Corrosion Science 63 (2012) 404–409

Contents lists available at SciVerse ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Surface roughness and the corrosion resistance of 21Cr ferritic stainless steel
Sang Mok Lee a, Wan Gyu Lee a, Yeong Ho Kim b, Ho Jang a,⇑
a
Department of Materials Science and Engineering, Korea University, Seoul, South Korea
b
POSCO Technical Research Lab, Incheon, South Korea

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

Article history: The influence of surface topography on corrosion resistance of 21Cr ferritic stainless steel was studied.
Received 4 January 2012 The surface topography was modified by changing the grinding conditions such as load, speed, and fluid
Accepted 23 June 2012 type and its correlation with corrosion resistance was examined. Results showed that the corrosion
Available online 29 June 2012
behavior was correlated well with the reduced valley depth of the surface, while other roughness param-
eters including roughness average, root mean square roughness average, skewness, and others showed
Keywords: poor correlations with corrosion behavior. Another salient observation was the strong influence of the
A. Stainless steel
type of the grinding-fluid on the topography of the ground surface.
B. Polarization
C. Pitting Corrosion
Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.

1. Introduction the local concentration of a solution was determined by the geom-

etry of surface pits and valleys, which affected the diffusion of
Corrosion resistance of stainless steel and its correlation with active ions during corrosion.
surface topography have been important concerns for product The surface topography of a metallic component that strongly
sustainability. This is because the dense chromium oxide layer affects the corrosion is determined by surface finishing processes.
on the stainless steel surface prevents corrosion and surface rough- In particular, the surfaces prepared by mechanical processes are
ness can affect this layer. The evidence supporting the correlation strongly affected by the wear processes. Different wear modes
between the surface topography and corrosion behavior of stain- result in different surface finishes with different surface textures.
less steel can be found in the literature. Pistorius and Burstein In order to describe the unique features of the surfaces, various
[1] showed more pits on rougher stainless steel surfaces and roughness parameters have been developed and among them the
suggested that deep depressions on the surface required a lower average surface roughness (Ra) and the root mean square average
potential to activate corrosion than more open pits because of roughness (Rq) are most frequently used parameters for general
different diffusion rates. Hong and Nagumo [2] reported that engineering purposes [12,13]. Investigations of corrosion resistance
smooth surfaces were more resistant to the propagation of corro- as a function of surface roughness have also used Ra (or Rq) as a
sion pits based on their AC impedance measurements of the stain- roughness parameter in many reports. However, they often mislead
less steel plates wet-ground with different abrasive papers. Sasaki the detailed information of the surface features that can affect the
and Burstein [3] also showed that pitting potential decreased electrochemical properties of a surface [12]. There are numerous
linearly with the grit number that determines the final surface parameters available to represent surface characteristics that more
roughness after abrasion. Other experiments using stainless steel precisely describe the surface in terms of local corrosion and passiv-
have reported similar conclusions of improved corrosion resistance ation. These measures include skewness, total roughness (Rt), re-
with surface smoothness [4–8]. Copper and magnesium alloys also duced valley depth (Rvk), reduced peak height (Rpk), and others [12].
showed analogous results [9,10]. Particularly noteworthy was the In this study, the corrosion resistance of 21Cr Ferritic Stainless
work by Li et al. investigating the correlation of surface roughness steel was investigated to examine the influence of surface topogra-
and the electron-work function that determined the corrosion and phy on corrosion resistance after a wet-grinding process. By chang-
electronic behavior of copper [10]. They reported that roughness ing the applied load, grinding speed, and grinding fluids, various
increased local fluctuations in the electron-work function, which surface topographies were produced and a roughness parameter
accelerated corrosion by promotion of microelectrodes on the that correlated well with corrosion resistance was suggested and
surface. The effect of surface roughness on the initiation of pit its mechanism was discussed.
corrosion was also reported by Sharland [11], who suggested that
2. Experimental procedure

⇑ Corresponding author. Tel.: +82 2 32903276. Commercial ferritic stainless steel (ASTM S44536, POSCO),
E-mail address: hojang@korea.ac.kr (H. Jang). which was developed for improved corrosion resistance, was used

0010-938X/$ - see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.corsci.2012.06.031
 S.M. Lee et al. / Corrosion Science 63 (2012) 404–409 405

for this study. The composition of 21Cr stainless steel was provided by ultrasonic cleaning using acetone and ethyl alcohol, respec-
by the manufacturer (POSCO) and listed in Table 1. tively, to remove possible residues in the surface.
A block-on-ring tribotester was used to produce wear tracks on
the stainless steel surface. Before the tribotest, the steel specimens
3. Results and discussion
were ground using SiC abrasive paper (#2000) to produce the same
initial surface roughness.
3.1. Surface roughness
Abrasion tests were carried out using SiC abrasive paper (PC221
Deerfos Co., equivalent to a #180 mesh size) attached around the
The surface topography of the stainless steel plates was ana-
ring in a ring-on-block tribometer sliding against a stainless steel
lyzed using a laser confocal microscope. Surfaces with different
plate. The ring was half-immersed in a grinding fluid. A schematic
roughness were produced by changing the applied load, sliding
of the ring-on-block tribometer is shown in Fig. 1. The plate spec-
speed, and grinding fluid. Fig. 2 shows that the surface roughness
imens were cleaned by washing them with acetone using an ultra-
average (Ra) slightly increased as a function of load, while sliding
sonic cleaner for 5 min followed by rinsing with distilled water,
speed affected more on Ra. The figure also indicates that the type
and drying.
of grinding fluid changed the surface roughness. The Ra values were
Two different commercial grinding fluids were used for abra-
higher when the hydrophilic grinding fluid was used, suggesting
sion tests to simulate the final polishing procedure that has been
that high wet ability of the hydrophilic grinding fluid promoted
carried out to clean stainless steel surfaces before shipping. The
the wear process by removing the wear particles more efficiently
two grinding fluids showed different wetting angles of 54° and
from the sliding interface. The coefficient of friction during the
11° on the steel surface, respectively, indicating dissimilar hydro-
abrasion test was also measured and shown in Fig. 3. The figure
phobicity. In this report, the grinding fluid with a high wetting
indicates that the friction coefficient did not change much as a
angle (54°) was called the hydrophobic fluid (LS2050/LS2080, LUB-
function of load or sliding speed. However, the high wet ability
TEC) and the other with a low wetting angle (11°) as the hydro-
of the hydrophilic grinding fluid provided better lubrication at
philic fluid (Mineral61 Polishing Oil, Houghton Co.). To produce
the sliding interface and led to lower friction levels, while the
various topological characteristics on the steel surface, the applied
hydrophobic grinding fluid increased friction levels due to partially
load (50–250 N) and sliding speed (0.05–0.2 m/s) were also varied.
unlubricated solid contact junctions.
After the tribotest, a laser confocal microscope (Wyko NT-1100,
Other surface roughness parameters were also calculated from
Veeco, USA) was used to measure various surface roughness
the stainless steel surfaces after abrasion tests. However, other
parameters and 3-dimensional topographical information from
parameters such as root mean square average roughness (Rq), total
the ground surface.
roughness (Rt), ten-point height (Rz), and reduced peak height
Corrosion tests were carried out using ground specimens after
(Rpk) showed little influence from grinding conditions (not shown),
cleaning in a potentiodynamic mode in a 1 M NaCl solution at
while the reduced valley depth (Rvk) increased as a function of the
room temperature using a computerized potentiostat/galvanostat
applied load and sliding speed as shown in Fig. 4. Rvk and Rpk repre-
(VMP3-CHAZ, Princeton Applied Research) to evaluate corrosion
sent the percentage of peaks and valleys on the surface, and they are
resistance of the steel specimens after abrasion tests. The cleaning
obtained by measuring the heights of the triangles near 0 and 100%
was carried out by washing the specimen with acetone using ultra-
in the bearing-area-ratio curve. The figure also showed that Rvk was
sonic cleaner for 5 min followed by rinsing with distilled water,
strongly affected by the type of the grinding fluid, suggesting that
and drying. The corrosion circuit used in this study consisted of
the formation of the deep valleys produced after abrasion tests could
three electrodes; a graphite counter electrode, a saturated calomel
be could be closely related with the wet ability of the grinding fluid.
electrode as a reference electrode, and a stainless steel specimen as
Surface topography after abrasion tests in Fig. 5 supports the
a working electrode. After the abrasion tests, the stainless steel
effects of the grinding fluids. The formation of the deep valleys pro-
specimens were thoroughly washed in distilled water followed
duced during the wet-grinding process using a hydrophilic grinding
fluid appeared to be closely related to the effectiveness of wear-
Table 1
debris removal from the valleys in the worn surface. This is because
The composition of the ferrite stainless steel (ASTM S44536, POSCO).
a grinding fluid with high wet ability can effectively remove the
Elements C Cr Ni Mo Ti Nb wear debris from the surface and prevents the valleys from clogging
wt.% <0.015 20  23 <0.5 0 0.3 0.1 by submicron-size wear particles. Another interesting finding from
these abrasion tests was that there was no correlation between Ra
and Rvk [Fig. 6], suggesting that the wear mechanisms determining
the two roughness parameters were different.

3.2. Corrosion resistance

The correlation of corrosion resistance with surface roughness

was examined by calculating the polarization resistance (Rp) of
the stainless steel after abrasion tests. The polarization resistance
was obtained using the linear polarization method, which is inver-
sely proportional to the corrosion rate as obtained from the equa-
tion : Rp = babc/2.3icorr(ba + bc), where ba and bc are cathodic and
anodic Tafel constants and icorr is the corrosion current density ob-
tained from a potentiodynamic curve [14].
The potentiodynamic curves obtained from the two different
steel plates ground with different grinding fluids were compared
in Fig. 7. The two specimens were ground at the same applied load
Fig. 1. A ring-on-block tribometer used in this study with the geometry of the steel (100 N) and speed (0.1 m/s). The roughness parameters after the
plate and the ring. The ring was half immersed into the grinding fluid. abrasion tests and the polarization resistance were compared in
406 S.M. Lee et al. / Corrosion Science 63 (2012) 404–409

2000 2000

1800 1800

1600 1600

1400 1400

Ra(nm)
Ra(nm)

1200 1200

1000 1000

800 800

600 Hydrophobic 600 Hydrophobic

Hydrophilic Hydrophilic
400 400
50 100 150 200 250 0 0.05 0.1 0.15 0.2 0.25
Load(N) Speed(m/s)

Fig. 2. Average surface roughness (Ra) as a function of load and speed tested with hydrophilic (square) and hydrophobic (circle) grinding fluids. The abrasion tests were
performed at the applied load ranging from 50 to 250 N and the sliding speed in the range from 0.05 to 0.2 m/s and sliding time was set to 5400 s.

1 1
Hydrophobic Hydrophobic
Hydrophilic Hydrophilic
0.8 0.8

0.6 0.6
COF

COF

0.4 0.4

0.2 0.2

0 0
0 50 100 150 200 250 300 0 0.05 0.1 0.15 0.2 0.25
Load(N) Speed(m/s)

Fig. 3. The coefficient of friction measured during the abrasion test as a function of load and speed using hydrophilic (square) and hydrophobic (circle) grinding fluids. The
tests were performed at the applied load ranging from 50 to 250 N and the sliding speed in the range from 0.05 to 0.2 m/s and sliding time was set to 5400 s.

250 250
Hydrophobic Hydrophobic
Hydrophilic Hydrophilic
200 200

150 150
Rvk(nm)

Rvk(nm)

100 100

50 50

0 0
0 50 100 150 200 250 300 0 0.05 0.1 0.15 0.2 0.25
Load(N) Speed(m/s)

Fig. 4. Reduced valley depth (Rvk) plotted as a function of applied load and sliding speed when hydrophilic (square) and hydrophobic (circle) grinding fluids were used.
Abrasion tests were performed at the applied load ranging from 50 to 250 N and the sliding speed in the range from 0.05 to 0.2 m/s and sliding time was set to 5400 s.

Table 2. Fig. 7 indicates that the current density of the steel ground while other roughness parameters showed little correlation with
with the hydrophilic grinding fluid is higher than ground with corrosion resistance.
hydrophobic fluid. Table 2 also suggests that the corrosion resis- The stainless steel surfaces after potentiodynamic tests showed
tance is closely related to the distribution of valleys on the surface, different number densities of corrosion pits depending on the type
 S.M. Lee et al. / Corrosion Science 63 (2012) 404–409 407

Fig. 5. Topography of stainless steel surfaces after abrasion tests after ground using (a) hydrophobic and (b) hydrophilic grinding fluids. Two steel plates were observed using
laser confocal microscope after same abrasion condition (150 N, 0.1 m/s, 5400 s).

300
Hydrophobic showed fewer etch pits compared to the one ground using the
250 Hydrophilic hydrophilic fluid. The micrograph also shows that the corrosion
pits were preferentially aligned along the grooves and that the
surface traces(asperities) produced during abrasion tests were
200 removed in the early stage of corrosion during pitting corrosion,
Rvk(nm)

suggesting that the deep valleys on the ground surface were favor-
150 able sites for pit nucleation.
Roughness parameters representing various features of surface
100 topography were compared with the polarization resistance of
the stainless steel after abrasion tests. Fig. 9 shows the polarization
resistance (Rp) plotted as a function of average surface roughness
50
(Ra). The figure indicates that corrosion resistance has a poor corre-
lation with average surface roughness (Ra), while hydrophilic
0 grinding fluid tended to produce the steel surfaces more vulnerable
500 1000 1500 2000 2500 3000
to corrosion. An analogous result was reported by Li et al. [10] in
Ra(nm)
their study of copper surfaces using a scanning Kelvin probe. They
Fig. 6. Average surface roughness (Ra) plotted as a function of reduced valley depth
reported that the corrosion rate increased as a function of the
(Rvk). No correlation between average surface roughness (Ra) and reduced valley surface roughness average (Ra) up to 100 nm but it was not af-
depth (Rvk) was observed. fected by Ra beyond 100 nm. The similar corrosion behavior
regardless of the rough surface at high Ra was attributed to the
similar level of the electron work function of Ra beyond 100 nm.
The poor correlation found in this work between the average
surface roughness (Ra) and the polarization resistance was, there-
fore, appeared to be related to the relatively high degree of surface
roughness in the range of 0.5–2.5 lm, as shown in Fig. 8. Shahryari
et al. [15] reported a similar result, showing a poor correlation
between polarization resistance and surface roughness of 316
stainless steel when Ra was relatively high. The polarization resis-
tance was also examined to obtain a possible correlation with the
surface roughness parameters other than Ra, such as Rq, Rz, Rt, Rv,
skewness, and Rpk. However, poor correlations between polariza-
tion resistance and these roughness parameters were found.
On the other hand, an apparent correlation between corrosion
resistance and surface roughness was found when the polarization
resistance (Rp) was plotted as a function of the reduced valley
depth (Rvk) (Fig. 10). This figure shows that the polarization resis-
tance is increased by more than three times when Rvk is reduced
from 250 nm to 25 nm in the case of using the hydrophilic grinding
fluid. The improvement in corrosion resistance was more pro-
Fig. 7. Potentiodynamic curves obtained from two different steel plates in a 1 M nounced when hydrophobic grinding fluid was used. This is an
NaCl solution at room temperature (25 °C). The two steel plates with different interesting finding because it indicates that the number of valleys
surface roughness were obtained after ground for 5400 s at 100 N and 0.1 m/s using on the surface can play a crucial role in the corrosion resistance of a
the grinding fluids with dissimilar wettability. These results suggest the difference steel surface in the range of Ra with no correlation with corrosion
of surface roughness affected to pitting potential, while material property related to
behavior. On the other hand, polarization resistance showed a poor
repassivation potential.
correlation with the reduced peak height Rpk, indicating that the
surface peaks were removed in the early stage of potentiodynamic
of grinding fluid used. Fig. 8 shows the micrographs obtained using tests and its contribution to polarization resistance was relatively
an optical microscope and a laser confocal microscope. The small. The significant influence of the valleys on corrosion resis-
corroded surface that was ground using the hydrophobic fluid tance appears to be closely related to the depth of the valleys that
408 S.M. Lee et al. / Corrosion Science 63 (2012) 404–409

Table 2
Roughness parameters and polarization resistance obtained from the stainless steel plates after abrasion tests using hydrophilic and hydrophobic grinding fluids. The two
specimens were ground at the same applied load (100 N), speed (0.1 m/s), and time (5400 s).

Grinding Fluid Ra(nm) Rq(nm) Rt(lm) Rpk(nm) Rvk(nm) Rv(lm) Rp(Xcm2)

Hydrophobic 1500 2140 100.0 80.0 268.8 30.1 1924.7
Hydrophilic 1580 2090 48.0 123.6 334.1 31.6 1015.0

Fig. 8. Stainless steel surfaces after potentiodynamic tests. Images were obtained using an optical microscope (top) and a laser confocal microscope (bottom).

6000
6000
Hydrophobic
Hydrophilic Hydrophobic
5000
5000 Hydrophilic

4000
4000
Rp( cm2)

Rp( cm2)

3000
3000

2000
2000

1000
1000

0
500 1000 1500 2000 2500 3000 0
Ra(nm) 0 50 100 150 200 250 300
Rvk(nm)
Fig. 9. Polarization resistance (Rp) as a function of average surface roughness (Ra).
No correlation was found between polarization resistance (Rp) and average surface Fig. 10. Polarization resistance (Rp) as a function of reduced valley depth (Rvk).
roughness (Ra). Polarization resistance (Rp) shows an inverse-proportional relation with reduced
valley depth (Rvk).

affect the diffusion of active ions during corrosion [11] and to IR 2. The coefficient of friction measured during abrasion tests was
drops in the deep valleys [16,17]. higher when the hydrophobic grinding fluid was used, due to
poor lubrication.
3. Corrosion resistance of the stainless steel showed poor correla-
4. Conclusions tions with surface roughness parameters such as Ra, Rt, Rz,
skewness, and Rpk when Ra was greater than 0.5 lm.
Corrosion resistance of 21Cr Ferritic Stainless steel and its rela- 4. The reduced valley depth (Rvk) showed a good correlation with
tion to surface roughness were investigated. By examining the corrosion resistance in the case of rough surfaces with Ra was
surface topography after abrasion tests under various grinding con- greater than 0.5 lm.
ditions, the correlation with the polarization resistance obtained
from potentiodynamic tests provided the following information:
References
1. The surface roughness after abrasion tests was significantly
affected by the wettability of the grinding fluid used, and the [1] P.C. Pistorius, G.T. Burstein, Metastable pitting corrosion of STS, Philos. Trans.
hydrophilic fluid produced greater reduced valley depth (Rvk) R. Soc. London, A. 341 (1992) 531–559.
[2] T. Hong, M. Nagumo, Roughness and pitting corrosion of 301STS, Corros. Sci. 39
than hydrophobic fluid. (1997) 1665–1672.
 S.M. Lee et al. / Corrosion Science 63 (2012) 404–409 409

[3] K. Sasaki, G.T. Burstein, The generation of surface roughness during slurry roughness on a structural AE44 magnesium alloy, Corros. Sci. 52 (2010) 1635–
erosion-corrosion and its effect on the pitting potential, Corros. Sci. 38 (1996) 1648.
2111–2120. [10] W. Li, D.Y. Li, Influence of surface morphology on corrosion and electronic
[4] P. Peyre, C. Braham, J. Lédion, L. Berthe, R. Fabbro, Corrosion reactivity of laser- behavior, Acta Mater. 54 (2006) 445–452.
peened steel surfaces, JMEPEG 9 (2000) 656–662. [11] S.M. Sharland, A review of the theoretical modeling of crevice and pitting
[5] G.T. Burstein, S.P. Vines, Repetitive nucleation of corrosion pits on stainless corrosion, Corros. Sci. 27 (1987) 289–323.
steel and the effects of surface roughness, J. Electrochem. Soc. 148 (12) (2001) [12] G.W. Stachowiak, A.W. Batchelor, Engineering tribology, 3rd ed., Elsevier
B504–B516. Butterworth-Heinemann, 2005.
[6] L.R. Hillert, D. Bagge-Ravn, J. Kold, L. Gram, Influence of surface roughness of [13] T.R. Thomas (ed.), Rough Surfaces, Longman Group Ltd, 1982.
stainless steel on microbial adhesion and corrosion resistance, Int. Biodeterior. [14] D.A. Jones, Principles and Prevention of Corrosion, Prentice-Hall, Inc., 1995, pp.
Biodegradation 52 (2003) 175–185. 148.
[7] M.H. Moayed, N.J. Laycock, R.C. Newman, Dependence of the critical pitting [15] A. Shahryari, S. Omanovic, J.A. Szpunar, Electrochemical formation of highly
temperature on surface roughness, Corros. Sci. 45 (2003) 1203–1216. pitting resistant passive films on a biomedical grade 316LVM stainless steel
[8] H. Lee, D. Kim, J. Jung, Y. Pyoun, K. Shin, Influence of peening on the surface, Mater. Sci. Eng., C 28 (2008) 98–106.
corrosion properties of AISI 304 stainless steel, Corros. Sci. 51 (2009) 2826– [16] N.J. Laycock, J. Stewart, R.C. Newman, The initiation of crevice corrosion on
2830. stainless steels, Corros. Sci. 39 (1997) 1791–1809.
[9] R.B. Alvarez, H.J. Martin, M.F. Horstemeyer, M.Q. Chandler, N. Williams, P.T. [17] G.F. Kennell, R.W. Evitts, K.L. Heppner, A critical crevice solution and IR drop
Wang, A. Ruiz, Corrosion relationships as a function of time and surface crevice corrosion model, Corros. Sci. 50 (2008) 1716–1725.