scans of polished 1214 steel samples were taken to identify inclusions and then
energy dispersive spectroscopy was performed on these sites, which indicated that the inclusions were MnS. Furthermore
measurements of the percentage of the surface of the steel samples covered by inclusions was carried out, the results of which
are shown in Table 3. The 1214 steel sample was found to have at least ×10 greater percentage of inclusions than any of the
other steel types studied. Finally the diameter of the pits observed were in the order of 1 m, which matches well the recorded
diameter of the inclusions.
Table 3 Percentage of surface coverage by inclusions measured for the different steel types tested.

Steel grade Inclusions (%)

1020 0.16  0.05
1045 0.12  0.08
1214 1.66  0.34
S-Marine 0.03  0.01
S-Land 0.06  0.02

A comparison of the surface of the different steel grades immersed in fresh Clarke’s solution for 2 min is provided in Figure 6.
The SEM images shown are for higher magnification than those in Figure 5 to provide some more detail of what is happening
at the surface. As was previously shown for the marine grade structural steel (see Figure 3) the 2 min immersion tends to cause
some localised pitting, believed to be at inclusion sites, but otherwise the steel surface remains relatively free of attack. Also of
particular interest are Figures 6 a and b which show the surface of 2 different 1020 grade steel coupons immersed separately
for 2 min. The difference between the testing of these coupons was believed to be the degree to which the coupons were rinsed
with water after removing from the Clarke’s solution, with the coupon in Figure 6b receiving only a very quick rinse while
more care was taken with the coupon shown in Figure 6a. It is clear that this has left some products of the cleaning process on
the surface of the coupon in Figure 6b. This difference can be easily seen by the eye, as shown in the photographs of the 2
coupons in Figure 7.

Corrosion & Prevention 2015 Paper 050 - Page 5

Print Main Menu Zoom In Zoom Out Previous View Next Page

(a) (b) (c)

1020 (cleaned well) 1020 (not cleaned well) 1045

(d) (e) (f)

1214 S-Marine S-Land

Figure 6 SEM images of coupons of various steel types cleaned for 2 min with freshly prepared Clarke’s solution
(all images ×5000 magnification).

Rinsed well Not rinsed well

Figure 7 Photographs of 1020 grade steel coupons cleaned for 2 min with Clarke’s solution.
Examples of SEM images of the different steel types after 15 min immersion in fresh Clarke’s solution are provided in Figure
8. These images show that most of the coupons are starting to show signs of increased level of surface attack for this length of
immersion, similar to what was seen earlier in Figure 3. When the spread of results are taken into account the relative mass
losses for the different steels for 15 min immersion (as given in Table 2) are comparable.

Corrosion & Prevention 2015 Paper 050 - Page 6

Print Main Menu Zoom In Zoom Out Previous View Next Page

(a) (b) (c)

1020 Location 1 1020 Location 2 1045

(d) (e) (f)

1214 S-Marine S-Land

Figure 8 SEM images of coupons of various steel types cleaned for 15 min with freshly prepared Clarke’s solution
(all images ×5000 magnification).

4. DISCUSSION
When the steels were immersed in Clarke’s solution for shorter durations (i.e. ≤ 5 min) the attack seemed to be mainly in the
form of pitting. These pits appear to match quite well with both the diameter and number of MnS inclusions present in the
different steel types tested. The preferential attack of MnS inclusions in carbon steels is not a new phenomenon, and has been
well documented by a number of authors (see references 8 and 9 for example). What is good however is that for these short
immersion durations the extent and morphology of attack should be able to be predicted relatively well. This means that for
corrosion tests of steels the attack due to the exposure of interest (e.g. seawater, atmospheric, etc.) should be able to be
distinguished from the changes that are due to the Clarke’s solution. Other than the higher density of pitting attack observed for
the 1214 steel grade the variation in the level of attack on the different steel types did not appear to be great.
The tests reported in this work were undertaken using uncorroded steel coupons to avoid the additional complication of
variations in mass loss due to the corrosion process itself if corroded coupons were used. Despite this a reasonable level of
variation in the extent of coupon attack was observed between tests performed on different occasions with what should have
been the same test parameters. The exact source of the variation is unclear at present and requires further study to resolve. It
does however reiterate the need to use uncorroded control specimens for each test as suggested in the ATSM standard [2].
In previous tests by Chaves et al [4, 5] looking at different steel cleaning methods bulky doughnut shaped features were
observed on the surface of steel coupons cleaned using Clarke’s solution. The source of these additional products was
somewhat unclear. The results of work presented in the current paper (see Figures 6a and b for example) would suggest that
these products are formed during the process of removal of the coupons from the Clarke’s solution and the subsequent rinsing
step to remove any remaining Clarke’s solution. It was found that the cleanest samples were obtained when extra care was
taken during the rinsing, e.g. agitating the coupons in relatively large volumes of water. Further work however to define the
best methods to clean/rinse the samples following removal from the Clarke’s solution would be helpful but haven’t been
undertaken so far due to time limitations.
The work performed also highlights two key issues related to the use of Clarke’s solution, namely that:
 Standards shouldn’t be followed blindly: the ASTM G1-03 [2] doesn’t mention anything about careful washing of
samples following immersion in Clarke’s solution, or about the potential problems with using old solution.
 Always need to read the original papers: a database check indicates that the 1936 paper by Clarke [1] has only been
referenced 5 times. This would tend to suggest that relatively few people have taken the time to read the paper in detail
and/or follow up on the work presented. Therefore it’s probably not that surprising that some of the key issues discussed
in the paper, such as the potential problem with ageing Clarke’s solution, appear possibly not to have been understood or
taken on-board by users of the solution.

Corrosion & Prevention 2015 Paper 050 - Page 7

Print Main Menu Zoom In Zoom Out Previous View Next Page

Finally another point worth discussing is the method used to agitate the Clarke’s solution during cleaning. In the paper by
Clarke [1] it suggests “vigorous agitation of the liquid, preferably by mechanical means” while the ASTM G1-03 standard [2]
notes that the “solution should be vigorously stirred or specimen should be brushed”. In this work we have used ultrasonication
to agitate the solution. There could be a question however as to whether this method provides an adequate level of mixing.
Further studies are required to investigate if other methods of agitating the Clarke’s solution are superior to ultrasonication.

5. CONCLUSIONS
The cleaning of steel samples using Clarke’s solution is a standard technique commonly used in corrosion research. Despite
this there appears to have been little work undertaken to investigate how the specifics of the cleaning procedure, the steel type
or the solution itself affect the level or type of attack on the steel. The current work reaffirms some of the critical, but
potentially overlooked, points made in the 1936 paper by Clarke, and offers the following advice to those using the solution:
 For shorter immersion times the steel inclusions are preferentially attacked by Clarke’s solution, and the morphology of
the inclusion attack is reasonably distinctive.
 While the pitting attack of the 1214 steel was proportionally higher for shorter duration immersion, overall there was
minimal variation in mass loss observed in this work between the different steel types.
 Fresh Clarke’s solution should always be used to minimise the extent of attack on the steel
 For the majority of tests performed using fresh Clarke’s solution the weight loss due cleaning was relatively small even
after 30 min immersion.
 To avoid surface contamination due to the cleaning process it is critical to wash coupons thoroughly with H2O after
immersion in Clarke’s solution

6. ACKNOWLEDGMENTS
The authors would like to thank Muhammed Awais Javed (Swinburne University) and Robert Melchers, Robert Jeffrey and
Igor Chaves (University of Newcastle) for useful discussions in relation to the use of Clarke’s solution for cleaning of steel
samples.

7. REFERENCES
1. Clarke SG, The use of inhibitors (with special reference to antimony) in the selective removal of metallic coatings and
rust, Trans. Electrochem. Soc. 69(1) (1936) 131-144.
2. ASTM, Standard practice for preparing, cleaning, and evaluating corrosion test specimens [standard G1-03], in: ASTM
Int. ASTM Handbook 3.02 Corrosion of Metals, Wear and Erosion, West Conshohocken, PA, 2011, 20-28.
3. Singh DDN and Kumar A, A Fresh Look at ASTM G 1-90 Solution Recommended for Cleaning of Corrosion Products
Formed on Iron and Steels, Corrosion 59(11) (2003) 1029-1036.
4. Chaves IA, Jeffrey R and Melchers RE, The effect of cleaning procedures on corrosion coupon surface topography, (In)
Proc. Corrosion and Prevention Sept 21-24, 2014 Darwin, Australia, Paper 45.
5. Chaves IA, Jeffrey R and Melchers RE, Rust removal from steel coupons after short term marine immersion, Corrosion
(NACE) (in press 2015).
6. Atmospheric corrosivity zones in Australia [standard AS4312], Standards Australia 2008.
7. Atlas Specialty Metals, Technical Handbook of Bar Products 2006.
8. Wranglen G, Pitting and sulphide inclusions in steel, Corrosion Science 14(5) (1974) 331-349.
9. Avci R, Davis BH, Wolfenden ML, Beech IB, Lucas K and Paul D, Mechanism of MnS-mediated pit initiation and
propagation in carbon steel in an anaerobic sulfidogenic media, Corrosion Science 76(7) (2013) 267-274.

Corrosion & Prevention 2015 Paper 050 - Page 8

Print Main Menu Zoom In Zoom Out Previous View Next Page

8. AUTHOR DETAILS
Dr Wade joined Swinburne University of Technology in 2009 and is a Senior Lecturer and
Program Coordinator of the Biomedical Engineering degree in the Faculty of Science,
Engineering and Technology. He has worked on various aspects of fundamental corrosion
research as well as collaborating with leading industry and research organisations to provide
information and guidance on applied corrosion problems. His current research interests include
microbiologically influenced corrosion, marine corrosion, biofouling, corrosion resistant coatings
and the development corrosion sensors.

Y Lizama joined Swinburne University of Technology as a Materials Characterisation Technician

in 2012. She has over 25 years of experience in the field of metallurgy including mechanical
testing and metallography of a diverse range of materials. Prior to Swinburne University she has
worked at the BHP Research Laboratories and Monash University (CAST CRC, Institute
Railway Technology and Centre of Excellence Design in Light Metals). She has an interest in
various aspects of corrosion as well as the development of new materials and microstructural
analysis.

Corrosion & Prevention 2015 Paper 050 - Page 9

View publication stats