EJERS, European Journal of Engineering Research and Science

Vol. 5, No. 3, March 2020

Fractal Dimension Analysis of Aluminum Corrosion

Roughness by Electrochemical and Optical Methods
Elsa C. Menchaca-Campos, Erika R. Villalba-Enciso, Victor Juarez-Nuñez, Miriam Flores-
Dominguez, Darwin Mayorga-Cruz, Rene Guardian Tapia, and Jorge Uruchurtu-Chavarín

 pattern interferometry [11], and structured light projection

Abstract—Fractal dimension is a versatile method to study [9- 13] to quantify the effects of metallic corrosion. In this
and evaluate corrosion from the point of view of metallic latter work it was applied the concepts of digital
conditions, namely: surface roughness, electrochemical interferometric and holographic microscopy were applied [9,
measurements and microscopy images. Aluminum corrodes at
12-17].
different rates, under different pH electrolytes with or without
the presence of chloride ions. In this work, corrosion and the A. EIS Fractal Analysis
surface roughness for aluminum corrosion at different pH
electrolytes: acid, with and without chlorides, neutral and basic Electrochemical impedance spectroscopy (EIS) evaluates
solutions, were obtained. It was measured and obtained using corrosion, providing kinetic information about the corrosion
electrochemical and optic techniques. The results of process and particularly, about parameters like: polarization
Electrochemical Impedance Spectroscopy (EIS) and Noise resistance (Rp), charge transfer resistance (Rct), double-layer
Measurements (ENM), Digital Holographic Microscopy capacitance (Cdl), solution resistance (Rs), etc.
(DHM), Scanning Electron Microscope (SEM) micrographs
Sometimes Nyquist plot obtained at free corrosion
and their respective Fractal Dimension analysis were obtained.
For the different experimental techniques and conditions, presents the shape of a depressed semicircle. The equivalent
fractal dimension was obtained and presented, reflecting the electrical circuit to represent this is the Randles circuit
surface condition of aluminum corrosion as a function of pH corresponding to a combination of a capacitor and a
solution. resistance in parallel. For this circuit, under certain
circumstances the ideal capacitor is replaced by a constant
Index Terms—Fractal Dimension, Electrochemical phase element (CPE) representing the double-layer
Techniques, Holographic Microscopy, Aluminum.
capacitance Cdl. The complex impedance Z (jω), of a
depressed semicircle is described as:
I. INTRODUCTION
Z =Rs+Rat / [1+ (jCdlRct)n] (1)
Euclidian geometry based on the properties of regular
forms such as point, lines, surfaces and volumes classifies For improved quality of electrical circuit fitting of
(ideal abstractions) them in dimensions shown by integer experimental data for depressed semicircles, CPE substitutes
numbers: 0, 1, 2 and 3, respectively. Fractal geometry the double-layer capacitance. The CPE is defined by the
considers real dimensions as fractional numbers relating to equation:
the form to be treated, hence its name fractal geometry. A
line may be spiky therefore the dimension should be Z CPE= Zo(j)-n = 1/Q(j)n (2)
between 1 and 2, a surface may be rough and dimension lies where ZCPE is the CPE impedance, Q is a proportionality
between 2 and 3 [1]. factor, j is (–1)½, ω is the angular frequency, and where n is
Corrosion is the gradual destruction of metallic materials related to the surface topographic calculation [18]. The CPE
by electrochemical reaction with their environment defines an electrode`s surface irregularities, promoting a
rendering the surface rough, porous or pitted, reducing the Nyquist plot semicircle depression [19-23]. The reaction
metal thickness and diminishing their mechanical properties. time constant (τ) and CPE element capacitance value are
Surface roughness is the measure of physical surface obtained through equations [5, 6]:
topography, i.e. slight irregularities over the surface. The
procedure to determine the surface roughness, is defining a Q =n Rtc (3)
height or depth that contains the profile information of the
surface. This can be obtained either by measuring the CCPE= [QRct1-n]1/n (4)
surface with a profile meter, through atomic force
microscopy (AFM) image, or by optical methods such as: where, CCPE is the obtained double layer capacitance from
optical interferometric measurements or light scattering [2- the CPE, and α is the semicircle`s depression angle:
8]. Roughness characteristic may be related to the corrosion
α= (1−n) × 90° (5)
attack observed [9].
More recently other methods were developed and applied where the n parameter is equal to 1 for an ideal capacitor.
like: digital in-line holography [9, 10], digital speckle In real systems, capacitive behavior is hardly observed
because of odd current distribution over the metal surface
Published on March 12, 2020. [24]. In the case when n = 1, the term (jωCdlRct)n reduces to
Authors are with Universidad Autónoma del Estado de Morelos (IICBA-
CIICAp), Cuernavaca, Morelos, Mexico. jωCdlRct. This is an indication of the level of heterogeneity

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 282

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

over the electrode [24, 25]. When the n value is higher than time dependence on the time-series and is the result of
0.5, it is related to a severe roughness of the metal surface, alterations that occurs in the data obtained. A time series is a
but when n is equal to 1, the surface is considered totally fractal when variations in the time data set over a specific
smooth. time interval known as the lag time are proportional to the
The degree of surface heterogeneity or roughness is lag time increased to the power H. According to Hurst’s
related to the fractal dimension [26]. A “fractal” is an object empirical law proposed [27]:
with complex structure, that reveals novel details as long as
its magnification degree is increased, it remains equal or R /S = ( / 2)H (8)
similar (self- similar) at different magnifications [27-29].
R is the difference between the maximum and minimum
Metals, with a fracture or surface heterogeneities could be
values of the variable, S is the standard deviation of the time
described through fractal geometry, by fractal dimension
(Dfs). series;  is the period of time measured and H is the Hurst
Taking into consideration the degree of the Nyquist exponent. Parameter H shows the appearance (“roughness”)
diagram semicircle`s depression, and to obtain the fractal of the time-series: when 0.5 < H < 1 represents an
dimension of the metal surface the following equation is undulating signal and “persistency” of the processed signal
utilized [21, 22]: that is concerned; when 0 < H < 0.5 is a `zig-zagged` signal
that indicates “anti-persistency”. When H is equal to 0.5 the
n = 1/ (Dfs–1) (6) process is random, independent of time lags. These
characteristics are associated to the physicochemical
where Dfs is the fractal dimension, taking values close to corrosion process [20-23].
2 for a smooth, or to 3 for a very rough surface. The fractal Therefore, fractal dimension Df , for signals as a function
dimension of an electrode obtained through electrochemical of time is:
impedance correlates with atomic force microscopy or other
optical techniques [9-11]. Df = 2 –H = (5 −) / 2 (9)
Electrochemical noise (ENM) technique is utilized to
determine the type of corrosion attack under different C. Phase-Shifting Digital Holographic Microscopy
corrosion conditions and its protection [30]. Data are Optical methods are widely used; one of such techniques
collected as a series of potential and/or current vs. time of known as optical interferometry is able to measure surface
sufficient length. Data analysis include: visual, statistical or
characteristics under 0.1 m without physical contact [37].
spectral analysis of time-records [31-35]. When localized
Also light scattering techniques have been used for
attack is present, the EN signal presents high-frequency
superficial roughness measuring at 0.1-3 m range [38-43].
transients of varying amplitudes.
More recently, laser speckle techniques were applied on
Spectral analysis of (ENM), has been used to determine
different science and engineering areas [44-52]; the spectral
the periodicity of the structure of EN signals [30, 31, 35].
speckle correlation being the most commonly used
The power spectrum is a plot of the spectral amplitude
technique for surface evaluations at 1.6-50 m interval [53-
against frequency of the electrochemical noise variable.
55].
Two types of behavior are commonly observed in the
K. Habib was probably the first in using holographic
spectra: white noise (random) independent of frequency, and
interferometric principles to monitor metallic electrodes
1/fβ behavior, related to the slope of the spectral density
immersed in saline solutions [56, 57]. Michelson
function (SDF) at higher frequencies. Different values of β
interferometry was applied to perform optical monitoring of
exponent have been obtained for specific processes [36].
corrosion processes on metallic samples under similar
Mandelbrot [26, 27] fractal geometry and mathematics
conditions [58]. Some other optical methods have been used
provide the resources for analysis and characteristics of the
and reported to quantify the effects of corrosion on metallic
variable in the time domain. The connection between the
surfaces. Digital holographic microscopy (DHM) is a
structure and scaling exponents of the EN time series, the
relatively new technique based on the concepts of optical
spectral density function (characterized by Df and β), and
holography, introduced by Dennis Gabor in 1948 [59].
oxidation reactions responsible for corrosion, can be thus
Digital Holography (DH) can be performed when an
obtained. For this case, the fractal dimension is defined as:
interference pattern is accomplished by superposition of an
Df = (5 − ) / 2 (7) object and reference beam, by digital sampling of a CCD
sensor and is finally sent to a computer as a numbers array,
preventing the need for chemical development utilized in
B. ENM Fractal Analysis
conventional holography. It also allows numerical
The ENM time records can be characterized by the information access to the amplitude and the phase data,
Rescaled Range or Hurst analysis and its Hurst exponent H; related to the object wave field [60-62]. Quantitative
being the variable time record a “random” fractal sharing the physical measurements of analyzed samples, like surface
same statistical properties and the level of details. The shape or optical thickness, are embedded in the phase
fractal geometry and its dimension Df , describes the distribution , of the digital hologram sampled by the
structure roughness of an EN time record and the
CCD. As such phase distributions suffer from module 
explanation for Df, H, and β values observed for the EN
ambiguities, phase unwrapping algorithms are necessary in
time-series spectra parameters.
order to obtain a 3D reconstruction of the samples [63].
The Hurst exponent related to β, shows the long-term
During a DH process the zero-order and the twin-image

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 283

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

terms that are always present, can be removed in a three initially binary (b/w) images established methodologies
steps stages during the computational numerical during the last decades, include applications to different
reconstruction, when the phase of a reference beam is tones of grey images, and the box- counting method can be
changed with a mirror installed on a piezoelectric transducer applied using free software such as DFRAC. Scanning
system. This method known as Phase-Shifting Digital electron microscope (SEM) micrographs were obtained
Holographic Microscopy (PSDHM), applies digital from different grey value images for fractal dimension D
holographic concepts to microscopy allowing a direct estimation, with an efficient application of converted binary
calculation of (phase shift) from three or more object [67].
subsequent interferograms with phase-shifts among them, Aluminum is an active metal, but develops a very stable
introduced by a piezoelectric transducer [64-66]. and protective passive film, decreasing corrosion for many
Regarding surface roughness characterization, the environs. The oxide is stable in neutral and some acid
statistical data processing has been used for determination of solutions, but is affected severely by alkali solutions.
certain parameters such as RMS roughness (Rrms) and Corrosion rates in acid media depend on the nature of
average roughness (Ra) where: L, Ra and Rrms represent anions; and its corrosion characteristics make it particularly
sampling length, arithmetic average of the absolute values of suitable in seawater and other chloride containing
heights z(x), and quadratic shifts average value with respect environments [68, 69].
to average height, respectively. However, as such values Pure aluminum is soft and weak, alloying and heat
depend on the analyzed characteristic length, it is necessary treatment of it improves the mechanical properties.
to follow another method to avoid this dependence such as Corrosion resistance and other characteristics that contribute
the self-affinity analysis. This method is usually used for to its wide application are: its colorless appearance, nontoxic
interface roughness and material surface studies, and corrosion products, electric and thermal conduction,
involves determining the Hurst exponent (H), also known as reflection, and good strength to weight ratio. All these
roughness exponent (mentioned above). properties make it attractive for new technological
Unlike former Ra and Rrms parameters, H is ideally applications in aerospace, energy, etc. Nowadays, with
independent from sampling size with H = 1 for a fully scarce resources, low budgets, care of the environment and
smooth profile (straight line). On the other hand, Hurst sustainable development, aluminum is fully recyclable
coefficient H is a quite versatile and simple concept, as it which makes it a very useful material.
has been used as a control parameter in roughness studies Several investigations have been reported [70-75] where
using speckle digital patterns obtained in return from the fractal analysis methods using different techniques were
laser beam scattering on a metallic rough surface [67], as used, related to metallic corrosion degradation, Corrosion
well as surface roughness determination on polymers [66, damage on the surface of an aluminum alloy used in the
67]. Therefore, within this work, the analysis of DHM is aircraft industry was characterized by applying fractal
focused on determination of the method`s sensitivity for dimension. Identification and classification of the type of
different roughness scales, where Hurst coefficient is corrosion damage were carried out through the fractal
obtained from optical profiles and digital holographic dimensions obtained. Also reports on the use of fractal
patterns acquired from scattered light coming from a known dimension to evaluate surface corrosion and inhibitors
rough surface. A detailed description of the method can be efficiency can be found [21, 22]. Another interesting
found on specialized literature [64]. application is the prediction and evaluation of general or
Fractal analysis is commonly used during the analysis of localized corrosion such as: pitting, inter-granular stress and
digital images demonstrating analogous structures at fatigue cracks, with fractal dimensions being different for
different ranges. Various approaches to calculate the fractal each case. This is especially in the case of the corrosion
dimension Df of objects recorded in these images have been resistant aluminum alloys and steels, although no direct
utilized in different scientific fields. Although a direct relationship has been found between the corrosion rate and
comparison between these different procedures has not been fractal dimension.
performed, however, reasonable outcomes have been
D. Objective
demonstrated. Even though, improvements in digital
imaging, enhanced resolution and Df calculation have been The aims of the present work was to show the possibility
attained, algorithms have not been thoroughly evaluated of fractal dimension relations obtained and compared from
with their application to high resolution imaging. different electrochemical techniques, of the corrosion
Nevertheless, these methods are close to or already included performance of aluminum samples, To determine the metal
in end-user applications that should provide quick and surface characteristics shown as the surface fractal
trusted results. dimension, calculated from EIS measurements and
Digital image figures are used for investigation, because compared with the one obtained through the H exponent
they are readily available for the purpose of storing, calculated from EN measurements, and evidenced by digital
processing and analysis. Various procedures to calculate holographic and SEM images and their fractal dimensions
fractal dimension Df of objects in digital imaging are used in obtained. The purpose was to evaluate surface corrosion
medicine, geology, geography or astronomy. The attack and corrosion resistance characteristics in different
commonest algorithm is the box-counting method (BCM), solutions.
because it straight forwardly implements the measurement
of object counts based on box size and number, representing
the pixel grid of digital images, as proposed [27, 28]. The

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 284

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

II. EXPERIMENTAL PROCEDURE coherent optical beams superposition on the plane of the
CCD sensor, are digitally stored in a PC where the CCD is
A. Materials
connected, for further 3D numerical reconstruction of the
Aluminum cylindrical samples with 2 cm in length and 6 samples. The fractal dimension of these images was
mm in diameter encapsulated in a commercial epoxy resin calculated using the box counting method applied to the
with an exposed surface of 1.13 cm2 to the solution. All obtained pixel matrix [80].
samples abraded with 600 SiC emery paper were rinsed with
distilled water followed by ethanol (C3H6O). Different pH CCD

solutions were prepared and used for the electrochemical

tests namely: sulfuric (pH 2 and 3) and hydrochloric (pH 5)
acids, potassium hydroxide and distilled water solutions.
Lens 3

B. Electrochemical Measurements microscopic

objective Mirror 3

Measurement of the electrochemical free corrosion Mirror 2 piezoelectric

actuator
potential of the working electrode, Ecorr, was performed
Lens 2
using a silver/silver chloride reference electrode. EIS Z
microscopic
measurements were performed with a Gamry 300 EIS objective

equipment in the frequency interval of 0.010 Hz to 10,000 sample

Hz with an amplitude of ±10 mV at the free corrosion

Pinhole
potential after reaching stability under immersion in the
solution, and a graphite rod was used as auxiliary electrode. Mirror 1
microscopic
objective He-Ne Laser
Afterwards, the depression angles from the Nyquist plots Lens 1
Fig. 1. The experimental PSDHM setup
analysis were obtained. Micrographs at different
magnifications were obtained from FESEM Jeol JSM-7600
model, to observe the corrosion attack over the metal III. RESULTS
surface, and fractal analysis over the images were performed Fractal analysis from the corroded metal holograms were
and obtained. performed and obtained, using the box counting method
The electrochemical noise measurements (ENM) were [27]. Reactive aluminum can develop a passive film
made gathering simultaneously the potential and current protecting it from electrochemical corrosion in neutral and
oscillations using two `identical` working electrodes and the some solutions but it is attacked in basic electrolytes [69].
reference electrode, at a sample rate of 0.5 second per read According to the Pourbaix diagram [76], aluminum
during a period of 1024 seconds. A computer controlled dissolves actively in acid or basic solutions while in neutral
ACM zero-resistance ammeter (ZRA) was used. DC signal media corrodes, forming a stable and protective corrosion
trend was removed from the raw noise data was through a products or passive film over the metal substrate. From Fig.
least-squares fitting method. The noise resistance (Rn) was 2 the free corrosion potential of aluminum immersed in
obtained as the ratio of the potential noise standard deviation different pH solutions can be observed, presenting different
over the current noise standard deviation (Rn = v/i). values aligned accordingly, as a function of pH: nobler or
From the EN time records, the Hurst coefficient and the less negative potentials for acid and active or more negative
fractal dimension of the electrochemical noise signals from for basic conditions. For neutral (pH 7) conditions, the free
the R/S analysis, were obtained [77- 79]. corrosion potential starts as active potential (-700 mV)
C. Phase-Shifting Digital Holographic Microscopy reaching the steady state at more positive values (-450 mV)
The experimental PSDHM setup is shown at Fig.1. An and above, probably reflecting the growth of the passive
film over the metal surface. For pH 10 and 11, values are
optical beam emission from a He-Ne laser of =633 nm
very similar and overlap in the active potential region.
wavelength and P0=35 mW output power, is firstly
amplified and collimated by means of a 25 m pinhole, 20X -300
microscopic objective and a 50 mm focal length lens (Lens
1). After being reflected from mirrors 1 and 2, light is
-600
passed through a second 50 mm focal length lens (Lens 2)
Potential (mV)

pH-2
and a beam splitter; this divides our beam in an object and a pH-3
-900 pH-5
reference beams with mutually orthogonal trajectories. The pH-7
pH-10
object beam is focused on the metallic sample by lens 2 and pH-11
pH-12
-1200
its reflection is focused on a CCD sensor by a 4X
microscopic objective and another 50 mm focal length lens
(Lens 3). The reference beam is reflected from mirror 3 and -1500
0 300 600 900 1200 1500 1800 2100
focused on the CCD sensor by another 4X microscopic Time (s)
objective and Lens 3, where interferes with the object beam. Fig. 2. Free corrosion potentials for aluminum as a function of pH
The third mirror is fixed onto a piezoelectric transducer set
on a moving platform, which operates as the mechanical Fig. 3 presents some electrochemical impedance (EIS)
phase-shifting system in a Z=5m range. results obtained, for different pH`s. From the Nyquist plots
Resulting holographic interferometric patterns, i.e. (Fig. 3) capacitive semicircles can be observed, from which
interferograms (HI), obtained by the object and reference electrochemical parameters were calculated including

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 285

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

charge transfer resistance and the depression angle, used for 0.010 pH-2
pH-3
the surface fractal dimension. Solution resistance varies for pH-5
0.008

current noise (mA/cm2)

H-7
different electrolytes as expected, but is not involved for pH-10
0.006
direct calculations of fractal dimensions. pH-11
0.004

0.002

0.000

-0.002

-0.004
500 1000 1500 2000
Time (s)
Fig. 4. Electrochemical current noise (ECN) time series as a function of pH

Fig. 5a presents the EIS charge transfer resistance and fig.

5b the ECN resistance as a function of pH obtained, both
showing similar trend but with approximately one order
difference values. The highest values observed were for the
hydrochloric acid (pH5) showing the effect of chlorides and
for neutral condition (pH 7) probably due to the presence of
Fig. 3. EIS Nyquist plots for aluminum immersed in different pH solutions the corrosion protective film. The lowest values were for
very acid or alkaline conditions, being the latter the lowest
Table I presents the electrochemical parameters obtained (pH 12) of them all, due to active dissolution. These results
from EIS measurements, in which the effect of the solution are in accordance to the known aluminum corrosion as a
pH can be clearly observed. Smaller charge transfers function of pH behavior, reported in the literature [76, 79].
resistances for lower and higher pH values, being the
highest for neutral solution and the smallest for the alkaline Impedance

pH 12 solution [76]. 10000

Impedance (ohm*cm2)

TABLE I: EIS PARAMETERS AS A FUNCTION OF PH

Df
pH n Depresión angle Θ Rct Cdl 1000
EIS
2 0.80 17.34 1.23E+02 1.02E-03 2.23
3 0.92 7.02 3.88E+03 7.96E-02 2.08
5 0.94 4.76 7.50E+03 5.35E-02 2.05 100
7 0.67 29.21 1.31E+04 1.18E-01 2.48
10 0.85 13.03 9.77E+02 1.10E-02 2.17
11 0.81 16.55 9.84E+02 1.10E-02 2.22 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
12 0.92 6.55 3.57E+01 8.36E-04 2.07
pH
a)
Fig. 4 presents the electrochemical current noise (ECN) 100000 Rn
time series results obtained, for different pH solutions. Some
ECN time series (Fig. 4) present distinctive stochastic
behavior namely: random noise (pH-11), low frequency
Rn (ohm.cm2)

(pH-5) or transient oscillations (pH-7) associated to film

10000
rupture-repassivation events; or a mixture of them. These
different oscillations relate to uniform, mixed, passive or
localized events over the metal surface [30].

1000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
b)
Fig. 5. Electrochemical a) EIS charge transfer resistance and b) ECN
resistance as a function of pH

As an example, Fig. 6 presents the fractal dimension

analysis using the FDim software, forthe SEM 500X sample
immersed in a solution pH 2, showing the graph to obtain
the Hurst coefficient and surface fractal dimension.

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 286

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

Fig. 6. SEM micrograph 500X for the sample immersed in a pH 2 solution, and the FDIM software graph to obtain the Hurst coefficient to calculate the
fractal dimension

Fig. 7 presents some SEM micrographs (500X) for Using FDim free software, micrographs analysis was
aluminum samples immersed at acid, neutral and basic pH performed and from the scale of grays, the fractal dimension
solutions. It can be observed the different surface conditions was calculated. Also, micrographs for different (1kX, 2kX
obtained after the samples were removed from the solution. and 4kX) magnifications were also obtained.

a)pH- 2 b) pH-7 c) pH-12

Fig. 7. SEM micrographs (1000X) for aluminum samples immersed under acid, neutral and basic solutions at different pH solutions

It can be seen in Fig. 8(a and b), that for acid or alkaline
solutions the fractal dimension changes slightly as a
function of SEM micrograph magnifications obtained for
2.1
different pH solutions. Being 2 for completely smooth and 3
for totally rough, for acid solutions the fractal dimension
Df SEM

diminishes for increased magnification. The opposite is true

for alkaline solutions where the fractal dimension increases pH-10
for increments in magnification. These could be due to the pH-11
formation of corrosion products and slower corrosion rates pH-12
in acid media, while faster dissolution rates render the
2.0
surface more uniform in alkaline conditions, being x500 x1000k x2000k x4000k
necessary greater magnification to observe changes in the Magnification
surface roughness. Nevertheless, fractal dimension values b)
Fig. 8. Fractal dimension changes as a function of SEM micrograph
suggest more or less smooth surfaces, reflecting the magnifications obtained for different pH solutions a) acid and b) alkaline
corrosion attack and state of the surface [80].
Some examples (Fig. 9 a and b) of fractal dimensions
obtained for SEM micrographs (500X), of samples
immersed in different pH solutions and compared to EIS and
2.1 ECN time series fractal dimensions calculated. SEM fractal
dimensions registered, suggest the surfaces being somewhat
Df SEM

smooth, presenting a direct relation obtained comparing

pH-2
pH-3 with EIS fractal dimensions obtained, using DFrac free
pH-5 software. As for ECN time series, the trends are similar but
showing an inverse relation was obtained, being the
difference greater for neutral solutions. The highest value
2.0 obtained was for the pH 5 solution showing the effect of
x500 x1000k x2000k x4000k
chlorides. In all cases low fractal dimensions were obtained,
Magnification
a) suggesting an apparent smoothness surface conditions.SEM
fractal dimension obtained seems to be underrated, maybe

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 287

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

due to the quality of the micrographs.

Df EIS
Df SEM 500X

2.6 2.14
2.13
2.5 2.12
2.11
2.10
2.4 2.09

Df SEM 500X
2.08
Df EIS

2.3 2.07
2.06
2.2 2.05
2.04
2.03
2.1 2.02
2.01
2.0 2.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
a) (b)
Df ECN
Df SEM 500X
1.14 2.14

2.13
1.12
2.12
1.10
2.11
Df SEM 500X

1.08
Df ECN

2.10

1.06 2.09

2.08
1.04
2.07
1.02 2.06
(c)
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
b)
Fig. 9. Fractal dimension of SEM micrograph (500X) obtained for different
pH solutions compared to a) EIS and, b) ECN time series fractal
dimensions

As an example, Fig. 10(a) presents microscopic image of

a corrosion pit produced on the metallic surface of the
sample which was previously immersed in a hydrochloric
acid solution (pH 5); Fig. 10(b) obtained by the PSDHM
technique shows the 3D mesh plot representation of the
unwrapped phase map calculated for the sample. Fig. 10(c- (d)
Fig. 10. (a) Pit over sample heterogeneous surface, (b) unwrapped phase
d) show corresponding x-y pixel profiles of Fig. 10(b). map mesh plot of the pit, (c) x-axis pixel profile, (d) y-axis pixel profile.
The sample was previously immersed in a hydrochloric acid solution at pH
5

Table II presents the Hurst exponent and fractal

dimensions from the surface reconstructions, calculated for
the samples immersed at different solutions. The lowest
fractal dimension obtained corresponds to the pH 3 and the
highest at pH 10 solutions, contrary to the SEM micrographs
fractal dimensions
TABLE II: FRACTAL DIMENSIONS AND HURST EXPONENT AS A FUNCTION
OF PH OBTAINED FROM DHM

(a) pH Df H
2 2.048 0.952
3 2.020 0.980
5 2.038 0.962
7 2.042 0.958
10 2.063 0.937
11 2.042 0.958
12 2.033 0.967

Comparison can be seen in Fig. 11 where both fractal

dimensions for different solutions can be observed. DHM
fractal dimensions obtained as a function of pH are similar,

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 288

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

presenting slightly lower values than the ones obtained IV. CONCLUSIONS
using SEM micrographs, that shows values related to Fractal dimensions obtained using different techniques
relatively smooth surfaces, but with an inverse trend. Bear in present low values for acid or basic conditions, being the
mind that surface dimension analysis was larger for SEM lowest (smooth surface) for the more basic the solution is.
micrographs than DHM reconstruction, being the possible The highest fractal dimension value belongs to the neutral
explanation for the results observed. condition for the EIS and SEM techniques, and the effect of
Df DHM
chlorides can be observed in the hydrochloric acid solution.
Df SEM 500X
However, ECM and DHM showed an opposite trend due to
2.13 2.13
2.12 2.12 the fact that in the former is the surface while in the latter
2.11 2.11 methods refer to the signal structure rather than the surface
2.10 2.10
2.09 2.09
itself, showing an inverse trend relation.
The aluminum surface roughness appears to depend on

Df SEM 500X
2.08 2.08
Df DHM

2.07 2.07 the surface finishing of the metallic sample and the solution
2.06 2.06
2.05 2.05
condition, showing correlation at a macro‐scale, through the
2.04 2.04 fractal dimension Df. In general terms the fractal dimension,
2.03 2.03 which is related to the statistical behavior of a measured
2.02 2.02
2.01 2.01 parameter, reflects the corrosion rate and attack condition.
2.00 2.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
pH
Fig. 11. Fractal dimensions as a function of pH for DHM and SEM
V. CURRENT AND FUTURE DEVELOPMENTS
Fractal analysis has proved to be a relatively simple and
Pure aluminum commonly presents an air formed passive useful analysis to characterize metal surface under different
film over the surface protecting it from corrosion. This conditions. Another advantage is the application for
passive film (Al2O3) is stable between pH 5 and 8, different experimental techniques results obtained, namely:
according to the Pourbaix [76] diagram. Below and above electrochemical, optics and image analysis. These may
active dissolution may occur at free corrosion conditions. reveal general, localized corrosion and fracture conditions
Fractal dimensions obtained from the different techniques at and morphology. Future developments into the
neutral conditions corroborates this rough surface condition incorporation of these techniques and analysis to practical
presenting the highest values since the passive film is working and in-service conditions is necessary from non-
formed over the finished abraded surface. destructive corrosion monitoring and routine inspection
Activation is the transition from passivity to the active instrumentation to correlate the fractal dimension to service
state after removal of the passive film, under the influence and failure events.
of a more negative electrode potential than the passive film
formation potential, or subject to acid or basic solutions due
to cathodic currents [25, 31, 32]. Under acid conditions, VI. CONFLICT OF INTERESTS
general corrosion proceeds and the corrosion rate diminishes
The authors confirm that this article content has no
as the pH increases. In general, the fractal dimensions
conflict of interest
obtained reflects this, observing lower values for more
acidic conditions. At neutral conditions (pH 7) the protective
film is formed, and in alkaline condition generalized
VII. ACKNOWLEDGEMENTS
corrosion proceeds with an increasing rate as pH becomes
more basic, being corrosion rates greater than in acid To CONACYT for grant Num. 605181 received by our
solutions [82]. The smallest fractal dimensions were students.
obtained at pH 12 suggesting a smooth surface. In the
presence of chlorides (hydrochloric acid pH 5) pitting
corrosion may proceed by disruption of the passive film at REFERENCES
local sites [32, 83], where fractal dimensions are relatively [1] B. B. Mandelbrot. The Fractal Geometry of Nature. New York: Ed.
W. H. Freeman & Co., 1991.
high probably due to mixed corrosion showing the effect of [2] A. J. V. Vaamonde. and J. J. D. González. Science and Engineering of
chlorides over the unstable film. the Surface Modification on Metallic Materials. Madrid: Ed. Consejo
The analysis of the fractal characteristics of the material`s Superior de Investigaciones Cientificas, 2001.
[3] M. P. Groover. Fundamental of Modern Manufacturing: Materials,
surface subjected to corrosion reflects these conditions. The Processes, and Systems. New York: John Wiley & Sons, 2010.
calculated values of fractal dimension D through [4] C. Galván. “New software for 2D and 3D roughness data analysis”.
electrochemical and optical measurements provide evidence Metrology Symposium, SM2008-S3D1-1027-1, 2008.
for the possible formation of a fractal structure of the [5] K.F. Sherwood. and J.R. Crookall. “Surface finish assessment by an
electrical capacitance technique”. Proc. Inst. Mech. Eng. 182 (3K).
surface during the corrosion attack. It also provides direct pp. 344-349.1967-1968.
evidence on the relation of corrosion rates and fractal [6] G.V. Blessing. and E. D. Eitzen. “Surface roughness sensed by
dimension, and compares favorably with the reported ultrasound”. Surf. Topogr. Vol. 1.pp. 143-158. 1998.
[7] C.W. O. Sosa. M. Sierra. C.A. P. Vargas. and L. A. Salcedo. “Analisis
literature [76, 82, 83]. de rugosidad por microscopia de fuerza atomica (AFM) y software
SPIP aplicado a superficies vítreas”, Revista Colombiana de Fisica.
Vol. 38, 2. pp.826-829.2006.

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 289

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

[8] R. P. Sierra. R. G. Romero-Paredes. and G. A. Rodriguez. “Estudio de [29] U. Cano. J. M. Malo. and J. Uruchurtu-Chavarín. “¿Es Caos o es
la morfología superficial e índice de refracción en películas Ruido Electroquímico? Corrosión del Aluminio en Medio Salino”.
nanométricas de silicio poroso”. Superficies y Vacio. Vol. 13. pp. 92- Rev. de Metal. Vol.28, 6. pp.348-352, 1992.
96. 2001. [30] E. Almeida. L. Mariaca. A. Rodríguez. J. Uruchurtu. and M. A.
[9] D. Mayorga. D. Uribe. R. Guardian. C. Menchaca-Campos. and Veloz. Electrochemical Noise Measurements for Corrosion
J. Uruchurtu. “Optical Interferometric Corrosion Protection Applications, ASTMSTP 1277.1996.
Evaluation of a Coated Copper in Sulphate/Chloride Solution”. [31] J. Uruchurtu. “Electrochemical Investigations of the Activation
Innovations in Corrosion and Materials Science (Formerly Recent Mechanism of Aluminum,” Corrosion.Vol. 47,6, pp. 472-479. 1991.
Patents on Corrosion Science). Vol. 7, 1. pp. 46-51. 2017. [32] J. Uruchurtu. NACE Conference in Localized Corrosion.Orlando:
[10] J. A. Marbán-Salgado. D. Mayorga-Cruz. J. Uruchurtu-Chavarín.and NACE-9, 1987.
O. Sarmiento-Martínez. “Quantification of corrosion products [33] J. M. Bastidas. and J. M. Malo. “Electrochemical Noise 1/f in the
formation onto a copper sample by digital holographic microscopy”. Study of Corrosion Inhibitor Efficiency”. Rev. Metal. Vol. 21, 6.
Opt. Pura y Apl.Vol. 46,1. pp. 49-54. 2013. pp.337-340. 1985.
[11] Oscar Sarmiento-Martinez. José A. Marban-Salgado. Darwin [34] C. T. Chen. and B. S. Skerry. “Assessing the Corrosion Resistance of
Mayorga-Cruz. Jorge Uruchurtu-Chavarin. and Estela Sarmiento- Painted Steel by Electrochemical Noise Technique”. Corrosion.Vol.
Bustos. “Rough Surfaces Profiles and Speckle Patterns Analysis by 47, 8. pp. 598-611. 1991.
Hurst Exponent Method”. Journal of Materials Science and [35] J. Uruchurtu-Chavarin. and J. M. Malo. “Electrochemical Noise as a
Engineering. Vol. B3, 12, pp.759-766. 2013. Powerful Electrochemical Technique for Corrosion Studies,”
[12] E.H. Sherrington and H. Smith. “Modern measurement techniques in Research Trends. Vol. 2. pp. 49-58. 1997.
surface metrology: Part I. Stylus instruments, electron microscopy [36] R. F. Voss. and J. Clarke. “Flicker (1/f) noise: Equilibrium
and non-optical comparators”. Wear. Vol. 125. pp.271-288.1988. temperature and resistance fluctuations”. Phys. Rev.Vol. B 13, 556-
[13] E.H. Sherrington. and H. Smith.” Modern measurement techniques in 561. 1976.
surface metrology: Part II. Optical instruments”. Wear. Vol. 125. pp. [37] K. Habib, F. Al-Sabti, H. Al-Mazidi, “Optical corrosion-meter”.
289-308.1988. Proc SPIE, 2577, pp. 210-217, 1995.
[14] L. Li. C. Wang. B. Yuan. and S. Chen. “Numerical reconstruction of [38] D. H. Hensler, “Light scattering from fused polycrystalline aluminum
digital holograms for the study of pitting dynamic processes of the oxide surfaces”. App. Optics.Vol. 11, pp. 2522-2528.1972.
X70 carbon steel in NaCl solution”, Electrochem. Commun. Vol. 10. [39] T. V. Vorburger. E. Marx. and T. R. Lettieri. “Regimes of surface
pp.103-107. 2008. roughness measurable with light scattering”. App. Optics.Vol. 32,
[15] P. E. Klages. In Situ Real-Time Visualization and Corrosion Testing 19.pp. 3401-3408.1993.
of Stainless Steel 316LVM with Emphasis on Digital in-line [40] U. Persson. “In-process measurement of surface roughness using light
Holographic Microscopy.PhD Thesis. Dalhousie University. Halifax. scattering”. Wear. Vol. 215.pp. 54-58.1998.
Nova Scotia. 2012. [41] C. J. Tay. S.H. Wang. C. Quan. and H. M. Shang. “In situ surface
[16] N. Andrés. J. Lobera. M. P. Arroyo, and L. A. Angurel. “Two- roughness measurement using a laser scattering method”. Optics
dimensional quantification of the corrosion process in metal surfaces Communications.Vol. 218.pp. 1-10.2003.
using digital speckle pattern interferometry”. Appl. Opt. Vol. 50, pp. [42] G.Y. Tian. R. S. Lu. and D. Gledhill. “Surface measurement using
1323-1328. 2011. active vision and light scattering”. Optics and Lasers in Eng. Vol. 45.
[17] P. S. Huang. F. Jin. and F. P. Chiang. “Quantitative evaluation of pp. 131-139.2007.
corrosion by a digital fringe projection technique”. Opt. Lasers and [43] R.L. Voti, G.L. Leahu, S. Gaetani, C. Sibilia, V. Violante, E.
Eng. Vol. 31, pp. 371-380. 1999. Castagna, and M. Bertolotti. “Light scattering from a roughmetal
[18] T. de J. Licona-Sanchez. G.A. Alvarez-Romero. L. H. Mendoza- surface: theory and experiment”. J. Opt. Soc.Am.Vol. B26, 8.
Huizar. C. A. Galan-Vidal. M. Palomar-Pardave. M. Romero-Romo. pp.1585-1593. 2009.
H. Herrera-Hernadez. J. Uruchurtu, and J.M. Juarez-Garcia. [44] U. Persson. “Surface roughness measurement on machined surfaces
“Nucleation and Growth Kinetics of Electrodeposited Sulfate-Doped using angular speckle correlation”. J. of Mats. Process Tech.Vol. 180.
Polypyrrole: Determination of the Diffusion Coefficient of SO42− in pp. 233-238.2006.
the Polymeric Membrane”. J. Phys.Chem. Vol.B114, 30. pp.9737- [45] M. A. Rebollo. M. R. Landau.E. N. Hogert. N. G. Gaggioli.and M.
9743. 2010. Muramatsu. “Roughness determination by direct visual observation of
[19] J. L. Guiñon.J. García-Anton. V. Perez-Herranz. and G. Lacaste. “Use the speckle pattern”. Optics and Laser Tech. Vol. 27, 6.pp. 355-
of Differential Pulse Polarography to Study Corrosion of Galvanized 356.1995.
Steel in Aqueous Lithium Bromide Solution”. Corros. Vol. 50, 2. pp. [46] D. Leger.and J. C. Perrin. “Real-time measurement of surface
91-97.1994. roughness by correlation of speckle patterns”. J. Opt. Soc.Am. Vol.
[20] D.M. García-García. J. García-Antón. M. A. Igual-Muñoz. and E. 66, 11.pp. 1210-1217.1976.
Blasco-Tamarin, “Effect of cavitation on the corrosion behaviour of [47] S. L. Toh. H. M. Shang.and C. J. Tay. “Surface-roughness study using
welded and non-welded duplex stainless steel in aqueous LiBr laser speckle method”. Optics and Lasers in Eng. Vol. 29.pp. 217-
solutions”. Corros. Sci. Vol.48, 9. pp. 2380-2405.2006. 225. 1998.
[21] E. Samiento-Bustos. J.G. González Rodriguez. J. Uruchurtu. G. [48] H. Nitta. and T. Asakura. “Measurements of fine particle size using a
Dominguez-Patiño. and V. M. Salinas-Bravo. “Effect of inorganic speckle correlation technique”.Meas. Sci. Technol. Vol. 1. pp. 131-
inhibitors on the corrosion behavior of 1018 carbon steel in the 135.1990.
LiBr + ethylene glycol + H2O mixture”. Corros. Sci. Vol. 50, 8. [49] M. Hamed. and M. Saudy.“Computation of surface roughness using
pp.2296-2303. 2008. optical correlation”. J. of Physics. Vol. 68, 5.pp. 831-842. 2007.
[22] E. Samiento-Bustos, J.G. González Rodriguez, and J. Uruchurtu, “A [50] H. Fujii. T. Asakura. and Y. Shindo. “Measurement of surface
study of the corrosion inhibition of carbon steel in a bromide solution roughness properties by using image speckle contrast”. J. Opt. Soc.
using fractal analysis”. Surf. Coat. Technol. Vol. 203. 1-2. pp. 46-51. Am.Vol. 66, 11. pp.1217-1222. 1976.
2008. [51] L.C. Leonard. and V. Toal.“Roughness measurement of metallic
[23] M. Saremi. C. Dehghanian. and M. Sabet. “The effect of molybdate surfaces based on the laser speckle contrast method”. Optics and
concentration and hydrodynamic effect on mild steel corrosion Laser in Eng. Vol. 30. pp. 433-440.1998.
inhibition in simulated cooling water”. Corros. Sci. Vol. 48, 6. [52] X.Z. Zhao. and Z. Gao.“Surface roughness measurement using
pp.1404-1412. 2006. spatial-average analysis of objective speckle pattern in specular
[24] K. Hladky. and J. L. Dawson. “The Measurement of Corrosion Using direction”. Optics and Lasers in Eng. Vol. 47. pp. 1307-1316.2009.
Electrochemical 1/f Noise”. Corros. Sci.Vol. 22, pp. 231-235. 1982. [53] B. Dhanasekar. N. K. Mohan. B. Bhaduri. and B. Ramamoorthy.
[25] C. Gabrielli. F. Huet. M. Keddam.and R.Oltra. NACE Conference in “Evaluation of surface roughness based on monochromatic speckle
Localized Corrosion. Orlando: NACE-9. 1987. correlation using image processing”. Precision Eng. Vol. 32. pp. 196-
[26] B. B. Mandelbrot. E. Dann. A. Passoja. And J. Paullay. “Fractal 206.2008.
character of fracture surfaces of metals”. Nature, Vol. 308, pp.721– [54] V. M. Huynh. S. Kurad. W. North.“Texture analysis of rough surfaces
722.1984 using optical Fourier transform”. Meas. Sci. Technol. Vol. 2. pp. 831-
[27] J. Feder. Fractals. New York: Plenum Press. 1989. 837.1991.
[28] U. Cano. J. M. Malo. and J. Uruchurtu-Chavarin. “Aluminum [55] S. L. Toh. C. Quan. K. C. Woo. C. J. Tay. and H. M. Shang.“Whole
Corrosion in Chloride Solution. Is it Chaos or is it Electrochemical field surface roughness measurement by laser speckle correlation
Noise?” Corrosao e Proteccao de Materiais. Vol. 13, 1. pp. 6-12. technique”. Optics and Laser Technology. Vol. 33. pp. 427-434.2001.
1994.

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 290

 EJERS, European Journal of Engineering Research and Science
Vol. 5, No. 3, March 2020

[56] K. Habib. “Holographic interferometric in predicting cathodic [70] F. Jin. And F. P. Chiang. “Nondestructive evaluation of corrosion by
deposition of metals in aqueous solution”. Proc. SPIE.Vol. 1230.pp. fractal geometry”. Res. Nondestr. Eval., Vol. 7, 4. pp. 229–238.1996.
293-296. 1990. [71] G. N. Frantziskonis. L. B. Simon. J. Woo. and T. E. Matikas.
[57] K. Habib. W. Mohammad. F. Karim. and J. Dutta. “Resistance Values “Multiscale characterization of pitting corrosion and application to an
of Aluminum Oxide Film in Situ during Anodization of Aluminum by aluminum alloy”. Eur. J. Mech. A/Solids.Vol. 19.pp. 309–318. 2000.
Fabry-Pérot Interferometry”. ECS Trans.Vol. 80.10, pp. 1221-1229. [72] Z. Chuan. C. Yanhui. And Y.Weixing. “The use of fractal dimensions
2017. in the prediction of residual fatigue life of pre-corroded aluminum
[58] Mayorga-Cruz, J. Uruchurtu-Chavarín, O. Sarmiento-Martínez, P. A. alloy specimens”. Int. J. Fatigue. Vol. 59. pp. 282–291.2014.
Márquez-Aguilar, J. Castrellón-Uribe. “Estimation of Corrosion [73] S. Xu. and B. Qiu. “Experimental study on fatigue behavior of
Parameters in Electrochemical Systems Using Michelson corroded steel”. Mater. Sci. Eng., A, Vol. 584. pp. 163-169.2013.
Interferometry”. Proceedings of the International Society for Optical [74] X.Wang. J.Wang. C.Fu. and Y.Gao. “Determination of corrosion type
Engineering (SPIE), 6046(0J), 115-120, 2006. by wavelet_based fractal dimension from electrochemical noise”. Int.
[59] D. Gabor, “A new microscopic principle”. Nature, 161. pp. 777-778. J. Electrochem. Sci. Vol. 8, pp. 7211–7222.2013.
1948. [75] S. Xu. And Y.Weng. “A new approach to estimate fractal dimensions
[60] T. Kreis. Handbook of Holographic Interferometry, Optical and of corrosion images”. Pattern Recognit. Lett. Vol. 27. 16. pp. 1942-
Digital Methods. Weinheim:Wiley-VCH Verlag GmbH & Co. KGaA. 1947.2006.
2005. [76] M. Pourbaix. Atlas of Electrochemical Equilibria in Aqueous
[61] M. Takeda. I. Hideki. and S. Ki. “Fourier-transform method of fringe- Solutions. Houston: NACE International, CEBELCOR. 1966.
pattern analysis for computer-based topography and interferometry”. [77] M. A. González-Núñez, and J. Uruchurtu-Chavarín. “R/S fractal
J. Opt. Soc. Am. Vol. 72. pp. 156-160. 1982. analysis of electrochemical noise signals of three organic coating
[62] J. W. Goodman. and R. W. Lawrence. “Digital image formation from samples under corrosion conditions”. J. of Corrosion Sci. and
electronically detected holograms”. Appl. Phys. Lett. Vol. 11. pp. 77- Eng.Vol. 6. C117. 2003.
79. 1967. [78] M. Moon. and B. Skerry. “Interpretation of Corrosion Resistance
[63] U. Schnars. and W. Jüptner. Digital Holography: digital hologram Properties of Organic Paint Films from Fractal Analysis of
recording, numerical reconstruction and related techniques. Berlin- Electrochemical Noise Data”. J. of Coat. Tech. Vol. 67, 843, pp. 35-
Heidelberg: Springer-Verlag. 2005. 41. 1995.
[64] I. Yamaguchi. and T. Zhang. “Phase-shifting digital holography”. [79] B. Zaid. D. Saidi. A. Benzaid. and S. Hadji. “Effects of pH and
Opt. Lett.Vol. 22. pp. 1268-1270.1997. chloride concentration on pitting corrosion of AA6061 aluminum
[65] T. Zhang, and I. Yamaguchi. “Three-dimensional microscopy with alloy”, Corr. Sci. Vol. 50, 7, pp. 1841-1847. 2008.
phase-shifting digital holography”. Opt. Lett.Vol. 23, 15. pp. 1221- [80] O. Sarmiento-Martinez, J. A. Marban-Salgado, D. Mayorga-Cruz, J.
1223.1998. Uruchurtu- Chavarin and E. Sarmiento-Bustos, “Rough Surfaces
[66] M. K. Kim. Digital Holographic Microscopy: Principles, Techniques Profiles and Speckle Patterns Analysis by Hurst Exponent Method”.
and Applications. New York: Springer-Verlag. 2011. J. of Mat. Sci. and Eng. B 3. Vol. 12. pp.759-766. 2013.
[67] M. Mayrhofer-Reinhartshuber and H. Ahammer. “Pyramidal fractal [81] J. Garcia-Sucerquia.C.Trujillo. and J. F. Restrepo. “Reconstrucción de
dimension for high resolution images”. Chaos. Vol. 26. 073109. 2016. hologramas de microscopía holográfica digital en línea a velocidad de
[68] H. Ahammer and T. T. J. DeVaney. “The influence of edge detection video”. Dyna.Vol. 79. 173. pp. 25-31. 2012.
algorithms on the estimation of the fractal dimension of binary digital [82] R. B. Mears, “Aluminum and Aluminum Alloys," Chapter 10.
images”. Chaos.Vol. 14, 183. 2004. Corrosion Handbook. Edited by H. H. Uhlig. New York: John Wiley
[69] M. G. Fontana. and N. D. Greene. Corrosion Engineering. New York; and Sons, Inc. 1948.
Mc Graw Hill. 1978, [83] H. Kaesche. Localized Corrosion, Ed. R. W. Staehle, Houston:
NACE, NACE-3. 1974.

DOI: http://dx.doi.org/10.24018/ejers.2020.5.3.1816 291