Corrosion-induced release of zinc and copper

in marine environments

Jan Sandberg

Licentiate thesis

Division of Corrosion Science

Department of Materials Science and Engineering

School of Industrial Engineering and Management

Royal Institute of Technology

SE-100 44 Stockholm, Sweden

Reviewer: Dan Persson, Ph.D., Corrosion and Metals Research Institute, Stockholm, Sweden

ISRN KTH/MSE--06/39--SE+CORR/AVH

ISBN 91-7178-364-4

Abstract

This licentiate study was initiated by copper, zinc and galvanized steel producers in Europe,

who felt a need to assess runoff rates of copper and zinc from the pure metals and commercial

products at marine exposure conditions. Their motive was the increasing concern in various

European countries and the on-going risk assessments of copper and zinc within the European

commission. Also the circumstance that available runoff rates so far, had been reported for

mainly urban exposure conditions, rather than marine. A collaboration was therefore

established with the French Corrosion Institute, which runs a marine test site in Brest, and a

set of vital questions were formulated. Their answers are the essence of this licentiate study.

Based on the ISO corrosivity classification and one-year exposures, the marine atmosphere of

Brest is fairly corrosive for zinc (class C3) and highly corrosive for copper (C4). Despite

higher corrosivity classifications for both metals in Brest compared to the urban site of

Stockholm, used as a reference site, nearly all runoff rates assessed for copper, zinc and their

commercial products were lower in Brest compared to Stockholm. This was attributed to a

higher surface wetting in Brest and concomitant higher removal rate of deposited chloride and

sulphate species from the marine-exposed surfaces. The comparison shows that measured

corrosion rates cannot be used to predict runoff rates, since different physicochemical

processes govern corrosion and runoff respectively.

For copper, the runoff rate in Brest was approximately 1.1 g m-2 yr-1 with cuprite (Cu2O) as

main patina constituent. During periods of very high chloride and sulphate deposition,

paratacamite (Cu2Cl(OH)3) formed which increased the runoff rate to 1.5 g m-2 yr-1. For zinc,

with hydrozincite (Zn5(CO3)2(OH)6) as the main patina constituent, the runoff rate was

relatively stable at 2.6 g m-2 yr-1 throughout the year, despite episodes of heavy chloride and

sulphate deposition.

The application of organic coatings of varying thickness on artificially patinated copper or on

different zinc-based products resulted in improved barrier properties and reduced runoff rates

that seem highly dependent on thickness. The thickest organic coating (150 µm thick), applied

on hot dipped galvanized steel, reduced the runoff rate by a factor of 100. No deterioration of

organic coatings was observed during the one-year exposures. Alloying zinc-based products

with aluminium resulted in surface areas enriched in aluminium and concomitant reduced zinc

runoff rates.

The release rate and bioavailability of copper from different anti-fouling paints into artificial

seawater was also investigated. It turned out that the release rate not only depends on the

copper concentration in the paint, but also on paint matrix properties and other released metal

constituents detected. Far from all copper was bioavailabe at the immediate release situation.

In all, the results suggest the importance of assessing the ecotoxic response of anti-fouling

paints not only by regarding the copper release, but rather through an integrated effect of all

matrix constituents.

Keywords: copper, zinc, marine environment, runoff rate, atmospheric corrosion, chloride

deposition, anti-fouling paints, chemical speciation, bioavailability, cupric ions

Preface

The following papers are included in this thesis:

I Corrosion-induced copper runoff from naturally and pre-patinated copper in a

marine environment

J. Sandberg, I Odnevall Wallinder, C. Leygraf, N. Le Bozec

Corrosion Science, in press

II Corrosion-induced zinc runoff from construction materials in a marine environment

J. Sandberg, I Odnevall Wallinder, C. Leygraf, N. Le Bozec

Submitted to Journal of the Electrochemical Society (April 2006)

III Release and chemical speciation of copper from anti-fouling paints with different

active copper compounds in artificial seawater

J. Sandberg, I Odnevall Wallinder, C. Leygraf, M. Virta

Submitted to Materials and Corrosion (April 2006)

Contents

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

1.1 Atmospheric corrosion of copper and zinc ...................................................... 2

1.2 Runoff rates of copper and zinc ....................................................................... 3

1.3 Copper release from anti-fouling paints .......................................................... 5

1.4 Chemical speciation and bioavailability of released copper ........................... 6

2. Experimental............................................................................................................ 7

2.1 Marine site and experimental setup................................................................ 7

2.1.1 Investigated materials ......................................................................... 8

2.1.2 Experimental approach....................................................................... 9

2.2 Copper release from anti-fouling paints .......................................................... 10

2.3 Copper bioavailability in runoff water and artificial sea water...................... 11

3. Summary .................................................................................................................. 12

3.1 Can corrosion rates be used to predict runoff rates at any given site? ......... 12

3.2 Does precipitation characteristics influence the runoff process? ................. 13

3.3 Do changes in patina composition influence the runoff rate?...................... 15

3.4 Do surface treatments and coatings influence the runoff rate?.................... 17

3.5 Do changes in alloy composition influence the runoff rate? ........................ 18

3.6 Do copper compounds of anti-fouling paints influence the copper

release rate? ................................................................................................... 20

3.7 How bioavailable is released copper in marine environments? .................... 22

3.8 Is the runoff rate different in a marine compared to an urban

environment? ................................................................................................. 24

4. Main conclusions ..................................................................................................... 27

5. Acknowledgements.................................................................................................. 28

6. References ................................................................................................................ 29

1. Introduction

Metal dispersion from diffuse sources in the society has received increased attention during

the last decade. The environmental awareness of corrosion-induced metal release from, e.g.

building materials, into urban storm water systems and ecological systems have resulted in

restrictions regarding the use of metals, due to potential adverse effects during transport into

soil and water bodies. Due to scarce corrosion-induced release data from metals and metal

alloy surfaces on external constructions in different environments, policies and restrictions are

often based on the precautionary principle and conservative deliberations.

Atmospheric corrosion of metals, such as zinc and copper, has been extensively studied in

terms of corrosion rates and patina formation in both field and laboratory investigations. Even

though the corrosion resistance of copper and zinc is well documented, there is a considerable

lack of knowledge related to patina formation and stability, and the metal release process

during precipitation or immersion.

The general aims of this thesis are to: i) provide an improved understanding of the metal

runoff process from copper and zinc-based materials exposed during one year in a marine

environment, and i) through a multi-analytical approach correlate changes in runoff rates to

patina formation, changes in barrier properties and deposition of corrosive species, and iii) to

elucidate the difference between the total copper release rate and the corresponding

bioavailable copper fraction from anti-fouling paints of different main copper compound.

The following 8 questions have been formulated to increase the knowledge of corrosion-

induced zinc and copper runoff from bare metals and alloys with and without surface

treatments or organic coatings, and copper bioavailability and chemical speciation in marine

environments.

1. Can the corrosion rate be used to predict the runoff rate at any given site?

2. Does precipitation characteristics influence the runoff process?

3. Do changes in patina composition influence the runoff rate?

4. Do surface treatments and coatings influence the runoff rate?

5. Do changes in alloy composition influence the runoff rate?

6. Do copper compounds of anti-fouling paints influence the copper release rate?

7. How bioavailable is released copper in marine environments?

8. Is the runoff rate different in a marine compared to an urban environment?

1

1.1 Atmospheric corrosion of zinc and copper

The corrosion process occurs as a part of the natural cycle of bringing the metal back to its

thermodynamically stable form as a mineral. This formation of adherent minerals on a metal

sheet result in a patina that often acts as a protective barrier between the atmosphere and the

metal substrate. The barrier properties and the development of the patina determine the

corrosion rate of the metal, i.e. the amount of the metal that is oxidized per unit surface area

and time period. When the patina has reached steady state thickness and chemical

composition, a process that may take years or even decades depending on prevailing

environmental conditions and material, the corrosion rate will reach steady-state, Figure 1.

[Odnevall-Wallinder et.al (1997), Verbiest et.al. (1997), He et.al. (2001b), Leuenberger et.al.

(2002), Faller et.al. (2005)].

20

40

60

80

100

120

0 5 10 15 20 25Exposure period / years

Corrosion rate

Rat

e / g

m-2

yr-1

Figure 1. Schematic description of the gradual decrease in the corrosion rate with time for zinc and

copper exposed at atmospheric conditions.

A prerequisite for a corrosion process is the presence of an aqueous ad-layer on the metal

surface, in which electrochemical reactions can take place. The aqueous ad-layer is normally

always present except at very low relative humidities [Marcus (2002)] or during dew, fog and

precipitation events. The water layer acts as solvent for gaseous constituent, e.g. SO2, NOx,

O2. Also aerosol particles, such as ammonium sulphate (common in urban and rural areas)

and sodium chloride (common in marine areas) deposit and dissolve or precipitate within the

water layer. The chemical composition of the patina depends on prevailing environmental

conditions, and the amount of deposited species.

2

Reaction sequences have been proposed for both copper and zinc exposed at atmospheric

conditions [Krätschmer et al (2002), Odnevall and Leygraf (1994)].

The intital patina component on copper in urban and rural environments is cuprite(Cu2O),

which forms an inner brown or black layer, often covered with a thicker and more porous

green layer of posnjakite (Cu4SO4(OH)6 x H2O) and brochantite (Cu4SO4(OH)6). Other copper

sulphates may also be part of the patina in more polluted environments [Krätschmer et al

(2002)]. In low SO2- and sulphate-polluted areas, cuprite is the main patina constituent

[Atrens et. al. (1996), Odnevall and Leygraf (1995), Odnevall Wallinder and Leygraf

(1997)]. In marine environments with high chloride deposition rates, a patina is evolved with

cuprite that with time, and repeated dissolution and precipitation sequences, is transformed

into a patina mainly composed of paratacamite. (Cu2Cl(OH)3) [Costas (1982), Veleva

et.al.(1995), Skennerton et. a. (1997), Nunez et. al.(2005)].

For zinc materials, hydrozincite (Zn5(CO3)2(OH)6) is the initial patina component in most

environments [Graedel (1987), Odnevall Walllinder and Leygraf (1994), He et. al. (2002),

Cramer et al. (1990)]. With time other corrosion products will precipitate or form due to

structural similarities [Odnevall Walllinder and Leygraf (1994)]. Basic zinc chloride,

simonkolleite (Zn5(OH)8Cl2xH2O) and a basic zinc chlorohydroxysulphate

(NaZn4SO4(OH)6*6H2O) are commonly detected in marine environments, whereas sites

polluted with high SO2 levels primarily show basic zinc sulphates (Zn4SO4(OH)6*nH2O (n=1-

5)) as main components of the patina [Odnevall and Leygraf (1993), Odnevall and Leygraf

(1994)].

1.2 Runoff rates of copper and zinc

The copper and zinc runoff process is dominated by chemical dissolution and erosion/wear of

corrosion products within the patina through a combination of deposited corrosive particles

and gases and the presence and impingement of water on the surface. This process, which

takes place at the interface between the patina and the environment, is fundamentally different

from the corrosion process, which is of an electrochemical nature and occurs at the interface

between the bulk metal and the patina. On a long-term perspective, months and years, the

runoff rate is relatively constant compared to the corrosion rate, which is highly time

3

dependent due to the gradual built-up of corrosion products [Odnevall-Wallinder and Leygraf

(1997), Verbiest et.al. (1997), He et.al. (2001b), Leuenberger et.al. (2002), Faller et.al.

(2005)]. The difference between the corrosion rate and the runoff rate is schematically

illustrated in Figure 2. With time, it is anticipated that the corrosion rate and the runoff rate

will be equal when the protective patina has reached its steady state thickness. The formation

of corrosion products will change the dissolution properties compared to the underlying metal

substrate.

20

40

60

80

100

120

0 5 10 15 20 25TimeExposure period / years

Corrosion rate

Runoff rate

Rat

e / g

m-2

yr-1

Figure 2. Schematic description of differences in corrosion rates and metal runoff rates on a long-term

perspective.

During a single rain event, the runoff rate is significantly higher during the initial rain portion

flushing the surface (first flush) compared to later periods of more constant release rate

(steady state) during the remaining rain event. The runoff process, and the magnitude of first

flush in particular, depends on a large number of interacting material and environmental

parameters including e.g. patina morphology, surface inclination and orientation, rain pH,

intensity and amount, duration of dry periods in-between rain episodes, and dry deposition of

gaseous pollutants and gases. The individual effect of these parameters on the runoff process

has been thoroughly studied in a laboratory investigation [He (2001a)].

Annual runoff rates of copper and zinc from different copper and zinc-based materials with

and without surface treatments or barriers have during recent years primarily been generated

for urban and rural environmental conditions [Bertling (2006b), Faller and Reiss (2005), He

(2001b), Verbiest et. al. (1997), Odnevall Wallinder et. al. (1998), Odnevall Wallinder et. al.

4

(2002), Veleva and Meraz (2005), Costa (1999), Matthes (2003), Quek and Förster (1993),

Boulanger and Nikolaidis (2003)]. However, only few studies are available that address the

marine influence on the runoff process of copper and zinc [Belghazi et al. (2002), Sullivan et.

al. (2002), Cramer et. al. (2002), Verbiest et. al. (1997)].

The improved knowledge on the runoff process has resulted in the generation of predictive

runoff rate models for copper and zinc, based on real runoff data and parameters that can

physically be explained [Cramer et.al. (1990), Odnevall Wallinder et. al. (1998), Odnevall

Wallinder et. al. (2004)].

1.3 Copper release from anti-fouling paints

Copper is a micronutrient for aquatic organisms, and commonly used in marine environments

due to its beneficial anti-fouling properties. However, elevated concentrations of the free

cupric ion can result in toxicity towards aquatic organisms. As a result, different copper

compounds of varying shape and chemical composition, such as copper oxides, copper

thiocyanates, and metallic copper flakes mixed with copper oxides, are often one major

component of anti-fouling paints. The paints are designed in such a way that copper, and other

metallic and organic species, are slowly released from the paint matrix in contact with water.

The release rate depends on many factors including paint matrix properties and the solubility

of the specific copper compound. Organic biocides, such as irgarol and dichlofluanid, are in

addition often included in anti-fouling paints to enhance the anti-fouling properties of the

specific paint [Voulvoulis (1999)].

Due to experimental difficulties to monitor the release rate of copper in-situ at field conditions

with reproducible results, and the need for fast and comparable test methods, a laboratory

release rate standard is available, ASTM D6442-03, [ASTM (2003)]. It should however be

noted that this protocol has not been validated for in-situ copper release rates from anti-

fouling paints [ASTM (2003)]. Generated results based on this standard, should be used with

caution within the framework of environmental risk assessments of copper. A recent study has

shown lower total copper release rates at in-situ field conditions compared to laboratory

measurements [Valkirs et. al. (2003)].

5

1.4 Chemical speciation and bioavailability of released copper

The aquatic ecotoxicity of copper depends on the chemical speciation and the bioavailability

of copper and copper complexes formed in natural environments. The free cupric ion,

Cu(H2O)62+, is generally considered to predominantly govern the toxic response [Landner and

Reuther (2004]. Complexation with inorganic ligands, such as chlorides, sulphates and

carbonates, reduces the toxicity. The pH dependence of this process is illustrated in Figure 3,

for copper in runoff water and artificial seawater. The prediction is made with the MinteqA2

water ligand model. It should be noticed that the presence of natural organic matter, e.g.

present in seawater would further reduce the bioavailable free cupric ion concentration [Donat

et.al. (1994), Bruland et. al (2000)].

A lower predicted free cupric ion fraction in artificial seawater is due to significantly higher

salt concentrations, compared to the runoff water. This increases the presence of inorganically

complexed copper species, such as copper chlorides and copper sulphates.

Free

cup

ric

ion

frac

tion

ofto

tal c

oppe

r co

ncen

trat

ion

/ %

pH

0

20

40

60

80

100

5.5 6.0 6.5 7.0 7.5 8.0 8.5

In artificial sea water

In runoff water

Figure 3. The pH dependence of the free cupric ion fraction of the total copper concentration in runoff

water (triangles) and in artificial sea water (circles) predicted by the MinteqA2 water ligand

model. No natural organic matter was used in the calculation.

6

2. Experimental

2.1 Marine site and experimental setup

One-year runoff rate measurements have been performed on copper and zinc materials at the

marine exposure site at Saint-Anne du Portzic, located on a small cape outside the city of

Brest, France. The French Corrosion Institute maintains the site. The test site is characterised

as moderately corrosive for zinc (ISO class C3) and highly corrosive for copper (ISO class

C4). All materials (sized 12x25cm) are fixed on Plexiglas fixtures equipped with inclined

gutters, into which impinging rainwater is collected and transported into polyethylene bottles.

The fixtures are exposed at 45° from the horizontal, facing south, in agreement to the ISO

9226 standard for corrosion rate measurements at atmospheric conditions. Environmental and

meteorological details are given in Papers I and II.

Runoffwater

PlexiglasfixturePanel

Runoffwater

PlexiglasfixturePanel

Figure 4. Test racks for runoff rate measurements at the marine exposure site at Saint-Anne du Portzic,

Brest, and a schematic illustration of the Plexiglas fixtures used for runoff water collection.

7

2.1.1 Investigated materials

Seven different copper based materials, naturally and pre-patinated copper sheet, with and

without surface treatments, were exposed during one year (June 2004-May 2005) at the

marine site, Figure 5. Initial patina components of the 200 year old naturally patinated copper

sheet, originating from Copenhagen, Denmark, were cuprite (Cu2O) beneath a greenish layer

of brochantite (Cu4SO4(OH)6). The bare pre-patinated surfaces were chemically oxidized and

mainly composed of a mixture of cuprite/tenorite (CuO) for Nordic Brown, and brochantite

for Nordic Green. In addition, Nordic Green surfaces were exposed with different surface

treatments of varying thickness, aimed for anti-graffiti protection. Additional details are given

in Paper I.

Coppersheet

Coppersheet

Copper sheet(Cu-new)

Naturally patinated200 years old

(Cu-200y)

Nordic Brownpre-patinated

(NB)

Nordic Greenpre-patinated

(NG)

Nordic Greenpre-patinated

+surface treatment

(NG1U)

Nordic Greenpre-patinated

+surface treatment

(NG1U)

Nordic Greenpre-patinated

+surface treatment

(NG2U)

Nordic Greenpre-patinated

+surface treatment

(NG2U)

Nordic Greenpre-patinated

+surface treatment

(NG2P)

Nordic Greenpre-patinated

+surface treatment

(NG2P)

Figure 5. Copper based materials with and without surface treatments exposed at unsheltered conditions

at the marine site from June 2004 to May 2005.

11 commercially available zinc-based materials were exposed between December 2004 and

November 2005, Figure 6. The exposure comprised bare, surface treated and coated materials

of continuously galvanized steel, hot-dipped galvanized steel, pre-weathered zinc sheet, and

zinc-aluminium alloys. Additional details are given in Paper II.

8

Continuouslygalvanized

+ chromate

(Galv)

Continuouslygalvanized

+ chromate

(Galv)

GalvalumeGalvalume

Zinc sheet(Zn)

Continuously galvanized

+ organic coating

(Galv E)

Hot dippedgalvanized(HD-galv)

Hot dippedgalvanized(HD-galv)

Hot dippedgalvanized

+ organic coating

(HD-galv D)

Hot dippedgalvanized

+ organic coating

(HD-galv D)

Hot dippedgalvanized

+ organic coating(HD-galv CD)

Hot dippedgalvanized

+ organic coating(HD-galv CD)

Pickled zinc sheet(PW Zn)

Pickled zinc sheet+

organic coating(PW Zn A)

Pickled zinc sheet+

organic coating(PW Zn B)

Pickled zinc sheet+

organic coating(PW Zn B)

GalfanGalfan

Figure 6. Zinc-based materials, with and without surface treatments and organic coatings, exposed at

unsheltered conditions at the marine site between December 2004 and November 2005.

2.1.2 Experimental approach

A multi-analytical approach, schematically illustrated in Figure 7, has been applied to provide

a comprehensive understanding of prevailing corrosion and runoff processes on exposed

copper- and zinc-based materials.

Start (copper)June 2004

Continuous metal runoff rate measurements (AAS) + runoff water characteristics (pH, volume, anions)

On-going exposure

Multi-analytical surface analysisEIS

(Barrier properties)

GDOES(Elemental depth profile)

XPS(Chemical state)

XRD(Crystalline phases)

SEM/EDS(Morphology)

0 2 4 26 52 weeks120 2 4 26 52 weeks12

Start (zinc)Dec 2004

Wetting angle(Wetting capacity) IRAS(Functional groups)

Corrosion rate measurements

Pourbaix diagramcalculations

Bio assay and chemical speciation in runoff water and artificial sea water

Figure 7. Experimental approach to provide an improved understanding of changes in corrosion and

runoff processes on copper and zinc-based materials during marine exposure.

9

The total rainwater volume impinging on the materials was collected continuously and

measured in terms of volume, runoff pH and total metal concentration using atomic

absorption spectroscopy (AAS). The concentration of anions, such as sulphate, chloride and

nitrate, in runoff water was measured by means of ion chromatography (IC).

The runoff rate investigation was conducted in parallel with a multi-analytical approach

aiming to provide information on changes in patina formation and barrier properties

(corrosion resistance) during the one-year exposure. These measurements were performed on

small samples (1 cm2) after 2, 4, 12, 26 and 52 weeks of exposure. Changes in barrier

properties due to patina formation, presence of surface treatments or coatings were measured

by means of electrochemical impedance spectroscopy (EIS). Different surface analytical

techniques was applied to provide the following information:

• X-ray photoelectron spectroscopy (XPS): Elemental and chemical state information

of the outermost surface layer,

• Infrared reflection spectroscopy (IRAS): Information of functional surface groups

and possible phase characterization,

• Scanning electron microscopy and energy dispersive spectroscopy (SEM/EDS):

Morphological studies and bulk composition analyses,

• X-ray powder diffraction (XRD): Identification of crystalline phases in the patina.

• Wetting angle measurements: Examination of the wetting capacity of the patina and

differences between similar materials, and

• Glow discharge optical spectroscopy (GDOES): Elemental depth profiling.

In addition, annual corrosion rate measurements were performed on pure zinc and copper,

respectively, for comparison with annual runoff rate measurements, and Pourbaix diagram

calculations [Puigdomenech, (2006)] were performed to provide information on stability

regions for corrosion products formed.

2.2 Copper release from anti-fouling paints - experimental approach

Copper release rates from seven different anti-fouling paints with different active copper

compound, Figure 8, were investigated in artificial sea water [ASTM (1998)]. A 5x5 cm large

Plexiglas surface was painted, dried and immersed in the water solution and rocked bi-linearly

10

during 14 days of exposure. After 1, 3, 6, 9 and 14 days the painted surfaces were removed

into fresh artificial seawater for a one-hour measurement of the copper release rate. The total

copper concentration was measured by Inductively Coupled Plasma – Sector Field Mass

Spectrometry (ICP-SFMS). The copper release rates were calculated according to ASTM

standard D6442-03 [ASTM (2003)]. More details are given in Paper III.

Figure 8. Main active copper compound present i he investigated anti-fouling paints.

ater and artificial seawater

he chemical speciation of copper at the immediate release situation was predicted using the

Copper thiocyanate[Cu(NCS)5]

Copper thiocyanate[Cu(NCS)25]

Copper oxide[CuO30/40-1]

Copper oxide[CuO30/40-2]

Copper oxide[CuO25/50]

Metalliccopper flakes

[Cu4]

Metallic copper flakes

[Cu18]

Copper thiocyanate[Cu(NCS)5]

Copper thiocyanate[Cu(NCS)25]

Copper oxide[CuO30/40-1]

Copper oxide[CuO30/40-2]

Copper oxide[CuO25/50]

Metalliccopper flakes

[Cu4]

Metallic copper flakes

[Cu18]

n t

2.3 Copper bioavailability in runoff w

T

water ligand model MinteqA2 [Allison et al. (1991)]. In addition, the bioavailability of

released copper species from naturally patinated copper and antifouling paints in artificial

seawater was investigated, for selected sampling periods, with bioassays, including bacterial

and fungal copper specific sensors. The sensors respond primarily to the presence of the free

cupric ion in water, Cu(H2O)62+, considered to be the most bioavailable copper species to

aquatic organisms. Differential pulse anodic stripping voltammetry (DPASV) was used to

detect the electrochemical active fraction (labile) of copper species in artificial sea water, to

elucidate differences between the total copper concentration, the fraction of labile copper

species and the fraction of free cupric ions. More details are given in Papers I and III.

11

3. Summary

3.1 Can corrosion rates be used to predict runoff rates at any given site?

(Papers I, II)

The corrosivity of an exposure site can be classified according to the ISO classification

system 9223 into different classes (C1-C5), where a C5-site is the most corrosive one with the

highest corrosion rates for metals. This classification system is useful to evaluate the

corrosivity of a specific environment to the most commonly used metals and to enable a

comparison between different sites. Corrosion rate measurements within this study showed

the marine site to be moderately corrosive (class C3) for zinc (Paper II) and highly corrosive

for copper (class C4) during the one-year exposure (Paper I). The difference between the

annual corrosion rate and the corresponding annual runoff rate is illustrated in Figure 9 for

zinc and copper, respectively.

0

5

10

15

20

Copper Zinc

Ann

ual r

unof

f and

cor

rosi

on r

ate

g m

-2ye

ar-1

Corrosion rate

Corrosion rate

Runoff rateRunoff rate

7%33%

Figure 9. Differences in annual corrosion rate and runoff rate for copper and zinc sheet exposed during

one year at unsheltered conditions at the marine site.

The results show unambiguously the runoff rate to be substantially lower compared to the

corrosion rate for both copper and zinc after one year of exposure to the marine environment.

For copper sheet, the runoff rate of total copper (1.4 g m-2 year-1) was approximately 13 times

lower compared to the corrosion rate (19 g m-2 year-1). For zinc sheet, the difference was not

as pronounced as for copper, but still 3 times lower (runoff rate: 2.6 g m-2 year-1, corrosion

12

rate: 7.9 g m-2 year-1). Similar results have been published in the literature for e.g. an urban

site showing the annual runoff rate during one year of exposure to be approximately 5 times

lower than the corrosion rate for copper and 1.6 times lower for zinc [He et al. 2002]. It is

anticipated that the corrosion rate will equal the runoff rate after the patina has reached

steady-state composition and thickness, a process that can take decades or longer depending

on the metal and prevailing environmental conditions.

Since different mechanisms govern the corrosion (electrochemical) and the runoff (chemical,

erosion/wear) process, and large differences and variations are seen between different years of

exposure, neither the corrosivity of the test site nor the corrosion rate can be used to estimate

the annual runoff rate of a specific metal.

In all, the metal runoff rate at a specific site cannot be estimated from the annual corrosion

rate or the corrosivity of the test site. The runoff rate is significantly lower compared to the

corrosion rate during one year of marine exposure.

3.2 Does precipitation characteristics influence the runoff process?

Any rain event results in time-dependent runoff processes with initially high runoff rates

during the first rain portion that decrease to fairly constant rates during the remaining rain

event. The magnitude of the first flush contribution depends primarily on rain pH, prevailing

environmental conditions in terms of wet conditions and dry deposition of corrosive species

in-between rain events (dry periods), rain intensity, which determine the residence time of

rainwater on the surface, and surface morphology, which affects the effective surface area and

enables the surface accumulation of water and corrosive species for longer time periods. The

steady-state release rate is primarily governed by rain pH. The runoff rate increases with

increasing rainfall quantity, decreasing rain pH, extended residence time for rainwater (low

rain intensity), extended dry periods of high deposition rates of corrosive species, and humid

conditions, and increasing porosity of the surface patina. The individual effect of these

parameters has been thoroughly studied in laboratory investigations [He et al (2001a)]. In

addition, the effect of surface inclination and orientation also has to be taken into account.

[Odnevall-Wallinder et. al. (2000), Lehmann (1995)].

13

Long periods of dry deposition in between rain events have shown to largely influence the

magnitude of the first flush runoff portion, whereas few, or short, dry periods result in a small

first flush contribution and hence a runoff rate that primarily is governed by steady-state

conditions, and vice versa. This is schematically illustrated in Figure 10.

-20

20

40

60

80

100

120

0 5 10 15 20 25 30 5 10 15 20 25 30 35 4

Exposure time / a.u.

Run

off r

ate

/ a.u

.

Long dry periods0

First flush

0

First flush

Steady state

Short dry periods-20

20

40

60

80

100

120

0 5 10 15 20 25 30 5 10 15 20 25 30 35 4

Exposure time / a.u.

Run

off r

ate

/ a.u

.

Long dry periods0

First flush

0

First flush

Steady state

Short dry periods

Figure 10. Schematic description of the importance of short and long dry periods in-between rain events on

the first flush runoff contribution.

Due to an integrated effect of different environmental, rain and material parameters on the

runoff process, any individual effect of a single parameter is difficult to interpret at field

conditions. However, the effect of dry periods is evident when comparing runoff rates

measured at the marine site, characterized by significantly more wet conditions compared to

similar measurements at an urban site, see section 3.8

The fact that also other parameters except rainfall quantities govern the runoff process is

evident when normalizing momentary runoff rates for individual sampling periods to the

corresponding rainfall quantity during this period. This is illustrated in Figure 11 for zinc

sheet and copper sheet, respectively. Despite sampling periods (filled symbols) of relatively

constant rainfall pH (4.8±0.1), large variations (19-55%) in runoff rates were observed for

both zinc and copper. This clearly shows that other parameters, except pH and annual

precipitation volume, have a significant effect on the metal runoff rate. The effect is attributed

to differences in rainfall intensities and dry periods.

14

0 100 200 300 400 5000.000

0.001

0.002

0.003

0.004

0.005

0.006

Exposure period / Days

Run

off q

uant

ity /

g m

-2m

m-1

June 2004 Nov 2005Dec 2004

Zn

Cu-new

0 0

2

4

6

8

0 100 200 300 400 500

pH

Exposure period / Days

Nov 2005Dec 2004

Figure 11. The runoff quantity normalized to the rainfall quantity for each sampling period during the one-

year marine exposure for zinc sheet (Zn), and copper sheet (Cu-new)-(left), and variations in

precipitation pH during the same time period. Filled symbols denote sampling periods with a pH

of 4.8±0.1.

In all, the runoff rate depends on the integrated effect of prevailing environmental conditions,

precipitation characteristics, and material parameters, including rainfall quantities, the

duration of dry periods with dry deposition, rain pH and intensity, and patina morphology.

3.3 Do changes in patina composition influence the runoff rate?

(Papers I, II)

The marine site is characterized by very high deposition rates of chlorides and sulphates,

normally during October-December, with significantly lower deposition rates during the

remaining part of the year. Monthly deposition rates can also vary significantly between

different years of exposure. This was evident when comparing data compiled from the field

exposure on copper materials staring in June 2004, with the exposure of zinc materials

starting in December 2004. The average deposition rate of chlorides was approximately 7

times higher during the summer period between June and August in 2004 (136-265 g m-2 yr-1)

compared to summer period in 2005 (20-37 g m-2 yr-1) (Paper II).

One year of unsheltered marine exposure resulted in a patina primarily composed of

hydrozincite (Zn5(CO3)2(OH)6) and cuprite (Cu2O), for the high purity zinc materials (zinc

15

sheet, hot dipped galvanized steel) and copper sheet, respectively. The presence of these

phases is in agreement with findings in urban and rural environments [He et. al. (2001b)].

High deposition rates of chlorides in October 2004 resulted in an extensive formation of

paratacamite (Cu2Cl(OH)3) on copper. On the other hand, scarce formation of chloride and/or

sulphate rich phases was observed on the high purity zinc materials from which followed that

the runoff rate remained relatively constant. Due to a substantial local deterioration of the

chromate layer of continuous galvanized steel, these phases were more abundant on this

material. The extensive formation of paratacamite on copper resulted in a somewhat increased

runoff rate from 1.1g m-2 yr-1 with cuprite as the main patina component to 1.4 g m-2 yr-1. For

zinc, the runoff rate remained relatively unchanged. A schematic description of changes in

patina composition and corresponding changes in runoff rates are given in Figure 12 for

copper and zinc, respectively

Figure 12. Schematic description of corrosion product formation and corresponding changes in runoff

rates for bare copper sheet (left) and zinc sheet (right) exposed at unsheltered conditions

during one year at the marine site.

Cu2OCu2OCu2O

0 26 52 weeks

Corrosion product formation

Cu2Cl(OH)3 Cu2Cl(OH)3

June 2004 May 2005

Run

off r

ate

/ g m

-2yr

-1

Exposure period / Days

Corrosion product formation Dec 2004 Nov 2005

0 26 52 weeks

ZnXClY(OH)ZSO4

0.0

0.6

1.2

1.8

2.4

3.0

0 50 100 150 200 250 300 350 400Exposure period / Days

Run

off r

ate

/ g m

-2yr

-1

0.0

0.4

0.8

1.2

1.6

2.0

0 50 100 150 200 250 300 350 400

Zn5(CO3)2 (OH)6Zn5(CO3)2 (OH)6 Zn5(CO3)2 (OH)6

16

In all, the marine exposure resulted in the initial formation of cuprite on copper. High

deposition rates of chlorides gave an extensive formation of paratacamite with somewhat

increased runoff rates as a result. Due to wet conditions and high removal rates of deposited

sulphates and chlorides on the zinc rich materials, the thin patina was predominantly

composed of hydrozincite and a relatively constant runoff rate during the time of exposure.

3.4 Do surface treatments and coatings influence the runoff rate?

(Papers I, II)

Annual runoff rates are compiled in Figure 13 for all zinc- and copper based materials

exposed during one year in the marine environment (Brest, France). The total annual rainfall

quantities impinging the panels were 740 and 700 mm yr-1 for copper- and zinc-based

materials, respectively. Differences in annual rainfall quantities are due to different starting

periods, June for copper and December for zinc

HD-galv

Zn

Galv

PW Zn

GalfanGalvalume

PW Zn B

PW Zn A

HD-galv D

Galv E

HD-galv CD

Runoff rate / g m-2 yr-10 0.5 1 1.5 2 2.5 3 3.5.50 0.5 1 1.5 2 2.5 3 3

NB

Cu-new

Cu-200y

NG

NG1U

NG2P

NG2U 0.12

0.14

0.4

0.9

1.2

1.4

1.6

Runoff rate / g m-2 yr-10.5 1.0 1.5

0.12

0.14

0.4

0.9

1.2

1.4

1.6

0.120.12

0.140.14

0.40.4

0.90.9

1.21.2

1.41.4

1.61.6

Runoff rate / g m-2 yr-10.5 1.0 1.5

2.9

2.62.6

1.71.7

1.61.6

1.21.2

0.550.55

2.9

1

0.140.14

0.0850.085

0.0590.059

0.0570.057

0.0290.029

Figure 13. Annual runoff rates of zinc and copper from zinc-based, left, and copper-based, right,

materials exposed at unsheltered conditions during one year at the marine site. The studies

comprise bare materials (Hot dipped galvanized steel, HD-galv; Zinc sheet, Zn; Galfan;

Galvalume, Cu sheet, Cu; 200 year old copper sheet, Cu-200y; Nordic Brown artificial patina,

NB; Nordic Green artificial patina, NG), and materials with surface treatments, 0.02-20 µm

thick (Chromated continuously Galvanized steel, galv; Pre-weathered zinc, PW Zn; Nordic

Green artificial patina, NG1U, NG2, NG2P), and organic coatings of various thickness, 1-150

µm (Pre-weathered zinc, PW Zn A, PW Zn B;Hot dipped galvanized steel, HD-galv D, HD-

galv CD; continuously galvanized steel, Galv E)

17

The presence of a surface treatment on pre-patinated copper, aimed for anti-graffiti protection,

decreases the runoff rates with 56-87% compared to bare pre-patinated copper. However, the

beneficial barrier properties were in some cases reduced with time. The presence of a thin

layer of chromate (0.02 µm) on continuously galvanized steel, aimed for temporary

protection, reduced the runoff rate with 35% compared to bare zinc sheet. However, the

chromate layer became locally deteriorated already within weeks of exposure, with a reduced

corrosion resistance, an increased runoff rate, and the formation of chloride and sulphate rich

corrosion products protruding the surface as a result. This is contradictory to previous

findings in urban environments showing lower runoff rates of zinc from continuously

galvanized steel with chromate up to 5 years of unsheltered exposure [Bertling et. al.

(2006b)].

As seen in Figure 13, the presence of organic coatings of varying thickness and composition,

on zinc-based materials, improved the corrosion resistance and reduced runoff rates of zinc

significantly (95-99%), compared to bare zinc sheet. The reduction was related to the coating

thickness, probably as a result of reduced numbers of defects and a longer penetration

distance for water to reach the metal substrate. No deterioration of any organic coating or

significant change in runoff rates was observed during the first year of exposure to the marine

environment.

In all, the application of a surface treatment or an organic coating reduces the metal runoff

significantly. Important factors that determine this protective behaviour are coating thickness

and porosity. Surface treatments with chromate are not beneficial for marine-exposed

materials due to a local deterioration, concomitant accumulation of corrosive species,

formation of chlorine- and sulphate zinc-rich corrosion products and hence

3.5 Do changes in alloy composition influence the runoff rate?

(Paper II)

The metal runoff process is predominantly governed by chemical processes, and to a certain

extent also by erosive processes. The influence of patina characteristics is assumed to be small

for the high purity zinc materials due to a limited patina formation, whereas its effect on

copper is more pronounced due to the formation of voluminous, and poorly adhesive

18

corrosion products. The chemical composition and dissolution properties of corrosion

products formed within the patina depends on prevailing environmental conditions and the

possibility for corrosive species such as chlorides and sulphates to accumulate and precipitate

on the surface.

The degree of metal runoff depends on whether the metal is in its pure form or in an alloy.

Differences in microstructure and surface coverage of phases composed of different alloy

constituents in the bulk material result in the formation of corrosion products of varying

composition and surface coverage. The metal runoff process can, hence, not be estimated

based on the nominal alloy composition. This is illustrated in Figure 14 for zinc-aluminum-

based alloys (Galfan and Galvalume) and bare zinc rich materials (zinc sheet and hot dipped

galvanized steel). Although a trend can be discerned, no linear relation can be established

between the bulk metal content and the annual metal runoff quantity. Since the one-year

exposure in the marine environment (Paper II) resulted in a very slow formation rate of

corrosion products in the patina, differences in annual runoff quantities can be explained by

the microstructure. Galfan, with an eutectic microstructure (two phase system), has one

aluminum rich phase (~99% Zn, ~1% Al) covering approximately 75% of the exposed area

and another phase of higher aluminum content (85% Zn, ~16% Al). Even though 75% of the

exposed surface area primarily contains zinc, the runoff rate was 1.6 times lower

(1.2 g m-2 yr-1) compared to zinc sheet (2.6 g m-2 yr-1). The reason is the formation of less

soluble aluminum-containing zinc rich corrosion products in both phases of Galfan.

Galvalume behaves in a similar way with zinc-rich interdendritic areas composed of 99%Zn

and ~1%Al (33 % of the surface area) and aluminum-rich dendrite branches containing more

than 15 wt% Al (66 % of the surface area). Due to a larger surface area of zinc-rich phases on

Galfan, compared to Galvalume, higher annual runoff rates were obtained from the former.

19

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0 20 40 60 80 100

Ann

ual r

unof

f rat

e

Weight % Zn

Galfan

Zn

HD-galv

Galvalume

Figure 14. Annual runoff rates of zinc from zinc-aluminum based alloys (Galfan and Galvalume) and zinc

rich materials (zinc sheet and hot dipped galvanized steel) compared to the nominal bulk

composition when exposed at unsheltered conditions during one year in a marine environment.

In all, the nominal bulk alloy composition cannot be used to predict metal runoff rates of

metals from metal alloys.

3.6 Do copper compounds of anti-fouling paints influence the copper release

rate? (Paper III)

Anti-fouling paints are used to prevent fouling on the hull of ships and pleasure crafts, which

will increase fuel consumption significantly if present. Different types of anti-fouling paints

are available on the market. Copper has beneficial anti-fouling properties and is therefore

often used in various compounds such as copper oxides and copper thiocyanate, in

combination with organic biocides. Release studies of metals other than copper from some

copper-containing paint matrices show significantly higher released concentrations of zinc,

nickel and iron compared to copper. These results imply the importance of assessing the

ecotoxic response of anti-fouling paints due to an integrated effect of not only copper, but also

of other paint matrix constituents released, including organic species, and other metals. It has

also recently been shown that newly developed, so-called "toxic free", anti-fouling paints that

do not contain e.g. copper, still are toxic towards marine organisms [Karlsson and Eklund,

2004].

20

Measured release rates of total copper from anti-fouling paints of different active copper

compounds, copper oxides (CuOX), copper thiocyanates (Cu(NCS)X or metallic copper

flakes combined with copper oxides (FlakeX), are presented in Figure 15 together with the

corresponding nominal copper content of the different paint matrices.

T

otal

mom

enta

ry c

oppe

r re

leas

e ra

tes /

μg

cm

-2da

y-1

Weight % Cu0 10 20 30 40 50 60 70 80

80

70

60

50

40

30

20

10

0

CuOX

Cu(NCS)X

FlakeX

Figure 15. Release rates of total copper from anti-fouling paints of different active copper compound

(copper oxides-CuOX, copper thiocyanates-Cu(NCS)X and metallic copper flakes with copper

oxide-FlakeX) after 14 days of exposure in artificial seawater compared to the nominal copper

content of the different paint matrices.

The total copper release rate cannot directly be related to the nominal copper content of the

paint matrix. This implies that other parameters, such as the dissolution properties of the

active copper compound, morphology and other properties of the paint matrix, strongly

influence the copper release process.

In all, the release rate of total copper from anti-fouling paints cannot solely be explained by

the active copper compound concentration. Paint matrix properties, the chemical form of

copper, as well as the release of other paint constituents including organic species and other

metals are also of importance.

21

3.7 How bioavailable is released copper in marine environments?

(Paper I, III)

Corrosion-induced copper runoff from construction materials constitutes one source of the

diffuse emission of copper in the society. Other sources with significantly higher

contributions are traffic, tap water systems and aerial lines [Bergbäck et. al. (2001)]. As

shown earlier, the environmental effect of dispersed copper depends on the chemical form of

released copper, which determines the bioavailability (potential uptake by organisms). The

total copper concentration is not a valid measure of any potential hazard. Parameters such as

pH and organic matter largely affect the chemical form of released metal, where an increase

of both parameters generally reduces the bioavailability of copper. The presence of organic

matter reduces the bioavailability by forming strongly bonded copper-organic complexes and

an increased pH results in the formation of inorganic or organic complexes of higher stability.

The chemical speciation of copper in runoff water collected from naturally patinated copper

exposed during one year at the marine test site was predicted with the Minteq A2 water ligand

model. Since no measurements were made on the amount of organic matter in the runoff

water, the model predicted the free cupric ion concentration to approximately 60 to 70% of

the total copper concentration during periods of high deposition rates of chlorides and

sulphates. However, the presence of organic matter reduces the bioavailable fraction

considerably. The bioavailability of copper in the runoff water was therefore determined with

a copper-specific bacterial strain at the immediate release situation (Paper I), Figure 16. The

bacteria are genetically modified to react to bioavailable copper, primarily the free cupric ion,

and will not react to copper in inorganic or organic complexes.

The bioavailable fraction varied between 14 and 54% of the total copper content in the runoff

water. During the period with very high deposition rates of chlorides and sulphates (172 days)

only 14% was measured as bioavailable, i.e. substantially lower compared to the model

prediction (60-70%). The reason is most probably related to the interaction of seawater

aerosols containing different amounts of inorganic and organic species that form complexes

with copper and thereby reduce the bioavailability. The results imply the importance of

evaluating other parameters such as organic matter present in the runoff water when assessing

the bioavailability, or the potential environmental effect of corrosion-induced copper release

from construction materials. More details are given in Paper I.

22

0

20

40

60

80

100

25 63 71 123 137 172 214 264 305 327 376

Bio

avai

labl

e fr

actio

n af

ter

the

imm

edia

te r

elea

se/

%

Exposure period / Days

Figure 16. The bioavailable fraction of the total copper concentration in runoff water immediately after

release from naturally patinated copper sheet exposed during one year at unsheltered conditions

in a marine environment (Paper I). The fraction is determined by means of a copper-specific

bacterial biosensor (MC1061(pSLcueR/pDNPcopAluc))

Recent investigations have shown copper containing runoff water to interact with natural

and/or man-made surfaces in the near vicinity of buildings in urban environments.

Investigated surfaces include concrete, limestone and soil systems [Bertling et. al. (2002)].

According to these studies, soil systems have a high capacity to retain copper in runoff water

(> 98%), showing breakthrough capacities ranging from 170 to 8000 years, depending on soil

systems. The investigations with soil and limestone showed a substantial reduction of the total

and the bioavailable fraction of copper during runoff water interaction.[Bertling et. al.

(2006a)]

If copper containing runoff water directly would be transported into seawater in a marine

environment, the bioavailable fraction of released copper would react with natural organic

matter and inorganic ligands forming strong complexes, thereby reducing the bioavailability

even further. Copper specific bacterial tests on copper released from pure copper sheet

exposed to artificial seawater, without any organic matter, showed the fraction of free cupric

ions to be less than 2% at comparable total copper concentrations, i.e. substantially lower

compared to the bioavailable fraction in runoff water collected at the immediate release

situation (14-54%). Bioassay testing with yeast predicted a bioavailable fraction below 8%, in

23

good agreement with the MinteqA2 model predictions, showing 6% of total copper to be

present as free cupric ions for similar copper concentrations. The presence of natural organic

matter in seawater would reduce the bioavailability even further by forming strong complexes

with copper. More details are given in Paper III.

In all, the total copper concentration at the immediate release situation does not equal the

bioavailable copper fraction, which depends on the chemical form of released copper and

parameters such as pH precipitation and presence of organic matter. The bioavailable

fraction, measured with copper specific bacterial bioassays, showed values between 14 and

54% at the immediate release situation. This fraction would be significantly reduced in

contact with seawater by forming strong complexes with natural organic matter and

inorganic ligands.

3.8 Is the runoff rate different in a marine compared to an urban environment?

(Papers I, II)

The metal runoff process is complex since it is governed by a combination of environmental

and material parameters. Detailed studies in urban and rural environments have shown the

annual rainfall quantity, rain pH and intensity and environmental conditions and deposition of

corrosive species in between rain events to largely influence the runoff rate [Odnevall-

Wallinder et. al. (1998), He et. al. (2000), He et. al. (2001a), Matthes et. al. (2003), Faller et.

al. (2005)]. In order to enable a comparison of the runoff process in a marine and an urban

environment, it is important to understand the influence of these parameters on the runoff

process.

Both sites within this comparison show a similar rain pH (~5.6), typical rainfall intensities

less than 4 mm hr-1, and a low gaseous pollutant level of SO2 (< 5 µg m-3). The main

difference between the sites is higher relative humidities, a higher annual precipitation

quantity and significantly higher dry and wet deposition rates of anions including chlorides

and sulphates from seawater at the marine site. More details are given in Papers I and II.

The annual runoff rate normalized against the annual rainfall quantity impinging on high

purity zinc materials (hot dipped galvanized steel, zinc sheet, continuously galvanized steel

24

with a chromate layer) and naturally patinated copper sheet exposed unsheltered during one

year in the marine (Brest) and the urban (Stockholm) environment, is illustrated in Figure 17.

The normalization eliminates the effect of differences in annual rainfall quantities between the

sites (Brest: 700 mm yr-1, Stockholm: 350 mm yr-1).

0

0.001

0.002

0.003

0.004

Cu-new Cu-200y

Run

off q

uant

ity /

Rai

nqu

antit

yg

m-2

yea

r-1 / m

mye

ar-1

30%

50%

UrbanMarineUrbanMarine

0

0.001

0.002

0.003

0.004

0.005

0.006

HD-galv Zn Galv

Run

off q

uant

iy /

Rai

n qu

antit

yg

m-2

yea

r-1

/ mm

yea

r-1

24%0%

24%

UrbanMarine

Figure 17. Annual runoff rates normalized to the annual rainfall quantity at a marine and an urban site

for high purity zinc materials including hot dipped galvanized steel (HD-galv), zinc sheet (Zn),

and continuous galvanized steel with a chromate layer (Galv) (left), and copper materials

including new copper sheet (Cu-new) and 200 years old copper sheet (Cu-200y) (right). All

materials have been exposed at unsheltered conditions during one year. Runoff rate

differences at comparable rainfall quantities are presented in each graph.

Despite a more corrosive environment with significantly higher deposition of chlorides, the

marine site shows lower runoff rates at comparable rainfall quantities. The effect is more

pronounced for copper, compared to zinc. One exception is bare zinc sheet (Zn) with similar

runoff rates at both exposure sites. The reason is unknown but may be related to a small first

flush effect and a low effect of dry and wet deposition in-between rain events on low-

corroded bare zinc sheet [He et. al. (2001a)].

Higher runoff rates at comparable rainfall quantities at the marine site compared to the urban

site are primarily the result of prevailing environmental conditions with higher relative

humidities and rainfall quantities resulting in longer periods of wet conditions on the surface

that dilute and disable deposited species to precipitate on the metal surface. In addition, the

presence of different sea-salts lowers the point of deliquescence, and prolongs the wet surface

conditions [Winkler and Junge (1972)]. Runoff rate measurements of chlorides and sulphates

from the exposed surfaces showed the majority of deposited species to be washed from the

25

surface on both zinc and copper. Long periods of wet conditions result in a small first flush

contribution at the marine site compared to the urban site with more dry and wet periods

enabling deposited species to accumulate and precipitate into corrosion products.

In all, lower runoff rates at comparable rainfall quantities at the corrosive marine site,

compared to the more benign urban site, are attributed to higher surface wetting and

concomitant higher removal rates of deposited chloride and sulphate species from the

marine-exposed surfaces.

26

4. Main conclusions

Corrosion-induced metal runoff studies of zinc and copper have been conducted in a marine

environment during one year for different zinc and copper materials with, and without,

surface treatments and organic coatings. The main focus has been on outdoor constructions.

Changes in patina formation, corrosion rates, barrier properties and runoff rates of copper and

zinc have been investigated through a multi-analytical approach, chemical speciation and

bioavailability monitoring, and correlated to changes in prevailing environmental conditions

and deposition of seawater aerosols. Main conclusions drawn from this investigation are

illustrated in Figure 18. Detailed results are provided in Papers I-II.

metals / alloys

surface treatmentsorganic

coatings

patina

θ

Main patina components:Cu2O, Cu2Cl(OH)3Zn5(CO3)2(OH)6, ZnxCly(OH)zSO4

Metal release ratesnot proportional

to bulk alloy composition

Metal runoff:

Bare patina >> surface treatments > coatings

Cu: 0.12 - 1.6 g m-2 yr-1Zn: 0.03-2.9 g m-2 yr-1

Biovailable copper coatings

Cu: 0.12 - 1.6 g m-2 yr-1Zn: 0.03-2.9 g m-2 yr-1

Biovailable copper

5. Acknowledgement

My time at Division of Corrosion Science is coming to an end, but I am very thankful for the

opportunity given by Professor Christofer Leygraf and my supervisor, Associate Professor

Inger Odnevall Wallinder two years ago. During the time I have gradually, with guidance and

support from my supervisor, gone through a transition being a material scientist into a

scientist with an environmental perspective.

I wish to thank all colleagues at the Division of Corrosion Science for laughter and happy

memories.

I would like to thank all persons that I have been in contact with during my projects:

Mariann Sundberg (SCDA, Outokumpu Copper) and Rolf Sundberg (Outokumpu) for their

enthusiasm when discussing copper topics.

Christer Halén, Outokumpu Copper for the help with several surface analyses.

Nathalie LeBozec and Cecile Hall, French Corrosion Institute, for the possibility to perform a

field study, and the time-consuming work to maintain the exposure site in Brest, France.

Lars Göthe, Stockholm University for X-ray powder diffraction analyses.

Marko Virta, Department of Applied Chemistry and Microbiology, University of Helsinki for

the help to perform bacterial and fungal biosensor tests.

Lena Höjvall at Attana AB for wetting angle measurements.

Per-Erik Augustsson, SSAB; Eric Pezennec, Arcelor; Anniki Hirn, Nordic Galvanizers;

Jostein Digernes, Örsta Stål AS; and Andreas Friedrich, Marianne Schönnenbeck at

Rheinzink are all gratefully acknowledged for in-valuable discussions and financial support.

The financial support from European Copper Institute, International Copper Association and

Outokumpu Copper is greatly acknowledged.

Finally I would like to express my sincere gratitude to my family, Ulrika and my daughter,

My, for understanding and help during the last month of hard work on my thesis.

28

6. References

Aastrup T., Wadsak M., Leygraf C., Schreiner M., (2000), In situ studies of the initial

atmospheric corrosion of copper. Influence of humidity, sulfur dioxide, ozone, and

nitrogen dioxide, Journal of The Electrochemical Society, 147(7), p. 2543

Atrens A., Fitzgerald K., Skeleton G., Nairn J., (1996), Composition and structure of copper

patinas, Proc.13th International Corrosion Congress, Melbourne, Australia

Allison J.D., Brown D.S., (1991), MinteqA2, A geochemical assessment model for

environmental systems, U.S. Environmental Protection Agency, Athens, GA,

EPA/6003-91021

ASTM, (1998), American Society for Testing Materials, Standard Practice for the

Preparation of Substitute Ocean Water, D1141-98

ASTM, (2003), American Society for Testing Materials, Standard Test Method for

Determination of Copper Release Rates From Anti-fouling Coating Systems in Artificial

Seawater, D6442-03

Antelo J.M., Arce F., Penedo F.J., (2000), Effect of pH on the complexation parameters for

organic matter dissolved in river water, Chemical Speciation and Bioavailability, 12(1)

Auckland Regional Council, (2005), Sources and loads of metals in urban stormwater,

Auckland, Technical Publication No. ARC04104

Belghazi, A., Bohm, S., Sullivan, J.H., Worsley, D.A., (2002), Zinc runoff from organically

coated galvanized architectural steel, Corrosion Science, 44, p. 1639

Bergbäck B., Johansson K., Mohlander U., (2001), Urban metal flows – A case study of

Stockholm, Water, Air and Soil Pollution: Focus 1: p.3

Bertling S., Odnevall Wallinder I., Berggren Kleja D., Berggren D., (2002), Immobilization of

copper in runoff water from roofing materials by limestone, soil and concrete,

Proceedings, 15th International Corrosion Conference, paper 44, Granada, Spain

Bertling S., Odnevall Wallinder I., Berggren Kleja D., Leygraf C., (2006a), Long-term

corrosion-induced copper runoff from natural and artificial patina and its

environmental impact, Environmental Toxicology and Chemistry 25(3), p. 891

Bertling S., Odnevall Wallinder I., Leygraf C., Berggren-Kleja D., (2006b), Occurrence and

fate of corrosion induced zinc runoff water from external structures, The Science of the

Total Environment, doi:10.1016/j.scitotenv.2006.01.008

Boulanger B., Nikolaidis P., (2003), Mobility and aquatic toxicity of copper in urban

watershed, Journal of the American Water Resources Association, April, p. 1

29

Bruland K.W., Rue L.E., Donat J.R. Stephen A.S., Moffet J.W., (2000), Intercomparison of

voltammetric techniques to determine the chemical speciation of dissolved copper in

coastal seawater samples, Analytica Chimica Acta, 405, p. 99

Cachet C., Wiart R., (1980), The kinetics of zinc dissolution in chloride electrolytes:

Impedance measurements and electrode morphology, Journal of Electroanalytical

Chemistry, 111, p. 235

Cachet C., Wiart R., (1981), Reaction mechanism for zinc dissolution in chloride electrolytes,

Journal of Electroanalytical Chemistry, 129, p. 103

Carter J.P, Lindstrom P.J., Flinn D.R., Cramer S.D., (1987), The effect of sheltering and

orientation on atmospheric corrosion on structural metals, Materials performance, July,

p. 25

Campbell, P.G.C, (1995), Interactions between trace metals and aquatic organisms: a

critique of the free-ion activity model in A. Tessier and D.R. Turner, eds., Metal

Speciation and Bioavailability in Aquatic Systems, IUPAC, New York, John Wiley and

Sons, p.45

Chebbo G., Gromaire M.C., Ahyerre M., Garnaud S., (2001), Production and transport of

urban wet weather pollution in combined sewer systems: the “Marais” experimental

urban catchment in Paris, Urban Water, 3, p.3

Chen R.Y., Willis D.J., (2005), The behavior of Silicon in the solidification of Zn-55Al-

1.6Si coating on steel, Metallurgical and Materials Transaction A, 36A(1), p. 117

Chen Z.Y., (2005), The Role of Particles on Initial Atmospheric Corrosion of Copper and

Zinc – Lateral Distribution, Secondary Spreading and CO2/SO2 influence, Doctoral

Thesis, School of Industrial Engineering and Management, Royal Institute of

Technology, Sweden, ISBN 91-7178-155-2

Costa J.M., Lorente J., Vilarrasa M., Codina B., (1999), Atmospheric corrosion: A field study

of metal runoff, Proc.14th International Corrosion Congress, Cape Town, South Africa

Costas L.P.( 1982), Atmospheric corrosion of copper alloys exposed for 15 to 20 years,

Atmospheric Corrosion of Metals, ASTM STP 767, S.W. Dean Jr., and E.C. Rhea, Eds,

American Society for Testing and Materials, p. 106

Cramer S.D., McDonald L.G., (1990), Atmospheric factors affecting the corrosion of zinc,

galvanized steel and copper, Corrosion testing and evaluation: Silver anniversary

volume ASTM STP 1000, American Society of Testing and Materials, p.241

30

Cramer S.D., McDonald L.G., Spence J. W, (1993), Effects of Acidic deposition on the

Corrosion of Zinc and Copper, Proc.12th International Corrosion Congress, Houston,

USA

Cramer S. D., Matthes S. A., Covino B.S., Bullard S.J., Holcomb G.R., (2002),

Environmental factors affecting the atmospheric corrosion of copper, Outdoor and

Indoor Atmospheric Corrosion, ASTM STP 1421, H. E. Townsend Ed., American

Society for Testing Materials, West Conshohocken, PA,

Donat J.R., Lao K.A., Bruland K.W., (1994), Speciation of dissolved copper and nickel in

South San Fransisco Bay: a multi-method approach, Analytica Chimica Acta, 284,

p. 547

Elvins J., Spittle J. A., Worsley D.A., (2005), Microstructural changes in zinc aluminium

alloy galvanising as a function of processing parameters and their influence on

corrosion, Corrosion Science, 47, p. 2740

Faller M., Reiss D., (2005), Runoff behaviour of metallic materials used for roofs and facades

– a 5-year field exposure study in Switzerland, Materials and Corrosion, 56(4), p.244

Falk T., Svensson J-E, Johansson L.G., (1998), The influence of CO2 and NaCl on the

atmospheric corrosion of zinc, Journal of The Electrochemical Society, 145(9), p. 2993

French Corrosion Institute, (2004), Atmospheric and corrosivity data of Sainte Anne field

station, Brest, France

French Corrosion Institute, (2005), Atmospheric and corrosivity data of Sainte Anne field

station, Brest, France

Friel J.J., (1986), Atmospheric corrosion products on Al, Zn AlZn metallic coatings,

Corrosion, 42(7), p. 422,

Graedel T.E., (1989), Corrosion mechanisms for zinc exposed to the atmosphere, Journal of

Electrochemical Society, 136(4), p. 193C

Hakkila K., Green T., Leskinen P., Ivask A., Marks R., Virta M., (2004), Detection of

bioavailable heavy metals in EILATox-Oregon samples using whole-cell luminescent

bacterial sensors in suspension or immobilized onto fibre-optic tips, Journal of Applied

Toxicology, 24, p.333

Hall L.W, Andersen R.D., (1999), A Deterministic Ecological Risk for Copper in European

Salt Water Environments, Marine Pollution Bulletin, 38(3), p.207

He W., Odnevall Wallinder I., Leygraf C., (2001a), A laboratory of copper and zinc runoff

during first flush and steady-state conditions, Corrosion Science, 43, p.127

31

He W., Odnevall Wallinder I., Leygraf C., (2001b), A comparison between corrosion rates

and runoff rates from new and aged copper and zinc as roofing materials, Water, Air

and Soil Pollution, 1, p. 67

He W., (2002), Atmospheric corrosion and runoff processes on copper and zinc as roofing

materials, Doctoral thesis, Royal Institute of Technology, Stockholm, Sweden, ISBN

91-7283-364-5

Huang S., Wang Z., (2003), Application of anodic stripping voltammetry to predict the

bioavailable/toxic concentration of Cu in natural waters, Applied Geochemistry, 18,

p.1215

Hullman H., editor, (2003), Naturally oxidizing metal surfaces – Environmental effects of

copper and zinc in building applications, Fraunhofer IRB Verlag, Stuttgart

Gagneten A.M., Vila I., (2001), Effects of Cu2+ and pH on the fitness of Ceriodaphnia dubia

(Richard 1894) (Crustacea, Cladocera) in microcosm experiments, Environmental

Toxicology, 16, p. 428

ISO standard 9223, (1992), Corrosion and metals and alloy – Corrosivity of atmospheres –

Classification

ISO standard 9225, (1992), Corrosion and metals and alloy – Corrosivity of atmospheres –

Measurement of pollution

ISO standard 9226, (1992), Corrosion and metals and alloys – Corrosivity of atmospheres –

Determination of corrosion rate of standard specimens for the evaluation of corrosivity

Ivask A, Virta M, Kahru A., (2002), Construction and use of specific luminescent

recombinant bacterial sensors for the assessment of bioavailable fraction of cadmium,

zinc, mercury and chromium in the soil, Soil Biology and Biochemistry, 34, p.1439

Ivask A., Francois M., Kahru A., Dubourguier H.C., Virta M., Douay F., (2004), Recombinant

luminescent bacterial sensors for the measurement of bioavailability of cadmium and

lead in soils polluted by metal smelters, Chemosphere, 22, p.147

Ivask A, Francois M, Kahru A, Dubourguier H.C, Virta M, Douay F., (2004), Recombinant

luminescent bacterial sensors for the measurement of bioavailability of cadmium and

lead in soils polluted by metal smelters, Chemosphere, 22, p.147

Karlsson, J., Eklund, B., (2004), New biocide-free anti-fouling paints are toxic, Marine

Pollution Bulletin, 49, p.456

Kasperek J., Lenglet M., (1997), Identification of thin film on zinc substrates by FTIR and

Raman spectroscopies, La Revue de Métallurgie, May, p. 713

32

Keddam M., Hugot- Le-Goff A., Takenouti H., Thierry D., Arevalo M.C., (1992), The

influence of a thin electrolyte layer on the corrosion process of zinc in chloride-

containing solutions, Corrosion Science, 33(8), p. 1243

Krätschmer A., Odnevall Wallinder I., Leygraf C., (2002), The evolution of outdoor copper

patina, Corrosion Science, 44, p.425

Landner L., Reuther R., (2004), Metals in society and in the environment, Kluwer Academic

Publishers, Dordrecht, ISBN 1-4020-2741-9

Lehmann B., (1995), Freibewitterungsverhalten von Dächern mit mettalldeckung –

Untersuchung zur Zinkabgabe von Dachdeckungen mit Titanzink, Doctoral thesis,

Universität von Hannover, Germany

Leskinen P, Virta M, Karp M., (2003), One-step measurement of firefly luciferase activity in

yeast, Yeast, 20, p.1109

Leskinen, P., Michelini, E., Picard, D., Karp, M. and Virta, M., (2005), Bioluminescent yeast

assays for detecting estrogenic and androgenic activity in different matrices.

Chemosphere, 61, p.259

Lenglet M., Lopitaux J., Leygraf C., Odnevall I, Carballeira M., Noualhaguet J.C., Guinement

J., Gautier, Boissel J., (1995), Analysis of corrosion products formed on copper in

Cl2/H2S/NO2 exposure, Journal of The Electrochemical Society, 142(11), p.3690

Leuenberger-Minger A.U., Faller M., Richner P., (2002), Runoff of copper and zinc caused by

atmospheric corrosion, Materials and Corrosion, 53, p.157

Lobnig R., Frankenthal R.P.,. Siconolfi D.J, Sinclair J.D., Stratmann M., (1994), Mechanism

of atmospheric corrosion of copper in the presence of submicron ammonium sulphate

particles at 300 and 373 K, Journal of The Electrochemical Society, 141(11), p. 2935

Lobnig R., Sinclair J.D., Unger M., Stratmann M., (2003), Mechanism of atmospheric

corrosion of copper in the presence of ammonium sulphate particles, Journal of the

Electrochemical Society, 150(6), p.A835

Malvault J.Y., Lopitaux J., Delahaye D., Lenglet M., (1995), Cathodic reduction and infrared

reflectance spectroscopy of basic copper(II) salts on copper substrate, Journal of

Applied Electrochemistry, 25, p.841

Mansfeld F., (1999), Analysis and interpretation of EIS data for metals and alloys, Solartron,

Technical report 26

Marcus P., (2002), Corrosion mechanisms in theory and practice, Marcel Dekker Inc., New

York,

33

Matthes S.A., Cramer S.D., Bullard S.J, Covino B.S., Holcomb G.R., (2003), Atmospheric

corrosion and precipitation runoff from zinc and zinc alloy surfaces, NACExpo,

Technical symposia, Atmospheric corrosion, paper 03598

Mendoza A.R., Corvo F., Gomez A., Gomez J., (2004), Influence of the corrosion products of

copper on its atmospheric corrosion kinetics in tropical climate, Corrosion Science, 46,

p.1189

Nùñez L., Reguera E., Corvo F., González E., Vasquez C., (2005), Corrosion of copper in

seawater and its aerosols in a tropical island, Corrosion Science, 47, p. 461

Odnevall I., Leygraf C., (1993), Formation of NaZn4Cl(OH)6SO4 * 6H2O in a marine

atmosphere, Corrosion Science, 34(8), p. 1213

Odnevall I., Leygraf C., (1994), Reaction sequence in atmospheric corrosion of zinc,

Atmospheric corrosion ASTM STP 1239, American Society of Testing and Materials

Odnevall I., Leygraf C., (1995), Atmospheric corrosion of copper in a rural atmosphere,

Journal of The Electrochemical Society, 142(11), p. 3682

Odnevall-Wallinder I., Leygraf C., (1997), A study of copper runoff in an urban environment,

Corrosion Science, 39 (12), p.2039

Odnevall-Wallinder I., Verbiest P., He W., Leygraf C., (1998), The influence of patina age

and pollutant levels on the runoff rate of zinc from roofing materials, Corrosion

Science, 40(11), p. 1977

Odnevall Wallinder I., He W., Augustsson P-E, Leygraf C., (1999), Characterization of

black rust staining of unpassivated 55% Al-Zn alloy coatings. Effect of temperature, pH

and wet storage, Corrosion Science, 41, p.2229

Odnevall Wallinder I., Unpublished data

Odnevall Wallinder I., Verbiest P., He W., Leygraf C., (2000), Effects of exposure direction

and inclination on the runoff rates of zinc and copper roofs, Corrosion Science, 42,

p. 1471

Odnevall Wallinder I., Leygraf C., (2001), Seasonal variations in corrosion rate and runoff

rate of copper roofs in an urban and rural atmospheric environment, Corrosion Science,

43, p.2379

Odnevall Wallinder I., Korpinen T., Sundberg R., Leygraf C., (2002), Atmospheric corrosion

of naturally and pre-patinated copper roofs in Singapore and Stockholm - Runoff rates

and corrosion product formation, Outdoor and Indoor Atmospheric Corrosion, ASTM

STP 1421, H. E. Townsend Ed., American Society for Testing Materials, West

Conshohocken, PA

34

Odnevall Wallinder I., Bertling S., Zhang X., Leygraf C., (2004), Predictive models of copper

runoff from external structures, Journal of Environmental Monitoring, 6, p.704

Opila R., (1987), Copper patinas: An investigation by Auger electron spectroscopy, Corrosion

Science, 27(7), p.685

Puigdomenech I., (2006), Hydra and Medusa chemical equilibrium software,

http://web.telia.com/~u15651596/, January

Quek U., Förster J., (1993), Trace metals in roof runoff, Water, Air and Soil Pollution, 68,

p.373

Riley J.P., Chester R., (1971), Introduction to marine chemistry, Academic Press, London

RHEINZINK GmbH & Co, (2006), Personal communication

Sandberg J., Odnevall Wallinder I., Leygraf C., Le Bozec N., Corrosion induced copper

runoff from naturally and pre-patinated copper in a marine environment, Corrosion

Science, in press

Seligman, P.F., Zirino, A., eds, (1998), Chemistry, Toxicity, and Bioavailability of copper

and its relationship to regulation in the marine environment, Technical

Document 3044, Space and Naval Warefare Systems Center, San Diego

Skennerton, G., Nairn, J., Atrens, A., (1997), Atmospheric corrosion of copper at Heron

Island, Materials Letters, 30, p.141

Stiles D.C., Edney E.O., (1989), Dissolution of zinc into thin aqueous films as a function of

residence time, acidic species and pH, Corrosion, 45(11), p. 896

Strandberg H., (1997), Perspective on bronze sculpture conservation. Modelling copper and

bronze corrosion, Doctoral Thesis, Dept. Inorganic Chemistry, Göteborg University,

Sweden, ISBN 91-7197-539-X,

Sullivan J.H., Worsley D.A., (2002), Zinc runoff from galvanised steel materials exposed in

industrial/marine environment, British Corrosion Journal, 37(4), p.282

Tipping, E., (1998), Humic-ion binding model VI: an improved description of the

interactions of protons and metal ions with humic substances, Aqueous Geochemistry,

4, p.3

U.S. EPA, (2002), United States Environmental Protection Agency, Annotated Bibliography

of Urban Wet Weather Flow Literature from 1996 through 2000

U.S. EPA, (2006), United States Environmental Protection Agency, Office of Water, Office

of Science and Technology (4304T), National recommended water quality criteria,

2006. http://www.epa.gov/waterscience/criteria/wqcriteria.html

35

Valkirs A.O., Davidson B.M., Kear L.L., Fransham R.L., (1994), Environmental Effects from

In-Water Hull Cleaning of Ablative Copper Anti-Fouling Coatings, Technical

Document 2662, Naval Command, Control and Ocean Surveillance Center RTD&E,

Division San Diego United States

Valkirs A.O., Seligman P.F., Haslbeck E., Caso J.S., (2003), Measurement of copper release

rates from antifouling paint under laboratory and in situ conditions: implications for

loading estimation to marine water bodies, Marine Pollution Bulletin, 46, p.763

Veleva L., Quintana P., Ramanauskas R., Pomes R.,. Maldonado L, (1995), Mechanism of

copper patina formation in marine environment, Electrochimica acta, 41(10), p.1641

Veleva L., Meraz E., (2005), Initial stages of zinc runoff in humid tropical climate, 16th

International Corrosion Congress, Beijing, China

Verbiest P., Waeterschoot H., Racek R., Leclercq M., (1997), Zinc and the environment,

Protective Coatings Europe, Sep, p.47

Vilche J.R., Varela F.E., Codaro E.N., Rosales B.M., (1997), A Survey of Argentinean

Atmospheric Corrosion II – Copper samples, Corrosion Science, 39(4), p.655

Voulvoulis, N., Scrimshaw, M.D., Lester, J.N., (1999), Review-Alternative antifouling

biocides, Applied Organometallic Chemistry, 13, p.135

Wang Z., Huang S., Liu Q., (2002), Use of anodic stripping voltammetry in predicting toxicity

of copper in river water, Environmental Toxicology and Chemistry, 21, p.1788

Winkler P., Junge C., (1972), The growth of atmospheric aerosol particles as a function of the

relative humidity- Part I: Method and measurements at different locations, Journal de

Recherches Atmosphériques, p.617

Zhang X., He W., Odnevall-Wallinder I., Pan J., Leygraf C., (2002), Determination of

instantaneous corrosion rates and runoff rates of copper from naturally patinated

copper during continuous rain events, Corrosion Science, 44, p.2131

Zhu F., Persson D., Thierry D., Taxèn C., (2000), Formation of corrosion products on open

and confined zinc surfaces exposed to periodic wet-dry conditions, Corrosion,

December, p.1256

36

Abstract Preface Contents 1. Introduction 1.1 Atmospheric corrosion of zinc and copper 1.2 Runoff rates of copper and zinc 1.3 Copper release from anti-fouling paints 1.4 Chemical speciation and bioavailability of released copper

2. Experimental 2.1 Marine site and experimental setup 2.1.1 Investigated materials 2.1.2 Experimental approach 2.2 Copper release from anti-fouling paints - experimental approach 2.3 Copper bioavailability in runoff water and artificial seawater

3. Summary 3.1 Can corrosion rates be used to predict runoff rates at any given site? (Papers I, II) 3.2 Does precipitation characteristics influence the runoff process? 3.3 Do changes in patina composition influence the runoff rate? (Papers I, II) 3.4 Do surface treatments and coatings influence the runoff rate? (Papers I, II) 3.5 Do changes in alloy composition influence the runoff rate? (Paper II) 3.6 Do copper compounds of anti-fouling paints influence the copper release rate? (Paper III) 3.7 How bioavailable is released copper in marine environments? (Paper I, III) 3.8 Is the runoff rate different in a marine compared to an urban environment? (Papers I, II)

4. Main conclusions 5. Acknowledgement 6. Referen