Electrochimica Acta 62 (2012) 4250

Contents lists available at SciVerse ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

Failure behavior of nano-SiO2 llers epoxy coating under hydrostatic pressure

Li Liu a , Yu Cui a,b , Ying Li a, , Tao Zhang b , Fuhui Wang a,b
a

State Key Laboratory for Corrosion and Protection, Institute of Metal Research, Chinese Academy of Sciences, Wencui Rd 62, Shenyang, 110016, China
Corrosion and Protection Laboratory, Key Laboratory of Superlight Materials and Surface Technology, Harbin Engineering University, Ministry of Education, Nantong ST 145, Harbin
150001, China
b

a r t i c l e

i n f o

Article history:
Received 8 July 2011
Received in revised form 28 October 2011
Accepted 15 November 2011
Available online 25 November 2011
Keywords:
High hydrostatic pressure
Deep ocean
Organic coating
Fractal dimension

a b s t r a c t
The failure of organic coating (epoxy resin lled with 5 mass% nano-SiO2 particles) on mild steel under
high hydrostatic pressure (35 atm) has been studied compared with that under atmospheric pressure
(1 atm), using impedance measurements, gravimetric testing, adhesion testing and scanning electron
microscopy (SEM). The results show that high hydrostatic pressure accelerated the failure of the organic
coating by promoting diffusion of water in the coating, which speeds up water spread and electrochemical
reactions at the interface. The roughness of the coating and steel has been discussed from point of view
of their respective fractal dimensions Dfc and Dfdl , as deduced from impedance measurements
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Extensive exploration of deep ocean resources is greatly hindered by the unique and special conditions in deep ocean
environments, including relatively low temperatures, diverse dissolved oxygen levels, and high hydrostatic pressure. Amongst
these unique properties, the high hydrostatic pressure is
the major factor in terms of the corrosion performance of
metallic structures deployed in service in the deep ocean
environments [1,2].
The use of organic coatings is one of the most important
methods for materials corrosion protection in marine environments. Organic coatings can signicantly restrict the transport of
water and corrosive species to base metals/alloys, thus offering
appreciable corrosion protection. From the foregoing, successful
application of coatings for corrosion protection in deep ocean
environments requires in-depth knowledge and understanding of
the effect of high hydrostatic pressures on coating stability and
general performance. However, at present, the studies of organic
coating under high hydrostatic pressure have not received much
attention.
Electrochemical impedance spectroscopy (EIS) is a well-known
electrochemical technique to evaluate the performance of organic
coatings. It has been proven to be a powerful tool for acquiring
specic parameters of systems of organic coatings [310]. Usually, the tted parameters by equivalent electric circuit such as

Corresponding author. Tel.: +86 24 2392 5323; fax: +86 24 2389 3624.
E-mail address: [email protected] (Y. Li).
0013-4686/$ see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2011.11.067

the lm resistance and the coating capacitance can provide some

quantitative information bout the barrier properties of the coatings. When the coating is aged in humid conditions, the aqueous
aggressive medium penetrates the coating layer through pores
and lls in the free volume. Impedance measurements are as
well useful for assessing coating characteristics under such conditions and as such could nd useful application in deep ocean
environments [11].
The failure of organic coating on metals under atmospheric conditions mainly involves three steps [5,6,11]. First of all, the coatings
act as a barrier by slowing down the diffusion of water and aggressive species towards the coating/metal interface. Secondly, water
or other aggressive species arrive at the interface of coating/metal
and electrochemical reactions take place. Finally, the electrochemical reactions proceed at a rapid rate, yielding large amounts of
corrosion products, followed by coating failure. A large amount of
studies have focused on the transport mechanism of water in and
through organic coatings [11,12]. Other studies have focused on the
corrosion reaction at the coating/metal interface [13] and the effect
of various inorganic llers on the coating performance [14]. Only
relatively few studies have addressed the various stages involved
in the failure of organic coatings.
In this study, the failure of organic coating on mild steel under
high hydrostatic pressure has been studied and compared with
that under atmospheric pressure. Test procedures by EIS measurement, gravimetric testing, adhesion testing and SEM observation.
The failure process has been discussed. The application of fractal
dimensions Dfc and Dfdl , both of which reect the roughness of the
coating and steel surface respectively and inferred from the EIS data
have also been discussed.

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

43

the hydrostatic pressure at 350 m depth of the ocean). The corresponding tests are in 1 atm, which is atmospheric pressure.
2.3. EIS measurements
For all the experiments, a three-electrode cell was used with
a counter electrode of platinum (20 mm 20 mm) and an Ag/AgCl
reference electrode. The solid reference electrode is put into the
polyuortetraethylene tube and sealed by special materials for
keeping closeness [1,2]. The potential of the Ag/AgCl reference electrode was calibrated (within the pressure range 1100 atm) with
that of identical electrode maintained at 1 atm. The exposed area
of the coupons to the NaCl solution was 9.08 0.02 cm2 . EIS results
were measured by a PAR2273 Potentiostat integrated with a PC system. The software employed was Powersuite electrochemical test
system. A sinusoidal ac perturbation of 20 mV amplitude coupled
with the open circuit potential was applied to the metal/coating
system. At the beginning period of immersion, the EIS test was performed in the high frequency range from 100 kHz to 100 Hz for
measurements at short immersion periods [15] and in the range of
100 kHz to 10 mHz for prolonged immersion periods. The EIS data
were analyzed with ZSimpwin software.
Fig. 1. Schematic diagram of deep ocean simulation device: (1) nitrogen vessel, (2)
valve, (3) solid reference electrode, (4) thermocouple, (5) working electrode, (6)
counter electrode, (7) pressure meter, (8) automatic elevator, and (9) temperature
controller.

2. Experimental procedures
2.1. Coatings preparation
In this paper, only round nano-SiO2 particles (diameter 20 nm)
was used to make a single component pigmented coating system,
i.e. an E44 epoxy resin (bisphenol A) with a pigment concentration
of 5 mass% was formulated. The round nano-SiO2 particles is lled
in the real coating using in deep ocean. The amount of it is low
(about 5 mass%). In this work, xylene was chosen as the solvent,
E44 was dissolved by additional xylene with the proportion of 5:3.
The resin was mixed with the calculated amount of polyamide and
then stirred using a commercial magnetic stirrer machine for 1.5 h
to mix sufciently.
A mild steel panel of composition (wt.%): 4.67% Ni, 0.60% Cr,
0.46% Mo, 0.065% Al, 0.54% Mn, 0.076% C, and remainder Fe was
used as a substrate and cut into test coupons of dimensions
50 mm 50 mm 2 mm. The coupons were ground to 600-grit nish, degreased in acetone and ethanol and dried in air.
Two different kinds of coated samples were prepared for measurements: one is the plate sample, in which the coating is applied
on the steel substrate (mild steel panel) using a hand brush; the
other is the free lm, produced by applying the coating on a silicon chip (20 mm 20 mm). Both samples were cured in an oven
under the following conditions: 40 C for 4 h, 60 C for 20 h, and
room temperature (25 C) for one week. The average coating thickness was measured by a hand-held electronic gauge (PosiTector
6000; Defelsko) according to ISO 2808 standard procedures. The
thickness of free lm was 200 10 m. Steel panels coated only
with epoxy varnish coating without any pigment were used as the
reference specimens. And the thickness of the dry varnish coating
was 300 10 m.
2.2. Experimental setup
The test pressure vessel is shown in Fig. 1. The vessel is pressurized with a hydropneumatic pump using high purity nitrogen.
The hydrostatic pressure in the vessel is kept at 35 atm (equal to

2.4. Gravimetric testing, adhesion testing and tensile testing

Gravimetric data were obtained from measuring the mass gain
of free lm after different immersion periods at 1 atm and 35 atm
pressure. After the samples were removed from the solution, dried
quickly with lter paper and re-weighed. The tests were undertaken using a Sartorius MC5 microbalance (1 g resolution) and
were all run in triplicate and the average values reported.
The adhesive ability was conducted by PosiTest Pull-Off Adhesion Tester according to ISO4624-2002, which is a test method for
assessing the intensity of the outer paint to separate from substrates. The substrates used for experiment was cut into plates
(30 mm 30 mm 2.5 mm), painted with epoxy resin (E-44 and
polyurethane, obtained from commercial sources). The test dolly
size was 20 mm.
3. Experimental results
3.1. EIS measurements
Typical experimental impedance spectra, which were collected
during 800 h immersion in 3.5% NaCl solution under atmospheric
pressure (1 atm) and high hydrostatic pressure (35 atm), are presented as Nyquist plots and Bode plots in Figs. 2 and 3, respectively.
EIS results changes with immersion time in both environments.
Under 1 atm (see Fig. 2), within the initial 45 h, there was only
one capacitive loop. Beyond this time two capacitive characteristics are obvious within the entire frequency range. From 45 h to
about 400 h, the impedance prole changes with immersion time,
but remains more or less uniform after 498 h. Our impedance data
reveal three stages during the entire immersion period.
Under high hydrostatic pressure, the impedance data at the
initial stages reveal two capacitive characteristics and after 2 h,
there were two capacitive loops in spectrum. Both capacitive loops
decreased with immersion time. After about 100 h, the capacitive
loop at high frequency did not change with time, and there was
diffusion character at low frequency.
From the EIS results, at initial stage, there are two capacitive characters and at nal stage there is diffusion character,
both of which indicate that the corrosion behavior of organic
coating on steel has been changed by high hydrostatic pressure.
The value of impendence module changed from about 2 1011
to 1 109 Ohm cm2 under atmospheric pressure, while the value

44

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

Fig. 2. Typical EIS measurement results with immersion time under atmospheric pressure. (a)(c) Nyquist plots of coating on steel plates; (d) and (e) Bode plots of coating
on steel plates. In the gure, the scatter points are data points, and line is the tting results according to corresponding equivalent models.

changed from about 1 109 to 5 106 Ohm cm2 under high hydrostatic pressure, respectively seeing Fig. 2(e) and Fig. 3(f). The results
under atmospheric pressure were about two orders of magnitude higher than those under high hydrostatic pressure, indicating
that the high hydrostatic pressure decreased the protection of the
organic coating and accelerated the failure of organic coating.

3.2. Gravimetric testing results

Gravimetric testing from free standing lms (Fig. 4) showed a
good reproducibility for the organic coatings under both environments. The variations of mass gain in both environments follow
the same trend: an initial region of rapid increase, followed by a
region of slow growth and nally an almost steady state region.
Mass gain for the free lm suggests that water penetrates the coating and eventually saturates the coated sample as reected by the
steady state region. The initial stages of rapid mass gain show no
great differences under both environments. High hydrostatic pressure signicantly promotes the increase of mass in both the slow
growth region and in the nal steady state stage. These results

indicate that high hydrostatic pressure promotes the transport of

water in the organic coating.
3.3. Wet adhesion testing results
The change of the wet adhesion characteristics of organic coating/steel with immersion time is shown in Fig. 5. The dry adhesion
of coatings was 9.32 MPa and 9.43 MPa for atmospheric pressure
(1 atm) and high hydrostatic pressure (35 atm), respectively, both
of which are high. However, once the coatings were immersed into
water, dry adhesion turned into wet adhesion which decreased dramatically. When the immersion time was only 48 h, wet adhesion
had dropped signicantly to 3.34 MPa and 2.67 MPa (see arrows in
Fig. 5) for 1 atm and 35 atm respectively. Fig. 6 shows the morphologies of coatings on steel after immersion for 50 h in both
environments. There was no obvious change on the morphologies
of the coating after immersion under atmospheric pressure. While
the coating easily peels off from the steel after the same time of
immersion at high hydrostatic pressure. Fig. 6 shows the visual
pictures of coating on the steel after immersion 50 h in 3.5%NaCl
solution under 35 atm by camera. No macroscopic corrosion

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

45

Fig. 3. Typical EIS measurement results with immersion time under high hydrostatic pressure. (a)(d) are Nyquist plots of coating on steel plates; (e) and (f) are Bode plots
of coating on steel plates. In the gure, the scatter points are data points, and line is the tting results according to corresponding equivalent models.

products blister or cathodic disbondment etc. could be observed

on the surface of steel depending on the visual observation, which
indicates that the adhesion after immersion under high hydrostatic
pressure should drop more critically than that under atmospheric
pressure. Adhesion is signicantly lost before the formation of
macroscopic corrosion products. When the immersion time was
72 h, electrochemical corrosion and blistering become obvious. The
corresponding wet adhesion declined to a very low level, 2.91 MPa
and 2.31 MPa, respectively (see in Fig. 5), and such poor wet adhesion cannot ensure the sustainable performance of the organic
coatings.

4.1. Effect of high hydrostatic pressure on the diffusion process

4.1.1. Diffusion behavior
Different models have been developed in order to describe
the moisture absorption behavior of the materials. For onedimensional moisture, each sample is exposed to the same
accelerated environment and the moisture content can be
expressed as [1619]:

1
(2j + 1) D2
8
Mt
=1 2
exp
t
2
M
 j=0 (2j + 1)
L2


(1)

4. Discussion
The failure of organic coating is mainly a two-step process: the
rst step involves diffusion of aggressive species (such as water, O2 ,
ions) within the coating, while the second step involves the various
reactions at the coating/steel interface [12]. The former step initiates the failure process, and the interface reactions, which occur
concurrently with transport of diffusive species, nally propagate
the failure process.

Where j = 0, 1, 2 . . ., Mt is the initial weight of the moisture in the

material and M is the weight of moisture in the material when
the material is fully saturated in an equilibrium environment. D is
the mass diffusivity or coefcient in the composite. L is thickness
of specimen, t is the times.
Eq. (1) gives a unique curve when plotted as Mt /M vs t1/2 ,
where t = Dt/L2 , is a dimensionless time. For relatively small

46

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

Fig. 4. Weight increase with immersion time diagrams measured when the free
lms are immersed in 3.5% NaCl solution under both environments.

values of t, corresponding to Mt /M < 0.6, that is for short periods

of time, Eq. (1) reduces to Eq. (2) [16,17]:

Mt
2 D
=
t
(2)
M
L 
Eq. (2) shows that as following:
Mt = Kt m

(3)

K is a constant, and 0 < m 1. when m = 0.5, the diffusion follows Fick

diffusion; when m =
/ 0.5, the diffusion is other diffusion behavior,
non-Fickian diffusion.
According to Eq. (3), the m and diffusion coefcient under both
environments have been calculated. The results indicate that the
increase at the initial stage correspond to m value of 0.43, which
is near to 0.5. The value suggests that the diffusion of water under
hydrostatic pressure (35 atm) is Fickian, which also implies that
hydrostatic pressure does not change the behavior of water in
the organic coating. The diffusion coefcient, D was calculated;
D1 (under 1 atm) = 1.45E-09 cm2 /s; D35 (under 35 atm) = 2.29E09 cm2 /s, and reveals that high hydrostatic pressure promotes the
diffusion rate of water in the organic coating.
The calculation of diffusion coefcient (D) is mainly based on the
initial mass gain results at t < 50 h. In this stage, the morphologies of

Fig. 5. Adhesion measurement results with immersion time in 3.5% NaCl solution
under both environments.

Fig. 6. Equivalent electrical circuit of EIS data for organic coating in 3.5%NaCl solution under both environments (a) for Model A, (b) for Model B and (c) for Model
C.

coatings on the steel specimens indicate that there are no obvious

corrosion products on the surface, which means an insignicant
presence of O2 and Cl species at the interface, hence slow rates of
electrochemical reactions. For that reason, the obtained diffusion
coefcients mainly correspond to diffusion of H2 O species.
4.1.2. EIS analysis
For the EIS data under atmospheric pressure (1 atm), display
three distinct stages as a function of time. At the initial periods of
immersion the impedance of the system changes rapidly, decreasing as water penetrates the coating. The spectrum corresponds to
a capacitiveresistive response, approximately described by the
equivalent circuit A (Fig. 6a), in which the coating capacitance
and coating resistance are included in the model [12]. Because
of scattering effect resulting from the non homogeneous nature
of the coating surface, the constant phase elements (CPE) replace
the capacitance elements. The coating resistance is mainly from
the pores in the coating. Therefore, CPEc and Rc are included in
the model. As the immersion time increases, the EIS data can no
longer be satisfactorily tted to model A. Although one capacitive loop is observed in the spectrum, a poor quality t is obvious
in the whole frequency range. Considering that water and oxygen molecules reach the substrate surface through micro-pores in
the coating and electrochemical reaction may start, model B (see
Fig. 6b) in which CPEdl , the double layer capacitance and Rt the
charge transfer resistance are introduced to t the experimental
data [12]. After oxygen comes onto the metals surface, the electrochemical reactions happen at the coating/metal interface. At that
time, in EIS, the characteristics of coating and electrochemical reactions should both appear. The reason that only one capacitive loop is
phenomenally shown in Nyquist diagram is probably that the electrochemical reaction area in coating/metal interface is yet small in
this immersion period, resulting in the same order of magnitude of
the time constant related to the coating impedance as associated
to the electrochemical reaction impedance [12].

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

Fig. 7. The coating pore resistance Rc as a function of the immersion time in 3.5%
NaCl solution under both environments.

For the EIS results obtained under high hydrostatic pressure,

there are two capacitive loops at the initial immersion period. The
equivalent circuit model contains electrochemical reaction parameters, i.e., CPEdl and Rt . The t at initial periods agree with model B
and as the immersion time increases to more than 100 h, there are
clear diffusion characteristics, which may be produced by accumulation of corrosion products at the interface of coating/metal.
Therefore, considering that the coating has a barrier performance
against the diffusion of corrosion products from metal surface
towards coatings, the diffusion process may become the control
procedure in the faradaic processes [12]. Model C in Fig. 6(c) is used
to t EIS results in which Zdiff contains the diffusion capacitance
(CPEdiff ) and the diffusion resistance (Rdiff ).
The coating pore resistance Rc of the organic coating under both
environments, as determined from the tting results of the experimental EIS, was plotted as a function of immersion time in Fig. 7. The
values of Rc for both environments gradually decreased from about
1.7 1011 to 8.7 108 and about 9.3 106 to 1.1 106 Ohm/cm2 ,
for atmospheric pressure and high hydrostatic pressure, respectively. The results under atmospheric pressure were about three
orders of magnitude higher than those under high hydrostatic
pressure, indicating that the barrier of organic coating under both
environments were gradually deteriorated due to water uptake,
and the high hydrostatic pressure decreased the protection of the
organic coating. The coating capacitance Cc was plotted as a function of immersion time in Fig. 8. The values of Cc were gradually

Fig. 8. The coating capacitance Cc as a function of the immersion time in 3.5% NaCl
solution under both environments.

47

Fig. 9. The acceleration of water diffusion in the coating calculated of mass gain
results under both environments.

increasing during the initial 100 h of immersion and seems to

attain a steady state thereafter in the vicinity of 5 1011 F cm2
and 3 1011 F cm2 under high hydrostatic pressure and atmospheric pressure, respectively. Obviously, the value of Cc under
high hydrostatic pressure is larger than that of under atmospheric
pressure.
The rapid increase of the coating capacitance in the initial period
of immersion suggests that the water may permeate the coatings
through the micro-pores formed by solvent volatilization [12]. After
a certain time of immersion, the coating capacitance reaches a
steady value, which shows that the water uptake of the coating
attains a saturated state. The time of both environments is about
100 h, which is in agreement with the gravimetric results.
According to the experimental results, high hydrostatic pressure signicantly promotes the transport of water and ions in the
organic coating, which is reected by additional mass gain, higher
value of Cc and lower value of Rc . In order to clarify the effect of
hydrostatic pressure, the calculation of mass gain results has been
made. The derivation of mass gain with time can reect the diffusion rate of water in the coating [20]. The second derivation of
it shows the acceleration of the diffusion. Fig. 9 shows the acceleration of water diffusion in the coating calculated of mass gain
results. The results show that the acceleration of water diffusion
signicantly increases under high hydrostatic pressure at initial
stage. The results suggest that hydrostatic pressure may act as an
impulse to improve water entering into the coating at initial stage.
Another effect of hydrostatic pressure is the increase of pores
size during the serving time. Fig. 10 shows the optical pictures of
organic coating before and after serving under both environments.
The results show that the initial pores exist in the coating (Fig. 10a).
After serving at 1 atm, the pores increase in the coating (Fig. 10b).
After serving at 35 atm, the pores signicantly increase in the coating, and more seriously than that at 1 atm. The hydrostatic pressure
obviously increases the pores size. Based on the above analysis,
high hydrostatic pressure increases the roughness of the coating
and decreased the coating pore resistance Rc , both of which promote the transport of diffusive species, i.e. water and ions. This may
explain the observed accelerating effect of high hydrostatic on the
failure of organic coatings.
From above results, the direct effect of hydrostatic pressure is to
impute the water diffusion into the coating at initial and increase
the pores size in the coating. Both of effects improve the water
diffusion and accelerate the failure of the coating.

48

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

Fig. 11. The double layer capacitance Cdl as a function of the immersion time in 3.5%
NaCl solution under both environments.

reactions may happen without visible evidence of corrosion products, which means that both steps are not distinguishable.
4.2.1. Wet adhesion lose
Wet adhesion measurement results indicate that wet adhesion
is lost with prolonged immersion time. The high hydrostatic pressure accelerates the loss of wet adhesion, possibly by promoting
the adsorption and transport of water within the coatings (see
the gravimetric test results). High amounts of water at the interface of coating/steels facilitate the loss of wet adhesion as well as
formation of corrosion products. However, the wet adhesion measurement results show no obvious presence of corrosion products
at the surface of steel that could induce the loss of wet adhesion.
Therefore, the loss of wet adhesion should mainly be due to the
presence of high amounts of adsorbed water at the interface of
coating/steels. From above results, the hydrostatic pressure leads
to higher capacitance and lower resistance of the coating, which is
likely due to enhanced adsorption of the electrolyte (water gives
higher capacitance, ions give low resistance), and thus more water
at the coating/metal interface. The wet adhesion loses rapidly. This
is consistent with the wet adhesion lose data.

Fig. 10. Optical images of organic coating before testing (a), after immersion 800 h
in 3.5% NaCl solution under 1 atm (b), and under 35 atm (c).

4.2. Effect of high hydrostatic pressure on the coating/steel

interface reaction
The migration of O2 and Cl with water into the coating/steel
interface should lead to signicant interfacial reactions. Some studies have focused on the electrochemical reactions at the interface
of coating/steel in this stage. However, in this study, the dramatic
decrease of wet adhesion occurred before the obvious formation
of corrosion products at the interface of coating/steel, which suggests that in fact the interfacial reactions involve two steps: water
spread at the interface; electrochemical reactions after O2 and Cl
transported to the interface. It is interesting that electrochemical

4.2.2. Electrochemical reactions

The double layer capacitance, Cdl is plotted as a function of
immersion time in Fig. 11. The values of Cdl under high hydrostatic
pressure are larger than those under atmospheric pressure, which
suggest that the reaction areas are larger under high hydrostatic
pressure.
The charge-transfer resistance, Rt , is plotted as a function of
immersion time in Fig. 12. The higher value of Rt for the coating,
the greater corrosion resistance and the slow development of corrosion under the coatings. The results of Rt obviously indicate that
change of Rt under both environments decrease with immersion
time, implying higher rates of electrochemical reactions. The values of Rt under high hydrostatic pressure are signicantly smaller
than those under atmospheric pressure. The high hydrostatic pressure promotes the electrochemical reactions at the interface of
coating/steel, as visually reected by the surface morphologies
of coatings after 800 h of immersion as shown in Fig. 13. Fig. 13
shows the visual pictures of coating on the steel after immersion
800 h under both environments by camera. The corrosion products
appear on the steel after serving at 1 atm. While, more corrosion
products form on the steel and water bubble appear in the interface of coating/metal. The results indicate that the corrosion of the
coating under high hydrostatic pressure is more severe than that at

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

49

atmospheric pressure. And the adhesion of coating loses obviously

under high hydrostatic pressure. The more water and other, insidious species O2 and Cl penetrate the interface of coating/steels
under the effect of high hydrostatic pressure, the more severe the
electrochemical corrosion process at the interface.
4.3. Discussion of fracture dimensions (Df ) obtained from EIS

Fig. 12. The charge-transfer resistance Rt as a function of the immersion time in

3.5% NaCl solution under both environments.

The transfer of electric charge across the interface between two

substances is a ubiquitous process in modern technology, where
the two substances are very dissimilar, e.g., a metal electrode and
an aqueous or solid electrolyte [21]. Physically, the electric current
encounters Ohmic resistance in both substances and a capacitance
across the interface [21]. The general fractional values arise from
the fractal nature of the interface, i.e., the geometrical property that
a rough surface is self-similar under a scale transformation [21]. A
point of view of CPE and its association with rough and porous electrodes is given in this regard [21], and a kind of fractal dimensions
(Df ) can be calculated from the n values.
But for the organic coating, it is as a pure capacitance from
EIS without other changes. After water entering into it, it changes
and deviates from the fully capacitive behavior. It is reasonable to
believe that the deviation comes from the entering of water. In the
paper, authors also use the way to analysis the inner roughness of
coating and interface roughness of coating/metal. The results agree
with the morphologies observation. If clearer relation of the fractal
dimension from EIS and real change of interface roughness can be
set up, a new rapid measuring technique may be found and come
into real testing machine by measuring EIS parameters in the future.
Now the results are the initial step, the possible relation may exist
between them.
The non-idea capacitive response measured from EIS is caused
by many different kinds of heterogeneities, not just surface roughness. But for organic coatings, in the initial time of immersion,
the coating acts as a pure capacitance [12]. After a short time of
exposure, the impedance spectrum starts to deviate from the fully
capacitive behavior and a resistive behavior is visible at low frequencies, which means that the water transports into the coating
and interface between water and coating appears. The existing of
water should be the defects, pores and so on. And then some relation can be set up between the deviation of capacitance and surface
roughness. Some researchers use this way to analyze the surface
roughness at the interface of metal/oxides [4,22]. Based on this,
the inner roughness of the coating can also be reected by a CPEc
(CPE of the coating), i.e. described by value of nc tted from CPEc .
The changing of nc is mainly due to the roughness of the surface
of the material, and Dfc is a physical variable corresponding to the
roughness of the coating. If the roughness increases, then Dfc also
increases. The relation between nc and Dfc is shown as follows [4]:
n=

Fig. 13. The visual pictures of coating on the steel after immersion 800 hours in
3.5%NaCl solution under 1 atm (a), and under 35 atm (b) by camera.

1
1
and Df =
Df 1
n+1

The value of Dfc changes with immersion time has been shown in
Fig. 14(a). The results indicate that with immersion time increase,
Dfc increased gradually, and the roughness of the coating should
increase. Compared with the Dfc under atmospheric pressure, Dfc
under high hydrostatic pressure is obviously larger, which suggests that the roughness of coating may be increased by hydrostatic
pressure. The surface morphologies of the coatings after 800 h
immersion under both environments conrm this point (shown
in Fig. 13). Therefore, the Dfc obtained from EIS may reect the
change of coating roughness. If the relation between the fractal
dimension obtained from EIS and real change of interface roughness can be set up, a new rapid measuring technique may be found
by measuring EIS parameters in the future.

50

L. Liu et al. / Electrochimica Acta 62 (2012) 4250

Firstly, high hydrostatic pressure does not change the diffusion

behavior of water, and it still follows Fick diffusion. However, it
enhances the diffusion rate. The direct effect of hydrostatic pressure is to impute the water diffusion into the coating at initial and
increase the pores size in the coating. Both of effects improve the
water diffusion and accelerate the failure of the coating.
High hydrostatic pressure also promotes the interface reaction including water spread and electrochemical reactions, both
of which promote the failure of the organic coating in 3.5% NaCl
solution.
By calculation of fractal dimensions (Df ) from EIS, the change
of coating roughness and the interface roughness of coating/metal
can be tested.
References

Fig. 14. The fractal dimensions corresponding to the roughness of the coating Dfc (a)
as a function of the immersion time in 3.5% NaCl solution under both environments,
and the roughness of the interface of coating/steel Dfdl (b) as a function of the initial
50 h immersion in 3.5% NaCl solution under 35 atm.

The water spread at the interface of coating/steel is not easy to

distinguish from the electrochemical reactions. The roughness of
the surface of the materials can also be reected by the Df calculated
from the CPE of the double layer (CPEdl ). The CPEdl comes from the
electrochemical reactions and exists at the interface of water and
steel. Therefore, Dfdl , which is calculated from CPEdl , should reect
the roughness of the steel surface. The variations in the surface
roughness can be related to the spread of water at the interface of
coating/steel. The Dfdl of the coating in the initial 50 h of immersion
as calculated from CPEdl is shown in Fig. 14(b). The results indicate a clear change of interface roughness. The value of Dfdl initially
increases with immersion time, suggesting increasing variations in
the surface roughness. This trend is in agreement with the trend
of water availability at the interface of coating/steel. The value of
Dfdl subsequently decreases with time to a relatively stable value,
which means that the variations in surface roughness progressively
become negligible.
5. Conclusions
The high hydrostatic pressure decreased the protectivity of the
organic coating and accelerated the failure of organic coating.

[1] B. Liu, T. Zhang, Y. Shao, G. Meng, F. Wang, Effect of hydrostatic pressure on

the nature of passive lm of pure nickel, Materials and Corrosion 62 (2011)
269274.
[2] Y. Yang, T. Zhang, Y. Shao, G. Meng, F. Wang, Effect of hydrostatic pressure on
the corrosion behavior of NiCrMoV high strength steel, Corrosion Science
52 (2010) 26972706.
[3] Y. Shao, H. Huang, T. Zhang, G. Meng, F. Wang, Corrosion protection of Mg-5Li
alloy with epoxy coatings containing polyaniline, Corrosion Science 51 (2009)
29062915.
[4] Z. Yao, Z. Jiang, S. Xin, X. Sun, X. Wu, Electrochemical impedance spectroscopy
of ceramic coatings on Ti6Al4V by micro-plasma oxidation, Electrochimica
Acta 50 (2005) 32733279.
[5] J. Hu, H. Meng, J. Zhang, C. Cao, Degradation mechanism of long service life
Ti/IrO2 Ta2 O5 oxide anodes in sulphuric acid, Corrosion Science 44 (2002)
16551668.
M.J. Teodorovic,
Electrolyte penetration through
[6] V.B.Miskovic, D.M. Drazic,
epoxy coatings electrodeposited on steel, Corrosion Science 37 (1995)
241252.
[7] J. Titz, G.H. Wagner, H. Sphn, M. Ebert, K. Jttner, W.J. Lorenz, Characterization of organic coatings on metal substrates by electrochemical impedance
spectroscopy, Corrosion 46 (1990) 221228.
[8] F. Deorian, L. Fedrizzi, S. Rossi, Degradation of protective organic coatings after
cupping test, Corrosion 55 (1999) 10031011.
[9] F. Mansfeld, C.H. Tsai, Determination of coating deterioration with EISI basic
relationships, Corrosion 47 (1991) 958963.
[10] A.S. Castela, A.M. Simes, Assessment of water uptake in coil coatings by capacitance measurements, Progress in Organic Coatings 46 (2003) 5561.
[11] A. Amirudin, D. Thierry, Application of electrochemical impedance spectroscopy to study the degradation of polymer-coated metals, Progress in
Organic Coatings 26 (1995) 128.
[12] J. Zhang, J. Hu, J. Zhang, C. Cao, Studies of impedance models and water transport
behaviors of polypropylene coated metals in NaCl solution, Process in Organic
Coatings 49 (2004) 293301.
[13] M. Niknahad, S. Moradian, S.M. Mirabedine, The adhesion properties and corrosion performance of differently pretreated epoxy coatings on an aluminum
alloy, Corrosion Science 52 (2010) 19481957.
[14] X. Shi, T. Nguyen, Z. Suo, Y. Liu, R. Avci, Effect of nanoparticles on the anticorrosion and mechanical properties of epoxy coating, Surface & Coatings
Technology 204 (2009) 237245.
[15] D.H. Weijde, E.P.M. Westing, J.H.W. Wit, EIS Measurements on articial blisters
in organic coatings, Electrochimica Acta 41 (1996) 11031107.
[16] F. Bellucci, L. Nicodemo, Water transport in organic coatings, Corrosion 49
(1993) 235247.
[17] C. Perez, A. Collazo, M. Izquierdo, P. Merino, X.R. Novoa, Characterisation
of the barrier properties of different paint systems Part I: Experimental
set-up and ideal Fickian diffusion, Progress in Organic Coatings 36 (1999)
102108.
[18] S.C. See, Z.Y. Zhang, M.O.W. Richardson, A study of water absorption characteristics of a novel nano-gelcoat for marine application, Progress in Organic
Coatings 65 (2009) 169174.
[19] V.N. Nguyen, F.X. Perrin, J.L. Vernet, Water permeability of organic/inorganic
hybrid coatings prepared by solgel method: a comparison between gravimetric and capacitance measurements and evaluation of non-Fickian sorption
models, Corrosion Science 47 (2005) 397412.
[20] M. Huldn, C.M. Hansen, Water permeation in the coatings, Progress in Organic
Coating 13 (1985) 171194.
[21] S.H. Liu, Fractal model for the ac response of a rough interface, Physical Review
Letters 55 (1985) 529532.
[22] L. Nyikos, T. Pajkossy, Fractal dimension and fractional power frequencydependent impedance of blocking electrodes, Electrochimica Acta 30 (1985)
15331540.