Electrochemically controlled pitting corrosion in Ni lm: A study of AFM

and neutron reectometry

Surendra Singh
a,
*
, Saibal Basu
a
, A.K. Poswal
b
, R.B. Tokas
b
, S.K. Ghosh
c
a
Solid State Physics Division, Bhabha Atomic Research Center, Mumbai, Maharastra 400085, India
b
Spectroscopy Division, Bhabha Atomic Research Center, Mumbai, Maharastra 400085, India
c
Material Processing Division, Bhabha Atomic Research Center, Mumbai, Maharastra 400085, India
a r t i c l e i n f o
Article history:
Received 15 August 2008
Accepted 17 December 2008
Available online 25 December 2008
Keywords:
B. Neutron reectometry
B. AFM
C. Pitting corrosion
C. Ni lm
a b s t r a c t
Electrochemical behavior of pitting corrosion of a Ni lm, grown on Si substrate by sputtering, prepass-
ivated in a chloride-free sulfuric acid solution and subsequently exposed to chloride above the pitting
potential is reported. Specular and off-specular unpolarized neutron reectometry and Atomic Force
Microscopy (AFM) techniques have been used to determine the depth prole of scattering length density
and morphology of as-deposited as well as corroded sample. Specular neutron reectometry measure-
ment of the lm after corrosion shows density degradation along the thickness of lm. The density prole
as a function of depth, maps the growth of pitting and void networks due to corrosion. The AFM and off-
specular neutron reectivity measurements has suggested that the morphology of the lm remains same
after exposure of the lm in chloride solution.
2008 Elsevier Ltd. All rights reserved.
1. Introduction
Metals such as Ni, Fe, Cr and Ni-based stainless alloys, are pro-
tected against corrosion by a passive lm [1,2]. Usually, passive
lms consist of an inner oxide layer and an outer layer of hydroxyls
or hydroxide [15]. Localized corrosion (pitting) of metals nor-
mally occurs in the presence of aggressive anionic species and
Cl

is the cause in most cases. Generally the pitting corrosion pro-

cess of metals in chloride-containing solutions involves: formation
of the passive lm, breakdown of the passive lm, growth of meta-
stable pits which can repassivate, and growth of stable pits. De-
spite a considerable amount of data obtained [15] the
fundamental aspects of the mechanisms operating in the early
stages of pitting (leading to passivity breakdown and growth of
pits which may or may not repassivate) must still be investigated
in detail. In particular, there is no atomic-scale data on the struc-
ture and morphology of growth of pits in these early stages of
growth. Chemical or physical heterogeneities at the surface (such
as inclusions, second particles, solute-segregated grain boundaries,
aws, dislocations, or mechanical damage) are, however, favored
locations for pit initiation. Moreover the study of pitting in thin
lms is of interest because of differences in behavior compared
to pitting in bulk alloys [68]. Also recently passivation and pitting
of thin lms [9,10] has received extensive attention because thin
lms are widely used in electronic and magnetic recording indus-
tries as well as other industries. A pit will quickly penetrate a thin
lm and, if the substrate is inert, become two dimensional [7],
which may behave differently from three-dimensional pits in bulk
materials. The pitting of thin lms also provides a convenient
opportunity for studying the phenomenon of pitting [11].
Ni, which can be considered as a prototypical example of a pas-
sive metal and has been subject to numerous studies. Corrosion of
Ni in aqueous electrolytes has been the subject of fundamental
understanding of the mechanism of metal corrosion as well as high
technological importance of Ni and its alloys, such as stainless
steels and brasses. The tendency for a metal to pit strongly depends
on the microstructure and composition [5]. The passivation and
pitting of Ni in Cl

environment has been extensively studied both

theoretically and experimentally [1116]. Despite of extensive
studies [1116] in the growth and propagation of pits and the com-
plex atomic-scale process underlying metal corrosion in electro-
chemical environment, the depth prole of pit growth along the
thickness of the lm and growth of interface morphology, which
might play important role in controlling further growth of the pits,
are considerably less understood and studied, mainly due to the
lack of methods suitable for the direct structural analysis of the
lms. In recent time it becomes possible to study such properties
by modern in situ techniques such as scanning probe microscopy
(STM, AFM), X-ray spectroscopy and surface X-ray diffraction.
These approaches, even though helpful in qualitatively explaining
the interface, lack direct depth proling of chemical and morpho-
logical information non-destructively. Also these techniques may
not exactly reveal formation of double layer interfaces on passiv-
ation. Neutron reectometry (NR) [17,18] in specular mode is a
non-destructive technique which provides a depth resolution as
0010-938X/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.corsci.2008.12.017
* Corresponding author. Tel.: +91 22 25594604; fax: +91 22 25505151.
E-mail address: surendra@barc.gov.in (S. Singh).
Corrosion Science 51 (2009) 575580
Contents lists available at ScienceDirect
Corrosion Science
j our nal homepage: www. el sevi er . com/ l ocat e/ cor sci
low as 0.1 nm and is used to investigate the depth prole of com-
position. NR determines the structural parameters, such as individ-
ual layer thickness, interface roughness, and density of each layer
averaged over the lateral dimension of the sample. Diffuse (off-
specular) neutron reectivity (DNR) [19] yields heightheight cor-
relation function, which determine detailed morphology (in-plane
structure) of the surfaces and buried interfaces.
In a previous study, using specular and off-specular neutron
reectivity we compared the change in morphology of the surface
caused by corrosion with respect to a buried interface in a Ni lm
grown on glass substrate using thermal evaporation technique
[20]. In this work, we report the depth prole and in plane mor-
phology of pitting corrosion in chloride solution of a passivated
Ni lm grown by sputtering on a Si substrate using neutron reec-
tivity and Atomic Force Microscopy (AFM) techniques. Detailed
scattering length density depth prole of an as-deposited Ni lm
and of the same lm after controlled electrochemical corrosion
was measured by unpolarized neutron reectometry. The attempt
has been to map the propagation of pits in the lm due to corro-
sion. The neutron reectometry measurements show density deg-
radation along the thickness of the lm, which maps the growth of
pitting and promulgation of void networks along the depth of the
lm. Interestingly, the density prole of the corroded lm shows
that the degradation in density is not a monotonic function of
depth in the lm as one tends to feel intuitively. AFM and DNR
study revealed that the morphology of the lm remains same on
pitting corrosion and shows a two dimensional Gaussian
interfaces.
2. Experimental details
The Ni samples used in the present study had been grown by r.f.
sputtering technique at a base pressure of 5 10
8
Torr and at
deposition pressure of 4 10
3
Torr. A Si (100) substrate was rst
cleaned for both organic and inorganic contaminations. The lm of
Ni with layer thickness of about 800 was grown on the Si sub-
strate. The sample thickness was monitored by quartz crystal sys-
tem during growth. The electrochemical passivation and pitting
corrosion on sputtered Ni surface was carried by using AUTOLAB-
30 Potentiostat/Galvanostat. The electrolytes were 0.05 M
H
2
SO
4
+ 0.095 M NaOH without and with 0.05 M NaCl, prepared
fromhigh purity chemicals and Millipore water. The pH of the elec-
trolyte was maintained at 3, making it strongly acidic. The proce-
dure of the corrosion tests was as follows. The passive lm was
rst produced in chlorine-free solution. After about 30 min of pas-
sivation chloride anions (NaCl) were added to the electrolyte at a
concentration of 0.05 M and current transient were recorded. Both
pitting and passivation experiments were performed under con-
stant potential conditions and the transient current behaviour
was monitored with time. All the potentials mentioned in the text
are referred to saturated calomel electrode (SCE). Both specular
and off-specular neutron reectometry experiments at room tem-
perature were performed on the polarized neutron reectometer at
Dhruva, BARC, India [21]. This instrument has been designed for
vertical sample geometry with a horizontal
2
He
3
-based linear posi-
tion sensitive detector (PSD), located normal to the incident beam
to capture the reected intensity. The specular reectivity was
measured as a function of wave-vector transfer Q (=4psin(h)/k,
where h and k are angle of incidence and wavelength of neutron,
respectively). The momentum transfer Q dependence of reectivity
prole is related to the Fourier components of the scattering length
density (SLD) depth prole, providing depth sensitivity. Speci-
cally, the neutron reectivity data can be tted [22,23] using
depth-dependent SLD prole q(z) (where z is the lm depth),
qz
P
i
N
i
zb
i
, where the summation is over each type of atom
in the system, N is the number density at a depth z and b is the
coherent nuclear scattering length. Quantitative information was
extracted by modeling to nd a q(z) model that reproduces the
data with the lowest possible value of v
2
using Genetic Algorithm
(GA) [24]. The scattering length density and thickness of each layer
are taken as variables for getting best t. The horizontal PSD col-
lected the DNR data along its length (x direction), integrated over
vertical direction, centered on a specular peak. This is equivalent
to detector scan in a conventional h2h spectrometer [25]. The sur-
face morphology of the sample was investigated using a NT-NDTs
Solver P-47 H multimode AFM instrument. The images were taken
in a non-contact mode with a Si
3
N
4
tip. The cantilever has a spring
constant of 0.6 N/m. AFM topographic images were recorded over
scan areas from 1 1 lm to 10 10 lm, each with a resolution
of 256 256 data points.
3. Results and discussion
The electrochemical behavior of Ni depends strongly on the
structure and morphology of the electrode surface as well as on
the preparation procedure used. We measured the potentiodynam-
ic polarization curve of Ni lm in 0.05 M H
2
SO
4
+ 0.095 M NaOH
solution with different NaCl concentrations (0.02, 0.05 and
0.07 M). From these measurements we observed that the pitting
potential of Ni shifts towards negative direction with the increase
of [Cl

] in the solution. The potentiodynamic polarization curve for

Ni lm in 0.05 M H
2
SO
4
+ 0.095 M NaOH solution with 0.05 M NaCl
is shown in Fig. 1. It can be seen clearly that Ni lm shows passiv-
ation over a wide potential range. The breakdown potential, or pit-
ting potential (E
pit
), of the Ni lm was 0.47 V
SCE
(volt). In order to
understand the pitting initialization in Cl

ion we have studied

the change of anodic current with time for Cl

free and Cl

contain-
ing solution. Fig. 2 shows a typical current transient recorded at 0.5
V
SCE
before the introduction of chloride. The continuous decrease of
the current density was observed in chlorine-free solution. The
sharp decrease in current density (Fig. 2) at the beginning followed
by gradual decrease with time indicates the formation of non-con-
ducting layer on conducting Ni surface. This behavior represents
formation of nickel oxide layer (passivation), which was conrmed
later by neutron reectivity measurements. The Ni lm exhibited
the fast kinetics of passivation and reached the steady-state cur-
10
-6
10
-5
10
-4
10
-3
-0.6
-0.4
-0.2
0.0
0.2
0.4
P
o
t
e
n
t
i
a
l
,

E
S
C
E

(
V
o
l
t
)
Current Density (A/cm
2
)
Fig. 1. Polarization curves for Ni lm in 0.095 M NaOH + 0.05 M H
2
SO
4
solution at
pH 3.0 + 0.05 M NaCl.
576 S. Singh et al. / Corrosion Science 51 (2009) 575580
rent density within 10 s. The composition of these passivating
oxide layers depends considerably on pH, applied potential and
on electrolysis time [26]. Fig. 3 shows a typical current transient
recorded at 0.5 V
SCE
after the introduction of chloride. The curve
shows an increase of current, which is a characteristic for the for-
mation of stable pits. After the introduction of chloride, the current
density shows steep increases from 0.43 to 0.63 mA cm
2
within
72 s of introduction of chloride. Thereafter current was found to
be constant for a longer duration with uctuation of smaller ampli-
tude. This increase of current density is characteristic of the effect
of chloride since a continuous decrease of current density was ob-
served in chloride-free solutions. This effect corresponds to an in-
creased dissolution in the passive state previously assigned to a
general corrosion process promoted in the presence of chloride
[2]. The small uctuation in current density in electrochemical
behavior of Ni lm in presence of chloride solution might have
the outcome of fast localized dissolution and with repassivation,
which is a signature of initialization of pitting growth, also termed
as metastable pitting. The breakdown of the passive lm and initi-
ation of the pitting process is probably the least understood aspect
of the pitting phenomenon. Breakdown is a rare occurrence that
happens extremely rapidly on a very small scale, making direct
observation extraordinarily difcult. In spite of that, the knowledge
of the kinetics and mechanism of Nickel corrosion, passivation and
passivity breakdown in chloride containing aqueous alkaline solu-
tions still requires further investigation to determine the reaction
mechanisms taking place at each stage.
Fig. 4A shows a 5 5 lm AFM image of as-deposited Ni lm on
Si substrate. Film shows that grains are uniformly distributed over
the entire substrate, representing the growth of a smooth surface.
Fig. 4B shows a 5 5 lm AFM image of the passivated Ni surface
exposed to chloride. The AFM image clearly demonstrates the local
areas where the passivated Ni surface has been attacked by chlo-
ride atoms resulting into a pit formation. No pits were observed
in the absence of chloride as shown in Fig. 4A. Pits with variable
sizes from few nm to 140 nm and shapes are distributed over the
entire lm, which makes a porous surface with reduction in aver-
age density of metal. So the AFM image of Ni surface under corro-
sive pitting conditions can reveal the initiation of localized
corrosion at surfaces on chloride solution, which might have prop-
agated by strong inuence of structural defects on pitting onset
sites (i.e. boundary separating different grains of the passive lm)
and morphology of the interfaces. The dimensions of observed pits
are small with few nm, which might have resulted due to simulta-
neous dissolution and repassivation of the metal surface.
Since the AFM does not provide any information about the local
chemistry and propagation of pits along the thickness of the lm as
well as along in the planes of the lm, specular neutron reectom-
0 50 100 150 200
0.300
0.325
0.350
0.375
0.400
0.425
C
u
r
r
e
n
t

d
e
n
s
i
t
y

(
m
A
.
c
m
-
2
)
Time(s)
Fig. 2. Current transient of Ni lm recorded at 0.5 V
SCE
in 0.095 M NaOH + 0.05 M
H
2
SO
4
solution at pH 3.0.
50 100 150 200
0.45
0.50
0.55
0.60
0.65
C
u
r
r
e
n
t

d
e
n
s
i
t
y

(
m
A
.
c
m
-
2
)
Time(s)
Fig. 3. Current transient of Ni lm recorded at 0.5 V
SCE
in 0.095 M NaOH + 0.05 M
H
2
SO
4
solution at pH 3.0 after the addition of 0.05 M NaCl after 30 min of
passivation.
Fig. 4. AFM images of the Niair interface as-deposited Ni sample grown by
sputtering technique (A) and the sample corroded electrochemically in chlorine
solution (B) of sizes 5.0 5.0 lm recorded in non-contact mode.
S. Singh et al. / Corrosion Science 51 (2009) 575580 577
etry measurements has been performed to study the detail layered
structure and depth prole of density of the as-deposited as well as
lm after corrosion. The scattering length density (chemical den-
sity) prole obtained from NR measurements is averaged over the
planer surface of the lm. Fig. 4 shows the unpolarized neutron
reectometry data from as-deposited sample (before corrosion)
as well as from sample after corrosion. Closed circles and open cir-
cles are the experimental data from as-deposited sample and sam-
ple after corrosion, respectively. The continuous lines are ts to
measured data. The nuclear SLD (q(z)) model used to t the data
are shown in the inset of Fig. 5. The dash line and continuous line
in the inset of Fig. 4 represents the SLD prole from as-deposited
sample and sample after corrosion, respectively. The average thick-
ness and scattering length density of as-deposited Ni lm on sub-
strate, obtained from the unpolarized NR measurements, was
790 10 and 8.45 0.20 10
6

2
, respectively. The SLD of Ni
in this layer is about 90% of its bulk value (9.4 10
-6

2
). Thick-
ness of the Ni lm symbolizes the oscillations (Keissing oscillation)
in the reectivity pattern for as-deposited Ni sample. Using NR we
also obtained the average roughness of Niair interface for as-
deposited Ni lm, which is about 18 2 .
The reectivity pattern for the lm exposed to chloride (after
corrosion) is completely different from that obtained for as-depos-
ited sample (i.e. before corrosion). In this case the reectivity pat-
tern shows a lower critical angle (h
c
k

qz=p
p
) for reection as
well as the modications in the oscillation of reectivity pattern
(Fig. 4). Since critical angle depends on the density of the media,
overall loss in density of the lm due to corrosion is apparent from
the lower critical angle of the lm after corrosion. This is due to the
modication of the interface structure resulted on corrosion. The
SLD model that produces the best t for reectivity data from
the lm after corrosion is shown in the inset of Fig. 5 (continuous
line), which shows an interesting peaked nature as a function of
depth. The best t q(z) model, for the lm after corrosion, consists
of seven layers with different thicknesses and SLDs: top 200 has a
SLD of 3.35 10
6

2
and six others layers each of thickness
about 100 and different SLDs: 6.21 10
6
, 8.01 10
6
,
7.20 10
6
, 6.04 10
6
, 5.56 10
6
and 3.74 10
6

2
, respec-
tively, along the depth of the lm. The top 200 thickness has a
SLD of 3.3510
-6

2
in the lm after corrosion. The SLD of NiO
is approximately 2.6010
-6

2
. The lower density on the surface
could be due to formation of non-conducting NiO as well as due
to pit formation. NiO layer might have resulted during the passiv-
ation of the lm before addition of chloride solution, as depicted in
electrochemical results (Fig. 2). If one considers that the propaga-
tion of the corrosion is solely due to the pit formation by the Cl

ion in the material, which is possible in case of bulk samples, then

intuitively one feels that the corrosion front due to Cl

ion migra-
tion should result in a SLD prole that monotonically increases to
SLD of Ni as one goes from surface to deeper in the lm. Instead the
present SLD prole shows that the SLD is lowest on the surface of
the lm and as one goes in the lm, SLD rises to a near-bulk value
(SLD for Ni in as-deposited lm) and then again goes down to a
lower value in the interior of the lm (inset of Fig. 5). This can
be explained by theory of pit growth in case of thin lms [7] where
pits grown in thin lm samples quickly penetrate through the lm
thickness and become two dimensional steady-state analogs to
small pits in bulk samples. Pits in thin lms imitate, to a certain ex-
tent, small pits (nm length scale) in three-dimensional samples
because of the small pit depth, which is limited to the metal lm
thickness once the pit perforates the metal lm. The small pit then
propagates along the plane of the lm at different depth, which can
result into formation of sublayers with different SLD as observed in
SLD prole for the lm after corrosion. Another reason for observed
SLD prole from neutron reectivity measurements of lm after
corrosion can be explained by voiding and collapse mechanism of
passivation breakdown on pitting [15]. The transfer of metal cat-
ions in electrolytes and their dissolution will create voids of differ-
ent sizes. The growth of bigger size voids will lead to its collapse;
where as small voids can propagate through grains boundaries or
pre-existing void networks along the depth as well as in plane of
the lm. The small void propagation in the lm can make it porous,
which will results in reduction in average density of the lm. The
increase in density and the subsequent decrease as one goes dee-
per in the lm can be attributed to the void pathways for chloride
ions in the lm. NR result predicts that the pitting corrosion in Ni
lm has grown two dimensionally along the surface of the lm and
moreover the propagation of pits and void networks along the
depth does not show any regular variation.
Another important aspect of understanding the mechanism,
dynamics and morphology of pitting growth on corrosion still re-
main subject of discussion. Though the importance of morphology
on corrosion process has been certainly recognized for a long time
but the main emphasis has been certainly devoted to the electro-
chemical aspects of the phenomena. Despite the advances that
have been made [1116], fundamental understanding of the
dynamics and growth of pitting corrosion is difcult because
events take place on a very small scale, with passive lm nanome-
ters in thickness and initiation sites of a similar size. Immediately
after initiation, the rate of pit growth can be extremely high, even
tens of A/cm
2
[6,7]. Therefore, the situation is extremely dynamic
with rapidly moving boundaries and rapidly changing chemistries.
As a result, it is not possible to predict exactly when and where
breakdown will occur, which makes high-resolution observation
of initiation events extremely difcult. To understand the mecha-
nism of growth of pitting in chloride solution we studied the cor-
rosion effect in identical condition on another sample (Ni lm)
with layer thickness of about 600 grown on the Si substrate in
identical condition. The sample was immersed in chloride solution
for 125 s, which is half the period of immersion time for previous
sample in identical corrosion environment. Fig. 6 shows the unpo-
0.016 0.032 0.048 0.064
10
-4
10
-3
10
-2
10
-1
10
0
Critical angle
S
L
D
(

-
2
)
Depth()
unpolarized neutron reflectivity profile
as-deposited Ni film
Ni film after corrosion
R
e
f
l
e
c
t
i
v
i
t
y
Q(
-1
)
0 200 400 600 800
2.0x10
-6
4.0x10
-6
6.0x10
-6
8.0x10
-6
as-deposited Ni film
Ni film after corrosion
Fig. 5. Specular unpolarized neutron reectometry measurements on sample
before (lled circles) and after (open circles) corrosion. Continuous lines are t to
measurements. Inset shows the scattering length density (SLD) prole of the sample
extracted from t to unpolarized neutron reectivity measurements. DashDash
line is SLD prole of as-deposited (before corrosion) sample whereas continuous
line is SLD prole of sample after corrosion.
578 S. Singh et al. / Corrosion Science 51 (2009) 575580
larized neutron reectometry data from as-deposited sample (be-
fore corrosion) as well as on sample after corrosion. Closed circles
and open circles are the experimental data from as-deposited sam-
ple and sample after corrosion, respectively. The continuous lines
are ts to measured data. The nuclear SLD (q(z)) model used to
t the data are shown in the inset of Fig. 6. The dash line and con-
tinuous line in the inset of Fig. 6 represents the SLD prole from as-
deposited sample and sample after corrosion, respectively. It is
clear from the SLD proles from two samples that the depth
dependent of scattering length density is similar on corrosion for
different time in chloride solution. This shows the growth of pitting
occurs rapidly on immersing the sample in chloride solution and
stable pit formation occurred after some time. This was also evi-
dent from current transient plot in Fig. 3. This predicts that the
dynamics of two consecutive stages of pit initiation and pit growth
is very fast in chloride solution. The fast growth of pits in preset
system might have occurred through grain boundary of passivated
layer. The resulting dynamic process of dissolution, operating layer
by layer and promoted by chloride, is thought to be responsible for
the growth of pitting corrosion of the sample after passivation,
which shows distinct density degradation along the thickness of
lm.
There has been substantial interest in the morphology of pitting
corrosion. Various types of corrosion models for pitting corrosion
of metal surfaces have been studied using theoretical models and
computer simulations [27,28]. A variety of morphologies has been
observed in the open-circuit pitting of evaporated Al thin layers in
NaCl and Fe
2
(SO
4
)
3
containing acidic medium [29]. It was shown
that the shape of the corrosion patterns depends on the layer thick-
ness, the pH, the concentration of aggressive anions (Cl

) as well as
the concentration of the given oxidizing agent (Fe
3+
). A smooth to
fractal transition of the corrosion front has been observed in case of
Ni lm [20] using AFM, specular and DNR techniques. Morpholog-
ical changes due to pitting at micrometer length scales on surfaces
of aluminum and copper lms have been studied by diffuse light
scattering [30] and Diffuse X-ray scattering [31], respectively. A
large variety of rough surfaces over a wide range of length scales
demonstrate self-afne fractal morphology [32]. Here we have
used DNR and AFM techniques to study the modication in mor-
phology of the pitting corrosion of Ni lm in Cl

environment.
In general, the roughness parameter r
2
is dened as
< h

h
2
> the variance of the height (h) distribution on the sur-
face of a thin lm. The angular brackets specify an ensemble aver-
age. This average roughness, r can be obtained from specular
neutron reectivity (NR) data, which comprises the effect of alloy-
ing, interdiffusion, intermixing [33]. The in-plane correlation for
two points separated by a distance r on a self-afne fractal surface
is quantied by the Height Difference Correlation Function [HDCF]
g(r), expressed as [19]:
gr hhr h0
2
i 2r
2
2Cr
2r
2
1 exp
r
n

2H
!
1
where n is a length, over whichthe heights remaincorrelated. His the
Hurst parameter, which describes the fractal dimension D (=3-H) of
the surface. r is true roughness and for distance r > >n, g(r) reaches
a saturation value of 2r
2
, when the heights become uncorrelated.
C(r) is called height-height correlation function. We have used a for-
malismdeveloped by Holy et al. [34] which calculates the off-specu-
lar scattering cross-section for a stratied thin lm sample (see Eq.
(16) in Ref. [34]), from the Fourier transform of layer specic
Height-Height correlation function, C(r) dened in Eq. (1).
Fig. 7i show background corrected DNR data on a semi-log scale
centered on the specular peaks at 0.028
1
. Closed circles in g-
ures (A) and (B) of Fig. 7i are experimental DNR data from as-
deposited and lm after corrosion, respectively. The continuous
lines are best t to measured data using Holy formalism [34].
The roughness parameters, r, n and H for Niair interface of as-
deposited Ni lm as obtained from best t to DNR data are 20 ,
800 and 0.95, respectively. The Hurst parameter H = 0.95 corre-
sponds a fractal dimension of 2.05, which represents a Gaussian
two dimensional surface. DNR measurements on corroded lm
didnt illustrate any drastic change in the morphology as compared
to that of as-deposited lm. The roughness parameters, r, n and H
for corroded lm obtained from the analysis of DNR data are 22 ,
1200 and 0.90, respectively, which again shows a Gaussian two
dimensional surface. We found higher value of correlation length
0.02 0.03 0.04 0.05 0.06
10
-4
10
-3
10
-2
10
-1
10
0
unpolarized neutron reflectivity profile
as-deposited Ni film
Ni film after corrosion
S
L
D
(

-
2
)
Depth()
R
e
f
l
e
c
t
i
v
i
t
y
Q(
-1
)
0 200 400 600
2.0x10
-6
4.0x10
-6
6.0x10
-6
8.0x10
-6
as-deposited Ni film
Ni film after corrosion
Fig. 6. Specular unpolarized neutron reectometry measurements another Ni lm
with a reduced thickness (see text), before (lled circles) and after (open circles)
corrosion. Inset shows the scattering length density (SLD) prole of the sample
extracted from t to unpolarized neutron reectivity measurements. Dash-dot-dot
line is SLD prole of as-deposited (before corrosion) sample whereas continuous
line is SLD prole of sample after corrosion.
0.0
6.0x10
-4
1.2x10
-3
10
-4
10
-3
10
-2
(B) After Corrosion
0.0
5.0x10
-4
1.0x10
-3
10
-4
10
-3
10
-2
10
-1
Q
x
(
-1
)
I
n
t
e
n
s
i
t
y

(
a
r
b
.

u
n
i
t
)
(A) Before Corrosion
Q
x
(
-1
)
100 1000
10
0
10
1
10
2
(A) Before Corrosion
10
0
10
1
10
2
(B) After Corrosion
g
(
r
)
r()
(ii) AFM Measurements (i) DNR Meaurements
Fig. 7. (i) Off-specular neutron reectivity measurements from as-deposited (A)
and corroded (B) Ni sample. Solid circles and continuous line are the measured off-
specular neutron reectivity data and the t to data, respectively. (ii) The
correlation function (see text) of the AFM data of scans area 2.0 2.0 l for as-
deposited (A) and corroded (B) Ni sample.
S. Singh et al. / Corrosion Science 51 (2009) 575580 579
(1200 ) for corroded lm as compared to as-deposited lm
(800 ), which is typical size of the width of the pit. DNR mea-
surements show that in a length scale of about pit diameter the
morphology of the lm remain two dimensional and the surface
morphology remains identical in pitted areas as in nonpitted areas
where the chemical dissolution of the passive lm is promoted by
chloride.
In the present analysis for the morphology at airlm interface,
we have used AFM data from the sample over scan areas ranging
from 1.0 1.0 l to 10.0 10.0 l with space resolution varying
from 39 to 390 , respectively. The AFM data were used to obtain
the HDCF (g(r)) for the as-deposited and corroded surface as shown
in gures (A) and (B) of Fig. 7ii, respectively, with open circles.
Using Eq. (1) for g(r), we obtained the roughness parameters, r, n
and H from the AFM data for as-deposited as well as for corroded
lm. The t gives RMS roughness (r) of 14 , a correlation length
(n) of 650 and the Hurst parameter (H) of 0.95 for as-deposited
lm. Similarly the roughness parameters, r, n and H for corroded
lm using AFM data are 15 , 950 and 0.92, respectively, which
are close to that measured from DNR measurements. Like DNR
measurements, AFM data also suggest similar kind of morphology
for the as-deposited and corroded lms, which are Gaussian and
two dimensional.
4. Summary and conclusions
In summary, we have presented the morphology and depth
prole of pitting corrosion, produced by exposing a prepassivat-
ed Ni lm to chloride above the pitting potential, using AFM,
specular and off-specular neutron reectometry techniques.
AFM image of corroded lm shows pits with variable sizes from
few nm to 140 nm and shapes are distributed over the entire
lm. This makes the surface porous with reduction in average
density of metal. The electrochemical analysis also show that
the rise of residual current density caused by exposure to chlo-
ride results from a general corrosion process promoted in the
presence of chloride. The potentiodynamic polarization curves
of Ni lm clearly show passivation over a wide potential range.
The breakdown potential, or pitting potential (E
pit
), of the Ni lm
was 0.47 V
SCE
(volt). Specular neutron reectivity measurements
showed formation of oxide layer during passivation of sample in
the electrolytic solution of 0.05 M H
2
SO
4
+ 0.095 M NaOH. Also
the scattering length density prole of the corroded lm, ob-
tained from the analysis of specular neutron reectivity mea-
surement, suggest growth of pitting and promulgation of void
networks along the depth of the lm, which changes the scatter-
ing length density along the depth of the lm. The results also
predict fast dynamics of pit initiation and pit growth in chloride
solution. The experimental results from AFM and off-specular
neutron reectivity show that the morphology associated with
the pitting corrosion of Ni in chloride-containing solutions is a
two dimensional Gaussian surfaces, similar to that of as-depos-
ited Ni lm.
References
[1] G.S. Frankel, J. Electrochem. Soc. 145 (1998) 2186.
[2] P. Marcus (Ed.), Corrosion Mechanisms in Theory and Practice, second ed.,
Marcel Dekker Inc., New York, 2002.
[3] D.E. Williams, R.C. Newman, Q. Song, R.G. Kelly, Nature (London) 350 (1991)
216.
[4] M.R. Ryan, D.E. Williams, R.J. Chater, B.M. Hutton, D.S. Mcphail, Nature
(London) 415 (2002) 770.
[5] Z. Szklarska-Smialowska, Pitting Corrosion of Metals, NACE, 1986.
[6] G.S. Frankel, Corros. Sci. 30 (1990) 1203.
[7] G.S. Frankel, J.O. Dukovic, B.M. Rush, V. Brusic, C.V. Jahnes, J. Electrochem. Soc.
139 (1992) 2196.
[8] G.S. Frankel, J.R. Scully, C.V. Jahnes, J. Electrochem. Soc. 143 (1996) 1834.
[9] T. Shahrabi, S. Sanjabi, E. Saebnoori, Z.H. Barber, Mater. Lett. 62 (2008) 2791.
[10] B. OBrien, W.M. Carroll, M.J. Kelly, Biomaterials 23 (2002) 1739.
[11] L. Balaz s, J.-F. Gouyet, Mater. Res. Soc. Symp. Proc. 407 (1996) 281. and
references therein.
[12] J. Scherer, B.M. Ocko, O.M. Magnussen, Electrochim. Acta 48 (2003) 1169.
[13] A. Seyeux, V. Maurice, L.H. Klein, P. Marcus, J. Electrochem. Soc. 153 (2006)
B453.
[14] SeJin Ahn, HyukSang Kwon, Digby D. Macdonald, J. Electrochem. Soc. 152
(2005) B482.
[15] P. Marcus, V. Maurice, H.-H. Strehblow, Corros. Sci. 50 (2008) 2698.
[16] N. Pineau, C. Minot, V. Maurice, P. Marcus, Electrochem. Solid-State Lett. 6
(2003) B47.
[17] J. Lekner, in: Theory of Reection of Electromagnetic and Particle Waves,
Martinus Nijhoff, Dordrecht, Netherlands 1987.
[18] Hartmut Zabel, Festkrperprobleme 30 (1990) 197.
[19] S.K. Sinha, E.B. Sirota, S. Garoff, H.B. Stanley, Phys. Rev. B 38 (1988) 2297.
[20] Surendra Singh, Saibal Basu, Surf. Sci. 600 (2006) 493.
[21] S. Basu, S. Singh, J. Neutron Res. 44 (2006) 109.
[22] C.F. Majkrzak, Physica B 173 (1991) 75.
[23] J.F. Ankner, C.F. Majkrzak, H. Homma, J. Appl. Phys. 73 (1993) 6436.
[24] Surendra Singh, Saibal Basu, Solid State Phys. (India) 44 (2001) 257.
[25] J.P. Schlomka, M. Tolan, L. Schwalowsky, O.H. Seeck, J. Stettner, W. Press, Phys.
Rev. B 51 (1995) 2311.
[26] J. Gomez Becerra, R.C. Salvarezza, A.J. Arvia, Electrochim. Acta 33 (1988) 613.
[27] T. Nagatani, Phys. Rev. Lett. 68 (1992) 1616.
[28] P. Meakin, T. Jossang, J. Feder, Phys. Rev. E 48 (1993) 2906.
[29] L. Balzs, L. Nyikos, I. Szab, R. Schiller, Fractals 1 (1993) 416.
[30] Y.P. Zhao, C.F. Cheng, G.C. Wang, T.M. Lu, Appl. Phys. Lett. 73 (1998) 2432.
[31] S.K. Sinha, Y.P. Feng, C.A. Melendres, D.D. Lee, T.P. Russel, S.K. Satija, E.B. Sirota,
M.K. Sanyal, Physica A 231 (1996) 99.
[32] Heinz-Otto Peitgen, Dietmar Saupe (Eds.), The Science of Fractal Images,
Springer-Verlag, New York, 1988.
[33] M. Wormington, I. Pape, T.P.A. Hase, B.K. Tanner, D.K. Bowen, Philos. Mag. Lett.
74 (1996) 211.
[34] V. Holy , J. Kubna, I. Ohldal, K. Lischka, W. Plotz, Phys. Rev. B 47 (1993) 15896.
580 S. Singh et al. / Corrosion Science 51 (2009) 575580