Corrosion Protection
by Selective Addressing of Polymer Dispersions to
Electrochemical Active Sites
Von der Fakultät für Naturwissenschaften
Department Chemie
der Universität Paderborn
zur Erlangung des Grades eines
Doktors der Naturwissenschaften
- Dr. rer. nat. -
genehmigte Dissertation
von
Sergej Toews, M.Sc.
aus Orenburg
Paderborn 2010
Erstgutachter: Prof. Dr. Wolfgang Bremser
Zweitgutachter: Prof. Dr.-Ing. Guido Grundmeier
Eingereicht am: 18.08.2010
Mündliche Prüfung am: 24.09.2010
für meine Familie: Marina,
Maximilian Abram und
Leonhard Jakob
v
Danksagung
Die vorliegende Arbeit wurde während meiner Tätigkeit als wissenschaftlicher
Angestellter am Institut für Polymere, Materialien und Prozesse der Universität
Paderborn im Zeitraum von November 2007 bis August 2010 angefertigt. Sie soll
mir an dieser Stelle Gelegenheit bieten, mich bei all denen zu bedanken, die
mich während meiner Promotionszeit begleitet und unterstütz haben.
Meinem Mentor und Betreuer, Prof. Wolfgang Bremser, gilt mein besonderer
Dank für die kontinuierliche Unterstützung, Förderung und Betreuung dieser
Dissertation sowie für seine stete Diskussionsbereitschaft und das Interesse am
Fortgang dieser Arbeit. Herrn Prof. Guido Grundmeier danke ich für die
freundliche Übernahme des Koreferats und für die vielen anregenden
Diskussionen.
Der BASF Coatings GmbH gebührt mein besonderer Dank für die fortwährende
Unterstützung durch mein gesamtes Studium, sowie für die Finanzierung der
vorliegenden Arbeit. Im speziellen möchte ich Herrn Dr. Horst Hintze-Brüning und
Herrn Dr. Michael Dornbusch für die interessante Thematik dieser Arbeit und für
die intensiven und konstruktiven Diskussionen danken. Mein spezieller Dank gilt
Herrn Dr. Sebastian Sinnwell für die anregenden Diskussionen zum Ende dieser
Arbeit.
Allen Mitarbeitern des Fachbereiches Chemie und Technologie der
Beschichtungsstoffe danke ich für die angenehme Atmosphäre, die Kollegialität,
die stete Hilfsbereitschaft und für viele heitere Stunden.
Herrn Dr. Manuel Lohrengel gilt mein besonderer Dank für die freundliche
Aufnahme in seiner Arbeitsgruppe für Mikroelektrochemie an der Heinrich-Heine-
Universität zur Durchführung von Mikro-Kapillar-Zellen-Messungen.
Mein größter Dank gilt meiner Familie für die anhaltende Unterstützung während
meiner Promotionszeit.
vii
Abstract
This study followed the scientific approach of identifying active sites for electrochemical,
corrosive processes on hot dipped galvanized steel and the creation of a material with
selective deposition properties that could inhibit the corrosive activity of these weak
spots. This study firstly involved high lateral resolution investigations on surface element
composition and electrochemical micro probe techniques that identified weak zones on
HDG steel substrate. Applying scanning electron microscopy and energy dispersive X-ray
spectroscopy it was found that aluminum in HDG alloys (ZnAl 0.5 w.-%) segregates not
only towards the zinc/iron and zinc/air interface but especially towards the boundaries of
zinc grains. Surface high resolution potential mappings utilizing scanning Kelvin probe –
force microscopy showed the electrochemical potential difference of grain boundaries
(lower potential) compared to the surrounding grain surfaces: a clear indication for higher
corrosive activity of these structures. The application of the micro capillary cell showed
that grain boundaries tend to dissolve more easily due to corrosive currents that were
measured at lower electrochemical potential (-910 mV) than on single grains (-830 mV).
The final corrosion test on a coil coated substrate showed the higher corrosion activity of
grain boundaries at the corrosion front. In this test corroded grain boundaries could be
observed leading up to 200 µm forward from the main corrosion front into the intact
coating/substrate interface. From these findings a model of different corrosion pathways
was derived, which suggests that the anodic part reaction quickly propagates forward
along the grain boundaries and is escorted by the local cathode. It can be said that on the
delaminated substrate surface the main corrosion front can spread out more easily.
Based on this model the corrosion propagation would decelerate by inhibiting the grain
boundary activity. The second part of this study focused on the design of polymeric
material and its selective application to grain boundaries. The anodic dissolution ability of
grain boundaries and the preferred release of aluminum cations were thus used for the
selective application of corrosion inhibiting materials on these active sites. An initial
applicability screening of materials towards grain boundaries showed promising results
with autophoresis of water-borne dispersions; among phosphating and surface
spontaneous polymerization. Based on these results water-borne block-co-polymer
dispersions were synthesized containing adhesion promoting groups such as carboxylic
acid, phosphonic acid and triethoxysilane. Application of these polymers exclusively on
grain boundaries could be realized with a controlled release of aluminum cations from
these weak spots. It was shown that in the pH region of 2.5 to 4.0 grain boundaries start
to dissolve and that coagulation of dispersed polymer particles is susceptible towards
triple charged aluminum cations. This combination led to selective deposition of polymer
particles on grain boundaries. The final proof of concept was provided by a comparison of
non-grain boundary treated versus grain boundary treated HDG substrates in a salt spray
test. The results showed a significant difference in the condition of the grain boundaries
at the corrosion front, where the grain boundaries selectively covered with a specifically
designed polymer reduced the anodic dissolution along the grain boundaries by a factor
of three to twenty. In conclusion, grain boundaries on HDG steel are highly corrosively
active and it was possible to block their corrosive activity by applying polymers only to
these weak zones. This new method of selective corrosion protection bears high potential
as a promising strategy towards smart and environmentally friendly pretreatment of steel
goods.
viii
Kurzfassung
Die vorliegende Studie verfolgte einen umfassenden experimentellen Ansatz die im
Korrosionsprozess elektrochemisch aktiven Schwachstellen auf feuerverzinktem (HDG)
Stahl zu identifizieren und Materialien selektiv auf diesen Stellen abzuscheiden um diese
dadurch zu inhibieren. Im ersten Teil dieser Studie wurden hochauflösende
Oberflächenanalysemethoden angewendet um die Elementzusammensetzung und die
elektrochemisch aktiven Stellen auf solchen heterogenen Oberflächen zu untersuchen.
Mittels der Rasterelektronenmikroskopie und der energiedispersiven Röntgenspektro-
skopie konnte gezeigt werden, dass sich Aluminium aus der Feuerverzinkungslegierung
(ZnAl 0,5 gew.-%) neben der Segregation an die Zink/Eisen- und Zink/Luftgrenzfläche
sich bevorzugt an den Korngrenzen der feuerverzinkten Stahloberflächen anreichert. Die
Raster-Kelvinsonde-Kraftmikroskopie detektierte einen Unterschied im elektrochem-
ischen Potential von Korngrenzen zu den Kornflächen. Dabei trat an den Korngrenzen
ein negativeres Potential auf, was auf eine höhere anodische Auflösung dieser Strukturen
hinweist. Mittels der Mikro-Kapillar-Zellen-Technik konnten Stromdichte-Potential-
Messungen an Korngrenzen und Kornflächen durchgeführt werden. Dabei konnten
bereits bei einer Polarisierung des Substrates von -910mV initiale Korrosions-ströme an
den elektrochemisch aktiven Korngrenzen detektiert werden. Auf den Kornflächen
wurden erste Korrosionsströme bei einem Potential von -830 mV gemessen. Dieser
Potentialunterschied bestätigte die anodische Auflösungsreaktion der Korngrenzen-
strukturen. Im abschließenden Korrosionstest an einem lackierten Substrat konnte
gezeigt werden, dass ausgehend von der aktiven Korrosionsfront eine korrosive
Schädigungen entlang der Korngrenzen von bis zu 200 µm unter den noch intakten
Grenzflächenbereich zwischen Beschichtung und Substrat auftritt. Aus diesen
Ergebnissen wurde ein Modell mit zwei verschiedenen Korrosionswegen für das
Fortschreiten der Korrosionsfront entlang dieser Grenzfläche entwickelt. Zunächst
schreitet die Lokalanode schnell entlang der Korngrenzen voran und wird begleitet von
dem komplementären kathodischen Bereich. Die nachfolgende Hauptkorrosionsfront
kann somit schneller auf dem vorgeschädigten Substrat voranschreiten. Ausgehend von
diesem Modell würde eine Inhibierung der Korngrenzenkorrosion auch eine
Verlangsamung der Hauptkorrosionsfront mit sich bringen. Im zweiten Teil dieser Studie
wurden Materialien für die gezielte Adressierung an die Korngrenzen synthetisiert und
untersucht. Als Steuerungselement wurde hierfür das anodische Auflösungsverhalten der
Korngrenzen und die damit verbundene lokale Freisetzung von Aluminiumkat-
ionenangewendet. Unter den zunächst untersuchten Techniken wie Phosphatierung und
spontane Oberflächenpolymerisation zeigte die autophoretische Polymerabscheidung die
vielversprechendsten Ergebnisse bezüglich der Selektivität des Abscheidemechanismus.
Basierend auf diesen Erkenntnissen wurden Block-co-polymer-Dispersionen mit
funktionalen Haftgruppen für oxidische Substrate (Carboxyl-, Phosphonsäure und
Triethoxysilan) für die gezielte Korngrenzenapplikation synthetisiert. Die Abscheidung
dieser Polymerdispersionen an den Korngrenzen konnte durch die kontrollierte
Freisetzung von Aluminiumkationen im pH-Wertbereich 2,5 - 4,0 erfolgen. Die
Destabilisierung der Dispersionspartikel durch dreiwertige Aluminiumkationen führte
schließlich zur selektiven Belegung der korrosiven Schwachstellen. Die abschließenden
Korrosionstests zeigten eine drei- bis zwanzigfache Verringerung der Schädigung entlang
der Korngrenzen, wenn diese gezielt mit den synthetisierten Block-co-polymeren belegt
und somit inhibiert wurden. Zusammenfassend konnte gezeigt werden, dass
Korngrenzen auf HDG-Stahl eine verstärkte Korrosionsaktivität aufweisen und dass diese
durch die gezielte Applikation geeigneter Polymere exklusiv an diesen Schwachstellen
herabgesetzt werden kann. Dieses Konzept für selektiven Korrosionsschutz birgt ein
hohes Potenzial als Strategie zur ressourcenschonenden und umweltfreundlichen
Vorbehandlung von Stahloberflächen.
ix
Content
Danksagung . . . . . . . . . v
Abstract . . . . . . . . .
vii
Kurzfassung . . . . . . . . .
viii
Chapter 1 – General Introduction . . . . . . 1
1.1 Corrosion impact on economy and society . . . . 1
1.2 Scientific approach . . . . . . . 3
1.3 Corrosion processes on coated substrates . . . . 4
1.4 Adhesion theory . . . . . . . 7
1.4.1 Mechanical adhesion . . . . . . 8
1.4.2 Primary and secondary adhesion . . . . . 9
1.5 Approaches to new conversion methods – a literature survey
. 14
1.5.1 Rare earth compounds and transition metals . . . 14
1.5.2 Self-healing function . . . . . . 15
1.5.3 Sol-Gel process . . . . . . . 16
1.5.4 Self-assembled monolayer . . . . . . 16
1.6 References . . . . . . . . 18
Chapter 2 – Applied Techniques . . . . . . 23
2.1 Scanning electron microscopy and
energy dispersive X-ray spectroscopy . . . . 23
2.2 Scanning Kelvin probe – force microscopy . . . . 24
2.3 Micro capillary cell . . . . . . . 25
2.4 Gel permeation chromatography . . . . .
28
2.5 Dynamic light scattering and ζ-potential . . . . 28
2.6 Surface plasmon resonance spectroscopy, measurements and
sensor preparation . . . . . . . 29
2.7 References . . . . . . . . 31
Chapter 3 – Results and Discussion . . . . . 33
3.1 Substrate characterization . . . . . . 33
3.1.1 Fundamentals . . . . . . . 33
3.1.1.1 Aluminum in HDG coating alloys . . . . . 33
3.1.1.2 Grain boundaries and weak zones . . . . . 36
x
3.1.2 Experimental procedures . . . . . . 38
3.1.2.1 Substrate . . . . . . . . 38
3.1.3 Experimental results . . . . . . . 39
3.1.3.1 Surface structure and element distribution on
HDG-Steel Al 0.5 w.-% . . . . . . 39
3.1.3.2 Surface potential . . . . . . . 43
3.1.3.3 Dissolution activity of grain boundaries vs. grains . . . 45
3.1.3.4 Corrosion behavior of grain boundaries on coated substrates . 50
3.1.3.5 Corrosion pathways as a model development . . . 51
3.1.4 Conclusion . . . . . . . . 55
3.1.5 References . . . . . . . . 56
3.2 Material survey for grain boundary application . . . 59
3.2.1 Fundamentals . . . . . . . 59
3.2.1.1 Phosphating . . . . . . . 60
3.2.1.2 Surface spontaneous polymerization . . . . 61
3.2.1.3 Polymer deposition . . . . . . . 62
3.2.2 Experimental procedure . . . . . . 63
3.2.2.1 Grain boundary dissolution . . . . . 63
3.2.2.2 Phosphating . . . . . . . 63
3.2.2.3 Surface spontaneous polymerization . . . . 64
3.2.2.4 Polymer deposition . . . . . . 64
3.2.3 Experimental results . . . . . . 65
3.2.3.1 Grain boundary dissolution . . . . . . 65
3.2.3.2 Phosphating . . . . . . . 66
3.2.3.3 Surface spontaneous polymerization . . . . 71
3.2.3.4 Polymer deposition . . . . . . . 72
3.2.4 Conclusion . . . . . . . . 76
3.2.5 References . . . . . . . . 77
3.3 Polymer design for grain boundary application . . . 81
3.3.1 Fundamentals . . . . . . . 81
3.3.1.1 Requirements and characteristics of the polymer of choice . 81
3.3.1.2 Synthesis of block-co-polymers . . . . . 82
3.3.1.3 Mechanistic understanding of the block-co-polymer formation
in the presence of DPE . . . . . . 85
xi
3.3.2 Experimental procedures . . . . . . 88
3.3.2.1 Materials . . . . . . . . 88
3.3.2.2 Synthesis of functional block-co-polymers in a hetero phase system
88
3.3.3 Experimental results . . . . . . 89
3.3.3.1 Overview of synthesized block-co-polymers . . . 89
3.3.3.2 GPC and particle size observations . . . . 90
3.3.3.3 Conversion ratios of the monomers . . . . 93
3.3.3.4 Dispersion stability . . . . . . . 96
3.3.4 Conclusion . . . . . . . . 97
3.3.5 References . . . . . . . . 98
3.4 Polymer application on grain boundaries . . . . 101
3.4.1 Fundamentals . . . . . . . 101
3.4.1.1 Colloid stability . . . . . . . 101
3.4.2 Experimental procedure . . . . . . 107
3.4.2.1 Titration on ion selectivity . . . . . . 107
3.4.2.2 SPR measurements . . . . . . . 107
3.4.2.3 Grain boundary selective polymer application . . . 107
3.4.2.4 Coating application and testing . . . . . 108
3.4.3 Experimental results . . . . . . 108
3.4.3.1 Polymer ion selectivity . . . . . . 109
3.4.3.2 Polymer particle stability in dependency of pH values . . 111
3.4.3.3 Polymer particle adsorption to aluminum and zinc oxide surfaces
112
3.4.3.4 Selective deposition on grain boundaries of HDG steel . . 119
3.4.3.5 Testing results . . . . . . . 124
3.4.4 Conclusion . . . . . . . . 128
3.4.5 References . . . . . . . . 129
Chapter 4 – Overall Conclusion and Outlook . . . . 131
Appendix . . . . . . . . . xiii
Abbreviations . . . . . . . . xiii
Publications . . . . . . . . . xv
xii
1
Chapter 1 –
General Introduction
Fig. 1.1: Example of corroded steel surface where the coating has lost its ability to protect
the surface from oxygen and humidity [1].
1.1 Corrosion impact on economy and society
Corrosion protection of steel surfaces plays an important role in everyday life.
The majority of steel manufactured goods and products are exposed to
atmospheric humidity and oxygen in which presence iron follows its natural role
to create the most thermodynamically stable compound. The steel oxidizes and
Chapter 1 – General Introduction
2
corrosion occurs. In order to protect steel from corrosion it is covered with a
multilayer coating system. In the case of galvanized steel, the first layer is
metallic zinc. When a corrosive attack on steel occurs that acts as a sacrificial
anode and protects the steel from oxidic dissolution. In general, the next layer is
a zinc phosphate layer which inhibits electrochemical corrosion processes on the
zinc surface and provides adhesion to the following organic coating layer. Aside
from the decorative demand for color and appearance, this organic coating layer
provides a barrier against humidity, oxygen and corrosive stimulants. Within
industry, the phosphating process is well established. It is easy to handle and
cheap in terms of the ingredients required. However the overall costs of the
phosphating process are in issue, due to a temperature of around 65°C and the
toxic sludge that is continuously generated in the phosphating bath. The toxicity
from the nickel and fluorides added into the phosphate bath for corrosion
improvement are also a major reason that industry is seeking new technologies
that would replace the phosphating process. This study will focus on these exact
issues. Existing corrosion concepts will be evaluated and a new concept based
on water borne polymers will be derived and proved in terms of its usability. In
addition, the reduction of the pretreated surface area from a continuous layer to a
smart and structured application that is only specific to the weak spots of the
substrate, will reduce the total amount of pretreatment material required, which
may in turn improve the economic and ecological efficiency.
Fig.1.1 provides an illustration of a corroded steel surface. The annual loss
resulting from corrosion in the industrial countries is estimated at 3-4% of GDP
[2]. According to these numbers, the annual economic loss for EU countries
alone can be calculated at 350 Billion Euro [3]. In addition to the high financial
costs the production of steel is very energy intensive and leaves a carbon
footprint of 1.35 tones CO
2
per ton of steel [4]. Prolonging the lifetime of steel
made goods will thus also lead to a reduction in carbon dioxide emissions. Even
though recent studies show that a quarter of these costs could be reduced
through the appropriate use of existing corrosion concepts, corrosion still remains
a tremendous loss factor. In addition, there is a strong need for the replacement
of traditional corrosion concepts based on heavy metals such as lead and
chromium, due to environmental and health responsibilities. There is a long
history of heavy metals having excellent corrosion protection for steel, but they
3
are toxic and in the case of hexavalent chromium even strongly carcinogenic
[5,6]. In the last decade huge efforts have been made by industry and research
institutions towards the development and implementation of new, environmentally
friendly corrosion protection systems that eliminate these substances from
technical applications. This change in technology is strongly enforced by political
directives. In 2000, the European Parliament prohibited the use of chromium
compounds for automotive industry, one of the major users of corrosion
protective materials [7]. For the aviation industry, a similar policy is being
prepared and will be enforced in the near future.
All of these aspects highlight the necessity for a better understanding of corrosion
processes and the development of new, environmentally friendly materials and
concepts for corrosion protection.
1.2 Scientific approach
The general consideration when inventing coatings for a specific type of
substrates e.g. steel, aluminum, or zinc coated steel is that the substrate surface
consists of a homogeneous distribution of all alloy containing elements. At the
same time, it is general knowledge that the surface of any technical substrate is
heterogeneous due to the alloying process. It is also well known that corrosion
prefers to start at weak zones. These areas are typically edges, vertices and
heterogeneous segregations [8]. It can be assumed that the electrochemical
process of corrosion starts at these spots. The aim of this study is to identify the
weak zones of a technical substrate of hot dipped galvanized steel (HDG),
investigate the electrochemical activity and influence on the corrosion process at
the substrate coating interface and develop a polymeric coating material to
specifically block and inhibit these weak zones. The approach of adapting the
coating material to the heterogeneous surface properties of technical substrates
would provide a new concept for environmentally friendly corrosion protection
coatings.
Chapter 1 –
General Introduction
1.3
Corrosion processes on coated substrates
E
lectrochemical processes of corrosion on a coated substrate were primarily
investigated by Evans in the
corrosion reactions are separated
local cells an
anodic and a cathodic
Fig. 1.2: Schematic illustration
of the cathodic delamination mechanism, a cross section
view [10].
Metal oxidizes i
n the area of the local anode
The released electrons from
conductive substrate to the local cathode. The oxidized
loose in the metal
lic grid structure and dissolve
Oxygen is reduced
to hydroxyl ions
General Introduction
4
Corrosion processes on coated substrates
lectrochemical processes of corrosion on a coated substrate were primarily
investigated by Evans in the
1960s [9].
Evans found that electrochemical
corrosion reactions are separated
locally into two part reactions; in the
so called
anodic and a cathodic
area (see Fig. 1.2).
of the cathodic delamination mechanism, a cross section
n the area of the local anode
.
The released electrons from
this part of the
reaction will flow through the
conductive substrate to the local cathode. The oxidized
metal atom
s become
lic grid structure and dissolve
into the ambient electrolyte.
to hydroxyl ions
at the local cathode in the
presence of water.
lectrochemical processes of corrosion on a coated substrate were primarily
Evans found that electrochemical
so called
of the cathodic delamination mechanism, a cross section
reaction will flow through the
s become
into the ambient electrolyte.
presence of water.
In a cascade of
the
In the case of iron
,
During the oxygen re
and peroxo radicals
the
outermost region of the defect
The radical s
pecies are highly
polymeric material of the organic coating.
T
he diffusion coefficient of water and electrolytes (
polymer/metal oxide interface
This leads to t
he destruction of t
between the coa
5
the
following reactions, metal ions and
hydroxyl
,
this process results in rust.
During the oxygen re
duction, many short-
lived intermediates of h
and peroxo radicals
are formed in the cathodic area which
is
outermost region of the defect
[11,12].
pecies are highly
reactive and therefore
prefer
polymeric material of the organic coating.
he diffusion coefficient of water and electrolytes (
10
-6
polymer/metal oxide interface
is 100 times higher
than through the bulk polymer
he destruction of t
he polymeric material particularly
ting and metal/metal oxide [13]. T
his process
hydroxyl
ions precipitate.
lived intermediates of h
ydroxy radicals
is
always located in
prefer
to react with the
cm²/s) along the
than through the bulk polymer
.
he polymeric material particularly
at the interface
his process
leads to de-
Chapter 1 –
General Introduction
adhesion of the organic coating and t
so called
cathodic delamination
atmospheric
humidity and occurs on metals
[10]. The most determining
factors for the
are highlighted as the pH-
stability of the metal oxide
the
adhesion of the organic coating
surrounding atmosphere
[13,15
sheets in order to improve
corrosion resistance by acting as a sacrificial anode
[21]. At the same time,
the zinc oxide surf
delamination. The
hydroxyl ions generated
value in the local cathode
area.
area the delaminated organic coating is still very close to the substrate which
results in a very small free volume. In this area a few hydro
to maximum values. Zinc/zinc oxide is
forefront of the delaminated area these pH values are exceeded which leads to
the dissolution of the amphoteric zinc oxide/hydroxide to zincate species
The benefit brought by
zinc coating
losses,
considering the corrosive
this fundamental understanding a model will be derived for corrosive
delamination on the heterogeneous substrate surface of HDG steel within this
study (see section 3.1).
In order to decrease
the delamination a
research is approaching
the improvement of corrosive
sides, by modifying the zinc all
of the zinc alloy approach
finds
doping the zinc alloy with metals of lower electrochemical potential e.g. titanium,
General Introduction
6
adhesion of the organic coating and t
o an uncovered metal substrate,
this is
cathodic delamination
[14]. Cathodic delamination
is dominant
humidity and occurs on metals
such as iron, but especially
factors for the
kinetics of the
cathodic delamination
stability of the metal oxide
,
the oxygen permeability and
adhesion of the organic coating
as well as
the relative humidity of the
[13,15
-20]. Zinc coating
s are applied to low alloy
corrosion resistance by acting as a sacrificial anode
the zinc oxide surf
ace is more affected by corrosive
hydroxyl ions generated
in the cathodic area ra
ise the pH
area.
In this outermost forefront of the delaminated
area the delaminated organic coating is still very close to the substrate which
results in a very small free volume. In this area a few hydro
xyl ions raise the pH
to maximum values. Zinc/zinc oxide is
stable up to the pH value of 13.
5 but in the
forefront of the delaminated area these pH values are exceeded which leads to
the dissolution of the amphoteric zinc oxide/hydroxide to zincate species
[22
zinc coating
to the steel substrate
is combined with
considering the corrosive
delamination of the organic coatings.
Based on
this fundamental understanding a model will be derived for corrosive
delamination on the heterogeneous substrate surface of HDG steel within this
the delamination a
nd corrosion process,
ongoing public
the improvement of corrosive
delamination
from two
sides, by modifying the zinc all
oy and
by improving the organic coating. Scoping
finds
recent research results that have
show
doping the zinc alloy with metals of lower electrochemical potential e.g. titanium,
this is
the
is dominant
in high
on zinc
cathodic delamination
the oxygen permeability and
the relative humidity of the
s are applied to low alloy
steel
corrosion resistance by acting as a sacrificial anode
ace is more affected by corrosive
ise the pH
In this outermost forefront of the delaminated
area the delaminated organic coating is still very close to the substrate which
xyl ions raise the pH
5 but in the
forefront of the delaminated area these pH values are exceeded which leads to
[22
-24].
is combined with
the
Based on
this fundamental understanding a model will be derived for corrosive
delamination on the heterogeneous substrate surface of HDG steel within this
ongoing public
from two
by improving the organic coating. Scoping
show
n that
doping the zinc alloy with metals of lower electrochemical potential e.g. titanium,
7
aluminum, or magnesium lower the electron conductivity of the oxide layer
significantly [11]. Electrons coming from the local anode are no longer able to
come up to the local cathode. Disrupting the cathodic part of the corrosion
process leads to a breakdown of the cathodic delamination. The anodic corrosion
reaction also then stops, when the electron flow is inhibited. The approach to
prevent cathodic delamination from the side of the organic coating faces the
major aspects of adhesion and degradation resistance of polymeric materials.
1.4 Adhesion theory
Adhesion theory in general describes adhesive forces at the interface of a solid
substrate and a second phase. The second phase can be a molecule, a particle,
a droplet or a continuous liquid or solid phase such as a coating [25]. Adhesion
theory is based on different models of mechanical, primary and secondary
adhesion, due to the different interaction mechanism of these two phases [26].
Considering coatings, adhesion is a complex phenomenon that can only be
partially explained by these simplified models; testing coating for adhesion visco-
elastic properties of the solidified coating material also provides a dominant factor
to the adhesion performance of the coating [27]. Nevertheless, the overall
adhesion performance of coatings is predominant for the protection of the metal
substrate and is also dependent on the molecular adhesion of the functional
groups incorporated into the coating material and addressed to the
substrate/coating interface. Molecular adhesion models will therefore be
discussed within this chapter. Based on these models, functional monomers have
been screened on their adsorption and desorption strength towards oxidic
surfaces [28]. These screening results lead to the monomer choice used for the
block-co-polymer design for grain boundary application of this study (see section
3.3).
Chapter 1 – General Introduction
8
1.4.1 Mechanical adhesion
The theory of mechanical adhesion describes the force-fit interlock of two
adhered phases. This mechanism takes place when one of the two adhesive
phases is a solid material with a rough or porous surface structure that comes
into contact with a penetrating liquid phase, which solidifies after spreading over
the surface. A good technical example would be a coating or a sealant applied to
a zinc phosphated or defined oxidized metal surface. Van den Brand et al.
reported to achieve a well-defined microstructure with good wetting properties on
aluminum by creating a pseudoboehmite oxyhydroxide layer through immersion
in boiling water [29]. This procedure creates a porous structure with an active
surface area increased 14 times [30]. In line with this work, mechanical adhesion
would occur in the rough structure of grain boundaries on coated HDG steel
samples (see section 3.1). A schematic understanding of the mechanical
adhesion models is illustrated in Fig. 1.3.
Fig.1.3: Schematic illustration of the relevant mechanical adhesion models.
On an absolutely planar surface this mechanism of force-fit interlock loses its
effectiveness and primary and secondary adhesion models have to be
considered. An exception to force-fit interlock on planar surfaces is found in the
adhesion of two thermoplastic surfaces. This model for adhesion is based on the
mechanical adhesion
(force-fit interlock)
interdiffusion
(entanglements of polymer
chains)
9
diffusion theory, due to inter-diffusion and entanglements of polymer chains.
Even though in corresponding literature polymer chain inter-diffusion is most
often considered as an independent model [31-33], in this study it will be
subordinated to the mechanical adhesion model based on a general
understanding of mechanics. The difference to the model previously discussed is
only that the mechanics are expanded to the sub nanometer scale on the
molecular level of polymer chains. Polymer chains in the bulk phase of the
material have some degree of mobility. This mobility is highly temperature
dependent, which makes this mechanism controllable and well suited to industrial
applications such as the welding of plastics. The mobility of polymer chains
becomes significant for diffusive interpenetration at temperatures above the glass
transition temperature. For example, hydrated poly polybutadiene chains with a
molecular weight of 105,000 g/mol and a glass transition temperature of 108 °C
have a diffusion coefficient of 2.22
.
10
-11
cm²/s at 125 °C. Raising the
temperature to 165 °C doubles the diffusion coefficient to 4.52
.
10
-11
cm²/s [34].
The welding process for plastics requires a self-diffusion coefficient above 10
-8
cm²/s in order to achieve an interpenetrated length of 50 µm within one second
[35]. Therefore the resulting entanglements of the single polymer chains can be
understood as force-fit interlock and associated to the theory of mechanical
adhesion.
1.4.2 Primary and secondary adhesion
Both primary and secondary adhesions are based on the fundamentals of
adsorption theory where chemical bonding takes place on a molecular level.
These types of adhesions play the most dominant role by designing polymers for
a strong bond to the metal/oxide substrate in order to reduce coating
delamination processes discussed in section 3.3 of this study. The concept of
primary adhesion includes covalent, ionic and metallic interaction. In Tab. 1.2. the
bonding type, the length and the energy of the molecular are consolidated. Fig.
1.4 schematically illustrates the chemistry of primary and secondary adhesion.
Chapter 1 – General Introduction
10
Tab.1.2: Bonding type, length and energy of molecular interaction in adsorption theory
[36].
bonding type length
[nm]
energy
[kJ/mol]
primary adhesion
ionic bonds
covalent bonds
metallic bonds
0.15 – 0.24
0.15 – 0.24
0.26 – 0.30
600 – 1100
60 – 700
110 – 350
secondary adhesion
hydrogen bridging bonds
dipole / dipole-interaction
(Keesom-Forces)
dipole / induced dipole-interaction
(Debye-Forces)
dispersive interaction
(London-Forces)
0.30 – 0.50
0.50 – 10.0
40 – 50
4 – 21
< 2
4 – 42
These chemical bonds typically have a relatively high bonding energy and are
therefore very stable. In the case of covalent bonding and the overlapping of
binding, molecular orbitals occur by sharing at least one pair of electrons from
one atom of the adhesive molecule and one from the substrate [37]. A good
example of covalent bonding in the interface chemistry is the application of thiols
or thiolfunctionalized polymers on gold surfaces [38,39]. In the present day,
covalent bonding to corrosion susceptible surfaces attracts a great deal of
attention in steel pretreatment chemistry. Covalent bonded molecules saturate
the chemical reactiveness of coordinative centers of the substrate and provide
them with a high resistance against further reactions. In addition, the bonded
molecules provide these reactive sites with steric hindrance and shield these
sites from corrosive attacks. Therefore the outermost atoms of the substrate
become stabilized against corrosive dissolution.
11
Fig.1.4: Schematic illustration of molecular interaction in the primary and secondary
adhesion.
Based on these concepts, over the last decade there have been general
developments of conversion systems using silane chemistry to passivate
aluminumoxide and zincoxide surfaces. It is postulated that the silanol group
when under cleavage of alcohol undergoes a condensation process with a
hydroxyl function of the substrate [40,41]. Investigations supporting this theory
were provided by finding
+
Si-O-Al and
+
Si-O-Fe fragments from ToF-SIMS
spectra on the substrate surface after removing the conversion layer [42].
However contra to this, recent investigations using the ToF-SIMS-technique
showed the same molecule fragment by treating the metal/oxide substrate with a
silicon oil polydimethylsiloxane (PDMS) [13]. It is very well known that PDMS is
chemically inert and does not react with the oxide surface of the substrate. This
means that the conversion chemistry of oxidic surfaces with silanols remain a
ionic bonding hydrogen bridging bonds
covalent bonding dipole / dipole-interaction
dipole / induced dipole-
interaction
Chapter 1 – General Introduction
12
research object of high interest. However, good adhesion of silanol conversion
layers to metal/metal oxide substrates and the improved corrosion resistance is
still undisputable [43,44]. This method of conversion is not fully understood but is
formulated to the stage where it is ready to be implemented in the steel coil and
automotive industry [45]. Based on the results of the monomer survey silane
monomers will also be co-polymerized and applied to grain boundaries within this
study.
Ionic bonds are very well known from organic and inorganic salts. The chemical
interaction in an ionic bond relies on positive and negative charges. In order to
describe the interaction between the adsorbing organic molecule/polymer and the
inorganic surface of the substrate, the HSAB-Model (Hard and Soft Acids and
Bases) has been developed by Pearson [46,47]. This model distinguishes acids
as electron pair acceptor molecules defined by Lewis or as proton donor
molecules as defined by Brönsted. Bases are the counter molecules to the acids
which make them electron pair donor molecules in the case of Lewis and proton
acceptor molecules in the case of Brönsted to [48]. The HSAB-Model also
distinguishes between hard (hard to polarize) and soft (easy to polarize) acids
and bases. Relatively stable acid-base-complexes result from the combination of
hard acids with hard bases and from soft acids with soft bases [49]. In the case of
a combination of a soft component with a hard component the resulting bond
strength is rather weak. Based on these principles the oxide surface of a
substrate can react in different ways with the polymeric coating material. Hydroxyl
terminated surfaces in general behave as a Brönsted base by donating their
protons to a proton acceptor such as an amino group coming from the coating
material [50]. In the case where the reactive group from the coating material is a
carboxylic acid, the hydroxyl function of the surface can be protonated and leave
the metal/oxide surface as a water molecule. The newly created and freely
available metal cation on the substrate surface reacts as a Lewis acid with the
carboxylic group, the Lewis base and creates a coordinative complex [51]. This
observation can be made on zinc substrates as well as aluminum [52]. Review of
the two substrates highlights the need to consider that the strengths of these
complexes cannot be the same, because carboxylic acids behave as strong
Lewis bases in their dissociated form [53]. In this way, the more stable complex
of the carboxylic acid function will be performed with the hard Lewis acid, the
13
Al
3+
-cation, rather than with the weak Lewis acid, the Zn
2+
-cation [54]. The
fundamental understanding of the interaction between cations and functional
groups of polymer dispersion will also be important for the selective deposition of
polymer particles on grain boundaries in section 3.4 of this study. As a results of
the most recent developments in single molecule desorption using atomic force
microscopy it is possible to measure the adhesion force of single functional
groups on the specific metal oxide surfaces. In fact not only the type of metal
oxide but also the surface orientation due to the crystal structure of the
metal/oxide leads to different adhesion strength of the adsorbed molecule.
Recent studies by M. Valtiner and G. Grundmeier have shown that the peal force
from an acrylic acid function on a hydroxide-stabilized polar ZnO(0001)-Zn
surface can be measured in the range of 60-80 pN and attributed to secondary
adhesion forces. When a carboxylic acid function was pealed off the edges of the
zinc oxide surface, peal forces of up to 700 pN were measured. The higher peal
force was attributed to coordinative bonding of the carboxylic function to the edge
of a polar surface ZnO(0001)-Zn surface [55]. These investigations are very
important to understand how adhesives adsorb to surfaces such as HDG steel in
order to evaluate the stability of the adsorbed functionalities under corrosive
conditions. In fact within this study, it will be found that HDG steel surfaces
appear as flat grains and grain boundaries can be described as multi edges
geometries (see section 3.1). Single molecule peel force investigations are still in
the early stages, but will provide important information and knowledge on the
adhesion strength of different functionalities and help to design specific polymeric
coating materials for the specific need of the different metal/oxide properties in
order to improve corrosion protection and resistance for cathodic delamination.
The metallic bond is characteristically known for creating a delocalized electron
gas which surrounds the atomic kernels and establishes the adhesive force
within a metal substrate. This type of chemical bond is of interest when
considering inter and intra-metallic phases. Because of its different focus, it will
not be discussed further in this thesis.
The secondary adhesion is based on secondary valences and distinguishes
between hydrogen bridging bonds, Keesom forces, Debye forces and London
forces. In the case of hydrogen bridging bonds, a hydroxyl terminated surface
interacts with the hydroxyl function of the coating polymer or simply with water.
Chapter 1 – General Introduction
14
The strength of the bond is dedicated to delocalization of the bridging proton
which is entropically stabilized [56]. In the case of Keesom forces, both the
adhesive and the substrate bear a dipole moment resulting in an attractive
interaction of the contrary charges. Debye forces interact similarly to the Keesom
forces with a marginal difference where only one of the reacting partners bears a
dipole moment but induces a contrary charge at the surface of its interacting
partner.
1.5 Approaches to new conversion methods; a literature
survey
Ambitious efforts to eliminate toxic and carcinogenic materials such as
hexavalent chrome have led to new approaches in passivation systems for steel,
galvanized steel and aluminum over the last few years. The supreme target of
these coating materials is to establish highly adhesive, oxygen and water
impermeable layers on the surface. The fundamentals of the previously
discussed adsorption theory provide the basis for the following innovative
conversion layer systems:
• Rare earth compounds and transition metals
• Self-Healing function
• Sol-Gel process
• Self-assembled monolayer
1.5.1 Rare earth compounds and transition metals
The idea of inhibiting corrosion through the use of rare earth compounds and
chromes analogous transition metals started where hexavalent chrome showed
its effectiveness. Choromium compounds have been used for corrosion
protection throughout the last century but were only classified as highly toxic and
carcinogenic two decades ago. In the corresponding literature, the properties in
terms of corrosion protection in conversion layers of elements such as Cerium,
Zirkonium, Molybdate or Vanadium have been investigated [57-59]. In all cases,
the oxidized form of the elements is the active corrosion inhibiting substance. The
15
use of trivalent chromium compounds is also reported to provide improved
corrosion protection as they are less reactive than the hexavalent chromium and
therefore exposure to humans is less harmful. However trivalent chromium could
oxidize to its hexavalent species under certain conditions [60]. Cerium for
example converts to ceriumhydroxyde at higher pH values, similar to those
obtained in the area of the local cathode. The ceriumhydroxide interrupts the
electron transport of the local cathode and the corrosion process stops. The
conversion of the substrate can be realized either through dipping it into a
solution of the corresponding salt or by adding these salts to the coating
pretreatment material that is applied to the substrate. Recently developed
processes add rare earth compounds to sol-gel coatings and combine both the
passivating abilities of the transition metals with the good barrier properties of the
sol-gel process. In this case, the oxidized form of the element is directly
implemented in the silica-oxygen matrix of the coating layer [58,61]. However,
some of the rare earth and transition metals are not sufficiently tested in terms of
long-term exposure to the human body. A broad application in industrial
manufacturing could lead to the classification of these compounds as critical for
the human body. It could be argued that the development of new metal
passivation materials should strive towards non-heavy metal compounds.
1.5.2 Self-healing function
The self-healing effect for coating materials in terms of corrosion protection was
firstly realized as a result of the addition of organic corrosion inhibitors such as 8-
hydroxychinoline and different types of benzotriazoles and benzothiazoles
[62,63]. The inhibitor molecules are not bonded to the polymeric matrix of the
coating and are able to diffuse. This mobility is necessary in the case where a
defect in the coating layer occurs. Inhibitor molecules can migrate to the bare
substrate spot and cover it by establishing chemical bonds to the metal surface.
But this mobility has one major disadvantage. The inhibitor is able to leak from
the coating and the system will thus within time, lose its self-healing properties.
The inhibitor covered defect can also only be protected for a limited quantity of
time. The latest developments in this area allow the encapsulation of the inhibitor
molecules in nano-container or nano-vesicles which release their content only in
Chapter 1 – General Introduction
16
the situation where damage to the coating occurs [64]. This method allows a
prolonged self-healing function of the coating layer, however the time period of
the healed status in the case of a defect occurring is still limited.
1.5.3 Sol-Gel process
A conversion layer based on the sol-gel technique provides improved corrosion
resistance properties on steel [65,66], galvanized steel [67], and aluminum [68].
The conversion layer can be created by any type of coating process e.g. dipping
or spraying with a silanol solution. In order to achieve a conversion layer with the
desired properties for corrosion protection, the silanol solution should be adjusted
to low pH values such as 4 – 5 [69]. The process of film formation is controlled by
the pH and is thus going through the steps of pre-condensation, generating
oligosilanole molecules and subsequently condensing with residual monomeric
silanole on the substrate surface. The deposited film is then finally cured at
raised temperatures in order to achieve the high cross link density. Even though
chemical bonding of this type of conversion layer is controversial within literature,
it provides a good protection toward cathodic delamination and corrosion
propagation. These abilities may also be attributed to the high cross link density
of the three dimensional network [70] as discussed in the adhesion theory.
1.5.4 Self-assembled monolayer
At an academic level, self-assembling molecules have been intensively
investigated as long chain alkylthioles (C
12
) on gold surfaces [71,72]. These
molecules adsorb to the gold surface and create covalent bonds of thiole
functionalities with gold atoms from the substrate surface. Fig. 1.4 shows a
schematic illustration of adsorbed SAM molecules to a metal substrate. SAM
molecules are tilting to an angle of around 15° and create a dense packaging due
to their long backbone chain. The densely packed and highly orientated
monolayer hermetically seals the substrate surface [73]. For modern corrosion
protection, applications of the molecules are designed with an anchoring group
on one side and a head group on the other. The anchoring group is designed for
the specific target surface which in the case of aluminum can be a phosphonic
17
acid [74]. Silane functional anchoring groups in analogy to the sol-gel process are
also terms of investigation for usability in terms of corrosion protection. Head
groups are designed to create covalent bonds to the top applied organic coating
system. A prominent example for such head groups is the amine functionality
which is very reactive with epoxy groups from the coating.
Fig.1.4: Schematic illustration of some of the intrinsic and extrinsic defects found in SAMs
formed on polycrystalline substrates. The dark line at the metal-sulfur interface is a visual
guide for the reader and indicates the changing topography of the substrate itself [75].
The same packaging angle that is responsible for the high impermeability of an
established SAM is also responsible for the biggest disadvantage of such
conversion systems in terms of corrosion protection. At ambient temperatures,
the formation of the self-assembled monolayer starts simultaneously at different
points and propagates in so called island growth [76]. Whenever two growing
SAM crystals hit each other they create an intrinsic defect and are not able to
protect the substrate sufficiently. In general, SAMs are weak in covering extrinsic
defects arising from the substrate surface such as grain boundaries, impurities
due to segregation and unevenness in topography (see Fig. 1.4). Therefore for
technical use, short chain molecules such as trimethoxysilyl propylamine (γ-APS),
a C
3
– chain molecule, are preferably applied. These molecules cannot create
high barrier properties but they act well as an adhesion promoter between the
substrate and coating [13].
Chapter 1 – General Introduction
18
1.6 References
[1] Annual Report 2003 IFAM Klebtechnik und Oberflächen (2003)
[2] Z. Zeng, K. Natesan, Z. Cai, S.B. Darling, Nature Mat. 7 (2008) 641
[3] Eurostat Press Release 55/2010 (2010)
[4] SteelConstructions.org, NSC 2010 (2010) 32
[5] W. Machle, F. Gregorius, Public Health Rep. 63 (1948) 1114
[6] T.F. Mancuso, W.C. Hueper, Ind. Med. Surg. 20 (1951) 358
[7] Directive 2000/53/EC of the European Parliament and of the Council of 18
September 2000 on end-of-life vehicles, Off. J. Europ. Comm. L269 (2002) 34
[8] E. Almeida, I. Alves, C. Brites, L. Fedrizzi, Prog. Org. Coat. 46 (2003) 8
[9] U. R. Evans, An Introduction to Metallic Corrosion, Hodder-Arnold, London
(1963)
[10] R.Posner, Thesis „Combined Spectroscopic and Electrochemical Studies of
Ion Transport and Corrosive de-Adhesion Processes at Polymer/Oxide/Metal
Interfaces”, Universität Paderborn (2009)
[11] R. Hausbrand, Steel Research Int., 74 (2003) 453
[12] A. Leng, H. Streckel, K. Hofmann, M. Stratmann, Corr. Sci. 41 (1999) 599
[13] K. Wapner, Thesis „Grenzflächenchemische und elektrochemische
Untersuchungen zur Haftung und Enthaftung an modifizierten Klebstoff/Metall-
Grenzflächen“ Ruhr-Universität Bochum (2006)
[14] P.A. Sorensen, K. Dam-Johansen, C.E. Weinell, S.Kiil, Prog. Org. Coat. 67
(2010) 107
[15] H. Leidheiser, Corrosion Control by Organic Coatings, Science Press,
Princeton NJ (USA), 1979
[16] H. Leidheiser, W. Wang, L. Igetoft, Prog. Org. Coat. 11 (1983) 19
[17] M. Stratmann, A. Leng, W. Fürbeth, H. Streckel, H. Gehmecker, K. H.
Grosse-Brinkhaus, Prog. Org. Coat. 27 (1996) 261
[18] W. Fürbeth, M. Stratmann, Prog. Org. Coat. 39 (2000) 23
[19] A. Leng, H. Streckel, M. Stratmann, Corr. Sci. 41(1999) 547
[20] A. Leng, H. Streckel, M. Stratmann, Corr. Sci. 41(1999) 579
[21] M. Rohwerder et al, Annual Report 2003, Max-Planck-Institut für
Eisenforschung GmbH, Düsseldorf (2003)
19
[22] W. Fürbeth et al, Corr. Sci., 41 (2001) 207
[23] W. Fürbeth et al, Corr. Sci., 41 (2001) 229
[24] W. Fürbeth et al, Corr. Sci., 41 (2001) 243
[25] A. J. Kinloch, Adhesion and Adhesives – Science and Technology, Chapman
and Hall, London (1987)
[26] D. E. Packham, J. Adhesion, 39 (1992) 137
[27] G. Meichsener, T.G. Mezger, J. Schröder, Coatings Compendien
„Lackeigenschaften messen und steuern“ Vincentz, Hannover (2003)
[28] S. Toews, Adsorption funktionaler Monomere an oxidischen Grenzflächen,
Master Thesis, Universität Paderborn (2007)
[29] J.van den Brand, S.van Gils, P.C.J.Beentjes, H.Terryn, V.Sivel, J.H.W.de
Wit, Prog. Org. Coat., 51 (2004) 339
[30] J.van den Brand, O.Blaiev, P.C.J.Beentjes, H.Terryn, J.H.W.de Wit,
Langmuir, 20 (2004) 6308
[31] A. J. Kinloch, Adhesion and Adhesives – Science and Technology, Chapman
and Hall, London (1987)
[32] W. Possart, Adhesion-Current Research and Applications, Wiley-VCH,
Weinheim (2005)
[33] A. V. Pocius, Adhesion and Adhesives Technology-An Introduction,2
nd
Ed.,Hanser, München, Wien (2002)
[34] C. R. Bartels, Macromolecules, 17 (1984) 2702
[35] H. Potente, Fügen von Kunststoffen - Grundlagen Verfahren Anwendung,
Hanser, München, Wien (2004)
[36] L.-H. Lee, Fundamentals of Adhesion, Plenum Press, New York, London
(1991)
[37] A. F. Holleman, E. Wieberg, Lehrbuch der Anorganischen Chemie, 101
Auflage, deGruyter, Berlin (1995)
[38] M. Büttner, J. Phys. Chem. B, 109 (2005) 5464
[39] L. D. Unsworth, J. Coll. Interf. Sci., 281 (2005) 112
[40] M.-L. Abel, J. Adhes., 80 (2004) 291
[41] J. F. Watts, Surf. Interface Anal., 29 (2000) 115
[42] J. F. Watts, Int. J. Adhes. Adhes, 16 (1996) 5
[43] M. F. Montemor et al, Electrochim Acta, 52 (2007) 7486
[44] W. J. van Ooij et al, Surf. Engineering, 16 (2000) 386
Chapter 1 – General Introduction
20
[45] P. Schubach, Chemetall GmbH, 2. Korrosionsschutz-Symposium, Grainau
(2007)
[46] A. J. Kinloch, Adhesion and Adhesives – Science and Technology, Chapman
and Hall, London (1987)
[47] A. V. Pocius, Adhesion and Adhesives Technology-An Introduction,2
nd
Ed.,Hanser, München (2002)
[48] A. F. Holleman, E. Wieberg, Lehrbuch der Anorganischen Chemie, 101
Auflage, deGruyter, Berlin (1995)
[49] R. G. Pearson, Coord. Chem. Rev.100 (1990) 403
[50] C. Fauquet, Appl. Surf. Sci. 81 (1994) 435
[51] N. Carciello, J. Mater. Sci., 19 (1984) 4045
[52] J. H. W. de Wit, Langmuir, 20 (2004) 6308
[53] K. P. C. Volhardt et al, Organische Chemie, dritte Auflage,Wiley-VCH,
Weinheim (2000)
[54] A. F. Holleman, E. Wieberg, Lehrbuch der Anorganischen Chemie, 101
Auflage, deGruyter, Berlin, (1995)
[55] M. Valtiner, G. Grundmeier, Langmuir 26 (2010) 815
[56] C. Tanford, The Hydrofphobic Effect: Formation of Micelles and Biological
Membranes, Krieger Publishing Comp. Malabar (1991)
[57] A. J. Davenport, Corr. Sci. 32 (1991) 653
[58] J. Krömer, Henkel AG & Co. KGaA,2. Korrosionsschutz-Symposium,
Grainau (2007)
[59] R. L. Cook, Corrosion, 56 (2000) 321
[60] A.D. Apte, V. Tare, P. Bose, J. Hazard. Mater. 128 (2006) 164
[61] P. Schubach, Chemetall GmbH, 2. Korrosionsschutz-Symposium, Grainau
(2007)
[62] M. L. Zheludkevich, Electrochem. Comm., 9 (2007) 2622
[63] S. H. Sanad, Surface Tech., 22 (1984) 29
[64] M. L. Zheludkevich, Chem. Mater., 19 (2007) 402
[65] M. Guglielmi, J. Sol-Gel Sci. Technol. 8 (1997) 443
[66] G. P. Sundararajan, Surf. Eng. 16 (2000) 315
[67] M. F. Montemor, Prog. Org. Coat. 51 (2004) 188
[68] A. Franquet, Surf. Interface Anal., 36 (2004) 681
[69] W. J. van Ooij, J. Adhes. Sci. Technol. 11 (1997) 29
[70] W. J. van Ooij, J. Adhes. Sci. Technol. 17 (2004) 2191
21
[71] J. P. Folkers, J. Adhes. Sci. Technol. 6 (1992) 1397
[72] G. M. Whitesides, J. Am. Chem. Soc., 113 (1991) 12
[73] E. Delamarche, Langmuir, 10 (1994) 4103
[74] P. Thissen, M. Valtiner, G. Grundmeier, Langmuir 26 (2010) 156
[75] J.C. Love, L.A. Estroff, J.K. Kriebel, R.G. Nuzzo, G.M. Whitesides, Chem.
Rev. 105 (2005) 1103
[76] B.J. Ravoo, D.N. Reinhoudt, S. Onclin, Angew. Chem. 117 (2005) 6438
Chapter 1 – General Introduction
22
23
Chapter 2 –
Applied Techniques
2.1 Scanning electron microscopy and energy dispersive
X-ray spectroscopy
Scanning Electron Spectroscopy (SEM) images were obtained by means of a
NEON 40 FE-SEM (Carl Zeiss SMT AG, Oberkochen, Germany). For Energy
Dispersive X-ray (EDX) spectroscopy analysis an UltraDry Silicon Drift X-ray
detector from Thermo Scientific was used. Both the imaging and element
analysis were obtained using an acceleration voltage of 5.0 kV and an electrode
working distance of 5.0 mm.
It is important in this section to highlight the resolution ability of the obtained EDX
spectra. With dependency to the accelerating voltage ܧ
(in keV), the critical
excitation voltage ܧ
(in keV) and the mean sample density ߩ (in g/cm³), the
spatial resolution of the obtained EDX analysis can be calculated with the
Equation 2.1 [1].
Chapter 2 – Applied Techniques
24
For the two main elements of interest to this thesis, aluminum and zinc, the
accelerating voltage of 5.0 keV and the mean sample density of 7.14 g/cm³
(density of zinc, as zinc is the major element of alloy) the spatial resolution is
shown in Tab. 2.1.
Tab. 2.1: Spatial resolution of the EDX detected element lines [2,3] for an acceleration
voltage of 5.0 keV.
Element
Line
Critical Exitation
Voltage (E
C
)
Spatial
Resolution
Al
(k-Line)
1.49 keV 0.04 µm
Zn
(l-Line)
1.01 keV 0.06 µm
Under these conditions the resolution of 60 nm was calculated for the analysis of
zinc substrates. This spatial resolution is rather high and provides the appropriate
dimension for the characterization of HDG steel surface heterogeneities in the
top layer of 60 nm and grain boundaries ranging between 100 to 1000 nm.
2.2 Scanning Kelvin probe – force microscopy
The Scanning Kelvin Probe – Force Microscopy (SKP-FM) potential mappings of
HDG steel surfaces were generated by a Veeco Dimension Icon (Veeco
Instruments Inc., Santa Barbara, USA) using the Veeco Surface Potential
imaging tool. The distance between the substrate and the conductive cantilever
tip was set to 50 nm and adapted to the surface topography. The excitation
25
voltage on the cantilever tip was 500 mV. This technique allows measurement of
the electrochemical potential of the substrate surface top layer.
2.3 Micro capillary cell
The micro capillary cell, also known as scanning droplet cell, is a tool for spatially
resolved electrochemical investigation of metallic surfaces and was developed
simultaneously by Lohrengel and Bohni [4,5]. The fundamental idea is to position
tiny electrolyte droplets (diameter of ≥10 µm) on the investigated surface. The
wetted area on the substrate forms the working electrode (WE), a gold wire within
the electrolyte filled capillary is the counter electrode (CE) and silver
silver/chloride precipitate provides the reference electrode (RE). A schematic
illustration of the micro capillary cell set up is shown in Fig. 2.1. The electrolyte
was provided by an acetate solution at a pH value of 6. With this three-electrode
arrangement, the complete range of common potentiostatic techniques such as
impedance, transients or cyclovoltammograms is measurable.
Fig. 2.1: The three-electrode arrangement of the micro capillary cell technique where the
substrate is the working electrode (WE), a gold wire within the electrolyte filled micro
Chapter 2 – Applied Techniques
26
capillary is the counter electrode (CE) and silver/silver chloride precipitate is the
reference electrode (RE) [6].
In the present study, the micro capillary cell measurements were carried out with
a micro capillary cell technique developed at the Institute of Physical Chemistry
and Electrochemistry of the Heinrich-Heine Universität Düsseldorf as illustrated in
Fig. 2.1. The major challenge when applying this technique on small structures
such as grain boundaries is to prepare a capillary with a working diameter in the
same range as the structure to be investigated. Therefore a conventional
capillary with an inner diameter of 1.8 mm is heated at a single spot and pulled
apart in a manner so that the melting glass on the spot taper becomes a cone
point as shown in Fig. 2.2. (left).
Fig. 2.2: Photograph of the pulled glass capillary to cone point (left) and polished tip of
the micro capillary to an mouth diameter of approx. 70 µm (right) [7].
The capillary end (capillary mouth) must be flat in order to provide closed contact
with the substrate. Therefore in the conventional capillary preparation method,
the capillary mouth must be micro polished. The capillary mouth that typically
results is shown in Fig. 2.2. (right). However due to the fragile nature of such
small glass pieces, it is very difficult to prepare smaller flat polished taper
capillary ends with a diameter of less than 50 µm; the reproduction of capillaries
with the same mouth diameter is also quite difficult.
27
Fig. 2.3: Capillary preparation by Focused Ion Beam cutting. Capillary end as obtained
after pulling with an inner diameter of the capillary end of ca.10 µm (top). Arrangement for
the Capillary in the SEM/FIB Chamber in the NEON 40 FE-SEM from Zeiss (middle).
Capillary end after cutting with the inner diameter 20 µm (bottom).
Chapter 2 – Applied Techniques
28
A diameter larger than 50 µm would be too large for the electrochemical analysis
of grain boundaries as found in HDG steel (see section 3.1). Therefore
preparation of the capillary mouth was carried out through the application of the
Focused Ion Beam technique (FIB). FIB is a popular tool for the cutting or drilling
of micro structures especially in the micro-chip industry [8,9]. It uses a gallium ion
beam to remove material from a sample to a precision of a few nanometers, this
cutting procedure is shown in Fig. 2.3. With this technique, capillaries with any
mouth diameter can be obtained. On the other hand, even though diameters
smaller than 5 µm could be obtained, such small diameters were difficult to
handle during the measurement process. When approaching the substrate
surface, breakage of the capillary at the fragile end occurred quite often and thus
only a few measurements could be obtained with each capillary. It was found that
the smallest diameter for the secure handling of the capillaries was 20 µm. Micro
capillaries with a diameter of 20 µm were thus used for the surface
characterization.
2.4 Gel permeation chromatography
Gel permeation chromatography (GPC, L-Series from Merck) was used to
determine molecular weights and molecular weight distributions, M
w
/M
n
, of block-
co-polymers synthesized in section 3.3: Polymer Design for Grain Boundary
Application. Molecular weights are calculated with respect to the polystyrene
standards. The measurements were obtained in Tetrahydrofuran (THF) at a
temperature of 25°C.
2.5 Dynamic light scattering and ζ
ζζ
ζ-potential
Dynamic light scattering (DLS) and ζ-potential were measured on a Zetasizer
Nano-ZS from Malvern. DLS was used to determine the particle size and its
distribution in the resulting polymer dispersion. ζ-potentials were measured under
standard conditions using the automatic Hückel approach at a temperature of
25°C.
29
2.6 Surface plasmon resonance spectroscopy
measurements and sensor preparation
Surface plasmon resonance (SPR) spectroscopy is an advanced method that
probes adsorption phenomenon on metal surfaces [10]. It is widely used on gold
surfaces and their variations that have some organic monolayer. This technique
has found broad application in biosensoring through the use of monolayer
chemistry as a receptor for biomolecules [11-12]. Studies monitoring the self-
assembly of long chain molecules on aluminum oxide layers using the SPR
technique have also been conducted [13,14]. Within section 3.4:Polymer
Application on Grain Boundaries of this study, this technique will be used to
investigate the adsorption and precipitation of polymer particles to aluminum
oxide and zinc oxide surfaces. However SPR devices are based on the detection
of refractive index changes in a thin dielectric layer found on top of a noble metal
surface and probed by the evanescent field of a laser beam [15]. The reflected
intensity of the beam is recorded as a function of incident angle and decreases
dramatically as light couples into the plasmon mode of the metal or the
waveguide modes of the dielectric. The evanescent tail of the plasmon is very
surface sensitive. When adsorption on the surface occurs the plasmon interacts
with the additional material. This results in the reflection intensity minimum
shifting to higher angles. This method can therefore be used to determine film
thickness. In the present study, the SPR technique will be used to detect the
adsorption of block-co-polymer dispersions to aluminum oxide and zinc oxide
surfaces with a pH dependency. Investigations in this work were carried out on a
SPR apparatus Res-Tec GmbH, Framershein. Fig. 2.5 provides a schematic of
the SPR flow cell set up and the resulting spectra information in dependency of
the thickness of the adsorbed polymer layer. Going from characters a) to c) in the
scheme the adsorbed polymer layer is increasing, resulting in the shift of the
reflection minima.
Chapter 2 – Applied Techniques
30
Fig. 2.5: Scheme of the SPR flow cell set up and the spectra obtained by varying the
adsorbed polymer layer thickness. 1-Prism; 2-sensor LaSFNN9 glass; 3-gold layer; 4-
metal oxide layer; 5-flow cell. [a. no adsorbed polymer layer; b. monolayer of polymer
particles; c. increased thickness of the adsorbed polymer layer] 6-laser beam.
SPR sensors based on LaSFN9 glass with defined gold layers were purchased
from MPI Polymer Research, Mainz. The film build was 0.5 nm Chromium as an
adhesion promoter for the 45 nm gold layer. The sensors were finished with a
layer of metal oxide obtained through physical vapor deposition (PVD). Findings
showed that a layer of zinc or aluminum deposited on the sensor had to be
smaller than 10 nm for proper plasmon activity. Higher film builds of zinc or
aluminum resulted in layers that were too high for plasmons to inform the sensor
surface. The film build recorded at around 6 nm were obtained with a PVD
chamber pressure of 5 x 10
-5
mbar and a deposition rate of 10 Å/s. After
removing the sensors from the PVD chamber the metal with a thickness of only a
few nanometers immediately oxidized to aluminum oxide or zinc oxide. The oxide
surface of the sensor imitates surface conditions of the hot dipped galvanized
steel. The exact film build of the sensors was calculated by fitting the resulting
SPR spectra using WinSpall Software V3.02, which is also provided by the
manufacturer of the SPR apparatus. The thickness of the layers of the prepared
sensors as used in section 3.4 for the polymer adsorption studies are gathered in
Tab. 2.1.
31
Tab. 2.1: Sensor film build as used for the SPR adsorption studies. Layer thickness is
measured with SPR and with calculated WinSpall V3.02, Res-Tec.
2.7 References
[1] S. J. B. Reed, Electron Microprobe Analysis, Cambridge Univ.
Press,Cambridge, England (1975)
[2] D.O. Flamini, S.B. Saidman, J. Appl. Electrochem 38 (2008) 663
[3] K. Tsuji, K. Nakano, H. Hayashi, K. Hayashi, C.-U. Ro, Anal. Chem. 80 (2008)
4421
[4] T. Suter, H. Bohni, Electrochim. Acta 42 (1997) 3275
[5] M.M. Lohrengel, Electrochim. Acta 42 (1997) 3265
[6] C.J. Park, M.M. Lohrengel, T. Hammelmann, M.Pilaski, H.S. Kwon,
Electrochim. Acta 47 (2002) 3395
[7] M.M. Lohrengel, A. Moehring, M. Pilaski, Electrochim. Acta 47 (2001) 137
[8] M.D. Henry, M.J. Shearn, B. Chhim, Nanotechn. 21 (2010)245303
[9] A. Shahmoon, O. Limon, O. Gishavitz, Microelectronic Engineering 87 (2010)
1363
[10] J. Homola, Chem. Rev. 108 (2008) 462
[11] W. Knoll, Annu. Rev. Phys. Chem. 49 (1998) 569
[12] K. S. Phillips, Q. Cheng, Anal. Bioanal. Chem. 387 (2007) 1831
[13] S. Heyse, O. P. Ernst, Z. Dienes, K. P. Hofmann, H. Vogel, Biochemistry 37
(1998) 507
[14] E. Jaehne, S. Oberoi, H.-J. Adler, Prog. Org. Coat. 61 (2008) 211
[15] S. Oberoi, E. Jaehne, H.-J. Adler, Macromol. Symp. 254 (2007) 284
[16] D. Kuckling, M.E. Harmon, C.W. Frank, Macromolecules 35 (2002) 6377
Chapter 2 – Applied Techniques
32
33
Chapter 3 –
Results and Discussion
3.1 Substrate characterization
3.1.1 Fundamentals
3.1.1.1 Aluminum in HDG coating alloys
In the process of hot dip galvanizing a steel strip runs through a bath of molten
zinc-aluminum alloy. There are in general, three main galvanized steel products
distinguishable by their aluminum content: HDG Al 0.5 w.-%, Galfan Al 5 % w.-%,
and Galvalume Al 55 w.-% [1]. This study will focus on the most prominent, HDG
Al 0.5 w.-%steel. The addition of Aluminum to the zinc bath in the galvanizing
process is known to result in zinc-coatings with improved ductility, brightness and
uniformity. Aluminum retards the surface oxidation of the molten zinc during the
coating application process [2]. Due to the low electrochemical potential of
aluminum at -1.66 V (electrochemical potential of zinc -0.76 V) it is highly reactive
towards oxygen and oxide formation. In addition, the density of aluminum oxide
(3.75 g/cm³) is less than the density of zinc (7.14 g/cm³) or zinc oxide (5.61
g/cm³) which leads to the formation of a fine aluminum oxide layer on the surface
of the molten zinc alloy bath and prevents the zinc alloy further oxidizing beneath.
All of these characteristics result in a smooth flow of the molten zinc alloy and
therefore a final smooth surface of the zinc coated steel sheet [3]. The high
Chapter 3 – Results and Discussion
34
reactivity of aluminum that prevents the zinc molten bath from oxidation is also
beneficial when the molten zinc alloy is applied onto the steel substrate.
Aluminum reacts preferably with the ferreous steel surface and creates a thin
layer of a FeAl-alloy. This reaction in turn disables the zinc iron reaction and the
formation of brittle FeZn-alloys. Therefore, aluminum acts as an adhesion
promoter between the steel and the zinc phases [4]. G. A. Lopez investigated the
solid solubility of aluminum in zinc [5]. He found that the solubility of aluminum is
negligible at room temperature and that during solidification of the zinc coating,
aluminum segregates towards the outermost surfaces of a solidifying single zinc
grain. The poor solubility of aluminum in zinc can be attributed to the different
atom radii (Zn 135 pm; Al 125 pm) and their different metal crystal lattice
structures. The incorporation of smaller aluminum atoms into the zinc crystal
lattice (hexagonal close-packed) results in the distortion of the crystal lattice and
a rise in lattice energy. In order to minimize the systems energy, aluminum atoms
are sorted out from the zinc lattice and are replaced by zinc atoms during the
solidification process. Therefore when aluminum atoms accumulate, they solidify
in the aluminum crystal lattice, face-centered cubic [6]. Weinberg et al. made
similar observations with Galfan hot-dipped galvanized steel. In their results
aluminum segregates to the surface of the zinc coating and provides a passive
aluminum oxide/hydroxide layer of a few nanometers thickness [7]. Surface
aluminum segregation was assumed as a reason for the enhanced corrosion
resistance of Galfan steel. Similar segregation could also be observed in HDG
steel with aluminum contents lower than 5 w.-% [8-11]. Selected images of
aluminum segregation in HDG steel are shown in Fig. 3.1.
Fig. 3.1: a) SEM image of a cross section of a HDG steel Al < 0.5 w.-%. b) EDX mapping
on Aluminum of the cross section. (bright color high Aluminum content) [12].
Substrate characterization
35
Nevertheless in these papers the segregation effect of aluminum to the surface is
observed and described as equally distributed on the surface. (see Fig. 3.2)
Fig. 3.2: Schematic illustration of the aluminum segregation in HDG coating [8].
Within this study, HDG steel with a zinc aluminum alloy of Al 0.5 w.-% will also be
characterized. The focus within this section will be on aluminum segregation
towards grain boundaries and their activity in the corrosion process, whereas the
following section will deal with inhibiting these electrochemically active spots.
However in those publications, the improved corrosion resistivity of HDG steel
was also attributed to the aluminum rich surface. Aluminum is very reactive due
to its negative electrochemical potential, as discussed earlier in this section. The
oxidation reaction of aluminum is rather fast, but due to its auto-passivation ability
of a short period; the oxidation of the surface top layer stops when a dense
passive aluminum oxide layer has been formed. The volume ratio of aluminum
and aluminum oxide can be calculated as 1:1. This ratio is most important so that
the oxide layer is stable and not brittle [13]. Therefore the aluminum oxide layer
seals the substrate surface and the oxidizing process stops. In comparison with
iron and its oxide, the volume ratio is 1:2. The oxide layer needs two times more
space on the substrate surface. Subsequently it succumbs to stress induced
cracks within the oxide layer. The brittle oxide layer uncovers a new iron surface
which reacts with oxygen.
The passive aluminum layer becomes distracting when phosphating is applied on
HDG steel. Phosphating is the current technology of surface conversion
chemistry and broadly processed in industry. The passive aluminum layer
Chapter 3 – Results and Discussion
36
protects zinc from corrosion and dissolution. At the same time, it is predominant
to dissolve the outermost zinc layer from the substrate in order to precipitate a
phosphate layer. [14-17]. Fundamentals of phosphating will be discussed in
section 3.2: Material Survey for Grain Boundary Application. In order to remove
aluminum from the substrate surface alkaline cleaning procedures have been
established. These procedures are carried out through simple dipping of the
substrate into an alkaline solution with a pH of 12. While aluminum and its oxides
dissolve at pH values above 10.5, zinc remains insoluble up to a pH of 13.5. This
etching process uncovers the zinc and therefore activates the surface for the
phosphating process [18,19].
Berger et al. have investigated the industrial
alkaline cleaning process [20] and found that the cleaning procedure does not
remove all of the aluminum. Even after a prolonged cleaning procedure,
aluminum can be found in some segregated islands and in grain boundaries of
the HDG steel surface. Similar observations will be made in the experimental
results of this section. It is postulated however on the one hand that the
aluminum oxide layer on HDG steel is passivating the substrate surface and
improving the corrosion resistance. On the other hand, current technology is
making an effort to remove this passive layer in order to apply a phosphate
conversion layer. A true benefit of the process would be the direct application of
the conversion chemistry on the passivated HDG surface. Following along the
lines of this study within this section, the passivated HDG steel surfaces will be
characterized towards their surface chemistry. The weak spots, in terms of
electrochemical corrosion activity, will be localized and discussed.
3.1.1.2 Grain boundaries and weak zones
Industrial surfaces are heterogeneous, as shown in some obvious cases such as
inter-granular corrosion, which is known from metal engineering (see Fig. 3.3).
For coated substrates it is postulated that corrosion starts mostly at weak spots ,
which are typically edges and vertices [21]. To date no one in open literature has
investigated this assumption for substrates such as HDG steel. It is also possible
that segregation spots could generate galvanic cell and contact corrosion.
Accumulated alloy elements are in electrochemical contact with the surrounding
alloy. In the case of aluminum alloys in near-neutral bulk solutions, it is well
Substrate characterization
37
established that pitting is influenced by intermetallic particles that exhibit different
surface characteristics to the matrix and may be either anodic or cathodic in
relation to the matrix [22-30]. Often in grain boundaries, both segregation of
intermetallics and some amorphous metallic structures with edges and vertices
on the interphase to air can be observed. Therefore, grain boundaries are known
to be electrochemically active species that can also function as transport
channels for small molecules and ions [31]. The phenomenon of intergranular
corrosion has been soundly investigated [32,33].
Fig. 3.3: Cross section of Ni-base superalloy after exposure to air at 550°C for 90h. The
segregated elements within the grain boundaries oxidize and cause the intergranular
corrosion [34].
A prominent example of where intergranular corrosion occurs is in stainless steel.
Selected images of intergranular corrosion are shown in Fig. 3.3. Chromium rich
carbides accumulate in the grain boundaries of the steel during welding or
general service leading to corrosion susceptibility. Chromium oxidation, starting
from the surface and moving inward, leads to de-adhesion of the grains and
intergranular corrosion thus occurs [35-37]. Trinidade et al. describe the role of
alloy grain boundaries as short-circuit diffusion paths for inward oxygen transport
[34]. Bredesen and Hussey et al. describe high diffusivity for molecular oxygen
through micro cracks occurring in the oxidizing grain boundaries [38,39].
However all this knowledge is based on engineering granular materials such as
steel where grain boundary reactions lead to embrittlement. But open literature
provides no information on how grain boundaries on non-brittle substrate
surfaces behave when coated with an organic coating. From observations in
Chapter 3 – Results and Discussion
38
industry it was reported that aluminum in HDG steel was migrating along grain
boundaries from the zinc/iron interface to the zinc/coating interface during the
baking of the organic coating [40]. In any case grain boundaries are very active
sites. Findings from the literature survey show that the grain boundaries also
found on HDG steel surfaces could have potential as weak spots. Within this
section, surface heterogeneities of HDG steel Al 0.5 w.-% will be investigated in
order to determine weak areas with high electrochemical activity.
3.1.2 Experimental procedures
3.1.2.1 Substrate
The HDG-Steel sheets (DX 53D + Z 100 NA 0.6 mm) were provided by
voestalpine AG, Austria. The aluminum content in the zinc alloy coating is
quantified by the manufacturer at < 0.5 w.-%. For surface characterization and
polymer application the steel samples were solvent cleaned in a three step
process (see Fig. 3.4). The first step, a thorough rinsing for 10 min in an
ultrasonicated bath of Tetrahydrofuran (THF) was followed by drying in a nitrogen
gas stream. Step one was repeated twice, once with iso-propanol and then with
ethanol, followed by nitrogen drying.
Fig. 3.4: Schematic illustration of substrate cleaning procedure.
Substrate characterization
39
3.1.3 Experimental results
3.1.3.1 Surface structure and element distribution on
HDG Steel Al 0.5 w.-%
In order to characterize the surface structure and chemistry of the HDG Steel
substrate high resolution SEM images and element mappings were obtained.
Fig.3.5. shows the typical surface structure of HDG steel. The texture arose from
the solidification of the molten zinc and is visible as grains, so called spangles,
with a diameter of approximately 100 to 200 µm. Where ever two grains are in
contact with each other, the grain boundaries are visible. At a higher
magnification the grain boundaries can be seen as a grid of connected micro
tranches all over the substrate surface. These trenches vary in their cross section
dimensions of width and depth from a few nanometers to a micrometer.
Fig. 3.5: a) Grain (spangle) structure at the surface of HDG steel Al <0.5 w.-%. Grain
boundaries can be seen as a grid of connected trenches. b) High resolution image of a
triple point of three grain boundaries.
Chapter 3 – Results and Discussion
40
In Fig. 3.5 the dimensions of the grain boundaries at the triple point have an
estimated width of 300 to 400 nm. The surface of a single grain can be seen as
rather smooth and flat but also with some micro distortions and scratches. The
scratches are probably due to the rolling transportation of the steel strip during
the zinc coating process. Grain boundaries occur as edges and vertices on the
substrate surface, which in literature, as discussed in the fundamentals of this
section, are described as the weak zones where the corrosion processes start. In
order to characterize the surface element composition EDX mappings were
obtained.
Fig. 3.6: a) SEM image and corresponding EDX element mapping of the HDG steel
surface. b) High resolutionSEM image and corresponding EDX element mapping
focusing a triple point of three grains and its grain boundaries. The mapping was obtained
with an accelerating voltage of 5.0 keV and a working distance of 5.0 mm.
Substrate characterization
41
Segregation effects of aluminum on HDG steel are most often described to the
surface and to the interface of zinc coating and steel. There are also a few
observations that have been made where aluminum was found enriched in the
grain boundaries of the zinc grains at aluminum concentrations higher than
5.0 w.-% [41,42]. In Fig. 3.6 high resolution EDX mappings clearly show
aluminum enriched in the surface grain boundaries, with aluminum
concentrations of 0.5 w.-% aluminum in the HDG coating alloy. Most of the
surface analytics dealing with coating development use homogeneous averaged
element compositions in order to determine the surface characteristics and select
the right adhesion promoters for such surfaces. One of the most common
arguments stated for not focusing on these heterogeneities is the application of
the alkaline cleaning process. As discussed in the fundamentals of this section,
alkaline cleaning is used to remove the passive aluminum layer so that the
conventional pretreatment procedures can be applied. Literature dealing with
alkaline cleaning processes finds only negligible amounts of aluminum on the
HDG steel surface after the alkaline cleaning. Therefore in the following, the
alkaline cleaning process was investigated with concern to its homogenization of
the HDG steel surface.
Fig. 3.7: EDX analysis of HDG steel surface 200 x 200 µm before (black line) and after
(red line) alkaline cleaning. The measurements were obtained with an accelerating
voltage of 5.0 keV and a working distance of 5.0 mm.
Chapter 3 – Results and Discussion
42
Fig. 3.7 provides element spectra of the solvent cleaned HDG steel surface
before and after alkaline cleaning as used in industry. The measurements were
taken from an area of 500 x 500 µm. This area contains some grains and also
some grain boundaries. In analogy to literature the aluminum content of the
outermost substrate surface could be reduced from 3.8 Atom-% to an
insignificant amount of 0.6 Atom-%. However, before undertaking the alkaline
cleaning, most of the aluminum was found in the grain boundaries. It is therefore
also important to investigate the element distribution after the alkaline cleaning
procedure. Fig. 3.8 shows a high resolution EDX element mapping of a partly
alkaline cleaned HDG substrate. In the top of the SEM image (solvent cleaned) a
smooth surface is visible.
Fig. 3.8: SEM image and EDX mapping of a solvent and partly alkaline cleaned HDG
steel substrate. Top of the substrate is solvent cleaned, bottom is alkaline cleaned.
In the lower part of the image (alkaline cleaned) the native oxide surface was
removed through alkaline etching. Similar to findings discussed in literature, the
rough morphology of grain surfaces is visible. The different morphologies of the
two grains imply that grains may have a different main crystallographic
Substrate characterization
43
orientation. Most importantly, by scoping the EDX mappings it can be seen that
there is little change in the oxygen intensity along the grain boundary. The
aluminum intensity decreases along the grain boundary on the bottom where the
alkaline has cleaned but does not disappear in total. These findings lead to the
conclusion that alkaline cleaning does not create a homogeneous surface on the
HDG steel as often stated by industry and in literature. The topography of the
grain boundary also does not change much through the procedure of alkaline
cleaning. This would imply that if grain boundaries on the HDG steel surface are
the weak zones of interest, they probably remain as the weak zones after the
alkaline cleaning procedure. Before alkaline etching, only the edges of grains
(grain boundaries) may be active while the mean surface of the grain is naturally
passivated by a compact aluminum oxide layer. After the alkaline cleaning, the
grain surfaces, in addition to the grain boundaries also become activated. The
aim of this study is to localize the weak zones of the HDG substrate and create
coating material that treats only those weak zones active in the corrosion
process. All these findings address the grain boundaries as possible weak zones
susceptible for corrosion. The alkaline cleaning process makes sense as an
activation process for the corrosive application of conventional pretreatments,
e.g. the phosphate layer. Alkaline cleaning does not create a homogeneous
surface. However it would be easier and more efficient to use the substrate as
produced by steel manufacturing. A coating process could be applied without
additional etching and activation of the steel surface that would also lead to
economic advantages in the process chain.
3.1.3.2 Surface potential
Scanning Kelvin probe force microscopy (SKP-FM) allows mapping of the
topography and Volta potential distribution on surfaces [43]. It combines the
classical Kelvin probe technique [44] with atomic force microscopy [45-47]. SKP-
FM operates at much smaller distances to the probe surface and uses the
cantilever tip as the electrode. This technique allows a higher lateral resolution
than the classical Kelvin probe technique. The lateral resolution of SKP-FM is
reported to be better than 0,1µm [48-50], compared to ~ 100µm for the standard
Kelvin probe technique [51,52]. This technique will be applied in order to
Chapter 3 –
Results and Discussion
determine the weak zones on
surface area of 20 x 20
µm containing a grain
shown in Fig. 3.9.
Results of the SKP
mappings in an atmosphere with 50% relative humidity at 20°C. Along the grain
boundaries and on some of the scratches the potential could be determined as
significantly lower than on the smooth
negative Volta potential between the grain boundary and the main grain surface
implies that the grain boundaries have anodic behavior relative to the
surrounding grains [53]. In other words, grain boundaries have a
oxidation reactions.
Fig. 3.9: SKP-
FM potential mapping over grain boundaries, scratches, and grains.
Andreatta and Terryn
et al. find such behavior in intermetallic particles
responsible for pitting corrosion
measure the Volta potential between intermetallics
measurements
of the HDG steel surface two ma
potential can be considered. The first one
grain boundaries.
According
aluminum is less noble than zinc due to its lower standard potential
result in lower potentials measured by SKP
Results and Discussion
44
determine the weak zones on
the HDG Steel.
A surface potential mapping of a
µm containing a grain
boundary and som
e scratches is
Results of the SKP
-
FM measurements show surface potential
mappings in an atmosphere with 50% relative humidity at 20°C. Along the grain
boundaries and on some of the scratches the potential could be determined as
significantly lower than on the smooth
surface of the grains. The existence of a
negative Volta potential between the grain boundary and the main grain surface
implies that the grain boundaries have anodic behavior relative to the
surrounding grains [53]. In other words, grain boundaries have a
higher affinity to
FM potential mapping over grain boundaries, scratches, and grains.
et al. find such behavior in intermetallic particles
responsible for pitting corrosion
[54,55]. In that case SKP-
FM was used to
measure the Volta potential between intermetallics
, the matrix and the
air.
of the HDG steel surface two ma
j
or effects for the lower surface
potential can be considered. The first one
is
the presence of aluminum in th
According
to the fundamentals of this section
aluminum is less noble than zinc due to its lower standard potential
which would
result in lower potentials measured by SKP
-FM. The second
effect could be
A surface potential mapping of a
e scratches is
FM measurements show surface potential
mappings in an atmosphere with 50% relative humidity at 20°C. Along the grain
boundaries and on some of the scratches the potential could be determined as
surface of the grains. The existence of a
negative Volta potential between the grain boundary and the main grain surface
implies that the grain boundaries have anodic behavior relative to the
higher affinity to
FM potential mapping over grain boundaries, scratches, and grains.
et al. find such behavior in intermetallic particles
FM was used to
air.
In the
or effects for the lower surface
the presence of aluminum in th
e
metallic
which would
effect could be
Substrate characterization
45
topography. A grain boundary is an edge with a rough nano-structure. Metal
atoms from edges are more readily available to chemical reactions. This implies
that grain boundaries may be highly reactive species in terms of metal oxidation
and dissolution resulting in corrosion also because of their geometry.
3.1.3.3 Dissolution activity of grain boundaries vs. grains
In the following part of this section, the corrosive ability of grain boundaries will
be investigated by measuring their susceptibility to electrochemical dissolution. A
higher dissolution activity would be an indication of higher corrosion activity. The
dissolution activity of grain boundaries in comparison to the flat area of the grains
could thus be observed using the micro capillary cell technique. Since the
establishment of this technique, many applications in corrosion research,
especially in intermetallic particles or single grains of some metallic texture, have
been reported [56,57]. Investigations on the dissolution behavior of single grains
on polycrystalline zinc have been shown to have different crystallographic
orientation of grains resulting in different reactivity in terms of metal dissolution,
oxidation and passivation [58]. Similar results could be obtained on ferritic steel
[59]. Schreiber et al. observed that grain boundaries of ferritic steel between
grains of different crystallographic orientation would also have different
dissolution kinetics [60]. However there is no information available on the
behavior of grain boundaries in the HDG steel surface. In order to investigate the
electrochemical behavior of grain boundaries in the HDG steel substrate, a
capillary with a diameter of 20 µm was prepared and set up as described in
Chapter 2 of this thesis. The capillary was filled with an acetate buffer of pH 6.
Fig.3.10 shows targeting of the specific surface area and how the target was
approached for measurement.
Chapter 3 – Results and Discussion
46
Fig. 3.10: Targeting the micro capillary on the grains and grain boundaries on the HDG
Steel substrate surface.
Grain boundary measurements were taken on spots in a triple point of three
grains. After touchdown of the capillary on the surface, a potential sweep was run
starting from -1.0 V up to + 1.5 V and the current density resulting from the metal
dissolution was obtained. Each spot was measured once only because of the
changing surface during the measurement. After each measurement the capillary
was rinsed with electrolyte before targeting the next spot. Fig. 3.11 shows a large
number of current density – potential curves obtained from measurements on the
grains and grain boundaries. The dashed lines always indicate current activity of
the grain boundaries and solid lines indicate the grains.
Substrate characterization
47
Fig. 3.11: Current density-Potential curves obtained from micro capillary measurements
on grain boundaries (dashed line) and on grains (solid line).
On the first view, a broad scattering of the obtained data from the specific spots
of the HDG steel surface can be observed. Starting from low potential first current
activities can be detected at potentials of around -900 to -800 mV. Due to the
oxidation and dissolution of the metal, the current density rises with the higher
anodization of the working electrode. In the range between 0 and 1.5 V the
slopes of the current density or the potential curves are decreasing. This effect
may be attributed to the fall out of metal oxides/hydroxides in the rather small
capillary volume. In the backward sweep current density immediately breaks
down due to the metal oxides/hydroxides and thus choking the capillary the
capillary mouth and interrupt the current flow. All of these observations however
Chapter 3 – Results and Discussion
48
lead to scoping the very beginning of the measurable current activity where the
substrate is negatively polarized. Looking closely at Fig. 3.11 one could assume
that the dashed lines indicating grain boundary measurements start a little earlier
than the solid lines indicating the flat grains. This observation becomes more
visible when applying statistics to the obtained curves (see Fig. 3.12).
Fig. 3.12: Averaged Current density/Potential plots measured on grain boundaries (bold,
dashed line), with the standard deviation (hatched area) and on grains (bold, solid line),
with the standard deviation (grey area).
Fig. 3.12 represents the average curves from the measurements taken. The chart
shows only a small range of the potential sweep where the starting point of
measurable current activity is of interest in order to distinguish the susceptibility
to metal dissolution or corrosion processes. In this survey it was found that the
starting point of measurable current activity on grain boundary spots is on a lower
potential than that of flat grain spots, where the difference can be accounted to
approximately 80 mV. These findings indicate that anodic dissolution and
oxidation occur more easily on grain boundaries. It should also be considered
that grain boundaries with an estimated width of 1µm and measuring on a triple
point would contribute 30 µm² to the measured surface of 314 µm² when
Substrate characterization
49
calculated by the capillary geometry (see Fig. 3.13). The grain boundary
contribution can be calculated to a maximum of 10 % because of the rather large
measured surface area of the micro capillary with an inner diameter of 20 µm.
Fig. 3.13: Illustration of grain boundary contribution to the measured area by
capillary diameter of 20 µm.
It is assumable that the potential difference when measuring just the grain
boundaries could be even higher, however due to capillary limitations it was not
possible to provide more precise measurements in this survey. From this point of
view the results obtained by the micro capillary cell again provide strong
evidence that grain boundaries are highly corrosive active species. One has to
consider that this activity can be due to the edge geometry, the alloy composition
and the crystallographic stability of the grain boundary area. This anodic
dissolution activity of grain boundaries is in analogy to the lower electrochemical
potential of the grain boundaries obtained from the SKP-FM measurements. In
the following section 3.2: Material Survey for Grain Boundary Application, this
dissolution activity of grain boundaries will be investigated in order to control the
selective material creation and deposition on grain boundaries for selective
inhibition of these corrosive sites. But before focusing on the material for grain
boundary corrosion inhibition, the last part of this section will investigate the
corrosion behavior of grain boundaries in HDG steel substrate when coated and
exposed to a corrosive environment.
Chapter 3 – Results and Discussion
50
3.1.3.4 Corrosion behavior of grain boundaries on coated substrates
Strong evidence was obtained in the previous experiments that grain boundaries
are the weak zones susceptible to corrosive reactions. In the following, the
question of whether it can be proved that grain boundaries are the preferred
pathways for corrosion will be answered. For this reason, the behavior of grain
boundaries on coated substrate under corrosive conditions will be investigated.
To this end, a HDG steel substrate was coated with a commercial coil coating
primer. A defect in the coating was created through a scratch and the sample
was exposed in a salt spray chamber for 504 h. After exposure, the delaminated
coating was peeled off and the corrosion front was investigated by focusing on
the corrosion front (Fig.3.14). The obtained images always showed that corrosion
products could be found along the grain boundaries as a forefront of the corroded
area.
Fig. 3.14: a) SEM images and corresponding EDX mappings of the corrosion front of
coated HDG steel substrate. It shows the corrosion propagation along the grain
boundaries. b) Scratched sample after exposure in a salt spray chamber for 504 h
indicating the area where the SEM image was obtained.
Substrate characterization
51
The EDX mappings of Fig. 3.14 identify oxygen, chlorine and zinc, all of which
identify that corrosion products along the grain boundary consist of zinc oxides
and zinc chlorine. This type of prolongation of corrosion in the interface grain
boundary/coating could be found on all grain boundaries and ranged between 50
and 200 µm along the grain boundary when measured from the main corrosion
front. These findings correlate closely with the assumptions made in the
fundamentals of this chapter and the results obtained from the surface analytics.
At this point, some additional thinking would enable the influence of mechanical
adhesion between coating and substrate in terms of corrosion protection to be
evaluated. Whilst flat grain surfaces do not provide the appropriate geometry for
mechanical adhesion, the porous topography of grain boundaries provides the
appropriate geometry for force-fit interlock adhesion. The fast corrosion
propagation along the grain boundaries shows that mechanical adhesion is less
relevant when the substrate is electrochemically active. The interlocks
established from the coating lose their basis when the substrate
electrochemically dissolves. However from these observations a model for
corrosion pathways on this particular substrate will be derived and discussed in
the following part of this chapter.
3.1.3.5 Corrosion pathways as a model development
The general cross-sectional view of corrosive delamination (see chapter 1)
becomes more complicated on heterogeneous surfaces. Therefore to visualize
the corrosive delamination process based on different corrosion pathways, a top
view of the corrosion front of the substrate surface is more appropriate. Fig. 3.15
illustrates the basic mechanisms that can be derived from experimental findings
within this section. The SEM image shows a corroded grain boundary in the
forefront of the corrosion front. The detailed explanation of this postulated
corrosion process is described based on the schematic illustration of Fig. 3.16.
Chapter 3 – Results and Discussion
52
Fig. 3.15: a) SEM image and EDX mapping of the corrosion front and grain boundary.
The sample was coated with a conventional coil coating, scratched and exposed for 504
h in a salt spray chamber. After corrosion exposure the coating was removed with THF.
The sketches in the image indicate the corrosion pathways on the HDG steel sample. b)
Schematically illustrates the tested sample.
Single steps more or less simultaneously occur, but in order to understand the
single parts of the process it is necessary to divide the corrosion process into
sequential steps as shown in Fig. 3.16.
Substrate characterization
53
Fig. 3.16: A model for delamination and corrosion propagation on HDG steel with
active/non-protected grain boundaries and protected/disabled grain boundaries.
Grain boundaries of HDG steel were found to be electrochemically very active
and connected to each other as a grid of trenches in the surface of the substrate.
It was also found that grain boundaries preferably undergo anodical dissolution.
When a defect in the coating induces a corrosive attack to the substrate, it is
assumed that the grain boundaries will propagate at a higher rate, the corrosion
process under the intact regions of the coating. The anodic corrosion reaction
along the grain boundaries would create local cathodes in the surrounding of the
grain boundary area whilst moving forward. Oxygen would be reduced to some
Chapter 3 – Results and Discussion
54
radical species in that area and within water molecules to hydroxyl ions. As
described in the general introduction, both species are known to migrate along
the substrate/coating interface and delaminate the coating from the substrate.
Herein the corrosive delamination would occur on flat areas of the grains. The
free and unprotected substrate surface now excessively corrodes as indicated by
the plateau corrosion. In summary, there may be two corrosion pathways, a fast
one along the grain boundaries and a slower one as represented by the main
plateau corrosion. The corrosion rate of the second one, the plateau corrosion,
should be influenced by the first one, the grain boundary corrosion. According to
these assumptions further investigation should highlight the existence of two time
constants for the two corrosion pathways. Rohwerder et al. have applied the
SKP-FM technique to investigate the electrochemical aspects of delamination on
a model gold substrate [61]. Even though the propagation of corrosion activity
along the grain boundary was not discussed in their paper, the obtained images
clearly showed corrosive activity in the grain boundaries of the gold substrate.
However SKP-FM could provide the appropriate tool to investigate the different
corrosion rates of the two assumed corrosion pathway models on HDG steel
substrate as postulated within this work. One must consider that the preparation
of appropriate samples as well as the experimental set up is very advanced and
that the measurement is very time consuming.
In the general introduction to this study it is postulated that corrosion starts at
some weak spots, which can be edges or vertices. Based on the results
obtained, corrosion can also propagate along these weak zones, when they are
grain boundaries and connected with each other all over the substrate surface.
Learning from this model, one would assume that disabling grain boundary
activity would lead to a slow-down of the corrosion propagation, known as
plateau corrosion. The following sections of this study will therefore investigate
the possibility of inhibiting grain boundary corrosion activity through the
application of inhibiting material exclusively to the grain boundaries.
Substrate characterization
55
3.1.4 Conclusion
Investigations with high lateral resolution on surface element composition and
electrochemical micro probe techniques lead to the identification of the weak
zones of the HDG steel substrate. It could be found that aluminum even at low
concentrations of 0.5 w.-% segregates not only to the zinc/iron and iron/air
interface but also to the grain boundaries of the zinc grains. It was found that the
conventional alkaline cleaning process contrarily to most literature does not
remove all of the aluminum from the substrate surface. Even where the overall
aluminum content becomes negligible, it still remains within the grain boundaries.
Surface potential mappings with the high resolution of the SKP-FM technique
discovered the lower potential of the grain boundaries when compared to the
surrounding grain surfaces. Spots with lower potential than the surrounding
matrix are known to be more corrosively more active. The application of the micro
capillary cell showed that the grain boundaries tend to dissolve more easily due
to the corrosive currents that could be measured at a lower potential than on the
single grains. The final corrosion test in a salt spray chamber on a coil coated
and scratched sample showed the higher corrosion activity of the grain boundary.
Corrosion products along the grain boundaries beneath the intact coating
material could be observed after removing the coating. Collectively these findings
led to the development of a corrosion model for heterogeneous surfaces; where
the anodic part reaction is quickly propagated forward along the grain boundaries
and escorted by the local cathode which delaminates the grains. The plateau
corrosion can follow more easily along the delaminated grains. The derivation
from this model would be a slowdown of the corrosion propagation through
disablement of the grain boundary activity. The proof of this model would lead to
new smart coating materials that treated only the weak zones of the substrate
which are susceptible to corrosion. The current practice of pretreating the entire
substrate surface with the same material would become obsolete and thus save
the pretreatment material. This will be the focus in the following chapters.
Chapter 3 – Results and Discussion
56
3.1.5 References
[1] P.Thissen, J.Wielant, M.Köyer, S.Toews, G. Grundmeier, Surf. Coatings
Tech., submitted
[1] A.Rodnyansky, Y.J. Warburton, L. D. Hanke, Surf. Interface Anal. 29 (2000)
215
[3] A.R. Marder, Prog. Mat. Sci. 45 (2000) 191
[4] V. Furdanowicz, C.R. Shastry, Metall. Trans. A 30 (12) (1999) 3031
[5] G.A. Lopez, Dissertation „Segregation and Phase Transformations at
Interfaces“ Universität Stuttgart, Max-Planck-Institut für Metallforschung Stuttgart
(2004)
[6] A. F. Holleman, E. Wieberg, Lehrbuch der Anorganischen Chemie, 101
Auflage, deGruyter, Berlin, (1995)
[7] F. Weinberg, M.Mager, L.Frederick, Can. Met. Quart. 29 (1990)163
[8] N. Fink, B. Wilson, G. Grundmeier, Electrochim. Acta 51 (2006) 2956
[9] J. Strutzenberger, J. Faderl, Metall. Trans. A 29 (2) (1998) 631
[10] S. Feliu Jr., V. Barranco, Acta Mater. 51 (2003) 5413
[11] L.A. Rocha, M.A. Barbosa, Corrosion 47 (7) (1991) 536
[12] A.Rodnyansky, Y.J. Warburton, L. D. Hanke, Surf. Interface Anal. 29 (2000)
215
[13] J. Ruf, „Organischer Metallschutz“ Vincentz, Hannover (1993)
[14] U. Koenig, J.W. Schultze, Galvanotechnik 98 (2007)1834
[15] D. Zimmermann, A.G. Munoz, J.W. Schultze, Surf. Coat. Tech. 197(2005)
260
[16] A.G. Munoz, J.W. Schultze, Electrochim. Acta 49 (2004) 293
[17] E. Klusmann, J.W. Schultze, Electrochim. Acta 48 (2003) 3325
[18] D. Zimmermann, A.G. Munoz, J.W. Schultze, Electrochim. Acta 48 (2003)
3267
[19] A. Losch,J.W. Schultze, J. Electroanal. Chem. 359 (1993) 39
[20] R. Berger, U. Bexell, N. Stavlid, T.M. Grehk, Surf. Interf. Anal. 38 (2006)
1130
[21] E. Almeida, I. Alves, C. Brites, L. Fedrizzi, Prog. Org. Coat. 46 (2003) 8
[22] N. Birbilis, R. G. Buchheit, J. Electrochem. Soc., 152 (2005) B140
Substrate characterization
57
[23] G. O. Ilevbare, O. Schneider, R. G. Kelly, J. R. Scully, J. Electrochem. Soc.
151 (2004) B453
[24] M. Büchler, T. Watari, W. H. Smyrl, Corros. Sci. 42 (2000) 1661
[25] O. Schneider, G. O. Ilevbare, R. G. Kelly, J. R. Scully, J. Electrochem.
Soc.,151 (2004) B465
[26] G. O. Ilevbare, J. R. Scully, J. Electrochem. Soc., 148 (2001) B196
[27] J. R. Scully, T. O. Knight, R. G. Buchheit, D. E. Peebles, Corros. Sci., 35
(1993) 185
[28] R. G. Buchheit, J. Electrochem. Soc., 142 (1995) 3994
[29] R. P. Wei, C.-M. Liao, M. Gao, Metall. Mater. Trans. A, 29 (1998) 1153
[30] N. Birbilis, R.G. Buchheit, J. Electrochem. Soc. 155 (2008) C117
[31] G. Ibe, K. Lucke, Am. Soc. Met. (1966) 434
[32] M. Shimada, H. Kokawa, Z.J. Wang, Y.S. Sato, I. Karibe, Acta Mater.50
(2002) 2331
[33] H. Liu, M. Gao, D. Harlow, R. Wei, J. Scr. Metall. Mater. 32 (1995) 1807
[34] V.B. Trinidade, U. Krupp, Ph. E.-G. Wagenhuner, H.-J. Christ, Mater. Corr.
56 (2005) 785
[35] E. Trillo, L. Murr, J. Mater. Sci., 33 (1998) 1263
[36] E.P. Bulter, M.G. Burke, Acta. Metal1., 34 (1986) 557
[37] M.S. Laws, P.J. Goodhew, Acta Metal. Mater. 39 (1991) 1525
[38] R. Bredesen, P. Kofstad, Oxidation of Metals 34 (1990) 361
[39] R. J. Hussey, M. Cohen, Corrosion Science 11 (1971) 713
[40] Interview R&D Group, BASF Coatings GmbH, Münster
[41] G.A. Lopez, E.J. Mittemeijer, B.B. Straumal, Acta Mater. 52 (2004) 4537
[42] B. Straumal, R. Valiev, O. Kogtenkova, P. Zieba, T. Czeppe, E. Bielanska,
M. Faryna, Acta Mater. 56 (2008)6123
[43] F.M. Ferry, K. Kjoller, J.T. Thornton, R.J. Tench, D.Cook, Electric Force
Microscopy, Surface Potential Imaging, and Surface Electric Modofocation with
Atomic Force Microscopy (AFM), Digital Instrument, Veeco Metrology Group,
Application Note 27 (1999)
[44] M. Stratmann, H. Streckel, Corr. Sci. 30 (1990) 681
[45] G. Binning, C.F. Quate, C. Gerber, Phys. Rev. Lett. 56 (1986) 930
[46] Q. Zhong, D. Inniss, K. Kjoller, V.B. Ellings, Surf. Sci. 290 (1993) L688
[47] A.P. Quist, J. Ahlbom, C.T. Reimann, B.U.R. Sandqvist, Nucl. Instr. Meth.
B88 (1994) 164
Chapter 3 – Results and Discussion
58
[48] P. Schmutz, G. Frankel, J. Electrochem. Soc. 145 (1998) 2285
[49] H.O. Jacobs, H.F. Knapp, S. Müller, A. Stemmer, Ultramicroscopy 69 (1997)
39
[50] M. Nonnenmacher, M.P. O’Boyle, H.K. Wickramasinghe, Appl. Phys. Lett. 58
(1991) 2921
[51] S. Yee, A. Oriani, M. Stratmann, J. Electrochem. Soc. 138 (1991) 55
[52] M. Rohwerder, F. Turcu, Electrochim. Acta. 53 (2007) 290
[53] B.S. Tanem, O. Lunder, A. Borg, J. Mardalen, Int. J. Adhes. Adhes. 29
(2009) 471
[54] F. Andreatta, M.M. Lohrengel, H. Terryn, J.H.W. de Wit, Electrochim. Acta 48
(2003) 3239
[55] F. Andreatta, A. Turco, I. de Graeve, H. Terryn, J.H.W de Wit, L. Fedrizzi,
Surf. Coat. Tech. 201 (2007) 7668
[56] A. Schreiber, C. Rosenkranz, M.M. Lohrengel, Electrochim. Acta 52 (2007)
7738
[57] K. Fushimi, S. Yamamoto, R. Ozaki, H. Habazaki, Electrochim. Acta 53
(2008) 2529
[58] C.J. Park, M.M. Lohrengel, T. Hamelmann, M. Pilaski, H.S. Kwon,
Electrochim. Acta 47 (2002) 3395
[59] K.A. Lill, A.W. Hassel, G. Frommeyer, M. Stratmann, Electrochim. Acta 51
(2005) 978
[60] A. Schreiber, C. Rosenkranz, M.M. Lohrengel Electrochim. Acta 52 (2007)
7738
[61] M. Rohwerder, E. Hornung, M. Stratmann, Electrochim. Acta 48 (2003) 1235
59
3.2 Material survey for grain boundary application
3.2.1 Fundamentals
The following section of this work will deal with the major question of what
material is appropriate to disable the corrosion activity of grain boundaries and
how to focus it on the identified weak zones. In section 3.1 the substrate surface
was characterized and it was found that grain boundaries are electrochemically
very active. As could be seen in the micro capillary cell measurements, they have
a strong drive towards anodical dissolution. It is assumed that the same strong
drive of grain boundaries to corrosion can be used as a trigger for the selective
application of inhibiting materials on grain boundaries. Promising known
materials from literature and industrial use with the potential for adaptation to an
exclusively grain boundary application were investigated and are discussed
within this chapter. Phosphating, surface polymerization and polymer deposition
were found to be the most promising for further evaluation. All three of these
techniques deposit or create the corrosion inhibiting material on the substrate
surface by partly dissolving metal ions from the substrate surface. However all of
these techniques cover the entire substrate surface as applied. This survey will
evaluate the controllability of addressing the material deposition or creation on
grain boundaries of:
• Phosphating;
• surface spontaneous polymerization; and
• polymer deposition.
Chapter 3 – Results and Discussions
60
3.2.1.1 Phosphating
The phosphating of metals for corrosion protection has been an established
technique since the 1960s [1-5]. This technique uses the anodic dissolution of the
surface metal e.g. zinc and the subsequent phosphate crystal growth on the
substrate surface controlled by the pH. The main components of phosphating
solutions are diluted phosphoric acid and zinc cations. State of the art
phosphating technology also uses nickel and manganese cations [6]. In a first
step, a pickling attack from the phosphoric acid (pH 2.5 – 3.5) on the substrate
metal consumes protons and zinc cations are released. This leads to a pH-
gradient and zinc cation gradient build up. The increased cation concentration
and the higher pH on the substrate surface shift the protolytic equilibrium of the
phosphoric acid to the phosphate anion which results in precipitation of hopeite
caused by its extremely low solubility. The precipitation of hopeite is basically
controlled by the amount of phosphate anions in the phosphating solution [7].
However this technique is always applied on the entire substrate surface. When
applying to HDG steel, the substrate surface has to first be activated by an
alkaline etching process (see section 3.1). Alkaline etching removes the native
aluminum oxide / hydroxide passive layer [8,9]. The remaining bare zinc surface
can be easily dissolved as it is needed in order to create a phosphate layer. In
the experimental results of this chapter it will be shown that exposure of HDG
steel to acidic solutions in the pH range of 3.6 to 4.0 dissolve only the grain
boundaries. The flat grain areas remain non-etched. This can be attributed to the
passive aluminum oxide / hydroxide layer and the flat topography of the grains.
The topography of grain boundaries is very rough with vertices and edges. The
high solubility of grain boundaries can be attributed to the high accessibility and
loose structure of the metal atoms. In this part of the study, the high solubility of
grain boundaries will be used to create phosphate crystals selectively on the
grain boundaries. Aside from phosphoric acid, vinylphosphonic acid (VPA) will
also be investigated in the phosphating process modified for the grain boundary.
VPA is attractive for such processes due to its polymerizable double bond.
Shannon et al. have investigated copper and zinc precipitates of VPA and its
radical polymerization. Their purpose was to immobilize the catalytically metal
atoms in order to use the newly created material as a heterogeneous catalyst
Material survey for grain boundary application
61
[10-12]. In this study the aim will be to use the reactive double bond in order to
react it with the UV coating system applied on top. The main idea is to establish a
covalent bond between the grain boundary applied phosphonate layer and the
top coating material.
3.2.1.2 Surface spontaneous polymerization
In the late 1990s Zhang, Agarwal and Bell developed an environmentally friendly
polymeric coating process for aluminum [13,14] and steel [15-17] substrates
through a surface spontaneous polymerization. The coating is directly
synthesized on the metal surface by simply dipping the metal sample into the
solution of monomers for a few seconds. The mechanism of spontaneous
polymerization is based on the high copolymerization reactivity of donor acceptor
monomers and has been studied by several researchers [18,19]. A monomer
with a negatively polarized double bond (donor) is able to combine to a donor-
acceptor complex with a positive polarized double bond of a monomer (acceptor)
under certain conditions [20,21]. Therefore spontaneous copolymerization can
take place only if there is a substantial polarity difference between the reacting
monomers. An addition of Lewis acids to the system then destabilizes the donor-
acceptor complex and spontaneous polymer chain propagation starts. Bell et al.
have investigated the surface spontaneous polymerization in a simple system of
styrene (St) and n-phenylmaleimide (NPMI). The electron donating character of
styrene comes from the presence of the phenyl ring next to the C=C bond while
the two carbonyl groups withdraw electrons from the double bond of NPMI. The
driving force for spontaneous polymerization in that system comes from the
aluminum surface, where Lewis acids such as Al
3+
-cations are generated when a
sample of aluminum is immersed in an acidic solution. The aluminum cation is
classified as a hard Lewis acid and therefore has a strong affinity to electron pair
acceptance (see section 1.4). Learning gained from the literature survey makes
the surface spontaneous polymerization very interesting when applying
selectively to grain boundaries. In section 3.1: Substrate Characterization, it was
shown that grain boundaries are enriched with aluminum. Therefore the
controlled release of aluminum cations and their ability to initiate the spontaneous
Chapter 3 – Results and Discussions
62
polymerization directly on grain boundaries of HDG steel will be investigated
within this chapter.
3.2.1.3 Polymer deposition
Among a broad variety of application methods for polymers, dip coating is the
most promising process for selective application on micro structures such as
grain boundaries. The two major dip coating applications for water based polymer
dispersions are driven by electrophoresis and autophoresis. Electrophoretic
coatings were introduced to the automotive industry in the 1960s. Nowadays
these primer systems are applied on almost all automotive bodies due to the
improved corrosion protection offered [22]. Autophoretic coatings were initially
invented by Amchem Products Inc. and became available in the US in the 1980s
[23-25]. Due to constant development throughout the last decades, these coating
systems found a broad application for corrosion protection on industrial goods
with less surface quality [26-33]. After Henkel AG & Co. KGaA acquired Amchem
Products Inc. in the late 1990s, autophoretic coatings also became available in
Europe [34-40]. In 2009 Henkel re-launched the autophoretic coating system
under the brand of Aquence© [41] Henkel’s coating system is environmentally
friendly due to its water based system containing no volatile organic compounds
(VOC). It reduces the application process steps because special substrate
pretreatment such as phosphating becomes redundant. There is also finally, no
electric current required in order to apply the polymeric coating material to a steel
part as it is in the electrophoretic coating system. The clear focus of Henkel is
entry into the automotive industry with its advanced technology. BASF Coatings
GmbH also recently patented some variations of autophoretic coatings systems
[42-47]. However, the principle of the autodeposition of the polymeric coating
material is based on the ability of the substrate to corrode. In this process, the
substrate to be coated is immersed in a stabilized latex bath containing a
hydrofluoric acid, oxygenated water and various other additives [48,49].
When the
substrate is immersed in the coating bath, the major chemical reaction that takes
place is the dissolution of the metal substrate surface. Subsequently the metal
cations generated on the substrate surface coagulate and precipitate the polymer
latex. The aggressive oxidants in the coating bath continue to penetrate the
Material survey for grain boundary application
63
deposited but still porous polymer particles and the metal oxidation and polymer
precipitation continues. The final coating thickness depends on the immersion
time, the solid content and the pH of the Bath [50,51]. The mechanism of the
autophoretic polymer deposition will be investigated for the exclusive application
to grain boundaries of HDG steel surfaces. Therefore the controlled release of
cation from the grain boundaries might also be of importance for the selective
application of polymer particles. A screening of a variety of water based polymer
dispersions will be provided within the following experiments.
3.2.2 Experimental procedure
3.2.2.1 Grain boundary dissolution
Grain boundary dissolution experiments were carried out by dipping the HDG
substrate in phosphoric acid solutions of different pH values and by varying the
dipping time.
3.2.2.2 Phosphating
Phosphoric Acid
The phosphating experiments were carried out in a 100 mL beaker. The starting
formulation and parameters for the phosphating bath were derived from Müller et
al [52]. A standard 100 mL phosphating bath contained 25 mmol phosphoric acid,
1.6 mmol zincoxide, and 100 mL of distilled water. The phosphating experiments
were carried out by varying the bath temperature, the pH and the dipping time of
the substrate.
Vinylphosphonic Acid
The phosphonating experiments with vinylphosphonic acid were carried out
following the same procedure of the experiments from phosphoric acid. A
Chapter 3 – Results and Discussions
64
standard formulation for the phosphating bath was derived from the phosphating
experiments with phosphoric acid and contained 14.7 mmol/100 mL of VPA, 9.2
mmol/100 mL of ZnO and 100 mL water.
3.2.2.3 Surface spontaneous polymerization
The formulation of the monomer solution and the conditions for carrying out the
surface spontaneous polymerization experiments were derived from Bell et al.
The experiments were carried out in a 100 mL flask. The following monomer
composition was dissolved in 21.5 g of water and 28.5 g n-methylpyrollidon
(NMP).
Monomer composition
n-phenylmaleimide (NMPI) 0.866 g (0.1 mol) acceptor
Styrol (St) 1.042 g (0.2 mol) donor
2-Methacryloxy-Ethylacetat (MEA) 1.071 g (0.1 mol) acceptor, adhesion
promoter
Bis-N-Methylmaleimid (BMI) 0.09 g (0.005 mol) acceptor, cross-linker
The polymerization experiments were carried out under an inert atmosphere. The
pH was varied within a range of 2.5 and 4.0.
3.2.2.4 Polymer deposition
A variety of exemplaric polymer dispersions were investigated in terms of their
selective deposition on grain boundaries of the HDG steel. The applications were
carried out by varying the dipping time and the pH.
Material survey for grain boundary application
65
3.2.3 Experimental results
3.2.3.1 Grain boundary dissolution
Grain boundaries were found to preferably undergo anodical dissolution (see
section 3.1). The ability of grain boundaries to corrode will be investigated in the
use of the selective application of corrosion inhibiting material onto grain
boundaries. It was shown that grain boundaries do not preferably dissolve when
exposed to alkaline solution of high pH values. Therefore the dissolution of grain
boundaries was investigated in acidic conditions as shown in Fig. 3.17.
Fig. 3.17: SEM images of grain boundaries. Top) grain boundaries selectively etched with
diluted phosphoric acid pH 3.6 for 30s. Bottom) natural grain boundaries on HDG steel
surface.
SEM images in Fig. 3.17 show dissolved grain boundaries in comparison to
natural grain boundaries of HDG steel. The substrate after etching in phosphoric
Chapter 3 – Results and Discussions
66
acid at pH 3.6 is shown in the top images. Selective grain boundary dissolution of
the grain boundaries could only be observed in the pH range of 3.6 to 4.0. Above
a pH of 4.0, no dissolution activity of the substrate was observed. When the pH
was lower than the value of 3.6, dissolution of grain surfaces also became visible
where the material is removed from the grain boundaries and they appear
excavated. Therefore the grain boundaries as shown in the bottom SEM images
of Fig. 3.17 should consist of an amorphous structure. The dissolution of the
grain boundaries releases cations that will be used for selective material
deposition on grain boundaries in the following part of this chapter.
3.2.3.2 Phosphating
Phosphoric Acid
In order to find the right parameters for grain boundary selective phosphating, the
precipitation parameter of the phosphating solution had to be investigated.
Therefore zinc phosphate precipitation was reviewed in relation to a dependency
of the pH and the temperature for different ion concentrations in the phosphating
bath (see Fig. 3.18).
Fig. 3.18: pH-Temperature dependency for zinc phosphate precipitation for a variety of
phosphating bath formulations.
Material survey for grain boundary application
67
Each of the phosphating bath formulations has an initial pH lower than 3.0, where
the phosphating solution appears as a clear liquid. At the specific temperature
the pH was raised slowly by the addition of 0.1 n sodium hydroxide solutions. The
pH value was transferred to the chart in Fig. 3.18 when visible precipitation
occurred. Overall the higher temperature resulted in a lower pH where
precipitation occurs, which can be attributed to the endothermic crystallization
behavior of zinc phosphate. A rise in the phosphate ion concentration from 25
mmol/100 mL to 50 mmol/100 mL to 100 mmol/100 mL results in a slightly
parallel shift of precipitating pH values to higher numbers. This is not surprising
because the phosphating bath is formulated with an excess of phosphate ions for
this very reason the zinc ion concentration can be used to control the
crystallization process. When varying the amount of zinc ions in the phosphating
bath, lower zinc ion concentration leads to more stability in the bath and a parallel
shift to higher pH values for precipitation. Higher zinc ion concentration led to
lower pH values for precipitation of zinc phosphate. These results are in line with
the phosphating process, where the dissolution of the outermost metal atoms of
the substrate consumes H
+
ions and releases Zn
2+
ions, resulting in a rise in the
pH and zinc. Within this section however it was observed that selective
dissolution of grain boundaries occurred in the pH range of 3.6 to 4.0.
The appropriate pH range for selective phosphating is assumed to also be in this
range. The release of cations from the grain boundaries should raise the ion
concentration locally. The pH would also rise in the grain boundary area due to
the reduction of protons. Both should lead to zinc phosphate crystal growth.
The experiments were carried out within these parameters. In addition, the
substrate exposure time in the phosphating bath varied from 10 to 240 s. The pH
of each phosphating bath was adjusted to 0.2 pH lower than the precipitation pH
value in Fig. 3.18. In phosphating experiments at pH values above 3.5, very few
crystals could be obtained on the substrate surface. These few crystals also
provided no specific selectivity to grain boundaries. In industrial applications, bath
parameters are adjusted to 65°C and pH 3.4. Under these conditions the
substrate surfaces are covered with zinc phosphate crystals within 60 s but this is
also very rare. This example is shown in Fig. 3.19.
Chapter 3 – Results and Discussions
68
Fig. 3.19: SEM image of zinc phosphated HDG steel in a standard zinc phosphating bath
(H
3
PO
4
25 mmol/100 mL; ZnO 1.6 mmol/100 mL) at 60°C and phosphating time of 60 s.
According to Fig. 3.19 there is no specific selectivity that can be observed
towards grain boundaries. The reason for these poor phosphating results can be
attributed to the aluminum passive layer that is inhibiting the pickling process
(see section 3.1). In addition dissolved aluminum is inhibiting the zinc phosphate
formation and precipitation due to the higher solubility product of
aluminumphosphates [23]. Therefore aside from removing the aluminum layer
from the substrate surface, in industrial application fluorine salts are always
added in order to capture remaining aluminum ions from the substrate. The high
aluminum content of the grain boundaries would therefore be counterproductive
to the selective grain boundary phosphating approach. The next step towards the
aim of selective inhibition of the corrosion active grain boundaries was to first
etch the grain boundaries and then apply the phosphating procedure as the
second step. The HDG steel substrate was first therefore exposed to an acidic
solution of pH 3.6 for 30 s. This procedure resulted in almost aluminum free grain
boundaries (see Fig. 3.17). Because in the phosphating results there was still no
improvement in grain boundary selectivity, an activator (Fixodine from Henkel AG
Material survey for grain boundary application
69
& Co. KGaA) was introduced to the phosphating process to act as seed crystals.
The size of the crystals is less than 5 nm and could not be resolved with
scanning electron microscopy. After etching the grain boundaries a Fixodine
solution of 5 g/L as recommended from Henkel was rinsed over the substrate
followed by a rinse with distilled water. The aim of this procedure was to deposit
the seed crystals within the grain boundaries. The following phosphating process
led to enhanced crystal growth starting from grain boundaries. Fig. 3.20 shows a
zinc phosphate HDG steel surface when etching the grain boundaries, depositing
seed crystals and choosing the application parameter in the pH range of 3.6 to
4.0. Similar results could be obtained for different concentrations and
temperatures.
Fig. 3.20: SEM image of zinc phosphated HDG steel in a standard zinc phosphating bath
(H
3
PO
4
25 mmol/100 mL; ZnO 1.6 mmol/100 mL) at 30°C and phosphating time of 70 s
after etching the grain boundaries rinsing with seed crystals.
However zinc phosphate crystals were obtained on grain boundaries as well as
on the grains. Not all grain boundaries were covered with crystals and the crystal
density along the grain boundaries was poor. Satisfactory selectivity towards
grain boundaries could thus not be achieved. When reducing the dipping time,
Chapter 3 – Results and Discussions
70
the dipping tendency for grain boundary selective phosphating also dropped. The
higher the exposure time of the substrate to the phosphate bath, the more
crystals were found on the grains.
Vinylphosphonic Acid
Grain boundary selective phosphating experiments with VPA were carried out in
analogy to the experiments with phosphoric acid. The best results could be
obtained with a phosphating bath formulation of 14.7 mmol/100 mL VPA, 9.2
mmol/100 mL ZnO and 100 mL water at pH of 3.5, temperature of 60°C and a
dipping time of 60s. To achieve this result the substrate was cleaned and grain
boundaries were etched and then rinsed with seed crystals as found in the
previously discussed phosphating experiment. The best result is shown in Fig.
3.21.
Fig. 3.21: SEM image of zinc phosphonated HDG steel in a standard zinc phosphonating
bath (VPA 14.7 mmol/100 mL; ZnO 9.2 mmol/100 mL) at 40°C and phosphating time of
60 s after etching the grain boundaries rinsing with seed crystals.
Material survey for grain boundary application
71
Replacing phosphoric acid by vinylphosphonic acid led to similar results in terms
of the selective deposition of corrosion inhibiting materials on grain boundaries.
The crystal density on the grain boundaries was again found to be rather poor.
Driving the bath parameters towards higher phosphating activity always led to
crystal growth all over the grain surface and thus not improving the crystal
density in the grain boundaries. In summary, selectivity towards grain boundaries
could only be partly achieved with the support of the activating seed crystals. The
process of etching and depositing seed crystals is inconvenient for industrial
applications. A truly innovative process should be applicable through a one-step
dipping process that takes only a few seconds.
3.2.3.3 Surface spontaneous polymerization
Spontaneous polymerization was attempted in order to adapt the grain
boundaries of HDG steel. For this reason the high aluminum presence in the
grain boundaries of the HDG steel surface was most promising. The procedure
and monomer composition was derived from Bell et al. By varying the pH of the
monomer solution the study tried to control the polymerization towards the grain
boundaries. Results on surface spontaneous polymerization on HDG steel are
shown in Fig. 3.22.
Fig. 3.22: SEM images of polymer on HDG steel surface obtained through spontaneous
polymerization at a pH value of 3.0 and polymerization time of 10 min.
Within the pH range of 3.6 to 4.0, where grain boundaries selectively dissolve,
polymerization on the HDG steel surface could not be observed; only when the
Chapter 3 – Results and Discussions
72
pH was lowered to the value of 3.0 for an exposure time of 10 minutes, could a
polymer be found on the substrate surface as is shown in Fig. 3.22. The high
resolution SEM image shows that along grain boundaries especially, the polymer
is less visible than on the flat grain surface. Within these experiments the
spontaneous surface polymerization could be observed on HDG steel substrates.
The polymer creation on HDG was very slow in comparison to the polymerization
on aluminum substrates as observed by Bell et al. However selective
polymerization on grain boundaries failed in the provided experiments.
3.2.3.4 Polymer deposition
A number of exemplaric water borne polymers were investigated in terms of their
applicability to grain boundaries of HDG steel. In order to apply the polymer
dispersion they were diluted to a solid content of 1.0 w.-%. It was considered for
the polymer application that a simple dip process would be the most convenient
in an industrial application process on fast moving steel strips in coil production.
Due to the very specific and small structure of the grain boundaries, only an
autophoretic coating process can provide a selective deposition of the polymer
on the desired substrate areas. The application process was carried out by
varying the pH and the dipping time. These most satisfying results could be
obtained from the primary anionically stabilized dispersions. Acronal 250D was
found to be stable in the entire pH range from 1.0 to 14. The best application
results could be obtained with a pH of 2.0 and a dipping time of 15 s (see Fig.
3.23).
Fig. 3.23: SEM image of Acronal 250D a primary anionically stabilized dispersion from
BASF SE applied to HDG steel substrate at the pH of 2.0 and a dipping time of 15 s.
(Polymer particles appear bright in the images.)
Material survey for grain boundary application
73
At lower pH levels or higher dipping times, the selectivity to the grain boundaries
was decreasing and more of the polymer was found on the entire substrate
surface. At a higher pH, fewer polymers could be deposited on the grain
boundaries even at a longer dipping time.
The next dispersion of interest to this study was Aquence from Henkel AG & Co.
KGaA which is developed especially for the autophoretic coating processes and
is based on an anionic stabilized primary dispersion. In the commercial
formulation Aquence contains highly oxidizing acids, fluorides and peroxides in
order to dissolve the outermost metallic layer so that the metal cations coagulate
and precipitate the polymer on the surface. However this commercially available
product was diluted to 1.0 w.-% prior to the experiments. In the corresponding
image, the surface coverage with the polymer after just one second of dipping is
shown in Fig. 3.24. Unfortunately, the polymer can be found on the entire surface
which can be attributed to the high reactivity of the oxidizing ingredients in the
commercial formulation. A longer dipping process resulted in coverage of the
entire surface with a film build up to 20 µm (30 s). Therefore Aquence as it is
commercially formulated cannot be selectively applied on grain boundaries in line
with the aim of this study.
Fig. 3.24: SEM image of Aquence from Henkel AG & Co. KGaA. A primary anionically
stabilized dispersion specifically developed and formulated for autophoretic coating
deposition. Applied to HDG steel substrate at the pH of 3.0 (as obtained through diluting
to 1.0 w.-%) and a dipping time of 1 s. (Polymer particles appear dark in the left image)
In the next experiment, a cationic secondary dispersion which can be found in a
typical electro coat binder was investigated. This dispersion was found to be
stable only in a small pH range from 4.0 to 8.0. Within this pH range no polymer
Chapter 3 – Results and Discussions
74
could be autophoreticaly applied to the substrate. The results shown in the
corresponding image (Fig. 3.25) were obtained at a pH of 4.0 with a dipping time
of 30s. The polymer on the substrate is applied non-selectively all over the
substrate. The coagulated polymer particles were obviously precipitated due to
the disability of the dispersion at a low pH. There is no evidence that the cation
release from the substrate surface triggered the precipitation. This experiment
shows that the dispersion must be stable in the pH range of the controlled
release of cations from the surface.
Fig. 3.25: SEM image of E-Coat Binder a secondary cationically stabilized dispersion
from BASF Coatings GmbH. Applied to HDG steel substrate at the pH of 4.0 and a
dipping time of 30 s. (Polymer particles appear dark in the left image and bright in the
right image)
The following polymer is a polyurethane dispersion which is a typical anionic
stabilized secondary dispersion and behaves in a similar way to the E-Coat
binder. The pH range where the dispersion is stable was found to be very narrow
between the pH of 5.0 to 8.0. Aside from this pH range, similar observations to
that of the image of the E-Coat binder can be made. The corresponding image
for the polyurethane dispersion in Fig. 3.26 shows the polymer application
experiment at a pH of 5.5 and a dipping time of 30 s. The image was taken at a
higher magnification to demonstrate that only very few polymer particles are
randomly distributed on the substrate surface with no specific selectivity for grain
boundaries.
Material survey for grain boundary application
75
Fig. 3.26: SEM image of Polyurethane Dispersion a secondary dispersion developed for
Coil Coating applications. Applied to HDG steel substrate at the pH of 5.5 and a dipping
time of 30 s. (Polymer particles appear bright in the right image)
The last polymer considered for the application screening process was a block-
co-polymer that is semi soluble in water. An explanation for the semi solubility of
the polymer in water may be derived from its highly hydrophilic block which
dominates the solubility in water. The polymer was synthesized by the RAFT
technique in solvent and dissolved afterwards in water. It shows some particle
structure but also a dominant behavior of a solubilized polymer e.g. viscosity. The
stability of this polymer could be observed on a broad pH range from 2.0 to 10.
But the application in a dipping process always showed smudgy covered surface
areas without any selectivity to the specific HDG surface characteristics (see Fig.
3.27).
Fig. 3.27: SEM image of a Block-co-Polymer semi soluable in water from Rhodia Co.
Developed for anti-corrosion primer application. Applied to HDG steel substrate at the pH
of 3.0 and a dipping time of 30 s. (Polymer appears dark images)
Chapter 3 – Results and Discussions
76
These results may be attributed to the high solubility of the polymer in water. For
this reason, the polymers in this dispersion exist as particles as well as solute
polymer chains. When some of the particles become destabilized during
application, the precipitating polymer particles drag more polymers with them.
This results in more polymers being deposited on the surrounding substrate area
but with less precision. Therefore sharply dispersed polymer particles provide
more precise deposition on grain boundaries, on which the focus should be when
considering the polymer design for grain boundary application.
3.2.4 Conclusion
The experiments within this section were dealing with two major questions, how
to apply corrosion inhibiting material exclusively on the grain boundaries and
what that material should be. In section 3.1 it was found that grain boundaries are
the weak spots of the HDG steel substrate and that they are highly
electrochemically active, which is attributed to their high aluminum content and to
their rough topography. It was also found that grain boundaries can be dissolved
in acidic solutions. Therefore the application of choice would be to use the
anodical solubility of grain boundaries for the application of the corrosion
inhibiting material. The material investigated for grain boundary selective
deposition was either inorganic precipitates (phosphates) or polymeric material.
Phosphating with phosphoric acid and then right away with vinylphosphonic acid
on the grain boundaries was not successful. It is assumed that the aluminum is
inhibiting the phosphate crystal growth. This is one of the major reasons that
aluminum is removed from the substrate surface in industrial phosphating
applications through alkaline etching. Only when selectively etching the grain
boundaries and applying the seed crystals (Fixodine) first, could some crystal
growth along grain boundaries be achieved. But the poor selectivity and crystal
density along grain boundaries excluded this approach from further
investigations.
Spontaneous surface polymerization can be triggered by strong Lewis acids such
as aluminum cations. It was assumed that the aluminum cations released from
the grain boundaries could initiate the spontaneous polymerization at the grain
boundaries. Unfortunately the polymerization was not very selective. One could
Material survey for grain boundary application
77
observe even fewer polymers on grain boundaries than on the grains
themselves. From these findings, spontaneous polymerization was evaluated as
a technique with low potential for grain boundary polymerization on the specific
substrate.
The best results could be obtained from the autophoritac deposition of
dispersions. Among the screened dispersions, the primary anionically stabilized
ones were the most stable at low pH values which they need for the release of
cations from the grain boundaries. Acronal 250 D was found to especially provide
some initial selectivity towards the grain boundaries. Henkel’s Aquence is based
on similarly polymer particles but is formulated for application on the entire
surface and can therefore not be used for selective grain boundary application.
As a result of this survey, the approach towards inhibited grain boundary of HDG
steel will be focused on local autophoresis of water borne polymer particles. In
the following section polymer dispersion with specific functionalities will be
synthesized and selectively deposited on grain boundaries.
3.2.5 References
[1] E.L. Ghali and R.J.A. Potvin, Corros. Sci., 12 (1972) 583 594.
[2] G.D. Cheever, J. Paint Technol., 39 (1967) 1 13.
[3] G. Lorin, Galvano-organo, Trait. Surf., 4 (1985) 385 387.
[4] J.B. Lakeman, D.R. Gabe and N.O.W. Richardson, Trans. lnst. Met. Finish.,
55 (1976) 47 53.
[5] B. Ptacek, F. Dalard, J.J. Rameau, Surf. Coat. Tech. 82 (1996) 277
[6] D. Zimmermann, A.G. Munoz, J.W. Schultze, Surf. Coat. Tech. 197 (2005)
260
[7] E. Klusmann, J.W. Schultze / Electrochimica Acta 48 (2003) 3325
[8] D. Zimmermann, A.G. Munoz, J.W. Schultze, Electrochim. Acta 48 (2003)
3267
[9] A. Losch,J.W. Schultze, J. Electroanal. Chem. 359 (1993) 39
[10] I.J. Shannon, New J. Chem., 26 (2002) 906
[11] A.D. Knight, J. Chem. Soc. Dalt. Trans., 6 (2002) 824
Chapter 3 – Results and Discussions
78
[12] I.J. Shannon, J. Mat. Chem., 12 (2002) 350
[13] R. Agarwal,J.P. Bell, Polym. Eng. Sci. 38 (1998) 299
[14] X. Zhang, R. Agarwal, J.P. Bell, Pat.: US 2003/0153704 A1 (2003)
[15] X. Zhang, J.P. Bell, J. Appl. Polym. Sci. 66 (1997) 1667
[16] X. Zhang, J.P. Bell, Mater. Sci. Eng. A, A257 (1998) 273
[17] X. Zhang, J.P. Bell, Polym. Eng. Sci. 39 (1999) 119
[18] Y. Shirota, H.J. Mikawa, Macromol. Sci. Rev. Macromol. Chem. C17 (1977)
129
[19] D.J.T. Hill, J.T. O’Donnell, P.W. O’Sullivan, Prog. Polym. Sci. 8 (1982) 215
[20] J. R Ebdon, Macromol. Chem. Macromol. Symp. 10 (1987) 441
[21] A.Saito, D. A. Tirrel, European Pdym J. 29 (1993) 343
[22] E. Almeida, Guide on anticorrosive protection in automotive industry, INETI,
Lisbon (200) 79
[23] W.S. Hall, H.M. Leister, US Patent 4,243,704 (6 January 1981), to Amchem
Products Inc.
[24] M. Chaker, US Patent 4,313,983 (2 February 1982), to Amchem Products
Inc.
[25] W.S. Hall, US Patent 4,318,944 (9 March 1982), to Amchem Products Inc.
[26] H.M. Leister, J.C. Donovan, W.S. Hall, US Patent 4,357,372 (2 November
1982), to Amchem Products Inc.
[27] W.S. Hall, US Patent 4,366,195 (28 December 1982), to Amchem Products
Inc.
[28] W.S. Hall, US Patent 4,414,350 (8 November 1983), to Amchem Products
Inc.
[29] K. Park, US Patent 4,457,956 (3 June 1984), to Union Carbide Co.
[30] B.M. Ahmed, US Patent 4,562,098 (31 December 1985), to Amchem
Products Inc.
[31] R.W. Broadbent, E.A. Stockbower, US Patent 4,632,851 (30 December
1986), to Amchem Products Inc.
[32] W.S. Hall, US Patent 4,637,839 (20 January 1987), to Amchem Products
Inc.
[33] K.C. Benton, Weinert Jr., J. Raymond, US Patent 4,657,788 (14 April 1987),
the Standard Oil Co.
[34] N. Shachat, US Patent 5,061,523 (29 October 1991), to Henkel Co.
[35] R.W. Broadbent, US Patent 5,080,937 (14 January 1992), to Henkel Co.
Material survey for grain boundary application
79
[36] B.M. Ahmed, R.W. Broadbent, US Patent 5,342,694 (30 August 1994), to
Henkel Co.
[37] W.S. Hall, US Patent 5,352,726 (4 October 1994), to Henkel Co.
[38] B.M. Ahmed, US Patent 5,385,758 (31 January 1995), to Henkel Co.
[39] O.E. Roberto, M.A. Maxim, US Patent 5,486,414 (23 January 1996), to
Henkel Co.
[40] B.M. Ahmed, R.M. Jayasuriya, T.R. Hopkins, US Patent 5,500,460 (19
March 1996), to Henkel Co.
[41] Henkel AG & Co. KGaA, Dürr AG, Metalfinishing 2 (2009) 67
[42] M. Dornbusch, DE 10 2005 023 728 A1 (23 May 2005) to BASF Coating
GmbH
[43] M. Dornbusch, A. Wiesmann, R. Wegner, DE 10 2005 023 729 A1 (23 May
2005) to BASF Coating GmbH
[44] M. Dornbusch, E. Austrup, H. Hintze-Bruening, DE 10 2006 053 291 A1 (13
November 2006) to BASF Coating GmbH
[45] M. Dornbusch, A. Wiesmann, DE 10 2006 053 292 A1 (13 November 2006)
to BASF Coating GmbH
[46] M. Dornbusch, A. Wiesmann, H. Hintze-Bruening, WO 2008/058586 A2 (22
May 2008) to BASF Coating GmbH
[47] M. Dornbusch, A. Wiesmann, WO 2008/058587 A1 (22 May 2008) to BASF
Coating GmbH
[48] E. Almeida, I. Alves, C. Brites, L. Fedrizzi, Prog. Org. Coat. 46 (2003) 8
[49] B. Pfeiffer, J.W. Schultze, J. Appl. Electrochem. 21 (1991) 877
[50] Zh. I. Bespalova, L.G. Miroshnichenko, Yu.A. Lovpache, Yu.D. Kudryavtsev,
Russ. J. Appl. Chem. 77 (2004) 1888
[51] S. Balova, M. Christov, Corr. Sci. 41 (1999) 1633
[52] N. Müller, Einfluss von Bad- und Substratkomponenten auf die Bildung von
Phosphatschichten, Dissertation, Universität Düsseldorf (2000)
[53] Interview BASF Coatings GmbH, Münster
Chapter 3 – Results and Discussions
80
81
3.3 Polymer design for grain boundary application
3.3.1 Fundamentals
3.3.1.1 Requirements and characteristics of the polymer of choice
In the previous section a survey on materials for selective application on grain
boundaries was provided. It was found that among different techniques polymer
particle deposition was the most promising in terms of disablement of the
corrosion activity of natural substrate defects such as grain boundaries and alloy
segregations. Therefore the need to find a polymeric coating material which is
highly selective for such surface heterogeneities is present. Considering the use
of this polymeric material as an environmentally friendly pretreatment, it must be
water borne without any volatile organic compounds (VOC) and it must in the first
instance, easily produced on an industrial scale. Polymers based in water can
exist in two main physical systems; in a molecular solution where the entire
hydrophilic polymer molecule is dissolved in water and in a physical system
where water borne polymers is a dispersion obtained either by emulsion
polymerization or by dispersing a solvent borne polymer solution into the water.
In the first case the monomers are emulsified with a surfactant in water into small
droplets. The polymerization then occurs within the droplets and the resulting
dispersion is stabilized by the amphiphilic surfactant. In this case the terminology
“latex” is frequently used, the most relevant of which are the styrene-co-polymers
established since the 1950s for use in coatings and adhesives. In the second
case, the polymer is synthesized in organic solvents and has preferably
integrated functionalities able to make the ionic solubilizing such as carboxylic
acids or amines. In the stripping process the organic solvent is removed from the
Chapter 3 – Results and Discussions
82
system and the simultaneous addition of water and the counterpart for the ionic
solubilizing results in a water dispersed polymer. Nevertheless in terms of
corrosion protection it seems to be a double play of the appropriate balance of
hydrophilic and hydrophobic properties within one polymeric molecule [1,2].
While the hydrophilic part of the polymer chain is interacting with the metal/oxide
surface, the hydrophobic part is establishing a barrier for oxygen, water and
electrolytes. Bringing in more hydrophilic functionality than is necessary for
covering and adhesion to the metal/oxide surface will lead to a higher
hydrophilicity within the polymer layer and therefore decrease the barrier
properties. The amphiphilic surfactant in dispersions also behaves in the same
way and can be counterproductive to corrosion protection.
In the material survey of the last chapter it was found that dispersions are more
precise in selective depositions on small structures such as grain boundaries
rather than soluble polymer systems. Primary anionically stabilized dispersions
also showed a better stability over a broad pH range, where the low pH regions
are especially of importance as the anodic dissolution of the grain boundaries on
HDG at a low pH will be used to precipitate the polymer particles. The
mechanism of the local autophoretical precipitation of polymer particles on grain
boundaries will be discussed in section 3.4. For the selective deposition the
synthesis of the polymer should allow the incorporation of functional groups that
strongly react with the dissolved cations from the grain boundaries and with the
specific surface chemistry of the substrate when deposited.
Based on these requirements and characteristics of the polymer of choice as
outlined in some literature, discussed synthetic approaches for the synthesis of
block-co-polymers will be evaluated towards primary, anionically stabilized and
surfactant free polymer dispersion in the following part of this section. This
synthesis route should allow a broad variation of functional monomers to be
incorporated in a fast and easy manufacturing process.
3.3.1.2 Synthesis of block-co-polymers
Functional block-co-polymers, due to the development of new polymerization
techniques and the preparation of a wide range of new polymeric materials have
Polymer design for grain boundary application
83
found in the last decade a broad field of applications such as surfactants,
lubricants, adhesives, additives, thermoplastic elastomers, as well as biomedical
and electronic applications [3-7]. When considering polymerization techniques
there are two major routes to such multifunctional polymers; the living anionic
polymerization and the controlled radical polymerization which in corresponding
literature is often described as the controlled living radical polymerization (CLRP)
[8] Living anionic polymerization offers high levels of control in terms of well-
defined polymers and precise molecular architectures, but the process is much
less flexible than radical polymerization as it is very sensitive to monomer
functionalities and impurities [9]. The difficulties in handling the anionic
polymerization process make it unusable for industrial production processes.
Radical polymerization on the other hand is of enormous industrial importance. It
is easy to handle, tolerant to impurities, compatible with water and a huge variety
of functional monomers and thus it can be implemented in an industrial plant.
Approximately 50% of all commercial polymers are produced by radical
polymerization. The major drawback of radical polymerization is that it is not
possible to prepare block-co-polymers or polymers of narrow molecular weight
distributions due to the high reactivity of the propagating radicals and their affinity
to undergo bimolecular termination, transfer and other side reactions. The lifetime
of a propagating radical is typically less than one second and chains are
continuously initiated throughout the polymerization [10]. The combination of the
robustness of the radical polymerization with the controllability offered by living
ionic polymerization was found in controlled radical polymerization (CRP). The
three most important ways to apply CRP as discussed in literature are the
nitroxide mediated radical polymerization (NMP) [11-13], the atom transfer radical
polymerization (ATRP) [15-17] and the reversible addition fragmentation chain
transfer polymerization (RAFT) [18-22]. Evaluating these techniques for the
stated anti-corrosion application purpose in the requirements and characteristics
for the polymer of choice these major three techniques fail due to provide
different reasons. The ideal block-co-polymer would be non-conventional, water
borne, easy to functionalize, surfactant free and it would be obtained through a
straight forward polymerization.
One of the major disadvantages of all three of these polymerization techniques is
their poor applicability in a straight forward polymerization in water. In all cases,
as known from classical emulsion polymerization, surfactants have to be used.
Chapter 3 – Results and Discussions
84
RAFT and ATRP agents are also sensitive to aqueous solution. RAFT is one of
the CRP techniques most described in literature and the number of papers
describing RAFT in aqueous dispersed systems is increasing rapidly. The
stability of the RAFT agent ‘dithiocarbonyl derivates’ in water is however poor
[23,24]. Depending on the experimental conditions the hydrolysis of the
dithiocarbonyl species may be significant as the rate of hydrolysis is dependent
on the pH parameters and temperature when in the aqueous phase [25,26]. In
the ATRP process, transition metal compounds mostly based on copper are used
as the control reagent. Most of the Cu-complex ligands are employed in bulk or
solution and are highly water soluble which leads to partitioning and deactivation
of the controlling agent [27-30]. Even though a series of hydrophobic ATRP
agents have been developed Cu(I) and Cu(II), reaction with water would occur
only with the loss of their controlling functionality [31].
However, a multitude of other CRP techniques exist, Bremser et al. for example
developed the DPE method using 1,1-diphenylethene as a control agent [32-34].
It was found that conventional radical polymerizations become controllable when
a small amount of 1,1-diphenylethene (DPE) is added. Even though the DPE
route does not provide the high control and livingness known from the typical
controlled radical polymerizations as stated above, it allows a build up of block-
co-polymers in the water phase without the use of any surfactants [35]. This
makes the DPE method a very interesting alternative to the other methods if the
block co-polymer formation only has the goal of controlling free radical
polymerizations. In this sense it is the only technique based entirely on
hydrocarbons, where no other ingredients such as halides, nitroxides or metal
ions are needed. Moreover, it can be easily adapted on an industrial scale
[36,37]. The mechanistic investigation of the block-co-polymer formation in the
presence of DPE was thoroughly investigated by Viala et al [38-40].
It was found
that the DPE route is basically a two-step procedure requiring in the first step, the
preparation of a precursor polymer in the presence of DPE. This precursor
polymer is then used as the active species in a second polymerization, where
block co-polymer formation takes place. The activity of the precursor polymer is
based on its unique semiquinoid structure where the α,p-dimer is formed by the
combined termination of two DPE-ended radical chains. The two-step procedure
can be carried out either as a one-pot reaction with consecutive monomer
Polymer design for grain boundary application
85
additions or spatially and timely separated. The two-step, one-pot procedure is
especially interesting from an economic perspective as it allows the adaptation of
the widely used semi-batch feeding procedures and it is practical for the
requirements of this study. So far the DPE route to functional block-co-polymers
is described only for a small variety of monomers such as Methyl Methacrylate
(MMA), Acrylic Acid (AA), Styrene (St) [38]
and Vinyl Acetate (VAc) [41]. In this
chapter the DPE technique will be introduced to a broader variety of functional
monomers such as Vinylphosphonic Acid (VPA), Triethoxyvilsilane (TEVS) and
Maleic Acid (MA) which are suited to all requested characteristics of the block-co-
polymer and its polymerization process as stated in section 3.3.1.1:
Requirements and Characteristics to the Polymer of Choice.
3.3.1.3 Mechanistic understanding of the block-co-polymer formation in the
presence of DPE
The mechanism for the DPE route to functional block-co-polymers in the following
is derived from Viala et al. As proposed, the synthesis is undergoing a two-step
process in a one-pot synthesis. In the first step the hydrophilic monomers
polymerize until the active end of a growing chain runs into the 1,1-
diphenylethylene molecule which then stabilizes the radical and stops the growth
of a chain while the radical remains alive as shown in Fig. 3.28.
Fig. 3.28: Trapping of a growing polymer chain by DPE.
The stabilized radial is inactive to participate in further polymerization reactions
and recombines with a second species of itself into a semiquinoid structure as
shown in Fig. 3.29. The pre-polymer consisting of two hydrophilic blocks is
forming a hetero phase system in the water. It can be isolated and reaches a
molecular weight of 1.000 g/mol to 3.000 g/mol.
Chapter 3 – Results and Discussions
86
Fig. 3.29: Stabilizing the radical by recombination of two DPE ended chains to a
semiquinoid structure.
For the one-pot synthesis the pre-polymer will be left in the reactor. After raising
the temperature to 90°C the monomers for the second, hydrophobic block are
added to the batch and then get polymerized by the residual initiator. The rise in
temperature destabilizes the semiquinoid structure of the pre-polymer. The newly
created and propagating hydrophobic polymer chain runs into the destabilized
semiquinoid structure of the pre-polymer and forms the block-co-polymer as
shown in Fig. 3.30.
Fig. 3.30: Formation of the block-co-polymer.
Polymer design for grain boundary application
87
Throughout the whole process there is no need for external emulsifiers. The
hydrophilic blocks of the block-co-polymers are intrinsically stabilizing the
dispersion particles in the water. An example of a resulting block-co-polymer
dispersion is shown by a SEM image in Fig. 3.31. The typical particle sizes
obtained ranged between 40 and 200 nm by varying the monomer compositions.
Fig. 3.31: SEM image of the block-co-polymer dispersion obtained via the DPE route
containing Vinylphosphonic Acid (VPA) and a scheme of the block-co-polymer
aggregated particle.
In this study the use of a broad variety of functional monomers such as
Vinylphosphonic Acid (VPA), Triethoxyvinylsilane (TEVS) and Maleic Acid (MA)
were successfully introduced to the block-co-polymer synthesis via the DPE
method. The monomers in this study were chosen as a result of their specific and
well known functionality to interact with inorganic surfaces such as zinc/zinc
oxide and aluminum/aluminum oxide. The screening of the adequate monomers
in terms of adhesion to these specific surfaces was carried out and reported in a
previous publication [42]. Molecules with anchoring groups such as phosphonic
acid, carboxylic acid or silane were found to be very good at adsorbing to metal
oxide surfaces. It is also very well-known through literature that these
functionalities form strong bonds with metal oxides and are used for example, to
produce self-assembled monolayers that protect steel from corrosion and
Chapter 3 – Results and Discussions
88
increase adherence of coatings or adhesive bonding [43]. Below, the results
obtained by the incorporation of these functionalities into block-co-polymers in
the aqueous phase will be reported.
3.3.2 Experimental procedures
3.3.2.1 Materials
All reagents were used without further purification. Methyl Methacrylic Acid
(MMA, 99%), Butylmethacrylate (BMA, 99%), Hydroxyethylmethacrylate (HEMA,
98%), 1,1-diphenylethylene (DPE, 97%), Ammonium Persulfate ((NH
4
)
2
S
2
O
8
,
98%), Maleic Acid (MA, 99%) and Triethoxyvinylsilane (TEVS, 97%) were
purchased from the Aldrich Chemical Co. Vinylphosphonic Acid (VPA, 97%) was
donated by BASF SE.
3.3.2.2 Synthesis of functional block-co-polymers in a hetero phase system
A two liter glass reactor equipped with an anchor stirrer was filled with 770g of
deionized water, 2.1g DPE and the monomers for the hydrophilic block of the
block-co-polymer 25g MMA and 25g of one of the anchoring groups containing
monomers MA or VPA. Under a flow of nitrogen the reactor was heated to 70°C
and held at that temperature for 60 minutes at which point the initiator solution of
3.1g (NH
4
)
2
S
2
O
8
in 67g H
2
O was dropped into the rector for further 30 minutes.
The reactor temperature was maintained at 70°C and raised to 90°C after 2
hours when the addition of the monomers for the hydrophobic block of the block-
co-polymer started. The solution of 250g BMA and 12.5g HEMA and in the case
of DPE-TEVS block-co-polymers Triethoxyvinylsilane was also added into the
reactor for a duration of 2 hours and the temperature of 90°C was kept for a
further 2 hours in order to complete the conversion of the monomers.
Polymer design for grain boundary application
89
3.3.3 Experimental results
3.3.3.1 Overview of synthesized block-co-polymers
Based on the results of the screening experiments for polymerizable monomers
and their ability to adsorb to metal oxide surfaces such as zinc oxide or aluminum
oxide in the preliminary of this study, a selection of these monomers were chosen
to be incorporated into the block-co-polymer dispersion within this section [42].
Variations of the functionalities in the block-co-polymer chains are shown in Fig.
3.32.
Fig. 3.32: Functionalities incorporated into the hydrophilic and hydrophobic region of the
block-co-polymer. Red (left) indicates the hydrophobic block, blue (right) indicates the
hydrophilic block.
Chapter 3 – Results and Discussions
90
Carboxylic acid is quite often described as a good adhesion promoter on
aluminum oxide surfaces [44,45]. This finding could also be obtained in the
previous screening experiments. In this study the carboxylic function will be
incorporated into the block-co-polymer by acrylic acid (AA) and maleic acid (MA).
It was also found that triethoxyvinylsilane is a good adsorbate. The last monomer
of choice with promising adhesion properties, as found in the adsorption
experiments is vinylphosphonic acid. Block-co-polymer dispersions with four
different functionalities could be obtained via the one-pot-two-step-DPE-route.
One must consider that each block of the polymer contains some basic monomer
compositions which remain the same through all variations. These are for the
hydrophilic block MMA and for the hydrophobic block HEMA and BMA. MMA was
copolymerized in the hydrophilic block in order to reduce or adjust the
hydrophilicity of the block-co-polymer in the way that the block-co-polymers are
able to form stable micelles. Creating the first block out of a hydrophilic monomer
such as acrylic acid only, will lead to a polymer solution in water in the first step
and to a fall out of the polymer in the second step. With this method no stable
micelles could be obtained. Triethoxyvinylsilane is barely soluble in water; it is
rather hydrophobic but still remains a good adhesion promoter to polar materials.
By establishing chemical bonding to such oxidic surfaces, it cleaves the organic
alcohol and obtains a polar anchoring group. Therefore TEVS could only
successfully be incorporated into the block-co-polymer in the second step where
the hydrophobic block is created. Due to stabilization issues in the first block
acrylic acid had to be added.
3.3.3.2 GPC and particle size observations
The molecular weight build up during the one-pot-two-step polymerization routine
was detected by GPC. Fig. 3.33 shows the molecular weight distribution after the
first block is created and after the block-co-polymer was formed with the example
of DPE-VPA(4.5%).
Polymer design for grain boundary application
91
Fig. 3.33: Molecular weight distribution of the hydrophilic di-block (dashed line) and of the
block-co-polymer (solid line) on the example of DPE-VPA(4.5%). The measurements
were obtained in THF at 25°C. The molecular weight was calculated in respect to narrow
polydispersity polystyrene standards.
The stepwise growth of the block-co-polymer was tracked by their molecular
weight. The smaller peak in Fig. 3.33 shows the molecular weight distribution of
the hydrophilic di-block obtained after the first step in the polymerization process.
According to the mechanism of the reaction process, one hydrophilic block is
connected to a second hydrophilic block via two DPE units also known as the
semiquinoid structure. This number would thus have to be divided in two in order
to estimate the molecular weight distribution for single hydrophilic blocks. The
second peal at the higher molecular weight values is obtained through the
complete block-co-polymer. Based on this procedure the molecular weights Mn,
Mw and its Polydispersity D, as well as the resulted dispersion particle sizes of
the synthesized block-co-polymer dispersions were analyzed (see Tab. 3.1). The
percentage indication in the labels of the polymer dispersions specify them by the
solid content of the functional monomer based on the solid polymer. E.g. in DPE-
AA(6.2%) from all monomer weights 6.2 % are acrylic acid. In the case of the
DPE-VPA the vinylphosphonic content varied from 1.3 to 8.2%.
Chapter 3 – Results and Discussions
92
Tab. 3.1: Molecular weights Mn, Mw, and its polydispersity D resulted from GPC
experiments. The measurements were carried out in THF at 25°C and calculated on
polystyrene with narrow polydispersity. The particle sizes are obtained via light scattering.
Based on the results shown in Tab. 3.1, molecular weights (Mn) for the
hydrophilic di-block range from 1.5x10
3
to 1,9x10
3
g/mol in all batches. When the
block-co-polymer synthesis is finished, molecular weights (Mn) rise to 1.8x10
4
to
7.0x10
4
g/mol. The polydispersities D (Mw/Mn) show rather narrow molecular
weight distributions in the pre-polymer obtained in the first step but quite broad
molecular weight distributions in the final block-co-polymer. The first block is
certainly a very short chain polymer with a high hydrophilicity. The second block
due to a great number of monomers must be polymerized and shows more
clearly the molecular weight distribution which is typical for the radical
polymerization process. These observations support the assumptions on radical
polymerization that were made in the literature survey of this chapter. But even
though the molecular weight is broadly scattered, the obtained particle sizes are
quite narrowly distributed, as shown in Fig. 3.31. At the same time the particle
sizes can vary on a wide range along the nanometer scale with different
monomer compositions. From Tab. 3.1 there is no correlation that can be derived
between the molecular weight and the final particle size of the dispersion. The
particle size variation obtained was from 40 to 230 nm. Reproductions of batches
always resulted in similar numbers for the particle sizes. These findings imply the
monomer composition and its ability to form and stabilize the particle micelle
must be responsible for the final particle size. In all cases it can be observed that
Polymer design for grain boundary application
93
the pre-polymer formed in step one has much higher particle sizes than the final
block-co-polymer. This is probably due to the high hydrophilicity of the hydrophilic
di-block and its less dense agglomerated structure. The hydrophobic core
creates the dense particle stabilized by a hydrophilic shell.
3.3.3.3 Conversion ratios of the monomers
Conversion ratios can be easily calculated from solids. The weight ratio of the
monomers, DPE and Initiator to water that have been weighed and added into
the polymerization reactor, should be equal to the weight ratio of the resulting
polymer to its water phase in the dispersion, the solids. All theoretical and
experimental solid ratios are shown in Tab. 3.2.
Tab. 3.2: Theoretical and experimental solids of the polymer dispersion and their
calculated conversion ratios
The weight difference can be attributed to the residual monomers in the polymer
dispersion. Across all cases, the calculated conversion ratios have not reached
100%. The best conversions were obtained for c)DPE-MA with the solid ratio of
97.7 % and for d
1
)DPE-VPA with the solid ratio of 96.6 %. The worst conversion
ratios were obtained for b)DPE-TEVS with the solid ratio of 92.2 % and for
d
3
)DPE-VPA with the solid ratio of 90.6 %. In any case experimental solids were
performed by evaporating water and all volatile compounds from 1 g of the
Chapter 3 – Results and Discussions
94
dispersion at 130°C for 2 hours at atmosphere pressure. However most of the
monomers such as TEVS, AA, BMA and HEMA are highly volatile and evaporate,
making the experimental solids technique reasonable. The high volatility of such
monomers also leads to a loss of these monomers by continuously purging the
reactor with a nitrogen flux. Monomers such as VPA with a boiling point of around
200°C would probably remain in the sample even if t hey had not reacted and
falsify the result. On the other hand the DPE-VPA with the highest VPA content
shows the lowest conversion ratio. The overall block-co-polymer yields of more
than 90% are consistent with experimental findings from Viala et al [38].
Therefore a number of polymer dispersions b)DPE-TEVS, d
1
)DPE-VPA, and
d
2
)DPE-VPA were analyzed for their residual monomers (see Tab. 3.3). The
experimental procedure was performed with Gas Chromatography (GC) by BASF
Coatings GmbH. The most interesting findings can be derived from the residual
monomeric VPA content of the corresponding dispersions. This data shows that
only small amounts of VPA can be copolymerized by this synthesis route.
Tab. 3.3: Residual monomers in polymer dispersions after block-co-polymer formation.
Data obtained with Gas Chromatography by BASF Coatings GmbH.
These observations become clearer when correlating the added VPA monomers
to the synthesis of the copolymerized VPA monomer and the residual VPA
monomers as illustrated in Fig. 3.34.
Polymer design for grain boundary application
95
Fig. 3.34: Correlation of VPA residual monomer ( ), VPA monomer co-polymerized ( )
on the weigh percentage of VPA monomer added to the synthesis batch ( ). *Value for
residual monomer estimated from the experimental ratios. ** Calculated from residual
monomer.
E.g. in the d
1
) DPE-VPA dispersion the residual VPA monomer in the dispersion
is 0.2 w.-% from 0.36 w.-% added to the synthesis batch only 0.16 w.-% can be
polymerized and incorporated into the block-co-polymer. Calculated on the solid
block-co-polymer, the incorporated VPA would only be 0.6 %. For the solid
d
2
)DPE-VPA(4.5%) the VPA content polymerized into the block-co-polymer can
be calculated at 1.0%. For the solid d
3
)DPE-VPA(8.2%) the VPA content can be
estimated only through the correlation of the two experimental results to 1.5%.
The derivation can thus be conclusively made that the polymerization of VPA is
limited, even though providing the system with an excess of VPA does not result
in significantly higher VPA content in the polymer. Connecting these results with
the solid ratios, one could also estimate that the higher VPA content may also
inhibit the polymerization of other monomers, as the highest VPA content results
gave the lowest overall yield. BMA is the only one of the other monomers that
could be detected in the experiment as significantly present. BMA has the highest
weight proportion (20% on dispersion) in the synthesis recipes of monomer
Chapter 3 – Results and Discussions
96
compositions. The aim in this section was to incorporate special functionalities
such as phosphonic acid into the block-co-polymer. Among other functional
monomers VPA was chosen for its ability to establish a strong bond to aluminum
oxide surfaces and its strong complexing ability with aluminum cations
(chapter 1). However residual monomers in dispersion could be distracting to the
selective application of polymer particles on grain boundaries in two different
ways. They could and preferably do, adsorb onto the metal oxide surface and
terminate it. Because of their small size, monomers should have higher mobility
and thus reach these active adsorption points before the polymer particles. The
second negative influence of the monomers could occur when they complex the
released cations of the grain boundary. These cations would then no longer be
available for complexing the phosphonic acid groups incorporated into block-co-
polymer chains. This would result in less effective precipitation of polymer
particles on the grain boundaries. The precipitation activity on aluminum and zinc
oxides as well as on grain boundaries of HDG steel will be investigated in the
following section of this study.
3.3.3.4 Dispersion stability
The long-term stability of dispersion was found to be rather high and thus particle
sizes were measured over a period of one year as shown in Tab. 3.4.
Tab. 3.4: Particle sizes of dispersions over one year period.
Polymer design for grain boundary application
97
It was found that particle sizes did not change significantly over time. There was
also no significant settling observed during this period. The high stability of
dispersion can mainly be attributed to its narrow particle size distribution. As the
particle sizes are similar to each other no “Ostwald-Ripening” can occur and
dispersion thus remains stable.
3.3.4 Conclusion
The aim of this part of the study was to synthesize water borne block-co-polymer
dispersions with specific functionalities. In line with this aim, it was found that the
DPE route to obtain such polymers can be provided in water without any use of
surfactants. The one-pot-two-step polymerization routine could easily be adapted
to new monomer compositions. Co-polymerization of monomers such as maleic
acid (MA), triethoxyvinylsilane (TEVS), and vinylphosphonic acid (VPA) could be
introduced to the above described block-co-polymerization procedure. The
specific functionality of MA, TEVS and VPA in previous experiments was found to
provide good adhesion to metal/metal oxide substrates as a result of its ability to
establish strong chemical bonds. These functionalities will also be used for their
good complexing abilities of aluminum cations. In the case of VPA only small
amounts of the monomer could be co-polymerized. Residual VPA monomers
could therefore negatively influence the selective deposition of the corresponding
block-co-polymer dispersions. The dispersion particle sizes were obtained in a
range from 40 to 230 nm. Even though the molecular weight distribution of each
batch was found to be significantly broad, the particle sizes showed a narrow size
distribution. The long-term stability of the block-co-polymer dispersions can also
be attributed to this phenomenon of narrow particle size distribution. The polymer
dispersions obtained in this part of the study will be introduced to the selective
application on grain boundaries of hot dipped galvanized steel in the next section
of this study.
Chapter 3 – Results and Discussions
98
3.3.5 References
[1] W.J. Soer, W.H. Ming, C.E. Koning, R.A.T.M. van Benthem, Prog. Org. Coat.
61 (2008) 224
[2] W. J. Soer, Dissertation „Bridging adhesion with good barrier properties on
function dispersions – Towards waterborne anti-corrosion coatings” Universiteit
Eindhoven (2007)
[3] Braunecker, W. A.; Matyjaszewski, K. Prog. Polym. Sci. 2007, 32, 93
[4] Matyjaszewski, K. Controlled/Living Radical Polymerization: From Synthesis
to Materials; Matyjaszewski, K., Ed.; ACS Symposium Series 944; American
Chemical Society: Washington, DC, 2006
[5] Matyjaszewski, K. In Advances in Controlled/Living Radical Polymerization;
Matyjaszewski, K., Ed.; ACS Symposium Series 854; American Chemical
Society: Washington, DC, 2003; p 2
[6] Davis, K. A.; Matyjaszewski, K. Adv. Polym. Sci. 2002, 159
[7] Matyjaszewski, K. Prog. Polym. Sci. 2005, 30, 858
[8] P.B. Zetterlund, Y. Kagawa, M. Okubo, Chem. Rev. 108 (2008) 3747
[9] M. Szwarc, M. Levy, R.J. Milkovich, J. Am. Chem. Soc. 78 (1956) 2656.
[8] A. Goto, T.Fukuda, Prog. Polym. Sci. 29 (2004) 329
[10] C.J. Hawker, A.W. Bosman, E. Harth, Chem. Rev. 101 (2001) 3661
[12] M.K. Georges, R.P.N. Veregin, P.M. Kazmaier, G.K. Hamer,
Macromolecules 26 (1993) 2987
[13] D.H. Solomon, E. Rizzardo, P.U.S. Cacioli, U.S. Patent 4, (1986)
[14] M. Kamigaito, T. Ando, M. Sawamoto, Chem. Rev. 101 (2001) 3689
[15] K. Matyjaszewski, J. Xia, Chem. Rev. 101 (2001) 2921
[16] M. Kato, M. Kamigaito, M. Sawamoto, T. Higashimura, Macromolecules 28
(1995) 1721
[17] J.S. Wang, K. Matyjaszewski, J. Am. Chem. Soc.117 (1995) 5614
[18] G. Moad, E. Rizzardo, S.H. Thang, Aust. J. Chem. 58 (2005) 379
[19] S. Perrier, P. Takolpuckdee, J. Polym. Sci., Part A: Polym. Chem. 43 (2005)
5347
[20] G. Moad, E. Rizzardo, S.H. Thang, Aust. J. Chem. 59 (2006) 669
Polymer design for grain boundary application
99
[21] C. Barner-Kowollik, M. Buback, B. Charleux, M.L. Coote, M. Drache, T.
Fukuda, A. Goto, B. Klumperman, A.B. Lowe, J.B. McLeary, G. Moad, M.J.
Monteiro, R.D. Sanderson, M.P. Tonge, P. Vana, J. Polym. Sci., Part A: Polym.
Chem. 44 (2006) 5809
[22] J. Chiefari, Y.K. Chong, F. Ercole, J. Krstina, J. Jeffery, T.P.T. Le, R.T.A.
Mayadunne, G.F. Meijs, C.L. Moad, G. Moad, E. Rizzardo, S.H. Thang,
Macromolecules 31 (1998) 5559
[23] J. Loiseau, N. Doerr, J.M. Suau, J.B. Egraz, M.F. llauro, C. Ladaviere, J.
Claverie, Macromolecules 36 (2003) 3066
[24] S.W. Prescott, M.J. Ballard, E. Rizzardo, R.G. Gilbert, Macromolecules 35
(2002) 5417
[25] J.F. Baussard, J.L. Habib-Jiwan, A. Laschewsky, M. Mertoglu, J. Storsberg,
Polymer 45 (2004) 3615
[26] C.L. McCormick, A.B. Lowe, Acc. Chem. Res.37 (2004) 312
[27] M. Li, K. Matyjaszewski, Macromolecules 37( 2004) 2106
[28] S.G. Gaynor, J. Qiu, K. Matyjaszewski, Macromolecules31 ( 1998) 5951
[29] K. Matyjaszewski, J. Qiu, D.A. Shipp, S.G. Gaynor, Macroml.
Symp.155(2000) 15
[30] H. Gao, K. Min, K. Matyjaszewski, Macromol. Chem. Phys.207 (2006) 1709
[31] J. Qiu, T. Pintauer, S.G. Gaynor, K. Matyjaszewski, B. Charleux, J.-P.
Vairon, Macromolecules 33(2000) 7310
[32] W. Bremser, F. Strickmann, M. Bendix, W. Paulus, R.B. Raether, D.I.
Christie, PAT: WO 2000037507 A1, BASF Coatings AG, 2000
[33] M. Bendix, R. Clauss, R. Nickolaus, W. Bremser, PAT: WO 2001042310 A2,
BASF Coatings AG, 2001
[34] M. Bendix, W. Bremser, D. Christie, PAT: WO 2001042311 A1, BASF
Coatings AG, 2001
[35] W. Bremser, B. Raether, Prog Org Coat 45 (2002) 95
[36] R. Clauss, W. Bremser, M. Bendix, M. Antonietti, PAT: WO 2001042379
(2001)
[37] P.C. Wieland, N. Stoeckel,O. Nuyken, J Macromol Sci, Pure Appl Chem A40
(2003) 11
[38] S. Viala, K. Tauer, M. Antonietti, R.-P. Krüger, W. Bremser, Polymer 43
(2002) 7231
[39] S. Viala, M. Antonietti, K. Tauer, W. Bremser, Polymer 44 (2003) 1339
Chapter 3 – Results and Discussions
100
[40] S. Viala, K. Tauer, M. Antonietti, I. Lacik, W. Bremser, Polymer 46 (2005)
7843
[41] D. Chen, Y. Shi, Z. Fu, J. Appl. Polym. Sci. 111 (2009) 1581
[42] S. Toews, Adsorption funktionaler Monomere an oxidischen Grenzflächen,
Master Thesis, Universität Paderborn (2007)
[43] M.-L. Abel, J.-F. Watts, J. Adhes. 80 (2004) 292
[44] J. Wielant, R. Posner, G. Grundmeier, H. Terryn, J. Phys. Chem. C 112
(2008) 12951
[45] W.J. Soer, W. Ming, C.E. Koning, R.A.T.M. van Benthem, J.M.C. Mol, H.
Terryn, Prog. Org. Coat. 65 (2009) 94
101
3.4 Polymer application on grain boundaries
3.4.1 Fundamentals
3.4.1.1 Colloid stability
Within this section all findings from the previously conducted experiments will be
combined into the approach of polymer particle dispersion deposition on grain
boundaries of HDG steel. Therefore it is important to review the basics of particle
or colloid stability. Further, it is even more important to focus on the controlled
destabilization of such systems.
Dispersed particles are attracted to and approach one other due to attraction van
der Waals and London forces [1,2].
Dispersed particles could coagulate without
repulsive stabilization mechanisms. The three stabilization mechanisms one
should consider are electrostatic, steric and electrosteric stabilization [3].
Electrostatic stabilization of colloids is described by the DLVO-theory and was
developed simultaneously by Derjaguin and Landau as well as Verwey and
Overbeek [4,5]. Within the DLVO-theory, dispersed particles can be seen as hard
spheres bearing an electric charge. In the case of anionic polymer particles such
as those obtained via the DPE dispersions in section 3.3, this charge is negative
due to the co-polymerized anionic functionalities such as carboxylic or
phosphonic acid. Negatively charged particles attract mobile positive charges
from the bulk aqueous matrix resulting in an electrical double layer surrounding
the particles. This charge cloud shields the particles from surface charge. The
distribution of the positive counter ions in the surroundings of the particle is
described by the Stern-Gouy-Chapman-theory in which the potential at the
surface is dropped across two layers, a compact inner layer (Stern Layer) and a
diffuse outer layer (Gouy Chapman Layer) as illustrated in Fig. 3.35 [6]. At the
Chapter 3 – Results and Discussion
102
shear plane between these two layers the ζ-potential is measurable. The shear
plane separates the adherent counter ion layer (basically Stern layer) from the
loose counter ion cloud which will not stick to the particle when it is in motion.
Fig. 3.35: Illustration of the distribution of electrical potential in the double-layer region
surrounding a charged particle showing the zeta potential [6].
The DLVO-Theory is basically a linear addition to the total potential (V
T
) of
attractive and repulsive forces and expresses that dispersed particles are only
stable when repulsive potential (V
R
) dominates the attractive potential (V
A
).
V
T
= V
A
+ V
R
(3.1)
The total interaction potential curve for two electrostatic stabilized particles is
illustrated in Fig. 3.36. When two particles approach each other their diffuse
charge clouds will overlap and as they have the same charge, the clouds
repulsive forces become dominant towards the attractive van der Waals forces
hindering any further approach. This repulsion is expressed through the total
Polymer application on grain boundaries
103
potential maximum. One must consider that the shape of the curve with regard to
the maxima and minima is always dependent on the system.
Fig. 3.36: Total interaction potential curve for two charge stabilized particles considering
all possible potential maxima and minima in dependency of the distance according to the
DLVO-theory [7].
The second minimum which leads to flocculation mostly appears under certain
salt concentration only, where dispersions due to the maximum repulsion remain
stable. At the same time, the thermodynamic drive of such stabilized systems is
always towards the first minimum, where particles coagulate. The height of the
total potential maximum determines the time period over which the dispersion
remains stable. Addition of multivalent counter ions to the system will eliminate
the maximum and drive the system directly into the first minimum where
aggregation and coagulation occurs for each collision of two particles.
Electrostatic stabilization is dependent on the chemical environment such as the
salt concentration, counter ion type and pH, while the ζ-potential can be used as
an indicator for stability [8].
Counter ions are dominant in the Stern and diffuse
layer. Their valency is therefore of major importance to particle stability.
Multivalent ions can bridge the Stern layer of two particles and cause them to
coagulate. (elimination of the total potential maximum), Schultze and Hardy
derived an expression for the critical coagulation concentration (ccc) as
dependent to the counter ion valency [9-11]:
Chapter 3 – Results and Discussion
104
~
(3.2)
In this equation
z
is the counter ion valency and
n
the exponent ranging between
2 and 6 dependent on the particle potential [8]. For example, aluminum counter
ions would be 10 times more efficient in coagulation than sodium ions and 4
times more efficient than zinc ions considering the Schultz-Hardy-Equation with
the lowest exponent of 2.
The ζ-potential is characteristic for the measurable charge density of a dispersed
particle and can be adjusted by the pH. In general, a system is no longer
electrostatically stabilized when the ζ-potential reaches the value of zero. In
which case, there is no charge on the particle surface to provide repulsion forces.
Experimental observations have shown that absolute values of ζ-potentials
smaller than
|20|
mV can cause instability and coagulation of anionic stabilized
dispersions [8]. In general it is suggested to maintain the ζ-potential above the
absolute value of
|30|
mV [12].
In section 3.1 and 3.2 of this study it was found that aluminum ion dissolution
from aluminum enriched grain boundaries could be of importance to the selective
deposition of polymer particles on grain boundaries. Therefore both the counter
ion type and pH required in order to dissolve the counter ions from the grain
boundaries will be investigated within the experimental part of this chapter.
From the 1950s onwards, steric stabilization of dispersed particles has been
described by a variety of scientists [13-17].
For this type of stabilization, freely
moving polymer chains have to be adsorbed by one end to the particle surface
[18,19]. The particles result in spherical polymer brushes (see Fig. 3.37).
Polymer application on grain boundaries
105
Fig.3.37: Illustration of particles with polymer brushes sterically hindered in their
approach.
When two particles approach one another their polymer brushes start to
interpenetrate. This interpenetration reduces the degree of freedom of polymer
chains and therefore the entropy of the system. According to the Gibbs
expression for free energy derived from the second law of thermodynamics (∆G =
∆H – T∆S), a physicochemical process can only spontaneously follow in a certain
direction when the free energy ∆G of that system becomes negative. The
decrease of entropy ∆S at constant temperature T and constant free enthalpy ∆H
would result in an increase in the free Gibbs energy ∆G. Because entropy is the
determining factor of steric stabilization it is also called entropic stabilization. This
mechanism of stabilization works only when the polymer brushes are soluble in
the surrounding solvent matrix [20]. Therefore Fischer’s solvency theory becomes
predominant to the stability of steric stabilized dispersions [21]. Fischer
characterized the θ-point of dissolved polymers where the polymer chain has
undisturbed free mobility and creates a coil conformation. Solubility above the θ-
point is provided by good solvents. In this case, the coil conformation is unrolled
and the polymer chain is craned towards one direction. A solubility level lower
than the θ-point causes the coil formation to collapse and precipitation of the
polymer to occur. Therefore steric stabilized dispersions can only be stable
above θ-point conditions for the polymer brushes on the particles. Steric
Chapter 3 – Results and Discussion
106
stabilized particles are also less sensitive towards multivalent ions than those
that are electrostatically stabilized. The most common polymer types for steric
stabilization in aqueous media are polyvinylalcohol and polyethylenoxide. These
polymers are most often designed as block-co-polymers with some specific
anchoring blocks in order to adsorb on the particle surface following the
adsorption mechanisms (see Chapter 1).
Grafted polyelectrolyte chains are also commonly used for particle stabilization in
dispersions [22,23], because polyelectrolyte chains are charged, they combine
the characteristics of electrostatic and steric stabilization resulting in the so called
electrosteric stabilization. An illustration of polyelectrolyte brushes is shown in
Fig. 3.38
Fig. 3.38: (left) Illustration of grafted polyelectrolyte brushes on a particle. (right) TEM
image of polyelectrolyte brushes on a polystyrene particle [24].
Several experiments have shown that stabilizing dispersed particles with
adsorbed or grafted polyelectrolyte brushes increases the stability towards
multivalent counter ions and pH values in aqueous matrix [25-28].
It is assumed that DPE block-co-polymer dispersions similar to those synthesized
within this study are also electrosterically stabilized [29,30]. The creation of stable
dispersion particles without any use of surfactants as described in section 3.2 of
this study is attributed to this stabilization method. It is therefore assumed that
Polymer application on grain boundaries
107
hydrophilic polyelectrolyte chains are grafting on the surface of the polymer
particle during polymerization.
While electrosteric stabilized particles are less dependent on the pH of the
surrounding aqueous matrix and have more tolerance towards multivalent
counter ions, it can be assumed that the obtained DPE dispersions would be
more controllable for the mechanism of selective deposition on grain boundaries.
3.4.2 Experimental procedure
3.4.2.1 Titration on ion selectivity
The influence of aluminum and zinc ions on the coagulation behavior of the
dispersion was investigated by titration combined with particle size detection.
Therefore the block-co-polymer dispersions were reduced to a solid content of
1.0%. The titration was carried out on 20 ml of the reduced dispersion. Aluminum
and zinc chloride solutions of 0.01 mol/L were added in 0.5 mL steps to the
dispersions whilst being stirred, using a burette. After each step the particle size
of the dispersion was measured with the Zetasizer Nano-ZS from Malvern.
3.4.2.2 SPR measurements
SPR studies were performed as described in Chapter 2: Applied Techniques.
3.4.2.3 Grain boundary selective polymer application
Grain boundary selective polymer application was carried out through a dip
coating process. The polymer dispersion was diluted to a solid content of 1.0%.
and the pH was adjusted with nitric acid to 2.2 – 2.4. The solvent cleaned
substrates were dipped in the prepared dispersion for 30 s. After 30 s no visible
difference in the resulted coated substrate surface was found. Within this
Chapter 3 – Results and Discussion
108
application window only the grain boundaries of the hot dipped galvanized steel
were covered with the polymer material.
3.4.2.4 Coating application and testing
Within in the salt spray test, HDG steel sheets were spin coated with a coil
coating primer, Coiltec Universal P CF. The spin coater P6700 from Specialty
Coating Systems was operated at 500 r/min. This setting resulted in a constant
coating thickness of 10 µm.
The salt spray tests were carried out in a salt spray chamber for 504 hours in
accordance with the German Association for Industrial Testing DIN EN ISO 9227-
2006.
3.4.3 Experimental results
In section 3.1: Surface Characterization it was found that grain boundaries are
electrochemically more active than the surrounding grain surfaces. They were
therefore identified as the weak zones of the HDG steel substrate surface. This
corrosive behavior was attributed to both the enriched aluminum content in the
grain boundaries and their non-compact structure. Aluminum is more
electronegative in its potential than zinc and would therefore preferably
electrochemically dissolve. The less compact structure provides a great
accessible surface for dissolution activity and seemingly the atomic structure in
the grain boundaries is more amorphous.
In section 3.2: Material Survey for Grain Boundary Application it was found that
grain boundaries can selectively be dissolved with a dependency on the pH
value. It was assumed that the dissolution causes aluminum cations, which could
then be used for selective deposition of material on the grain boundaries. The
most promising results were obtained by polymer particle deposition.
Within the following experimental part of this section the specifically obtained
block-co-polymer dispersions (chapter 3.3) will be investigated in terms of their
selective applicability to the grain boundaries of HDG steel.
Polymer application on grain boundaries
109
However in terms of the systematic in this study, the deposition of DPE
dispersions on grain boundaries will be described in this section. Historically the
first observations of DPE dispersions selectively covering the grain boundaries
were nevertheless made in the very beginning of this study and initiated the
fundamental investigation around this phenomenon.
3.4.3.1 Polymer ion selectivity
The pH value and multivalent counter ion concentrations were described as the
predominant parameter for instability of electrostatic particles. These parameters
are also dominant to the stability of electrosteric stabilized particles even though
they are less in power. In order to deposit the obtained block-co-polymers
selectively on the grain boundaries one has to investigate the coagulation activity
towards the specific ions in the system, aluminum and zinc. This was provided by
titration of the block-co-polymer dispersions with aluminum and zinc ions (see
Fig. 3.39).
Fig. 3.39: Selectivity of block-co-polymer dispersions to aluminum and zinc ions
determined by dispersion particle coagulation in dependency of the ion content.
Whilst as part of the process the step of adding multivalent counter ions to the
diluted polymer dispersions included tracking of the particle size, it was found
Chapter 3 – Results and Discussion
110
that dispersions quickly coagulate in the presence of small amounts of aluminum
ions. Less than 0.05 mmol of aluminum ions raise the particle size from around
one hundred nm (single particles) up to thousands of nm as detected until the
entire dispersion precipitates. At the same time these dispersions are more
tolerant to the presence of zinc ions. The particle size starts to significantly rise
between 0.1 and 0.2 mmol of zinc ions added to the batch of 20 mL dispersion.
After 0.2 mmol the dispersion precipitates. The different coagulation behavior of
the dispersion particles can be attributed to different charge densities of
aluminum and zinc ions. Aluminum cations are triple charged on a rather small
ion radius which makes them highly charged atoms, able to attract the anionic
polymer particles and act as a bridging charge between the polymer particles.
This quickly leads to agglomeration and destabilization of the block-co-polymer
particles. Zinc ions on the other hand are positively double charged on a rather
large ion radius which results in a lower charge per radius relation. The anionic
stabilized polymer particles can tolerate more of the zinc ions before charge
quantity is reached and the stabilization of the polymer particles collapse. This
aligns with the Schultze-Hardy theory for critical coagulation concentration of
multivalent counter ions.
According to the assumptions of electrosteric stabilization of block-co-polymer
particle dispersions, the polyelectrolyte brushes due to their spatial extension into
the water matrix would also very quickly form a complex between the functional
group and the cation. There are basically two possible scenarios that could
occur. The first would be that polymer brushes from one particle catch out the
cation from the surrounding water matrix and create a complex within the
brushes. This particle then remains stable until the polyelectrolyte brushes
become saturated with the multivalent counter ions. In the second scenario the
complex would be created between one counter ion and different polymer chains
from different particles. In this case the counter ion would bridge the two particles
and they would destabilize and coagulate. In both scenarios aluminum cations
would also provide a stronger complex with the hard Lewis bases such as
phosphonic and carboxylic acid (see Chapter 1).
However these results open a window of individual cation quantities that must be
locally present for a selective coagulation and deposition of polymer particles on
the grain boundaries. In the next step it is important to investigate the stability of
Polymer application on grain boundaries
111
these polymer dispersions in relation to dependency of the pH values, as this is
also an important factor and the selective release of aluminum cations from the
grain boundaries was found to occur under acidic conditions.
3.4.3.2 Polymer particle stability in dependency of pH values
The stability of polymer dispersions in regions of low pH values is of importance
considering the desired application mechanism on the grain boundaries. Whilst
the creation of cations locally in the grain boundary region should precipitate the
polymeric material selectively, the dispersion should remain stable where no
aluminum cations are present but the pH value is low. Fig. 3.40 shows the
obtained results for particle size and ζ-potential in dependency of the pH value.
Fig. 3.40: Particle size and
ζ
-potential in dependency of pH values. *(pzc) point of zero
charge theoretically correlated from experimental data.
Chapter 3 – Results and Discussion
112
All dispersions remain stable down to the pH value of 1.0. When crossing the pH
1.0 value, particle sizes from DPE-MA and DPE-TEVS significantly rise, whereas
the DPE-VPA dispersions remain stable.
At this point particle size may be of significant importance. Cosgrove derives
calculations from the DLVO theory where the maximum of the total potential (V
T
)
(Fig. 3.33) has a different relation to the particle size [8]. For particles below 100
nm the radius of the particle is directly proportional to V
T
. For particles sizes
above 100 nm the relationship is more complicated but less dependent on the
radius. This means that down to particle sizes of 100 nm the repulsion maximum
decreases slower than the particle size. Below the particle size of 100 nm the
decrease of the repulsion maximum is steeper which results in a smaller barrier
to be overcome for the coagulation of smaller particles. This may be the reason
why DPE-TEVS and DPE-MA particles start to aggregate even though the ζ-
potential indicates a stabilized system.
A significant change towards smaller absolute values in the ζ-potential starts
below the pH value of 2.0. This behavior aligns with the theory discussed in the
fundamental part of this chapter. The point of zero charge can be estimated by
correlation functions of the obtained data to a range between 0.5 and 1.0 pH for
the individual dispersions. However down to pH values of 1.0 all dispersions
remain stable. This may be of importance when it comes to the selective
application process on grain boundaries at low pH values.
3.4.3.3 Polymer particle adsorption to aluminum and zinc oxide surfaces
Within the following experiments polymer particle adsorption to aluminum and
zinc oxide will be investigated. These experiments aim to provide specific
application characteristics and to find the appropriate parameter window for the
selective grain boundary deposition of polymer particles. The trigger for
aluminum release from grain boundaries was found to be the pH value. Therefore
the variation parameter will be the pH value using the surface plasmone
resonance spectroscopy as a surface sensitive detector for polymer adsorption.
The resulting spectra are shown Fig. 3.41 with the example of the DPE-
VPA(4.5%) on an aluminum oxide SPR sensor. The gathered spectra highlight
Polymer application on grain boundaries
113
the shift of the reflection minima to higher angles. This indicates an adsorption of
polymer particles on the surface of the sensor. The greater the shift of the minima
towards higher angles, the higher the thickness of the adsorbed layer on the
sensor. In all of these measurement cases only the relative shift will be
considered and evaluated. The information about the reflectivity index of the
adsorbed polymer particle layer will remain unknown in this study. But
considering that different polymer dispersions have a reflectivity index in the
same range, the relative angle shift will be comparable and provide information
on the adsorption process of the particles to each of the substrates.
Fig. 3.41: Step scan spectra of DPE-VPA(4.5%) on aluminum oxide at different pH values. The
reflection minima shift is indicated by character from a) deionized water to g) polymer dispersion
DPE-VPA(4.5%) adjusted to the pH of 1.80.
The reflectivity minimum measured in water before each measurement with the
varied pH is the scaling point in order to calculate the angle shift. In the case of
DPE-VPA(4.5%) the angle has shifted from water 61.18° to 62.03°. At the angle
of 61.18°, no adsorption occurred while at the angle of 62.03° (pH 1.80) a dense
polymer film was adsorbed to the sensor surface. The adsorption of the polymer
Chapter 3 – Results and Discussion
114
could be controlled by SEM imaging and the SPR sensor after each adsorption
measurement was run. Fig. 3.42 shows high resolution SEM images of the
surface of a bare aluminum oxide sensor after measurement in the water, the
polymer DPE-VPA(4.5%) adsorbed on the sensor at the pH of 2.38, and at the
pH of 1.80.
Fig. 3.42: High resolution SEM images of sensor surfaces: a) aluminum oxide sensor
after measured in distilled water, b) aluminum oxide sensor after measured the
adsorption of DPE-VPA(4.5%) at the pH of 2.38 and c) at the pH of 1.80.
One must consider that the detection area on the sensor is the same as the
diameter of the laser beam, which is a few 100 µm. The area where the
reflectivity information can be gathered is around the size of one of the SEM
Polymer application on grain boundaries
115
images as shown in Fig. 3.42. The minimum shift is dependent on the adsorbed
layer thickness. When this area is not homogeneously covered with the polymer
the obtained information has to be considered as an averaged thickness of the
adsorbed particles over the sensor surface. A good example for such a
heterogeneously covered sensor surface is provided in Fig. 3.40b, with the SEM
image of the adsorbed polymer dispersion DPE-VPA(4.5%) at the pH of 2.38. It
shows that the entire surface is not covered with polymer particles. The sensor
surface remains visible all over the image but the SPR spectra shows a minimum
shift towards higher angles as shown in Fig. 3.41. The image obtained after the
adsorption at the lowest pH of 1.80 shows an entirely covered sensor surface
with partly coagulated polymer particles. One must consider that coagulation of
these particles occurs also whilst the SEM image is obtained as initiated through
the energy from the electron beam.
The reflection minima shifts of all the investigated polymer dispersions on the
aluminum oxide and zinc oxide sensors by variation of the pH are provided in
Tab. 3.4. From this data it is apparent that all polymer dispersions roughly follow
the same trend. While on the aluminum oxide surface a minimum shift occurs at
the highest pH levels, there are no adsorption events that occur on the zinc oxide
surface. Only for pH values of around 2.0 does some adsorption of polymer
particles to the zinc oxide surface occur and the reflectivity minimum shifts
towards higher angles.
Chapter 3 – Results and Discussion
116
Tab. 3.4: SPR reflectivity minima shifts of investigated dispersions at different pH values
on aluminum oxide and zinc oxide sensors.
Dispersion pH
[
]
2 3
Al O
Δθ °
[
]
ZnO
Δθ °
DPE-VPA(8.2%)
1.76 0.3219 0.1649
1.90 0.0698 0
2.05 0.0537 0
2.31 0.0549 0
3.36 0.0269 -
5.53 0.057 -
DPE-VPA(4.5%)
1.80 0.8539 0.7241
1.98 0.5069 0.1902
2.13 0.3853 0
2.38 0.122 0
3.58 0.0548 -
5.23 0.0479 -
DPE-VPA(1.3%)
1.89 1.1318 0.126
2.10 0.1927 0.0082
2.27 0.1043 0
2.62 0.0604 0
3.32 0.0819 -
5.24 0.1107 -
DPE-TEVS(4.5%)
1.89 - 0.2296
2.10 1.7203 0.0875
2.38 0.6538 0
2.68 0.1776 -
3.13 0.0353 -
5.50 0.0217 -
DPE-MA(4.5%)
1.89 0.3252 0
2.09 0.1113 0
2.32 0.1597 0
3.08 0.119 0
The numbers from the minima shifts provide a potential application window for
selective deposition of polymer particles on grain boundaries when using the pH
as a trigger. Fig. 3.43 shows the correlation between the minima shift to higher
angles when lowering the pH. Overlaying the charts for aluminum oxide and zinc
Polymer application on grain boundaries
117
oxide visualize the pH range window where adsorption on the aluminum oxide
surface occurs but not for the zinc oxide surface.
Fig. 3.43: SPR reflectivity minima shift in dependency of the pH on aluminum oxide (top)
and on zinc oxide (bottom).
In Chapter 3.2: Material Survey for Grain Boundary Application, the best pH
range for grain boundary dissolution and therefore the release of cations from the
grain boundary was found to be 3.6 to 4.0. SPR measurements found the best
application range to be in the pH region of 2.1 and 2.5. This difference may arise
from the different topography of the sensor (aluminum oxide or zinc oxide)
compared to the grain boundaries (see section 3.1). The porous structure with
edges and vertices may expose the metal atoms and their oxidic species in a
manner more accessible for anodic dissolution into the surrounding acidic water.
Chapter 3 – Results and Discussion
118
The sensor is rather flat and therefore it may remain stable towards dissolution of
the outermost ions at lower pH values. The selective deposition on grain
boundaries will most probably be found in the pH region between 2 and 4 pH.
However one interesting observation that can be made is that aside from DPE-
VPA(8.2%), all dispersions provide different adsorption behavior towards the two
different oxide surfaces. Only DPE-VPA(8.2%) shows significant polymer particle
adsorption at the same pH value of 1.78 for both substrates. This dispersion
obviously does not distinguish between the surface types.
An explanation for this observation may come from the findings in Chapter 3.3
where VPA was found to be limited in copolymerization in block-co-polymer
synthesis. In DPE-VPA(8.2%) the residual VPA monomer content can be
estimated at 1.9 w.-%. As discussed in the previous chapter one can assume that
the small VPA monomer preferably complex the released cations from the sensor
surface. In this scenario polymer particles would not access the amount of
counter ions they need to coagulate at mild conditions. One could imagine that
only when counter ions are released excessively would polymer particles also
coagulate and precipitate, and because the zinc dissolution at a lower pH is
higher than aluminum, there is a loss of selectivity towards the different surfaces.
Another explanation in terms of the residual VPA monomers would also be the
surface stabilizing effect of VPA. As described in Chapter 1, phosphonic acid
derivates are able to anchor on aluminum/aluminum oxide surfaces and
assemble into a stable monolayer. This monolayer would protect the substrate for
dissolution at moderate pH levels. Thissen et al. investigated the stability of long
chain (C
18
) phosphonic acid SAMs on different aluminum oxide surfaces [31].
Besides the polar Al
2
O
3
(0001) surface the adsorbed monolayer remained stable
when it was exposed to water. This was also the case on aluminum oxide
surfaces created through physical vapor deposition, in the same way the SPR
sensors were prepared for the experiment in this section. Liakos et al. found that
C
18
– alcyle phosphonic acid self-assembled monolayers remain stable down to
the pH value of 1.0 on aluminum oxide surfaces [32]. It was also discussed that
the stability of SAM is most stable at a pH of 3.0. In the same paper it was found
that short chain (C
4
) phosphonic acid SAMs could be washed off the surface with
pure water after several rinsing cycles. However, VPA is a C
3
phosphonic acid.
Polymer application on grain boundaries
119
Since in the system of SPR experiments of this study there was no rinsing, one
could assume that some adsorption of VPA to the surface occurred. However by
evaluating the literature survey one would assume that the adsorbed monolayer
would desorb above the value of 1.0 pH. Nevertheless both scenarios could
explain the lack of selectivity of dispersions bearing a high amount of residual
functional monomers such as vinylphosphonic acid. In the next step the findings
from these model substrates will be transferred to the real technical substrate,
the HDG steel.
3.4.3.4 Selective deposition on grain boundaries of HDG steel
In the following experiments all understanding gained in terms of the substrate
characteristics, polymer design and polymer-substrate interaction, as generated
in the previous investigations will be transferred to the selective deposition of
polymers on grain boundaries of the technical substrate of interest, hot dipped
galvanized steel. The application process is illustrated in Fig. 3.44.
Fig.3.44: Illustration of the selective polymer application process on grain boundaries of
HDG steel.
Running through the application process the substrate has to first be degreased.
It was therefore dipped in a cascade of each solvent bath for 10 minutes and
Chapter 3 – Results and Discussion
120
ultrasonically enhanced. This procedure was found to be very effective in
removing all organic compounds from the surface and leaving an analytically
clean substrate [33]. However for industrial application the cleaning process
could probably be optimized in terms of time and solvent type. One could assume
that washing with aqueous detergent solutions could be introduced into a fast
cleaning process. After cleaning, the selective polymer deposition on the grain
boundaries occurs. Therefore the polymer dispersions were diluted to the solid
content of 1.0 w.-% and adjusted to the pH with a step-wise variation in
accordance with the most promising application window for selective polymer
deposition on the grain boundaries. The application is carried out through a
dipping of the substrate into the adjusted polymer dispersion. After dipping for 30
s the sample is rinsed with pure water and dried in a nitrogen stream. At this
stage of the process the grain boundary selective pretreatment is completed. For
end finishing the samples were coil coated. The resulting surface of the grain
boundary pretreated substrate was monitored via SEM imaging. An example of a
result obtained from selective deposition on the grain boundaries of DPE-
VPA(4.5%) is provided in Fig.3.45.
Fig. 3.45: SEM images of HDG steel substrate with the polymer dispersion DPE-
VPA(4.5%) selectively applied to the grain boundaries of the substrate. a) overview, b)
grain boundary triple point. The polymer dispersion was adjusted to solid content of 1.0
w.-% and the pH of 2.4. The dipping time was 30 s.
The SEM images in Fig. 3.45 visualize how polymer particles are concentrated
on the grain boundaries whilst almost no polymer can be found on the
Polymer application on grain boundaries
121
grains/spangles. Such results could be obtained with all of the tested DPE
polymer dispersions besides DPE-VPA(8.2%). For all polymer dispersions a good
window in the application process on the real HDG substrate was also found with
pH values of 3.0 to 2.2. The results of dip coating for 30 s within this application
window always led to similar images as shown in Fig. 3.45. The dispersion DPE-
VPA(8.2%) was the only one not able to be applied selectively on the grain
boundaries. Down to the pH value of 2.2 there was almost no dispersion found
on the HDG steel substrate. Therefore at the pH value 2.0 the entire substrate
was covered with a monolayer of polymer dispersion particles. These findings are
consistent with the adsorption results obtained via the SPR technique and
discussed in the earlier part of this chapter. These results show that DPE-
VPA(8.2%) has similar adsorption behavior on both aluminum oxide and zinc
oxide sensors.
It was assumed that either residual VPA monomers were stabilizing the substrate
surface by creating a self-assembled monolayer or that the small VPA molecules
complex the released cations from grain boundaries and deactivate these cations
from the destabilization of the polymer particles. With regard to the theory of SAM
formation as discussed in Chapter 1 of this thesis and the topography of grain
boundaries as found in section 3.1:Surface Characterization, it is most likely that
the loose structure of edges and vertices of grain boundaries would not be
stabilized by any type of self-assembled monolayer. Recent studies have also
shown that phosphonic acid derivates precipitate to zinc phosphonates on HDG
steel with an aluminum content of 0.5 w.-% rather than create a self-assembled
monolayer [34]. Stable SAM of organophosphonic acids could be obtained only
on galvanized steel with aluminum contents above 5 w.-%. Zinc phosphonates on
HDG steel could easily be washed off from the substrate surface with pure water.
It is therefore most likely that residual monomers catch the released cations so
they cannot destabilize the polymer particles of the dispersion. In addition, the
co-polymerized ratio of VPA in DPE-VPA(8.2%) must be greater than that for the
other dispersions as discussed in section 3.3, even though the residual monomer
values are at the highest. The higher the amount of VPA incorporated into the
block-co-polymer, the higher the negative charge of the particle and rise in the
repulsion maximum according to the DLVO theory, which makes the dispersion
more stable towards counter ions. This goes along with the slightly higher ζ-
Chapter 3 – Results and Discussion
122
potential and point of zero charge that was found for DPE-VPE(8.2%). This factor
may enhance the lack of counter ions due to residual monomers and both factors
cause very poor selectivity in the application process. For all other polymer
dispersion applied to the HDG substrate Fig. 3.44 summarizes the selective
application of the polymer dispersion on the grain boundaries.
Fig. 3.44: Scheme of the polymer application results on HDG steel substrates by dip
coating in dependency of the pH value.
When coming from higher pH value regions down to 3.0, there is no autophoretic
polymer application possible for the investigated dispersions. Between the pH
values of 2.0 and 3.0 the grain boundaries are selectively coated with the block-
co-polymer particles. Below the pH value of 2.0 polymer particles cover the entire
substrate surface. The selective polymer deposition on the grain boundaries
could be achieved due to their different electrochemical behavior. The high
aluminum concentration within the grain boundaries and their topography led
them in turn to have higher electrochemical activity. While the surface of grains is
Polymer application on grain boundaries
123
rather flat and smooth the structure of grain boundaries is rough and consists of
multiple edges and some sort of amorphous sponge like surface within the
tranches. The higher susceptibility of the grain boundaries to the corrosion
process leads to a faster dissolution of the metal of the grain boundary material
and the release of multivalent metal cations. Aside from aluminum cations it is
most likely that zinc cations will also be dissolved as zinc/zinc oxide (theoretical
dissolution at pH of 5.8) is less stable under acidic conditions than
aluminum/aluminum oxide (theoretical dissolution at pH of 3.8) [35]. As discussed
in section 3.1: Surface Characterization grain surfaces are also terminated with a
layer of aluminum oxide that is a few nano-meters thick. In addition to the flat
geometry, this layer may also provide the grain surface with more stability
towards electrochemical dissolution in acidic environments; even though the
aluminum surface density on grains is too low to create a stable SAM out of
phosphonic acid derivates. A scheme illustrating this process is shown in Fig.
3.45 where a local release of cations drives the coagulation and precipitation of
the polymer particles directly to the grain boundaries in the second step. The
entire process can be described as a local autophoresis.
Fig. 3.45: Scheme of the local autophoresis driven polymer precipitation on the grain
boundaries.
Chapter 3 – Results and Discussion
124
So far the selective polymer deposition on grain boundaries of HDG steel has
been discussed as a local autophoresis driven by the dissolution of cations from
grain boundaries. However it might be worth discussing this process under the
consideration of alternative possible processes.
It could also be speculated that by anchoring the block-co-polymer functional
groups to the solid edges of grain boundaries (see chapter 1). Recent studies
have shown that functional groups such as carbocylic acids from polyelectrolytes
similar to the block-co-polymers used in this study preferably attach to the edges
of polar terminated ZnO (0001)-Zn surfaces [36]. Assuming that all grain surfaces
are polar terminated and grain boundaries are the edges of those surfaces, one
should also observe polymer particle adsorption on the grain boundaries. Even
though the polymer brushes of the DPE dispersions would reach out as
anchoring tentacles, only a single particle could cover one adsorption spot that is
the size of the particle. As shown in Fig. 3.45 there is an aggregation and
stacking of particles on each other on grain boundaries. Therefore the deposition
of the particles may prefer to be driven by local autophoresis. However, the
anchoring to the edges of the grain boundaries might be the determining factor
for stabilizing them towards electrochemical activity after deposition.
3.4.3.5 Testing results
In order to prove the concept an accelerated corrosion test in a salt spray
chamber was provided on non-pretreated and grain boundary treated HDG steel
samples. Both the pretreated and non-treated substrates were coated with a coil
coating primer from BASF Coatings GmbH, Münster. The test samples were
scratched and exposed to the salt spray test for 504 hours. After the exposure
the entire coating material was removed from the sample surface and scanning
electron microscopy was used to investigate the corrosion propagation along the
grain boundaries. Fig. 3.48 shows the SEM images and combined EDX
mappings of non-grain boundary treated and grain boundary treated substrate
samples after the corrosive exposure. On the non-treated substrate the grain
boundaries connected to the corrosion front show some destruction along the
grain boundary stemming from a corrosive attack. The damage along the grain
boundary can be measured by chlorine traces detected in the EDX mappings. On
Polymer application on grain boundaries
125
the substrate with the non-treated grain boundaries, the extension of the
corrosion path can be measured from 100 to 200 µm based on the border
between the plateau corrosion front and the intact HDG surface.
The best results for disabling the grain boundary corrosion were achieved by
grain boundary selective deposition of the DPE-TEVS(4.5%) dispersion (see Fig.
3.46). To track the corrosion propagation along the grain boundaries the chlorine
trace from the EDX mappings was used. In this case, the chlorine trace along the
grain boundaries can only be measured to 10 to 30 µm. This would be three to
twenty times less than was found on the substrate without selective polymer
deposition on the grain boundaries. There is also no visible damage on the
treated grain boundaries than can be seen on the non-treated samples in the
SEM images.
Chapter 3 – Results and Discussion
126
Fig. 3.48: SEM images and EDX mappings of the corrosion front coming from the
scratch. Left) Image is obtained from a non-grain boundary treated substrate. Right)
Image is obtained from a grain boundary tread substrate with DPE-TEVS(4.5%). The
focus is in the border between the corroded and intact surface and on the grain
boundaries of that region.
Polymer application on grain boundaries
127
According to the model developed, (see section 3.1), the elimination of grain
boundary corrosion should slow down overall corrosion. The results within this
chapter could show that a selective polymer deposition on grain boundaries leads
to less corrosion along the grain boundaries as is sketched in Fig. 3.49.
Fig. 3.47: illustrates the results from corrosion on a HDG sample not treated (left) and
polymer deposited exclusively on grain boundary (right).
However, on a micro scale these findings support the theory raised about the two
corrosion pathways. Unfortunately the overall corrosion performance on a macro
scale showed different results. Salt spray tests were carried out and evaluated by
total creep from the scratch. While the initial tests showed small statistical
improvements on the overall corrosion when grain boundaries are selectively
treated with the block-co-polymers, the following results were statistically
scattering a lot [37]. The salt spray test itself also scatters with a standard
deviation of around 25%. This leads to the assumptions that the grain boundary
contribution to corrosion may be within this range and that the salt spray test is
not able to resolve the influence of grain boundaries to the overall performance.
In terms of overall performance, the silane functionalized polymer dispersion
showed the best results. This behavior is not surprising and is based on the
discussions within chapter 1 of this study.
Chapter 3 – Results and Discussion
128
At this point one could also consider that the synthesized polymer dispersions in
this study are model polymers for selective grain boundary application. In order to
enhance corrosion protection the polymer architecture might bear some potential
for improvement.
3.4.4 Conclusions
The main goal of this chapter was to apply the polymer dispersions selectively on
the grain boundaries of the hot dipped galvanized steel substrate. Knowing that
grain boundaries are aluminum enriched and have a specific structure and
therefore a higher susceptibility to electrochemical dissolution the specifically
designed water based polymer dispersions were investigated on their coagulation
behavior in the presence of aluminum and zinc ions. It was found that the triple
charged aluminum ions immediately cause a destabilization of the polymer
particles followed by their precipitation. Towards the double charged zinc ions the
polymer dispersion showed a tolerance up to a specific ion concentration but
then also coagulated and precipitated. This varied influence on the polymer
particle stability was further used to evaluate the application window for a
selective deposition of the polymer on the grain boundaries. The pH level as a
trigger for the dissolution of the aluminum cations was therefore investigated
using the surface plasmone resonance spectroscopy. It was found that an ideal
window for a selective application exists in the pH value range of 2.0 to 3.0. At
higher pH polymer particles could not be deposited on either the aluminum oxide
or zinc oxide surfaces. At a pH below that window the polymer dispersion were
precipitating on both surfaces. Only in that window were the polymer dispersions
precipitating exclusively on the aluminum oxide surfaces. After defining these
application parameters the goal of selectively coating grain boundaries of the
industrial HDG steel substrate could be realized in a dip coating process. The
final proof of concept was provided through a comparison of non-grain boundary
treated but coil coated versus grain boundary treated and coil coated HDG
substrates in a salt spray test. The results by SEM/EDX imaging showed a
significant difference in the condition of the grain boundaries at the corrosion
front. It was found that when selectively cover the grain boundaries with a
specifically designed polymer the corrosive damage along the grain boundaries is
Polymer application on grain boundaries
129
reduced by a factor of three to twenty. On the micro scale it could be shown that
grain boundaries on HDG steel are highly corrosively active. It has also been
shown that it is possible to block the grain boundaries and reduce their corrosive
activity.
3.4.5 References
[1] M. Smoluchowski, Phys. Chem. 92 (1917) 129
[2] H. C. Hamaker, Physica 4 (1937) 1058
[3] H.-J. Butt, M. Kappl, Surface and Interface Forces, Wiley-VCH (2010)
[4] Derjaguin, B.V. and Landau, L. (1941) Acta Physicochim. (URSS), 14, 633.
[5] Verwey, E.J. and Overbeek, J.T.G. (1948) Theory of the Stability of Lyophobic
Colloids. Elsevier, Amsterdam.
[6] K.S. Birdi, Handbook of Surface and Colloid Chamistry, CRC Press, London
(2009)
[7] P.W. Atkins, Physikalische Chemie, Wiley-VCH, Weinheim (2006)
[8] T. Cosgrove, Colloid Science – Principles, Methods and Applications,
Blackwell Publishing, Bristol UK (2005)
[9] Schulze, J. (1882) pr. Chem., 25, 471.
[10] Schulze, J. (1883) pr. Chem., 27, 320.
[11] Hardy,W.B. (1900) Proc. R. Soc., 66a, 110.
[12] Malvern Instruments
[13] E. L. Mackor, J. Colloid Sci., 6 (1951) 492
[14] D. J. Meier, J.Phys.Chem., 71 (1967) 1861
[15] F. Hesselink, J.Phys.Chem., 75 (1971) 65
[16] W. Heller, T. L. Pugh, J. Chem. Phys. 22 (1954) 1778
[17] W. Heller, Pure Appl. Chem. 12 (1966) 249
[18] D.H. Napper, Trans. Faraday Soc. 64 (1968) 1701
[19] R. Evans., J.B. Davison,D. H. Napper, J. Polymer Sci. B10 (1972) 449
[20] R. Evans, D.H. Napper, Kolloid-Z. Z. Polym. 251 (1973) 409
[21] E.W. Fischer, Kolloid-Z. 160 (1958) 120
[22] P. Pincus, Macromolecules, 24 (1991) 2912
[23] J. Hang, L. Shi, X. Feng, L. Xiao, Powder Technology, 192 (2009) 166
Chapter 3 – Results and Discussion
130
[24] M. Schrinner, B. Haupt, A. Wittemann, Chem. Eng. J. 144 (2008) 138
[25] H. Ahrens, S. Förster, C. A. Helm, Phys. Rev. Lett. 81 (1998) 4172
[26] T. Abraham,S. Giasson, J.F. Gohy, R. Jerome, Langmuir 16 (2000) 4286
[27] M. Balastre, F. Li, P. Schorr, J. Yang, J.W. Mays, M.V. Tirrell,
Macromolecules 35 (2002) 9480
[28] J. Forsman, Curr. Opin. Coll. Interf. Sci. 11 (2006) 290
[29] S. Viala, Dissertation „Kontrollierte radikalische Heterophasenpolymerisation
mit Anwesenheit des Diphenylethylens“ Universität Potsdam (2002)
[30] B. Weber, Dissertation „ Selbststrukturierende Hybridmaterialien für polymer
Werkstoffe“ Universität Paderborn (2009)
[31] P. Thissen, M. Valtiner, G. Grundmeier, Langmuir 26 (2010) 156
[32] I.L. Liakos, R.C. Newman, E. McAlpine, M.R. Alexander, Langmuir 23 (2007)
995
[33] N. Fink, B. Wilson, G. Grundmeier, Electrochim. Acta 51 (2006) 2956
[34] P. Thissen, J. Wielant, M. Köyer, S. Toews, G. Grundmeier, Surf. Coat. Tec.
204 (2010) 3578
[35] A. F. Holleman, E. Wieberg, Lehrbuch der Anorganischen Chemie, 101
Auflage, deGruyter, Berlin, (1995)
[36] M. Valtiner, G. Grundmeier, Langmuir 26 (2010) 815
[37] Salt Spray Test Reports, Internal Documents, BASF Coatings GmbH
(2009/10)
131
Chapter 4 –
Overall Conclusion and Outlook
The motivation for this study is based in contrasts of the general understanding of
corrosion and delamination of coated substrates and current technology in
substrate treatment for corrosion protection. While it is postulated that corrosion
and coating delamination on coated technical substrates starts with and follows
weak zones on that substrate, industrial surface treatment applies the same
chemistry to the entire substrate surface. The present study followed the
scientific approach of identifying the weak spots of corrosive electrochemical
processes on hot dipped galvanized steel and created material for selective
deposition directly to these spots, with the aim of inhibiting their corrosive activity.
In the first part of this study, investigations of high lateral resolutions on surface
element composition and electrochemical micro probe techniques led to the
identification of the HDG steel substrate weak zones. It was found that aluminum
in HDG alloys segregates not only towards the zinc/iron and zinc/air interface but
also towards the grain boundaries of the zinc grains, even at low concentrations
of 0.5 w.-%. Furthermore it could be shown that the conventional alkaline
cleaning process in contrary to the majority of literature, does not remove all
aluminum from the substrate surface. Even though the overall aluminum content
becomes negligible it still remains within the grain boundaries. Surface potential
mappings utilizing high resolutions of the SKP-FM technique discovered the
lower potential of grain boundaries when compared to the surrounding grain
surfaces. Spots with lower potential than the surrounding matrix are known to be
more corrosively active. The application of the micro capillary cell showed that
Chapter 4 – Overall Conclusion and Outlook
132
grain boundaries tend to dissolve more easily as a result of corrosive currents
that could be measured at potential lower than that on the single grains. The final
corrosion test on a coil coated and scratched sample in a salt spray chamber,
showed higher corrosion activity of the grain boundaries. Corrosion products
along grain boundaries beneath the intact coating material could be observed
after the coating was removed. These findings collectively led to the development
of a model that can predict the occurrence of corrosion on such surfaces; Where
the anodic part reaction is quickly propagated forward along grain boundaries
and is escorted by the local cathode which delaminates the grains; one could
assume that plateau corrosion could follow more easily on delaminated grains. A
derivation from this model would be to slowdown the overall corrosion
propagation by disabling the grain boundary activity. The proof of this model
would lead to new smart coating material that only treated the weak zones of the
substrate which are susceptible to corrosion. The pretreatment of the entire
substrate surface with the same material would then become obsolete and save
the pretreatment material.
In the second part of this study, the design and application of material selectively
on grain boundaries was the focus. It was here found that the anodic dissolution
ability of grain boundaries and their preferable release of aluminum cations
thereof could be used for the selective application of corrosion inhibiting materials
exclusively on grain boundaries.
In the first instance a material survey on their controllability of the grain boundary
deposition was conveyed. Aside from phosphating procedures and surface
spontaneous polymerization, the selective polymer particle deposition showed
the most promising results. Selective phosphating could only be achieved with
poor crystal densities on grain boundaries and showed poor selectivity towards
grain boundaries. Both phosphate and phosphonate precipitation on grain
boundaries could only be achieved after special etching of the grain boundaries
and the placement of seed crystals. The phosphating route towards inhibited
grain boundaries was thus not pursued further in this study.
Spontaneous surface polymerization can be triggered by strong Lewis acids such
as aluminum cations. It was assumed that the aluminum cations released from
the grain boundaries could initiate the spontaneous polymerization at the grain
133
boundaries. Unfortunately the polymerization was rather non-selective. One
could observe even fewer polymers on the grain boundaries than on the grains
themselves. These findings thus also led to an exclusion of the spontaneous
polymerization approach for this particular substrate.
The most promising results could be obtained from the autophoretic deposition of
dispersions. Among the screened dispersions, the primary anionically stabilized
dispersions were the most stable at low pH values as required for the release of
cations from the grain boundaries. Acronal 250 D was found to especially provide
to some extent initial selectivity towards grain boundaries. Henkel’s Aquence is
based on similarly polymer particles but is formulated for application on the entire
surface and thus it was not possible to deposit Aquence selectively on the grain
boundaries. From this survey the approach towards inhibited grain boundaries of
HDG steel was focused on local autophoresis of water borne polymer particles.
The most suitable route towards waterborne dispersions was found in the DPE
block-co-polymerization process. In some previous work it had been found that
monomers such as maleic acid, triethoxyvinylsilane and vinylphosphonic acid
provide strong adhesion to metal/metal oxide substrates. For this reason these
monomers were incorporated into block-co-polymer dispersions. In the case of
vinylphosphonic acid only small amounts, ranging between 17% and 50% of the
added monomer could be co-polymerized. The different polymer dispersions
ranged in particle size from 40 nm to 230 nm and presented a narrow particle
distribution within one batch. The narrow particle distribution was also attributed
to the long-term stability of more than one year for these dispersions.
In the last part of this study however the polymer dispersions obtained were
investigated in terms of their applicability towards the weak zones of HDG steel
substrate surfaces. Considering that grain boundaries are aluminum enriched
and have a specific structure and therefore a higher susceptibility to
electrochemical dissolution, the specifically designed water based polymer
dispersion were investigated for their coagulation behavior in the presence of
aluminum and zinc ions. It was found that the triple charged aluminum ions
immediately cause a destabilization of the polymer particles followed by their
precipitation. Towards the double-charged zinc ions the polymer dispersion
showed tolerance up to a specific ion concentration. This varied influence on the
Chapter 4 – Overall Conclusion and Outlook
134
polymer particle stability was further used to evaluate the application window for
a selective deposition of the polymer on grain boundaries. The pH was therefore
investigated as a trigger for the dissolution of the aluminum cations using surface
plasmon resonance spectroscopy. It was found that an ideal window for a
selective application exists in the pH range of 2.0 to 3.0. At higher pH polymer
particles neither could be deposited on the aluminum oxide surface nor on the
zinc oxide surface. At pH below this window, polymer dispersions were
precipitating on both surfaces. Only in that window were polymer dispersions
precipitating exclusively on the aluminum oxide surface. After defining these
application parameters the goal of selectively coating grain boundaries on the
industrial HDG steel substrate could be realized through a dip coating process.
The final proof of this concept was provided through a comparison of a treated
non-grain boundary versus a treated grain boundary of HDG steel substrates in a
salt spray test where both samples were coil coated and scratched. The results
captured by SEM/EDX imaging showed a significant difference in the condition of
the grain boundaries at the corrosion front. It was found that selectively covered
grain boundaries with a specifically designed polymer reduced the corrosive
damage along the grain boundaries by a factor of three to twenty. On the micro
scale it could be shown that grain boundaries on HDG steel are highly corrosively
active. It has been also shown that it is possible to block grain boundaries and
reduce their corrosive activity. The new method for selective corrosion protection
as introduced within this study could save a tremendous amount of material and
therefore provide cost benefits. Although the corresponding results from the
industrial scale salt spray tests could not show a clear improvement in corrosion
creep reduction of grain boundary inhibited HDG steel substrates, the focus on
the specific chemistry of industrial surfaces on the micro scale may provide new
smart corrosion protection systems.
Based on the results of this study one would recommend that the approach of
selective material application for corrosion protection to the specific surface
characteristics of technical substrates should be pursued. For HDG steel, as
used in this study, one could recommend that different colloidal material towards
grain boundaries such as corrosion inhibitor filled nano-containers or capsules
should be addressed. One could also combine polymer application for grain
boundaries with self-assembling molecules as it is known that SAMs do not
135
provide adequate barrier properties on the rough structure of grain boundaries,
yet they are appropriate for flat topographies as can be found on grains.
All investigations in this study were carried out on one substrate, namely HDG
steel Al 0.5 w.-%. Even though this substrate is one of those most commonly
used in industrial applications, there are a tremendous number of other technical
surfaces that require coating and protection from corrosion degradation. One
could make the assumption that benefits could be gained in the transfer of these
basic findings to further substrates, creating advanced coating materials adapted
to their specific substrate characteristics. Therefore this study underlines the
importance of focusing on specific substrate chemistry in order to develop smart
coatings and thus improved corrosion protection.
Chapter 4 – Overall Conclusion and Outlook
136
xiii
Appendix
Abbreviations
Acronyms
ATRP Atom Transfer Radical Polymerization
BMA Butylmethacrylate
CE Counter Electrode
CRP Controlled Radical Plymerization
DLS Dynamic Light Scattering
DPE 1,1-diphenylethylene
EDX Energy Dispersive X-ray spectroscopy
Eq Equation
FIB Focussed Ion Beam
Fig Figure
GPC Gel Permeation Chromatography
HDG hot dipped galvanized
HEMA Hydroxyethylmethacrylate
HSAB Hard and Soft Acids and Bases
MA Maleic Acid
MMA Methyl Methacrylic Acid
NMP Nitroxide Mediated Radical Polymerization
pH potentia Hydrogenii
PVD Physical Vapor Deposition
RAFT Reversible Addition Fragmentation Chain Transfer
Polymerization
RE Reference Electrode
SEM Scanning Electron Spectroscopy
SKP-FM Scanning Kelvin Probe-Force Microscopy
SPR Surface Plasmon Resonance spectroscopy
Tab Table
TEVS Triethoxyvinylsilane
THF Tetrahydrofurane
VOC Volatile Organic Compounds
VPA Vinylphosphonic Acid
WE Working Electrode
Appendix
xiv
Symbols
% percent
°C degree Celsius
cm centimeter
cm² square centimeter
D polydispersity
E
c
critical excitation voltage
E
0
accelerating voltage
g gram
g/cm³ density in gram per cubic centimeter
g/mol molecular weight in gram per mole
keV kilo electron Volt
kV kilovolt
mm millimeter
Mn weight average Molecular weight
Mw number average Molecular weight
mV millivolt
nm nanometer
s second
ρ density
µm micrometer
θ theta
ζ zeta
xv
Publications
M. Dornbusch, H. Hintze-Brüning, S.Toews, W. Bremser, „Verfahren zur
Autophoretischen Beschichtung“ Pat.: DE 10 2010 019 245.7 (2010)
S. Toews, W. Bremser, H. Hintze-Bruening, M. Dornbusch, „Smart Functionalized
Polymer Dispersions for effective mapping of heterogeneous metal surfaces: Part
1 - Substrate Characterization” submitted to Prog. Org. Coat.
S. Toews, W. Bremser, H. Hintze-Bruening, M. Dornbusch, “New concepts for
corrosion protection” EUROCORR 2010 – The European Corrosion Congress,
Moscow, RUS (2010), Oral Presentation
S. Toews, W. Bremser, “Smart functionalized polymer dispersions for selective
adsorption to metal oxide surfaces” 240
th
ACS National Meeting, Boston, MA,
USA (2010), Poster
S. Toews, W. Bremser, H. Hintze-Bruening, S. Sinnwell, M. Dornbusch, R.
Bautista-Mester, W. Kreis, „Smart Functionalized Polymer Dispersions for
effective mapping of heterogeneous metal surfaces: New Concepts for Corrosion
Protection” Coatings Science International Conference Noordwijk, NL (2010),
Oral Presentation
S. Toews, W. Bremser, “Smart Corrosion Protection” 8
th
Spring Meeting of the
International Society of Electrochemistry, Columbus OH, USA (2010),
Oral Presentation
S. Toews, N. Pollmann, O. Seewald, W. Bremser, „1,1-diphenylethylene a new
route to functional block-co-polymers” 239
th
ACS National Meeting, San
Francisco, CA, USA (2010), Poster
S. Toews, N. Pollmann, O. Seewald, W. Bremser, “Functional Block-co-Polymers
via the DPE Route” Coatings Science International Conference, Noordwijk, NL
(2009), Poster
