Corrosion Protection Properties of Formed,
Organically Coated Electro-galvanised Steel
PhD Thesis
Dr. rer. nat.
Faculty of Science
at the
University of Paderborn
Submitted by
Ingo Klüppel
from Düsseldorf
Paderborn, 2008
Defence on December 5, 2008
First referee: Prof. Dr.-Ing. Guido Grundmeier
Second referee: Prof. Dr. Wolfgang Bremser
Acknowledgement
5
Acknowledgements
The present work was carried out between June 2004 and June 2007 at the Dortmunder
Oberflächencentrum (DOC), a company of ThyssenKrupp Steel and the Max-Planck-Institut
für Eisenforschung (MPIE).
I want to thank Dr.-Ing. Michael Steinhorst and Prof. Dr. Martin Stratmann for giving the
opportunity to work at both the DOC and the MPIE and Dr. Christoph Filthaut for working in
his group.
My special thanks go to Prof. Dr.-Ing. Guido Grundmeier, my workgroup leader at the MPIE
and my first referee at the University of Paderborn. He was always ready to listen to questions
and gave productive advice for solving problems that occurred during this thesis.
I also would like to thank my tutor at the DOC Dr.-Ing. Bernhard Schinkinger who introduced
me to the industrial work and gave me a lot of hints for following in his foodsteps.
Prof. Dr. Bremser gets my acknowledgement for taking over the refereeing of my thesis and
the support with polyelectrolytes synthesised by Dr. Oliver Seewald.
Further thanks go to my roommates Dr. Krasimir Nikolov, Dipl.-Ing. Richard Weinhold,
Dipl.-Phys. Tobias Titz and Dipl.-Phys. René Vlasak for their discussions which were
sometimes not content-related but always interesting and rewarding. These discussions went
on with Dr. Miroslaw Giza during our travel times by train and car, I also thank him.
There are a lot more people at both DOC and MPIE I would like to thank but the space would
not suffice to name all of them.
I wish to thank the Fraunhofer Group at the DOC for supporting me by the laser cutting of the
miniature and corrosion samples. I am grateful to the group of Dr. Kessler for its help in
analysing the forming behaviour of the different samples. Many thanks go to Mr. Petzold for
some of the FE-SEM images.
Last but not least I would like to thank Lirija, my parents and my sister for their support and
patience during the preparation of this work.
The work was carried out with a financial grant from the Research Fund for Coal and Steel of
the European Community (contract no. RFS-CE-04031).
Acknowledgement
6
Index
7
Index
Acknowledgements _________________________________________________________ 5
Index_____________________________________________________________________ 7
Symbols _________________________________________________________________ 11
1 Motivation ___________________________________________________________ 13
2 Fundamentals ________________________________________________________ 16
2.1 Forming of polymer coated metals _________________________________________ 16
2.1.1 Forming of metals ____________________________________________________________ 16
2.1.1.1 Stress-Strain diagram_____________________________________________________ 16
2.1.1.2 Elastic and plastic deformation of single crystals _______________________________ 18
2.1.1.3 Forming of zinc and electrogalvanised zinc surfaces_____________________________ 20
2.1.1.4 Forming degree _________________________________________________________ 20
2.1.1.5 Forming-limiting-diagram _________________________________________________ 21
2.1.2 Forming of polymers and thin polymer films _______________________________________ 24
2.1.2.1 Molecular understanding __________________________________________________ 24
2.1.2.2 Relevance of interfaces ___________________________________________________ 25
2.1.2.3 Theory of crack formation in polymers and at polymer / oxide interfaces ____________ 26
2.1.3 Repair mechanisms of defects in protective coatings _________________________________ 27
3 Fundamentals of applied microscopic, spectroscopic and electrochemical methods_ 29
3.1 Electrochemical Impedance Spectroscopy ___________________________________ 29
3.1.1 General description of the impedance _____________________________________________ 29
3.1.2 Plotting of impedance data _____________________________________________________ 31
3.1.3 Pseudo-Linearity of the system __________________________________________________ 32
3.1.4 Elements of the equivalent circuit ________________________________________________ 33
3.1.4.1 Electrolyte resistance _____________________________________________________ 33
3.1.4.2 Double layer capacitance __________________________________________________ 34
3.1.4.3 Polarisation Resistance ___________________________________________________ 34
3.1.4.4 Coating capacitance ______________________________________________________ 35
3.1.5 Modelling of equivalent circuits _________________________________________________ 35
3.1.5.1 Ideal coating____________________________________________________________ 35
3.1.5.2 Metal electrode__________________________________________________________ 36
3.1.5.3 Non-ideal coating________________________________________________________ 37
3.2 Quartz crystal mircobalance ______________________________________________ 39
3.3 Vibration spectroscopy for thin film analysis ________________________________ 42
3.3.1 DRIFT and FTIR-ATR spectroscopy _____________________________________________ 42
Index
8
3.3.1.1 Fundamentals ___________________________________________________________42
3.3.1.2 DRIFT _________________________________________________________________43
3.3.1.3 FTIR-ATR spectroscopy___________________________________________________43
3.3.2 Raman spectroscopy ___________________________________________________________44
3.3.2.1 Surface Enhanced Raman Spectroscopy (SERS) ________________________________45
3.4 Scanning electron microscopy _____________________________________________ 47
3.5 Focused Ion Beam (FIB)__________________________________________________ 49
4 Experimental _________________________________________________________ 50
4.1 Materials, Electrolytes and Parameters _____________________________________ 50
4.1.1 Substrate and sample preparation _________________________________________________50
4.1.1.1 Preparation of corrosion protection primer samples ______________________________50
4.1.1.2 Preparation of polyelectrolyte coated samples __________________________________50
4.1.2 Forming of samples ___________________________________________________________51
4.1.2.1 Forming for Ex-Situ Analysis _______________________________________________51
4.1.2.2 GOM® Grid Evaluation____________________________________________________51
4.1.2.3 Finite Element Simulation__________________________________________________52
4.1.3 Preparation of biaxial formed samples for corrosion testing ____________________________52
4.1.4 Preparation of plane strain formed samples for scanning in-situ Raman / EIS investigation ____52
4.1.5 Phosphating and ED-paint application of formed corrosion protection primer coated samples __53
4.1.6 FE-REM investigations and FIB preparation ________________________________________53
4.1.7 Raman and IR spectroscopy _____________________________________________________54
4.1.8 Glow Discharge Optical Emission Spectroscopy (GDOES)_____________________________54
4.1.9 Laser Optical Emission Spectroscopy (Laser-OES) ___________________________________54
4.1.10 Electrochemistry and electrolytes ______________________________________________55
4.1.11 Corrosion Testing___________________________________________________________55
4.2 In-situ Electrochemical Impedance Spectroscopy (EIS) and Forming Setup _______ 57
4.3 In-situ Quartz Crystal Microbalance / Raman Setup __________________________ 60
4.4 Scanning in-situ Raman / Electrochemical Impedance spectroscopy Setup ________ 62
5 Results ______________________________________________________________ 64
5.1 Characterisation of corrosion protection primers (CPP) _______________________ 64
5.2 Uniaxial, biaxial and plane strain forming of corrosion protection primers________ 69
5.2.1 Strain evaluation of formed samples_______________________________________________69
5.2.2 Forming behaviour of the electro galvanised steel substrate ____________________________74
5.2.3 Uniaxial forming of dry corrosion protection primers _________________________________75
5.2.4 Uniaxial forming of wetted corrosion protection primers_______________________________77
5.2.5 Biaxial forming of corrosion protection primers _____________________________________78
5.2.6 Plane strain forming of corrosion protection primers __________________________________79
Index
9
5.2.7 Cleaning and phosphating of biaxially formed primers________________________________ 80
5.2.8 Open Circuit Potential during Phosphating _________________________________________ 84
5.2.9 ED-paint application on corrosion protection primers_________________________________ 86
5.3 Corrosion of corrosion protection primers___________________________________ 88
5.3.1 Corrosion of formed and unformed primers during standard corrosion testing______________ 88
5.3.2 Salt spray testing of formed primers ______________________________________________ 91
5.4 In-Situ Electrochemical Impedance Spectroscopy during stretch forming_________ 94
5.4.1 Forming characteristics of miniature stretching samples_______________________________ 94
5.4.2 In-Situ Electrochemical Impedance Analysis during tensile testing ______________________ 95
5.5 Scanning In-situ EIS / Raman investigations of formed corrosion protection primers
102
5.5.1 Scanning In-Situ EIS / Raman investigation of corrosion products on plane strain formed primers
102
5.6 Modification of corrosion protection primers _______________________________ 105
5.6.1 Organo silane surface modification of zinc particles implemented in corrosion protection primers
105
5.6.2 In-Situ Electrochemical Impedance Analysis during tensile testing of a modified primer ____ 108
5.7 Corrosion model of formed and unformed corrosion protection primers_________ 114
5.8 In-situ QCM / Raman investigations of the inhibitor adsorption on metals _______ 117
5.8.1 Ex-situ SERS analysis of Mercaptobenzothiazol (MBT) adsorption_____________________ 117
5.8.2 Combined in-situ Raman / QCM measurement of the adsorption of MBT on silver ________ 119
5.8.3 Combined in-situ Raman / QCM measurement of the adsorption of MBT on gold _________ 121
5.8.4 Theoretical calculation of the frequency shift by a monolayer _________________________ 124
5.9 Forming and water uptake of polyelectrolyte layers__________________________ 127
5.9.1 Molecular composition of polyelectrolytes ________________________________________ 127
5.9.2 FT-IR analysis of the curing state of polyelectrolyte layers ___________________________ 128
5.9.3 Hardness of polyelectrolyte layers_______________________________________________ 133
5.9.4 Water uptake of polyelectrolyte layers dried at different temperatures___________________ 135
5.9.5 FE-SEM investigation of formed polyelectrolyte layers ______________________________ 140
5.9.6 In-situ Electrochemical Impedance Spectroscopy during stretch forming of polyelectrolyte layers
142
5.9.6.1 Formability of a polyelectrolyte layer dried at room temperature __________________ 143
5.9.6.2 Formability of a polyelectrolyte layer dried at 130°C ___________________________ 145
5.9.6.3 Formability of a polyelectrolyte layer dried at 230°C ___________________________ 147
6 Overall conclusions ___________________________________________________ 150
7 Outlook_____________________________________________________________ 152
Index
10
8 Publications_________________________________________________________ 153
9 Literature___________________________________________________________ 154
Symbols
11
Symbols
A cross section, breaking elongation, area
B rotation constant
C capacitance
c concentration, speed of light
Cf quartz sensitivity factor
D spring constant
d thickness, piezoelectric tension module
E voltage, absorbance
Eloc electromagnetic field
e Euler constant
F force
F(J) rotation term
f frequency
h Planck constant, height
I current, intensity of light
Icorr corrosion current
J rotation quantum number
k absorption factor
L elongation, length, induction
m mass
N number of molecules
n harmonic number
P power of ion beam
R resistance
Re elastic limit of deformation
Rp polarisation resistance
Rp,0.2 0.2 % plastic deformation
Rm maximum strength during deformation
r radius
t time
w width
Symbols
12
x extension, length
Z impedance
∆m mass change
∆f frequency change
∆J quantum number of rotation
∆T temperature difference
∆v quantum number of vibration
Φ phase shift
α polarisability
βa anodic coefficient
βc cathodic coefficient
ε decade absorption coefficient
ε0 dielectric constant in vacuum
εr dielectric constant of a certain material
εel sum of plastic and elastic deformation
εus area of uniform deformation
εH2O dielectric constant of water
κ thermal conductivity
λ wave length
µ reduced mass, dipole moment
µq effective piezoelectric stiffened shear modulus
ν frequency
π Pi
ρ solution resistivity
ρq quartz density
φ logarithmic forming degree, water uptake
ω radial frequency
Motivation
13
1 Motivation
Coil coated steel is a composite material of a normally galvanised steel coated with one or
more organic layers. It is one of the premium products of the steel industry and gives large
benefits to the costumer like the effort reduction in their paint shops and further
improvements in corrosion resistance. Furthermore the coil coating facilities are equipped
with comprehensive waste water recycling and exhaust gas cleaning systems which allow a
much more environmentally friendly production than the coating of bulk goods.
The continuous production process is based first on an alkaline cleaning step to provide an
optimal surface of the galvanised steel coil. Afterwards the pretreatment and the organic
lacquer layers are applied by roll coaters in the thickness of some nm for the pretreatment and
some µm for the organic coating. The cross linking of the coating can be done by classic
thermal curing or in recent times by radiation curing where e.g. UV or IR radiation is used.
Finally the coated steel is coiled again and ready for the further production steps at the
customer facilities.
The flexible lacquer choice allows the production of optimised products for the different
demands which find a wide use in the household, building and automobile industry. Typical
household products that are made of coil coated steel are so called white goods (named
according to their colour) such as dish washers, refrigerators and washing machines. The
building industry uses pre-coated steel for the construction of dividing walls, cladding and
roof elements. In both industries, household and building, the colour appearance, corrosion,
UV and chemical resistance are the major demands.
The automotive industry uses coil coated material mainly in the form of corrosion protection
primers. These are thin pigmented, organic coatings (about 2 – 4 µm) which are applied on
galvanised steel and give a further improvement of the corrosion protection properties of the
substrate by the formation of a barrier layer. This allows the improvement of guaranteeing or
the reduction of secondary corrosion protection measures for the built cars. Next to the good
corrosion properties the coatings have to be weldable, glueable, must show good forming
features and must not disturb the further production steps during the car manufacturing
process, like phosphating and ED-paint application. Fig. 1 shows the typical production cycle
of coil coated material by the example of corrosion protection primers.
Motivation
14
Fig. 1. Production and application way of weldable corrosion protection primers for the automotive
industry
The formability of organic coatings on steel is a precondition for the application of pre-coated
(coil-coated) steel in various technical applications. Especially at high forming degrees or at
cut edges, defects appear within the organic coating. These defects influence the ensuing
processing technology as well as the long-term stability of the material.
The mechanical stress induced by the forming process leads to a reduction of the coating
thickness and the formation of cracks inside the organic matrix as well as at the
pigment/binder interfaces [1, 2]. This induces a loss of barrier properties [3, 4]. The loss of
corrosion protection of the layered material depends on the defect size, location, change in the
adhesion to the substrate and also on the surface modification steps applied subsequently.
It is therefore extremely important to correlate the forming degree with the change in the
barrier and structural properties of the coating.
Methods of investigating these barrier properties are common industrial tests like salt-spray or
cyclic corrosion tests, during which the lifetime of the coating is investigated in relation to the
forming degree [5]. These tests have the disadvantage of a long testing period and high rates
of systematic errors resulting from the visual evaluation by the operator.
Another well established way of monitoring the barrier properties of organic coatings in
contact with corrosive solutions is Electrochemical Impedance Spectroscopy (EIS) [6-28].
Recently Bastos et al. were able to show that ex-situ EIS is a powerful tool for correlating the
barrier properties of an organic coating with the local degree of forming [29].
Dynamic Electrochemical Impedance Spectroscopy was used by Darowicki et al. for the in-
situ monitoring of the passive layer cracking on 304L stainless steel and for evaluating the
performance of organic coatings under periodical stress [30, 31].
Motivation
15
The aim of the present work was to investigate and understand the defect formation and self-
healing properties of organic coatings like corrosion protection primers. Therefore, a new
electrochemical setup had to be established that allows the monitoring of electrochemical
impedance spectra with a common three electrode arrangement during the uniaxial forming of
a miniaturised model sample. This setup should allow a detailed correlation between the
forming degree and the barrier properties of the organic coating. The correlation is necessary
for predicting the maximum forming degree suitable for an organic coating during the
construction.
Furthermore, new setups for the combined spectroscopic and electrochemical analysis should
be developed which allow investigating the formation and characterisation of corrosion
products and the absorption kinetic of inhibitors used in organic coatings. This allows a fast
and comprehensive analysis of the efficiency and reactivity of new inhibitors and helps to
follow the self-healing potential of polymeric coatings.
Finally a new coating system the so called polyelectrolytes were investigated with respect to
their formability and curing state in order to understand the defect formation of lower cross
linked coatings which should be highly flexible. They are promising lacquer systems that
could provide huge benefits for corrosion protection as they can be specially tailored to the
different applications [32-46]. These flexible coatings should lead to a reduced defect
formation within the coating and thereby increase barrier properties.
The electrochemical measurements will be complemented by the Field Emission Scanning
Electron Microscopic (FE-SEM) and Focused Ion Beam (FIB) analyses to visualise the
location and size of the formed defects. The local reactivity should be correlated with the
formation of defects by means of a phosphating process and the common corrosion test.
Fundamentals
16
2 Fundamentals
2.1 Forming of polymer coated metals
2.1.1 Forming of metals
In the solid state metal atoms are regularly arranged and form a lattice of close-packed balls.
Within the material the atoms form small crystals (grains); these grains have slightly different
orientations to each other. The size of grains can vary from some micrometres up to some
millimetres. During the forming process the atoms move against each other within the lattice.
At grain boundaries between the single crystals the movement is hindered and requires a
much higher strength. Therefore, micro crystalline materials show a much higher maximum
strength.
The mechanical properties of metals can be evaluated by so-called elongation-strain diagrams.
The amount of elongation in the different directions in space is expressed by the forming
degree which is furthermore used to define the forming limit expressed in the forming
limiting curve (FLC) of the material under investigation [47, 48].
2.1.1.1 Stress-Strain diagram
Forming is the shape change of a material induced by an external force. Forming of metals
can be divided into a reversible, elastic and irreversible plastic deformation. The mechanical
properties of a material are shown in a Stress-Strain diagram. Fig. 2 shows the strain-
elongation diagram of a DC 06 ZE 75/75 steel sheet (thickness 0.8mm), a typical material
used for deep drawing. The diagram can be separated into the following parts:
• Hook’s straight, area of elastic tension. The material behaves like a spring and returns
to its original form according to Hook’s law
lDF ∆⋅= (eq. 1)
with F the force, D the spring constant and ∆l the elongation.
• The end of Hook’s straight is called elastic limit Re. At this point the plastic
deformations starts and the material will get a permanent shape change.
Fundamentals
17
• Most metals don’t show a defined elastic limit. Therefore, a limit at the 0.2 % plastic
deformation will be defined (Rp0.2). The overall elongation is thereby the sum of the
elastic elongation (εel) and the plastic deformation. At the point Rp0.2 the elastic
elongation can be calculated accordingly.
∆
ε
el =
ε
RP0.2 −0.2% (eq. 2)
• The maximum in the diagram is called tensile strength (Rm). Until Rm the sample will
be formed without constriction and the area is called uniform strain (εus).
• Beyond Rm the samples constrict and the further forming until breaking elongation (A)
only appears in that area.
Typical mechanical properties for different deep drawing steel grades are given in Tab. 1.
Tab. 1. Mechanical properties of typical deep drawing steel grades according to DIN EN 10130
Steel grade Rp0.2
/ MPa
Rm
/ MPa
A
/ %
DC 01 140 - 280 270 - 410 ≥ 28
DC 03 140 - 240 270 - 370 ≥ 34
DC 04 140 - 210 270 - 350 ≥ 38
DC 05 140 - 180 270 - 330 ≥ 40
DC 06 140 - 180 270 - 350 ≥ 38
Fundamentals
18
0 10203040
0
50
100
150
200
250
300
Stress / MPa
Strain / %
Fig. 2. Stress-Strain diagram of a tensile tested DC 06 ZE 75/75 sample (sheet thickness 0.8mm)
2.1.1.2 Elastic and plastic deformation of single crystals
The deformation of metals can happen in an elastic and inelastic way. During the elastic
deformation the induced stress leads according to Hook’s law to a linear translation (eq. 1). In
the case of a single crystal the induced stress causes deformation in the lattice. The
deformation is so small, that after removing the external force the atoms return to their origin
(Fig. 3).
Fig. 3. Elastic deformation of a single crystal lattice during application of external stress
(according to [48])
With increasing strain the elastic deformation leads to the plastic deformation of single
crystals which consist of two mechanisms, the formation of twins and the sliding.
The formation of twins appears along the twin plane, thereby a part of the lattice will be
transformed into a mirror image of the original one (Fig. 4).
Fundamentals
19
Fig. 4. Formation of twins in single crystals due to external stress (according to [48])
Sliding can be separated into the fixed sliding and the sliding by moving of dislocations.
Fixed sliding only appears if all atoms on both sides of the sliding plane are simultaneously
moved (Fig. 5). The length of one step is similar to the lattice constant (b).
Fig. 5. Fixed sliding along the sliding plane by the length of the lattice constant b (according to [48])
Technical materials differ from ideal crystals by the formation of lattice defects. The
reduction of bonds between the atoms leads to a decrease of the force which is necessary for
moveing the atoms along the sliding plane (Fig. 6).
Fig. 6. Defect migration during the formation of non-ideal crystals (according to [48])
Fundamentals
20
This explains the 100 – 1000 times smaller value found for the shear stress of materials
containing dislocations [48].
2.1.1.3 Forming of zinc and electrogalvanised zinc surfaces
Zinc crystallises into a hexagonal lattice structure with a height to edge ratio (c/a) of 1.856
[49]. Due to the complex and high anisotropic structure hexagonal systems show only 6 basal
slip systems [48]. These preferred sliding planes (in direction c) can be activated by a tension
of at least 0.3 MPa. For the activation of non-basal sliding (pyramidal sliding or twinning) a
shear stress which is 30 times higher has to be applied perpendicular or parallel to the basal
plane. During the uniaxial elongation a heterogeneous forming with sliding on the grains,
recrystallisation and the formation of twins occur [49].
The electrogalvanised zinc coating of steel shows the formation of parallel terraces with a
0001 orientation. Two cases of microstructure can be divided depending on the deposition
conditions. For hot electrolytes and low current densities the basal plane of the zinc lattice are
tilted between 35 – 55° to the steel substrate. In contrast to that for low temperatures and high
current densities the basal planes are mainly parallel to the substrate. Due to the crystal
structure and the ductility mainly basal plane slipping appears during plastic deformation.
Here the orientation of the sliding planes to the direction of strain is of great importance.
Grains with their basal planes in the direction of the strain can undergo a much higher local
forming than the global forming degree. The zinc coating shows no formation of defects
during the forming until the rupture of the steel substrate [50].
2.1.1.4 Forming degree
The shape change of a body by deformation is quantified by the deformation degree. Two
deformation degrees can be separated. First the logarithmic forming degree
ϕ
and second the
based forming degree
ε
which is normally written as a technical deformation in percent. In
this work only the logarithmic forming degree will be used.
In eq. 3 and 4 the derivation of the logarithmic deformation degree is shown. The change of
ϕ
is based on the actual dimensions of a body l and the length change dl. By solving the
integration of eq. 3
ϕ
is defined as the logarithm of the ratio before (l0) and after deformation
(l1).
Fundamentals
21
l
dl
d=
ϕ
(eq. 3)
0
1
ln
1
0l
l
l
dl
l
l
== ∫
ϕ
(eq. 4)
For the description of a forming process three forming degrees are defined (one for each
direction in space). The different degrees are marked by the index 1, 2 and 3. Index 1
indicates the major strain, this is the direction in which the highest dimension change
happens. In case of a simple tensile test it is the direction of the elongation. Index 2 displays
the minor strain which is the width change of the sample. The last deformation degree with
index 3 describes the thickness change of the sheet.
Using the logarithmic deformation degree leads to certain advantages. The sum of the
deformation degrees is independent of the numbers of deformation steps (eq.5). Due to the
fact that the volume of the formed body is constant the sum of all three forming degrees is
zero (eq. 6) [48].
ϕ
tot =
ϕ
i
i=1
n
∑ (eq. 5)
ϕ
1+
ϕ
2+
ϕ
3=0 (eq. 6)
2.1.1.5 Forming-limiting-diagram
The forming of metal sheets can be divided into four major groups: uniaxial, biaxial, plane
strain and deep drawing. These are separated by the different interconnections of the
deformation degrees.
The simplest way of forming is the uniaxial deformation. It appears during tensile testing.
Thereby the sample will be elongated in the direction of 1
ϕ
while the width 2
ϕ
decreases. In
eq. 9 the interaction between the major and the minor strain is given. Uniaxial forming often
appears during the production of real automotive parts.
21 2
ϕ
ϕ
−= (eq. 7)
Fundamentals
22
Biaxial forming is also known as stretch-forming. The sample will not only be elongated in
one direction as known from the uniaxial forming but in two directions 1
ϕ
and 2
ϕ
(eq. 8).
Biaxial forming leads to high stresses in the material and is unusual for most automotive
parts. This way of deformation is used for material investigations and needs especially shaped
samples like Marciniak or Nakajima specimen where the material is hindered by the blank
holder from flowing into the form [48].
21
ϕ
ϕ
= (eq. 8)
Plane strain deformation is often found in real automotive parts. The forming only appears in
direction of the major strain 1
ϕ
and by a change in the sheet thickness 3
ϕ
. The minor strain
2
ϕ
remains constant during the whole process (eq. 9). In this forming process the material
flow is similar to the biaxial forming which is hindered by the blank holder.
ϕ
2=0 (eq. 9)
During deep drawing the metal sheet will be formed by a stamp similar to the biaxial and
plane strain forming. The difference is that the blank holder force is negligible or adjusted to
such small forces that the material flow is not hindered. Thereby the width 2
ϕ
will be reduced
by the same amount as the length 1
ϕ
will be increased (eq. 10). The thickness of the sheet
will be constant during the forming procedure.
ϕ
1=−
ϕ
2 (eq. 10)
The whole forming process is described by three deformation degrees and the connection
between them is completely different for the four major forming procedures. This makes it
extremely difficult to compare different deformations just by a look at the deformation
degrees. Therefore von Mises defined the equivalent strain (eq. 11) which allows an easy
comparison of the forming processes [47, 48].
ϕ
v=2
3
ϕ
1
2+
ϕ
2
2+
ϕ
3
2
()
(eq. 11)
Fundamentals
23
For the forming process it is of great interest to get detailed information about the maximum
strain that can be applied to the material before it breaks. In the forming limiting diagram the
major strain 1
ϕ
is plotted versus the minor strain 2
ϕ
and the forming limiting curve shows the
maximum strain before rupture (Fig. 7).
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
forming limit
deep drawing
ϕ1 = −ϕ2
biaxial
ϕ1 = ϕ2
plane strain
ϕ2 = 0
Major forming degree ϕ1
Minor forming degree ϕ2
uniaxial
ϕ1 = -2ϕ2
Fig. 7. Forming limiting diagram indicating the maximum forming degrees possible for the material
(black line) and the major forming operations
During the production of formed parts it is important to be 10 % below this curve to guarantee
the necessary process reliability.
Fundamentals
24
2.1.2 Forming of polymers and thin polymer films
2.1.2.1 Molecular understanding
In contrast to metals that consist of single atoms which form a crystalline grid, polymers
consist of long molecule chains. Depending on the chemistry and the production parameters
of the polymer these chains can also form partly crystalline areas. Molecules with only few
branches have a high tendency to form crystalline areas while strongly cross-linked and
branched molecules are mainly amorphous. Furthermore the cooling speed after the
production influences the arrangement of the molecules. Low cooling rates strongly increase
the number of crystalline areas in the polymer. The different crystalline structures of polymers
show also different behaviours during forming. On the one hand they can be brittle and break
already at small elongations on the other hand they behave elastically and can be highly
elongated. The forming behaviour is strongly dependent on temperature and can change from
brittle at low degrees to elastic at higher degrees [51].
Brittle polymers show just a linear increase in the strain until they break. Partly crystalline
polymers show a high stretch ability before rupture due to the intricate structure of the
molecules. During elongation the molecule fibres orient themselves into the direction of the
applied force and reach a stress maximum (Fig. 8).
Fig. 8. Stress-strain diagram of a polymer under tensile stress showing the different areas of elastic and
plastic deformation until rupture (according to [51])
Fundamentals
25
With further elongation the sample starts to constrict and the stress decreases until it reaches a
plateau. Further elongation doesn’t lead to an increase in the stress over a large strain range.
In this region the crystalline parts break up and the molecule fibres become parallel. When all
fibres are arranged in parallel the strain strongly increases again until the polymer ruptures.
Elastomeres don’t show a specific maximum and only a slightly defined plateau. The strain
increases nearly continuously until the rupture.
2.1.2.2 Relevance of interfaces
Beside the formability of the polymer itself the interface between the coating and the metallic
substrate is of great importance. The interface is responsible for the adhesion of the coating to
the substrate. Weak adhesion leads to detachment or the delamination of the organic layer and
thereby to a loss of the barrier properties which accelerate corrosion. Typical industrial tests
for analysing the adhesion after forming are the Erichsen cupping and the ball impact [52].
Both lead to a biaxial forming of the specimen and an adhesive tape is used for testing the
adhesion.
A lot of work was done to stabilise the coating / substrate interface. One of the best-known
and widely used processes is phosphating [53]. Hereby the etching attack of the metallic
surface leads to a precipitation layer of phosphate crystals with a thickness between 1 and 50
µm. The micro rough surface structure of the phosphate layer allows a good mechanical
interlocking with the following coating and thereby increases the adhesion. The phosphate
crystals also act as a further barrier and thereby improve the corrosion resistance of the
system. Commonly this effect is increased by a passivation step after phosphating. Recently it
was changed from Cr (VI) passivation to Chromium free passivation.
In the case of coil coatings the pretreatment consists of ultra thin layers in the range of some
nm, which allows a Chromium free passivation of the substrate and increases the paint
adhesion [54-58] (Fig. 9).
Fundamentals
26
Fig. 9. Schematical illustration of the pretreatment (Granodine 1456) layer formation on a zinc surface
including an organo silane adhesion promoter
For the surface passivation mainly salts of Manganese, Titanium and Zirconium in
combination with complexing agents and phosphoric acid are used. Together they form with
the dissolved zinc ions a protective precipitation layer on the zinc surface. The adhesion
promotion is achieved by chemical bonding of bi-functional organo silanes between the
organic coating and the zinc substrate.
2.1.2.3 Theory of crack formation in polymers and at polymer / oxide
interfaces
The defect formation within polymers or at the polymer / oxide interface strongly depends on
the adhesion along the interface. In the case of a weak adhesion the coating delaminates from
the substrate and an adhesive failure is formed. If the adhesion between the coating and the
substrate is good the stress inside the coating can reach high values until the polymer cracks
by cohesive failure. If the load is applied uniaxially to the system the cohesive defects will
grow perpendicular to the stress direction and new cracks appear mainly in the middle of
existing fragments [59].
In the case of welded polymers the crack formation mechanism depends strongly on the
interface width between the polymers. For most polymers three regimes were found which
can be attributed to different failure mechanisms. For small interface widths the interdiffusion
Fundamentals
27
of polymer chains is only very weak and they are pulled out of the opposite polymer. With
increasing interface widths the separation is done by pull out and by chain scission of the
polymers. For high values of interdiffusion the main mechanism is the chain scission and the
force reaches the value of the bulk adhesion [60].
Moloney et. al. show that in front of the crack tip propagating within a highly crosslinked,
unfilled epoxy resin a zone of micro cracks is formed [61]. The filling of polymers can lead to
a toughening of brittle polymers. Thereby the particles initiate a high number of very small
local yield events. Examples are inorganic particles which lead to a large number of
microcracks and a reduction of the local stress. Another way is the implementation of ductile
particles which are plastically formed by the crack tip and help bridging the two sides of the
crack [62].
The propagation of a crack can be stopped if the crack hits e.g. a local volume of higher
strength, a weak second phase or another microcrack and the propagation into an unloaded
region between other cracks or strong fibres [62].
2.1.3 Repair mechanisms of defects in protective coatings
The repair mechanisms of defective coatings can be divided into two major groups. On the
one hand are the traditional ways of repairing the defects by e.g. welding and patching, on the
other hand are the recently developed self-repair systems that heal the defects without
external support [63-73]. The traditional repair processes are mainly used for macroscopic
defects while the self-repair focuses on the healing of microscopic defects as they might occur
during forming [74].
Self-healing can appear if the two sides of a defect are still in contact with each other and the
system is above its glass temperature. Than the crack heals due to molecular diffusion across
the interface [75]. Other mechanisms are the recombination of chain ends by the catalytic
redox reaction under oxygen atmosphere, the use of living polymers with radical functions at
the chain end and the inclusion of reservoirs that contain a healing agent [75].
Two different approaches were developed as reservoirs, the hollow fibre and the
microencapsulation. In the first case the glass fibres ranging from of 15 to 60 µm in diameter
were filled with a healing agent and embedded into the coating. If a defect hits such a fibre it
breaks and the crack will be filled with the healing monomer that cures within. The
microencapsulation works similar to the hollow fibre but instead of a fibre micro capsules are
used which are embedded together with a catalyst into the coating. The propagating crack
Fundamentals
28
front ruptures the microcapsules and the released healing agent reacts with the embedded
catalyst and seals the defect [63, 74, 76-78].
A further recently described method for healing defects in protective coatings is the use of
expandable materials like phyllosilicates (clays) [79]. If they are exposed to air inside a crack
these materials will absorb moisture and implement the water into their crystal structure. This
leads to a lattice growth and allows the defect to be filled.
Another approach is the intelligent and controlled release of inhibiting agents into defects to
passivate these corrosion sites. Shchukin et. al. describe the pH controlled release of
benzotriazol from nano particles and tubes due to the response of polyelectrolytes to a pH
shift. At low or high pH, as they appear at the cathodic or anodic corrosion sites, the
polyelectrolytes become permeable for the inhibitor. The release rate can be adjusted by the
properties of the used polyelectrolytes [80]. Paliwoda-Porebska et. al. show that the inhibitor
[PMo12O40]3- can be potentially released from conductive polymers (polypyrrole) [71] and
leads to a reduction of the cathodic delamination of coatings. Thereby the reduction of the
polymer film due to the corrosive reaction leads to the release of the inhibitor.
Fundamentals
29
3 Fundamentals of applied microscopic, spectroscopic and
electrochemical methods
3.1 Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy (EIS) is widely used in material and surface science
to investigate the properties of e.g. metallic substrates, batteries and fuel cells [81-95]. A
special focus is thereby set on the characteristics of polymer coatings. One of the key features
is the barrier properties of coatings over time which are in contact with a corrosive electrolyte
on order to gain information about the corrosion protection. Mansfeld describe in detail the
used equivalent circuit and the interpretation of data [96]. Further focus was laid on the
evaluation of the water uptake of organic coatings [19, 97-102]. In recent times EIS was used
to investigate ultra thin conversion coatings [55, 58] and zinc rich coatings [103-105]. The
combination of EIS with spectroscopic tools like FTIR to observe the water uptake was also
developed by different groups [106-108].
3.1.1 General description of the impedance
The behaviour of an ideal resistor which hinders the current flow in a circuit is described by
Ohm’s law (eq. 12). The resistor R follows the law at all currents I and voltages E and its
value is independent of the applied frequency. The current answer to an applied AC voltage is
always in phase [109-111].
I
R
E
⋅= (eq. 12)
The description of Ohm’s law can only be used in the case of the ideal resistor. A much
broader approach to handle the resistance of a circuit is the impedance which is not limited to
the simplifications mentioned above. The electrochemical impedance is measured by applying
a sinusoidal potential excitation to an electrochemical cell. The current answer will be
analysed and the impedance can be calculated.
The frequency f in Hz used for the voltage modulation can also be written by the radial
frequency ω in radians / second (eq. 13).
f⋅=
π
ω
2 (eq. 13)
Fundamentals
30
The applied excitation voltage Et has as a sinusoidal function of time and can be described by
eq. 14.
)sin(
0tEEt
ω
= (eq. 14)
The response of the current I0 will be shifted in phase by a certain angle
φ
(eq. 15).
)sin(
0
φ
ω
+= tIIt (eq. 15)
The dependence between the applied voltage and the phase shifted current answer is shown in
Fig. 10.
Fig. 10. Phase shift between a sinusoidal excitation voltage and the current response (taken from [112])
The impedance can be described similar to Ohm’s law with the time dependent voltage and
current. The magnitude of the impedance is expressed by Z (eq. 16).
)sin(
)sin(
)sin(
)sin(
)(
0
0
φω
ω
φω
ω
ω
+
=
+
== t
t
Z
tI
tE
I
E
Z
t
t (eq. 16)
The Euler relation allows the transformation of the voltage and current into a complex
function (eq. 17-20). It can be described by a real and an imaginary part.
)(
0
ti
teEE
ω
⋅= (eq. 17)
Fundamentals
31
)(
0
φω
+
⋅= ti
teII (eq. 18)
)sin(cos)( )(
φφω
φ
iZeZZ i+=⋅= (eq. 19)
ZiZZ ImRe)( +=
ω
(eq. 20)
The real part Re and the imaginary part Im are given by eq. 21 and 22.
φ
cos)Re( ZZ = (eq. 21)
φ
sin)Im( iZZ = (eq. 22)
3.1.2 Plotting of impedance data
Two ways of plotting the impedance data are common in electrochemistry. The Nyquist plot
displays the imaginary part on the y-axis and the real part on the x-axis (Fig. 11). The plot is
based on a simple vector plot; every impedance data on the semicircle can be displayed by a
vector. The length of the vector is the impedance Z while the angle between the x-axis and
the vector is the phase shift
φ
. The major disadvantage of the Nyquist plot is that the
frequency at any data point can not be taken from the diagram.
-Im
Z
Re
Z
|Z|
ω=0
φ
Fig. 11. Schematical Nyquist plot of a metal electrode indicating the phase shift (
φ
) and the impedance
shown as vector ( Z)
Another way of presenting the impedance data is the Bode plot (Fig. 12).
Fundamentals
32
0.01 0.1 1 10 100 1000 10000 100000
10
100
1000
10000
-80
-60
-40
-20
0
|Z| / Ω
f / Hz
Phase / °
Fig. 12. Schematical Bode plot of metallic substrate showing the impedance (black) and
the phase shift (red)
The logarithm of the impedance Z is plotted on the y-axis while the logarithm of the
frequency is plotted on the x-axis. The phase shift
φ
is thereby plotted on a second y-axis.
The Bode plot allows an easy interpretation of the electrolyte resistance (at high frequencies)
and the polarization resistance (low frequencies) in simple equivalent circuits.
3.1.3 Pseudo-Linearity of the system
The measurement of impedance spectra can take hours depending on the selected frequency
range. During the whole measurement it is necessary that the system is in a steady state and
will not change with time. Changes in the system can be caused by temperature changes,
adsorption of molecules, degradation of the coating and so on. These effects might lead to
misinterpretations of the results.
A further quite important point is the linearity of the system under investigation. In the case of
an ideal ohmic resistor the current increases linearly with the applied voltage. For an
electrochemical cell this linearity is not strictly given (Fig. 13). Therefore it is necessary to
create a pseudo linear system [112].
Fundamentals
33
Fig. 13. Pseudo linearity of an electro chemical system if a small excitation voltage is used
(taken from [112])
If the excitation of the system is very small (1-10 mV) the system can be supposed as pseudo
linear.
3.1.4 Elements of the equivalent circuit
For the interpretation of impedance data commonly known electrical elements are taken and
an equivalent circuit will be modelled. Each element stands for a certain electrochemical
property. The most common properties are described in Tab. 2 [112, 113]:
Tab. 2. Common electrical element
Component Current vs. Voltage Impedance
Resistor R IRU ⋅=
R
Z
=
Capacitor C dt
dU
CI = Ci
Z
ω
1
=
Inductor L dt
dI
LU = LiZ
ω
=
3.1.4.1 Electrolyte resistance
Even by the use of Haber-Luggin-Capillaries, which allow the positioning of the reference
electrode directly in front of the working electrode, the resistance of the electrolyte between
the reference and the working electrode cannot be completely eliminated. This impedance is
found in the spectra as an ideal ohmic resistor which dominates at high frequencies and
Fundamentals
34
ranges from 10 - 100 2
cm⋅Ω for common setups. The uncompensated electrolyte resistance Ru
is defined as
A
l
Ru
ρ
= (eq. 23)
With
ρ
the solution resistivity, A the area and l the length.
3.1.4.2 Double layer capacitance
If an electrode is in contact with an electrolyte, ions from the solution can stick to the
electrode surface by van der Waals or Coulomb forces. The charges in the electrolyte and at
the electrode boundary are of opposite sign and form a capacitor. Due to the small distance of
just a few Ångström between the charge carriers extremely high capacitances of 10 to 40
µF/cm² can appear.
3.1.4.3 Polarisation Resistance
Whenever an electrode is polarised a current flow at the electrode surface can occur due to an
electrochemical reaction. This leads to either an anodic or cathodic current whose density is
controlled by the kinetics of reaction and diffusion of the reactants. In case of a metal
electrode that corrodes at the open circuit potential two electrochemical (a cathodic and an
anodic) reactions appear at the electrode surface. At the open circuit potential the current of
both reactions are equal and it is known as the corrosion current. The current of a kinetically
controlled reaction can be described by the Butler-Volmer equation. In the case of a mixed
electrode it can be written as eq. 24.
() ()
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
−=
−−−
c
EE
a
UU
corr
OCOC
eeII
ββ
303.2303.2
(eq. 24)
With I the cell current, Icorr the corrosion current, Eoc the open circuit potential, a
β
the anodic
coefficient and c
β
the cathodic coefficient.
If only a small potential signal is applied to the cell the eq. 24 can be simplified to eq. 25 with
Rp the polarisation resistance.
Fundamentals
35
()
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
⋅
+
=
Pca
ba
corr R
I1
303.2
ββ
ββ
(eq. 25)
3.1.4.4 Coating capacitance
The capacitance of a coating can be described by a plate condenser. The condenser consists of
two plates formed by the surface or the electrode and the electrolyte. Between these two
plates the coating exists as a dielectric. The capacity C of the condenser is proportional to the
plate area A, the dielectric constant of the vacuum ε0, the dielectric constant of the material
between the plates εr and inverse proportional to the thickness of the dielectric d (eq.26).
d
A
Cr⋅⋅
=
ε
ε
0 (eq. 26)
3.1.5 Modelling of equivalent circuits
For the interpretation of impedance data so-called equivalent circuits are built from simple
electrical elements like resistor, capacitor and inductor. The equivalent circuits reflect the
physical properties of the electrode under investigation. Some of the most common circuits
are shown in the following chapters [8, 13, 28, 111-118].
3.1.5.1 Ideal coating
A metal electrode covered by an ideal coating without any defects can be described by a pure
capacity Cc. In contact with an electrolyte an ohmic resistor for the electrolyte resistance Ru
must be added in series to the capacitor (Fig. 14).
R
u
C
C
Fig. 14. Equivalent circuit for an ideal coating in contact with an electrolyte
Fundamentals
36
At high frequencies the impedance of the capacitor can be neglected and the cell impedance is
just dominated by the ohmic resistance of the resistor (parallel to the x-axis, phase shift 0)
(Fig. 15).
0.01 0.1 1 10 100 1000 10000 100000
10
100
1000
10000
100000
1000000
1E7
1E8
1E9
-80
-60
-40
-20
0
|Z| / Ω
f / Hz
Phase / °
Fig. 15. Bode plot of an ideal coating with a capacity of 10 nF and an uncompensated electrolyte resistance
of 100 Ohm
In the medium and low frequency the capacitor dominates and the impedance of the cell
increases linearly with a slope of -1 in the bode plot. The phase shift of an ideal coating is
90°.
3.1.5.2 Metal electrode
Metal electrodes are described by a parallel circuit of an ohmic resistor that reflects the
polarisation resistance RCt and a capacitor which describes the double layer capacity CDl of
the electrode. Furthermore, the uncompensated electrolyte resistance Ru is added in row.
R
u
R
C
t
C
Dl
Fig. 16. Equivalent circuit of a metal electrode in contact with an electrolyte
In the bode plot the domination of the electrolyte resistance is shown as a parallel to the x-
axis at high frequencies (Fig. 17).
Fundamentals
37
0.01 0.1 1 10 100 1000 10000 100000
10
100
1000
10000
-80
-60
-40
-20
0
|Z| / Ω
f / Hz
Phase / °
Fig. 17. Bode of a metal electrode with CDl = 30 µF, RP = 5000 Ohm and Ru = 100 Ohm
At the medium range the double layer capacitance shows a straight line with a slope of -1
while at low frequencies a series connection of the electrolyte and the polarisation resistance
prevails.
3.1.5.3 Non-ideal coating
For a defective coating the equivalent circuit for an ideal coating can no longer be taken.
Defects occur by pores within the coating and allow the electrolyte to penetrate and reach the
surface of the metal electrode. The penetrating electrolyte is reflected similar to the
uncompensated electrolyte resistance Ru by an ohmic resistor RC that is parallel to the coating
capacitance Cc. If the electrolyte reaches the metal surface it will react like an electrode in
contact with an electrolyte and can be described by a normal metal electrode circuit with
double layer capacitance CDl and polarisation resistance RCt. This sub-circuit is in series with
the pore resistance and parallel to the coating capacitance (Fig. 18).
Ru
RC
RCt
CDl
CC
Fig. 18. Equivalent circuit of a non-ideal coating in contact with an electrolyte
Fundamentals
38
The bode plot shows two time constants at high and medium frequencies displaying the
coating and the double layer capacitance (Fig. 19).
0.01 0.1 1 10 100 1000 10000 100000
10
100
1000
10000
100000
-80
-60
-40
-20
0
|Z| / Ω
f / Hz
Phase / °
Fig. 19. Bode plot of non-ideal coating with CC = 10 nF, RPore = 10000 Ohm, CDl = 30 µF, RP = 5000 Ohm
and Ru = 100 Ohm
The two time constants are separated by the ohmic behaviour of the pore resistance while at
low frequencies a series circuit of the polarisation, the pore and the electrolyte resistance
dominates.
Fundamentals
39
3.2 Quartz crystal mircobalance
The quartz crystal microbalance is a sensor that allows the detection of mass changes within a
few ng/cm2. It has been used for over 40 years for measuring the deposition of materials
within gas and liquid phases.
The system is based on the piezoelectric effect found by Pierre and Jacques Curie [109]. If a
piezoelectric material is deformed by an external force the charge carrier will be dislocated
and the material will be polarised which leads to an electrical charge at the surface. A dipole
moment along the polar axis appears which is proportional to the applied force. For the QCM
the inverse piezoelectric effect is used. In that case an external voltage will deform the
material. For small amplitudes the extension is proportional to the applied voltage (eq. 27).
x
=dE (eq. 27)
With x the extension, d the piezoelectric tension module and E the applied voltage
The quartz used for the QCM setup is so-called AT-cut quartz [119, 120]. The advantage of
these special cuts is the small temperature dependence of the resonance frequency in the range
from 0 – 50 °C. Metal electrodes applied by physical vapour deposition cover the parallel
surfaces of the quartz. These electrodes are used for applying the external AC voltage and for
the detection of the quartz vibration.
During the measurement the quartz is driven at its resonance frequency. At this frequency a
stationary wave orthogonal to the electrode surface appears. The condition for the wave
length
λ
is given by the thickness of the quartz d (eq. 28). For n = 1 the quartz is driven at its
fundamental mode while higher numbers lead to overtones which don’t have the same
frequency stability.
λ
=2d
n (eq. 28)
The first to describe the principle of the QCM was Sauerbrey in 1959 [121]. The resonance
frequency of an oscillating quartz is proportional to its mass. Sauerbrey described the change
in resonance frequency by the deposition of a film on the quartz surface. He pointed out that
the frequency change ∆
f
is directly proportional to the applied mass change ∆m (eq. 29).
mCf f∆⋅−=∆ (eq. 29)
Fundamentals
40
Cf is the quartz sensitivity factor which can be calculated from the quartz density
ρ
q, the
number of harmonic n, the resonance frequency f and the effective piezoelectric stiffened
shear modulus
µ
q according to eq. 30.
qq
f
fn
C
µρ
⋅
⋅
=
2
2 (eq. 30)
The mass change can be calculated after eq. 31 if eq. 30 will be inserted into eq. 29.
()
2
2fn
ff
C
f
mqqq
f⋅
⋅−
=
∆−
=∆
µρ
(eq. 31)
Van Dyke and Butterworth showed that the oscillating quartz can be described by a simple
equivalent circuit (Fig. 20) [122]. The setup is a parallel coupling of a capacitor and a
vibration branch. R, C and L describe the ideal resonance of the pure quartz. Here R
represents the energy lost by friction, L represents the mass load (inductance) and C the
mechanical elasticity of the material (capacitance). Cs represents the capacity of the two gold
electrodes and the quartz as dielectric in between.
R
C
C
S
L
Fig. 20. Van Dyke and Butterworth equivalent circuit of an oscillating quartz crystal
The impedance Z of L and C is given by
C
LZZZ CL
ω
ω
1
+=+= (eq. 32)
At the resonance frequency both impedances compensate each other (eq. 33). Therefore the
resonance frequency can be calculated according to eq. 34.
Fundamentals
41
CL ZZ = (eq. 33)
f0=1
2⋅
π
⋅L⋅
C
(eq. 34)
At frequencies below the resonance frequency the capacitive part of the circuit dominates and
the phase shift of the current is positive. At higher frequencies the inductance dominates and
the phase shift becomes negative.
The mission of the QCM setup is to keep the quartz at its resonance frequency while the load
changes and to transfer the measured frequency to a computer for storage. This is done by the
so-called Phase Lock Oscillator (PLO) which uses an internal voltage controlled oscillator to
drive the crystal. The PLO also includes a phase detector continuously monitoring the phase
differences between the crystal’s current and voltage. The output of the phase detector is
connected to an integrator. The integrator summarizes the phase error a positive phase shift
causes the output to climb and a negative phase shift to fall. If no phase error occurs the
integrator output holds steady. The integrator output is connected to the voltage controlled
oscillator and changes the frequency of the oscillator until the phase error becomes zero at the
resonance frequency.
Fundamentals
42
3.3 Vibration spectroscopy for thin film analysis
In recent times the classical spectroscopic methods like Raman and IR have often been used
for in-situ experiments. Especially Surface Enhanced Raman Spectroscopy (SERS) came into
focus as it allows to investigate even mono layers on the SERS active substrate. Furthermore,
Raman and IR spectroscopy are combined with other investigation methods to gain a more
comprehensive knowledge of the system under investigation. A widely used setup is ATR
FT-IR which allows an easy combination with liquid or electrochemical cells.
Different authors used ATR to study the diffusion of water through polymer film [123, 124].
Wapner et. al. used it to investigate the water transport along an adhesive / metal interface
[125]. The combined Impedance and FT-IR spectroscopic analysis of the water uptake were
carried out by Ohman et. al. on a metal / polymer while Vlasak et. al. used it on a
semiconductor / polymer interface [106-108].
In the past SERS was used to study the adsorption of organic molecules on coin metals and
the inhibition effect of these layers [126-130]. Recently Gu et. al. showed that SERS can also
be extended for the use with zinc surfaces [131]. Furthermore, the combination with
electrochemical methods for comprehensive in-situ investigations came into focus analysing
the potential dependence of the adsorbed films [132, 133].
3.3.1 DRIFT and FTIR-ATR spectroscopy
3.3.1.1 Fundamentals
IR spectroscopy allows the detection of rotation and vibration transitions of molecules by the
adsorption of light within the range of 4000 to 400 cm-1. Only transitions of molecules with a
permanent dipole are IR active or those where the dipole moment changes during the
transition. The periodic change of a dipole moment can only happen at discrete frequencies.
The adsorption intensity depends on the dipole moment change and on the orientation of the
dipole to the light vector. Maximum adsorption appears with maximum dipole change and if
the dipole and the light vector are parallel.
Every molecule has about 3N degrees of freedom with N the number of atoms. These degrees
are separated into vibration, rotation and translation. Every molecule has 3 translation degrees
one for each direction in space independent from the shape. Linear molecules have 3N-5
vibrations while non-linear molecules have 3N-6. The remaining degrees of freedom belong
to the rotations of the molecule [134-136].
Fundamentals
43
If light interacts with material the intensity will be weakened by the absorption, reflexion and
scattering. The loss of intensity –dI is proportional to the intensity of the beam I, the
absorption factor k and the length dx and the concentration c of the illuminated material
[137].
dxcIkdI ⋅⋅⋅=− (eq. 35)
The vibrations of a molecule can be divided into two classes. First the valance vibration
ν
which leads to deformations along the bonding axis of the atoms. On the one hand this
vibration can be symmetric and on the other hand asymmetric. The second class is the
deformation vibration
δ
which can be separated in the rocking, twisting and waging vibration.
Further details concerning IR spectroscopy can be found in the general literature [135, 136].
3.3.1.2 DRIFT
During classic infrared reflection absorption spectroscopy the infrared beam hits the flat
sample in a certain angle of incidence and is reflected with the same angle. In the case of
rough surfaces or powders the light is scattered in all directions which means that it can also
be reflected a few times inside the sample before it returns to the surface. The angle of
reflection at which the light leaves the sample is hardly the same as the angle of incidence.
Large spherical mirrors are therefore necessary for collecting the light from the sample and
focusing it to the inlet of the detector. The high number of reflections that occur before the
light hits the detector leads to a loss in the signal intensity. For a standard analysis the sample
needs no special preparation but has to be comparable for the quantitative analysis of the
roughness or the particle size of the samples. Most often the semi-quantitative Kubelka-Munk
equation is used for analysing the spectrum [138].
3.3.1.3 FTIR-ATR spectroscopy
Attenuated Total Reflection (ATR) FTIR spectroscopy is a versatile tool to investigate
gaseous, liquid and solid specimen without special preparation steps. The ATR setup consists
of a crystal (often ZnSe, Ge, Si or diamond) which is pressed against the sample. If the
medium in contact with the crystal has a lower refractive index the incident light will be
totally reflected under certain angles. Depending on the angle of incident and the size of the
crystal the light can be more than one time reflected which improves the signal from the
Fundamentals
44
sample. Behind the reflective interface an evanescent wave penetrates the sample and
interacts with it, leading to a frequency dependent absorption as it is known from transmission
spectroscopy. The penetration depth depends on the angle of incident and the wave length of
the used light. In a rough estimation the penetration depth d can be seen as in the order of the
incident light but can be calculated in detail with eq. 36.
2
2
22
1sin2 nn
d
−Θ⋅⋅
=
π
λ
(eq. 36)
With λ the wave length, n1 the refractive index of the ATR crystal, n2 the refractive index of
the material under investigation and Θ the angle of incidence [138-140].
3.3.2 Raman spectroscopy
For the detection of an IR spectrum a change of the dipole moment during the adsorption of
light is necessary. In the case of a Raman spectrum the sample is irradiated with
monochromatic light and the scattered light is detected. The frequency of the light can vary in
wide ranges but it mustn’t be adsorbed by the sample. The demand for the Raman Effect is
the polarisability of the molecule. This means the repositioning of the electron hull. The
influence of the electromagnetic field Eloc induces a dipole moment µ to a molecule with the
polarisability α (eq. 45) [134-136, 141].
loc
E⋅=
α
µ
(eq. 37)
Light with the frequency υ0 causes a periodic change of the dipole moment and the molecule
emits light with the same frequency (Rayleigh scattering) (eq. 38).
tE 00 2sin
π
ν
α
µ
⋅= (eq. 38)
Rotation and vibration of the molecule in the electromagnetic field cause a periodic change of
the polarisability.
For vibration: tvvvv
π
α
α
α
2sin
10 += (eq. 39)
For rotation: tvrrr 22sin
10
π
α
α
α
+= (eq. 40)
Fundamentals
45
with α0 the average and α1 the maximum polarisability, υv the vibration frequency and υr the
rotation frequency.
Combining the equation for vibration and rotation with equation 38 leads to the expression for
the vibration and rotation transition.
[]
tvvtvvEtvE vvvovi )(2cos)(2cos
2
1
2sin 000100 +−−+=
ππαπαµ
(eq. 41)
[]
tvvtvvEtvE rrrori )2(2cos)2(2cos
2
1
2sin 000100 +−−+=
ππαπαµ
(eq. 42)
From equation 41 and 42 the quantum number for the vibration and rotation follows:
1±=∆v
2;0 ±=∆J
3.3.2.1 Surface Enhanced Raman Spectroscopy (SERS)
The Raman scattering width of a molecule is normally very small so that either a high number
of molecules or a high laser power is necessary to get an adequate signal. In 1974
Fleischmann et al. [142] found an increased signal during their investigations of pyridine
adsorped on a rough silver sample. They attributed this effect to the increase of the active
surface by the roughening. Later on other researchers realised that the significant signal
increase cannot only result from the larger sample surface.
The so-called Surface Enhanced Raman Spectroscopy (SERS) is based on the effect that the
signal of molecules in contact or close to SERS active material undergoes an enhancement by
up to 8 orders of magnitude. SERS active substrates are certain metal substrates (e.g. gold and
silver) which show a nano rough or island structure with a size between 5 and 100 nm. Mainly
two models which are also complementary to each other are used to explain this effect, the
electromagnetic enhancement and the chemical enhancement [143-145].
During the electromagnetic enhancement the molecule comes close to the hemispheric nano
structure of the metallic substrate. The molecule thereby influences the electric field of
incident light and the induced dipole field of the sphere. Both fields add each other and lead to
a magnification of the signal by up to 8 orders of magnitude.
Fundamentals
46
For the chemical enhancement the molecule needs to have direct contact with the metal
sphere. The incident light can create an excited electron and a corresponding hole. If the
Fermi level is in between the HOMO and the LUMO of the adsorbed molecule the electron
can transfer into the LUMO of the molecule, resulting in an exited charge transfer state. The
electron-hole recombination leads to the emission of light of a certain frequency. This effect
should lead to a lower magnification of the signal than the electromagnetic enhancement but
this has not yet been clarified.
Fundamentals
47
3.4 Scanning electron microscopy
The scanning electron microscope (SEM) allows a detailed surface analysis of condensed
materials. During the analysis a focused electron beam that is formed by the interaction with
the surface secondary and back-scattered electrons scan the specimen. These electrons are
detected and used for imaging. Two main classes of SEMs exist which are separated by the
electron formation. The conventional SEM uses a hot cathode for the electron generation
while the latest Field Emission devices use either cold or Schottky Field Emission [146, 147].
After the generation the electrons are accelerated with up to 50 kV through the vacuum
column of the SEM where they are focused by electro-magnetic lenses.
Depending on the acceleration voltage ∆E the theoretical resolution can be calculated with the
de-Broglie wavelength of the electron λ (eq. 51).
λ
=h
m⋅v=h
2me⋅e⋅∆E (eq. 51)
with λ the de-Broglie wavelength, h the Planck constant, m the mass, v the speed, me the
electron mass, e the electron charge and ∆E the acceleration voltage.
The theoretical resolution for electrons accelerated with 1 kV is 38 pm. Due to the fact that
the energy source is not infinitely small and by lens errors the real resolution is about 10 nm.
When the electrons hit the sample surface they can interact elastically or inelastically with the
material. The interaction area and depth depends on the acceleration voltage and the atomic
number of the sample material. From the sample surface emitted electrons are separated by
their energy into secondary electrons (below 50 eV) and back-scattered electrons (50 eV up to
the acceleration voltage). Secondary electrons are formed by the inelastic interaction
(ionisation) of surface atoms. Further reactions of the electron beam with the sample are the
formation of Auger-electrons and X-Rays, both are not used for the standard SEM imaging.
Either a chamber or an in-lens detector can perform the detection of secondary electrons.
Chamber detectors mainly consist of an Everhart-Thornley detector that uses a grid with an
applied voltage between -200 V to +200 V to collect the electrons. The electrons hit the
scintillation counter and generate photons, which are amplified by a photomultiplier. The
signal is converted into an electrical signal and used by the image processor. High electron
yields thereby lead to lighter and lower electron yields and to darker pixels in the image.
Fundamentals
48
In-lens detectors also collect the electrons by an applied voltage but the detection of the
electrons happens by a semi-conductor. When an electron hits the detector it generates
electron-hole pairs that lead to an electric signal. In-lens detectors allow a much smaller
working distance and collect the electrons at the point of impact. These advantages lead to
higher resolutions in contrast to the chamber detectors.
Fundamentals
49
3.5 Focused Ion Beam (FIB)
The Focused Ion Beam (FIB) is very similar to the scanning electron microscope mentioned
above, except that the beam scanning the sample surface consists of ions rather than of
electrons. The FIB can be either used to image the sample surface or to process the surface by
sputtering it with ions. Secondary ions are generated during the scanning by the interaction of
the ions with the sample atoms and can be detected in a similar way as in the SEM.
In most commercial FIBs Gallium is used as ion source for the beam. The low melting point
of Ga allows the easy setup of liquid metal ion sources based on a tungsten needle from where
the Ga is extracted by field emission (1010 V/m) with a current of up to 2 µA. Similar to the
SEM the columns in which the ion beam is accelerated by 5 – 30 kV consist of lenses which
focus the ion beam. Due to the fact that the focusing strength of electromagnetic lenses is
directly related to the charge / mass ratio electrostatic devices are used to reduce the weight.
During the investigation the ion beam scans over the sample and interacts by different
processes like ion backscattering, electron emission, sputtering, sample damaging and heating
[148, 149].
As already mentioned the formation of secondary electrons can be used to image the surface
but thereby the sample will be contaminated by Ga-ions. This effect can be used to investigate
non-conductive samples as they are plated by a conductive Ga layer.
Highly accelerated ions can be used to sputter the surface and thereby create structures in the
micro and nano meter scale. Sputtering is often used to prepare lamellae for TEM
applications. During the milling the sample is heated due to the current flow by the ion beam.
The maximum power reached by commercial systems is in the range of 1 mW. Depending on
the thermal conductivity of the material negligible or significant temperature changes appear.
The temperature increase can be calculated accordingly eq. 52.
()
κπ
⋅⋅
=∆ r
P
T (eq. 52)
with
T
∆ the temperature change, P the power of the ion beam, r the radius of the ion beam
and
κ
the thermal conductivity of the sample.
In commercial FIB P/a values of 1 W/m up to 1000 W/m can be archived, causing a
T
∆ for
iron (
κ
= 80 W/mK) to be about 4°C while for polymers (
κ
= 0.2 W/mK)
T
∆ can reach up to
1600°C.
Experimental
50
4 Experimental
4.1 Materials, Electrolytes and Parameters
4.1.1 Substrate and sample preparation
4.1.1.1 Preparation of corrosion protection primer samples
The material used for coating applications consists of ductile electro-galvanised steel
recommended for high forming degrees (DC 06 ZE 75/75, thickness 0.8mm), coated with a
7.5 µm zinc layer on both sides. Before coating with organic layers, the steel sheets (500 x
200 mm) were cleaned using an alkali brush cleaning line (Wesero, Germany) and rinsed with
deionised water.
Afterwards all steel samples were first treated with a chromate-free conversion layer
(Granodine 1456, Henkel, Germany). Additionally, the metal substrates were coated with a
3.5 µm thick primer containing microscopic zinc particles and corrosion inhibition pigments
(Granocoat ZE, Henkel, Germany) and applied using a lab roller coater (Mathis, Switzerland).
The coating was cured in a continuous furnace at the peak metal temperature of 260 °C.
In-situ uniaxial forming was applied to specially shaped, miniature stretching samples. The
samples were cut from the coated steel sheets using a pulsed 300 W YAG laser (Lasag Vega,
Switzerland).
For the open circuit measurements during the phosphating step, ex-situ stretched samples (see
below) were cut to 50 x 30 mm and sealed in the non-coated areas with a corrosion protection
lacquer. The laquer was dried at room temperature for 24 h.
4.1.1.2 Preparation of polyelectrolyte coated samples
Steel substrate used for the preparation of polyelectrolyte (PET complex synthesised by O.
Seewald, University of Paderborn) coated samples are similar to those used for the
preparation of corrosion protection primer samples (DC 06 ZE 75/75, thickness 0.8 mm). The
laser cut miniature stretching samples and the manually cut samples (20 x 40 mm) were first
cleaned by a three step solvent cleaning and afterwards by an alkaline cleaning bath.
The solvent cleaning consists of ultra sonic cleaning in “pro analysis” tetrahydrofuran,
isopropanol and ethanol (Merck, Germany). The alkaline cleaning was undertaken in a mild
Experimental
51
alkaline bath at 55 °C (Ridoline 1553, Henkel, Germany), afterwards the samples were rinsed
with ultra-purified water (Elga Purelab, Germany) and dried in a nitrogen stream.
The coating procedure took place with a home-built dip coater at a dipping speed of 2 mm / s.
The PET complex was diluted 1:3 by weight with butylglycol. The curing of the coating was
carried out at room temperature (24h and 30d), 130 °C and 230 °C (both 2h) under standard
and nitrogen atmosphere.
4.1.2 Forming of samples
4.1.2.1 Forming for Ex-Situ Analysis
Biaxially formed, so-called Marciniak samples were prepared from 225 x 225 mm coated
steel sheets using a hollow round shaped stamp with a diameter of 100 mm on an Erichsen
deep drawing machine (Erichsen, Germany). The blank holder was adjusted to maximum
force and forming degrees of φv = 0.10, 0.15 and 0.25 were prepared.
Plane strain formed samples were prepared from round shaped steel sheets (diameter 200
mm) with special laser cut-outs at the edges (Fig. 22). The same setup as for Marciniak
samples was used and forming degrees of φv = 0.10, 0.15 and 0.25 were prepared.
Uniaxial forming with elongations between 5 and 20 % was performed by means of a Z 100
(Zwick, Germany) tensile testing device using 500 x 80 mm organically coated steel strips.
The positions of the odometer were marked on the strips and the samples were cut from
between the marks to guarantee reproducible forming.
These samples were used for FE-SEM investigations, focused ion beam cross section
measurements and open circuit potential measurements during phosphating.
4.1.2.2 GOM® Grid Evaluation
Forming degree evaluation of the stretched samples was performed using an optical grid
evaluation system (GOM®, Gesellschaft für Optische Messtechnik, Germany). An
electrochemically etched grid with equidistant spots (1 mm distance) was applied to the
sample surface and optically analysed with a digital camera after forming. Following the grid
detection, all three forming degrees can be calculated using the PC-based software when
comparing the original with the formed grid.
Experimental
52
4.1.2.3 Finite Element Simulation
The different elongations of the miniature stretching sample were also analysed with regard to
homogeneity and forming degree by means of finite element simulation (INDEED®,
Gesellschaft für numerische Simulation, Germany). For easier analysis, the three forming
degrees (ϕ1, ϕ2, ϕ3) were combined to the equivalent strain.
4.1.3 Preparation of biaxial formed samples for corrosion testing
For the phosphating, ED-paint application and finally corrosion testing of biaxially formed
samples a special mounting was constructed (Fig. 21). First the samples were formed like a
standard Marciniak sample with a defined forming degree (φv = 0.25). Afterwards the round
shaped, flat middle part was cut out with a diameter of 75 mm by a 300 W Nd:YAG laser
(Lasag Vega, Switzerland).
Fig. 21. Sample mounting for corrosion testing of formed corrosion protection primers
The circular blank was perforated at one edge and mounted to a carrier sheet (100 x 200 mm).
These carrier sheets fit into the mountings of the phosphating and ED-paint station. After the
phosphating and ED-paint application the samples were placed in racks inside the corrosion
chamber.
4.1.4 Preparation of plane strain formed samples for scanning in-situ
Raman / EIS investigation
For the scanning in-situ Raman / EIS investigation plane strain formed samples were
prepared. Before the forming operation round shaped specimen with defined cut-outs were
prepared from the coated samples by laser cutting (Fig. 22).
Experimental
53
Fig. 22. Preparation of plane strain formed samples for in-situ Raman / EIS analysis
Afterwards the samples were mounted in an Erichsen device and formed similar to the
Marciniak samples. Again the round shaped middle part of the sample was cut out with a laser
and used for further investigations. The circular blanks were cut into three pieces with the
middle part orthogonal to the homogeneously formed area of the sample (see forming of plane
strain samples for details).
4.1.5 Phosphating and ED-paint application of formed corrosion
protection primer coated samples
The phosphating of the stretched SEM samples was performed in a car phosphating simulator
based on the real automotive process with cleaning, activation and phosphating steps at 55°C
(Chemetall, Germany). If applied the ED-paint application followed the phosphating step. The
paint was applied in a laboratory scale by cathodic deposition (Cathoguard 310, BASF,
Germany).
The in-situ open circuit potential measurement during phosphating was conducted after
activation in a standard phosphating bath at 55 °C (Henkel Surface Technologies, Germany).
4.1.6 FE-REM investigations and FIB preparation
FE-SEM pictures of surfaces and cross sections were taken for different forming degrees
using a LEO-Zeiss 1530 Gemini (Germany). The acceleration voltage was set at 0.5 kV and
the in-lens detector was used to achieve a high surface resolution and to monitor the
formation of cracks in the organic coating. Focused ion beam (FIB) cross sectioning was
prepared on a LEO-Zeiss 1540 XB. The cut was milled by a gallium ion beam with a current
of 30 nA. Finally surface images of ex-situ stretched, phosphatised samples and of
Experimental
54
polyelectrolyte samples were taken by a LEO-Zeiss 1550 VP with a secondary electron
detector using different acceleration voltages.
4.1.7 Raman and IR spectroscopy
Raman spectroscopy was carried out using a Raman microscope system (Dilor, LabRAM,
ISA Instruments, France). It consists of a red HeNe laser (632.8 nm) with 20 mW power and a
green Ar+ laser (514.5 nm) with adjustable output power. Four objectives with 10x and 100x
magnification were selectable. During the measurements the laser power was set at 2 mW for
the adsorption measurements of MBT (Ar+ laser) and 20 mW (HeNe laser) for the corrosion
product identification of CPP. The pinhole was set to 1000 µm and the size of the slit to 100
µm. The integration time was set to 10 seconds and 10 spectra were taken to optimize the
signal to noise ratio.
For the IR spectroscopy a Fourier transformation infrared spectrometer (Biorad, USA) was
used with a microscope ATR unit. A liquid nitrogen cooled MCT detector was used and the
spectra were acquired with 512 scans at a resolution of 4 cm-1.
4.1.8 Glow Discharge Optical Emission Spectroscopy (GDOES)
The Glow Discharge Optical Emission Spectroscopy (GDOES) was performed with a LECO
GD-OES 750 (LECO, USA) system. The system was driven with a current of 20 mA at 700 V
DC voltage and a sampling rate of 20 Hz.
4.1.9 Laser Optical Emission Spectroscopy (Laser-OES)
Laser Optical Emission Spectroscopy (Laser-OES) was conducted with a home-built scanning
spectrometer which allows scanning the sample surface. The sample size was chosen to be 5 x
5 mm and scanned with 500 x 500 spots. For the pretreatment distribution only titanium was
taken as tracer element.
Experimental
55
4.1.10 Electrochemistry and electrolytes
During the in-situ impedance analysis the whole setup was placed in a Faraday cage and
connected to a highly sensitive potentiostat (Gamry Femtostat FAS2, USA) with an
implemented frequency analyser.
The in-situ Raman / EIS measurements were also done with a Gamry Femtostat.
All measurements were performed at room temperature between 100 kHz – 0.1 Hz (or 1 Hz)
using 0.05 M NaCl or borate buffer as electrolyte. 10 points per decade were recorded at open
circuit potential.
Pro analysis grade chemicals (Merck, Germany) and ultra-purified water (Elga Purelab,
Germany) were used for the preparation of the borate buffer electrolyte (pH 8.3) and the 0.05
M NaCl solution.
For the open circuit measurements during the phosphating step the ex-situ stretched samples
were measured versus a 3 molar Ag/AgCl reference electrode (Metrohm, Switzerland).
4.1.11 Corrosion Testing
Two accelerated corrosion tests were used to investigate the protective properties and the
corrosive behaviour of the formed corrosion protection primers. The first test in the common
salt spray test according to DIN 50021-SS due to the continuous load with corrosive medium
galvanised samples undergoes a fast attack (metal removal of ≈ 1µm Zn / 10h). The following
conditions are given during the testing:
• Salt concentration: 50 ± 5 g/l NaCl
• Amount of salt solution: 1.5 ± 0.5 ml/h
• pH-value of the salt solution: 6.6 – 7.2
• Temperature: 35 ± 2 °C
• Sample orientation: 60 – 70 °
The salt spray test shows a strong corrosive attack which is uncommon for the corrosive load
appearing during the use of an automobile. Cyclic corrosion tests show a much better
correlation between the field results and the laboratory testing. The standard test for cyclic
corrosion testing is the VDA 621-415. It is a combination of salt spray testing, wet and dry
periods. The testing details are given below and in Tab. 3:
Experimental
56
• 24 h salt spray testing according to DIN 50021-SS
• 8 h at 40 ± 3 °C with 100 % relative humidity, 16 h at 18 – 28 °C with 55 – 65 %
relative humidity according to DIN 50017
• 24 h at 18 – 28 °C with 55 – 65 % relative humidity according to DIN 50014
Tab. 3. One cycle (one week) of the accelerated corrosion testing according to VDA 621-415
Monday Tuesday Wednesday Thursday Friday Saturday Sunday
Salt spray
testing
DIN 50021
Climate change testing DIN 50017 Constant climate testing
DIN 50014
Experimental
57
4.2 In-situ Electrochemical Impedance Spectroscopy (EIS) and Forming
Setup
The major disadvantage of the ex-situ EIS of formed materials is the time-consuming sample
preparation and experiment procedure. Each sample has to be prepared and measured
individually for each forming degree.
A much faster approach is measuring the impedance at the same time as the forming of the
sample. Capillary cells provide a useful electrochemical technique for reducing the sample
preparation time and achieve a high local resolution [150-152]. Pilaski et al. show that they
are also suitable for EIS experiments on metal substrates [153].
During the process, a glass capillary is pressed to the sample surface. The inner area forms the
working electrode. The sample needs no special preparation or handling and due to this fact,
the capillary diameter can range from millimetres to micrometres. The sample size can be
much smaller and the required quantity of material is reduced. One limiting factor for the
reduction of the capillary tip is the input impedance and resolution of the potentiostat (Gamry
Femtostat FAS2, USA), which works in the femto ampere range and therefore restricts the
total maximum detectable impedance of the system to about 1012 Ω (with 10 mV amplitude)
[154]. The capacity of the wires used becomes more and more important as the working
electrode area is reduced, since it reaches the same range as the monitored signal and the two
values can no longer be distinguished. The capacity of 50 cm long connection wires can be
calculated to be nearly 50 pF. This means that the lowest detectable capacity must be larger
than 50 pF in order to remain separate.
A few disadvantages arise as a result of combining capillary cells with standard stretching
devices. The sample holder is mainly erected vertically, which leads to problems when
positioning the capillary. Moreover, the advantage of the smaller sample size is invalidated
since the sample holder is too large to handle these small samples. Custom-built miniature
stretching devices avoid these problems and allow horizontal sample mounting. Furthermore,
the setup size can be reduced to take advantage of the sample size reduction achieved by the
use of a capillary cell. Such devices need specially shaped, smaller samples to avoid an
overload and to enable homogeneous material flow over a wide forming range. A special
point of interest is the formed area in the middle of the sample where the measurement data
for the electrochemical experiment is obtained. Here it is crucial to have a uniform strain
distribution in order to achieve high reproducibility for the experiments.
Experimental
58
For the in-situ measurment of EIS during tensile testing, a new setup was designed, which
combines the advantages of localised electrochemistry with those of miniature stretching
samples. The setup consists of three main parts:
− Custom-built, miniature tensile testing device with home-made, isolated sample holder
− Home-made three-electrode electrochemical capillary cell
− Home-made mounting for capillary cell and stretching device
The miniature tensile testing device is a custom-built setup (Kammrath & Weiß, Germany)
which allows the uniaxial forming of samples with a load of up to 5 kN. The sample holder is
specially designed (home-made) in order to allow the electrical isolation of the sample from
the device (Fig. 23).
Fig. 23. Miniature stretching device with isolated sample holder and mounted sample; (right) detailed
view of a capillary cell positioned on a sample during in-situ EIS analysis.
The external controller displays the elongation and stretching force and can send the data to a
personal computer for storage and further processing. A new stretching sample was designed
which features both a small size and a homogeneous forming characteristic over a large
elongation range (Fig. 24).
Experimental
59
Fig. 24. Specially designed miniature stretching sample
For highly reproducible spectra during EIS analysis, it is important to measure in an area with
a homogeneous forming degree. This requires a special sample geometry and a high local
resolution. The high local resolution was achieved by means of a capillary-based cell where
the working electrode is formed by the inner area of the capillary. The cell setup used is
related to one described by Lohrengel et al. [155]. It consists of a Plexiglas® carrier including
a reference electrode, gold counter electrode and a glass capillary which forms the junction
between the counter/reference and the working electrode (sample surface). The area of the
working electrode is formed by the inner area of the capillary and was calculated to be 8 ⋅ 10-3
cm2. To avoid leaking and evaporation of the electrolyte and to form a reproducible working
electrode area, the capillary tip is equipped with a flexible silicon gasket. The gasket is
pressed against the sample surface during the measurement process. A force sensor measures
the level of contact and allows the force between the capillary and the sample to be
controlled. Two different kinds of reference electrodes can be used depending on the
electrolyte: mercury electrodes such as mercury/mercury sulphate electrodes or
environmentally friendly silver/silver chloride or silver electrodes.
The mounting consists of a stainless steel tripod and z-axis stage (Owis-Staufen, Germany)
with 3 µm resolution and a maximum lift of 30 mm. The stretching device is placed on top of
the stage. The electrochemical cell is fixed to the mounting and the stretching unit, and with
the sample in place, is pressed against the capillary tip during the measurement process. Fig.
23 shows the complete setup with tripod, capillary cell and stretching device.
Experimental
60
4.3 In-situ Quartz Crystal Microbalance / Raman Setup
The adsorption kinetics of monolayers is of great interest for the effectiveness of inhibitors
and conversion layers and for the crystal growth in general. A perfect tool that has enough
sensitivity for detecting mass changes in the ng/cm² range is the Quartz Crystal Microbalance
(QCM). The QCM consists of quartz oscillating at its resonance frequency, additional load to
the quartz leads to a frequency shift that is proportional to mass changes (eq. 31). The QCM
can be used either in a gas or in a liquid environment where it can be combined with
electrochemical tools for detecting the mass change under potential control.
Raman Spectroscopy on the other hand gives information about the structure of molecules
and can be used to identify them. In the case of monolayers the Raman signal is weak and
cannot be detected by common systems. Fleischmann et. al. [142] show that certain metal
nano particles can lead to an amplification of the signal. The so-called Surface Enhanced
Raman Spectroscopy (SERS) allows the detection even of monolayers attached to the nano
porous material. The combination of both tools allows the in-situ investigation of the
adsorption process of molecules.
For the combined in-situ measurement a new QCM flow cell was designed which allows the
simultaneous Raman investigation of the quartz crystal surface. The used QCM (Maxtek
RQCM, USA) is specially designed to work with high mass loads and in a liquid
environment. The setup consists of an oscillation circuit, a Teflon quartz holder and a
polished gold coated 5 MHz quartz. The holder of the quartz crystal was extended by the
newly designed and home-built flow cell.
The flow cell consists of a circular Teflon® body which is pressed against the quartz crystal
by an o-ring. The setup forms a volume of about 100 µl between the quartz and the Teflon®
body. For a constant electrolyte flow an in-let and out-let are adapted and connected by
Teflon® tubes to a computer controlled pulsation free syringe pump (MDSP3f, Micro
Mechatronic GmbH, Germany). The electrolyte can either flow from a beaker through the
pump and the cell back to the beaker, or from a reservoir beaker through the setup into a
waste compartment. The middle of the cell consists of an 11 x 11 mm quartz glass window to
couple the laser beam from the Raman microscope (Dilor, LabRAM, ISA Instruments. SA,
INC, France). The laser is focused on the surface of the quartz crystal and the emitted
radiation of the molecules is scattered back to the spectrometer. To allow the detection of
adsorbed monolayers the SERS effect is used. Therefore the gold electrode of the quartz was
Experimental
61
covered by a nano porous SiO2 and a 100 nm thick silver or gold layer [156]. Fig. 25 shows a
schematic sketch of the measurement setup.
Fig. 25. Schematic sketch of the in-situ Raman / QCM flow through concept. The electrolyte cycles
through the system and the inhibitor under investigation can be added after the system comes to a steady
state
The cell can be used in two configurations on the one hand just for in-situ Raman / QCM
measurements and on the other hand for in-situ Raman / QCM measurements under potential
control. For investigations under potential control the cell body can be further equipped with a
counter and a reference electrode (Fig. 26). As reference electrode a micro silver / silver
chloride electrode similar to the one described by Hassel et. al. [157] can be connected.
Fig. 26. Photo of the newly developed in-situ Raman / QCM flow cell, showing the electrolyte inlet / outlet
and the connections for the counter and reference electrode
The counter electrode consists of a gold wire which is mounted around the Raman window in
the middle of the cell. The working electrode will be in this case formed by the gold electrode
of the quartz crystal.
Beaker
with stirrer
PC controlled
syringe pump
Ar or He/Ne laser
beam from Raman
spectrometer
Quartz crystal with SERS
active layer
Quartz glass
window
Electrolyte flow
Experimental
62
4.4 Scanning in-situ Raman / Electrochemical Impedance spectroscopy
Setup
The investigation of localised corrosion phenomena appearing on organically and / or
metallically covered substrates is of great interest for understanding the corrosion mechanism
in small defects and on impurities. In this case investigation methods confined to small areas
allow gaining detailed information about the processes. The demand of high local resolution
is fullfilled by both Raman spectroscopy and electrochemical capillary cells. While Raman
spectroscopy delivers detailed information about the corrosion products formed,
electrochemistry and in particular impedance spectroscopy can give detailed information
about the protective properties of organic and metallic coatings.
The combination of both techniques delivers a detailed view of the processes that occur
during corrosion. Therefore an electrochemical capillary cell was designed which can be
mounted into a Raman microscope (Dilor, LabRAM, ISA Instruments, France) and allows the
spectroscopic and electrochemical investigation in the same area of the sample surface. The
basic cell used is related to the capillary cell described in chapter 4.2. It consists of a
Plexiglas® carrier including a reference electrode, gold counter electrode and a glass capillary
which forms the junction between the counter/reference and the working electrode (sample
surface). The area of the working electrode is formed by the inner area of the capillary and
was calculated to be 8 ⋅ 10-3 cm2. The sample is placed on a special stage equipped with a
force sensor which measures the level of contact between the capillary and the sample. The
reference electrode can either consist of a mercury electrode like e.g. a mercury/mercury
sulphate electrode or silver/silver chloride electrode. The complete capillary cell is mounted
as an exchange for one of the objectives in the Raman microscope and is connected to a
potentiostat (Gamry Femtostat FAS2, USA) (Fig. 27).
Experimental
63
Fig. 27. Photo of the in-situ Raman / impedance setup showing the capillary cell mounted to the Raman
spectrometer and in touch with a formed sample
During the analysis the same spot on a sample can be investigated by either Raman
spectroscopy or electrochemistry just by turning the objective holder of the microscope.
Results
64
5 Results
5.1 Characterisation of corrosion protection primers (CPP)
Corrosion protection primers (CPP) are applied on all grades of electrogalvanised steel which
are covered by 5 or 7.5 µm zinc. The zinc layer leads to a homogeneous, micro-structured
surface without the typical grain boundaries known from hot dip galvanised steel. Fig. 28
shows the undirected terrace structure of the zinc layer on high ductile steel (DC 06). The
layered structure leads to a good paint adhesion and thereby in a macroscopic view to an
excellent formability of the applied CPP.
0411A01752 1000 : 1 20 µm
0411A01755 3000 : 1 5 µm
Fig. 28. FE-SEM image of an unstrained DC 06 ZE 75/75 (electro-galvanised steel) used for the coil
coating application showing the parallel layers of the tilted zinc crystals
Before CPP application the sample was coated with a chromium free pretreatment containing
silane, Zirconium and Titanium which increases the paint adhesion and leads to a reduction of
the under paint corrosion by a passivating layer. The pretreatment (8 mg/m2 Titanium) forms
a thin layer of some nano meter on top of the zinc structure which is visible by the roughening
of the terraces (Fig. 29). Furthermore, the zinc layer edges are rounded by etching attack of
the pretreatment.
Results
65
Fig. 29. FE-SEM image of an unstrained DC 06 ZE 75/75 (electro-galvanised steel) used for the coil
coating application covered with a chromium free pretreatment of few nm thickness. The detailed view
(right) shows only a very thin roughening by the layer
For an optimal adhesion of the corrosion protection primer to the steel substrate the uniform
application of the pretreatment with an average amount of 6 – 10 mg/m² Titanium as trace
element is necessary. The distribution profile acquired by Laser-OES of a 5 x 5 mm sample
shows a homogeneous surface coverage of the pretreatment applied by the roll coater. The
average Titanium amount taken as indicator for the pretreatment is about 8 mg/m².
Fig. 30. Laser-OES measurement of the lab coil coated pretreatment distribution on a 5 x 5 mm electro
galvanised steel sample using titanium as tracer element. The measurement shows a homogeneous
coverage of the complete sample
The top view SEM picture of an unstrained CCP sample with a layer thickness of 3.5 µm is
shown in Fig. 31. The coating shows a rough surface caused by the mostly spherical zinc
particles with a size of 2 – 7 µm. These particles are homogenously distributed in the organic
0606A04033 10000 : 1 5 µm 0606A04041 50000 : 1 500 nm
Results
66
matrix. Furthermore, angularly shaped silicate pigments are visible which are also distributed
within the organic binder system.
The detailed view indicates that the particles are mainly embedded within the resin matrix,
only a small hemispheric part sticks out of the coating and is covered by a porous oxide film.
0508A00864 3000 : 1 10 µm
0508A00871 10000 : 1 2 µm
Fig. 31. FE-SEM image overview (left) and detail view (right) of an unstrained zinc pigmented coil coating
primer, showing round shaped zinc particles mainly embedded in the organic binder matrix
The cross section shows the steel substrate covered with the electro galvanised zinc layer (7
µm thickness) and the CPP layer on top (Fig. 32). The boundary between the organic binder
of the CPP and the resin for the cross section preparation is clearly visible.
The cross section of the CPP displays the roundly shaped zinc particles and the angularly
shaped corrosion inhibition pigments already known from the top view. Most of the zinc
particles are completely covered and float within the organic matrix. Only very few have
contact with the underlying zinc coating, but none of them leads to a direct conducting path
from the base material to the coating surface.
Results
67
0508A01511 5000 : 1 5 µm
Fig. 32. FE-SEM cross section of an unstrained zinc pigmented coil coating primer, showing roundly
shaped zinc particles and angularly shaped corrosion inhibition particles
The cross section prepared by the Focused Ion Beam (FIB) has the advantage of sample
preparation without the smearing effect of the resin known from the preparation of a
metallographic cross section. Here, the occurrence of defects within the coating can be
studied. Unfortunately the gallium ion beam can locally lead to a high thermal influence
within the coating which might result in a local shrinking of the polymer film. External force
during the handling of the coated steel (coiling, cutting etc.) can cause a stretching of the
polymer film and thus an energetically non-optimised molecule orientation. At higher
temperatures the molecule chains can reorient due to the higher mobility and might form a
closer, energy optimised package. The calculated local heating of the sample according to eq.
52 (chapter 3.5) can increase by up to 1600 K for a pure polymer coating. Due to the good
thermal conductivity of the steel substrate (maximum heating of steel ∆T = 4 K, see chapter
3.5) the heating of the applied polymer film should be lower than 1600 K but cannot be
clearly calculated.
The FIB cross section shows the steel matrix covered by the electro galvanised zinc layer and
the CPP on top (Fig. 33). The roundly shaped zinc particles are distributed within the organic
binder. Small defects are visible in the higher magnification at the pigment / binder interface
which can be caused by the heating of the ion beam. At the substrate / binder interface no
defects are visible which proves the good paint adhesion.
Results
68
Fig. 33. Focused ion beam cross section of an unstrained zinc pigmented coil coating primer, indicating a
good adhesion of the coating to the substrate and just local detachment from the zinc particles
The GDEOS depth profile of a CPP covered steel substrate is shown in Fig. 34. Up to three
µm the profile is dominated by the ingredients of the CPP and the pretreatment. Due to the
inhomogeneous sputtering of materials with different atomic numbers the maximum of C, O
and P is at about 200 nm while Si and Ca have their maximum at 1.2 µm. Ti and Mn are
found in the pre-treatment and show their highest value at 2.2 µm.
Fig. 34. GDOES element and depth profile analysis of a corrosion protection primer
As the two pretreatment atoms Ti and Mn have the maximum at nearly the same sputtering
depth the pre-treatment can be well distinguished from the CPP. The galvanised zinc layer
dominates from 3 to 10 µm with an increasing amount of iron for longer sputter times. The
non-uniform sputtering leads furthermore to a fluent crossover between the different layers
(CPP, galvanised zinc and steel) and the sputter depth can only be used as a rough orientation.
Results
69
5.2 Uniaxial, biaxial and plane strain forming of corrosion protection
primers
5.2.1 Strain evaluation of formed samples
Uniaxially, biaxially and plane strain stretched samples were used to study the behaviour of
CPPs during forming. The homogeneous forming in the whole sample area is crucial for a
reproducible sample preparation. Therefore a set of model samples was formed and
afterwards analysed with respect to their forming state.
The GOM® evaluations of uniaxial stretched samples (5, 10, 15 and 20 % elongation) show at
all elongations a homogeneous strain distribution over the complete sample area (Fig. 35).
Even at high forming degrees the sample is still in the range of uniform strain and has no
tendency for constriction. The forming degrees range from φv = 0.05 for 5 % elongation to φv
= 0.18 for 20 % elongation.
Results
70
Fig. 35. GOM® analysis of uniaxial stretched steel samples with 5 % (upper left), 10 % (upper right), 15
% (lower left) and 20 % elongation (lower right hand side) showing a homogeneous strain distribution
over the complete sample
The forming degrees of the biaxially stretched Marciniak samples were investigated by finite
element simulation and by in-situ GOM® analysis during forming in the Erichsen device. The
same stamp way was used in both the experiment and the simulation. Three forming degrees
φv = 0.10, 0.15 and 0.25 were chosen for evaluation from the experiment. All samples show
in the middle part a flat area with a homogeneous strain distribution. At the outer rounding the
forming degrees are significantly higher. The simulated samples show a slightly lower
forming degree especially at the highest forming degree (∆φv ≈ 0.2). This difference might
come from an inaccuracy of the position encoder in the experimental system but all formed
φv φv
φv φv
Results
71
samples showed a reproducible forming. Due to the constant strain in the inner part of the
sample this area was cut out and used for further investigations.
Fig. 36. Optical evaluation (left) and FEM-simulation (right) of a Marciniak samples; top φv = 0.10, middle
φv = 0.15 and bottom φv = 0.25 indicating a homogeneous strain distribution in the middle area of the
sample
φv
φv φv
φv
φv φv
Results
72
In-situ analysis of plane strain samples shows significant differences in contrast to the
biaxially formed samples (Fig. 37). The area of homogeneous forming is limited to a stripe in
the middle of the round shaped, flat area of the specimen. The samples were formed with an
equivalent forming degree (φv) of 0.10, 0.15 and 0.25 in the middle part of the sample.
Results
73
Fig. 37. Optical evaluation of plane strain formed samples (φv = 0.10 top, φv = 0.15 middle, φv = 0.25
bottom), left hand side the equivalent forming degree φv right hand side the forming degrees φ1 and φ2.
The analysis shows only a small area in the middle with a homogeneous strain distribution
φv
φv
φ2
φ1
φ1
φ1
φ2
φ2
Results
74
Close to the edges where the cut-outs are located (top and bottom of the sample), the major
forming degree 1
ϕ
decreases, while the minor forming degree 2
ϕ
rises from zero indicating
non-plane strain forming. Also on the left and right sides of the sample the forming becomes
inhomogeneous, which becomes visible by the increase of the major forming degree. With
higher deformations the transition from the homogeneously to the inhomogeneously formed
area becomes more defined and the area of plane strain forming decreases.
5.2.2 Forming behaviour of the electro galvanised steel substrate
The forming behaviour of the electro galvanised steel substrate is of great importance for the
corrosion resistance of the complete sample system including the CPP. Defects within the
zinc layer lead to conduction pathways and a much faster occurrence of red rust during
corrosion tests. Furthermore, defects in the metal coating can result in defects within the
organic coating due to an inhomogeneous strain distribution. Therefore, the most demanding
forming operation (biaxial forming) was used to investigate the forming performance of the
steel substrate.
The SEM images (Fig. 38) of biaxially formed samples (φv = 0.10, 0.15 and 0.25) show no
damages within the zinc layer.
Results
75
0601A04127 3000 : 1 5 µm
0601A04136 3000 : 1 5 µm
0601A03604 3000 : 1 5 µm 0601A03610 10000 : 1 2 µm
Fig. 38. FE-SEM images of the biaxially formed sample; top left φv = 0.10, top right φv = 0.15 and bottom
φv = 0.25 (overview and detail). The images show no damages of the zinc layer due to the forming even for
high forming degrees
With increasing strain the surface looks smoother due to the sliding along the terraces of the
electrogalvanised zinc structure.
5.2.3 Uniaxial forming of dry corrosion protection primers
The surface morphology of a dry, ex-situ uniaxially stretched corrosion protection primer was
studied by FE-SEM (Fig. 39). The samples were elongated to 5, 10, 15 and 20 % (ϕv,max =
0.18). Even at very low extensions small defects occur within the organic matrix. These
defects mainly arise perpendicular to the elongation direction at the interface between the
binder and the zinc particles. With increasing strain the defect size rises from a few hundred
nm at 5 % elongation to about some µm at 20 % elongation. Furthermore, the number of
Results
76
defects rises with increasing strain and provides conduction pathways for the electrolyte
through the coating.
0508A02426 3000 : 1 10 µm 0508A02446 3000 : 1 10 µm
0508A02455 3000 : 1 10 µm 0508A02474 3000 : 1 10 µm
Fig. 39. FE-SEM images of uniaxially formed coil coated electrogalvanised steel samples; top left 5% (φv =
0.05), top right 10% (φv = 0.08), bottom left 15% (φv = 0.15) and bottom right 20% elongation (φv = 0.18).
Crack formation occurs already at the lowest forming degree. The defect size increases with increasing
forming
The FIB cross section of 20 % elongated sample shows the steel substrate covered with the
electrogalvanised zinc layer and the corrosion protection primer on top (Fig. 40). The
elongated sample not only shows defects at the surface between the particles and the binder
but also within the coating. Most of the zinc particles are mainly detached from the binder and
float in the organic matrix. Some of these defects might arise by the shrinking of the binder
system caused by the thermal input from the gallium ion beam (compare FIB cross section of
an unformed sample) but most of them are generated by the sample forming. Apart from the
defects visible inside the coating, the binder locally detaches from the zinc surface and forms
caverns underneath the coating.
Results
77
Fig. 40. Focused ion beam cross section of a uniaxially (20% elongation, φv = 0.18) formed primer on
electrogalvanised steel, indicating the formation of defects along the particle / binder interface and the
binder / substrate interface
In case of good adhesion forces at the particle / binder and the binder / substrate interface
some strain might be still unloaded and keep the coating under tension. This could lead to a
fast formation of further defects if some external force will be applied as the interface is
already weak.
5.2.4 Uniaxial forming of wetted corrosion protection primers
For further in-situ electrochemical investigations during stretch forming it is of great
importance that the forming behaviour of a wetted corrosion protection primer is similar to a
dry one. Therefore a sample was stored in wet conditions (wrapped in wet tissues and sealed
with a plastic foil) for 48 h and was elongated under these wet conditions to 20 %. Afterwards
the sample was dried and FE-SEM images were taken (Fig. 41).
3 µm 1 µm
Results
78
0509A04855 3000 : 1 10 µm
Fig. 41. Uniaxially stretched (20 % elongation, φv = 0.18) corrosion protection primer
stored for 48 hours under wet conditions. The primer shows the same defect formation as a dry one (see
Fig. 39)
The sample shows the same defect formation at the particle / binder interface in the range of
some µm which is also visible for a dryly formed CPP. The highly cross linked binder seems
to show only a very low water uptake and swelling of the organic film which would cause a
softening of the coating and lead to less or no defect formation. Instead the coating shows
similar forming properties even under wet conditions. This allows the estimation that the
much shorter electrolyte contact (some minutes) during in-situ EIS / stretch forming will have
no influence on the forming behaviour of the coated sample.
5.2.5 Biaxial forming of corrosion protection primers
The biaxial forming (Marciniak sample) of corrosion protection primers to ϕv = 0.10, 0.15
and 0.25 leads to similar defects at the pigment binder interface as already known from the
uniaxial forming (Fig. 42).
Results
79
0601A04472 3000 : 1 10 µm 0601A04499 3000 : 1 10 µm
0601A04510 3000 : 1 10 µm
Fig. 42. FE-SEM images of a biaxially formed primer; top left ϕv = 0.10, top right ϕv = 0.15 and bottom
ϕv = 0.25. The formed defects show no preferred orientation and grow with increasing strain
The size of the defects increases with the level of strain applied to the sample and reaches up
to 15 µm at an equivalent strain ϕv of 0.25. In contrast to the uniaxial forming the defects
don’t show a preferred orientation anymore as the biaxial strain forces the coating to elongate
in equal amounts into the x and y direction. The coating failures lead to partly uncoated zinc
particles and to the formation of electrolyte pathways directly to the zinc coated substrate
underneath (compare cross section of phosphated sample 5.2.7).
5.2.6 Plane strain forming of corrosion protection primers
Plane strain is the forming step that is placed between the biaxial and uniaxial forming in the
forming limiting curve. This means that there is a major forming direction ϕ1 in which the
material flows. Unlike the uniaxial forming the minor forming degree ϕ2 is in this case zero.
Results
80
The only elongation along the major forming degree leads to the formation of cracks
perpendicular to the forming direction which reduces the strain within the coating.
The FE-SEM images show for an equivalent strain of ϕv = 0.10 initial defects at the particle
binder interface in a range of up to one µm (Fig. 43).
Fig. 43. FE-SEM images of plane strain formed primer, top left ϕv = 0.10, top right ϕv = 0.15 and bottom
ϕv = 0.25. The images show the formation of defects along the particle / binder interface which combine at
high forming degrees to large defects of some 10 µm
With increasing elongation (ϕv = 0.15) the number and size of the cracks rises until the
separated defects are combined and form large valleys with a size of some 10 µm (ϕv = 0.25).
5.2.7 Cleaning and phosphating of biaxially formed primers
The automotive tri-cation phosphating is used to increase the paint adhesion on galvanised
steel and to passivate the surface. It leads to the homogeneous formation of phosphate crystals
on top of the metal surface. The formation of crystals is coupled to an etching process of the
Results
81
free zinc and the resulting pH-shift of the solution in front of the surface. This leads to a
precipitation layer on the metal.
On corrosion protection primers the deposition of phosphate crystals can only happen in areas
where the zinc particles of the coating or the zinc layer of the steel substrate are in contact
with the phosphating solution to provide the necessary reaction sites. Another important point
for the reactivity of the activator (titanium phosphate) and thereby the nucleation of crystals is
the surface morphology and the oxide thickness of the zinc. The cleaning procedure
(commonly an alkaline cleaning) before the phosphating influences the zinc oxide layer and
maybe the morphology of the particles. Therefore, the influence of benzene, mild alkaline and
strong alkaline cleaning to the surface structure and the phosphate crystal appearance on
formed CPPs was investigated.
The FE-SEM images of the biaxially formed and differently cleaned samples (Fig. 44) show
no visible change of the surface structure. The forming induced defects still have a similar
size and direction compared to the non-cleaned (see above). The free zinc area of the particles
shows a smooth surface and no influence due to the cleaning procedure. Changes in the zinc
oxide thickness cannot be assessed by the FE-SEM investigation.
Results
82
0602A01221 3000 : 1 10 µm 0602A01238 3000 : 1 10 µm
0602A01256 3000 : 1 10 µm
Fig. 44. SEM images of benzene (top left), mild alkaline (top right) and strong alkaline (bottom) cleaned
coil coated samples. The cleaning procedure shows no visible influence on the formed primer
The images of the formed and phosphated samples show influences due to the different
cleaning procedures (Fig. 45). The benzene cleaning leads to the formation of large phosphate
crystals on top of the zinc particles and inside the forming-induced defects. The crystal
formation only occurs on free zinc surfaces and defects are partly sealed by the crystals. The
mild alkaline cleaning leads to the formation of smaller crystals inside the defects, while the
crystals on top of the particles still have the same size compared to benzene cleaning. For the
strong alkaline cleaning the crystals inside the defect areas are comparable to the mild
alkaline cleaning but in contrast to that the crystals on top of the particles also become much
smaller.
First the alkaline treatment seems to influence the free zinc area in the formation induced
defects and with increasing pH also attacks the partly organically covered top of the zinc
Results
83
particles. The reduction of the oxide layer by the alkaline treatment seems to increase the
number of nucleation sites resulting in the formation of smaller crystals.
0602A01232 3000 : 1 10 µm 0602A01243 3000 : 1 10 µm
0602A01270 3000 : 1 10 µm
Fig. 45. SEM images of phosphate crystal nucleation on benzene (top left), mild alkaline (top right) and
strong alkaline (bottom) cleaned coil coated samples. Slight differences in the crystal formation for the
benzene and alkaline cleaned sample re visible
In the FE-SEM cross section the formation of phosphate crystals on free zinc surfaces is
visible. The forming induced defects that occur along the pigment / binder interface lead to
additional uncovered zinc areas and thereby to the formation of phosphate crystals not only on
top but also on the side of the pigments (Fig. 46).
Results
84
0602A03019 3000 : 1 10 µm
Fig. 46. FE-SEM cross section of a formed and phosphatised corrosion protection primer indicates that
the defect formation goes down to the zinc layer and leads to the precipitation of phosphate crystals
Some large defects form conductive pathways through the coating to the zinc layer on top of
the steel and enable the crystal growth at the bottom of the defect. The crystal formation leads
to a partial sealing of the forming induced defects.
5.2.8 Open Circuit Potential during Phosphating
Unformed and ex-situ stretched samples were phosphated in accordance with a standard
automotive process and the open circuit potential (OCP) was measured simultaneously. Fig.
47 displays the development of the OCP during the phosphating step. The coated samples
initially show a drop of the OCP to a minimum value between -964 mV and -970 mV (vs.
Ag/AgCl) followed by a sharp increase with a maximum value between -905 mV and -892
mV. Finally, the potentials undergo a slight negative shift until they reach a plateau at -920
mV after about 250 s. For more highly strained samples, the slope after the first minimum
value rises and the peak OCP is reached in shorter time. In comparison with the coated
samples, the uncoated electrogalvanised steel sheet shows a more negative starting potential
of -990 mV, followed by a sharp rise with a maximum of -450 mV. The maximum value and
the plateau potential of -575 mV are achieved in even shorter times.
The increase of the OCP during phosphating is an indicator of the increasing surface coverage
of the metal surface with blocking phosphate crystals similar to the CV experiments
conducted by Losch et. al. [158].
When comparing the slope and the peak potential of the uncoated and coated samples, it is
important to remember that the phosphating rate depends on the free zinc area. During the
Results
85
phosphating the zinc surface is etched which leads to a dissolution of zinc ions and a pH-shift
in front of the surface. Due to the increase of the zinc concentration and the change in the pH-
value the solubility product of the phosphate crystals is quickly achieved and the precipitation
starts.
The change in the slope and the peak potential shift for the formed coated samples indicate
that the phosphating speed increases at higher elongations due to the quicker achievement of
the maximum solubility of the zinc phosphate. As the increase is related to the uncoated zinc
surface formed by the uncovered zinc particles in the coating and the free sample surface, an
acceleration of precipitation can be correlated with an increase of the free zinc surface by the
formation of defects. The differences in the plateau potential arise from an unclosed
phosphate crystal layer on top of the coated samples. No dense packing of phosphate crystals
occurs within the defect because in most cases only one crystal fits into a defect and it does
not completely cover the defect area. The remaining zinc surface is still reactive, which leads
to the more negative potential in comparison with the phosphated zinc coated steel substrate.
The acceleration of the phosphating process with increasing strain and therefore increasing
defect size is a synergistic effect. The increased amount of reactive zinc surface and the
improved adsorption of the activated Ti-phosphates due to the zinc oxide removal by the
alkaline cleaning lead to a faster nucleation of crystals.
0 50 100 150 200 250 300
-1.00
-0.99
-0.98
-0.97
-0.96
-0.95
-0.94
-0.93
-0.92
-0.91
-0.90
-0.89
-0.60
-0.55
-0.50
-0.45
0% elongation
10% elongation
15% elongation
20% elongation
ZE steel substrate
OCP / V (vs. Ag/AgCl)
t / s
Fig. 47. Open circuit potential of Zn particle containing organically coated and uncoated electrogalvanised
steel at different forming degrees during tri-cation phosphating. With increasing strain the phosphating
kinetic is accelerated and the final potential is shifted to more positive values
Results
86
The uniaxially formed and phosphated sample shows phosphate crystal growth only on free
zinc surfaces (Fig. 48).
Fig. 48. Overview (right) and detailed top view (left) of a 20 % elongated, organically coated electro-
galvanised steel sample after phosphating. The phosphate crystals grow on the free zinc area of the
particles (on top and inside the defects)
These surfaces are formed on top of large zinc particles that protrude from the coating and by
the formation of cracks at the pigment / binder interface. The SEM pictures clearly support
the OCP transient evaluation by illustrating that the remaining defect-free areas range around
the formed 3 to 5 µm phosphate crystals.
5.2.9 ED-paint application on corrosion protection primers
The application of corrosion protection primers may allow the reduction of ED-paint
thickness while achieving the same corrosion resistance. For the application on outer body
parts it is essential that the appearance of the ED-paint is smooth and without defects.
Biaxially formed CPPs (φv = 0.25) with reduced ED-paint thickness (11 and 6 µm) were
investigated concerning their surface morphology and the defect formation (Fig. 49). The top
view shows for both thicknesses a rippled surface. The sample with a thicker ED-paint layer
shows a much smoother surface but the structure of the CPP is still visible while the thinner
coated sample clearly badges the roughness of the zinc particles.
The uncovered zinc surfaces lead to a local increase of the current density during ED-paint
application which increases the deposition rate in these areas. The phosphate crystals on the
other hand act as an isolator and hinder the paint deposition. Both effects induce an
inhomogeneous thickness distribution of the ED-paint (see cross section Fig. 49).
Results
87
0607A06222 3000 : 1 10 µm 0607A06256 3000 : 1 10 µm
Fig. 49. FE-SEM top views and cross sections of a 12 µm (left) and a 6 µm (right) thick ED-painted primer
showing a flattening of the rough surface structure and an inhomogeneous paint thickness
The cross sections show that the inhomogeneous ED-paint thicknesses flatten the surface of
the CPP and heal the forming induced defects. But both ED-paint thicknesses still show a
very structured surface which might affect the appearance of the following processes.
0607A06050 500 : 1 50 µm 0607A06045 500 : 1 50 µm
Results
88
5.3 Corrosion of corrosion protection primers
5.3.1 Corrosion of formed and unformed primers during standard
corrosion testing
The influence of cleaning and forming the corrosion resistance of CPPs was investigated by
standard corrosion testing according to the VDA 621-415. The biaxially formed samples were
benzene, mild alkaline or strong alkaline cleaned and compared with an unformed mild
alkaline cleaned sample. Photos of the sample after different cycles in the corrosion chamber
are shown in Fig. 50. The quantitative analysis of the red rust formation is given in Fig. 51.
The optical investigation shows no differences between the differently cleaned, formed
samples concerning the white rust formation after one cycle and the red rust formation after 5
and 9 cycles. In contrast to the formed samples the unformed sample shows less formation of
white rust after 1 cycle and less red rust after 5 cycles of VDA 621-415 testing. After 9 weeks
in the testing chamber the corrosive load was so high that the differences between the formed
and the unformed sample are no more visible.
Results
89
Cycle
Benzene
cleaned
φv = 0.25
Mild alkaline
cleaned
φv = 0.25
Strong alkaline
cleaned
φv = 0.25
Mild alkaline
cleaned
unformed
1
5
9
Fig. 50. Biaxially formed and unformed corrosion protection primers after VDA 621-415 testing. The
different cleaning procedures show no differences in the formation of white and red rust. The forming of
the CPP shows an acceleration of corrosion in comparison to the unformed sample
The quantitative analysis of the red rust formation of the corroding samples confirms the first
impressions from the visual interpretation (Fig. 51). The red rust formation is not influenced
by the cleaning procedure of the formed samples and already starts after 3 VDA 621-415
cycles. After 5 cycles the maximum of the red rust formation is achieved which means that
nearly the complete zinc layer on top of the steel is dissolved. For the unformed sample the
first appearance of red rust is also after 3 weeks, but the amount is much less than for the
formed samples. Over the whole examination time the amount is clearly less than for the
formed sample and only reaches similar values after 8 weeks.
The forming induced defects allow an easier penetration of the corrosive electrolyte and lead
to a reduction of the protective properties of the CPP by 3 weeks in the VDA 621-415 test.
Results
90
12345678
0
20
40
60
80
100
benzine
mild alkaline
strong alkaline
unformed, mild alkaline
red rust / %
Time in VDA Test / weeks
Fig. 51. Amount of red rust formed on a formed corrosion protection primer sample in VDA Test. The
unformed sample shows a later and slower formation of red rust, while the different cleaning procedures
show no influence
For the use of CPPs with reduced ED-paint thickness especially on the outer body parts it is
of great interest when the first blistering appears. Biaxially formed samples were benzene,
mild or strong alkaline cleaned and reference samples consisted of unformed mild alkaline
cleaned CPP samples and of unformed electro galvanised steel samples. All samples were
coated with different ED-paint thicknesses (4, 12, 17 and 22 µm for the CPP samples and 4,
10, 15 and 20 µm for the steel samples). For very low ED-paint thicknesses of 4 µm the
unformed CPP and the electrogalvanised sample show in contrast to the formed samples a
slightly better resistance with two weeks VDA testing before blistering (Fig. 52).
Results
91
4 6 8 10121416182022
0
2
4
6
8
10
12
14
16
18
20
benzine
mild alkaline
strong alkaline
unformed
electrogalvanised steel
Time before bubble evolution
/ weeks
ED-Paint Thickness /
µ
m
Fig. 52. Blistering of ED-paint applied on corrosion protection primers in dependence on the thickness
At higher thicknesses the results are no more consistent and a strong scattering is visible. This
may be caused by the thin CPP layer (3 - 4 µm) in relation to the ED-paint layer (10 – 22 µm)
which is thus no more the time limiting step of the blistering. This leads to a strong
dependence of the barrier properties on the ED-paint thickness and the homogeneous
application.
5.3.2 Salt spray testing of formed primers
For a fast investigation of the corrosion mechanism of formed corrosion protection primers,
coated electrogalvanised steel salt spray tests according to DIN 50021 were performed. Due
to the high corrosion load the samples were only stored for 24h inside the salt spray chamber.
In Fig. 53 the FE-SEM top view of a uniaxially formed (20% elongation, φv = 0,18) and
phosphatised CPP is shown. Already after 24 h the surface is covered with thin corrosion
products (white rust). These precipitations don’t form a compact film on top and show a
porous structure at the island edges. The corrosion products mainly appear at areas where
defects were formed at the particle / binder interface. Furthermore, the phosphate crystals
covering the defects and the top of the zinc particles show a corrosive attack. The smooth,
roundly shaped particles undergo an etching attack which leads to a rough and porous surface.
Results
92
0602A04094 1000 : 1 20 µm 0602A04100 3000 : 1 10 µm
Fig. 53. FE-SEM images of uniaxially formed (20% elongation, φv = 0.18), phosphatised sample after 24h
of salt spray testing. The sample is partly covered by corrosion products and the phosphate crystals
undergo a corrosive attack
The cross section underlines the corrosion of the phosphate crystals and the formation of
corrosion products on top of the CPP (Fig. 54 left). At certain local areas the corrosive attack
after 24 h of salt spray testing lead to a completely dissolved zinc layer (Fig. 54 right). For
macroscopic investigations the average loss of a blank electrogalvanised sample is about 1
µm zinc / 10 h [159]. In both areas the zinc particles within the coating are mainly intact
while the steel zinc layer is strongly corroded. As the particles float within the coating they
have no conductive connection to the substrate underneath and can thereby not support the
cathodic protection properties of the zinc layer. The local formation of defects and therefore
conductive pathways through the coating seems to simplify the corrosive attack and leads to a
local increase of the corrosion rate by the creation of sheltered corrosion sites. These
corrosion sites can galvanically couple and act either as anode or cathode. The one acting as
cathode locally reduces oxygen to hydroxide and significantly increases the pH-value inside
the defect. With high pH-values the formation of stable zinc corrosion products is not possible
which could passivate the zinc layer and reduce the corrosion rate [160].
Results
93
0602A04581 3000 : 1 10 µm 0602A04585 3000 : 1 10 µm
Fig. 54. Cross sections of uniaxially formed (20% elongation, φv = 0.18), phosphatised and 24h exposed in
salt spray test corrosion protection primer. In both areas corrosion occurs while in some areas the strong
corrosive attack leads to a degradation of the complete zinc layer,
The FIB cross section supports the impressions of the metallographic cross section (Fig. 55).
Corrosion products are formed on top of the CPP, while the zinc particles are mainly intact.
Fig. 55. FIB cross section of uniaxially formed (20% elongation, φv = 0.18) corrosion protection primer
after 24h of exposition to a salt spray test. The images indicate the local detachment of the primer from
the sample surface and the corrosion of the zinc layer. Most of the embedded zinc particles show no
corrosive attack
The coating is detached from the steel substrate in large areas and forms caverns underneath
the coating. In the area where large caverns are formed the degradation of the zinc layer is
much faster (left side of the image) than in comparison to the smaller caverns (middle part of
the image). The corrosive attack of the zinc layer after 24 h in the salt spray chamber leads to
a complete removal of the typically tilted electrogalvanised zinc structure.
Results
94
5.4 In-Situ Electrochemical Impedance Spectroscopy during stretch forming
5.4.1 Forming characteristics of miniature stretching samples
The evaluation of the uniaxially stretched miniature samples by means of optical grid
measurements (GOM®) for 5 and 15 % elongation shows a homogeneous strain distribution in
the area of analysis (Fig. 56). The major strain ϕ1 induced by 15% elongation increases up to
0.16. Even at such high forming degrees, no necking in the middle of the sample becomes
obvious. The resulting changes in the minor and equivalent strain can be calculated in
accordance with equation 2 and 4 as -0.08 for ϕ2 and 0.15 for ϕv.
Furthermore, the forming characteristics of the specimen used were analysed by means of the
finite element simulation. For 5, 10, 15 and 20 % uniaxial elongation, the corresponding
equivalent strain ϕv is plotted in Fig. 57. In addition to that ϕv also shows a uniform
distribution here, especially in the middle of the sample where the working electrode of the
electrochemical setup is positioned.
Fig. 56. Major strain distribution of miniature stretching samples measured by GOM® grid evaluation for
5% and 15 % elongation. The sample shows a homogeneous forming especially in the middle where the
capillary will be placed
15%
ln
ϕ
1 5%
Results
95
Fig. 57. Finite element simulation of equivalent strain distribution for 5% (top left), 10% (top right), 15%
(bottom left) and 20% (bottom right) elongation of miniature stretching sample. Also the simulation
indicates a homogeneous forming in the middle of the sample
Both independent investigation methods indicate a uniform elongation with a homogeneous
strain distribution for up to 20% stretching.
The comparison of the strain values at 5% strain and 15% elongation shows that the FE
simulation and the GOM® grid evaluation concur. For 5% elongation, a ϕv of 0.07 was found
for the GOM® evaluation whereas 0.06 was determined for the FE simulation. At high
forming degrees, the equivalent strain was calculated to be 0.15 for the GOM® evaluation and
0.16 for the FE simulation.
5.4.2 In-Situ Electrochemical Impedance Analysis during tensile testing
Prior to the in-situ forming experiment, the electrochemical impedance spectra of the
uncoated and organically coated substrate in the non-stretched state were measured by means
of the capillary cell (Fig. 58).
Results
96
0.1 1 10 100 1000 10000 100000
1
10
100
1000
10000
100000
1000000
1E7
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
coated sample
uncoated sample
|Z| / Ω cm2
f / H
z
Phase / °
Fig. 58. Uncoated and primer-coated electrogalvanised steel sample in the unformed state, measured by
the in-situ EIS setup
The impedance spectrum of the corrosion protection primer sample could be analysed using
the equivalent circuit of a defect containing polymer coated metal (Fig. 59). The circuit
reflects the properties of the coating by the coating capacitance CC and the resistance of the
coating RC which is affected by nano and micropores. The double layer capacitance CDL and
the polarisation resistance RP display the properties of the metal electrode underneath the
coating that gets into contact with the electrolyte due to the penetration of the electrolyte
through the coating defects. The coated galvanised steel showed a capacitive behaviour in the
high frequency range between 100 KHz and 100 Hz dominated by the coating capacitance CC.
In the unformed state the delaminated coating area which is responsible for the double layer
capacitance is small. This leads to capacitance in the range of the weldable primer (some
nF/cm2). Furthermore, the metal electrode within the equivalent circuit is in series with the
high pore resistance of the coating which becomes only dominant at low frequencies due to
the parallel connection to the coating capacitance. It is thus not possible for this system to
separate the capacitance of the coating (CC) from that of the double layer capacitance (CDL).
At low frequencies the resistance of the coated sample becomes nearly 100 times larger
compared to the uncoated sample. In this range the two capacitive terms show extremely large
impedance values resulting in the dominating influence of the resistive terms. This leads to a
series connection of the coating resistance (RC), the polarisation resistance (RP) and the
uncompensated electrolyte resistance (RU) with the dominating influence of RC as the
delaminated area is small.
Results
97
During the measurement of the high absolute impedance values of the coated sample (up to
250 MΩ for a 1 mm in diameter capillary) the potentiostat showed some scattering when
switching the amplifier ranges. Furthermore, the potentiostat showed a phase shift in the low
frequency range of about 10° due to the high impedance. The uncoated galvanised steel was
used as reference indicating the typical behaviour of a metal in contact with an electrolyte
represented by an equivalent circuit without the coating elements represented by CC and RC
[96]. The measurement indicated that the capillary cell setup is suitable for both the uncoated
and even the high impedance of the coated metal substrate.
Fig. 59 Equivalent circuit of an organically coated metallic sample
The in-situ EIS evaluation during the stretching of the primer-coated substrate in borate buffer
is shown in Fig. 60. In the unformed state and for small forming degrees (≤5 % elongation),
the coating shows good barrier properties with an almost capacitive behaviour above 100 Hz
and a phase angle of 70 to 80°. In the low frequency range the coating shows resistive
behaviour by the combination of RU, RC and RP. At these small forming degrees the
delaminated area is still small and the system is dominated by the coating capacitance and the
high coating / pore resistance. With increasing strain, the barrier properties of the coating
decrease due to the formation of cracks and new electrolyte pathways. Furthermore, the
delaminated area at the coating substrate interface increases. These effects lead to a decrease
of the coating resistance (RC) and due to a rise in the contact area of the metal electrode with
the electrolyte to an increase of the polarisation resistance (RP), the double layer capacitance
(CDL). These result in an increase of the phase shift at high frequencies reaching 0° for an
equivalent strain of 0.22. At this forming degree the delamination and the defect formation of
the coating lead to a nearly complete breakdown of the barrier properties. The measured
Results
98
impedance and phase values thus become comparable to those of the bare zinc coated
substrate elongated to 25%. For all elongation degrees, the low frequency resistance at 0.1 Hz
shows a resistive behaviour by a combination of RU, RC and RP which decreases by two
orders of magnitude for high strain values. The small final impedance value indicates that the
dominating influence of the coating resistance in the equivalent circuit decreases and the
overall impedance is mainly dominated by the polarisation resistance.
0.1 1 10 100 1000 10000 100000
1
10
100
1000
10000
100000
1000000
1E7 0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
25% elongation
substrate 25% elong.
|Z| / Ω cm2
f / H
z
0,1 1 10 100 1000 10000 100000
-90
-80
-70
-60
-50
-40
-30
-20
-10
0 0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
25% elongation
substrate 25% elong.
|Z| / Ω cm2
f / Hz
Fig. 60. In-situ EIS evaluation during stretch forming of a Zn particle containing coating in borate buffer
(impedance (top) and phase shift (bottom))
The same in-situ experiment using 0.05 M NaCl as electrolyte shows similarly shaped
impedance spectra for the unstrained sample (Fig. 61) which can be interpreted by the same
Results
99
equivalent circuit as for the borate buffer experiment. Differences appear in the shift of the
capacitive behaviour to higher frequencies between 100 kHz and 1000 Hz and by total
impedance which is one order of magnitude smaller. These differences seem to arise as a
result of the strong corrosive effect of the chloride ions causing a fast attack on the passive
film on zinc and a faster delamination of the coating. Even at elongations of about 5-10 %, a
significant decrease in the coating barrier properties becomes obvious with a drop in the low
frequency resistance by one order of magnitude. At high forming degrees, the decay of
impedance becomes slower until it reaches about 3000 Ω cm2 at 1 Hz, which is again a value
close to that for the bare galvanised substrate at this forming degree.
1 10 100 1000 10000 100000
1
10
100
1000
10000
100000
1000000
0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
substrate 20% elong.
|Z| / Ohm cm²
f / Hz
1 10 100 1000 10000 100000
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
substrate 20% elong.
Phase / °
f / Hz
Fig. 61. In-situ EIS evaluation during stretch forming of a Zn particle containing coating in 0.05 M NaCl
(impedance (top) and phase shift (bottom))
Results
100
A complete evaluation cycle between 0 and 25 % elongation takes about 30 min during which
the sample is stretched and the impedance spectra are analysed. During the whole
measurement process the sample is in contact with the electrolyte and can undergo a corrosive
attack even without forming. For this reason, the impedance spectra of an unformed,
organically coated specimen were measured in borate buffer and 0.05 M NaCl over a period
of 30 minutes to obtain a reference measurement.
The low frequency impedance at 1 Hz is characterised by the series of the uncompensated
electrolyte, the coating resistance and the polarisation resistance. The electrolyte resistance
can be regarded as constant. While the coating resistance decreases during the degradation of
the coating and the polarisation resistance increases with the amount of delaminated area. The
coating resistance is in the initial state about 2 orders of magnitude larger than the polarisation
resistance. Thus, significant changes of the cumulated resistance should be mainly caused by
a change of the coating resistance. Therefore the impedance at 1 Hz was plotted as a function
of time and elongation (Fig. 62).
0 5 10 15 20 25
100
1000
10000
100000
1000000
1E7 0 10203040
|Z| / Ohm cm² at 1 Hz
elongation / %
forming in 0.05 M NaCl vs. elongation
forming in borate buffer vs. elongation
unformed in 0.05 M NaCl vs. time
unformed in borate buffer vs. time
t / min
Fig. 62. Decrease of the resistance at 1 Hz for primer-coated samples in relation to the forming degree and
for unformed samples in relation to the electrolyte contact time
Due to the corrosive properties and the delamination, the sample in contact with sodium
chloride electrolyte showed a slight decrease in the combined resistance after 30 min of
Results
101
immersion. The immersion in the chloride-free borate buffer leads to a nearly constant value
of the low frequency resistance over time.
Furthermore, Fig. 62 shows the decrease of the coating resistance as a function of the induced
strain. For both electrolytes, the forming induced changes of the impedance values are much
more significant than the corrosive and delaminating attack of the sodium chloride over time.
At high forming degrees, the coating loses almost completely its protective properties due to
the formation of defects. In the case of the NaCl electrolyte, the contribution of the corrosive
attack leads to a slightly increased degradation of the coating performance. Thus, the
corrosion and delamination of the coating and the forming induced defects cannot be clearly
separated, while the change in borate buffer is caused by the increase of the defect size and
density. For the in-situ measurements of the forming-induced defect formation and growth the
measurements should be carried out in chloride-free environments.
Results
102
5.5 Scanning In-situ EIS / Raman investigations of formed corrosion
protection primers
The simultaneous analysis of Raman and impedance spectra generates information about the
barrier properties and the formation and change of corrosion products. The combination of an
electrochemical capillary cell with a confocal Raman spectrometer allows the switching
between the two investigation methods on the same sample spot.
5.5.1 Scanning In-Situ EIS / Raman investigation of corrosion products on
plane strain formed primers
The new setup was used to investigate the barrier properties and the corrosion formation of a
plane strain formed (φv = 0.25) corrosion protection primer. The Raman spectra were taken
before and after 150 min of impedance analysis (electrolyte 0.5 M NaCl) with a HeNe Laser
(Fig. 63).
4000 3500 3000 2500 2000 1500 1000 500
0.00
0.02
0.04
0.06
0.08
0.10 before EIS
after EIS (150min)
Intensity
Wave number / cm-1
Fig. 63. Raman spectra of a plane strain formed (φv = 0.25) corrosion protection primer before and after
impedance measurements in 0.5 M NaCl
The CPP shows strong fluorescence in the complete range from 4000 to 250 cm-1 with two
maxima at about 2400 and 1100 cm-1 and were baseline corrected for further investigation.
Results
103
However, peaks of the organic binder matrix and zinc are still visible and are assigned in Tab.
4. The implemented zinc particles and / or the electrogalvanised zinc layer underneath the
coating which might be visible through the forming-induced defects already show the
formation of corrosion products before electrolyte contact. For both spectra (before and after
impedance measurement) the same signals appear indicating no significant change in the
formed corrosion products. The signal intensity on the other hand shows a small decrease in
intensity after the impedance measurement which might occur due to the dissolution and
formation of soluble corrosion products by the chloride containing electrolyte. For the
formation and precipitation of significant amounts of corrosion products the corroding time of
2.5 hours seems to be too short.
Tab. 4. Assignment of signals gained by Raman spectroscopy of a plane strain formed (φv = 0.25)
corrosion protection primer before and after impedance measurements in 0.5 M NaCl
Wave number / cm-1 Assignment
3055 C-H
2875 CH2
2280 N=C=O
1529
1452 C-C
1341 N-H
1145 Zn-O
750
681 O-H
595 Zn-O
485 Zn-O
In contrast to the Raman spectra the impedance measurement shows dramatic changes over
the investigation time (Fig. 64). Starting with an already low frequency impedance of 2000 Ω
cm2 due to the forming induced defects the value drops by nearly one order of magnitude
during the measurement. In the range from 100000 to 200 Hz the coating shows especially at
the beginning a strong capacitive behaviour with a phase shift of -65 ° decreasing down to
-20 ° over time. The double layer capacitance visible with its phase minimum at 60 Hz is
overlaid by the coating capacitance at the beginning and later on shows only a phase shift of
-20 ° due to the corrosive attack of the zinc layer.
Results
104
0,1 1 10 100 1000 10000 100000
10
100
1000
10000
0 min
11 min
22 min
33 min
44 min
55 min
99 min
154 min
|Z| / Ω cm2
f / Hz
0,1 1 10 100 1000 10000 100000
-80
-60
-40
-20
0
0 min
11 min
22 min
33 min
44 min
55 min
99 min
154 min
|Z| / Ω cm2
f / Hz
Fig. 64. Combined Electrochemical Impedance / Raman spectroscopy of plane strain formed (φv = 0.25)
corrosion protection primer showing a fast decrease of the impedance with time
The strong loss of barrier properties of the corrosion protection primer can only be explained
by the penetration of the electrolyte into the coating and consequently the corrosive
deadhesion along the binder / zinc particle interface which leads to the formation of new
electrolyte pathways. Such a deadhesion is a very fast process which can reach a speed of up
to 100 µm / h [161] and is thereby visible in the time scale of the experiment.
Combined Raman / EIS measurements provide complementary information about the barrier
and corrosion properties of coatings and enable a comprehensive view of the system.
Results
105
5.6 Modification of corrosion protection primers
5.6.1 Organo silane surface modification of zinc particles implemented in
corrosion protection primers
The impedance spectra and SEM images of formed corrosion protection primers indicate the
formation of defects within the forming process. These defects occur mainly at the binder /
zinc particle interface. One way to reduce the amount and size of defects is to increase the
adhesion force of the particle / binder interface. A surface modification of the zinc pigments
with a bi-functional organo silane should increase the stability of the interface. The organo
silane bonds on one side with the silane function to the oxide covered metal surface and with
the organic group to the binder.
Before modification, the particles were investigated concerning their surface morphology.
The FE-SEM images (Fig. 65) show a homogeneous size distribution between 2 and 6 µm. In
the detailed view the pigments show differences in the surface structure. Some show a very
smooth surface while others are covered by a rough, amorphous layer. This structure can lead
to either a better bonding of the binder to the metal surface or to a preferred defect area if the
amorphous layer breaks.
Fig. 65. FE-SEM image of zinc particles implemented into corrosion protection primers, showing a rough
and amorphous layer on top
In order to reduce the influence of the porous layer the particles were cleaned for one minute
with a strong alkaline cleaner before the surface modification for the reduction of the oxide
scale and for the preparation of a homogeneous, fresh surface (Fig. 66).
Results
106
Fig. 66. Schematic drawing of the zinc oxide scale on implemented particles before and after alkaline
cleaning
The FE-SEM images of the cleaned particles (Fig. 67) mainly show a much smoother surface
which shows only a nano rough structure. As the alkaline solution mainly removes the zinc
oxide the roughness of the particles seems to reflect the thickness of the zinc oxide layer on
top.
Fig. 67. Alkaline cleaned and organo silane modified zinc particles for implementation in corrosion
protection primers. Most of the rough oxide film on top of the particles is removed
The particles were modified with 3-Glycidyloxypropyltrimethoxysilane (Glymo) a bi-
functional organo silane. The silane part allows the bonding to a metal oxide surface [162-
164] while the epoxy function can bond to the organic binder system. The Si-O function
furthermore leads to the formation of a network covering the oxide surface (Fig. 68).
The pigments were immersed in a 1 % Glymo solution for 1h, afterwards rinsed with ethanol
and dried at room temperature or 60 °C for 1h.
Results
107
SiH3CO O
O
OCH3
OCH3
Zinc
ZnO
ZnOH
Si
O
R
HO OSi
R
O
O
R
1
ZnOHZnOH
Si
OH
O
O
HO OH
OH
+- H O
2
Fig. 68. Structure and network formation of 3-Glycidyloxypropyltrimethoxysilane (Glymo) on a zinc
surface
The layer formation was analysed on the dried powder using DRIFT (Diffuse Reflectance
Infrared Fourier Transform spectroscopy). The spectra of the modified particles dried at room
temperature and 60 °C are shown in Fig. 69.
Results
108
1600 1400 1200 1000 800
-0.005
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
Si-O-Si
vibration
dried at 25 °C
dried at 60 °C
Absorbance
Wave number / cm-1
Fig. 69. FT-IR DRIFT spectra of 3-Glycidyloxypropyltrimethoxysilane (Glymo) modified zinc particles
used in corrosion protection primers dried at room temperature and 60 °C. Only at 60 °C the formation of
Si-O-Si bond becomes visible while the epoxy ring seems already to be opened
The RT dried layer only shows a signal of the epoxy ring at 910 cm-1 while the 60 °C dried
layer shows the formation of Si-O-Si bond with two broad peaks at 1050 and 1200 cm-1. The
IR spectra reveal that the formation of a Si-O-Si network needs a curing temperature of 60 °C
while at this temperature the epoxy ring already seems to open. Due to the different curing
temperatures needed to get either the network formation or an intact epoxy ring the bi-
functionality of Glymo is limited. Anyhow the modified particles were implemented as
substitution for the unmodified zinc particles in a CPP.
5.6.2 In-Situ Electrochemical Impedance Analysis during tensile testing of
a modified primer
The Glymo modified particles were used as substitute for the ordinary zinc particles in the
corrosion protection primer. After application of the coating the formability was tested by
means of in-situ EIS during tensile testing.
In the unformed state and for small forming degrees (≤10 % elongation), the coating with the
particles dried at room temperature shows capacitively dominated behaviour above 100 Hz
with a phase angle of -70 to -80° (Fig. 70). With increasing strain, the barrier properties of the
Results
109
coating decrease due to the formation of cracks as indicated by the phase shift at high
frequencies, reaching 0° for an equivalent strain of 0.18. At this forming degree, the measured
impedance and phase values reach those of the bare zinc coated steel elongated to 25% (φv =
0,22). For all elongation degrees, the low frequency resistance at 0.1 Hz shows ohmic
behaviour and decreases by two orders of magnitude for high strain values. The forming
behaviour shows no significant differences to the unmodified coating.
0.1 1 10 100 1000 10000 100000
10
100
1000
10000
100000
1000000
1E7 0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
25% elongation
|Z| / Ω cm2
f / Hz
0.1 1 10 100 1000 10000 100000
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
25% elongation
Phase / °
f / Hz
Fig. 70. In-Situ impedance spectra during forming of 3-Glycidyloxypropyltrimethoxysilane (Glymo)
(dried at room temperature) modified corrosion protections primers, showing a fast reduction of the
barrier properties with increasing strain
Results
110
The coating with particles dried at 60 °C shows a similar behaviour as with the particles dried
at room temperature (Fig. 71). In the forming state between 0 and 10% the impedance above
100 Hz is capacitively dominated by a phase shift of -70 to -80°. With increasing strain the
coating shows crack formation indicated by the decrease of the phase angle shift to 0° in the
high frequency range and a strong drop of the ohmic resistance at low frequencies. Again at
high forming degrees the impedance spectrum of the blank steel substrate is measured and the
coating shows nearly no more protective properties.
0.1 1 10 100 1000 10000 100000
10
100
1000
10000
100000
1000000
1E7 0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
25% elongation
|Z| / Ω cm2
f / Hz
0.1 1 10 100 1000 10000 100000
-90
-80
-70
-60
-50
-40
-30
-20
-10
0 0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
25% elongation
Phase / °
f / Hz
Fig. 71. In-Situ impedance spectra during forming of 3-Glycidyloxypropyltrimethoxysilane (Glymo)
(dried at 60°C / 1h) modified corrosion protections primers, showing similar behaviour to the one with
particle dried at RT a fast reduction of the barrier properties with increasing strain
Results
111
The formed samples were further investigated by means of FE-SEM to follow the crack
formation at different elongation steps. The images (Fig. 72) show the formation of initial
small cracks in the range of some nm already at 5 % stretching. With increasing strain the
crack size and amount rise with cracks of up to 5 µm. The defects mainly occur at the
interface zinc particle / binder similar to the unmodified coating. The interface is obviously
still the weakest part of the system and cannot withstand the force applied to the coating by
the uniaxial forming procedure.
Fig. 72. SEM images of formed corrosion protection primers with 3-Glycidyloxypropyltrimethoxysilane
(Glymo) modified particles (organo silane dried at room temperature on top and at 60°C for 1h on the
bottom, 5% elongation left and 20% elongation right). A similar crack formation for both modified
particles is visible
Results
112
For a better comparison between the coatings and in order to avoid mistakes from the
scattering of measurement points the impedance spectra for each elongation step were fitted
by a polynomial curve. Afterwards the low frequency (ohmic) resistance was calculated from
the fit and is plotted in Tab. 5.
Tab. 5. Resistance at 0.1 Hz taken from non-linear impedance fit
CCP in
borate buffer
/ Ω cm²
CCP in
NaCl
/ Ω cm²
Modified CPP
(Particles dried
at RT)
in borate buffer
/ Ω cm²
Modified CPP
(Particles dried
at 60 °C)
in borate buffer
/ Ω cm²
0 % elongation 2211210 100168 4096909 4102455
5 % elongation 1169940 18568 826994 1653030
10 %
elongation 247783 5032 104900 206309
15 %
elongation 75052 859 53916 44212
20 %
elongation 34793 447 22611 12801
25 %
elongation 20096 273 14095 7261
The resistances calculated for the different forming steps are plotted in Fig. 73 versus the
different elongation steps. The direct comparison of the modified and unmodified coatings
show a slight increase of the barrier properties for the modified coatings at small elongations
due to the higher coating thickness (5µm instead of 3.5µm).
Results
113
0 5 10 15 20 25
0
1000000
2000000
3000000
4000000
unmodified
Organo silane dried at RT
Organo silane dried at 60 °C
|Z| / Ω cm2 at 1 Hz
elongation / %
Fig. 73. Comparison of corrosion protection primers with organo silane modified and unmodified zinc
pigments
The primer shows with particles dried at 60 °C a slightly better performance as the one dried
at room temperature. The differences between the modified coatings indicate that the
formation of a Si-O-Si network is of great importance for the performance of the organo
silane at the particle / binder interface while the reaction of the epoxy ring shows no visible
influence on the impedance spectra. The performance loss between the unmodified and the
modified CPP at elongations ≥5 % can be traced back to the fact that the cleaning procedure
leads to a much smoother surface of the particles and thereby to a worse connection between
the particle and the binder. Even the interface stabilisation by the particle modification can
not equalise this disadvantage.
Results
114
5.7 Corrosion model of formed and unformed corrosion protection primers
The impedance spectra of formed and unformed CPP in NaCl electrolyte show a fast
degradation of the coating barrier properties. This can be caused by the corrosive deadhesion
along the particle / binder interface which is a very fast process [161]. These new pathways
allow the electrolyte to penetrate the coating and thereby reduce the resistance and the
corrosion protection properties.
In the case of formed CPP the major loss of the corrosion protection results from forming-
induced defects that open new penetration ways for the electrolyte. Further on the coating
detaches from the electrogalvanised steel substrate underneath, which leads to the formation
of small caverns. These caverns act as initial sites for corrosive deadhesion along the binder /
zinc layer interface and lead thereby due to the oxygen reduction to a local, alkaline pH shift
(Fig. 74). Furthermore the coupling of defects leads to separated defects that act as anode and
cathode and can thereby also reduce the local pH value.
1/2O
2
+H O
2
2e
-
2OH
-
Zn
2+
Zinc
Steel
Delamination Delamination
Stretching
direction
HO / NaCl
2
O
2
Delamination along
the particle / binder interface
HO / NaCl
2
Corrosion protection
primer
Pretreatment
Delamination
Formed Area Unformed Area
Formation of defects and
delamination along
the zinc layer / binder interface
1/2O
2
+H O
2
2OH
-
Zn
2+
2e
-
2e
-
1/2O
2
+H O
2
2OH
-
Fig. 74. Model of the corrosion process occurring at formed and unformed corrosion protection primers.
In formed areas the defects act as initial corrosion sites and lead to a fast delamination and corrosion at
the binder / zinc interface. In the unformed state a fast delamination along the particle / binder interface
occurs which leads to new conductive pathways through the coating
This is also visible by the strong etching attack of the phosphate crystals which are not stable
in alkaline and acidic solutions. The shift in the pH accelerates local corrosion of the zinc
layer as the formation of protective precipitation layers is not possible [159, 160]. The
Results
115
Pourbaix diagram shows that at high pH-values (> pH 11) the formation of soluble zincate or
bizincate becomes possible while at low pH-values the formation of Zn2+ occurs (Fig. 75).
Fig. 75. Theoretical conditions of corrosion, immunity and passivation of zinc (taken from [160])
In chloride containing electrolytes as used for the salt spray testing the formation of insoluble
Simonekolleite (Zn5Cl2(OH)8·H2O) is possible at intermediate pH and high Cl- concentrations
according to eq. 43 [165, 166]. These would form a passivating precipitation layer which
reduces the corrosive attack.
5 Zn2+ + H2O + 8 OH- + 2 Cl- → Zn5Cl2(OH)8·H2O (eq. 43)
At high or low pH or at low Cl- concentrations the Zn5Cl2(OH)8·H2O becomes unstable and
dissolves under the release of chloride (Fig. 76). In case of the caverns that act as cathode the
high pH due to oxygen reduction and the reduced OH- diffusion leads to the formation of ZnO
or zincate. In the caverns acting as anodic sites the diffusion of chloride into them is hindered
and thereby falls below the critical concentration necessary for the formation of
Simonkolleite. In both cases the formation of passivating layers is not possible.
Results
116
Fig. 76 Stability diagram of simonkolleite in aerated solutions with varying pH and Zn2+
concentration of
0.1 M at 25 °C (taken from [165])
The high corrosion rate is visualised by the strong local corrosion found already after 24
hours of salt spray testing in the FE-SEM cross section. Furthermore, the FIB cross sections
show a fast dissolution of the zinc layer and precipitation layers of corrosion products only on
top of the CPP where the OH- concentration is lowered by the exchange of the electrolyte.
Results
117
5.8 In-situ QCM / Raman investigations of the inhibitor adsorption on
metals
Organic and inorganic inhibitors are widely used to improve the protective properties of
organic coatings. Inhibitors can act either by the formation of barrier layers on top of the
surface or by precipitation of low soluble products. In both cases the detection of corrosion
and precipitation products, the kinetic of the inhibiting effect and the concentration
dependence are of great interest. The newly developed in-situ QCM / Raman cell allows the
simultaneous detection of the frequency change of a sample which is caused by a dissolving
or precipitation reaction and the measurement of the Surface Enhanced Raman Spectrum of
the sample (see 4.3). In this case the adsorption of the inhibitor Mercaptobenzothiazol (MBT)
onto a silver and a gold sample was studied as a model system but the method can be
expanded to all SERS active substrates.
5.8.1 Ex-situ SERS analysis of Mercaptobenzothiazol (MBT) adsorption
Prior to the in-situ experiments the spectral properties of Mercaptobenzothiazol (MBT) were
investigated ex-situ. First tests on a flat non-SERS active silver substrate immersed in an
MBT / ethanol solution showed no measurable peaks. In Fig. 77 the powder spectrum and the
spectrum of a SERS active silver sample after immersion into a 10-3 M MBT / ethanol
solution are shown. Nearly all peaks that are found in the powder spectrum are also present in
the SERS spectrum, mainly with slight changes in the peak position and intensity. The most
significant peak at 1254 cm-1 in the spectrum b) strongly decreases in intensity in the SERS
spectrum while the one at 1395 cm-1 shows a strong increase. Further differences are the
peaks of the benzene ring at 397 cm-1, 1076 cm -1 and 3070 cm-1 which show much higher
intensities in the powder spectrum than in the SERS spectrum. This might be a consequence
of the preferred orientation of the MBT molecule during the adsorption. Known from
literature is that the non-cyclic sulphur atom bond to the sample which leads to a
perpendicular orientation of the molecule to the surface and can thereby lead to a reduction of
the Raman absorption of the benzene bonds. In contrast to that the powder displays a random
distribution of molecule orientations.
In Tab. 6 the peak positions and assignment of the powder and SERS spectrum are given.
Results
118
400 600 800 1000 1200 1400 1600
0
2000
4000
6000 SERS spectrum of MBT (a)
MBT powder (b)
(a)
Intensity
Wave number / cm-1
397
501
607
663
705
1031
1076
1133
1254
1273
1320
1460
1498
1585
1600
395
508
602
717
1012
1135
1246
1281
1315 1395
1459
1564
1590
865
1010
(b)
527
1425
2400 2800 3200 3600 4000
0
500
1000
1500 SERS spectrum of MBT (a)
MBT powder (b)
Intensity
Wave number / cm-1
3070
2645
(a)
(b)
Fig. 77 Ex-situ SERS spectrum of (a) MBT adsorbed on the quartz crystal-based SERS-active substrates
and (b) normal Raman spectrum of MBT powder, showing the amplification of the spectrum by the
change of the substrate structure
Tab. 6. Assignments of the Raman peaks of MBT powder and MBT adsorbed on quartz crystal-based
SERS-active substrates
Peak position /cm-1
Powder SERS
Assignment
3070 3070 CH stretching in Bz ring
2645 SH stretching
1600 1590 Bz ring stretching
1585 1564 Bz ring in plane stretching
1498 N-C=S stretching
1460 1459 C-C stretching
1425 1395 NCS ring stretching
1320 1315 NH bending
1273 1281 CH in-plane bending
1254 1246 CH bending or CN stretching
1133 1135 CH in-plane bending
1076 Bz ring or SCS antisymmetric stretching
1031 CH in-plane deformation
1011 1013 C-C-C bending
867 CH out-of-plane bending
705 717 C-S stretching
663 NH deformation
607 602 CS stretching in heterocyclic ring system
527 Bz ring deformation
501 508 Bz ring deformation
397 396 Bz ring deformation
Results
119
5.8.2 Combined in-situ Raman / QCM measurement of the adsorption of
MBT on silver
The in-situ MBT adsorption study started with a reference measurement of pure ethanol
running through the setup (first spectrum in Fig. 78). After the system reached the equilibrium
the injection of MBT happened and spectra were taken every 10s in the initial state. The
pumping and tube system lead to a delay of nearly 40s before the first MBT reaches the
sample and forms a peak at 1395 cm-1 (Fig. 78). With an increasing measurement period the
intensity of the peak rises and new peaks at 396, 717, 1315, 1564 and 1593 cm-1 appeared.
400 600 800 1000 1200 1400 1600
0
1000
2000
3000
4000
∗
∗
∗
∗
∗
150S
140S
130S
120S
110S
100S
90S
80S
70S
60S
Injection
Intensity
Wave number cm-1
717
1399
1564 1593
EtOH
50S
1459
1013
∗
400 600 800 1000 1200 1400 1600
0
1000
2000
3000
4000
∗
∗
∗
∗∗
150s
180s
210s
240s
270s
300s
330s
360s
Intensity
Wave number / cm-1
717
1399
1564
1593
1459
(a) 0-100 seconds (b) 100-240 seconds
Fig. 78. In-situ SERS spectra of adsorbed MBT as a function of adsorption time: (a) from the beginning to
100 s with an interval of 10 s; (b) from 100 to 240 s with an interval of 30 s. The spectra show an increase
of absorbed MBT with time
The intensity of a Raman peak is proportional to the number of molecules illuminated by the
laser and thereby reflects the amount of molecules adsorbed to the SERS substrate. Hence the
peak at 1395 cm-1 was chosen for the MBT adsorption and the ethanol peak at 887 cm-1 as
reference for controlling the focus of the system.
The intensity shows after the initial delay of about 1 min a strong increase which takes nearly
2 min until it passes over to a plateau value for longer times (Fig. 79). The ethanol peak stays
nearly constant over the whole experiment, indicating a focused sample and a uniform laser
power.
Results
120
01234567
500
1000
1500
2000
2500
Intensity
Time (min)
I887 ethanol
I1395 MBT
Fig. 79. Evolution of Raman intensity of ethanol and MBT adsorbed on SERS-active silver as a function of
time, indicating a constant signal for the ethanol while the MBT signal rises for some minutes until it
reaches a plateau value
The combined plotting of the Raman and QCM data (Fig. 80) shows a similar delay before
the MBT adsorption starts. In the beginning both show a steep slope with a change after about
one minute. In the case of the Raman measurement the turning point indicates the change for
reaching the final plateau while for the QCM measurement the linear decrease just gets a
different slope and further decreases linearly.
8 9 10 11 12 13 14 15
-25
-20
-15
-10
-5
0
500
1000
1500
2000
2500
3000
Intensity at 1395 cm-1
∆f / Hz
Time / min
Injection point Starting point
Fig. 80. In-situ QCM / Raman measurement of MBT adsorption on SERS-active silver. The Raman signal
reaches a plateau value after a few minutes while the QCM signal decreases further due to the ageing
effects of the silver substrate
Results
121
As the frequency shift in the QCM measurement continued even after the Raman intensity
became constant the data acquisition was continued until a plateau value was also observed
(Fig. 81). The frequency shift became constant after nearly 48 hours with a final value of
-200 Hz.
Fig. 81. Long time QCM experiment of the MBT absorption on SERS-active silver to investigate the
maximum frequency shift. After about 24 hours the systems become stable with a frequency shift of about
200 Hz
The long adsorption time indicated by the QCM experiment is contrasts with the Raman data
and the calculation for the frequency shift of a monolayer of MBT (see 5.8.4). The strong
decrease might occur by some ageing effects that occur on the silver surface in the ethanol /
MBT solution.
5.8.3 Combined in-situ Raman / QCM measurement of the adsorption of
MBT on gold
The adsorption of MBT on a silver layer shows discrepancies between the Raman and the
QCM data which might arise from aeging effects of the silver surface. These effects could be
avoided by using gold as SERS substrate. Therefore gold SERS active samples were prepared
in the same manner as the silver ones.
The SERS spectra of MBT on a silver and gold based sample are compared in Fig. 82. Both
spectra show similar peak formations of the adsorbed MBT which allows to analyse the gold
spectrum in a similar way as already undertaken with the silver based one.
Results
122
400 500 600 700 800 900
0
200
400
600
800
(b)
Intensity
Wave number / cm-1
507 602
716
860
525
390
(a)
900 1000 1100 1200 1300 1400 1500 1600 1700
0
1000
2000
3000
4000
5000
6000
Intensity
Wave number / cm-1
1010
1132
1241
1276
1394
1457 1565
1592
1316
b
a
Fig. 82. Spectra of MBT adsorbed on SERS-active gold (a) and on silver (b). Both substrates show a
strong amplification of the characteristic peaks
During the in-situ experiment the Raman signal at 1395 cm-1 (MBT) again showed a delay of
a few seconds before the MBT solution reached the cell (Fig. 83). Then a strong increase of
the peak intensity for about 6 minutes took place. Finally the intensity reached a constant
plateau value. Also in this experiment the ethanol reference value stayed constant over the
complete measurement time. The adsorption kinetic is in the case of gold three times slower
as on silver where the plateau value is reached after 2 minutes.
0 3 6 9 12 15
0
400
800
1200
I885 ethanol
I1395 MBT
Intensity
time / min
Fig. 83. Evolution of Raman intensity of ethanol and MBT adsorbed on SERS-active gold as a function of
time, indicating a constant signal for the ethanol while the MBT concentration grows for some minutes
until it reaches a plateau value
Results
123
The combined QCM and Raman data show in the case of gold a consistent behaviour with an
inverted slope of the two curves, indicating that both analytical tools enable the correct
identification of the kinetic of the adsorption process (Fig. 84). In contrast to the silver
experiment also the QCM data show a plateau after 6 minutes at about -10 Hz indicating that
no further reactions on the gold surface take place. As already mentioned the gold based
experiment shows a three times slower rate of the MBT adsorption than for silver which
might be due to a different adsorption mechanism on gold surfaces which was already
discussed in the literature [167].
051015
-12
-10
-8
-6
-4
-2
0
0
200
400
600
800
1000
1200
1400
∆f / Hz
time / min
Intensity
Fig. 84. In-situ QCM / Raman measurement of MBT adsorption on SERS-active gold. Both signals reach
a plateau value after a few minutes and show a consistent curve progression
A comparison of the gold and silver based SERS active surfaces shows for the adsorption
kinetics measured by Raman spectroscopy comparable results, while the QCM data
completely differ. The QCM measurement gives integral information for the mass change of
the surface while the selected Raman peak is sensitive to the adsorption of the MBT. It seems
that the contact of the silver surface with the MBT solution leads to further changes of the
SERS substrate which are not detected by the Raman measurement. Therefore the optimal
substrate for an in-situ investigation is a gold coated quartz.
Results
124
5.8.4 Theoretical calculation of the frequency shift by a monolayer
It is known from literature that Mercaptobenzothiazol forms a monolayer during the
adsorption process on the substrate. The frequency shift of a monolayer can be calculated
theoretically and compared to the experimental data of the adsorption on silver and gold.
For the calculation of the frequency shift the amount of molecules that adsorb per cm2 is
necessary which can be calculated from the three dimensions of the MBT molecule. The ideal
molecule structure was simulated with Gaussian [168] and is shown in Fig. 85 indicating a
completely flat structure.
S
N
SH
Fig. 85. Chemical and 3D structures (simulated with Gaussian) of Mercaptobenzothiazol indicating MBT
as a planar molecule
The length, width and height of the molecule were calculated to be l = 7.6 Å, w = 5.0 Å and h
= 2.0 Å. These dimensions can be used to describe the MBT molecule as a brick which can
either adsorb with the front face (A1), the long side (A2) or planar (A3) on the sample surface
(Fig. 86). The different orientations of the adsorbed molecule lead to a different mass load on
the surface and thereby to differences in the frequency shift.
Fig. 86. Schematic drawing of the MBT adsorption planes taken from the 3D simulation
The covered surface areas for the different absorption orientations were calculated to be:
2152201010
11011010102105 cmmmmhwA −−−− ⋅=⋅=⋅⋅⋅=⋅= (eq. 44)
A1
A3
A2
7.6 Å
5.0 Å
2.0 Å
Results
125
215
21052.1 cmhlA −
⋅=⋅= (eq. 45)
215
3108.3 cmlwA −
⋅=⋅= (eq. 46)
With the assumption that the molecules adsorb in a closed package without any free volume
the number of molecules and with the Avogadro constant (6.022 · 1023 molecules / mol) and
the molar mass of MBT (167.24 g / mol) the mass for a monolayer per cm2 can be calculated
(done here exemplarily for orientation A1) :
Number of molecules per cm²:
2
15
1
1101
1
cm
molecules
A
N⋅== (eq. 47)
Mol per cm²:
2
9
23
1
11066.1
10022.6 cm
mol
mol
molecules
N
n−
⋅=
⋅
= (eq. 48)
Mass per cm²:
22
9
11 27724.1671066.1 cm
ng
mol
g
cm
mol
Mnm MBT =⋅⋅=⋅= − (eq. 49)
Sauerbrey [121] shows that the frequency change of the QCM is proportional to the mass load
change of the quartz crystal. Using this equation with the known mass change for a monolayer
MBT and the proportional factor cf for the used quartz crystal allows the calculation of the
theoretical frequency change for the different orientations (A1, A2 and A3).
mcf f∆⋅−=∆ (eq. 29)
ng
cmHz
cf
2
056,0 ⋅
−= (eq. 50)
Hz
cm
ng
ng
cmHz
fA5,15277056.0 2
2
1−=⋅
⋅
−=∆ (eq. 51)
Results
126
HzfA3.10
2−=∆ (eq. 52)
HzfA04.4
3−=∆ (eq. 53)
This calculation shows that the frequency change of a monolayer MBT is found for the gold
substrate in the range of the experimental data. The orientation A2 already discussed in
literature [167] as the preferred adsorption orientation comes close to the measured value. The
results show that the frequency shift found for the adsorption on silver with about 200 Hz has
to be influenced by parallel processes which were already indicated by the slope change
observed after a few seconds.
Results
127
5.9 Forming and water uptake of polyelectrolyte layers
5.9.1 Molecular composition of polyelectrolytes
Polyelectrolytes are polymers with ionisable groups that can dissociate in a polar solvent to be
either positively or negatively charged. Due to the mixture of oppositely charged polymers the
formation of networks connected by coulomb forces between the dissociated groups is
possible. This allows a layer by layer formation of polymers which can attach to a charged
sample surface. Due to the non-covalent bonding between the groups highly flexible layer
systems are formed. In the case of e.g. carboxylic and amino groups they can be transferred
by heating into covalent bonded systems [36, 169]. The layer formation of oppositely charged
polyelectrolytes with carbonyl and amine functions before heating is schematically given in
Fig. 87.
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
NH
3+
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
COO
-
Fig. 87. Schematic drawing or the layer formation of oppositely charged polyelectrolytes
by coulomb forces
Equation 54 shows the amidation reaction of the carboxylate and ammonium function.
R-NH3+ + R-COO- → R-NH-CO-R + H2O (eq. 54)
For the further forming investigations especially by Wolfgang Bremser and Oliver Seewald
(University of Paderborn) synthesised polyelectrolytes were used. They consisted of the
functional groups already mentioned above. The schematic composition of the molecules is
Results
128
given in Fig. 88. In contrast to the commonly used layer by layer formation the two
components were already mixed during application.
O
O
C
4
H
9
O
C
4
H
9
OCOOH COOH
4-5
N N
H
2
N
H
2
N
NH
2
NH
2
n
Fig. 88. Schematic drawing of the polyelectrolytes (polyanion left and polycation right) used
for forming experiments
The amount of polycation groups (amine functions) was thus chosen twice as high as the
number of polyanion groups (carboxylic functions), Butlyglycol and Methylisobutylketon
were used as solvent. The layers were applied on electrogalvanised steel by dip-coating and
afterwards dried at room temperature, 130 °C and 230°C for 2h using standard or nitrogen (N2
stream) atmosphere.
5.9.2 FT-IR analysis of the curing state of polyelectrolyte layers
The influence of different temperatures and the storing of the coated samples on the curing
state were monitored by Infrared Attenuated Total Reflection spectroscopy (ATR).
Furthermore, the effect of modified atmospheres on the coating properties was studied. First
of all, a freshly prepared PET sample which was dried for 24h at room temperature was
compared to the solvent spectra (Fig. 89). The spectra indicate that already after 24h hours of
drying most of the solvent is evaporated which is furthermore indicated by a coating, that is
no longer sticky. The PET system shows characteristic peaks at 830 (C-O), 1037 (C-O), 1180
(C-O), 1246 (-O-H), 1458 (C-H), 1508 (-NH3+), 1730 (C=O), from 2872 – 2958 (CH2 and
CH3) and from 3080 – 3900 cm-1 (-OH).
Results
129
4000 3500 3000 1750 1500 1250 1000 750
0.000
0.025
0.050
0.075
0.100
0.125
0.150 dried at room temperature for 24 h under air
Butylglycol
Methylisobutylketon
Absorbance
Wave number / cm-1
Fig. 89. Comparison of freshly prepared PET complex dried at room temperature for 24 hours vs. the
used solvents (Butylglycol and Methylisobutylketon). The spectra show the nearly complete removal of the
solvents after 24 hours
The spectra of PET films dried at room temperature for 24 hours and 30 days were compared
with respect to their curing states. Both coatings showed the characteristic peaks already
mentioned before. Furthermore, the coatings which were dried for 30 days showed the
formation of new peaks at 1419 (N-H stretching), 1577 and 1604 cm-1 (amide peaks) (Fig.
90). This indicates that the cross linking reaction between the anionic and cathodic groups
took already place after the evaporation of the solvent and doesn’t necessarily need a thermal
step.
Results
130
4000 3500 3000 1750 1500 1250 1000 750
0.000
0.025
0.050
0.075
0.100
0.125
0.150 dried at room temperature for 24 h under air
dried at room temperature for 30 d under air
Absorbance
Wave number / cm-1
Fig. 90. Comparison of PET films dried at room temperature for 24 hours and for 30 days. The spectra
show the formation of amide bonds after 20 days of storage
The coatings dried at room temperature, 130 °C and 230 °C show all the characteristic peaks
of the PET complex (Fig. 91). Additional peaks appear for the two coatings cured at higher
temperatures which show the formation of amide bonds. With increasing temperature also this
peak intensity increases indicating a higher cross linking of the applied film. It seems that at
high temperatures of 230 °C the coatings degrade by a saponification according to eq. 54 and
leads to more reactive groups.
R1-COOR2 + H2O → R1-COOH + R2-OH (eq. 55)
This is visible by the further increase of the amide peak which needs more carboxylate groups
to form amid bond. The necessary water for the reaction can be provided either from humidity
or from the water formed by the amide formation.
Results
131
4000 3500 3000 1750 1500 1250 1000 750
0.00
0.05
0.10
0.15
0.20
dried at room temperature for 24 h under air
dried at 130 °C under air
dried at 230 °C under air
Absorbance
Wave number / cm-1
Fig. 91. Comparison of polyelectrolyte coated samples dried at different temperatures. Rising curing
temperature leads to an increase in the number of amide bonds
For the film preparation it is of great interest if the curing under standard atmosphere leads to
a degradation of the applied coating. Fig. 92 and Fig. 93 show the comparison of films cured
at 130 °C and 230 °C under air or nitrogen atmosphere. For the medium temperature the
coatings show no difference in peak formation and thereby no influence of the curing state or
higher decomposition in one of the atmospheres. In contrast to that the high temperature cured
systems show differences for the one cured under air atmosphere as already mentioned before.
The amide peak (1604 and 1577 cm-1) for the air cured system is much higher than for the one
under nitrogen. This is a result of the decomposition of the ester functions within the coating
by the formed water. In case of the N2 system a continuous nitrogen stream flowed over the
samples and immediately removed the developed water. Thus the decomposition reaction
cannot proceed. The complete list of the peaks appearing for the different curing conditions is
given in Tab7.
Results
132
4000 3500 3000 1750 1500 1250 1000 750
0.00
0.05
0.10
0.15
0.20 dried at 130 °C under air
dried at 130 °C under nitrogen
Absorbance
Wave number / cm-1
Fig. 92. Comparison of polyelectrolyte coated samples dried at 130 °C (under air and N2) and room
temperature. The curing conditions do not lead to changes in the polyelectrolyte film
4000 3500 3000 1750 1500 1250 1000 750
0.00
0.05
0.10
0.15
0.20 dried at 230 °C under air
dried at 230 °C under nitrogen
Absorbance
Wave number / cm-1
Fig. 93. Comparison of polyelectrolyte coated samples dried at 230 °C (under air and N2) and room
temperature. The curing under air increases the amount of formed amide bonds
Results
133
Tab. 7. Assignment of the absorption peaks of the PET complexes dried at different temperatures
PET dried
at RT (24h)
/cm-1
PET dried
at RT (30d)
/cm-1
PET dried
at 130 °C
/cm-1
PET dried
at 230 °C
(under air)
/cm-1
PET dried
at 230 °C
(under
nitrogen)
/cm-1
Vibration
3080 - 3900 3080 - 3900 3080 - 3900 3080 - 3900 3080 - 3900 O-H
2872 - 2958 2872 - 2958 2872 - 2958 2872 - 2958 2872 - 2958 CH2 and CH3
1730 1730 1730 1730 1730 C=O
1604 1606 1600 Amide
1577 1577 1577 1570 Amide
1509 1509 1509 1509 1509 -NH3+
1457 1457 1457 1450 1448 C-H
1419 1419 1419 1419 N-H
1246 1246 1246 1236 1239 -O-H
1182 1182 1182 1180 1181 C-O
1043 1037 1043 1083 1037 C-O
833 833 833 832 832 C-O
5.9.3 Hardness of polyelectrolyte layers
From the FT-IR measurements it is already visible that variations in curing temperature lead
to differences in the forming of amide bonds and thereby to cross linking of the polymer film.
Changes in the number of cross links should influence the hardness of the coating. The
coating hardness was studied at curing temperatures of 130, 180 and 240 °C under air
atmosphere (Fig. 94).
Results
134
120 140 160 180 200 220 240
110
120
130
140
150
160
170
Hardness / N mm-2
T / °C
Fig. 94. Hardness of polyelectrolyte layers cured at different temperatures [taken from I. Klueppel, O.
Seewald, R.Regenspurger, W. Bremser, G. Grundmeier, to be submitted]
With increasing curing temperature the coating hardness rises from 115 N/mm2 at 130 °C to
165 N/mm2 at 240 °C. The temperature dependence of the hardness underlines the FT-IR
measurements and indicates the differences in the amount of amide bonds formed during the
curing process. This should also influence further physical properties of the coating like
formability and water uptake.
Further on the time dependent FT-IR measurements for room temperature cured
polyelectrolyte layers show an increase of amide bonds even without heating. The ability of
post curing of polyelectrolyte layers was also studied by hardness measurements over time at
a layer cured at 130 °C (Fig. 95).
Results
135
0 50 100 150 200 250 300
100
110
120
130
140
150
160
Hardness / N mm-2
t / h
Fig. 95. Post curing of polyelectrolyte layer cured at 130 °C [taken from I. Klueppel, O. Seewald,
R.Regenspurger, W. Bremser, G. Grundmeier, to be submitted]
The already cross linked system shows a further increase of hardness from about 107 N/mm2
to about 155 N/mm2 after 300 hours. The final hardness correlates with a coating cured at
220 °C (compare Fig. 94). The increase shows that the curing at 130 °C still leaves some
reactive groups which can lead to a further cross linking with time.
5.9.4 Water uptake of polyelectrolyte layers dried at different
temperatures
The protective properties of organic coatings against corrosion are based mainly on barrier
properties blocking the transport of corrosive media. These barrier properties depend on the
organic structure of the polymer and also on the cross linking state of the macro molecules
within the coating. With increased cross linking the water uptake decreases and the barrier
properties increase. Electrochemical Impedance Spectroscopy allows to follow the water
uptake of organic coatings based on the Brasher-Kingsbury equation (eq. 56) using the
coating capacitance as a function of immersion time in an electrolyte [28, 97, 98, 101, 170].
100
log
loglog
2
0⋅
−
=
OH
tCC
ε
ϕ
(eq. 56)
Results
136
With ϕ the water uptake in percent, Ct the coating capacitance at time t, C0 the initial coating
capacitance and OH2
ε
(78.5) the dielectric constant of water.
The impedance spectra of PET complex coated samples dried at different temperatures and
under different atmospheres were recorded for some hours and afterwards used for the water
uptake calculation.
The impedance spectrum of a PET coated sample dried at room temperature shows in the high
frequency range from 100 kHz to 100 Hz a capacitive behaviour while at low frequencies the
ohmic resistance of the coating is predominant (Fig. 96). After 20 hours of immersion in a
borate buffer solution a change in barrier properties became visible by an increase of the
resistance by nearly one order of magnitude in the low frequency range. Furthermore, the
capacitively dominated range extended down to 10 Hz while the impedance in this range is
significantly lower than in the initial state.
0,1 1 10 100 1000 10000 100000
100
1000
10000
100000
1000000
1E7
-80
-60
-40
-20
0
Impedance t = 0h
Impedance t = 20h
|Z| / Ω cm-2
f / Hz
Phase t = 0h
Phase t = 20h
Phase / °
Fig. 96. Impedance spectra of PET complex dried at room temperature and immersed in borate buffer for
different times, indicating a strong change in barrier properties over time
The PET coating dried at 130 °C shows a capacitive behaviour down to 10 to 1 Hz (Fig. 97).
Only at very low frequencies a change in the ideal coating behaviour becomes visible. Over
the immersion time the coating resistance only slightly increases indicating a well cross
linked coating.
Results
137
1 10 100 1000 10000 100000
100
1000
10000
100000
1000000
1E7
1E8
1E9
-80
-60
-40
-20
0
Impedance t = 0h
Impedance t = 20h
|Z| / Ω cm-2
f / Hz
Phase t = 0h
Phase t = 20h
Phase / °
Fig. 97. Impedance spectra of PET complex dried 130 °C under standard atmosphere and immersed into
borate buffer for different times, indicating only slight changes of the barrier properties over time
The coating dried at 230 °C shows an ideal behaviour over the complete frequency range with
a phase shift of 90 ° (Fig. 98). The immersion for 9 hours leads to no significant changes in
the coating properties indicating a highly cross linked polymer.
0,1 1 10 100 1000 10000 100000
100
1000
10000
100000
1000000
1E7
1E8
1E9
1E10
-80
-60
-40
-20
0
Impedance t = 0h
Impedance t = 9h
|Z| / Ω cm-2
f / Hz
Phase / °
Phase t = 0h
Phase t = 9h
Fig. 98. Impedance spectra of PET complex dried at 230°C under standard atmosphere and immersed
into borate buffer for different times, indicating nearly no change of the barrier properties over time
The coating dried at 230 °C under nitrogen atmosphere shows differences in the low
frequency resistance in comparison to the one dried under air atmosphere. At frequencies
Results
138
below 10 Hz the coating changes from a purely capacitive behaviour to one with a clearly
ohmic behaviour. Long time immersion leads to only slight changes in the phase shift ranging
to higher frequencies and a less dominant ohmic behaviour. The barrier property indicated by
the low frequency resistance seems not to be affected.
0,1 1 10 100 1000 10000 100000
1000
10000
100000
1000000
1E7
1E8
1E9
1E10
-80
-60
-40
-20
0
Impedance t = 0h
Impedance t = 20h
|Z| / Ω cm-2
f / Hz
Phase t = 0h
Phase t = 20h
Phase / °
Fig. 99. Impedance spectra of PET complex dried at 230°C under nitrogen and immersed into borate
buffer for different times, indicating nearly no change of the barrier properties over time
The water uptake calculated from the time dependent impedance spectra taken for the
differently cured coatings shows significant differences (Fig. 100). The coating dried at room
temperature shows a high water uptake of about 19.5 % indicating just a low cross linking
with only weak barrier properties. With increasing curing temperature a clear change in the
coating properties becomes visible. At 130 °C the water uptake strongly decreases to 5.5 %
and the initial slope increases very strongly leading to a quicker achievement of the plateau.
The further increase of temperature to 230 °C leads to even less water uptake in the range of
1.5 % with a similar slope like the one dried at 130 °C.
Results
139
0 5000 10000 15000 20000 25000 30000 35000
0
1
2
3
4
5
6
15
16
17
18
19
20
dried at room temp.
dried at 130 °C under air
dried at 230 °C under air
water uptake / %
t / s
Fig. 100. Water uptake of a polyelectrolyte film dried at room temperature, 130°C and 230°C under air.
The water uptake decreases with increasing curing temperature and cross linking of the system
The comparison of the two coatings dried at 230 °C but under different atmospheres (air and
nitrogen) shows nearly no differences in the slope and total amount of water uptake (Fig.
101).
0 5000 10000 15000 20000 25000 30000 35000
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
dried at 230 °C under air
dried at 230 °C under nitrogen
water uptake / %
t / s
Fig. 101. Water uptake of a polyelectrolyte film dried at 230°C under air and nitrogen. The different
atmospheres show no influence on the water uptake of the system
The water uptake for the systems cured under air atmosphere was plotted versus curing
temperature and is given in Fig. 102.
Results
140
0 50 100 150 200 250
0
2
4
6
8
10
12
14
16
18
20
Water uptake / %
T / °C
Fig. 102. Water uptake versus curing temperature for PET system cured under standard atmosphere. The
water uptake follows an exponential decay with increasing curing temperatures
The reduction of the water uptake with increasing temperature follows a 1st order exponential
decay indicating a strong cross linking during drying even for medium temperatures.
5.9.5 FE-SEM investigation of formed polyelectrolyte layers
The formed and unformed polyelectrolyte coated samples were analysed by FE-SEM to
identify the formation of defects. The images of the unformed samples show a formation of
compact films that follow the surface roughness of the electrogalvanised substrate (Fig. 103).
The sample dried at room temperature (24h) shows certain pores within the coating while the
two other films are free of defects. The defect formation might occur due to the brittle
properties of the coating if the solvent is evaporated while no bond formation between the
polymer chains happens.
Results
141
Fig. 103. SEM image of unformed polyelectrolyte coated samples dried under standard atmosphere at
room temperature (top left), at 130 °C (top right) and 230 °C (bottom). At room temperature the
formation of certain pores is visible while at higher temperatures smooth films are formed
The elongated samples show formation of defects perpendicular to the direction of the major
strain (Fig. 104). There were no favourite defect sites as in the pigmented coatings and the
defects are statistically distributed over the complete sample.
Results
142
Fig. 104. SEM image of 20 % elongated polyelectrolyte coated samples dried under standard atmosphere
at room temperature (top left), at 130 °C (top right) and 230 °C (bottom). For all curing temperatures the
forming leads to defects within the coating
The defects size allows a ranking of the different curing conditions with 130 °C on top
showing on average the smallest defects, the RT dried sample with larger defects and finally
the sample cured at 230 °C. The last one shows a homogeneous distribution of large defects
that would dramatically reduce the barrier properties of the coating.
5.9.6 In-situ Electrochemical Impedance Spectroscopy during stretch
forming of polyelectrolyte layers
The in-situ Electrochemical Impedance Spectroscopy during the stretch forming of the
polyelectrolyte coated samples were performed with similar miniature stretching samples as
already described for the testing of corrosion protection primers. The samples were cut out by
a laser, three step solvent cleaned and afterwards alkaline cleaned by a coil coating cleaner.
The coating was applied by dip coating with a home-built dip coater (dipping speed 2 mm/s).
Results
143
The samples were dried / cured at different temperatures (RT 24 h drying, 130 °C and 230 °C)
and under standard or nitrogen atmosphere.
5.9.6.1 Formability of a polyelectrolyte layer dried at room temperature
The room temperature dried polyelectrolyte coating shows in the unformed state a broad
capacitive behaviour from 100000 Hz down to 10 Hz (Fig. 105). The low frequency resistance
reaches 1 MΩ / cm2 indicating high barrier properties. At an elongation of 5% the separation
into two time constants becomes visible. In the high frequency range the capacitive
characteristic ranges from 100000 Hz down to 1000 Hz. The low frequency part has a
capacitive part from 10 to 0.1 Hz. The first range has its minimum at 10000 Hz while the
second one shows a minimum between 2 and 3 Hz. With increasing elongation the
distribution into the two time constants becomes more and more dominant. In the high
frequency area the minimum in the phase shift changes from -75° to -40° while the low
frequency minimum increases from -30° to -50°. The low frequency resistance decreases by
one order of magnitude due to an elongation to 20%.
Results
144
0,1 1 10 100 1000 10000 100000
100
1000
10000
100000
1000000
0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
|Z| / Ω cm-2
f / Hz
0,1 1 10 100 1000 10000 100000
-80
-60
-40
-20
0 0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
Phase / °
f / Hz
Fig. 105. In-situ EIS during the forming of PET layer on electrogalvanised steel
dried at room temperature for 24 h. The forming leads to a decrease in the barrier properties of the
coating by the formation of defects
The thickness reduction of the coating seems to lead to a separation of the two time constants
(organic coating and double layer). The increase of the low frequency minimum can be
caused by the easier diffusion of water molecules through the coating and the formation of a
more significant double layer.
Results
145
5.9.6.2 Formability of a polyelectrolyte layer dried at 130°C
The polyelectrolyte layer shows capacitive behaviour in the range of 10000 to 10 Hz in the
unformed state with a minimum for the phase shift of -75° at 1000 Hz (Fig. 106). The low
frequency resistance shows good barrier properties with a value of about 1 MΩ / cm2. The
increase of the forming degree leads to a shift of the capacitive range to 1000 and 1 Hz and
also to a shift of the phase minimum to 50 Hz at 20% elongation. At higher elongations the
minimum of the phase shift drifts from -75° to -65° and shows a lesser broad structure. The
loss in low frequency resistance is less significant than for the room temperature dried sample
with just half an order of magnitude.
Results
146
0.1 1 10 100 1000 10000 100000
100
1000
10000
100000
1000000
0% elongation
5% elongation
15% elongation
20% elongation
|Z| / Ω cm2
f / Hz
0.1 1 10 100 1000 10000 100000
-90
-80
-70
-60
-50
-40
-30
-20
-10
0 0% elongation
5% elongation
15% elongation
20% elongation
Phase / °
f / H
z
Fig. 106. In-situ EIS during forming of PET layer on electrogalvanised steel dried at 130 °C. The forming
leads to a decrease in the barrier properties of the coating by the formation of defects
The slight change in the low frequency resistance even at high forming degrees indicates a
good formability of the coating with excellent barrier properties. The shift of the capacitive
range and the absolute minimum is caused by the thickness reduction of the coating during
the forming process.
Results
147
5.9.6.3 Formability of a polyelectrolyte layer dried at 230°C
The system dried under standard conditions shows capacitive properties in the range from
1000 Hz down to 0.1 Hz with a low frequency resistance of 400000 Ω / cm2 for the unformed
state (Fig. 107). The phase shift shows a broad distribution with a minimum of -80°at 10 Hz.
Elongation of only 5% leads to a dramatic loss of the barrier properties by nearly two orders
of magnitude. The capacitive area decreases to the range of 400 to 3 Hz with a sharp
minimum of -65° at 40 Hz. With further elongation the minimum shifts to -50° at 50 Hz,
simultaneously the low frequency resistance decreases further to 2000 Ω / cm2.
0,1 1 10 100 1000 10000 100000
10
100
1000
10000
100000
1000000
0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
|Z| / Ω cm-2
f / Hz
0,1 1 10 100 1000 10000 100000
-80
-60
-40
-20
0
0% elongation
5% elongation
10% elongation
15% elongation
20% elongation
|Z| / Ω cm-2
f / Hz
Fig. 107. In-situ EIS during forming of PET layer on electrogalvanised steel dried at 230 °C under air.
The forming leads to a decrease in the barrier properties of the coating by the formation of defects
Results
148
The impedance spectra of the high temperature cured coating (standard atmosphere) show
high barrier properties in the unformed state. But the first forming step leads to a dramatic
loss of these properties with a final spectrum similar to the one of an uncoated
electrogalvanised steel sample. The high temperature seems to lead to a highly dried and
hardened system with a brittle structure. Such a system completely loses its formability and
cannot withstand the applied forces during forming.
The impedance spectra were non linearly fitted to calculate reproducible values for the low
frequency resistance at 0.1 Hz (Tab. 8). This procedure avoids errors from scattered
measurement point due to noise uptake by the potentiostat. The data were normalised to the
initial value at an elongation of 0% and plotted in Fig. 108. The initial values for the polymer
layer dried at room temperature and at 130 °C show a significant higher resistance resulting in
much better barrier properties. The high temperature cured coatings, especially the one dried
under nitrogen, show much lower resistances and significant lower final values.
Tab. 8. Resistance of PET complex layer at 0.1 Hz taken from non-linear impedance fit
PET dried at RT
/ Ω cm²
PET dried at
130 °C
/ Ω cm²
PET dried at
230 °C (standard
atmosphere)
/ Ω cm²
0 % elongation 961575 771375 319020
5 % elongation 202337 291096 7171
10 %
elongation 249472 - 2734
15 %
elongation 99016 340988 2053
20 %
elongation 123676 299192 1848
The comparison of the PET layers dried at room temperature, 130 °C and 230° under standard
atmosphere shows a clear ranking with regard to formability (Fig. 108). The coating dried at
130 °C shows only a loss of about 50 % of its initial barrier property even at high forming
degrees. The room temperature dried one loses about 80 % of its initial value while the one
dried at 230 °C loses nearly its complete corrosion protection properties. In all cases the final
loss of the protective properties occurs in the first forming step and stabilises at higher
forming degrees.
Results
149
0 5 10 15 20
0
200000
400000
600000
800000
1000000
dried at room temperature
dried at 130 °C (standard atmosphere)
dried at 230 °C (standard atmosphere)
|Z| / Ω cm2 at 1 Hz
elongation / %
Fig. 108. Comparison of the formability of PET complexes cured at different temperatures. The system
dried at 130 °C shows the best forming performance with a decrease of the barrier properties to 40 %.
The information about the optimal curing conditions is of great interest as the formability and
thereby the barrier properties directly correspond with them. Too low temperatures would not
lead to the necessary cross linking while a too high temperature may lead to degradation of
the coating. It becomes clear that the optimal formability is reached at a medium temperature
of 130 °C for the investigated PET system.
Overall conclusions
150
6 Overall conclusions
The scope of this thesis was the development of new analytical tools to investigate the
formability, corrosion and barrier properties of organic coatings applied on electrogalvanised
steel. Thus the work focused on innovative, combined in-situ methods which generate
detailed information about the system under investigation. A new electrochemical capillary
cell setup was built, combining impedance spectroscopy with simultaneous forming of a
sample. Detailed information about the loss of the barrier properties during stretch forming by
the formation of defects and the corrosive attack of the electrolyte are gained by this setup.
Further results of the local loss of barrier properties and the corrosion product formation can
be obtained from the combination of a confocal Raman spectrometer with an electrochemical
capillary cell. Finally the new in-situ Raman / QCM flow cell allows the measurement of
adsorption kinetics and dissolution processes of e.g. inhibitors and further on allows the
identification of the products involved.
The lab coated pretreatment and corrosion protection primer show homogeneous distribution
and layer formation. The zinc particles embedded in the coating sizes from 2 – 7 µm are
covered with a rough oxide layer and form no direct, conducting contact between the coating
surface and the steel substrate. All uniaxial, biaxial, plane strain formed and also the
miniature stretching samples show a uniform strain distribution over the complete area under
investigation. The forming of the corrosion protection primer leads to the formation of defects
at the zinc particle / binder interface growing with increasing strain while the underneath zinc
structure shows no defects. Furthermore, the formed coating detaches from the substrate and
generates small sheltered caverns.
The cleaning procedure of the formed CPPs shows no visible influence on the surface
structure of the implemented zinc particles while the tendency to the formation of small
phosphate crystals rises with increasing pH of the cleaning solution. The precipitation of the
phosphate crystals appears on the free zinc area on top of the pigments and in the defects
induced by the forming procedure. Thus the existence and size of defects influences the
kinetic of the phosphating process which is accelerated by the amount of the free zinc surface.
The cleaning procedure shows no influence on the corrosion performance of the CPP in the
VDA climate test while the forming leads to a loss of the protective properties due to the
defect formation. The blistering of ED-paint applied on CPP additionally is just dependant on
pre-forming for very thin thicknesses of about 4 µm. With common thicknesses of about 15 –
Overall conclusions
151
20 µm the barrier properties of the ED-paint are the corrosion rate limiting step and no
distinction between formed and unformed samples can be observed.
FE-SEM and FIB analysis of formed (20 % elongated) CPP after 24 h salt spray test indicate
a strong local corrosion in the area of the forming induced caverns. The corrosion rate in these
caverns is significantly higher than on flat samples corresponding to a complete degradation
of the zinc layer after almost 24 h on local sites.
The in-situ Impedance Spectroscopy during stretch forming shows the formation of initial
defects already at small forming degrees. Further elongation leads to a significant loss of the
barrier properties while at high forming degrees (25 % elongation) the protection of the
coating is nearly completely destroyed which is indicated by the impedance value as of the
pure substrate.
Combined Raman / Impedance spectra indicate only a increase of in the formation of
corrosion products within two hours. While the barrier property degradation of the formed
CPP is a very fast process showing a significant change in the impedance.
The surface modification of the zinc particles to strengthen the bonding between the particles
and the binder matrix showed no reduction of the forming induced defects due to the high
strain inside the system.
The overall results indicate that a fast corrosive deadhesion along the particle/ binder interface
appears almost in the unformed state. After forming the induced defects and caverns allow an
easy penetration of the electrolyte and act as initial sites for a corrosive attack of the substrate
and the delamination of the coating.
The in-situ Raman / QCM measurements show a comprehensive view of the MBT inhibitor
adsorption on SERS active gold. The formation of a monolayer was completed after 6
minutes and could be correlated with the theoretically calculated frequency shift.
The investigated polyelectrolyte layers show the formation of covalent bonds after the drying
for several days at room temperature or after curing at temperatures of 130 or 230 °C. The
water uptake is reduced due to curing from 19.5 % for the room temperature dried coating to
5.5 % at 130 °C to finally 1.5 % for the one dried at 230 °C. Forming leads to the formation
of defects perpendicular to the elongation direction. The best formability is found for the
coatings cured at 130 °C while the dried at RT and 230 °C shows a significantly worse
formability.
Outlook
152
7 Outlook
In future studies the embedded zinc particles might be exchanged by other metal or alloy
pigments that guarantee the electrical conduction for spot welding and further on have a better
adhesion to the binder matrix and provide passivating properties. E.g. the formation of
voluminous, inhibiting corrosion products would allow a complete closure of the forming
induced defects at the particle / binder interface.
Other aspects to be considered are the passivation of the zinc particles by e.g. similar
solutions already used as pretreatment for the galvanised steel substrate. This could lead to a
reduction of corrosive delamination along the particle / binder interface resulting in a fast
decrease of the coating barrier properties.
The design of binder systems with low glass transition temperatures that allow closure of
forming induced defects at temperatures of about 180 °C, as they appear during curing of the
ED-paint, could probably significantly increase the barrier properties of the coating.
For polyelectrolyte coatings the kinetic of the cross linking should be further investigated as
this is of great importance for further technical applications concerning the formability after
the storage of the coated coils. For these systems also the double curing after application and
after ED-paint application might be suitable for improving the barrier properties. Further on
the electrical conductivity needed for spot welding might be increased by the implementation
of conductive polymers.
The setups developed within this thesis can be used to support the coating design by the fast
testing of new products concerning the formability and influence of implemented inhibitors.
But they are not only limited to coatings due to the high local resolution, they can also be
used to indentify and analyse the properties of e.g. inhomogeneous particles embedded in
metals or other technical products.
Publications
153
8 Publications
[1] C. Stromberg, P. Thissen, I. Klueppel, N. Fink and G. Grundmeier, “Synthesis and
characterisation of surface gradient thin conversion films on zinc coated steel”,
Electrochimica Acta, 52, 804 (2006)
[2] R. Vlasak, I. Klueppel and G. Grundmeier, “Combined EIS and FTIR–ATR study of
water uptake and diffusion in polymer films on semiconducting electrodes”,
Electrochimica Acta, 52, 8075 (2007)
[3] I. Klueppel, B. Schinkinger, G. Grundmeier, “In-Situ Electrochemical Studies of the
Formability of Organic Coatings on Metals”, submitted to Electrochimica Acta
[4] I. Klueppel, G. G. Sun, G. Grundmeier, “In-situ investigation of the kinetic adsorption
process of 2-mercaptobenzothiazole by combining surface-enhanced Raman
spectroscopy and quartz crystal microbalance”, in preparation
[5] I. Klueppel, O. Seewald, R. Regenspurger, W. Bremser, G. Grundmeier, in preparation
Literature
154
9 Literature
[1] B. F. Chen, J. Hwang, I. F. Chen, G. P. Yu and J. H. Huang, Surf. Coat. Technol., 126,
91 (2000)
[2] B. Baumert, M. Stratmann and M. Rohwerder, Z. Metallk., 95, 447 (2004)
[3] A. C. Bastos and A. M. P. Simoes, Progress in Organic Coatings, 46, 220 (2003)
[4] L. Fedrizzi, A. Bergo and M. Fanicchia, Electrochimica Acta, 51, 1864 (2006)
[5] B. S. Skerry and C. H. Simpson, Corrosion, 49, 663 (1993)
[6] M. Delucchi, A. Barbucci and G. Cerisola, Electrochimica Acta, 44, 4297 (1999)
[7] E. P. M. van Westing, G. M. Ferrari and J. H. W. de Wit, Corrosion Sci., 36, 979
(1994)
[8] F. Mansfield and M. W. Kending, Werkstoffe und Korrosion, 36, 473 (1985)
[9] F. Mansfeld and C. H. Tsai, Corrosion, 47, 958 (1991)
[10] U. Rammelt and G. Reinhardt, Progress in organic coatings, 21, 205 (1992)
[11] E. Cavalcanti and O. Ferraz, Progress in organic coatings, 23, 185 (1993)
[12] H.-R. Prause, Anwendbarkeit der Impedanzspektroskopie zur Beurteilung des
Korrosionsschutzes durch organische Beschichtungen, in Chemie, Stuttgart (1994)
[13] F. Mansfeld,
J. Appl. Electrochem., 25, 187 (1995)
[14] P. L. Bonora, F. Deflorian and L. Fedrizzi, Electrochimica Acta, 41, 1073 (1996)
Literature
155
[15] D. H. van der Weijde, E. P. M. van Westing and J. H. W. de Wit, Electrochimica Acta,
41, 1103 (1996)
[16] F. Mansfeld, H. Xiao;, L. T. Han and C. C. Lee, Progress in organic coatings, 30, 89
(1997)
[17] N. Tang, W. J. v. Ooij and G. Górecki, Progress in organic coatings, 30, 255 (1997)
[18] A. Miszczyk and K. Darowicki, Corrosion Sci., 40, 663 (1998)
[19] F. Deflorian, L. Fedrizzi, S. Rossi and P. L. Bonora, Electrochimica Acta, 44, 4243
(1999)
[20] A. S. L. Castela, A. M. Simoes and M. G. S. Ferreira, Progress in organic coatings,
38, 1 (2000)
[21] R. M. Souto, V. Fox, M. M. Laz and S. González, J. Adhes. Sci. Technol., 14, 1321
(2000)
[22] J. Vogelsang and W. Strunz, Materials and Corrosion, 52, 462 (2001)
[23] V. Lavaert, M. Moors and E. Wettinck, J. Appl. Electrochem., 32, 853 (2002)
[24] G. Bierwagen, D. Tallman, J. P. Li, L. Y. He and C. Jeffcoate, Progress in Organic
Coatings, 46, 148 (2003)
[25] J. M. Hu, J. T. Zhang, J. Q. Zhang and C. N. Cao, Journal of Materials Science, 39,
4475 (2004)
[26] V. Barranco and S. Feliu, Corrosion Sci., 46, 2203 (2004)
[27] V. Barranco and S. Feliu, Corrosion Sci., 46, 2221 (2004)
[28] D. Loveday, P. Peterson and B. Rodgers, Jct Coatingstech, 1, 88 (2004)
Literature
156
[29] A. C. Bastos, C. Ostwald, L. Engl, G. Grundmeier and A. M. Simoes, Electrochimica
Acta, 49, 3947 (2004)
[30] K. Darowicki, J. Orlikowski and A. Arutunow, Electrochimica Acta, 48, 4189 (2003)
[31] K. Darowicki and M. Szocinski, J. Solid State Electrochem., 8, 346 (2004)
[32] G. Decher, Y. Lvov and J. Schmitt, Thin Solid Films, 244, 772 (1994)
[33] F. Caruso, K. Niikura, D. N. Furlong and Y. Okahata, Langmuir, 13, 3422 (1997)
[34] J. B. Schlenoff, H. Ly and M. Li, J. Am. Chem. Soc., 120, 7626 (1998)
[35] D. Yoo, S. S. Shiratori and M. F. Rubner, Macromolecules, 31, 4309 (1998)
[36] J. J. Harris, P. M. DeRose and M. L. Bruening, J. Am. Chem. Soc., 121, 1978 (1999)
[37] P. Bertrand, A. Jonas, A. Laschewsky and R. Legras, Macromol. Rapid Commun., 21,
319 (2000)
[38] J. H. Dai, A. W. Jensen, D. K. Mohanty, J. Erndt and M. L. Bruening, Langmuir, 17,
931 (2001)
[39] B. Schoeler, G. Kumaraswamy and F. Caruso, Macromolecules, 35, 889 (2002)
[40] T. R. Farhat and J. B. Schlenoff, Electrochem. Solid State Lett., 5, B13 (2002)
[41] G. Ibarz, L. Dahne, E. Donath and H. Mohwald, Chem. Mat., 14, 4059 (2002)
[42] B. Schoeler, E. Poptoschev and F. Caruso, Macromolecules, 36, 5258 (2003)
[43] A. V. Dobrynin and M. Rubinstein, Prog. Polym. Sci., 30, 1049 (2005)
Literature
157
[44] M. D. Miller and M. L. Bruening, Chem. Mat., 17, 5375 (2005)
[45] G. Francius, J. Hemmerle, J. Ohayon, P. Schaaf, J. C. Voegel, C. Picart and B. Senger,
Microscopy Research and Technique, 69, 84 (2006)
[46] A. Schneider, G. Francius, R. Obeid, P. Schwinte, J. Hemmerle, B. Frisch, P. Schaaf,
J. C. Voegel, B. Senger and C. Picart, Langmuir, 22, 1193 (2006)
[47] K. Lange,
Umformtechnik, Springer, Berlin (1984)
[48] E. Doege and B.-A. Behrens, Handbuch Umformtechnik, Springer, Berliner (2007)
[49] R. Parisot, S. Forest, A. F. Gourgues, A. Pineau and D. Mareuse, Comput. Mater. Sci.,
19, 189 (2000)
[50] K. P. Imlau, K. D. Ehlers, S. M. Goklu, K. P. Mohr and D. Wolfhard, Stahl Und
Eisen, 113, 69 (1993)
[51] B. Tieke,
Makromolekulare Chemie, Wiley-VCH, Weinheim (2005)
[52] Stahl-Eisen-Prüfblatt 1160, Stahlinstitut VDEh, Düsseldorf (2004)
[53] T. Narayanan,
Rev. Adv. Mater. Sci., 9, 130 (2005)
[54] P. Puomi, H. M. Fagerholm, J. B. Rosenholm and K. Jyrkas, Surf. Coat. Technol.,
115, 70 (1999)
[55] C. Stromberg, P. Thissen, I. Klueppel, N. Fink and G. Grundmeier, Electrochimica
Acta, 52, 804 (2006)
[56] B. Wilson, N. Fink and G. Grundmeier, Electrochimica Acta, 51, 3066 (2006)
[57] N. Fink, B. Wilson and G. Grundmeier, Electrochimica Acta, 51, 2956 (2006)
Literature
158
[58] G. Klimow, N. Fink and G. Grundmeier, Electrochimica Acta, 53, 1290 (2007)
[59] U. A. Handge, I. M. Sokolov, A. Blumen, E. Kolb and E. Clement, J. Macromol. Sci.-
Phys., B38, 971 (1999)
[60] C. Lorenz-Haas, P. Muller-Buschbaum, T. Ittner, J. Kraus, B. Mahltig, S. Cunis, G. V.
Krosigk, R. Gehrke, C. Creton and M. Stamm, Phys. Chem. Chem. Phys., 5, 1235
(2003)
[61] A. C. Moloney and H. H. Kausch, J. Mater. Sci. Lett., 4, 289 (1985)
[62] G. H. Michler, J. Macromol. Sci.-Phys., B38, 787 (1999)
[63] R. P. Wool,
Soft Matter, 4, 400 (2008)
[64] S. R. White, N. R. Sottos, P. H. Geubelle, J. S. Moore, M. R. Kessler, S. R. Sriram, E.
N. Brown and S. Viswanathan, Nature, 409, 794 (2001)
[65] K. Aramaki,
Corrosion Sci., 44, 1621 (2002)
[66] K. Aramaki,
Corrosion Sci., 44, 1375 (2002)
[67] K. Aramaki,
Corrosion Sci., 45, 2361 (2003)
[68] M. R. Kessler, N. R. Sottos and S. R. White, Compos. Pt. A-Appl. Sci. Manuf., 34, 743
(2003)
[69] A. N. Khramov, N. N. Voevodin, V. N. Balbyshev and M. S. Donley, Thin Solid
Films, 447, 549 (2004)
[70] K. Aramaki,
Corrosion Sci., 47, 1285 (2005)
[71] G. Paliwoda-Porebska, M. Stratmann, M. Rohwerder, K. Potje-Kamloth, Y. Lu, A. Z.
Pich and H. J. Adler, Corrosion Sci., 47, 3216 (2005)
Literature
159
[72] A. Kumar, L. D. Stephenson and J. N. Murray, Progress in Organic Coatings, 55, 244
(2006)
[73] M. L. Zheludkevich, K. A. Yasakau, A. C. Bastos, O. Karavai and M. G. S. Ferreira,
Electrochem. Commun., 9, 2622 (2007)
[74] W. Feng, S. H. Patel, M. Y. Young, J. L. Zunino and M. Xanthos, Adv. Polym.
Technol., 26, 1 (2007)
[75] D. Y. Wu, S. Meure and D. Solomon, Prog. Polym. Sci., 33, 479 (2008)
[76] N. G. Hoogeveen, M. A. C. Stuart and G. J. Fleer, Langmuir, 12, 3675 (1996)
[77] D. G. Shchukin, M. Zheludkevich, K. Yasakau, S. Lamaka, M. G. S. Ferreira and H.
Mohwald, Adv. Mater., 18, 1672 (2006)
[78] D. G. Shchukin and H. Mohwald, Small, 3, 926 (2007)
[79] F. Micciche, H. Fischer, R. Varley and S. van der Zwaag, Surf. Coat. Technol., 202,
3346 (2008)
[80] D. G. Shchukin and H. Mohwald, Adv. Funct. Mater., 17, 1451 (2007)
[81] C. Gabrielle and M. Keddam, Corrosion, 48, 794 (1992)
[82] M. M. Lohrengel, Materials Science & Engineering R-Reports, 11, 243 (1993)
[83] J. Colreavy and J. D. Scantlebury, Journal of Materials Processing Technology, 55,
206 (1994)
[84] M. Itagaki, A. Matsuzaki and K. Watanabe, Corrosion Sci., 37, 1867 (1995)
[85] P. Schmuki, H. Bohni and J. A. Bardwell, 142, 1705 (1995)
Literature
160
[86] W. P. Gomes and D. Vanmaekelbergh, Electrochimica Acta, 41, 967 (1996)
[87] A. Nishikata, Y. Ichihara and T. Tsuru, Electrochimica Acta, 41, 1057 (1996)
[88] J. H. W. de Wit and H. J. W. Lenderink, Electrochimica Acta, 41, 1111 (1996)
[89] M. Gaberscek and S. Pejovnik, Electrochimica Acta, 41, 1137 (1996)
[90] P. Birke and W. Weppner, Electrochimica Acta, 42, 3375 (1997)
[91] Z. Kerner and T. Pajkossy, Electrochimica Acta, 47, 2055 (2002)
[92] V. A. Alves and C. M. A. Brett, Electrochimica Acta, 47, 2081 (2002)
[93] I. Betova, M. Bojinov, P. Kinnunen, T. Laitinen, P. Pohjanne and T. Saario,
Electrochimica Acta, 47, 2093 (2002)
[94] M. M. Lohrengel, S. Heiroth, K. Kluger, M. Pilaski and B. Walther, Electrochimica
Acta, 51, 1431 (2006)
[95] R. M. Souto, L. Fernandez-Merida, S. Gonzalez and D. J. Scantlebury, Corrosion Sci.,
48, 1182 (2006)
[96] F. Mansfeld,
J. Appl. Electrochem., 25, 1145 (1995)
[97] D. M. Brasher and A. H. Kingsbury, Journal of Applied Chemistry, 4, 62 (1954)
[98] E. P. M. van Westing, G. M. Ferrari and J. H. W. de Wit, Corrosion Sci., 36, 957
(1994)
[99] L. D. Rosa, T. Monetta and F. Bellucci, Materials Science Forum, 289-292, 315
(1998)
Literature
161
[100] A. S. Castela and A. M. Simoes, Corrosion Sci., 45, 1631 (2003)
[101] A. S. Castela and A. M. Simoes, Corrosion Sci., 45, 1647 (2003)
[102] J. T. Zhang, J. M. Hu, J. Q. Zhang and C. N. Cao, Progress in Organic Coatings, 51,
145 (2004)
[103] H. Marchebois, C. Savall, J. Bernard and S. Touzain, Electrochimica Acta, 49, 2945
(2004)
[104] A. Meroufel and S. Touzain, Progress in Organic Coatings, 59, 197 (2007)
[105] A. Meroufel, C. Deslouis and S. Touzain, Electrochimica Acta, 53, 2331 (2008)
[106] M. Ohman and D. Persson, Electrochimica Acta, 52, 5159 (2007)
[107] M. Ohman, D. Persson and C. Leygraf, Electrochem. Solid State Lett., 10, C27 (2007)
[108] R. Vlasak, I. Klueppel and G. Grundmeier, Electrochimica Acta, 52, 8075 (2007)
[109] H. Vogel,
Gerthsen Physik, p. 1262, Springer, Berlin (1997)
[110] A. J. Bard and L. R. Faulkner, Electrochemical Methods Fundamentals and
Application, p. 833, John Wiley & Sons, Hoboken (2001)
[111] J. R. Macdonald, Impedance Spectroscopy
Emphasizineg solid Materials ad Systems, p. 346, John Wiley & Sons, New York (1987)
[112] Gamry, Basics of Electrochemical Impedance Spectroscopy, in (2008)
[113] D. Loveday, P. Peterson and B. Rodgers, Jct Coatingstech, 1, 46 (2004)
[114] D. Loveday, P. Peterson and B. Rodgers, Jct Coatingstech, 2, 22 (2005)
Literature
162
[115] C. Gabrielli, Identification of Electrochemical Processes by Frequency Response
Analysis, p. 94, Paris
[116] G. W. Walter, Corrosion Sci., 26, 681 (1986)
[117] R. Hirayama and S. Haruyama, Corrosion, 47, 952 (1991)
[118] E. P. M. van Westing, G. M. Ferrari and J. H. W. de Wit, Corrosion Sci., 34, 1511
(1993)
[119] W. D. Hinsberg and K. K. Kanazawa, Review of Scientific Instruments, 60, 489 (1989)
[120] K. K. Kanazawa and O. R. Melroy, Ibm Journal of Research and Development, 37,
157 (1993)
[121] G. Sauerbrey,
Zeitschrift Fur Physik, 155, 206 (1959)
[122] Maxtec, Maxtek RQCM Quartz Crystal Microbalance Research System Manual, in
(2005)
[123] G. T. Fieldson and T. A. Barbari, Polymer, 34, 1146 (1993)
[124] M. Ohman, D. Persson and C. Leygraf, Progress in Organic Coatings, 57, 78 (2006)
[125] K. Wapner, M. Stratmann and G. Grundmeier, Electrochimica Acta, 51, 3303 (2006)
[126] T. Sueoka, J. Inukai and M. Ito, Journal of Electron Spectroscopy and Related
Phenomena, 64-5, 363 (1993)
[127] E. Johnson and R. Aroca, Journal of Physical Chemistry, 99, 9325 (1995)
[128] A. Bruckbauer and A. Otto, J. Raman Spectrosc., 29, 665 (1998)
[129] K. Ataka and M. Osawa, Journal of Electroanalytical Chemistry, 460, 188 (1999)
Literature
163
[130] J. K. Lim and S. W. Joo, Surf. Interface Anal., 39, 684 (2007)
[131] R. A. Gu, X. Y. Shen, G. K. Liu, B. Ren and Z. Q. Tian, J. Phys. Chem. B, 108, 17519
(2004)
[132] D. Gonnissen, A. Hubin and J. Vereecken, Electrochimica Acta, 44, 4129 (1999)
[133] H. F. Yang, X. J. Sun, J. Zhu, J. H. Ji, X. L. Ma and Z. R. Zhang, J. Phys. Chem. C,
111, 7986 (2007)
[134] G. Wedler,
Lehrbuch der physikalischen Chemie, Wiley VCH (1997)
[135] M. Hesse, H. Meier and B. Zeeh, Spektroskopische Methoden in der organischen
Chemie, Georg Thieme Verlag, Stuttgart (1991)
[136] P. W. Atkins,
Physikalische Chemie, p. 1106, VCH, Weinheim (1996)
[137] G. Wedler,
Lehrbuch der Physikalischen Chemie, p. 924, VCH, Weinheim (1987)
[138] E. Almeida, M. Balmayore and T. Santos, Progress in Organic Coatings, 44, 233
(2002)
[139] J. Grdadolnik,
Acta Chimica Slovenica, 49, 631 (2002)
[140] M. Milosevic,
Appl. Spectrosc. Rev., 39, 365 (2004)
[141] K. Herzog, S. Salzer and B. Ziegler, In-situ Untersuchungen thermischer Vorgänge in
Polymeren mittels Kopplung von FTIR-Spektroskopie und Differenz-Thermo Analyse
(DTA), in, TU, Institut für Analytische Chemie, Dresden
[142] M. Fleischmann, P. J. Hendra and McQuilla.Aj, Chemical Physics Letters, 26, 163
(1974)
Literature
164
[143] A. Otto, I. Mrozek, H. Grabhorn and W. Akemann, Journal of Physics-Condensed
Matter, 4, 1143 (1992)
[144] A. Otto,
Physica Status Solidi a-Applied Research, 188, 1455 (2001)
[145] R. Aroca,
Surface-Enhanced Vibrational Spectroscopy, Wiley, West Sussex (2006)
[146] G. Bertsche, Das Rasterelektronenmikroskop, in
[147] H. Jaksch and J.-P. Vermeulen, New developments in FESEM Technology, in
[148] C. A. Volkert and A. M. Minor, MRS Bull., 32, 389 (2007)
[149] W. J. MoberlyChan, D. P. Adams, M. J. Aziz, G. Hobler and T. Schenkel, MRS Bull.,
32, 424 (2007)
[150] T. Suter and H. Bohni, Electrochimica Acta, 42, 3275 (1997)
[151] T. Suter and H. Bohni, Electrochimica Acta, 47, 191 (2001)
[152] M. M. Lohrengel, C. Rosenkranz, I. Kluppel, A. Moehring, H. Bettermann, B. Van
den Bossche and J. Deconinck, Electrochimica Acta, 49, 2863 (2004)
[153] M. Pilaski, T. Hamelmann, A. Moehring and M. M. Lohrengel, Electrochimica Acta,
47, 2127 (2002)
[154] N. Birbilis, B. N. Padgett and R. G. Buchheit, Electrochimica Acta, 50, 3536 (2005)
[155] M. M. Lohrengel, A. Moehring and M. Pilaski, Fresenius Journal of Analytical
Chemistry, 367, 334 (2000)
[156] G. G. Sun and G. Grundmeier, Thin Solid Films, 515, 1266 (2006)
[157] A. W. Hassel, K. Fushimi and M. Seo, Electrochem. Commun., 1, 180 (1999)
Literature
165
[158] A. Losch, J. W. Schultze and H. D. Speckmann, Applied Surface Science, 52, 29
(1991)
[159] X. G. Zhang, Corrosion and Electrochemistry of Zinc, p. 474, Plenum
[160] M. Pourbaix,
Atlas of Electrochemical Equilibria in Aqueous Solutions., NACE
(National Association of Corrosion Engineers), Houston (1974)
[161] K. Wapner and G. Grundmeier, Adv. Eng. Mater., 6, 163 (2004)
[162] M. L. Abel, A. Rattana and J. F. Watts, J. Adhes., 73, 313 (2000)
[163] Z. D. Q. van Ooij W. J., Prasad G., Jayaseelan S., Fu Y., Teredesai N., Surface
Engineering, 16, 386 (2000)
[164] M. L. Abel, J. F. Watts and R. P. Digby, J. Adhes., 80, 291 (2004)
[165] T. Falk, J. E. Svensson and L. G. Johansson, J. Electrochem. Soc., 145, 2993 (1998)
[166] R. Lindstrom, J. E. Svensson and L. G. Johansson, J. Electrochem. Soc., 147, 1751
(2000)
[167] N. Sandhyarani, G. Skanth, S. Berchmans, V. Yegnaraman and T. Pradeep, Journal of
Colloid and Interface Science, 209, 154 (1999)
[168] I. Gaussian, Gaussian 03, in, Gaussian, Inc., Wallingford (2008)
[169] G. Decher and J. B. Schlenoff, Multilayer Thin Filmy, Wiley-VCH, Weinheim (2003)
[170] V. Barranco, P. Thiemann, H. K. Yasuda, M. Stratmann and G. Grundmeier, Applied
Surface Science, 229, 87 (2004)
