Atmospheric corrosion of zinc-aluminum and copper-based alloys in chloride-rich environments

Microstructure, corrosion initiation, patina evolution and metal release

 

Xian Zhang

Doctoral Thesis

Stockholm, Sweden 2014

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan framlägges till offentlig granskning for avläggande av teknologie doktorsexamen fredagen den 26 september 2014 klockan 10:00 i Kollegiesalen, Kungliga Tekniska Högskolan, Brinellvägen 8, Stockholm. Avhandlingen presenteras på engelska. Atmospheric corrosion of zinc-aluminum and copper-based alloys in chloride-rich environments. -Microstructure, corrosion initiation, patina evolution and metal release. Xian Zhang (xianzh@kth.se) Doctoral Thesis KTH Royal Institute of Technology School of Chemical Science and Engineering Department of Chemistry Division of Surface and Corrosion Science Drottning Kristinas väg 51 SE-100 44 Stockholm, Sweden TRITA-CHE Report 2014:27 ISSN 1654-1081 ISBN 978-91-7595-203-1 Denna avhandling är skyddad enligt upphovsrättslagen. Alla rättigheter förbehålles. Copyright © 2014 Xian Zhang. All rights reserved. No part of this thesis may be reproduced by any means without permission from the author. The following items are printed with permission: PAPER I: © 2012 Elsevier PAPER II: © 2013 Elsevier PAPER III: © 2013 Elsevier PAPER IV: © 2014 Elsevier PAPER V: © 2014 Elsevier Printed at Universitetsservice US-AB

  

 

 

 

 

 

 

Make things as simple as possible,

but not simpler.

Albert Einstein

  

 

 

 

 

 

 

 

 

 

 

 

 

i  

Abstract

Fundamental understanding of atmospheric corrosion mechanisms requires an in-depth understanding on the dynamic interaction between corrosive constituents and metal/alloy surfaces. This doctoral study comprises field and laboratory investigations that assess atmospheric corrosion and metal release processes for two different groups of alloys exposed in chloride-rich environments. These groups comprise two commercial Zn-Al alloy coatings on steel, Galfan™ (Zn5Al) and Galvalume™ (Zn55Al), and four copper-based alloys (Cu4Sn, Cu15Zn, Cu40Zn and Cu5Zn5Al). In-depth laboratory investigations were conducted to assess the role of chloride deposition and alloy microstructure on the initial corrosion mechanisms and subsequent corrosion product formation. Comparisons were made with long-term field exposures at unsheltered marine conditions in Brest, France.

A multitude of surface sensitive and non-destructive analytical methods were adopted for detailed in-situ and ex-situ analysis to assess corrosion product evolution scenarios for the Zn-Al and the Cu-based alloys. Scanning electron microscopy and energy dispersive spectroscopy (SEM/EDS) were employed for morphological investigations and scanning Kelvin probe force microscopy (SKPFM) for nobility distribution measurements and to gain microstructural information. SEM/EDS, infrared reflection-absorption spectroscopy (IRAS), confocal Raman micro-spectroscopy (CRM) and grazing incidence x-ray diffraction (GIXRD) were utilized to gain information on corrosion product formation and possibly their lateral distribution upon field and laboratory exposures. The multi-analytical approach enabled the exploration of the interplay between the microstructure and corrosion initiation and corrosion product evolution.

A clear influence of the microstructure on the initial corrosion product formation was preferentially observed in the zinc-rich phase for both the Zn-Al and the Cu-Zn alloys, processes being triggered by microgalvanic effects. Similar corrosion products were identified upon laboratory exposures with chlorides for both the Zn-Al and the Cu-based alloys as observed after short and long term marine exposures at field conditions. For the Zn-Al alloys the sequence includes the initial formation of ZnO, ZnAl2O4 and/or Al2O3 and subsequent formation of Zn6Al2(OH)16CO3·4H2O, and Zn2Al(OH)6Cl·2H2O and/or Zn5(OH)8Cl2·H2O. The patina of Cu sheet consists of two main layers with Cu2O predominating in the inner layer and Cu2(OH)3Cl in the outer layer, and with a discontinuous presence of CuCl in-between. Additional patina constituents of the Cu-based alloys include SnO2, Zn5(OH)6(CO3)2,

ii

Zn6Al2(OH)16CO3·4H2O and Al2O3. General scenarios for the evolution of corrosion products are proposed as well as a corrosion product flaking mechanism for some of the Cu-based alloys upon exposure in chloride-rich atmospheres.

The tendency for corrosion product flaking was considerably more pronounced on Cu sheet and Cu4Sn compared with Cu15Zn and Cu5Al5Zn. This difference is explained by the initial formation of zinc- and zinc-aluminum hydroxycarbonates Zn5(OH)6(CO3)2 and Zn6Al2(OH)16CO3·4H2O on Cu15Zn and Cu5Al5Zn, corrosion products that delay the formation of CuCl, a precursor of Cu2(OH)3Cl. As a result, the observed volume expansion during transformation of CuCl to Cu2(OH)3Cl, and the concomitant flaking process of corrosion products, was less severe on Cu15Zn and Cu5Al5Zn compared with Cu and Cu4Sn in chloride-rich environments. The results confirm the barrier effect of poorly soluble zinc and zinc-aluminum hydroxycarbonates Zn5(OH)6(CO3)2 and Zn6Al2(OH)16CO3·4H2O, which results in a reduced interaction between chlorides and surfaces of Cu-based alloys, and thereby reduced formation rates of easily flaked off corrosion products. From this process also follows reduced metal release rates from the Zn-Al alloys.

Keywords: atmospheric corrosion, chloride deposition, Zn-Al alloy coatings on steel, Cu sheet and Cu alloys, microstructure, corrosion initiation, corrosion product evolution, metal release, SEM, IRAS, CRM.

iii  

Sammanfattning

Bättre molekylär förståelse för metallers atmosfäriska korrosion kräver en fördjupad kunskap i det dynamiska samspelet mellan atmosfärens korrosiva beståndsdelar och metallytan. Denna doktorsavhandling omfattar laboratorie- och fältundersökningar av korrosions- och metallfrigöringsprocesser av två grupper av legeringar som exponerats i kloridrika atmosfärsmiljöer: två kommersiella Zn-Al beläggningar på stål, Galfan™ (Zn med 5% Al, förkortat Zn5Al) och Galvalume™ (Zn55Al), samt fyra kopparbaserade legeringar (Cu4Sn, Cu15Zn, Cu40Zn och Cu5Zn5Al). Undersökningar har genomförts i renodlade laboratorie-miljöer med för-deponerade NaCl-partiklar i en atmosfär av varierande relativ fuktighet. Syftet har varit att utvärdera betydelsen av kloriders deposition och legeringarnas mikrostruktur på korrosionsmekanismen samt bildandet av korrosionsprodukter. Jämförelser av korrosionsmekanismer har även gjorts efter flerårsexponeringar av samma legeringar i en marin fältmiljö i Brest, Frankrike.

Undersökningarna har baserats på ett brett spektrum av analysmetoder för detaljerade studier dels under pågående atmosfärisk korrosion (in-situ), och dels efter avslutad korrosion (ex-situ). Legeringarnas mikrostruktur och tillhörande variation i ädelhet hos olika faser har undersökts med svepelektronmikroskopi och energidispersiv röntgenmikroanalys (SEM/EDS) samt med en variant av atomkraftsmikroskopi (engelska: scanning Kelvin probe force microscopy, SKPFM). Korrosionsprodukternas tillväxt har analyserats in-situ med infraröd reflektions-absorptionsspektroskopi (IRAS), samt morfologi och sammansättning av bildade korrosionsprodukter ex-situ med SEM/EDS, konfokal Raman mikro-spektroskopi (CRM) samt röntgendiffraktion vid strykande ifall (GIXRD). Det multi-analytiska tillvägagångssättet har medfört att det komplexa samspelet mellan de skilda legeringarnas mikrostruktur, korrosionsinitiering och bildandet av korrosionsprodukter kunnat studeras i detalj.

En tydlig påverkan av mikrostruktur på det initiala korrosionsförloppet har kunnat påvisas. Korrosionsinitieringen sker företrädesvis i mer zinkrika faser för såväl Zn-Al- som Cu-Zn-legeringar och orsakas av mikro-galvaniska effekter mellan de mer zinkrika, mindre ädla, faserna och omgivande faser. Deponerade NaCl-partiklar påskyndar den lokala korrosionen oberoende av mikrostruktur. Snarlika sekvenser av korrosionsprodukter har kunnat påvisas såväl efter laboratorie- som fältexponeringar. För Zn-Al-legeringar bildas först ZnO, ZnAl2O4 och/eller Al2O3, därefter Zn6Al2(OH)16CO3·4H2O och Zn2Al(OH)6Cl·2H2O och/eller Zn5(OH)8Cl2·H2O. På ren koppar bildas ett inre skikt dominerat av Cu2O, ett mellanskikt av CuCl och ett

iv

yttre skikt med i huvudsak Cu2(OH)3Cl. Beroende på legeringstillsats har även SnO2

och Zn5(OH)6(CO3)2 kunnat identifieras.

En mekanism för flagning av korrosionsprodukter på kopparbaserade legeringar i kloridrika atmosfärer har utvecklats. Tendensen för flagning har visat sig vara mycket mer uttalad på ren Cu och Cu4Sn än på Cu15Zn och Cu5Al5Zn. Skillnaden kan förklaras med hjälp av det tidiga bildandet av Zn5(OH)6(CO3)2 och Zn6Al2(OH)16CO3·4H2O på Cu15Zn och Cu5Al5Zn som fördröjer bildandet av CuCl, en föregångare till Cu2(OH)3Cl. Därigenom hämmas även den observerade volymexpansionen som sker när CuCl omvandlas till Cu2(OH)3Cl, en process som visar sig vara den egentliga orsaken till att korrosionsprodukterna flagar. Resultaten bekräftar barriäreffekten hos de mer svårlösliga faserna Zn5(OH)6(CO3)2 och Zn6Al2(OH)16CO3·4H2O, vilken dels resulterar i en minskad växelverkan mellan klorider och de legeringsytor där dessa faser kan bildas, och dels i en reducerad metallfrigöringshastighet.

v  

Preface

This doctoral thesis provides a comprehensive understanding of atmospheric corrosion and metal release properties of Zn-Al alloy coatings and Cu-based alloys after both short-term laboratory, and long-term field exposures in chloride-rich environments. The main focus is placed on multi-analytical investigations to assess the influence of microstructure on corrosion initiation and corrosion product evolution. Investigated materials and main topics of each scientific paper of this thesis are schematically presented below.

vi

List of papers 

I. The initial release of zinc and aluminum from non-treated Galvalume and the formation of corrosion products in chloride containing media

X. Zhang, T.-N. Vu, P. Volovitch, C. Leygraf, K. Ogle, I. Odnevall Wallinder

Applied Surface Science, 258 (2012) 4351-4359

II. Atmospheric corrosion of Galfan coatings on steel in chloride-rich environments

X. Zhang, C. Leygraf, I. Odnevall Wallinder

Corrosion Science, 73 (2013) 62-71

III. Selected area visualization by FIB-milling for corrosion-microstructure analysis with submicron resolution

X. Zhang, C. Leygraf, I. Odnevall Wallinder

Materials Letters, 98 (2013) 230-233

IV. Corrosion and runoff rates of Cu and three Cu-alloys in marine environments with increasing chloride deposition rate

I. Odnevall Wallinder, X. Zhang, S. Goidanich, N. Le Bozec, G. Herting, C. Leygraf

Science of the Total Environment, 472 (2014) 681-694

V. Mechanistic studies of corrosion product flaking on copper and copper-based alloys in marine environments

X. Zhang, I. Odnevall Wallinder, C. Leygraf

Corrosion Science, 85 (2014) 15-25

Results from the following paper is partly included in the summary of this thesis:

The role of microstructure on the initial corrosion of Cu-Zn alloys in a chloride-containing laboratory atmosphere 

X. Zhang, C. Leygraf, I. Odnevall Wallinder 

Manuscript.

In addition, some selected unpublished work is presented in the summary of this thesis.

vii  

Author contribution to the papers

The contribution of the respondent to the papers is listed below: 

Paper I. Major part of experimental work, except for AESEC and XPS/AES measurements. Major part of data interpretation and manuscript preparation.

Paper II. Main part of experimental work, except for FEG-SEM/EDS measurements on unexposed surfaces. Main part of data interpretation and manuscript preparation.

Paper III. Major part of experimental work, except for FIB-SEM and SKFPM measurements. Main part of data interpretation and manuscript preparation.

Paper IV. Part of experimental work, active contribution in SEM/EDS, CRM and GIXRD measurements. Part of data interpretation and manuscript preparation.

Paper V. Main part of experimental work, except for FEG-SEM/EDS measurements on cross-sections. Main part of data interpretation and manuscript preparation.

viii

Abbreviations

AAS Atomic absorption spectroscopy AES Auger electron spectroscopy

AESEC Atomic emission spectroelectrochemistry AFM Atomic force microscopy BSE Backscattered electrons CRM Confocal Raman micro-spectroscopy

DTGS Deuterated triglycine sulfate EDS Energy dispersive x-ray analysis EMF Standard electromotive force ESEM Environmental scanning electron microscopy

FEG Field emission gun FIB-SEM Focused ion beam-scanning electron microscopy FTIR Fourier transform infrared GDOES Glow discharge optical emission spectroscopy

GF Graphite furnace GIXRD Grazing incidence x-ray diffraction IR Infrared IRAS Infrared reflection absorption spectroscopy

MCT Mercury cadmium telluride NHE Normal hydrogen electrode

N-VDA New revised VDA corrosion test method Standard SEP 1850 VDA 621-415 B

OCP Open circuit potential OM Optical microscopy SCE Saturated calomel electrode SE Secondary electrons

SEM Scanning electron microscopy SHE Standard hydrogen electrode SKFPM Scanning Kelvin probe force microscopy XPS X-ray photoelectron spectroscopy

XRD X-ray diffraction  

ix  

Table of Contents

Abstract ................................................................................................................................... i

Sammanfattning .................................................................................................................. iii

Preface .................................................................................................................................... v

List of papers ........................................................................................................................ vi

Author contribution to the papers..................................................................................... vii

Abbreviations .................................................................................................................... viii

1 Introduction ..................................................................................................................... 1

1.1 Motivation and scope .................................................................................................... 1

1.2 Atmospheric corrosion .................................................................................................. 3

1.2.1 Metal surface interaction with the atmosphere ...................................................... 3

1.2.2 Atmospheric gases and particles ............................................................................ 4

1.3 Zn-Al and Cu-based alloys ............................................................................................ 5

1.3.1 Zn-Al alloy coatings ............................................................................................... 5

1.3.2 Cu-based alloys ...................................................................................................... 6

1.4 Atmospheric corrosion of Zn-Al and Cu-based alloys .................................................. 7

1.4.1 Corrosion product formation and metal release .................................................... 7

1.4.2 Microstructure-related galvanic corrosion ............................................................ 9

2 Materials and methods ................................................................................................. 12

2.1 Materials and surface preparation ............................................................................... 12

2.1.1 Materials ............................................................................................................... 13

2.1.2 Surface preparation .............................................................................................. 13

2.2 Exposure conditions .................................................................................................... 14

2.2.1 Laboratory wet/dry cycle exposure (Papers II, V) ............................................... 14

2.2.2 Laboratory immersion and flow-cell tests (Paper I) ............................................ 15

2.2.3 Long-term field exposure (Papers I, II, IV, V)...................................................... 15

2.2.4 Accelerated N-VDA test (non-published data) ..................................................... 16

3 Analytical techniques .................................................................................................... 17

3.1 Scanning electron microscopy with x-ray microanalysis (SEM/EDS) ....................... 18

3.1.1 Environmental - SEM (ESEM) ............................................................................. 19

3.1.2 Focused ion beam - SEM (FIB-SEM) ................................................................... 20

x

3.2 Infrared reflection absorption spectroscopy (IRAS) ................................................... 20

3.3 Confocal Raman micro-spectroscopy (CRM) ............................................................. 22

3.4 Grazing incidence x-ray diffraction (GIXRD) ............................................................ 23

3.5 Scanning Kelvin probe force microscopy (SKPFM) .................................................. 24

3.6 X-ray photoelectron spectroscopy/ Auger electron spectroscopy (XPS/AES) ........... 25

3.7 Glow discharge optical emission spectroscopy (GDOES) .......................................... 25

3.8 Optical microscopy (OM)/ Stereomicroscopy ............................................................ 26

3.9 Atomic absorption spectroscopy (AAS) ..................................................................... 26

3.10 Atomic emission spectroelectrochemistry (AESEC) ................................................ 27

4 Influence of microstructure on corrosion initiation ................................................... 28

4.1 The eutectic structure of Galfan consists of an η-Zn matrix and β-Al lamellas and rods of lower surface nobility compared with the matrix, and are separated by β-Al grain boundaries in which ZnO and Al2O3 preferentially form. (Paper III) ............................... 28

4.2 Corrosion initiation observed for Galfan in the zinc-richer η-Zn phase adjacent to the less zinc-rich β-Al phase. Both carbonate and chloride-containing phases are formed in humidified air and in the presence of NaCl. (Paper II) ..................................................... 30

4.3 Selective zinc release and corrosion initiation in the zinc-rich phase observed for Galvalume in chloride containing media. Long-term correlation observed between the released zinc fraction and the surface coverage of zinc and aluminum-rich corrosion products. (Paper I) ............................................................................................................. 34

4.4 The dual-phase structure of Cu40Zn consists of zinc-richer β-phase crystals of lower surface nobility than the α-phase. Corrosion initiation is observed in the β-phase at low pre-deposition of NaCl. (non-published data) ................................................................... 37

4.5 Microgalvanic effects on a Cu-Zn patterned sample with pre-deposited chlorides result in a radial distribution of corrosion products from the Cu cathode to the Zn anode upon cyclic exposures in humidified air. (non-published data) ........................................ 39

5 Corrosion product evolution and characteristics......................................................... 43

5.1 Severe corrosion product flaking observed for Cu and Cu4Sn in chloride-rich environments is primarily connected to the presence of nantokite. Minor effects observed for Cu15Zn and Cu5Al5Zn. (Paper V).............................................................................. 43

5.2 Transformation of nantokite to paratacamite results in volume expansion within the patina causing corrosion product flaking for Cu and Cu4Sn. (Paper V) ........................... 46

5.3 The initial formation of Zn- and Zn/Al-hydroxycarbonates reduces the sensitivity of Cu15Zn and Cu5Al5Zn to chloride-induced corrosion, and also the release of zinc from Galfan at marine conditions. (Papers II, IV, V) ................................................................ 51

xi  

5.4 Similar corrosion products form on Galfan and bare Zn sheet and on Galvalume and bare Al sheet, respectively, upon accelerated chloride test conditions. (non-published data) ................................................................................................................................... 53

5.5 Laboratory set-ups with exposures to chloride-rich environments were able to successfully reproduce the predominating corrosion products formed at marine outdoor conditions for the Zn-Al coatings, bare Cu sheet and the Cu-based alloys. (Papers I, II, IV, V) ................................................................................................................................ 56

5.5.1 Zn-Al alloy coatings ............................................................................................. 56

5.5.2 Cu and Cu-based alloys ........................................................................................ 57

5.6 General scenarios for patina evolution established for Zn-Al coatings and corrosion product flaking mechanisms proposed for Cu-based alloys in chloride-rich atmospheric environments. (Papers II, V) ............................................................................................. 57

5.6.1 Zn-Al alloy coatings ............................................................................................. 57

5.6.2 Cu and Cu-based alloys ........................................................................................ 59

6 Summary and outlook .................................................................................................. 61

Acknowledgements ............................................................................................................. 65

References ............................................................................................................................ 67

 

 

 

 

 

 

 

 

 

 

 

 

 

xii

 

 

 

 

 

 

 

 

 

1  

1 Introduction

1.1 Motivation and scope

Atmospheric corrosion of materials at outdoor conditions is a corrosion process that results in large economic losses in the society. Increased levels of corrosive pollutants have in different parts of the world resulted in dramatic deterioration of metal surfaces used in outdoor constructions, in vehicles and other surfaces of the cultural heritage [1], whereas corrosion effects have been less severe in areas of reduced pollutant levels [2]. Prevention measures against atmospheric corrosion in high-technology societies have been reported to account for almost half the total estimated cost for corrosion protection [3].

Since the nature of atmospheric corrosion is inherently complicated, in-depth fundamental understanding of prevailing corrosion processes is of great importance for the society, e.g. for regulators and the industry [1]. The use of metals and alloys at outdoor conditions is of high necessity in the modern society due to their excellent mechanical properties and good corrosion resistance. Atmospheric corrosion of alloys is even more complex compared with the pure metals, as the mechanisms depend not only on prevailing environmental conditions but also on alloying elements and differences in microstructure and surface characteristics. The presence of e.g. secondary phases, grain boundaries, and inclusions may negatively affect the overall corrosion performance [4].

This doctoral thesis comprises extensive studies that contribute to a more comprehensive understanding of atmospheric corrosion and metal release processes of commercial Zn-Al alloys (of relevance for automotive applications) and Cu-based alloys (used in outdoor construction applications) in chloride-rich environments. The thesis includes both short-term laboratory and long-term marine field exposures aiming to assess initial atmospheric corrosion mechanisms and the evolution of corrosion products in chloride-rich environments, and their link to microstructural features of the investigated alloys. The research approach is summarized in Fig.1.1.

The influence of humidity and chlorides has been elucidated through successive short-term controlled laboratory exposures and the use of different near surface- and bulk sensitive analytical tools. Long-term data from marine field exposures has been used to ensure realistic laboratory simulations. Microstructure-related corrosion initiation and corrosion product evolution were investigated in-situ using IRAS (infrared reflection absorption spectroscopy) and ESEM/EDS (environmental

2

scanning electron microscopy with x-ray microanalysis), and ex-situ with SEM/EDS, CRM (confocal Raman micro-spectroscopy), GIXRD (grazing incidence x-ray diffraction), SKPFM (scanning Kelvin probe force microscopy), XPS/AES (x-ray photoelectron spectroscopy/ Auger electron spectroscopy), and GDOES (glow discharge optical emission spectroscopy). Metal release processes were evaluated in-situ with AESEC (atomic emission spectroelectrochemistry) and ex-situ with AAS (atomic absorption spectroscopy).

Research activities within this doctoral study are connected to an EU project (Autocorr, RFSR-CT-2009-00015) focusing on corrosion of heterogeneous metal-metal assemblies in the automotive industry, and to a long-term international industry consortium project focusing on atmospheric corrosion and environmental metal dispersion from outdoor construction materials. The test materials were supplied via Arcelor Mittal, France, KME, Germany and Aurubis, Finland. The experimental work was performed at the Division of Surface and Corrosion Science at KTH in collaboration with colleagues at Ecole Nationale Supérieure de Chimie de Paris, France and Politecnico di Milano, Italy.

Figure 1.1. Summary of main aspects investigated in this doctoral thesis.

3  

1.2 Atmospheric corrosion

The science of atmospheric corrosion was evolved by the pioneering work of Vernon [5] almost a century ago. Atmospheric corrosion is the result of an interaction between a material, mostly a metal, and its surrounding atmosphere. The process is triggered by relative humidity levels that result in a thin aqueous layer at the surface of the material [1]. The water layer may vary from monomolecular thickness to clearly visible water films depending on prevailing humidity conditions [3]. Environmental pollutants will interact with the aqueous film and influence the corrosion process in different ways. As an interdisciplinary field of science, atmospheric corrosion has become more widely investigated during the past decades assessing an improved molecular understanding of corrosion processes [6], and environmental effects induced by released metals from corroded surfaces [7, 8].

1.2.1 Metal surface interaction with the atmosphere

The field of atmospheric corrosion integrates diverse subjects such as chemistry, electrochemistry, material science and physics. Atmospheric corrosion is complex since prevailing processes take place in several regimes, at the interfaces between the gaseous phase and the liquid phase, and between the liquid and the solid phase [1]. Figure 1.2 schematically demonstrates the multi-regime (gas, liquid and solid phases separated by two interfaces) involved in atmospheric corrosion for a given metal surface.

Figure 1.2. Different regimes involved in atmospheric corrosion.

4

In the initial stage of atmospheric corrosion, water vapor instantly reacts with the metal surface that becomes hydroxylated. Further exposure in humid air results in the adsorption of water as monomolecular layers, or as a thin aqueous adlayer [1].

Atmospheric constituents, gaseous pollutants and airborne salt particles deposit at the intermediate stage on the metal surface and dissolve to different extent within the thin adlayer, processes that result in a variety of chemical and electrochemical interfacial reactions [1]. Unevenly distributed water droplets result in local electrochemical corrosion cells that form spatially separated anodes and cathodes [9]. The dominating electrochemical reactions for most corrosion processes include metal oxidation and oxygen reduction:

M → M n+ + ne- Anodic reaction

O2 + 2H2O + 4e- → 4OH- Cathodic reaction

Dissolved metal ions can coordinate with counterions present in the aqueous layer, and eventually precipitate into a solid phase (nucleation of corrosion products), when the concentration of ion pairs in the aqueous layer reaches supersaturation.

In the final stage of atmospheric corrosion, the number and size of precipitated nuclei increase with prolonged exposure until eventually they completely cover the metal surface, often referred to as a corrosion product layer.

1.2.2 Atmospheric gases and particles

The most important atmospheric constituents of importance for atmospheric corrosion include H2O and CO2, dominantly present in the atmosphere as corrosion stimulators. Other corrosive constituents include atmospheric gases such as sulfur dioxide (SO2), nitrogen dioxide (NO2), ammonia (NH3) and ozone (O3), organic acids, and airborne particles containing e.g. chlorides (Cl-), sulfates (SO4

2-) and nitrates (NO3

-) [10, 11].

Humidity is a measure of the quantity of water vapor present in the atmosphere. The relative humidity (RH) is defined as the ratio between the absolute humidity and the saturation quantity [11]. Due to climatic variations, RH significantly fluctuates during the day, and from day to day in open atmosphere, important factors together with temperature variations for the resulting aqueous layer on a metal surface. This layer is essential to initiate atmospheric corrosion as it acts as a medium for electrochemical and chemical reactions as well as a solvent for atmospheric constituents, as discussed above [1].

CO2 is a natural constituent of the atmosphere and also emitted by anthropogenic activities such as fossil fuel combustion. The average atmospheric concentration of

5  

CO2 is about 350-400 ppm. Even though it is relatively unreactive in ambient air, it is soluble in water and forms bicarbonate (HCO3

-) and carbonate ions (CO32-) that for

many metals have a major influence on the corrosion process [1]. The most significant effect of CO2 on atmospheric corrosion is its participation in the formation of corrosion products on non-ferrous metals [9]. Prevailing corrosion processes have been extensively investigated and been validated in the scientific literature, in particular for chloride-rich environments [12-16].

Chloride-rich aerosols are the dominating pollutants in marine environments that for many metals/alloys accelerate the corrosion rate by several orders of magnitude [17]. These aerosols are suspensions of small liquid or solid particles originating from salt spray and fog in the vicinity of the seashore [11]. Deposition of such aerosols provides a relatively corrosive aqueous surface layer of high conductivity in which the chloride ion interacts with the surface initiating corrosion, e.g. pitting. Depending on metal/alloy surface oxide properties, chlorides can locally destroy native surface oxides and hydroxides of different passive properties. The importance of chlorides on the atmospheric corrosion of metals has been extensively investigated in the scientific literature [18-22].

Gaseous SO2 is a common environmental pollutant in the atmosphere that is corrosive for many metals [1], and originates predominantly from combustion of coal, oil, and gasoline. SO2 is predominantly present in urban and industrial atmospheres and is easily dissolved in water forming different sulfur species including e.g. HSO3

- and SO4

2- both in airborne water droplets and in the aqueous surface layer at the metal surface [11]. This layer often obtains a reduced local pH (often below 4.5) that can, for some metals, reduce the oxide stability and increase the corrosion rate [1, 9]. Its effects have been extensively studied for different metals, also in combination with other pollutants such as CO2, NO2, chlorides and O3, due to its significance at atmospheric conditions [18, 20, 23, 24].

1.3 Zn-Al and Cu-based alloys

1.3.1 Zn-Al alloy coatings

Hot-dip galvanized zinc coatings containing aluminum on steel substrates have been frequently used in many applications of atmospheric relevance due to their sacrificial protection of steels and beneficial corrosion resistance [25, 26]. Two different commercial Zn-Al alloy coatings on steel, GalfanTM (Zn-5wt% Al) and GalvalumeTM (Zn-55wt% Al), of widespread use in automotive, construction and industrial applications [27], have been investigated in this thesis.

6

The Galfan coating contains 5 wt% Al, which is very close to the eutectic point in the Zn-Al system [28, 29], and up to 0.05 wt% mischmetal (lanthanum and cerium). Due to its high corrosion resistance, superior ductility and forming properties it is often used in automotive applications [30]. The microstructure is characterized by a two-phase structure with a zinc-rich proeutectoid phase (Al/(Al+Zn) <5wt%) and a eutectic phase (Al/(Al+Zn): 5-10 wt%) consisting of beta (β) aluminum (5-25 wt% Al) and eta (η) zinc-rich lamellas (<5 wt% Al) [31-34].

The Galvalume coating has a significantly higher content of aluminum compared with Galfan and low amount of silicon (55wt% Al-43.4 wt % Zn-1.6 wt % Si [35]). The corrosion resistance of Galvalume coatings on steel is remarkably improved compared with a bare galvanized steel surface of equivalent coating thickness upon exposure in marine and industrial atmospheres. The coating efficiently protects the substrate at scratches and cut edges.[36]. The Galvalume coating has a two-phase microstructure with an aluminum-rich dendritic phase (Al/(Al+Zn): 60-80 wt%) and a zinc-rich interdendritic phase (Al/(Al+Zn): 20-40 wt%) [14, 37].

Differences in cooling rate, annealing time and nucleation temperature upon solidification influence the microstructure, and subsequently the corrosion resistance of both Galfan and Galvalume [38-40].

1.3.2 Cu-based alloys

Copper in its pure or alloyed state forms a large group of industrially very important materials. Depending on alloying elements, the desirable properties range from high electrical conductivity, corrosion resistance, wear resistance, tensile strength, soldering and joining characteristics, to an appealing visual appearance. Below follows a few examples of copper-based alloys used in atmospheric applications of relevance for outdoor constructions such as roofs, claddings and facades.

Bronze alloys often contain tin as the main element. This alloy was developed already four thousand years ago. The tensile strength and corrosion resistance of the alloys are for most applications superior compared to bare copper sheet. Bronze alloys contains, similar to bare copper, often small amounts of phosphorus that further increase the alloy hardness and wear resistance [41].

Brass alloys consist primarily of copper and zinc in different proportions and have

good strength and ductility [41]. Alpha () brass with up to 35 wt% zinc consists

primarily of one phase with face-centred cubic crystal structure. The - brasses

(sometimes called duplex brasses) contain 35-45 wt% zinc and consist of -dendrites

in a body-centred cubic -matrix [42]. The dual-phase structure consists, according to

literature findings, of precipitated -phase crystals in a β-phase matrix [43, 44]. The -

7  

phase, of higher zinc content than the -phase, exhibits a characteristic plate-type morphology with plates orientated in different directions [42].

Cu-based alloys with aluminum (5 wt%) and zinc (5 wt%) have during recent years found an increasing use. Their welding and soldering properties are very good and they are malleable, regardless of temperature and rolling direction [45].

1.4 Atmospheric corrosion of Zn-Al and Cu-based alloys

1.4.1 Corrosion product formation and metal release

Atmospheric corrosion of zinc at natural weathering conditions has been widely investigated [19, 46-52]. A general evolution scheme of corrosion product formation on zinc was established by Odnevall and Leygraf [53]. A thin layer of amorphous hydroxycarbonate (Hydrozincite, Zn5(OH)6CO3) forms rapidly in any humid-containing atmosphere. In humid, low polluted environments, this phase often gradually evolves into its crystalline form. If chlorides are also present in the atmosphere, zinc hydroxychloride (Simonkolleite, Zn5(OH)8Cl2·H2O) is locally formed and eventually evolves into a sodium zinc hydroxychlorosulfate (Gordaite, NaZn4Cl(OH)6SO4·6H2O). The presence of SO2 and sulfates results in the formation of different zinc hydroxysulfates (Zn4SO4(OH)6·nH2O) and, in the presence of chlorides, also zinc chlorohydroxysulfate (Zn4Cl2(OH)4SO4·5H2O) may form. Atmospheric corrosion processes of zinc at laboratory-simulated conditions have been investigated in a large number of studies [54-59]. In-situ surface analyses reveal Zn5(OH)8Cl2·H2O as the main corrosion product formed on bare zinc sheet at chloride-rich conditions [18].

Extensive investigations are also reported in the scientific literature on corrosion rates and long-term corrosion product formation on zinc-aluminum coatings on steel at atmospheric conditions [19, 36, 60-64]. Zn6Al2(OH)16CO3·4H2O as well as Zn2Al(OH)6Cl·2H2O [65] have, in addition to zinc hydroxychlorides and carbonates, been identified in several investigations [66-68] for Zn-Al coatings. Al(OH)3 has been identified on Galvalume at different exposure conditions [60, 69] and Al2(OH)6·4H2O and AlZn sulfate hydrates after long term field exposures [60].

Significant knowledge exists in the scientific literature on patina formation on copper at marine exposure conditions, and to some extent also for some copper-based alloys in chloride-rich environments [50, 70-77]. Cuprite (Cu2O), and to some extent CuO, are the initial phases in the evolution of the copper patina at sheltered marine exposure conditions. Interaction with chlorides results in the formation of nantokite (CuCl) which typically transforms to atacamite or the isomorphous phase paratacamite (Cu2(OH)3Cl) as end corrosion products [70]. These constituents have also been

8

identified within the patina on bare copper at unsheltered marine exposure conditions [71], and also been observed at laboratory conditions in humidified air with pre-deposited NaCl [12, 72].

As evidenced from a recent review [78] has the atmospheric corrosion of Cu-alloys such as bronze or brass in chloride-dominating environments been significantly less extensively studied compared with bare copper. Similar patina constituents as formed on bare copper have also been observed within the patina on restored ancient bronzes [79-81]. Nantokite (CuCl), trapped within the patina of archeological bronzes, and paratacamite (Cu2(OH)3Cl) are important constituents that govern the so-called bronze disease, which is a progressive corrosion process mainly based on the conversion of nantokite to paratacamite. This process leads to volume expansion and subsequent cracking or flaking of formed corrosion products [80, 81]. Tin oxide (SnO2) has been identified within the patina in different environments [82, 83]. Amorphous and/or crystalline zinc hydroxycarbonate, hydrozincite (Zn5(OH)6(CO3)2) and zinc oxide (ZnO) that cover large surface areas have been observed on brass at sites of low chloride deposition rates [75]. Zinc oxide has been identified as the main corrosion product formed at laboratory conditions in humidified air with reduced CO2-concentrations [84].

The interaction of metal surfaces and surrounding environments results not only in corrosion and formation of corrosion products and their evolution but also in the release of dissolved metals from the patina. This portion can be transported from the surface by means of atmospheric precipitation. Metal release (patina dissolution) occurs at the interface between the corrosive media and the patina, and should be distinguished from corrosion (oxidation) [85] taking place at the bare metal surface. The concern of dispersed metals into the environment and potential adverse effects has gradually increased world-wide during the past decades [86], primarily as a consequence of considerably reduced emissions of metals from point sources.

Several investigations have during recent decades explored the effect of metal release of zinc and copper from outdoor construction materials [8, 61, 76, 86-92]. Results from long-term field exposures and parallel laboratory studies have confirmed that the metal release process is largely dependent on e.g. the corrosion product composition and characteristics and to prevailing environmental conditions. It has been reported that the zinc release pattern from Galvalume during long-term field exposures follows the same time-dependence as the surface coverage of zinc-rich and aluminum-rich corrosion products [14]. The release of copper from copper alloys exhibits typically lower rates compared with bare copper sheet [75]. Similar observations have been reported for the release of zinc from Galvalume compared with bare zinc sheet [14]. It should be taken into account that corrosion products within the patina on alloys

9  

become enriched in specific alloying elements that to different extent improve the solubility and barrier properties of the patina and thereby in reduced release rates.

1.4.2 Microstructure-related galvanic corrosion

Galvanic corrosion is an electrochemical process occurring when two dissimilar metals or alloys are in electrical contact in the presence of an electrolyte. The driving force for the process is the difference in electrode potential between the materials [11]. The general principle of galvanic corrosion is displayed in Fig.1.3 [93].

Figure 1.3. Schematic illustration of the principle of galvanic corrosion.

When two metals are coupled in a conductive electrolyte, the polarity and direction of electron flow can be determined from their difference in potential, indicating their galvanic behavior [11]. Table 1.1 provides examples of standard electrode potentials for different metals and predictions of relative nobilities in the galvanic series [94]. The metal of lower potential usually acts as anode and will initially preferentially undergo corrosion in the galvanic couple. However, since the electromotive series (EMF) only considers pure metals in pure water without considering the presence of surface oxides or the behavior of alloys, the actual behavior of the galvanic couple at atmospheric conditions may be very different from what is proposed by the EMF-series. As an example, Al is considered more active than Zn (-1.66 V vs. -0.76 V, see Table 1.1) in the standard EMF series. However, the presence of the native Al2O3 surface oxide on Al makes it more passive compared with Zn in a Zn-Al couple, from which follows that Zn becomes more active and protects the aluminum surface, in particularly in seawater (-0.79 V vs. -1.03 V, , see Table 1.1) [27, 95].

Galvanic corrosion cells can be established both on a macroscopic and a microscopic level. On a microstructural level microgalvanic couples can be formed between different phases or other microstructural features within an alloy, in terms of

10

coupling of two metals. The active phase or feature in the microgalvanic cell acts as a micro-anode and is preferentially corroded, which is the effect of microgalvanic corrosion [96]. The differences in nobility of an alloy between different microstructural features and phases, and sites prone to corrosion initiation can be observed by measuring differences in relative surface Volta potential [97-99] using the scanning Kelvin probe technique. The method measures differences in electrostatic potential over a surface, if any. The advancement of local probing techniques has enabled the exploration of corrosion phenomena and the interplay between local galvanic corrosion processes and microstructural features with high lateral resolution [14, 84, 100].

Table 1.1. Standard electromotive force (EMF) series of different metals relative to the standard hydrogen reference electrode (SHE) (Standard electrode potentials in H2O at 25 and 1 bar), and the corresponding galvanic series in seawater, relative to the saturated calomel reference electrode (SCE) [27, 95].

Electrode Half-cell reaction Potential /VSHE Potential /VSCE

in seawater

Galvanic series in seawater

Cu Cu+ + e- → Cu 0.52

-0.36 Noble

Active

Cu2+ + 2e- → Cu 0.34

Fe Fe3+ + 3e- → Fe -0.04

-0.61 Fe2+ + 2e- → Fe -0.45

Al Al3+ + 3e- → Al -1.66 -0.79 Zn Zn2+ +2e- → Zn -0.76 -1.03

Scarce literature findings exist regarding corrosion during early stages of atmospheric corrosion of Galfan and the relation to its microstructure. This may be connected to its complex and fine (sub-micron sized) and heterogeneous microstructure. An improved corrosion resistance of the proeutectoidic η-Zn phase compared with the eutectic colonies has been reported by Tang et al. [101]. According to Yang et al. [102] is the eutectic phase of superplastic Zn-Al alloys (4, 8, 12, 16 wt% Al) prone to preferential corrosion upon immersion in simulated acid rain.

The spangled microstructure of Galvalume has been extensively investigated and reported in the scientific literature. It consists of aluminum-rich dendritic areas and interdendritic zinc-rich regions containing silicon [35, 37]. Corrosion initiation has been reported to take place in zinc-rich interdentritic areas [35, 36, 60, 68], whereas corrosion of aluminum-rich dendritic phases preferentially takes place after longer exposure periods, or at corrosive conditions [35]. Corrosion of Galvalume results in selective release of alloy constituents [103] that typically involve selective release of the less noble element, leaving a surface enriched in the more noble element. Prevailing local surroundings will possibly be able to predict whether aluminum or zinc is the most

11  

active element. However, as corrosion products gradually evolve on the surface, selective release results in the enrichment of one or more alloying element in the alloy phase beneath the surface oxide [104-108].

Several studies in the scientific literature have shown the initial corrosion of -brass and the formation of micrometer-sized zinc-rich granules and the preferential release of zinc at atmospheric conditions [75, 84, 109]. Forslund et al. have reported initiation of local dissolution in zinc-rich areas of lower surface potential compared with the matrix on Cu20Zn in diluted NaCl solution [110]. Microgalvanic effect may also result in selective leaching of zinc from brass, dezincification, followed by weakened porous copper-rich areas [109, 111]. However, limited scientific understanding

exists on the microgalvanic corrosion between the -phase and the more zinc-rich -phase of duplex brass at initial stages of atmospheric corrosion.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12

2 Materials and methods

2.1 Materials and surface preparation

Different bare metals and alloys were investigated to assess initial atmospheric corrosion mechanisms and metal release processes upon exposure to different chloride-rich environments. Table.2.1. summarizes the materials, their corresponding bulk nominal composition and exposure conditions that have been investigated within this doctoral thesis (with reference to the corresponding Paper).

Table. 2.1. Summary of materials used within the scope of the different Papers of this thesis. (* non-published data presented in the thesis summary only.)

Name Composition (wt %) Exposure conditions Used in Papers

Galfan™ 95Zn-5Al

(0.05 Ce and La)

Wet/dry cycle II, III 

Marine field N-VDA tests

Galvalume™ 55Al-43.4Zn-1.6Si

Immersion and flow-cell tests

I Marine field N-VDA tests

Zn sheet 99.99 Zn N-VDA tests * Al sheet 99.99 Al N-VDA tests *

Cu sheet 99.98 Cu Wet/dry cycle

IV, V Marine field

Cu4Sn (Bronze)

96Cu-4Sn Wet/dry cycle

IV, V Marine field

Cu15Zn (Brass)

85Cu-15Zn Wet/dry cycle

IV, V Marine field

Cu5Al5Zn 89Cu-5Al-5Zn-1others Wet/dry cycle

IV, V Marine field

Cu40Zn (Brass)

60Cu-40Zn Wet/dry cycle *

CuZnP Zn substrate with

patterned deposited Cu islands

Wet/dry cycle *

13  

2.1.1 Materials

Commercial Zn-Al coated steel samples (Galfan and Galvalume) and bare zinc and aluminum sheets were supplied via Arcelor Mittal, France. The Zn-Al coatings were applied on steel via a hot-dip process that generated a coating thickness of 7 and 25 µm for Galfan and Galvalume, respectively [30].

Commercial Cu and Cu alloys samples (bare Cu sheet, Cu4Sn, Cu15Zn, Cu40Zn and Cu5Al5Zn) were provided via KME, Germany and Aurubis, Finland. More detailed information on their nominal bulk alloy composition (except for Cu40Zn) is provided in Table 1 in Paper IV [112].

The patterned copper-zinc samples (CuZnP) were kindly provided by Xiamen

University, China. Sputtered copper islands sized of 2020 µm (100 nm in height), with a spacing of 20 µm, were deposited onto a bare zinc substrate. The overall elemental surface coverage was theoretically calculated to 74.3% for zinc, and 25.4% for copper [113]. Details regarding the patterned copper-zinc samples are given elsewhere [113].

2.1.2 Surface preparation

Galfan samples for laboratory wet/dry cycle exposures were cut to a dimension of

11 cm for in-situ IRAS and ESEM exposures. The utmost high temperature oxide layer formed during the hot-dip process was removed for most samples by gentle polishing using a 0.25 μm diamond paste. All samples were cleaned ultrasonically in analytical grade ethanol for 10 min, dried by cold nitrogen gas and stored in a desiccator overnight.

Galvalume samples for laboratory immersion tests were cut to a dimension of 40.5 cm. The reverse side and all edges of the samples were sealed with a transparent non-metal containing polish three times. All flow-cell tests were performed using an

exposed surface area of 10.5 cm.

Bare Zn and Al sheet, Galfan and Galvalume samples for the N-VDA tests were

cut to a dimension of 25 cm. The Zn and Al samples were mechanically wet ground down to 1200 grit, and the Galfan and Galvalume samples were gently polished using a 0.25 μm diamond paste.

Bare Cu sheet and the Cu-based alloy samples were cut to a dimension of 11 cm

for in-situ IRAS exposures and 25 cm for climatic chamber exposures. Each sample was mechanically wet ground down to 2400 grit and polished with diamond paste down to 1 or 0.25 μm. All freshly polished samples were, prior to analyses,

14

ultrasonically cleaned in analytical grade ethanol for 10 min and dried by cold nitrogen gas before stored in a desiccator overnight.

No polishing was possible for the patterned copper-zinc samples. After ultrasonic cleaning in ethanol, the samples were immersed in 5 wt% amidosulfonic acid (ASA, H3NSO3) for approximately one second to chemically remove the outermost surface oxide. The samples were dried by cold nitrogen gas and stored in a desiccator overnight.

A synthetic nantokite layer (CuCl) was grown on the bare Cu surface according to a procedure described elsewhere [114, 115]. After the same surface preparation as above, bare Cu samples were immersed for 1 h in a saturated CuCl2·2H2O solution (>78g of CuCl2·2H2O per 100mL of deionized water). The samples were then rinsed with deionized water followed by immediate drying using cold nitrogen gas and stored in a desiccator overnight.

2.2 Exposure conditions

2.2.1 Laboratory wet/dry cycle exposure (Papers II, V)

Parallel in-situ experiments on Galfan samples were carried out by means of ESEM (environmental - scanning electron microscopy) and in-situ IRAS (infrared reflection absorption spectroscopy) upon exposure to cyclic exposure conditions, schematically displayed in Fig.2.1 (NaCl pre-deposition (0 or 4/0.1 μg NaCl/cm2), the first cycle 4 h (RH 90%) and 2 h (RH 0%), the second cycle 16 h (RH 90%) and 2 h (RH 0%)). These cycles were repeated several times. Parallel exposures were conducted at 70% RH. SEM imaging of surface features and IRAS spectra of corrosion product formation were generated during the dry cycles.

Parallel experiments on bare Cu sheet and Cu alloys were conducted by means of in-situ IRAS and climatic chamber exposures. Wet/dry cycle experiments were carried out following the same cyclic exposure conditions as shown in Fig.2.1. All climatic chamber exposures were conducted in a WEISS WK1000 climatic chamber. Different samples (Cu sheet, Cu4Sn, Cu15Zn, Cu40Zn and Cu5Al5Zn) were attached on Plexiglas fixtures and exposed at 45˚ from the horizontal. All samples were exposed in parallel, withdrawn after 1, 2, 6 and 14 cycles (corresponding to 6 h, 1, 3, 7 days, respectively) and analysed ex-situ.

NaCl (in a saturated 99.5% ethanol solution) was applied prior to exposure using a transfer pipette providing a relatively homogeneous distribution of NaCl crystals upon ethanol evaporation. The amount of deposited NaCl was weighed (Mettler Toledo Excellence microbalance) and normalized to the geometric surface area.

15  

Figure 2.1. Schematic illustration of wet/dry cycles during the laboratory exposures (Paper II).

2.2.2 Laboratory immersion and flow-cell tests (Paper I)

Immersion experiments and flow-cell tests were performed on Galvalume samples in synthetic rain water (pH 4.4.) [116] of modified chloride content (0.01 and 0.3 mM) and in synthetic seawater (pH 8.1) [117]. Chemical composition of each media is given in Tables I and II in Paper I. These acid- and chloride containing media were selected to enable comparison with atmospheric field data, and of particular relevance for automotive applications.

For the immersion tests, the samples were attached at the bottom of polypropylene

vessels (2965 cm). Triplicate samples and one blank (no sample) were exposed in parallel. The vessels were gently agitated using a bi-linear shaking table, moving the media over the sample surface. Similar exposure time periods and specific solution volumes were selected for the immersion tests as investigated in the flow-cell test.

For the flow-cell tests, AESEC (atomic emission spectroelectrochemistry) was employed for real-time measurements of released metals as a function of time downstream from an electrochemical flow cell at open circuit potential (OCP) and at an applied anodic potential (-400 mV vs. NHE). A sample area of 0.5 cm2 was exposed to the continuously renewed flowing test media for 3000 s (50 min). These investigations were performed at Ecole Nationale Supérieure de Chimie de Paris, France.

2.2.3 Long-term field exposure (Papers I, II, IV, V)

Non-treated Galfan and Galvalume samples were exposed at unsheltered conditions at the marine site of Brest (Sainte Anne du Portzic), France for 5 years (SO2 < 3 µg/m3, precipitation: 800-1000 mm/y, chlorides deposition rates: 7-8757

16

mg/m2 d). The evolution of corrosion products was studied after 2, 4, 12, 26 and 52 weeks and 5 years. All surfaces were exposed 5-10 m from the waterline, and exposed at an inclination of 45º from the horizontal, facing south, according to the ISO 17752 standard for metal runoff rate measurements [118].

Copper sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn were exposed at the same site and exposure conditions (45° from the horizontal, facing south) for 3, 6 months, and 1, 2 and 3 years (starting from Nov, 2009). The surfaces were exposed at four sites of increasing distance from the coastal line (site 1 - Military harbour: < 5m; site 2 - St. Anne: 20-30 m; site 3 - St Pierre: 1.5 km; site 4 - Langonnet: 40 km). Corrosion product formation studies of this thesis include measurements at the site closest to the coastal line, i.e. site 1 with measured deposition rates of chlorides varying between 300-1500 mg/ m2 d (large seasonal differences).

2.2.4 Accelerated N-VDA test (non-published data)

Bare Zn and Al sheet, Galfan and Galvalume samples were exposed in a WEISS WK1000 climatic chamber, following the cycles of accelerated N-VDA tests (New revised VDA Corrosion Test Method, Standard SEP 1850 VDA 621-415 B). The N-VDA test is a new accelerated corrosion test developed by the German Association of automotive industry (VDA) and the German Steel Association (VDEh) [119]. The test is schematically illustrated in Fig. 2.2. Duplicate samples of each material, attached on Plexiglas fixtures and exposed at 45˚ from the horizontal, were withdrawn after each cycle (day) of the N-VDA test during one week and analysed ex-situ. The N-VDA accelerated test was conducted at VoestAlpine, Austria.

Figure 2.2. Schematic illustration of the accelerated N-VDA test cycles. (cited from appendix 1.1 in ref. [120])

17  

3 Analytical techniques

Highly surface sensitive and non-destructive analytical techniques were employed for in-depth analysis and time-dependent measurements of changes in microstructure, corrosion product initiation and evolution, surface composition and release of metals. Table.3.1. summarizes the different techniques and main information provided.

Table 3.1. Compilation of analytical techniques (and abreviations) used within the scope of the different Papers of this thesis.

Abbreviation Name Main information Used in Papers

SEM/EDS (-mapping)

Scanning electron microscopy/ Energy

dispersive x-ray analysis

Corrosion product morphology/ elemental,

compositional distribution I-V

ESEM/EDS Environmental scanning

electron microscopy/Energy dispersive x-ray analysis

In-situ corrosion product morphology/ elemental

distribution II

FIB-SEM Focused ion beam-scanning

electron microscopy Milling of trenches to visualize

analyzed surface area III

(In-situ) IRAS

(In-situ) Infrared reflection absorption spectroscopy

(In-situ) Functional groups I, II, IV, V

CRM Confocal Raman micro-

spectroscopy Lateral distribution of

functional groups I-V

GIXRD Grazing incidence x-ray diffraction

Crystalline phases I, II, IV, V

SKPFM Scanning Kelvin probe

force microscopy Surface topography and Volta

potential mapping III

XPS/AES X-ray photoelectron spectroscopy/ Auger

Electron spectroscopy

Elemental composition and chemical state information

I, IV

GDOES Glow discharge optical emission spectroscopy

Elemental depth distribution IV

OM Optical microscopy /

Stereomicroscopy Surface morphology IV

AAS Atomic absorption

spectroscopy Release of metals I, IV

AESEC Atomic emission

spectroelectrochemistry In-situ release of metals I

18

3.1 Scanning electron microscopy with x-ray microanalysis (SEM/EDS)

Scanning electron microscopy (SEM) is a kind of electron microscope that generates images by scanning the sample by means of a focused beam of electrons. When an electron beam strikes onto a material, accelerated electrons can undergo elastic and inelastic scattering. This results in various signal emissions that depend on the incident energy of the electron beam [121, 122]. Typical signals include secondary electrons (SE), backscattered electrons (BSE), Auger electrons and x-rays. The electron scattering and photon- and x-ray-production develops in a volume, the electron interaction volume, within the material where interactions occur by impinging accelerated electrons, schematically illustrated in Fig.3.1 [123]. The spatial resolution depends on the generated interaction volume of the emitted energy. Several interactions of these signals are utilized for imaging, quantitative and semi-quantitative investigations of materials.

Figure 3.1. Interaction volumes for different electron-sample interactions [123].

Secondary electrons and backscattered electrons for sample imaging, and characteristic x-rays are the most commonly used signals for compositional analysis.

19  

Secondary electrons that are emitted from the atoms occupying the top surface are useful to gain morphological and topographical information. Backscattered electrons are incident electrons which are scattered by the atoms in the solid. Therefore the contrast of the image indicates differences in average elemental composition. A SEM image is generated when the electron beam is scanned in a raster scan pattern and the beam position is combined with the detected signal [122].

The interaction of the electron beam with atoms in the sample contributes to the shell transitions, and the energy difference between the two shells may be released in the form of an x-ray. Energy dispersive x-ray spectroscopy (EDS) allows elemental compositional analyses of the surface, facilitated by the detection and measurement of the energy of the x-rays that are characteristic of the atomic structure of the element from which they were emitted [123]. EDS provides rapid qualitative or quantitative elemental information and is able to generate elemental maps and line profiles.

High resolution images were generated by means of a FEG-SEM instrument, a LEO 1530 field emission gun SEM with a Gemini column, upgraded to a Zeiss Supra 55 column (equivalent). An accelerated voltage of 15kV and an aperture size of 60 µm were used to record the images. Elemental distribution information was obtained by means of an EDS X-Max SDD (silicon drift detector) 50 mm2 detector from Oxford Instruments.

A table-top SEM instrument (Hitachi TM-1000) equipped with a Hitachi EDS facility was also utilized to analyse the surface morphology and provide compositional information, using an accelerating voltage of 15 kV.

3.1.1 Environmental - SEM (ESEM)

An environmental scanning electron microscopy (ESEM) is a SEM that permits examination at near atmospheric pressures unlike conventional SEM, which operates in vacuum. The ESEM allows high resolution imaging of wet or dry, insulating or conducting samples, and allows in-situ analysis at specific environments generated within the chamber. The ESEM system utilizes differential pump systems and specialized electron detectors that permit electron beam transfer from the high vacuum of the column to the high pressure in the chamber [121]. A state of art ESEM can generate electron micrographs at pressures as high as 50 Torr, and at temperatures as high as 1000 °C, which makes it a unique instrument to image realistic surface conditions [124].

In-situ analyses were performed using a FEI-XL 30 Series ESEM instrument. The acquired relative humidity (RH) was obtained by controlling the temperature and pressure by means of a thermoelectric stage controller. Images were generated using a gaseous secondary electron (GSE) detector in the environmental mode.

20

Ex-situ measurements were also carried out by means of the normal SE/ BSE detector. All images (75% SE and 25 % BSE) were collected using an accelerating voltage of 15, 20 or 30 kV at high vacuum conditions. Elemental analysis and mapping were performed using an EDAX Phoenix EDS system with an ultra-thin window Si-Li detector.

3.1.2 Focused ion beam - SEM (FIB-SEM)

A focused ion beam-scanning electron microscope (FIB-SEM) is a SEM equipped with a FIB system. A finely focused beam of usually gallium ions can be operated at low beam currents for imaging or high beam currents for sputtering or milling in local areas [125]. In a FIB-SEM system is both the electron column and the ion column embedded in the same chamber and focused at the same spot at the sample surface [126].

FIB milling was accomplished by using a Quanta 3D instrument from FEI with a Ga+ ion source, which operated at an acceleration voltage of 30 kV and ion probe

currents of 30 nA. A selected 5050 μm area was marked as the area of interest for

surface analysis by framing two trenches, each sized 5015μm and with a depth of 5 μm.

3.2 Infrared reflection absorption spectroscopy (IRAS)

Infrared spectroscopy is a vibrational technique used to identify the presence of functional surface groups based on their specific frequencies that are characteristic of their structures. The technique can be used for several types of samples including solids, liquids and gases. IR spectroscopy mostly investigates the interaction between an external electromagnetic field and a dipole from the vibration of a molecule [6]. When irradiated with infrared light (photons), a sample can absorb the incident radiation. This excites molecules into a higher-energy vibrational state, resulting in a change in dipole moment [10, 127]. The process takes place when the photon energy corresponds to the energy difference between two vibrational states.

The main selection rule in IR spectroscopy is required for the molecular vibration being infrared active, i.e. during a vibration should the dipole moment (μ) change with respect to the normal coordinate (Q), normally described as:

0

Fourier transform infrared (FTIR) spectrometry is a commercial technique that allows infrared spectra to be recorded using an interferometer for which the intensity of each wavelength of light at a different audio frequency could be modulated [127].

21  

Infrared reflection absorption spectroscopy (IRAS) is an IR technique based on reflection, suitable for investigations of thin films on e.g. metal surfaces. Measurements of a single external reflection at near-grazing incidence angle enable adsorbed monolayers on metal surfaces to be detected. Only those dipole transition moments of adsorbed molecules perpendicular to the surface are observed by means of p-polarised light [6]. IRAS provides information on the molecular vibrations in the surface film from which surface species can be identified [128]. This information is collected as the peak absorbance, presented in the unit of (-log (R/R0)), where R is the reflectance of the exposed sample surface and R0 the reflectance of the non-exposed sample, used as background [55].

IRAS analyses were performed using a commercial Digilab 4.0 Pro FTIR spectrometer, equipped with a mercury cadmium telluride (MCT) detector. The p-polarized infrared beam strikes the sample surface at a grazing angle of around 78°, are reflected into the nitrogen cooled MCT detector via the CdTe window of an external chamber and recorded. The set-up is schematically displayed in Fig.3.2.

IRAS spectra can be difficult to obtain for conditions when the surface patina is too rough or thick. FTIR measurements were then employed using a Thermo Nicolet 6700 FTIR spectrometer equipped with a Deuterated triglycine sulfate (DTGS) detector. FTIR spectra were generated from powder scraped from the surface.

Figure 3.2. The experimental chamber used for IRAS experiments. (Cited with permission from Dr. Aastrup from Fig. 2. in ref. [55])

In-situ IRAS analyses of samples with or without pre-deposited NaCl were performed in a chamber inside the Digilab 4.0 Pro FTIR spectrometer with humidified air following the wet/dry cycles, schematically displayed in Fig.2.1. Controlled humidified conditions were obtained by mixing dry and wet pre-cleaned compressed

22

air of reduced CO2 (lower than 20 ppm), for exposures of Galfan samples in the experimental chamber, shown in Fig.3.2. In order to obtain ambient CO2 concentrations (350 ppm) for exposure conditions of Cu sheet and the Cu-based alloys, a small flow of air with 1.17% CO2 from a CO2 cylinder was added into the humidity chamber. In-situ IRAS spectra of formed corrosion products were generated during the dry cycles. All spectra presented in this thesis were collected acquiring 1024 scans with a resolution of 4 cm-1.

3.3 Confocal Raman micro-spectroscopy (CRM)

Raman spectroscopy is another technique used to obtain vibrational spectra, similar to infrared spectroscopy, however with different mechanisms and selection rules. The Raman effect, a form of inelastic scattering, means that the molecules are excited from their ground state to a virtual energy state by an incident photon with an energy significantly larger than the vibrational transition. A new photon is scattered from the virtual state when the molecule relaxes and it returns to a different rotational or vibrational state [127]. The energy difference between the original and new states results in a shift in frequency of the scattered photon away from the excitation wavelength.

Figure.3.3. IR absorption, Rayleigh, stokes and anti-Stokes Raman scattering processes. υ0 and υ1 represent the vibrational ground state and the first excited state, respectively, ΔEv is the energy difference between the vibrational states, hv represents the energy of incident photons, and Ev the energy of scattered photons [128].

There are three kinds of scattering phenomena: Rayleigh, Stoke and anti-Stoke scattering, schematically illustrated in Fig.3.3, and compared with normal infrared adsorption. In Rayleigh scattering are photons elastically scattered, which results in

23  

the same energy as the incident photons [128]. A Stoke scattering process means that the scattered photon shifts to a lower frequency when the final vibrational state is higher in energy than the initial state. Anti-Stoke scattering occurs when the final vibrational state is lower in energy than the initial state, and the scattered photon shifts to a higher frequency.

The absorption band position in IR spectroscopy depends on the absolute frequency of the incoming photons whereas the band positions in Raman spectroscopy rather are related to the difference in frequency between excited and scattered photons [128]. Different from infrared spectroscopy where the selection rule is determined by the dipole moment during a vibration, the selection criteria depend on the polarizability (α):

0

Confocal Raman micro-spectroscopy (CRM) is Raman spectroscopy combined with confocal microscope. When the laser light strikes the sample surface through a microscope setup, the excited laser light is filtered to reach the detection system in the focal plane. Only signals from the image plane are collected by the confocal optical system, contributing to a high lateral resolution [129]. Chemical analysis can be conducted by means of CRM if used combined with parallel vibrational information [6].

CRM analyses were performed by utilizing a WITec alpha 300 system, equipped with a laser source of wavelength 532 nm. A Nikon objective 100, Nikon NA0.9 NGC, and a pinhole of 100 µm diameter were used for the measurements. Raman spectra were generated within the scanned area with lateral and vertical resolutions of approximately 300 nm and 2 μm, respectively, and an integration time of 50 ms for each Raman spectrum.

3.4 Grazing incidence x-ray diffraction (GIXRD)

X-ray diffraction (XRD) is an analytical technique able to identify crystalline phases and to provide information on unit cell dimensions of a material. A monochromatic x-ray irradiates a crystalline material that produces diffracted x-rays at various angles, according to Bragg’ law [130]. This law connects the wavelength of the primary beam (λ), the angle of diffraction (θ) and the lattice spacing in a crystalline sample (d):

nλ = 2d sinθ

24

The crystalline phases are identified either directly via observed 2θ positions, or recalculated into d-spacings and compared with standard reference patterns [127].

Grazing incidence x-ray diffraction (GIXRD) uses small incident angles for the incoming x-rays, which significantly increase the path travelled by the x-rays. The small incidence angle makes diffraction surface sensitive, and obtains information primarily from the thin film or surface layers, unlike normal XRD which typically collects bulk information [131].

GIXRD analyses were carried out by means of an X’pert PRO PANALYTICAL system, equipped with an x-ray mirror (CuKα or MoKα radiation) and a 0.27° parallel

plate collimator on the diffracted side. The scanned area was 11 cm at a grazing angle of 88º versus the surface.

3.5 Scanning Kelvin probe force microscopy (SKPFM)

Scanning Kelvin probe force microscopy (SKPFM) is a variant type of atomic force microscope (AFM), widespread used to study the electrical properties of metal surfaces. SKPFM measures the work function difference at atomic or molecular scales, determined by electrostatic forces between a conductive AFM tip and the sample. The work function represents the difference between the Fermi level and the vacuum level for each material. If two materials are brought in contact, a net electric current flows between them until they reach the same Fermi level [132]. A voltage is applied between the tip, which acts as a reference electrode that forms a capacitor with the surface, and the sample in order to measure the difference in work function, denoted the Volta potential or contact potential difference. The normal topographic scan methods in AFM can be used independently, thereby the topography and Volta potential mapping can be determined simultaneously on the same surface area [133, 134].

SKPFM is a local probing technique useful for characterization of the corrosion processes associated with local inhomogeneities on metal surfaces [97]. To some extent, the relative nobility of microstructural features or phases of alloys, which is significantly related to microgalvanic interactions, can be suggested based on the Volta potential findings [98, 99]. However, interpretations must be made with caution as the technique measures absolute surface potentials on polished surfaces, which is possibly different from the real surface characteristics.

SKPFM measurements were employed using an Agilent 5500 scanning probe microscope equipped with a MAC III unit in single-pass mode, and a Nanoscope IV AFM with facilities for Volta potential measurements in two-pass mode. The probe was PtIr-coated Si, supplied by Bruker, with a nominal spring constant of 1-5 N/m and a resonance frequency of 60-100 kHz.

25  

3.6 X-ray photoelectron spectroscopy/ Auger electron spectroscopy (XPS/AES)

X-ray photoelectron spectroscopy (XPS) is a powerful technique used for elemental analysis and interpretation of oxidation states of the outmost surface layers (5-10 nm). By irradiating a surface with a beam of x-rays, electrons with binding energies lower than the excitation beam can be ejected. XPS spectra are obtained while measuring the kinetic energy and analyzing the number of escaped electrons from the material simultaneously [135].

Auger electron spectroscopy (AES) is another surface sensitive technique used for analysis of elements and oxidation states of the outmost surface layers. When a primary beam strikes on a sample surface, the Auger process results in the emission of secondary electrons from excited atoms. Information for elemental identification and quantification is obtained via of the kinetic energy and intensity of generated Auger peaks [127].

XPS measurements were performed with a Kratos AXIS UltraDLD x-ray photoelectron spectrometer (Kratos Analytical, Manchester, UK) system combined with an Auger unit. A monochromatic Al x-ray source (1486.6 eV) operated at 300 W (15 kV/20 mA) was used during the measurements. The analysis area was

approximately 1 mm2 (most of the signal from a 700300 µm surface area). Wide spectra and detailed high resolution spectra (pass energy: 20 eV) were collected. AES mapping was generated using an acceleration voltage of 10 kV and a beam current of 300 nA.

3.7 Glow discharge optical emission spectroscopy (GDOES)

Glow discharge optical emission spectrometry (GDOES) is widely used for elemental analysis and in-depth profiles on solid conductive materials. The sample is etched by cathodic sputtering in a glow discharge source. Sputtered atoms are excited by an argon plasma into which characteristic photons are emitted and collected by photomultipliers [136].

GDOES analyses were made using a Leco GDS 850 instrument. By using Ar plasma at a potential of 700 V and a current of 20 mA, a circular area with a diameter of 4 mm was continuously sputtered. The Ar pressure varied between 6 and 7.5 Torr.

26

3.8 Optical microscopy (OM)/ Stereomicroscopy

Optical microscopy is a fundamental technique for image magnification of samples by using visible light and a system of lenses. With help of normal light-sensitive cameras, the generated magnified images can be captured [121]. Stereomicroscopy uses light reflected from the sample surface instead of transmitted light which enables a three-dimensional visualization of the surface.

Optical microscopy imaging was employed to investigate surface morphological changes by using a Leica DM 2700M microscope that combines high-quality Leica optics with state-of-the-art universal white light LED illumination.

Stereomicroscopy imaging was performed to document the surface appearance by means of a M205C Stereo microscope with a Leica DFC 290 video camera, using a D65 reference light source applied at 10° at a magnification of 40×.

3.9 Atomic absorption spectroscopy (AAS)

Atomic absorption spectroscopy (AAS) is an analytical technique that quantitatively determines elements in solution. By absorbing an optical radiation (light) of a given wavelength, the electrons of the free atoms in the gaseous state in the atomizer can be transited to higher energy levels [137]. This defined quantity of energy (wavelength) is measured as photons of light transmitted by the sample, which is specific to a particular electron transition in a particular element. By comparing this wavelength to the wavelengths which originally passed through the sample, the total concentration of a specific element can be obtained [127].

Standards of known concentration for the element of interest are necessary to establish the connection between the measured absorbance and the element concentration of interest, therefore relying on the Beer-Lambert Law. This law connects the absorbance (A) proportional to the wavelength-dependent molar absorptivity coefficient (ε), the path length of the sample (b) and the concentration of the compound in solution (c):

A = εbc

All metal runoff measurements were conducted by means of AAS using a Perkin Elmer AAnalyst 800 instrument. Analyses of released zinc were performed using the flame mode AAS, whereas released concentrations of aluminum were determined with graphite furnace AAS (GF-AAS) due to low concentrations (sub-µg/L). Prior to analysis, the runoff water was acidified to a pH less than 2 to ensure complete dissolution of potentially formed complexes. Three replicate readings were made for

27  

each sample and control samples (every 10th sample), successively measured during the analysis.

3.10 Atomic emission spectroelectrochemistry (AESEC)

Atomic emission spectroelectrochemistry (AESEC) is a novel instrument that combines inductively coupled plasma optical emission spectrometry (ICP-OES) [107] downstream an electrochemical flow cell. Real-time measurements of released metals as a function of time can be achieved, both at open circuit potential (OCP) conditions and at applied potentials [138]. Details regarding AESEC is given elsewhere [107].

Emission signals from the ICP-OES and the open circuit potential were constantly measured and recorded after averaging over a user defined integration period. Release rates of metals were calculated based on the measured metal concentrations, Cm, times the flow rate per surface area. Metal concentrations were determined using a polychromator of Ultima 2C Horiba JobinYvon ICP-OES spectrometer (focal distance 50 cm), using wavelength 213.85 nm (Zn) and 167.08 nm (Al). All electrochemical experiments were carried out using a potentiostat (EG&G Princeton Applied Electronics M273A) functioning in potentiostatic mode in connection with ICP-OES. All AESEC measurements were performed by colleagues at Ecole Nationale Supérieure de Chimie de Paris, France.

 

28

4 Influence of microstructure on corrosion initiation

4.1 The eutectic structure of Galfan consists of an η-Zn matrix and β-Al lamellas and rods of lower surface nobility compared with the matrix, and are separated by β-Al grain boundaries in which ZnO and Al2O3 preferentially form. (Paper III)

The selected eutectic area was easily identified via FIB-milled trenches on a diamond polished Galfan surface, displayed in Fig.4.1a. Surface analyses were carried

out on the very same surface area (3232 μm) by utilizing SEM/EDS-mapping for morphology and elemental composition analyses, SKPFM for surface topogragphy and Volta potential, and CRM to obtain information of lateral differences in surface oxide composition.

Figure 4.1. SEM images of trenches milled by FIB on a diamond polished Galfan surface

(a), the selected surface area (3232 µm) between the trenches (b), and the corresponding EDS-mapping images of elemental distribution of zinc (c) and aluminum (d).(Paper III)

29  

The microstructure of the selected surface area between the trenches is shown in Fig.4.1b, and the corresponding elemental distribution, as obtained by EDS-mapping, in Fig.4.1c (zinc) and Fig.4.1d (aluminum), respectively. The structure of the eutectic area consists of a zinc-rich η-Zn matrix (<5wt-% Al/(Al+Zn)) and aluminum-rich β-Al lamellas or rods (Al/(Al+Zn) >5 wt-%), separated by β-Al grain boundaries (Al/(Al+Zn) >10 wt-%). It has been reported that the grain boundaries form during the coarsening process of the β-Al phase precipitation in the η-Zn phase, which is more pronounced at higher temperature and annealing time upon production [34].

Complementary measurements by SKPFM were conducted on the same surface area. The AFM-image (Fig.4.2a) displays a relatively flat topography whereas some parts of the grain boundaries show a lower height than the surrounding matrix. The corresponding SKPFM-image (Fig.4.2b) displays that the β-Al areas exhibit lower Volta potential than the matrix, indicating a lower relative nobility of the β-Al lamellas/ rods. They are therefore possibly more susceptible to corrosion initiation compared with surrounding η-Zn areas [97]. The β-Al grain boundaries exhibit a strongly varying Volta potential, alternatively, lower Volta potential at some areas and higher at other areas than the surrounding. These variations seem to some degree related to variations in topography along the grain boundaries.

Figure 4.2. AFM-based topography (a) and Volta potential mapping (b) obtained with

SKPFM of the same Galfan surface area (3232 µm) as in Fig.3.1. (Paper III)

Evident residues of thicker films of Al2O3 and ZnO, especially along the β-Al grain boundaries [139], were locally observed by means of CRM even after diamond-polishing. This is most likely a result of oxides formed during the high temperature manufacturing process.

A lower Volta potential of the β-Al lamellas or rods compared with the surrounding η-Zn matrix is expected from a thermodynamic perspective, and indicates that the origin of the Volta potential variation predominantly is due to variations in bulk

30

composition between different phases in the microstructure. It suggests furthermore that layers of Al2O3 and ZnO, expected to form at the surface besides in the grain boundaries, are too thin to influence the Volta potential difference as determined by SKPFM. This conclusion is contradictory to findings for Galvalume coatings (Zn-55wt% Al) of high aluminum content, where the continuous and thick aluminum oxide significantly enhances the Volta potential, in particularly in areas of aluminum-rich dendrites [14].

4.2 Corrosion initiation observed for Galfan in the zinc-richer η-Zn phase adjacent to the less zinc-rich β-Al phase. Both carbonate and chloride-containing phases are formed in humidified air and in the presence of NaCl. (Paper II)

The gradual deliquescence of pre-deposited NaCl particles (4 μg/cm2) on Galfan were observed by ESEM at in-situ conditions. Subsequent wet/dry cyclic exposures of Galfan, pre-deposited with 4 μg/cm2 NaCl, were performed, and the corrosion products were analyzed by means of SEM and in-situ IRAS. Generated findings on the corrosion product formation were compared with findings for Galfan exposed to short and long-term outdoor marine exposure conditions.

Figure 4.3. ESEM images of a Galfan surface pre-deposited with 4 μg/cm2 NaCl particles and exposed to increasing relative humidities (RH). (a) 0% RH; (b) 40% RH; (c) 70% RH and (d) 90% RH. (Paper II)

31  

Figure 4.3a shows the distribution of individual micrometer-sized NaCl crystals pre-deposited on a Galfan surface at high vacuum conditions. Figure 4.3b, obtained at 40% RH, displays how the NaCl crystals seem to slightly change in shape and in some cases also increase in size. At 70% RH, close to the point of deliquescence of NaCl, the crystals completely turned into droplets, Fig.4.3c. At 90%RH, the NaCl droplets further spread, probably due to secondary spreading effects [12], and gradually covered most of the surface, Fig.4.3d.

Figure 4.4a displays the same Galfan surface as shown in Fig.4.3 after exposure to one wet/dry cycle at 90% RH. Figure 4.4b reveals a selected area in Fig.4.4a at higher magnification and shows a high frequency of small granular corrosion products, primarily formed in the η-Zn phase between areas of the β-Al phase. After the exposure to two wet/dry cycles, Fig.4.4c, corrosion products of relatively uniform thickness completely covered the surface, whereas clusters of signficantly thicker corrosion products formed in some areas, Fig.4.4d. EDS analysis revealed Zn, Al, O and Cl to be the main elements of these clusters.

Figure 4.4. SEM images of a Galfan surface pre-deposited with 4 μg/cm2 NaCl particles and exposed to wet/dry cycles at 90% RH. (a) the same surface as in Fig.4.3 after 1 wet/dry cycle exposure at 90% RH; (b) a magnification of a specific corroded area observed in (a); (c) corrosion products formed after 2 wet/dry cycles, and (d) a chloride-rich corrosion product cluster after 2 wet/dry cycles. (Paper II)

32

The corresponding in-situ IRAS spectra were obtained after one (lower curve) and two (upper curve) wet/dry cycles, displayed in Fig.4.5. Several peaks were observed upon exposure with pre-deposited NaCl, including the peaks assigned to ZnAl2O4 at 560-670 cm-1, the peak tentatively assigned to OH- in Zn(OH)2 or Al(OH)3 at 1143 cm-1 [18, 140, 141] and the major peak at 1370 cm-1 identified as CO3

2- in Zn6Al2(OH)16CO3·4H2O [66]. Three major peaks in the range from 700 to 1100 cm-1 were assigned for Zn -O-H and Al-O-H vibrations, possible evidence for the formation of a layered double hydroxide such as Zn2Al(OH)6Cl·2H2O [142] and/or Zn5(OH)8Cl2·H2O [18]. However, it was not possible to distinguish these phases from each other.

Figure 4.5. In-situ IRAS spectra obtained on Galfan after exposure for 1(a) and 2 (b) wet/dry cycles at 90% RH pre-deposited with 4 μg/cm2 NaCl. (Paper II)

The rounded shape of most NaCl droplets observed with ESEM under in-situ conditions, Fig.4.3d, indicates that the wetting process was the same along the whole Galfan surface, regardless of the microstructural features. In addition, when analyzing the same corroded surface with SEM under vacuum conditions, Fig.4.4a, the circular features that stand for the droplets in Fig.4.3d seem to cover parts of the Galfan surface, independent of the grains and phases beneath. However, when studying a corroded area at higher magnification, Fig.4.4b, visible granular corrosion products of larger size and occurrence were primarily observed in the η-Zn phase between areas of the β-Al phase. Due to its slightly higher Al-content, the β-Al phase could be expected to be more susceptible for corrosion initiation. However, oxide formation seems to inverse these conditions so that the β-Al phase acts more passive than η-Zn phase during the exposures. Local clusters of corrosion products shown in Fig.4.4d, were by

33  

IRAS shown to be either composed of Zn2Al(OH)6Cl·2H2O and/or Zn5(OH)8Cl2·H2O. The IRAS-spectra furthermore revealed the formation of Zn6Al2(OH)16CO3·4H2O, another layered double hydroxide predominantly formed upon exposure without pre-deposited NaCl [143].

Similar corrosion product morphologies were observed by SEM on Galfan surfaces after short term (2 and 4 weeks) marine exposures, Fig.4.6, as after the laboratory exposures, Fig.4.4. Corrosion was preferentially initiated in the η-Zn phase and developed into clusters of corrosion products, Fig.4.6c. These cluster-like corrosion products mainly contained Zn, Al, O and C, as evidenced by EDS-analysis, and may be attributed to ZnO or Zn6Al2(OH)16CO3·4H2O. Different sheet-like corrosion products were observed after two weeks, Fig.4.6d, and after four weeks, Fig.4.6d. These corrosion products contained significant amounts of Cl, beside Al, Zn and O, and consisted possibly of chloride-rich corrosion products, such as NaZn4Cl(OH)6SO4·6H2O, Zn2Al(OH)6Cl·2H2O and/or Zn5(OH)8Cl2·H2O [144].

Figure 4.6. (a) SEM image of corrosion products on Galfan exposed for 2 weeks at the marine site of Brest, France; (b) a selected area of (a) at higher magnification; (c) carbonate-rich corrosion products after 2 weeks, and (d) chloride-rich corrosion products after 4 weeks. (Paper II)

34

4.3 Selective zinc release and corrosion initiation in the zinc-rich phase observed for Galvalume in chloride containing media. Long-term correlation observed between the released zinc fraction and the surface coverage of zinc and aluminum-rich corrosion products. (Paper I)

Short-term metal release from Galvalume and the formation of corrosion products were investigated when exposed to synthetic rainwater and seawater by means of two different set-ups, immersion tests and flow-cell tests at OCP conditions. The relevance of generated results obtained at given laboratory conditions were validated with findings from long-term atmospheric outdoor exposures.

Fig.4.7a presents release rates of zinc and aluminum from Galvalume after immersion tests determined by means of AAS. Zinc was primarily released compared with aluminum, independent of test media. Observed differences in zinc release rates were small up to 1800s, but were significant (factor of two) after 3000 s of exposure in seawater, containing significantly more chloride but of a more alkaline pH compared with rainwater. These results indicate that the protective ability of aluminum oxide was locally destroyed by chloride ions after a certain time period, a process that took place in a faster rate in seawater compared with rainwater, respectively. However, observed release rates of aluminum were for most conditions very low, being below or close to the limit of detection. Similar to findings of the immersion tests were significantly higher release rates of zinc compared to aluminum observed for all media in the flow cell test at OCP conditions, as illustrated in Fig.4.7b for seawater. Preferential release of zinc was evident in both synthetic rainwater and in seawater.

Figure 4.7. Release rates of zinc and aluminum from Galvalume during immersion tests (a) in rainwater and seawater, respectively, and during flow-cell tests (b) in seawater. (Paper I)

35  

Corrosion products of relatively similar surface morphology were formed on Galvalume upon immersion and flow cell tests at OCP conditions, Fig.4.8. Consistent with literature findings [6-8], were corrosion products preferentially formed in zinc-rich interdendritic areas, an effect particularly pronounced upon seawater exposure. Minor morphological differences were observed between samples exposed to the rainwater media of different chloride content, suggesting only slight and locally occurring corrosion. Al2O3 was identified as the only crystalline corrosion product by means of GIXRD at all exposure conditions, an oxide that, according to literature findings [14], is favoured by the presence of chlorides.

Figure 4.8. SEM images of Galvalume during immersion tests for 3000 s in artificial rain water ((a) 0.01 mM and (b) 0.3 mM Cl- ), and (c) in artificial sea water (560 mM Cl-) and during flow-cell tests in artificial sea water (d). The average mass ratio Al/ (Al+Zn) refers to XPS compositional measurements of two separate areas (each sized 0.4 mm2). (Paper I)

The dominance of a thin layer of Al2O3 was clearly demonstrated by means of XPS,

showing a Al/(Al+Zn) mass ratio varying between 0.85 and 0.99 in the outermost surface layer of non-exposed bare Galvalume. Al2O3 forms at high temperatures during alloy production and was present on both zinc-rich and on aluminum-rich areas. Exposure in artificial rainwater during the immersion tests resulted in relatively small changes of the Al/(Al+Zn) surface mass ratio at 0.01 mM Cl- (mean ratio: 0.96)

36

and 0.3 mM Cl- (mean ratio: 0.84). Exposures in seawater resulted on the other hand in significant changes in surface distribution between aluminum-rich and zinc-rich corrosion products, elucidated by a reduced mean Al/(Al+Zn) surface mass fraction to 0.63 (0.58-0.68) after the immersion test, and to 0.23 after the flow-cell test. Furthermore, an IRAS peak was identified at around 1467 cm-1, assigned to CO3

2-, possibly associated with Zn6Al2(OH)16CO3·4H2O and/or Zn5(OH)6(CO3)2.

Findings of the relative surface distribution of aluminum- and zinc-rich corrosion products were in agreement with observed metal release data showing increased release rates of zinc with time, and a more rapid release rate in seawater after the longest period investigated (3000 s) for both experimental set-ups. This can tentatively be explained by a faster interaction and local destruction of the aluminum oxide by the high chloride content in seawater compared with artificial rainwater. Both sulfate and chloride were observed by means of XPS/AES on Galvalume surfaces exposed to seawater conditions. According to literature findings may sulfate be associated to the local formation of basic zinc sulfates and or chlorides in zinc-rich interdendritic areas under atmospheric conditions [35, 36, 60, 68, 145].

Figure 4.9. Average release rates of zinc and aluminum (a) and corresponding annual release Al/(Al+Zn) ratio for each year of exposure (b) for Galvalume exposed to unsheltered conditions at the marine site of Brest, France for five years. (Paper I)

Release rates of zinc and aluminum were continuously monitored for Galvalume surfaces exposed at unsheltered conditions for 5 years in a marine environment[14]. Observed release rates obtained under laboratory conditions revealed the same trend as observed at field conditions. Significantly more zinc was released compared with aluminum throughout the exposure period, Fig.4.9a, with an annual Al/ (Al+Zn) release ratio varying within 0.04±0.01 for individual years of exposure, Fig.4.9b. Field data revealed, similar to laboratory findings, a preferential release of zinc with increased rates during the first year of exposure, Fig.4.9a. Al2O3 was predominantly

37  

present over the entire Galvalume surface before exposure, and was subject to local destruction upon interaction with chloride ions during exposure. Gradual formation of zinc-rich corrosion products, most likely non-crystalline basic zinc chlorides and/or sulfates, was preferentially taking place in the zinc-rich interdendritic phases. This resulted in gradually increased release rates of zinc during the first year of exposure. However, due to the gradual formation and integration of aluminum rich corrosion products also in zinc-rich interdendritic areas [14], reduced release rates of zinc were observed with time during the 5 year exposure.

4.4 The dual-phase structure of Cu40Zn consists of zinc-richer β-phase crystals of lower surface nobility than the α-phase. Corrosion initiation is observed in the β-phase at low pre-deposition of NaCl. (non-published data)

A dual-phase brass alloy, Cu40Zn, was introduced to explore the influence of microstructure on corrosion initiation. Prior to exposure was the alloy surface analyzed by means of optical microscopy (OM) to gain metallographic information, SEM/EDS for morphology and elemental compositional analyses, and SKPFM to assess surface topography and to map differences in Volta potential over the surface. The surfaces were pre-deposited with 0.1 μg/cm2 NaCl and subsequently exposed to wet/dry cycles. Corrosion initation was primarily analyzed by means of CRM.

Figure 4.10. Images showing the microstructure of an unexposed Cu40Zn surface by optical microscopy (a), and by SEM after etching (2M HCl+ 0.2M FeCl3) for 15 s to reveal the alloy microstructure (b).

The microstructure of the diamond-polished Cu40Zn surface is shown in Fig.4.10a, as viewed by OM. The dual-phase structure consists of β-phase irregular grains typically sized from 5 to 20 µm in an α-phase matrix, and result from the solidification processes during casting [43]. A SEM image of the etched surface further reveals the α-β dual phase characteristics, Fig.4.10b. The β-phase exhibits a platelet-type

38

morphology and occurs in different grains with an estimtated surface coverage fraction of approximately 10%. Elemental compositional EDS analysis revealed a more zinc-rich β-phase (42-46 wt-% Zn/(Zn+Cu)) compared with the α-phase (37-41 wt-% Zn/(Zn+Cu)). The α-phase solidifies firstly during the cooling process and obtains hence a higher Zn-content, compared with the subsequent solidification of the β-phase of lower zinc content [43]. The SEM/EDS investigation revealed further the presence of granular sulfur-rich inclusions (0.5-10% wt-% S/(S+Zn+Cu)). Inclusions with a high content of sulfur (5-10% wt-% S/(S+Zn+Cu)) were associated with a higher content of zinc compared with adjacent surface areas, and indicate a chemical composition similar to ZnS [110].

 

Figure 4.11. AFM-based topography (a) and Volta potential mappings (b) obtained with SKPFM of an unexposed diamond polished Cu40Zn surface.

Figure 4.12. Optical image (a) and combined Raman mapping images (b, c) obtained with CRM of the squared-sized area in (a). Raman map of Cu2O ((b), integrated between 150 and 250 cm-1) and of ZnO ((c), integrated between 400 and 600 cm-1) in corrosion products formed on Cu40Zn with pre-deposited NaCl (0.1μg /cm2) after one wet/dry cycle at 90% RH.

Complementary measurements by SKPFM were performed on the diamond-polished surface observe variations in topography and Volta potential, Fig.4.11. The

39  

AFM-image (Fig.4.11a) shows that the β-phase exhibits a lower height than the α-phase, which may be a result of the polishing process. Grains of the β-phase exhibit a lower Volta potential compared with the surrounding α-phase, as revealed by the corrsponding SKPFM-image (Fig.4.11b). This suggests a lower relative nobility of the more zinc-rich β-phase with possibly a higher susceptibility for corrosion initiation. Furthermore, a granular area of circular shape of even lower Volta potential than both the α- and the β-phase was observed with SKPFM (Fig.4.11b). This area most likely correspond to zinc-enriched sulfur inclusions seen in Fig.4.10a. Their presence is consistent with previous findings [14, 75, 113].

Figure 4.12a displays the morphology of the Cu40Zn surface pre-deposited with low amounts of NaCl (0.1 μg /cm2) after exposure to one wet/dry cycle at 90% RH. A circular darker area appears in the image, which indicates the interaction of a NaCl droplet with the surface upon deliquescence of deposited NaCl. The slightly zinc-richer β-phase crystals inside the circular area exhibit a darker feature compared with the matrix. CRM measurements were conducted on the square-sized area seen in Fig.4.12a. The results revealed a lateral distribution of cuprite (Cu2O) and of zinc oxide (ZnO), Fig.4.12b-c. The microstructure evidently influences the initial formation of corrosion products as Cu2O and ZnO are preferentially formed in the slightly zinc-richer β-phase grains of lower Volta potential compared with the surrounding α-phase matrix.

The results are in agreement with recent literature findings [110], in which similar microstructural effects between grains of slightly different Zn-content could be observed when a Cu20Zn alloy was exposed to a dilute NaCl solution. It should be added, however, that this microstructural influence could not be observed when the same Cu40Zn alloy was exposed to wet/dry conditions with a significantly higher amount of pre-deposited NaCl (4 μg/cm2). In this case the chloride-induced corrosion effect seemed to be evenly distributed along the corroded Cu40Zn surface, rather than preferentially occur in the β-phase.

4.5 Microgalvanic effects on a Cu-Zn patterned sample with pre-deposited chlorides result in a radial distribution of corrosion products from the Cu cathode to the Zn anode upon cyclic exposures in humidified air. (non-published data)

To further explore the effect of structural heterogeneity of CuZn substrates on chloride-induced atmospheric corrosion a Cu-Zn patterned sample was prepared that consists of 20 by 20 µm-sized squared Cu-islands regularly deposited on a pure Zn-substrate, each square at a distance of 20 µm from the next. This Cu-Zn patterned

40

sample, from now on designated CuZnP, was pre-deposited with 4μg/cm2 NaCl and exposed to wet/dry cycles at 90% RH. The lateral distribution and composition of corrosion products were then analyzed by means of SEM/EDS and CRM, and their location related to the Cu-Zn surface pattern.

Figure 4.13. SEM images of corrosion products formed on CuZnP pre-deposited with 4μg /cm2 NaCl and exposed for 1(a) and 6 cycles (b-d) at 90% RH.

Fig. 4.13 reveals the morphology, by means of SEM, of corrosion products formed on CuZnP after cyclic exposures. After exposure to 1 cycle, relatively uniform corrosion products preferentially formed on the Zn-matrix surrounding each Cu-island, Fig.4.13a. After exposure to 6 cycles, a characteristic circular-shaped pattern of corrosion products formed on the Zn-matrix at a certain distance from the Cu-islands, Fig.4.13b. At higher magnification (Figs. 4.13c and d) the circular features are seen as small granular corrosion products, Fig.4.13c, whereas clusters of sheet-like corrosion products are seen outside the circles, at the center of each of the four Cu islands seen in Fig.4.13d.

In order to identify the phases in the corrosion products and their location CRM measurements were performed after exposure to 6 wet/dry cycles, Fig.4.14. CRM mapping, Fig.4.14b, reveals that cuprite, Cu2O (main peaks at 219, 424 and 636cm-1)

41  

[146] formed on the Cu-island, zinc oxide, ZnO (main peaks for crystalline ZnO at 426 cm-1 and amorphous ZnO at 560 cm-1) [147] on the Zn-matrix, and nantokite, CuCl (main peaks at 290, 613 and 1110 cm-1) [146] along the boundary in-between. An area of the Zn-matrix was selected between the four Cu-islands for further CRM-analysis, Figs.4.14c-d. Besides ZnO and Cu2O, was hydrozincite, Zn5(OH)6(CO3)2 (main peaks at 386, 1070, 1378, 1586, 2938 and 3450 cm-1) [148, 149] formed close to the periphery of the circle whereas sheet-like corrosion products of simonkolleite, Zn5(OH)8Cl2·H2O (main peaks at 270, 400, 910, 2936 and 3486 cm-1) [149-152] formed outside the circular features at the centre of the four Cu-islands.

Figure 4.14. Optical images ((a) and (c)) and combined Raman mapping images ((b) and (d)) of corrosion products formed on CuZnP pre-deposited with 4 μg /cm2 NaCl and exposed during 6 cycles at 90% RH. CRM on Cu2O (blue, integrated between 150 and 250 cm-1), CuCl (red, integrated between 250 and 350 cm-1), ZnO (yellow, integrated between 400 and 600 cm-1, (e)), the CO3

2- band in Zn5(OH)6(CO3)2 (cyan, integrated between 1000 and 1100 cm-1, (f)) and the OH band in Zn5(OH)8Cl2·H2O (purple, integrated between 3400 and 3550 cm-1, (g)).

The lateral distribution of corrosion products formed on both Cu and Zn are believed to be the result of microgalvanic effects between each Cu-island and its surrounding Zn matrix, and be a result of their relatively large difference in electrode potential. The circular feature of formed corrosion products and the radial distribution of corrosion products imply, based on the basic corrosion cell theory, the formation of a gradient in potential and in chemical composition from each Cu-island (cathodic area characterized by high local pH and low local chloride concentration in the aqueous adlayer) to the anodic area in the Zn-matrix, centrally located between the four Cu-islands (anodic area, low local pH and high local chloride concentration).

42

Cu2O is formed on the cathodic Cu-island, CuCl along the Cu/Zn boundary, ZnO and Zn5(OH)6(CO3)2 on the Zn-matrix closer to the Cu-islands and Zn5(OH)8Cl2·H2O in the centre of the anodic area that exhibits a high chloride concentration and low pH. The distribution of corrosion products may be compared with those found by Cole et al [152, 153], who studied formation of corrosion products on pure zinc beneath a single NaCl-drop. As a consequence of the radial distribution in potential and local chemistry in the NaCl-drop, Zn5(OH)6(CO3)2 was found to preferentially precipitate in the secondary spreading area of the droplet characterized by a higher pH, while Zn5(OH)8Cl2·H2O formed in the centrally located anodic area, characterized by a higher chloride ion concentration and lower pH [152, 153].

43  

5 Corrosion product evolution and characteristics

5.1 Severe corrosion product flaking observed for Cu and Cu4Sn in chloride-rich environments is primarily connected to the presence of nantokite. Minor effects observed for Cu15Zn and Cu5Al5Zn. (Paper V)

The extent of flaking of poorly adherent corrosion products on Cu sheet and three Cu-based alloys was examined by means of SEM (surfaces and cross-sections). The hypothesis of flaking being connected with the formation of nantokite was further elucidated by CRM measurements.

Figure 5.1. Top view SEM images of Cu sheet (a) and Cu4Sn (b) surfaces showing severe corrosion product flaking, and of Cu15Zn (c) and Cu5Al5Zn (d) surfaces showing minor, or no flaking. All samples were exposed for 1 year at unsheltered marine field conditions. (Paper V)

SEM-top view images of the morphology of the patina formed on Cu sheet and the copper-based alloys (Cu4Sn, Cu15Zn and Cu5Al5Zn) after 1 year at the marine site

44

are shown in Fig.5.1. Loosely and heterogeneously adherent corrosion products that have detached as flakes were observed on both Cu sheet (a) and Cu4Sn (b), while more adherent corrosion products were evident on both Cu15Zn (c) and Cu5Al5Zn (d) displaying minor or no flaking at all, respectively. In general, relatively severe flaking was observed for Cu and Cu4Sn throughout the entire 3 year marine field exposure, whereas the degree of flaking was only minor for Cu15Zn and no flaking at all for Cu5Al5Zn during the same time period.

Figure 5.2. SEM images of cross-sections of the corrosion patina formed on Cu sheet (a), Cu4Sn (b), Cu15Zn (c), and Cu5Al5Zn (d), after 3 years of marine unsheltered field exposure. (Paper V)

Cross-sectional SEM images of the patina of Cu sheet (a), Cu4Sn (b), Cu15Zn (c) and Cu5Al5Zn (d) are displayed in Fig.5.2 after 3 years of exposure. In general, the patina formed on all materials consisted of a two-layer structure, except for Cu4Sn showing a striped-layer structure. The total average thickness of the patina decreased according to Cu4Sn ˃ Cu ≈ Cu15Zn ˃ Cu5Al5Zn, a trend remaining during the entire exposure period. Further EDS line-analyses were conducted on the cross-sections of all Cu and Cu-based alloys to identify the corresponding elemental distribution. The patina contained very similar main constituents, and revealed a non-uniform inner layer primarily composed of Cu and O, probably as cuprite (Cu2O), and a porous outer

45  

layer predominantly rich in Cl, besides Cu and O, probably as paratacamite/atacamite, Cu2(OH)3Cl [71, 79]. In particular, there was an interfacial discontinuous region enriched with Cu and Cl between the two inner layers of patina on Cu4Sn, Fig.5.2b, probably attributed to nantokite (CuCl). Literature findings by SEM/EDS have reported the possible presence of nantokite between the inner layer and the substrate, or within the outer layer after long term field and/or laboratory exposures [71, 79, 154]. Its presence has also been confirmed by XRD [70-73, 155]. The presence of cuprite (Cu2O) and paratacamite (or possibly atacamite, Cu2(OH)3Cl) were confirmed by means of GIXRD.

Figure 5.3. Optical images ((a)-(d)) and combined Raman mapping images (laser source of 532 nm, (e)-(h)) obtained with CRM on Cu2O (blue inner layer, integrated between 150 and 250 cm-1, (i)), CuCl (red intertwined layer, integrated between 250 and 350 cm-1, (j)) and the OH band in Cu2(OH)3Cl (green outer layer, integrated between 3300 and 3500 cm-1, (k)) of cross-sections of patina formed on bare Cu sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn exposed at unsheltered conditions for 3 years at a marine site. (Paper V)

Based on the SEM/EDS findings it was proposed that flaking was related to the formation of nantokite. This hypothesis was further elucidated by CRM measurements performed on cross-sections of the patinas formed on bare Cu and Cu-based alloys

46

exposed for 3 years at the marine site, displayed in Fig.5.3. The layered structure of the patina with its main constituents was clearly distinguished by the combined CRM images. Cuprite, Cu2O (main peaks at 219, 424 and 636 cm-1) [156], was the dominant constituent of the inner layer, and paratacamite, Cu2(OH)3Cl (main peaks at 373, 519, 930, 977, 3364 and 3447 cm-1) [157] the predominant constituent of the outer layer of the patina on bare Cu sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn. A discontinuous thin layer of nantokite was intertwined between the inner and outer layers, supported by the Raman peak positions characteristic of artificial nantokite, CuCl (main peaks at 290, 613 and 1110 cm-1) [146]. By comparing the CRM images generated for the different materials, Fig.5.3, it was obvious that the occurrence of nantokite was significantly more pronounced for Cu sheet and Cu4Sn compared with Cu15Zn (minor) and Cu5Al5Zn (non-significant). Its presence coincided with the tendency for corrosion product flaking, which indicates that nantokite plays an important role in the flaking process.

5.2 Transformation of nantokite to paratacamite results in volume expansion within the patina causing corrosion product flaking for Cu and Cu4Sn. (Paper V)

A layer of synthetic nantokite was grown on bare Cu sheet and subsequently exposed to wet/dry cycles at 90% RH to test the hypothesis that nantokite transforms at atmospheric conditions to a basic copper chloride. The transformation process was primarily verified by GIXRD measurements. Different interactions between the NaCl droplets and the surfaces were elucidated by parallel exposures of the Cu sheet and the Cu-based alloys, pre-deposited with NaCl (4 µg/cm2), to the same wet/dry exposures. Formed corrosion products were analyzed by means of SEM/EDS and CRM.

The synthetic nantokite layer grown on bare Cu sheet was exposed to wet/dry cyclic conditions in the climatic chamber at 90% RH and subsequently analyzed with GIXRD. The thickness of the corrosion product layer, determined with a microprocessor coating thickness gauge, increased continuously from an average thickness of approximately 10 µm prior to exposure to approximately 50 µm after 14 cycles. In agreement with field exposed samples was the gradual transformation of artificial nantokite to paratacamite confirmed by GIXRD-analyses, Fig.5.4. It was evident that this transformation process is associated with a significant volume expansion, a process that creates internal physical stresses within the patina and possibilities for subsequent corrosion product flaking [80].

47  

Figure 5.4. GIXRD diffraction patterns illustrating the transformation of a synthetic nantokite (CuCl) layer on bare Cu sheet, pre-deposited with 4 μg/cm2NaCl, to paratacamite (Cu2(OH)3Cl), after 0 (unexposed), 2 and 14 wet/dry cycles at 90% RH. (Paper V)

Bare Cu sheet, Cu4Sn, Cu15Zn and Cu5AlZn, pre-deposited with NaCl (4 µg/cm2) were further exposed to wet/dry cycles in the climatic chamber at 90% RH and the morphology and elemental composition of formed corrosion products were investigated using SEM/EDS. A few ring-formed corrosion features were present on Cu sheet and on all Cu-based alloys after 1 cycle (6 h), Figs.5.5a-d, features that were induced by the spreading of deliquescent NaCl droplets [143]. Whitish small granular corrosion products revealed a higher content of Cl and O compared with the surrounding areas. After 14 cycles (7 days) were observed corrosion effects on Cu sheet and Cu4Sn (Fig. 5.5e and 5.5f) quite severe with a high frequency of granular corrosion products that almost completely covered the surface, independent of the original positions of initially formed NaCl droplets. Aggregation of white granular corrosion products, randomly observed on Cu sheet and Cu4Sn containedmainly Cu, O and Cl (based on EDS findings). Less severe corrosion effects and ring-like surface features with more uniformly formed corrosion products were observed on Cu15Zn and Cu5Al5Zn after 14 cycles (Figs.5.5g and 5.5h) compared with bare Cu sheet and Cu4Sn.

48

Figure 5.5. SEM images of patina formed on bare Cu sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn pre-deposited with NaCl (4 μg /cm2)and exposed for 1 (a-d) and 14 (e-h) cycles at 90% RH. (Paper V)

49  

Complementary CRM measurements of local areas (25x25 µm) were conducted on Cu4Sn after exposures to the wet/dry cycles described above. A ring-like corrosion feature present after 1 wet/dry cycle was selected for further investigation, Fig.5.6a. The CRM investigation revealed an evident formation of cuprite surrounding the periphery of the ring, Fig.5.6f. The characteristic morphology of cuprite followed the ring-like feature, seen in Fig.5.5b). The formation of cuprite is electrochemically driven and has been studied in more detail elsewhere [22]. Except for cuprite was nantokite observed in the ring-like features after 2 cycles (1 day), Figs.5.6b and 5.6g. Cuprite covered most of the surface while nantokite was preferentially concentrated along and inside the periphery of the ring. The morphology and distribution of nantokite with white granular-shaped corrosion products seen in Fig.5.5b, are consistent with literature findings [80]. After 6 cycles (3 days) was a large part of the surface covered with circular clusters of nantokite surrounded by cuprite, Figs.5.6c and 5.6h. According to literature findings is nantokite formed through a reaction between chloride and cuprous ions dissolved from cuprite [70, 73]. After 14 cycles were the circular features no longer visible and local clustered features were randomly observed, Fig.5.6d. Due to the volume expansion of the corrosion products, the focus of the optical camera in the CRM was lifted with approximately 1µm to enable a sharp image to be obtained, Fig. 5.6e. Generated CRM images provided evidence of cuprite in the surrounding lower region (Fig.5.6i), while the corrosion products mostly consisted of paratacamite in the upper region (Fig. 5.6j). The height difference between nantokite and paratacamite further confirmed the concomitant volume expansion within the patina during this transformation process, Figs.5.6d-j.

Both cuprite and nantokite were identified with CRM on bare Cu sheet, Cu15Zn and Cu5Al5Zn surfaces pre-deposited with NaCl and exposed to wet/dry cycles. The occurrence of nantokite was, in accordance with findings from the marine field exposure, significantly more abundant on bare Cu sheet and Cu4Sn compared with Cu15Zn and Cu5Al5Zn. Furthermore, no paratacamite was observed by means of CRM on either Cu15Zn or Cu5Al5Zn, whereas this phase was readily detected on Cu sheet and Cu4Sn after 14 cycles of exposure. When comparing all results was the same patina reaction pathway evident in the laboratory exposures with humid air for synthetically grown CuCl and surfaces pre-deposited with NaCl, and at marine field conditions.

50

Figure 5.6. Optical images ((a)-(e)) and corresponding combined Raman mapping images (laser source of 532 nm, (f)-(j)) obtained with CRM of the Cu2O band (blue, integrated between 150 and 250 cm-1), CuCl band (red, integrated between 250 and 350 cm-1) and the OH band in Cu2(OH)3Cl (green, integrated between 3300 and 3500 cm-1), of the patina formed on Cu4Sn pre-deposited with NaCl (4 μg/cm2) and exposed for 1, 2, 6 and 14 wet/dry cycles at 90% RH. (e) image obtained at the same area as (d) but with the focus of the optical camera lifted 1µm. (Paper V)

51  

5.3 The initial formation of Zn- and Zn/Al-hydroxycarbonates reduces the sensitivity of Cu15Zn and Cu5Al5Zn to chloride-induced corrosion, and also the release of zinc from Galfan at marine conditions. (Papers II, IV, V)

Complementary IRAS analyses were performed to examine corrosion product formation on bare Cu sheet and the Cu-based alloy surfaces pre-deposited with NaCl and exposed to wet/dry cycles. Comparative analyses were obtained via e.g. GDOES for samples exposed to the marine outdoor test site. Improved surface properties induced by the formation of hydroxycarbonates were further elucidated for both the Cu-based alloys and for the Zn-Al alloys.

Figure 5.7 displays ex-situ IRAS spectra obtained after exposures to 14 wet/dry cycles for bare Cu sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn, respectively. Starting from lower wavenumber and going upwards was the peak at 651 cm-1, attributed to the vibration of Cu2O, cuprite [158], identified for all materials. A broad band ranging from 1300 to 1600 cm-1 with two resolved peaks located at 1395 and 1515 cm-1 was evident in the spectra of Cu15Zn and Cu5Al5Zn, but barely observed for Cu sheet and Cu4Sn. These bands are most likely assigned to anti-symmetric stretching modes of carbonate (CO3

2-) [148], possibly in hydrozincite (Zn5(OH)6(CO3)2) and/or hydrotalcite (Zn6Al2(OH)16CO3·4H2O). In the higher wavenumber range was a broad band observed for both Cu15Zn and Cu5Al5Zn in the range between 3000 to 3700 cm-1, a band that commonly is attributed to hydroxide ions (OH-) or water.

Figure 5.7. Ex-situ IRAS spectra obtained on bare Cu sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn pre-deposited with NaCl (4 μg /cm2) and exposed to 14 wet/dry cycles at 90% RH. (Paper V)

4000 3500 3000 2500 2000 1500 1000 5000.0

0.1

0.2

0.3

0.4

0.5

Cu

Cu15Zn

Cu5Al5Zn

Cu2O

OH-/H2O CO

3

2-

Cu4Sn

Ab

sorb

ance

(-

log

(R/R

0))

Wavenumber (cm-1)

52

Figure 5.8 illustrates the depth profiles of zinc with GDOES for both Cu15Zn and Cu5Al5Zn after 1 year of marine exposure. The presence of zinc-rich corrosion products was evident and predominantly enriched within the patina when compared with the bulk content. These observations further supported the presence of hydrozincite (Zn5(OH)6(CO3)2) and/or hydrotalcite (Zn6Al2(OH)16CO3·4H2O), primarily formed within the patina on Cu15Zn and Cu5Al5Zn, respectively.

Figure 5.8. Enrichment of zinc in the patina compared with its bulk content in relation to copper for Cu15Zn (a) and Cu5Al5Zn (b) after 1 year of unsheltered marine exposure, based on GDOES measurements. (Paper IV)

The presence of hydrozincite is important for the barrier properties on bare zinc sheet and galvanized steel in chloride-rich environments [159]. Literature findings report accelerated atmospheric corrosion of zinc in marine environments during an initial period of NaCl spreading, combined with a promoted formation of hydrozincite in the secondary spread zone characterized by a more alkaline pH. Hydrozincite is reported to protect the surface from further corrosion to a much larger extent compared with other basic zinc compounds due to its resistance to chloride diffusion and relatively high stability [160]. Dissociated protons or metal ions at the surface contribute to a negative surface charge possessed by hydrozincite at pH lower than 7 [159]. The negatively charged surface of hydrozincite is capable to repel chloride ions, which may result in the prevention of chloride-induced atmospheric corrosion for such surfaces [159]. These findings are in good agreement with the earlier conclusions of less occurrence and formation of nantokite on CuZn15 and Cu5Al5Zn compared with bare Cu sheet and Cu4Sn, providing an explanation for the different extent of corrosion product flaking of the alloys investigated.

53  

Similar effects of hydrotalcite have been elucidated in the literature [65] with improved barrier properties of corrosion products as observed on Zn-Al coatings compared to bare Zn/galvanized steel in chloride-rich environments [161]. Zinc release patterns induced by the interaction of humidity and rainwater on the surface during long-term field exposures are given as an example. Regardless if zinc-rich phases dominated the patina on both zinc and Galfan after longer exposure period, a significantly lower (approximately twice as low) released amount of zinc was observed from Galfan compared with bare zinc sheet throughout the 5-year exposure period. Detailed corrosion product analysis revealed Zn5(OH)6(CO3)2 as one of the main corrosion products formed on bare zinc sheet [147] and Zn6Al2(OH)16CO3·4H2O on Galfan.

5.4 Similar corrosion products form on Galfan and bare Zn sheet and on Galvalume and bare Al sheet, respectively, upon accelerated chloride test conditions. (non-published data)

Accelerated N-VDA-tests were performed on bare Zn and Al sheet, Galfan and Galvalume. Corrosion product formation was analyzed by means of SEM/EDS, IRAS and GIXRD. The aim was to evaluate whether the accelerated corrosion test in a realistic way could mimic the behavior (similar corrosion products and formation mechanisms) of Zn-Al coatings on steel as in automotive environments of high chloride content (or at chloride-rich marine conditions) [119].

Figure 5.9. SEM images of corrosion products formed at N-VDA test conditions on bare Zn sheet exposed for 1 (a) and 7 (b) days, forming carbonate-rich corrosion products (c), and on Galfan exposed for 1 (d) and 7 (e) days, forming chloride-rich corrosion products (f).

54

The morphology of corrosion products formed on bare Zn sheet and Galfan during the first week of the VDA test exposure is presented in Fig.5.9. Galfan revealed very similar corrosion behavior as bare Zn sheet. Randomly distributed corrosion products were present to a large extent on the surfaces already after 1 day of exposure. After 7 days were layer-structured patinas observed for both bare Zn sheet, Figs.5.10a-b and Galfan, Figs.5.10d-e. ZnO, Zn(OH)2 and Zn5(OH)8Cl2·H2O were the main corrosion products identified on both bare Zn sheet and Galfan with Al2O3, Zn6Al2(OH)16CO3·4H2O and/or Zn2Al(OH)6Cl·2H2O additionally detected on Galfan during the initial stages of the test. Sheet-like corrosion products were observed on Galfan after 4 days of exposure, Fig.5.9f, and were predominantly composed of Cl, Zn, Al and O, possibly assigned as Zn5(OH)8Cl2·H2O and/or Zn2Al(OH)6Cl·2H2O. After longer exposure periods was Zn5(OH)6(CO3)2 the dominating constituent of the patina, possibly present in the round-clustered features containing significant amounts of Zn, C and O, demonstrated for bare Zn sheet exposed for 5 days, Fig.5.9c.

Figure 5.10. SEM images of corrosion products formed on bare Al sheet at N-VDA test conditions exposed for 1 (a) and 7 (b) days, forming aluminum-rich corrosion products (c) and Galvalume exposed for 1 (d) and 7 (e) days, with corrosion products forming locally occurring platelets (f).

Galvalume revealed a similar corrosion behavior as bare Al sheet due to its high alloy content of aluminum. Circular corrosion products were formed after 1 day of exposure and were largely dependent on the NaCl interaction. These corrosion products gradually covered the whole surface after 7 days of accelerated testing of both bare Al sheet, Figs.5.10a-b, and of Galvalume, Figs.5.10d-e. Deposition of chlorides highly influenced the corrosion of both Galvalume and Al due to the rapid

55  

local damage of the compact native thin film of Al2O3. AlOOH and Al(OH)3 were preferentially formed on bare Al sheet during initial stages of the accelerated testing, and on Galvalume during intermediate stages. The bare Al surface revealed a fine-grain surface patina when exposed for 4 days (Fig.5.10c) and micron-sized crystalline corrosion products (Fig.5.10b), containing significant amount of Al and O. In addition, Zn-Al containing corrosion products, Zn6Al2(OH)16CO3·4H2O and/or Zn2Al(OH)6Cl·2H2O were identified on the Galvalume surface. Islands with platelet-like morphologies were preferentially formed in zinc-rich interdendritic regions after 5 days of exposure, Fig.5.10f. SEM observations, and proposals in the scientific literature [68], suggest these corrosion products to consist of Zn6Al2(OH)16CO3·4H2O. This could not be verified in this study.

A comparison of the corrosion behavior of bare metals (Zn and Al) and alloys (Galfan and Galvalume) is schematically given in Fig.5.11. Galfan revealed a similar corrosion behavior as bare Zn sheet, whereas Galvalume exhibited similar behavior as bare Al sheet.

Figure 5.11. Comparison of the corrosion behavior of bare Zn and Al sheet and the Zn-Al alloys (Galfan and Galvalume) after exposure to accelerated N-VDA test conditions.

56

5.5 Laboratory set-ups with exposures to chloride-rich environments were able to successfully reproduce the predominating corrosion products formed at marine outdoor conditions for the Zn-Al coatings, bare Cu sheet and the Cu-based alloys. (Papers I, II, IV, V)

5.5.1 Zn-Al alloy coatings

To obtain a comprehensive understanding of corrosion product formation and their evolution for Zn-Al alloys exposed to chloride-rich environments, are all corrosion products identified on Galfan and Galvalume upon laboratory, unsheltered marine field and accelerated exposures compiled and displayed in Table. 5.1 based on the multi-analytical approach employed in this thesis.

The relatively simple laboratory exposures with pre-deposited NaCl and humidity or chloride-containing media were able to generate the same predominating corrosion products formed of Galfan and Galvalume after long-term marine field exposures, even though in different amounts and relative proportions. The only exception was the lack of formation of NaZn4Cl(OH)6SO4·6H2O, present after long-term marine field exposure conditions. This was primarily attributed to the lack of SO2 or sulfate deposition involved in the laboratory wet/dry cyclic set-ups and accelerated test.

Table 5.1. Compilation of corrosion products identified on Galfan and Galvalume after laboratory-, marine field- and accelerated exposures for different time periods. (GF-Galfan, GV-Galvalume, *- tentative phases)

Corrosion product Laboratory

exposure Field exposure (Marine site)

Accelerated exposure (N-VDA-test)

ZnO GF GV GF GV GF GV

Zn(OH)2 GF* GV* GF* GV* GF GV*

Zn5(OH)6(CO3)2 GF GV*

Zn5(OH)8Cl2·H2O GF GV* GF GV* GF GV

NaZn4Cl(OH)6SO4·6H2O GV* GF GV*

Al2O3 GF GV GF GV GF GV

AlOOH GV* GV

Al(OH)3 GF* GV* GF* GV GF* GV

ZnAl2O4 GF* GF* GF*

Zn6Al2 (OH)16CO3·4H2O GF* GV* GF GV* GF GV

Zn2Al(OH)6Cl·H2O GF* GV* GF* GV* GF* GV*

57  

 

5.5.2 Cu and Cu-based alloys

A compilation of corrosion products identified within the patina of Cu sheet and the Cu-based alloys exposed to chloride-rich laboratory conditions via wet/dry cyclic sequences and outdoor marine conditions is given in Table 5.2 based on the multianalytical investigation presented earlier.

Exposures in the relatively simple laboratory set-ups with pre-deposited NaCl and humidified air were able to successfully reproduce the same predominating corrosion products, although in different relative proportions and surface coverage, formed during long-term outdoor marine conditions.

Table 5.2. Compilation of corrosion products identified on bare Cu sheet and Cu-based alloys exposed to chloride-rich laboratory and marine field exposures. (*- tentative phases)

Corrosion products Laboratory exposure Field exposure (marine site)

Cu2O Cu Cu4Sn Cu15Zn Cu5Al5Zn Cu Cu4Sn Cu15Zn Cu5Al5Zn

CuCl Cu Cu4Sn Cu15Zn Cu5Al5Zn* Cu Cu4Sn Cu15Zn Cu5Al5Zn*

Cu2(OH)3Cl Cu Cu4Sn Cu Cu4Sn Cu15Zn Cu5Al5Zn

SnO2 Cu4Sn

ZnO Cu15Zn* Cu15Zn

Zn(OH)2 Cu15Zn Cu15Zn*

Zn5(OH)6(CO3)2 Cu15Zn* Cu15Zn*

Zn5(OH)8Cl2·H2O Cu15Zn* Cu15Zn* Cu5Al5Zn*

Al2O3 Cu5Al5Zn* Cu5Al5Zn

Zn6Al2 (OH)16CO3·4H2O Cu5Al5Zn* Cu5Al5Zn

Zn2Al(OH)6Cl·H2O Cu5Al5Zn* Cu5Al5Zn*

5.6 General scenarios for patina evolution established for Zn-Al coatings and corrosion product flaking mechanisms proposed for Cu-based alloys in chloride-rich atmospheric environments. (Papers II, V)

5.6.1 Zn-Al alloy coatings

Previous investigations have established general sequences for the evolution of the most commonly occurring corrosion products on bare Zn sheet [53] and pure Al sheet [60, 162] at different atmospheric conditions.

58

The established evolution scenarios of relevance for chloride-rich conditions (sheltered conditions) are presented in Fig.5.12 and compared with findings and proposed scenarios from this study for the investigated Zn-Al alloy coatings (unsheltered conditions) [14].

In moist atmospheres are zinc- and aluminum oxides and/or hydroxides formed immediately on all materials. Atmospheric CO2 dissolves in the aqueous adlayer and reacts with hydroxides, rapidly forming hydrozincite (Zn5(OH)6(CO3)2) and hydrotalcite (Zn6Al2(OH)16CO3·4H2O). Triggered by chlorides, hydroxychlorides, simonkolleite (Zn5(OH)8Cl2·H2O) is subsequently formed on bare Zn sheet, and Zn2Al(OH)6Cl·2H2O and/or Zn5(OH)8Cl2·H2O on the Zn-Al alloy coatings. The evolution from one phase to another can in addition to chemical transformations proceed via previously proposed ion-exchange mechanisms [53], since all hydroxycarbonates and hydroxychlorides (and hydroxysulfates) possess structural resemblance and layered structures held together by different anions such as carbonate, chloride and/or sulfate.

Figure 5.12. Established sequences of corrosion product formation on bare zinc sheet/galvanized steel (left) and bare Al sheet (right) in chloride-rich atmospheric environments compared with the proposed sequence for Al-Zn-coatings of Galfan and Galvalume (middle) based on findings in this study. Vertical arrows indicate the evolution of one phase into the next phase and phases without vertical arrow can exist simultaneously next to each other. Underlined phases are also observed in corresponding sequences for bare Zn and bare Al. (Paper II)

Marine environments are not only characterized by abundant chloride deposition, but also by sulphate deposition, e.g. from sea spray and gaseous SO2 [144]. Consequently, NaZn4Cl(OH)6SO4·6H2O, another phase with a layered structure, has been identified after long-term exposure perods on both bare Zn [144] and Galfan [143]. The predominating corrosion product formed on bare Al sheet is Al(OH)3, a

59  

predominating phase also on Galvalume after long-term exposures in marine environments [14]. Besides being of importance from a more fundamental atmospheric corrosion mechanistic perspective is the proposed evolution sequence for Al-Zn coatings important for elucidating long-term barrier properties and corrosion characteristics of these and similar alloys in chloride-rich environments.

5.6.2 Cu and Cu-based alloys

Bare Cu sheet and Cu-based alloys were investigated in chloride-rich environment with the goal of revealing the possible mechanism behind corrosion product flaking. The established patina evolution scenario on bare Cu sheet [70] was shown to be valid also for the Cu-based alloys upon exposures in chloride-rich conditions, both at laboratory and field conditions. In addition, other phases were formed and identified related to reactions between the main alloying elements (Sn, Zn and Al) and given atmospheric constituents. These corrosion products turned out to be essential for the extent of corrosion product flaking for the different materials investigated at chloride-rich conditions.

A summary of all findings are schematically displayed in Fig.5.13, showing the sequence of corrosion products formed on bare Cu sheet and the Cu-based alloys in chloride-rich environments. Cuprite (Cu2O) forms rapidly on all materials at humid conditions, while zinc oxide (ZnO) and/or aluminum oxide (Al2O3) form on Cu15Zn and Cu5Al5Zn. Hydroxides react with dissolved atmospheric CO2 and form hydrozincite (Zn5(OH)6(CO3)2) and/or hydrotalcite (Zn6Al2(OH)16CO3·4H2O) on Cu15Zn and Cu5Al5Zn, respectively. High deposition rates of chloride result in the local formation of nantokite (CuCl) through reactions between chloride ions and dissolved cuprous ions. This phase was in these studies significantly more abundant on bare Cu sheet and Cu4Sn compared with Cu15Zn and Cu5Al5Zn. As a consequence revealed the former materials a significantly higher sensitivity to chloride-induced corrosion compared with the latter materials. Nantokite easily transforms to e.g. paratacamite (Cu2(OH)3Cl) through reactions with water and oxygen. Paratacamite is a more voluminous phase, with a molar volume (molar mass divided by the density) of 61.0 cm3 mole-1 compared with nantokite, of 23.9 cm3 mole-

1. The transformation process of nantokite to paratacamite hence contributes to a significant volume expansion. The discontinuous presence of nantokite within the patina and its continuous formation and transformation to paratacamite induce internal physical stresses, which result in a separation between the inner patina layer (mainly cuprite) and the outer patina layer (mainly paratacamite), causing corrosion product flaking primarily observed on bare Cu sheet and Cu4Sn.

60

Figure 5.13. Sequence of corrosion products formed on bare Cu sheet, Cu4Sn, Cu15Zn and Cu5Al5Zn upon long-term unsheltered exposures in a marine environment, and during short-term laboratory exposures in chloride-containing humid environments. Key processes governing the mechanism of flaking are indicated in blue. (Paper V)

Minor occurrence and lack of corrosion product flaking for Cu15Zn and Cu5Al5Zn, respectively, were attributed the formation and presence of hydroxycarbonates (hydrozincite and/or hydrotalcite) within the patina that altered the surface chemistry and surface charge in chloride-rich environment. As a consequence the formation of nantokite and subsequent transformation to paratacamite were suppressed for these alloys.

61  

6 Summary and outlook

The main objective of this doctoral thesis work has been to provide a fundamental understanding of the link between microstructure and corrosion initiation and corrosion product evolution and sequences for Zn-Al coatings on steel and of Cu-based alloys in chloride-rich atmospheric conditions via short-term laboratory and long-term outdoor marine environments, summarized in Fig.6.1. The investigated alloys are all commercially available and used in different automotive applications (Galfan (Zn5Al) and Galvalume (Zn55Al)) and for outdoor building applications (Zn55Al, Cu4Sn, Cu15Zn, Cu40Zn and Cu5Zn5Al). A multianalytical approach has been employed using a combination of surface sensitive in-situ and ex-situ analytical techniques to investigate microstructural characteristics and corrosion product initiation, formation, evolution and surface distribution. The influence of surface heterogeneity on corrosion initiation was explored based on microgalvanic effects and corrosion product evolution elucidated via short-term laboratory and long-term field exposures. Main conclusions are summarized below with an outlook for possible further investigations.

An obvious influence of the microstructure on corrosion initiation has been observed for Galfan and Galvalume upon laboratory and field exposures in chloride-rich environments. Initial corrosion was triggered by microgalvanic effects between active zinc-rich phases and passive aluminum-rich phases where local corrosion cells were created. Based on cyclic laboratory exposures of Galfan pre-deposited with NaCl in humidified air it was apparent that corrosion product initiation predominantly takes place in the zinc-richer microstructural phase adjacent to the aluminum-richer phase. These corrosion products extend with time following the direction of lamellas or rods, and form local clusters covering the surface.

High-temperature formed Al2O3 is the predominant surface oxide of Galvalume prior to exposure, an oxide being subject to local destruction upon interaction with chloride ions. Corrosion initiation and propagation of zinc-rich corrosion products primarily take place in interdendritic zinc-rich microstructural phases. With time also aluminum-rich corrosion products form and evolve and dominate the entire surface after long-term exposure periods in chloride-rich environments. Zinc is preferentially released from the surface compared with aluminum from Galvalume, both during the immersion test and the flow cell test at open circuit potential conditions, and upon the long-term unsheltered marine field exposure.

62

Figure 6.1. Compilation of main research findings of this doctoral thesis.

An evident microstructural influence on the initial formation of corrosion products has been observed for the Cu40Zn alloy upon laboratory exposures with humidified air and low amounts of pre-deposited NaCl. Cu2O and ZnO preferentially form in the slightly zinc-richer β-phase crystals adjacent to the α-phase matrix. A microgalvanic cell forms for each Cu island and its surrounding Zn matrix on the Cu-Zn patterned sample. This leads to the separation of Cu acting as the cathode and Zn as the anode beneath the NaCl droplet. Local formation of Cu2O, ZnO and CuCl were identified respectively on the Cu island, Zn matrix and the boundary in-between, whereas circularly distributed Zn5(OH)6(CO3)2, predominantly forms in the secondary spreading areas and clusters of Zn5(OH)8Cl2·H2O form outside these circles.

The laboratory set-up with humidified air and NaCl was able to reproduce the same corrosion products on both investigated alloy groups, although in different proportion and surface coverage, as identified at unsheltered outdoor marine conditions. Main corrosion products formed on Galfan and Galvalume, during atmospheric exposure in the chloride-rich laboratory and marine environments include the initial formation of ZnO, Al2O3 and/or ZnAl2O4 and subsequent formation of Zn6Al2(OH)16CO3·4H2O,

63  

and Zn2Al(OH)6Cl·2H2O and/or Zn5(OH)8Cl2·H2O. Corrosion products identified within the patina on bare Cu sheet and commercial Cu-based alloys exposed at similar chloride-rich conditions include an outer layer of Cu2(OH)3Cl and an inner layer of Cu2O, separated by an discontinuous layer of CuCl. Additional minor patina constituents include SnO2 (Cu4Sn), Zn5(OH)6(CO3)2 and Zn5(OH)8Cl2·H2O (Cu15Zn and Cu40Zn), Zn6Al2(OH)16CO3·4H2O/ Zn2Al(OH)6Cl·2H2O/Zn5(OH)8Cl2·H2O and Al2O3 (Cu5Al5Zn).

The marine exposure resulted in a patina of poorly adherent corrosion products on bare Cu sheet and Cu4Sn that easily flaked from the surface, more adherent corrosion products with less or non-significant extent of flaking were formed on Cu15Zn and Cu5Al5Zn, respectively. The tendency for flaking of corrosion products was significantly connected to the occurrence of CuCl, and attributed to a volume expansion and internal stresses within the patina induced by the gradual transformation of CuCl to Cu2(OH)3Cl. The fact that the Cu15Zn and Cu5Al5Zn alloys were significantly less sensitivity to chloride-induced corrosion product flaking compared with bare Cu sheet and Cu4Sn can be explained by a reduced interaction of chlorides with the former two alloys due to the initial formation and presence of Zn- and Zn/Al-hydroxycarbonates that hinder chloride surface interactions.

General scenarios for corrosion product evolution on Zn-Al coatings and the corrosion product flaking mechanism on Cu sheet and Cu-based alloys in chloride-rich environments are proposed based on findings from both laboratory and field exposures. The findings are of importance for the understanding of corrosion-related properties of these materials. Zn- and Zn/Al-hydroxycarbonates, Zn5(OH)6(CO3)2 and Zn6Al2(OH)16CO3·4H2O were two very important corrosion products identified on the Zn-Al and the Cu-based alloys. Their presence reduces, or hinders, the formation of CuCl, and hence chloride-induced corrosion product flaking for the Cu15Zn and Cu5Al5Zn alloys, and reduce the long-term annual release of zinc from Galfan compared with bare zinc sheet at marine conditions.

This thesis work provides novel in-depth understanding of corrosion initiation and corrosion product evolution at atmospheric conditions for Zn-Al coatings on steel and for Cu-based alloys in chloride-rich environments. However, there are still aspects of large importance for such environments that require further investigations. Even though chlorides are the main corrosion stimulators at marine conditions are effects of other pollutants such as aerosol particles, SO2/SO4

2-, NO2 and O3 on corrosion initiation and corrosion product formation and evolution important to investigate at atmospheric conditions. Acidified marine aerosols are believed to be of increasing importance for the atmospheric corrosion of metals [160]. Such effects could be further investigated via systematic laboratory exposures and evaluated using e.g. in-situ

64

IRAS. Moreover, due to the heterogeneous distribution of pre-deposited NaCl on the surface, local corrosion was largely triggered and the effect of the microstructure and microgalvanic cells on corrosion initiation difficult to assess. Future studies could be facilitated by using highly diluted NaCl solutions.

The multi-analytical approach employed in this doctoral thesis provides a generic tool box that is applicable for detailed studies to assess the relation between corrosion initiation and complex microstructures of any alloy systems, e.g. Zn-Al-Mg alloy coatings on steel.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

65  

Acknowledgements

The work presented within this thesis would not be possible without the help and support from many people. I would like to express my sincere gratitude to those have made contributions to my doctoral studies during the past four years:

Prof. Inger Odnevall Wallinder, my main supervisor; I really appreciate your professional supervision on me with immense supporting, endless patient and continuous encouragement. Your extensive knowledge and brilliant suggestions always guide me to the right direction for my graduate study. Special thanks for the extraordinary efforts on finalizing my thesis.

Prof. Christofer Leygraf, my co-supervisor; thank you so much for providing the opportunity for me to join this great group. Your creative ideas and scientific comments are of great significance for the progress during my doctoral study and also the discussion with you, no matter on science or life, always inspire and encourage me and help me grow up.

Prof. An Lin, Prof. Dihua Wang and Prof. Fuxing Gan, my former supervisors in Wuhan University, China; I am sincerely grateful for supporting and helping me to study abroad. Special thanks to Prof. An Lin for immense help on my future career.

Prof. Jiazhu Li; many thanks for introducing me to Christofer and helping me to get this precious chance to start a new journey.

Prof. Jinshan Pan; thank you a lot for not only sharing your professional knowledge in science but also encouragement in life.

Ping Qiu and Klara Midander; great thanks for guiding me into my doctoral study and teaching me the basic knowledge and the operation of several techniques, which makes my first year here much easier.

Prof. Kevin Ogle, Prof. Polina Volovitch and Thanh-Nam Vu at Ecole Nationale Supérieure de Chimie de Paris, France, my co-authors in the first paper; thanks for the scientific input of AESEC measurements and invaluable discussions.

Sara Goidanich at Politecnico di Milano, Italy and Gunilla Herting, my co-authors in the copper paper; thanks for your important contribution. Special thanks to Gunilla Herting for AAS training and great helping with different issues.

Rodrigo Robinson and Birgit Brandner at SP Technical Research Institute, Sweden for ESEM and CRM training and also informative discussions.

66

Fan Zhang, Mattias Forslund, Eleonora Bettini and Majid Sababi for invaluable help with SKPFM measurements and suggestions.

Oscar Karlsson, Niklas Pettersson and Mats Randelius at Swerea KIMAB, Sweden for effort on FEG-SEM, FIB-SEM and GDOES measurements.

Gerald Luckeneder at VoestAlpine, Austria for helping with the N-VDA test.

Neda Mazinanian, Tao Jiang and Xin Wang for kindly helping with the roof work.

Xiaoyan Liu, Olga Krivosheeva, Jonathan Liljeblad and Zahra Besharat, my present and former room-mates; thank you for sharing an office with me.

All the present and past colleagues in the Div. Surface and Corrosion Science; many thanks for creating the friendly working atmosphere and helping with different matters.

All my Chinese friends in Stockholm, especially those living in Kungshamra; great thanks for creating an enjoyable environment and relax times. Life would be boring without you.

CSC, the China Scholarship Council, RFCS, European Union’s Research Fund for Coal and Steel research programme, and ECI, the European Copper Institute are gratefully acknowledged for financial support.

My big family, both in Zhang and Yang; thank you so much for your endless understanding and support.

Miao Yang, my husband; I couldn’t be happier with your presence in my life. Your love, consideration and unlimited support mean the world to me.

I would like to express my sincere appreciation to my father, Bingshu Zhang, and my mother, Libai Xiong (1956-2006), for loving and supporting me in all situations throughout all years as well as guiding me to become a strong and independent person.

 

 

 

 

 

 

 

67  

References

[1] C. Leygraf, T. Graedel, Atmospheric corrosion, John Wiley & Sons., New York, 2000. [2] S. Feliu, M. Morcillo, B. Chico, Effect of distance from sea on atmospheric corrosion rate, Corros., 55 (1999) 883-891. [3] C. Leygraf, Atmospheric corrosion, J.O. P. Marcus (Ed.) Corrosion mechanisms in theory and practice, Marcel Dekker, Inc., New York, 1995, 421-455. [4] V. S. Sastri, E. Ghali, M. Elboujdaini, Practical Solutions, Corrosion Prevention and Protection, John Wiley & Sons, Ltd, 2012, 461-551. [5] W. H. J. Vernon, Effect of sulphur dioxide on the atmospheric corrosion of copper, Trans. Faraday Soc., 27 (1933) 255-277. [6] J. Hedberg, A molecular view of initial atmospheric corrosion - In situ surface studies of zinc based on vibrational spectroscopy, Ph.D. thesis, KTH Royal Institute of Technology, Stockholm, 2009. [7] S. Bertling, Corrosion-induced metal runoff from external constructions and its environmental interaction - A combined field and laboratory investigation of Zn, Cu, Cr and Ni for risk assessment, Ph.D. thesis, KTH Royal Institute of Technology, Stockholm, 2005. [8] Y. Hedberg, J. Hedberg, G. Herting, S. Goidanich, I. Odnevall Wallinder, Critical review: Copper runoff from outdoor copper surfaces at atmospheric conditions, Environ. Sci. Technol., 48 (2013) 1372-1381. [9] Z. Y. Chen, The role of particles on initial atmospheric corrosion of copper and zinc - Lateral distribution, secondary spreading and CO2 -/ SO2 - influence, Ph.D. thesis, KTH Royal Institute of Technology, Stockholm, 2005. [10] P. Qiu, Quantified in situ analysis of initial atmospheric corrosion - surface heterogenity galvanic effects and corrosion product distribution on zinc, brass and Galvalume, Ph.D. thesis, KTH Royal Institute of Technology Stockholm, 2011. [11] Corrosion: Fundamentals, testing, and protection, S.D. Cramer, B.S. Covino, Jr. (Eds.) ASM Handbook, ASM International, 2003. [12] Z. Y. Chen, D. Persson, F. Samie, S. Zakipour, C. Leygraf, Effect of carbon dioxide on sodium chloride-induced atmospheric corrosion of copper, J. Electrochem. Soc., 152 (2005) B502-B511. [13] Z. Y. Chen, D. Persson, A. Nazarov, S. Zakipour, D. Thierry, C. Leygraf, In situ studies of the effect of CO2 on the initial NaCl-induced atmospheric corrosion of copper, J. Electrochem. Soc., 152 (2005) B342-B351. [14] P. Qiu, C. Leygraf, I. Odnevall Wallinder, Evolution of corrosion products and metal release from Galvalume coatings on steel during short and long-term atmospheric exposures, Mater. Chem. Phys., 133 (2012) 419-428. [15] N. LeBozec, D. Thierry, M. Rohwerder, D. Persson, G. Luckeneder, L. Luxem, Effect of carbon dioxide on the atmospheric corrosion of Zn-Mg-Al coated steel, Corros. Sci., 74 (2013) 379-386. [16] S. Velu, V. Ramkumar, A. Narayanan, C. Swamy, Effect of interlayer anions on the physicochemical properties of zinc-aluminium hydrotalcite-like compounds, J. Mater. Sci., 32 (1997) 957-964. [17] M. Morcillo, B. Chico, L. Mariaca, E. Otero, Salinity in marine atmospheric corrosion: its dependence on the wind regime existing in the site, Corros. Sci., 42 (2000) 91-104. [18] Z. Y. Chen, D. Persson, C. Leygraf, Initial NaCl-particle induced atmospheric corrosion of zinc - Effect of CO2 and SO2, Corros. Sci., 50 (2008) 111-123. [19] T. H. Muster, A. Bradbury, A. Trinchi, I.S. Cole, T. Markley, D. Lau, S. Dligatch, A. Bendavid, P. Martin, The atmospheric corrosion of zinc: The effects of salt concentration, droplet size and droplet shape, Electrochim. Acta, 56 (2011) 1866-1873. [20] Q. Qu, C. Yan, Y. Wan, C. Cao, Effects of NaCl and SO2 on the initial atmospheric corrosion of zinc, Corros. Sci., 44 (2002) 2789-2803.

68

[21] D. Liang, H.C. Allen, G.S. Frankel, Z.Y. Chen, R.G. Kelly, Y. Wu, B.E. Wyslouzil, Effects of sodium chloride particles, ozone, UV, and relative humidity on atmospheric corrosion of silver, J. Electrochem. Soc., 157 (2010) C146-C156. [22] Z. Y. Chen, S. Zakipour, D. Persson, C. Leygraf, Effect of sodium chloride particles on the atmospheric corrosion of pure copper, Corros., 60 (2004) 479-491. [23] J. Itoh, T. Sasaki, T. Ohtsuka, The influence of oxide layers on initial corrosion behavior of copper in air containing water vapor and sulfur dioxide, Corros. Sci., 42 (2000) 1539-1551. [24] S. Oesch, M. Faller, Environmental effects on materials: The effect of the air pollutants SO2, NO2, NO and O3 on the corrosion of copper, zinc and aluminium. A short literature survey and results of laboratory exposures, Corros. Sci., 39 (1997) 1505-1530. [25] W.-B. Chen, P. Chen, H.Y. Chen, J. Wu, W.-T. Tsai, Development of Al-containing zinc-rich paints for corrosion resistance, Appl. Surf. Sci., 187 (2002) 154-164. [26] GALVALUME® sheet steel, U. S. Steel Corporation, 2005. [27] X. G. Zhang, Corrosion and elctrochemistry of zinc, Plenum Press, New York, 1996. [28] S. T. Bluni, A. R. Marder, J. I. Goldstein, Surface characterization of hot-dip Galfan coatings, Mater. Charact., 33 (1994) 93-97. [29] J. Murray, The Al−Zn (Aluminum-Zinc) system, J. Phase Equilib., 4 (1983) 55-73. [30] Steels coated with Galfan zinc-aluminium alloy - Zinc coatings and thin organic coatings ArcelorMittal, 2008. [31] A. R. Marder, The metallurgy of zinc-coated steel, Prog. Mater. Sci., 45 (2000) 191-271. [32] J. Elvins, J.A. Spittle, D.A. Worsley, Relationship between microstructure and corrosion resistance in Zn-Al alloy coated galvanised steels, Corros. Eng., Sci. Technol., 38 (2003) 197-204. [33] M. Żelechower, J. Kliś, E. Augustyn, J. Grzonka, D. Stróż, T. Rzychoń, H. Woźnica, The microstructure of annealed Galfan coating on steel substrate, Arch. Metall. Mater., 57 (2012) 517-523. [34] G. A. López, E. J. Mittemeijer, B. B. Straumal, Grain boundary wetting by a solid phase; microstructural development in a Zn-5 wt% Al alloy, Acta Mater., 52 (2004) 4537-4545. [35] A. R. Moreira, Z. Panossian, P. L. Camargo, M. F. Moreira, I. C. da Silva, J. E. R. de Carvalho, Zn/55Al coating microstructure and corrosion mechanism, Corros. Sci., 48 (2006) 564-576. [36] E. Palma, J. M. Puente, M. Morcillo, The atmospheric corrosion mechanism of 55% Al-Zn coating on steel, Corros. Sci., 40 (1998) 61-68. [37] J. Selverian, M. Notis, A. Marder, The microstructure of 55 w/o Al-Zn-Si (Galvalume) hot dip coatings, J. Mater. Eng., 9 (1987) 133-140. [38] J. Elvins, J.A. Spittle, D. A. Worsley, Microstructural changes in zinc aluminium alloy galvanising as a function of processing parameters and their influence on corrosion, Corros. Sci., 47 (2005) 2740-2759. [39] A. E. Ares, L. M. Gassa, Corrosion susceptibility of Zn-Al alloys with different grains and dendritic microstructures in NaCl solutions, Corros. Sci., 59 (2012) 290-306. [40] W. R. Osório, C. M. Freire, A. Garcia, The effect of the dendritic microstructure on the corrosion resistance of Zn-Al alloys, J. Alloys Compd., 397 (2005) 179-191. [41] A guide to working with copper and copper alloys, Copper Development Association, New York. [42] Metallography and microstructures, G.F. Vander Voort (Ed.) ASM Handbook, ASM International 2004. [43] G. Pantazopoulos, A. Vazdirvanidis, Characterization of the microstructural aspects of machinable α-β phase brass, Microsc. Anal., 22 (2008) 13-16. [44] J. Dutkiewicz, F. Masdeu, P. Malczewski, A. Kukula, Microstructure and properties of α+β brass after ECAP processing, Arch. Mater. Sci. Eng., 39 (2009) 80-83. [45] Production data sheet - TECU® Gold, KME Germany GmbH & Co. KG, 2012. [46] E. Almeida, M. Morcillo, B. Rosales, Atmospheric corrosion of zinc Part 1: Rural and urban atmospheres, Brit. Corros. J., 35 (2000) 284-288. [47] E. Almeida, M. Morcillo, B. Rosales, Atmospheric corrosion of zinc Part 2: Marine atmospheres, Brit. Corros. J., 35 (2000) 289-296.

69  

[48] D. de la Fuente, J. G. Castaño, M. Morcillo, Long-term atmospheric corrosion of zinc, Corros. Sci., 49 (2007) 1420-1436. [49] T. E. Graedel, Corrosion mechanisms for zinc exposed to the atmosphere, J. Electrochem. Soc., 136 (1989) 193C-203C. [50] M. Morcillo, B. Chico, D. de la Fuente, E. Almeida, G. Joseph, S. Rivero, B. Rosales, Atmospheric corrosion of reference metals in Antarctic sites, Cold Regions Sci. Tech., 40 (2004) 165-178. [51] J. Castaño, C. Arroyave, M. Morcillo, Characterization of atmospheric corrosion products of zinc exposed to SO2 and NO2 using XPS and GIXD, J. Mater. Sci., 42 (2007) 9654-9662. [52] A. K. Neufeld, I. S. Cole, A. M. Bond, S. A. Furman, The initiation mechanism of corrosion of zinc by sodium chloride particle deposition, Corros. Sci., 44 (2002) 555-572. [53] I. Odnevall, C. Leygraf, Reaction sequences in atmospheric corrosion of zinc, W.W. Kirk, H.H. Lawson (Eds.) ASTM STP 1239 Atmospheric Corrosion, American Society for Testing and Materials, Philadelphia, PA, 1995, 15. [54] F. Zhu, X. Zhang, D. Persson, D. Thierry, In situ infrared reflection absorption spectroscopy studies of confined zinc surfaces exposed under periodic wet-dry conditions, Electrochem. Solid-State Lett., 4 (2001) B19-B22. [55] T. Aastrup, C. Leygraf, Simultaneous infrared reflection absorption spectroscopy and quartz crystal microbalance measurements for in situ studies of the metal/atmosphere interface, J. Electrochem. Soc., 144 (1997) 2986-2990. [56] N. S. Azmat, K. D. Ralston, B. C. Muddle, I. S. Cole, Corrosion of Zn under acidified marine droplets, Corros. Sci., 53 (2011) 1604-1615. [57] J. E. Svensson, L. G. Johansson, A laboratory study of the initial stages of the atmospheric corrosion of zinc in the presence of NaCl; Influence of SO2 and NO2, Corros. Sci., 34 (1993) 721-740. [58] J. E. Svensson, L. G. Johansson, The temperature-dependence of the SO2-induced atmospheric corrosion of zinc; a laboratory study, Corros. Sci., 38 (1996) 2225-2233. [59] S. Li, L. H. Hihara, Aerosol salt particle deposition on metals exposed to marine environments: A study related to marine atmospheric corrosion, J. Electrochem. Soc., 161 (2014) C268-C275. [60] J. J. Friel, Atmospheric corrosion products on Al, Zn and AlZn metallic coatings, Corros., 42 (1986) 422-426. [61] S. A. Matthes, S. D. Cramer, S. J. Bullard, B. S. Covino, B. S. Covino Jr, G.R. Holcomb, Atmospheric corrosion and precipitation runoff from zinc and zinc alloy surfaces, NACE Corrosion Conference, Paper No. 3598. (2003). [62] I. Odnevall Wallinder, C. Leygraf, C. Karlén, D. Heijerick, C.R. Janssen, Atmospheric corrosion of zinc-based materials: runoff rates, chemical speciation and ecotoxicity effects, Corros. Sci., 43 (2001) 809-816. [63] J. C. Zoccola, H. E. Townsend, A. R. Borzillo, J. B. Horton, Atmospheric corrosion resistance of aluminum-zinc alloy-coated steel, S.K. Coburn (Ed.) Atmospheric Factors Affecting the Corrosion of Engineering Materials, ASTM STP 646, American Society for Testing and Materials, Philadelphia, Pennsylvania, 1978. [64] E. Palma, B. Fernández, M. Morcillo, Long-term atmospheric cathodic protection of 55% Al-Zn coating on steel and its comparison with galvanized steel, Mater. Corros., 48 (1997) 765-769. [65] P. Volovitch, T. N. Vu, C. Allély, A. Abdel Aal, K. Ogle, Understanding corrosion via corrosion product characterization: II. Role of alloying elements in improving the corrosion resistance of Zn-Al-Mg coatings on steel, Corros. Sci., 53 (2011) 2437-2445. [66] D. Persson, D. Thierry, N. LeBozec, Corrosion product formation on Zn55Al coated steel upon exposure in a marine atmosphere, Corros. Sci., 53 (2011) 720-726. [67] Y. Li, Formation of nano-crystalline corrosion products on Zn–Al alloy coating exposed to seawater, Corros. Sci., 43 (2001) 1793-1800. [68] I. Odnevall Wallinder, W. He, P. E. Augustsson, C. Leygraf, Characterization of black rust staining of unpassivated 55% Al-Zn alloy coatings. Effect of temperature, pH and wet storage, Corros. Sci., 41 (1999) 2229-2249.

70

[69] H. E. Townsend, J. C. Zoccola, Chromate Passivation Protection of Zn- and Al-Zn-Coated Steel Sheet Against Wet-Storage Stain, J. Electrochem. Soc., 125 (1978) 1290-1292. [70] A. Krätschmer, I. Odnevall Wallinder, C. Leygraf, The evolution of outdoor copper patina, Corros. Sci., 44 (2002) 425-450. [71] D. de la Fuente, J. Simancas, M. Morcillo, Morphological study of 16-year patinas formed on copper in a wide range of atmospheric exposures, Corros. Sci., 50 (2008) 268-285. [72] H. Strandberg, L.G. Johansson, Some aspects of the atmospheric corrosion of copper in the presence of sodium chloride, J. Electrochem. Soc., 145 (1998) 1093-1100. [73] M. Watanabe, E. Toyoda, T. Handa, T. Ichino, N. Kuwaki, Y. Higashi, T. Tanaka, Evolution of patinas on copper exposed in a suburban area, Corros. Sci., 49 (2007) 766-780. [74] L. Núñez, E. Reguera, F. Corvo, E. González, C. Vazquez, Corrosion of copper in seawater and its aerosols in a tropical island, Corros. Sci., 47 (2005) 461-484. [75] S. Goidanich, J. Brunk, G. Herting, M.A. Arenas, I. Odnevall Wallinder, Atmospheric corrosion of brass in outdoor applications: Patina evolution, metal release and aesthetic appearance at urban exposure conditions, Sci. Total Environ., 412-413 (2011) 46-57. [76] J. Sandberg, I. Odnevall Wallinder, C. Leygraf, N. Le Bozec, Corrosion-induced copper runoff from naturally and pre-patinated copper in a marine environment, Corros. Sci., 48 (2006) 4316-4338. [77] M. Morcillo, E. Almeida, M. Marrocos, B. Rosales, Atmospheric corrosion of copper in Ibero-America, Corros., 57 (2001) 967-980. [78] C. A. C. Sequeira, Copper and Copper Alloys, R.W. Revie (Ed.) Uhlig´s Corrosion Handbook, John Wiley & Sons, 2011. [79] M. Ghoniem, The characterization of a corroded Egyptian bronze statue and a study of the degradation phenomena, Int. J. Conserv. Sci., 2 (2011) 95-108. [80] D. A. Scott, Chlorides and basic chlorides, D.A. Scott (Ed.) Copper and bronze in art - corrosion, colorants, conservation, Getty Publications, Los Angeles, 2002, 122-144. [81] D. A. Scott, A review of copper chlorides and related salts in bronze corrosion and as painting pigments, Studies in Conservation, 45 (2000) 39-53. [82] M. Serghini-Idrissi, M.C. Bernard, F.Z. Harrif, S. Joiret, K. Rahmouni, A. Srhiri, H. Takenouti, V. Vivier, M. Ziani, Electrochemical and spectroscopic characterizations of patinas formed on an archaeological bronze coin, Electrochim. Acta, 50 (2005) 4699-4709. [83] F. Ospitali, C. Chiavari, C. Martini, E. Bernardi, F. Passarini, L. Robbiola, The characterization of Sn-based corrosion products in ancient bronzes: a Raman approach, J. Raman Spectrosc., 43 (2012) 1596-1603. [84] P. Qiu, C. Leygraf, Initial oxidation of brass induced by humidified air, Appl. Surf. Sci., 258 (2011) 1235-1241. [85] I. Odnevall Wallinder, Y. Hedberg, P. Dromberg, Storm water runoff measurements of copper from a naturally patinated roof and from a parking space. Aspects on environmental fate and chemical speciation, Water Res., 43 (2009) 5031-5038. [86] I. Odnevall Wallinder, P. Verbiest, W. He, C. Leygraf, Effects of exposure direction and inclination on the runoff rates of zinc and copper roofs, Corros. Sci., 42 (2000) 1471-1487. [87] I. Odnevall Wallinder, C. Leygraf, C. Karlén, D. Heijerick, C.R. Janssen, Atmospheric corrosion of zinc-based materials: Runoff rates, chemical speciation and ecotoxicity effects, Corros. Sci., 43 (2001) 809-816. [88] W. He, I. Odnevall Wallinder, C. Leygraf, A comparison between corrosion rates and runoff rates from new and aged copper and zinc as roofing material, Water Air Soil Poll., 1 (2001) 67-82. [89] J. Sandberg, I. Odnevall Wallinder, C. Leygraf, N. Le Bozec, Corrosion-induced zinc runoff from construction materials in a marine environment, J. Electrochem. Soc., 154 (2007) C120-C131. [90] S. D. Cramer, S. A. Matthes, B. S. Covino, S. J. Bullard, G. R. Holcomb, Environmental factors affecting the atmospheric corrosion of copper, H.E. Townsend (Ed.) Outdoor Atmospheric Corrosion, American Society Testing and Materials, West Conshohocken, 2002, 245-264. [91] I. Odnevall Wallinder, C. Leygraf, Seasonal variations in corrosion rate and runoff rate of copper roofs in an urban and a rural atmospheric environment, Corros. Sci., 43 (2001) 2379-2396.

71  

[92] S. Bertling, I. Odnevall Wallinder, D. B. Kleja, C. Leygraf, Long-term corrosion-induced copper runoff from natural and artificial patina and its environmental impact, Environ. Toxicol. Chem., 25 (2006) 891-898. [93] Basic corrosion technology for scientists and engineers, 2nd ed., Institute of Materials, 2001. [94] Uhlig's Corrosion Handbook third ed., John Wiley & Sons, Inc., 2011. [95] Corrosion: Materials, S. D. Cramer, B. S. Covino, Jr. (Eds.) ASM Handbook, ASM International, 2003. [96] G. -L. Song, Corrosion prevention of magnesium alloys, Woodhead Publishing, 2013. [97] V. Guillaumin, P. Schmutz, G. S. Frankel, Characterization of corrosion interfaces by the scanning Kelvin probe force microscopy technique, J. Electrochem. Soc., 148 (2001) B163-B173. [98] P. Schmutz, G. S. Frankel, Characterization of AA2024-T3 by scanning Kelvin probe force microscopy, J. Electrochem. Soc., 145 (1998) 2285-2295. [99] A. E. Coy, F. Viejo, P. Skeldon, G. E. Thompson, Susceptibility of rare-earth-magnesium alloys to micro-galvanic corrosion, Corros. Sci., 52 (2010) 3896-3906. [100] A. Davoodi, J. Pan, C. Leygraf, S. Norgren, The role of intermetallic particles in localized corrosion of an aluminum alloy studied by SKPFM and integrated AFM/SECM, J. Electrochem. Soc., 155 (2008) C211-C218. [101] N. -Y. Tang, Y. Liu, Corrosion performance of aluminum-containing zinc coatings, ISIJ Int., 50 (2010) 455-462. [102] L. Yang, Y. Zhang, X. Zeng, Z. Song, Corrosion behaviour of superplastic Zn–Al alloys in simulated acid rain, Corros. Sci., 59 (2012) 229-237. [103] M. Mokaddem, P. Volovitch, K. Ogle, The anodic dissolution of zinc and zinc alloys in alkaline solution. I. Oxide formation on electrogalvanized steel, Electrochim. Acta, 55 (2010) 7867-7875. [104] K. Ogle, A. Tomandl, N. Meddahi, M. Wolpers, The alkaline stability of phosphate coatings I: ICP atomic emission spectroelectrochemistry, Corros. Sci., 46 (2004) 979-995. [105] K. Ogle, M. Mokaddem, P. Volovitch, Atomic emission spectroelectrochemistry applied to dealloying phenomena II. Selective dissolution of iron and chromium during active-passive cycles of an austenitic stainless steel, Electrochim. Acta, 55 (2010) 913-921. [106] M. Mokaddem, P. Volovitch, F. Rechou, R. Oltra, K. Ogle, The anodic and cathodic dissolution of Al and Al-Cu-Mg alloy, Electrochim. Acta, 55 (2010) 3779-3786. [107] K. Ogle, J. Baeyens, J. Swiatowska, P. Volovitch, Atomic emission spectroelectrochemistry applied to dealloying phenomena: I. The formation and dissolution of residual copper films on stainless steel, Electrochim. Acta, 54 (2009) 5163-5170. [108] K. Ogle, M. Serdechnova, M. Mokaddem, P. Volovitch, The cathodic dissolution of Al, Al2Cu, and Al alloys, Electrochim. Acta, 56 (2011) 1711-1718. [109] G. Herting, S. Goidanich, I. Odnevall Wallinder, C. Leygraf, Corrosion-induced release of Cu and Zn into rainwater from brass, bronze and their pure metals. A 2-year field study, Environ. Monit. Assess., 144 (2008) 455-461. [110] M. Forslund, C. Leygraf, C. J. Lin, J. S. Pan, Radial spreading of localized corrosion-iduced selective leaching on alpha-brass in dilute NaCl solution, Corros., 69 (2013) 468-476. [111] H. H. Rehan, N. A. Al-Moubarak, H. A. Al-Rafai, A model for prolonged dezincification of α-brasses in 3% sodium chloride buffer solutions at different pH values, Mater. Corros., 52 (2001) 677-684. [112] I. Odnevall Wallinder, X. Zhang, S. Goidanich, N. Le Bozec, G. Herting, C. Leygraf, Corrosion and runoff rates of Cu and three Cu-alloys in marine environments with increasing chloride deposition rate, Sci. Total Environ., 472 (2014) 681-694. [113] M. Forslund, C. Leygraf, P. M. Claesson, C. J. Lin, J. S. Pan, Micro-galvanic corrosion effects on patterned copper-zinc samples during exposure in humidified air containing formic acid, J. Electrochem. Soc., 160 (2013) C423-C431. [114] M. Dowsett, A. Adriaens, C. Martin, L. Bouchenoire, The use of synchrotron x-rays to observe copper corrosion in real time, Anal. Chem., 84 (2012) 4866-4872.

72

[115] A. Adriaens, M. Dowsett, G. Jones, K. Leyssens, S. Nikitenko, An in-situ x-ray absorption spectroelectrochemistry study of the response of artificial chloride corrosion layers on copper to remedial treatment, J. Anal. At. Spectrom., 24 (2009) 62-68. [116] G. Herting, I. Odnevall Wallinder, C. Leygraf, A comparison of release rates of Cr, Ni and Fe from stainless steel alloys and the pure metals exposed to simulated rain events, J. Electrochem. Soc., 152 (2005) B23-29. [117] ASTM standard D1141-98, Standard practice for the preparation of substitute ocean water, 2003. [118] ISO 17752:2012, Corrosion of metals and alloys - Procedures to determine and estimate runoff rates of metals from materials as a result of atmospheric corrosion, 2012. [119] K. -H. Stellnberger, S. Geisler, Neuer laborkorrosionstest - relevante Ergebnisse, Journal für Oberflächentechnik, 49 (2009) 32ff. [120] C. Allély, A. Coffigny, G. Luckeneder, S. Reiter, S. Krebs, F. Beier, K. Ogle, C. Leygraf, I. Odnevall Wallinder, Corrosion of heterogeneous metal-metal assembling in the automotive industry, publishable report of AutoCorr project (RFS-CT-2009-00015), Arcelor Research Maizières-les-Metz, VASL Linz, DOC Dortmund, SZMF Salzgitter, ENSCP Paris, KTH Stockholm, 2013. [121] P. J. Goodhew, J. Humphreys, R. Beanland, Electron microscopy and analysis, 3rd ed., Taylor & Francis, London, 2001. [122] Y. Leng, Materials characterization: Introduction to microscopic and spectroscopic methods, John Wiley & Sons, Singapore, 2010. [123] J. Goldstein, D. E. Newbury, D. C. Joy, C. E. Lyman, P. Echlin, E. Lifshin, L. Sawyer, J.R. Michael, Scanning electron microscopy and x-ray microanalysis, 3rd ed., Springer, 2007. [124] L. Khouchaf, Gaseous scanning electron microscope (GSEM): Applications and improvement V. Kazmiruk (Ed.) Scanning electron microscopy, InTech, 2012. [125] P. R. Munroe, The application of focused ion beam microscopy in the material sciences, Mater. Charact., 60 (2009) 2-13. [126] V. G. M. Sivel, J. van den Brand, W. R. Wang, H. Mohdadi, F. D. Tichelaar, P. F. A. Alkemade, H. W. Zandbergen, Application of the dual-beam FIB/SEM to metals research, J. Microsc-Oxford, 214 (2004) 237-245. [127] Materials characterization, A.I.H. Committee (Ed.) ASM Handbook, ASM International USA, 1986. [128] M. Johnson, Vibrational sum frequency and infrared reflection/absorption spectroscopy studies of the air/liquid and liquid/metal interfaces, Ph.D. thesis, KTH Royal Institute of Technology, Stockholm, 2005. [129] G. J. Puppels, Confocal Raman microspectroscopy, W.T. Mason (Ed.) Fluorescent and luminescent probes for biological activity, Academic Press, London, 1999, 377-406. [130] B. C. Giessen, G. E. Gordon, X-ray diffraction: new high-speed technique based on x-ray spectrography, Science, 159 (1968) 973-975. [131] P. Dutta, Grazing incidence x-ray diffraction, Curr. Sci., 78 (2000) 1478-1483. [132] W. Melitz, J. Shen, A. C. Kummel, S. Lee, Kelvin probe force microscopy and its application, Surf. Sci. Rep., 66 (2011) 1-27. [133] B. S. Tanem, G. Svenningsen, J. Mårdalen, Relations between sample preparation and SKPFM Volta potential maps on an EN AW-6005 aluminium alloy, Corros. Sci., 47 (2005) 1506-1519. [134] J. H.W. de Wit, Local potential measurements with the SKPFM on aluminium alloys, Electrochim. Acta, 49 (2004) 2841-2850. [135] C. S. Fadley, X-ray photoelectron spectroscopy: Progress and perspectives, J. Electron. Spectr. Relat. Phenom., 178-179 (2010) 2-32. [136] M. R. Webb, V. Hoffmann, G. M. Hieftje, Surface elemental mapping using glow discharge - optical emission spectrometry, Spectrochim. Acta B, 61 (2006) 1279-1284. [137] R. García, A. P. Báez, Atomic absorption spectrometry (AAS), M. A. Farrukh (Ed.) Atomic absorption spectroscopy, Intech, 2012.

73  

[138] X. Zhang, T. -N. Vu, P. Volovitch, C. Leygraf, K. Ogle, I. Odnevall Wallinder, The initial release of zinc and aluminum from non-treated Galvalume and the formation of corrosion products in chloride containing media, Appl. Surf. Sci., 258 (2012) 4351-4359. [139] X. Zhang, C. Leygraf, I. Odnevall Wallinder, Selected area visualization by FIB-milling for corrosion-microstructure analysis with submicron resolution, Mater. Lett., 98 (2013) 230-233. [140] D. Wu, Y. Jiang, J. Liu, Y. Yuan, J. Wu, K. Jiang, D. Xue, Template route to chemically engineering cavities at nanoscale: A case study of Zn(OH)2 template, Nanoscale Res Lett, 5 (2010) 1779-1787. [141] H. W. Lee, B. K. Park, M. Y. Tian, J. M. Lee, Relationship between properties of pseudo-boehmite and its synthetic conditions, J. Ind. Eng. Chem., 12 (2006) 295-300. [142] N. Thomas, M. Rajamathi, Intracrystalline oxidation of thiosulfate-intercalated layered double hydroxides, Langmuir, 25 (2009) 2212-2216. [143] X. Zhang, C. Leygraf, I. Odnevall Wallinder, Atmospheric corrosion of Galfan coatings on steel in chloride-rich environments, Corros. Sci., 73 (2013) 62-71. [144] I. Odnevall, C. Leygraf, Formation of NaZn4Cl(OH)6SO4·6H2O in a marine atmosphere, Corros. Sci., 34 (1993) 1213-1229. [145] I. Odnevall, Characterization of corrosion products formed on rain sheltered Aluzink TM and aluminum in a rural and an urban atmosphere, Proc. 13th Int. Corr. Congr., 1 (1996) 1-8. [146] X. Zhang, I. Odnevall Wallinder, C. Leygraf, Mechanistic studies of corrosion product flaking on copper and copper-based alloys in marine environments, Corros. Sci., 85 (2014) 15-25. [147] J. Hedberg, N. Le Bozec, I. Odnevall Wallinder, Spatial distribution and formation of corrosion products in relation to zinc release for zinc sheet and coated pre-weathered zinc at an urban and a marine atmospheric condition, Mater. Corros., (2011) 1-9. [148] M. C. Hales, R. L. Frost, Synthesis and vibrational spectroscopic characterisation of synthetic hydrozincite and smithsonite, Polyhedron, 26 (2007) 4955-4962. [149] T. Ohtsuka, M. Matsuda, In situ Raman spectroscopy for corrosion products of zinc in humidified atmosphere in the presence of sodium chloride precipitate, Corros., 59 (2003) 407-413. [150] M. C. Bernard, A.H.-L. Goff, N. Phillips, In situ Raman study of the corrosion of zinc-coated steel in the presence of chloride, J. Electrochem. Soc., 142 (1995) 2167-2170. [151] H. Marchebois, S. Joiret, C. Savall, J. Bernard, S. Touzain, Characterization of zinc-rich powder coatings by EIS and Raman spectroscopy, Surf. Coat. Technol., 157 (2002) 151-161. [152] I. S. Cole, T. H. Muster, S. A. Furman, N. Wright, A. Bradbury, Products formed during the interaction of seawater droplets with zinc surfaces: I. Results from 1- and 2.5-day exposures, J. Electrochem. Soc., 155 (2008) C244-C255. [153] I. S. Cole, T. H. Muster, D. Lau, N. Wright, N. S. Azmat, Products formed during the interaction of seawater droplets with zinc surfaces: II. results from short exposures, J. Electrochem. Soc., 157 (2010) C213-C222. [154] I. Constantinides, A. Adriaens, F. Adams, Surface characterization of artificial corrosion layers on copper alloy reference materials, Appl. Surf. Sci., 189 (2002) 90-101. [155] V. Hayez, V. Costa, J. Guillaume, H. Terryn, A. Hubin, Micro Raman spectroscopy used for the study of corrosion products on copper alloys: study of the chemical composition of artificial patinas used for restoration purposes, Analyst, 130 (2005) 550-556. [156] T. Kosec, P. Ropret, A. Legat, Raman investigation of artificial patinas on recent bronze - part II: urban rain exposure, J. Raman Spectrosc., 43 (2012) 1587-1595. [157] R. L. Frost, Raman spectroscopy of selected copper minerals of significance in corrosion, Spectrochim. Acta A, 59 (2003) 1195-1204. [158] Z. Y. Chen, S. Zakipour, D. Persson, C. Leygraf, Combined effects of gaseous pollutants and sodium chloride particles on the atmospheric corrosion of copper, Corros., 61 (2005) 1022-1034. [159] T. H. Muster, I. S. Cole, The protective nature of passivation films on zinc: Surface charge, Corros. Sci., 46 (2004) 2319-2335. [160] I. S. Cole, N. S. Azmat, A. Kanta, M. Venkatraman, What really controls the atmospheric corrosion of zinc? Effect of marine aerosols on atmospheric corrosion of zinc, Int. Mater. Rev., 54 (2009) 117-133.

74

[161] S. Schürz, G. H. Luckeneder, M. Fleischanderl, P. Mack, H. Gsaller, A.C. Kneissl, G. Mori, Chemistry of corrosion products on Zn-Al-Mg alloy coated steel, Corros. Sci., 52 (2010) 3271-3279. [162] T. E. Graedel, Corrosion mechanisms for aluminum exposed to the atmosphere, J. Electrochem. Soc., 136 (1989) 204C-212C.