Document [original]

Mobilization Mechanisms of Soluble

and Dispersible Heavy Metals and

Metalloids in Soils

vorgelegt von

Umweltwissenschaftlerin

Sondra Klitzke (M. Sc.)

von der Fakultät VI – Planen Bauen Umwelt

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktor der Naturwissenschaften

-Dr. rer. nat.-

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr. Dieter Scherer (TU Berlin)

Berichter: Prof. Dr. Martin Kaupenjohann (TU Berlin)

Berichter: Prof. Dr. Rainer Schulin (ETH Zürich)

Tag der wissenschaftlichen Aussprache: 20. Dezember 2007

Berlin 2007

D 83

Dedicated to my parents

Ich höre und vergesse,

ich sehe und behalte,

ich handle und verstehe.

(alte konfuzianische Weisheit)

vii

Table of Contents

Table of Contents....................................................................................................vii

List of Figures…………. ...........................................................................................xi

List of Tables ..........................................................................................................xiv

List of Abbreviations...............................................................................................xv

Abstract...................................................................................................................xvi

Zusammenfassung...............................................................................................xviii

1 General Introduction......................................................................................1

1.1 Soil colloid characteristics, characterization and mobilization .....................................2

1.1.1 The role of colloid composition and size in metal binding to colloids..........2

1.1.2 Colloid mobilization in soils...............................................................................3

1.2 Mobilization of soluble and colloidal metals in soils ....................................................4

1.2.1 The impact of drying on the mobilization of soluble and colloidal metals....4

1.2.2 The impact of liming on the mobilization of soluble and colloidal

metal(loid)s............................................................................................................5

1.3 Objectives of this thesis.....................................................................................................6

2 A Method for the Determination of Hydrophobicity of suspended Soil

Colloids............................................................................................................9

2.1 Abstract................................................................................................................................9

2.2 Introduction........................................................................................................................10

2.3 Material and Methods........................................................................................................11

2.3.1 Approach for method testing ...........................................................................11

2.3.2 Hydrophobic phases tested for the separation of hydrophobic colloids....12

2.3.2.1 Dichlormethane-Method...................................................................................12

2.3.2.2 C18-Method......................................................................................................13

2.3.3 Calculation of hydrophobicity..........................................................................13

2.3.4 Further optimization of the C18-method.........................................................14

2.4. Results and Discussion...................................................................................................15

2.4.1 Dichlormethane-Method ...................................................................................15

2.4.2 C18-Method ........................................................................................................15

2.5 Conclusion.........................................................................................................................17

viii

3 Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects

of Drying ...................................................................................................... 19

3.1 Abstract ............................................................................................................................. 19

3.2 Introduction....................................................................................................................... 20

3.3 Material and Methods....................................................................................................... 22

3.3.1 Soil samples...................................................................................................... 22

3.3.2 Experimental setup........................................................................................... 24

3.3.3 Analyses ............................................................................................................ 25

3.4 Results and Discussion................................................................................................... 27

3.4.1. Influence of drying on the composition of the dissolved phase ................ 28

3.4.2. Influence of drying on the composition and properties of dispersible

colloids............................................................................................................... 32

3.5 Conclusion ........................................................................................................................ 37

4 Increasing pH releases colloidal Lead in a highly contaminated Forest

Soil................................................................................................................. 39

4.1 Abstract ............................................................................................................................. 39

4.2 Introduction....................................................................................................................... 40

4.3 Material and Methods....................................................................................................... 42

4.3.1 Soil sampling and soil characterization ......................................................... 42

4.3.2 Dispersion experiments................................................................................... 43

4.3.3 Effect of pH increase in the presence of a monovalent counterion............ 44

4.3.4 Effect of pH increase – comparison of counterion valency......................... 45

4.3.5 Analyses ............................................................................................................ 46

4.4 Results and Discussion................................................................................................... 47

4.4.1 Effect of a pH increase in the presence of a monovalent base.................... 47

4.4.1.1 Effects of pH on Pb mobilization...................................................................... 47

4.4.1.2 Effects of pH on colloid properties................................................................... 50

4.4.1.3 Mechanisms of colloid mobilization ................................................................. 53

4.4.2 Effect of pH increase – comparison of counterion valency......................... 54

4.5 Environmental Relevance................................................................................................ 57

5 Lead, Antimony and Arsenic in dissolved and colloidal Fractions from

an amended Shooting Range Soil as characterize by Multi-stage

tangential Ultrafiltration and Centrifugation.............................................. 59

5.1 Abstract ............................................................................................................................. 59

5.2 Introduction....................................................................................................................... 60

5.2.1 Separation methods for the size fractionation of colloids........................... 60

5.2.2 The impact of liming on shooting range sites................................................61

5.2.3 Aim and scope of the work...............................................................................63

5.3 Material and Methods........................................................................................................64

5.3.1 Soil collection and characterization................................................................64

5.3.2 Soil batch extractions .......................................................................................65

5.3.3 Colloid fractionation – membrane filtration...................................................65

5.3.4 Colloid fractionation – centrifugation..............................................................67

5.3.5 Estimation of the fraction of organic- and mineral-dominated colloids –

theoretical background.....................................................................................68

5.3.6 Further characterization of the colloids..........................................................69

5.4 Results and Discussion....................................................................................................70

5.4.1 Multi-stage tangential ultrafiltration ................................................................70

5.4.1.1 Instrumental......................................................................................................70

5.4.1.2 Colloid size and element distribution as determined by MTUF........................70

5.4.2 Centrifugation....................................................................................................76

5.4.2.1 Colloid size and element distribution as determined by centrifugation ............76

5.4.2.2 Turbidity and zeta potential ..............................................................................76

5.4.3 Colloid characterization by comparison and combination of results

obtained by centrifugation and MTUF.............................................................77

5.4.3.1 Estimation of the fraction of organic- and mineral-dominated colloids.............77

5.4.3.2 Particle size measurements by dynamic light scattering (DLS).......................81

5.5 Conclusion.........................................................................................................................83

6 Mobilization of soluble and dispersible Pb, As, and Sb in a polluted,

organic-rich Soil – Effects of pH Increase and Counterion Valency ........85

6.1 Abstract..............................................................................................................................85

6.2 Introduction........................................................................................................................86

6.3 Material and Methods........................................................................................................88

6.3.1 Soil sampling and soil characterization..........................................................88

6.3.2 Dispersion experiments....................................................................................90

6.3.3 Analyses.............................................................................................................91

6.4 Results and Discussion....................................................................................................91

6.4.1 Colloid mobilization...........................................................................................91

6.4.2 Mobilization of soluble and dispersible Pb, As, and Sb................................93

6.4.2.1 Lead..................................................................................................................93

6.4.2.2 Arsenic .............................................................................................................94

6.4.2.3 Antimony...........................................................................................................96

6.5 Conclusion.........................................................................................................................98

7 Synthesis and general Conclusions........................................................... 99

7.1 The governing role of soil solid phase composition in soluble and colloidal

metal(loid) mobilization in the context of drying and liming ....................................... 99

7.1.1 Drying................................................................................................................. 99

7.1.2 Liming .............................................................................................................. 100

7.1.2.1 Lead ............................................................................................................... 103

7.1.2.2 Arsenic ........................................................................................................... 104

7.1.2.3 Antimony ........................................................................................................ 104

7.1.3 Is there a critical zeta potential for colloid mobilization?........................... 105

7.2 General conclusions and derived implications for the management of contaminated

sites.................................................................................................................................. 106

7.3 Outlook ............................................................................................................................ 108

8 References ................................................................................................. 111

Acknowledgements.............................................................................................. 121

Curriculum Vitae................................................................................................... 123

Appendix ................................................................................................................A-1

List of Figures

Figure 2.1: Hydrophobicity of differently treated goethite (as determined with

dichlormethane) as a function of the zeta potential.................................15

Figure 2.2: Hydrophobicity of Al- and P-treated goethite (as determined with C18

microparticles) as a function of the zetapotential....................................16

Figure 2.3: Capacity test of C18 microparticles........................................................17

Figure 3.1: Relation between contact angle, hydrophobicity and wettability ............26

Figure 3.2: Dissolved and colloidal C

org

concentrations in suspensions of field-moist

and air-dried samples ............................................................................29

Figure 3.3a: Dissolved and colloidal Cd concentrations in suspensions of field-moist

and air-dried samples...........................................................................30

Figure 3.3b: Dissolved and colloidal Zn concentrations in suspensions of field-moist

and air-dried samples...........................................................................30

Figure 3.3c: Dissolved and colloidal Cu concentrations in suspensions of field-moist

and air-dried samples...........................................................................31

Figure 3.3d: Dissolved and colloidal Pb concentrations in suspensions of field-moist

and air-dried samples...........................................................................32

Figure 3.4: Turbidity in the soil suspensions of field-moist and air-dried samples....34

Figure 3.5: Contact angle of field-moist and air-dried samples.................................35

Figure 3.6: Hydrophobicity of colloidal suspensions of field-moist and air-dried soil

samples..................................................................................................35

Figure 4.1: Total mobilized Pb concentration as a function of shaking time (soil 1;

pH of soil suspension: 3.9) .............................................................................44

Figure 4.2a-c: Mobilized Pb (a), C (b) and zeta potential (c) of soil suspensions

extracted from soil 1 at various pHs of the KOH treatment (after 1.2

µm filtration).......................................................................................48

Figure 4.3: Optical density of soil suspensions extracted from soil 1 at various pHs

of the KOH treatment (after 1.2 µm filtration).........................................50

Figure 4.4a+b: EDX spectra of colloids dispersed at (a) pH 7 and (b) pH 4 of soil 1

(KOH treatment) ...............................................................................51

Figure 4.5: SEM image of colloids dispersed at pH 4 (soil 1, KOH treatment) .........52

xii

Figure 4.6: Size of mobilized colloids of soil 1 at various pHs of the KOH treatment

(after 1.2 µm filtration)............................................................................ 53

Figure 4.7a: Concentrations of colloidal Pb of soil suspensions extracted from

soil 2 at various pHs by KOH and Ca(OH)

(after 1.2 µm filtration) ......55

Figure 4.7b: Concentrations of dissolved Pb of soil suspensions extracted from

soil 2 at various pHs by KOH and Ca(OH)

(after 1.2 µm filtration) .....56

Figure 5.1: Elemental size distribution of the control sample (pH 4.7) following

colloid separation by MTUF ................................................................... 71

Figure 5.2: Elemental size distribution of the KOH treatment (pH 6.5 ± 0.1))

following colloid separation by MTUF..................................................... 72

Figure 5.3: EDX spectra of the KOH treatment........................................................ 73

Figure 5.4: Elemental size distribution of the Ca(OH)

treatment (pH 6.5 ± 0.1))

following colloid separation by MTUF..................................................... 75

Figure 5.5: EDX spectra of the Ca(OH)

treatment.................................................. 76

Figure 5.6: Comparison of the elemental size distribution of the control treatment

between results gained from MTUF and centrifugation.......................... 78

Figure 5.7: Comparison of the elemental size distribution of the KOH treatment

between results gained from MTUF and centrifugation.......................... 80

Figure 5.8: Comparison of the elemental size distribution of the Ca(OH)

treatment

between results gained from MTUF and centrifugation.......................... 81

Figure 6.1: Turbidity as a function of pH for the KOH and Ca(OH)

treatment.......... 92

Figure 6.2: Zeta potential as a function of pH for the KOH and Ca(OH)

treatment. 92

Figure 6.3a: Mobilized dissolved Pb as a function of pH for the Ca(OH)

and KOH

treatment.............................................................................................. 93

Figure 6.3b: Mobilized colloidal Pb as a function of pH for the KOH and Ca(OH)

treatment............................................................................................. 94

Figure 6.4: Mobilized dissolved and colloidal As as a function of pH for the KOH

and Ca(OH)

treatment.......................................................................... 95

Figure 6.5: Mobilized dissolved and colloidal Sb as a function of pH for the KOH and

Ca(OH)

treatment................................................................................. 97

xiii

Figure 7.1: Turbidity of soil suspensions of soil 1 and 2 with a high and a low

concentration of organic matter as a function of pH

(Ca(OH)

treatment)..............................................................................102

Figure 7.2:

Zeta potential of soil 1 and soil 2 with a high and a low concentration of

organic matter as a function of pH (Ca(OH)

treatment).......................102

Figure 7.3:

Turbidity as a function of the zeta potential for all soils.........................106

Figure 7.4:

Relation between the soil solid phase, soil sorbents and colloidal and

dissolved metal(loid)s...........................................................................107

xiv

List of Tables

Table 2.1: Result of the comparison between goethite particles sorbed on DOC-

treated and untreated C18 microparticles .....………………………………17

Table 3.1: Properties of the soil samples used for the experiments……………....... 24

Table 3.2: Significance of the effects of soil drying on the given soil characteristics.27

Table 3.3: pH and electrical conductivity in suspensions (s:w ratio: 1:10) of field-

moist and air-dried soil samples.....………………………………………....28

Table 4.1: Some properties of the analysed soils.....………..…………………………42

Table 5.1: Some physico-chemical parameters, total heavy metal and oxalate-

extractable Fe (Fe

) and Al (Al

) concentrations [mg kg

-1

] of the

used soil, determined as duplicate.…………………………………………64

Table 5.2: MTUF recovery rates for the individual elements, averaged over the

entire experiment..................................................................................... 70

Table 5.3: Dissolved and colloidal concentrations of the individual elements in the

different treatments as obtained by MTUF....…………………...…………71

Table 5.4: Zeta potential, Fe and DOC concentration of the supernatants of the

different treatments (centrifugation).....……………..………………...…….77

Table 5.5: Percentage of the individual element associated with mineral-

dominated and organo-dominated colloids in the respective size

fractions of MTUF and centrifugation as obtained by duplicate samples

of the KOH and Ca(OH)

treatment....…………………….…….…….…....79

Table 5.6: Cut-off for different densities based on the same centrifugation

parameters.............................................................................................. 82

Table 6.1: Soil properties ....……………………………………….……………..………89

Table 7.1: C

org

concentration and C

org

/Fe

tot

-ratio of the soils used in chapter 4

and 6.......……………………………………………..………………………100

List of Abbreviations

B: Buch soil

COC: colloidal organic carbon

DCM: Dichlormethane

DOC: dissolved organic carbon

DOM: dissolved organic matter

FFF: field-flow fractionation

FNU: Formazin nephlometric unit (unit of turbidity)

GS: Gütersloh soil

MED: Molarity of Ethanol droplet

MTUF: Multi-stage tangential ultrafiltration

OM: organic matter

TG: Tiergarten soil

TOC: total organic carbon

WDPT: Water droplet penetration time

xvi

Abstract

Colloidal heavy metals are known to be easily translocatable along preferential flow paths in soils and

therefore pose a significant risk of groundwater pollution. To estimate potential leaching of metals

from contaminated sites not only in dissolved but also in colloidal form an understanding of underlying

mechanisms governing dispersible metal mobilization is crucial. In addition the knowledge of colloid

composition is important. The presented thesis investigates these physico-chemical mechanisms of

soluble and dispersible metal mobilization in two examples, namely (1) the impact of drying and

rewetting of soils on the mobilization of soluble and dispersible Cd, Cu, Pb, and Zn and (2) the impact

of soil liming on the mobilization of soluble and dispersible Pb, As, and Sb. The quality of colloids was

considered in particular.

I tested the hypothesis that 1) drying of soil samples increases not only the hydrophobicity of the solid

phase but also the hydrophobicity of the dispersed colloids and that 2) drying of soil samples leads to

an increase in colloidal and dissolved metal concentrations. In the context of liming, I tested the

hypothesis that 3) the pH-induced increase in particle charge leads to a greater dispersion of colloids

with increasing pH and thus increases colloidal Pb, As and Sb. Furthermore I hypothesized that 4) the

addition of Ca

reduces this pH effect by enhancing colloid flocculation. In addition, I tested the

assumption that 5) both a change in pH and counterion valency impact on the quality and size

distribution of colloids. I investigated 6) the influence of a pH increase on the mobilization of soluble

and dispersible Pb, As, and Sb in a polluted organic-rich soil and tested the hypothesis that 7)

divalent cations attenuate the pH effect by a) “neutralizing” negative colloid charge and by b)

immobilizing dissolved As and Sb by the formation of inorganic precipitates. Before I could test my

hypotheses the development of methods to characterize colloid properties was essential: The C18-

method determines the hydrophobicity of dispersed colloids and the development of a multi-stage

tangential ultrafiltration (MTUF) method serves the characterization of the separated colloids and

associated metal(loid)s in soil suspensions. In addition, I developed an approach to estimate the

percentage of dispersed organic-dominated and mineral-dominated colloids by combining results

gained by MTUF and centrifugation. I tested all hypotheses on filtered aqueous extracts (1.2 µm) of

batch experiments.

To test hypotheses 1 and 2 I determined turbidity, zeta potential, hydrophobicity and size of dispersed

colloids of field-moist and air-dried topsoil samples of a former sewage farm. I measured dissolved

and colloidal concentrations of heavy metals and organic C in the suspensions and determined the

hydrophobicity of the soil solid phase.

Whereas drying of soil led to an increase in hydrophobicity of the solid phase, it did not increase the

hydrophobicity of dispersed colloids. Drying led to the mobilization of organic colloids, but immobilized

(organo-)mineral colloids. While concentrations of the investigated dissolved heavy metals increased

Abstract

xvii

in all soil samples, concentrations of colloidal heavy metals decreased in almost all soil samples for

Cd and Zn, in some soils for Cu and Pb.

I conducted the experiments related to liming with topsoil samples from former shooting ranges and

increased the pH of the batch samples by adding dilute solutions of either a monovalent (KOH) or a

divalent Ca(OH)

base. I used an organic-rich soil (C

org

: 21 %) to test hypotheses 3 and 4.

Hypotheses 3 – 5 were tested on a soil with a low C/Fe-ratio (C

org

: 9 %) while experiments related to

6 and 7 were conducted on a soil with a high C/Fe-ratio (C

org

: 8 %). Soil colloid fractionation was done

using the developed MTUF (hypothesis 5). In the suspensions, I measured zeta potential, particle

size, turbidity and dissolved and colloidal concentrations of Pb, As, Sb, Fe, and organic C. Results of

the organic-rich soil showed that colloids may be stabilized by charge and by steric effects. The

combination of both effects renders dispersed organic colloids at pH 4 much more stable than colloids

dispersed from soils with less organic C. While concentrations of colloidal Pb increased continuously

in the presence of KOH, they only increased at pH > 5.8 in the presence of Ca(OH)

due to charge

compensation of the negatively charged colloids by the divalent Ca cation. Colloid characterization by

MTUF showed that the valency of the counterion controls quantity, elemental composition and size of

the colloids. Increasing the pH using KOH led to the dispersion of sesquioxides and organic colloids

(9 nm – 220 nm). Colloidal As and Pb were found to be associated with sesquioxides which may in

part be stabilized by organic C. Increasing the pH using Ca(OH)

suppressed the dispersion of

sesquioxides and induced the formation of larger colloids (0.22 – 1.2 µm) such as precipitation

products of Pb and “bridging”-products of Ca and DOM, which may also include some Pb. While my

data of the high C/Fe-soil excludes the formation of inorganic precipitates it suggests that the

counterion valency controls the mobility of the sorbents (i.e. colloids, DOM) and the sorption capacity

of the sorbents. A liming-induced pH increase does not pose any major risk of dispersible As and Sb

mobilization as long as divalent cations dominate in solution. However, the mobilization of soluble Sb

needs to be considered.

My results showed that different mechanisms are responsible for the mobilization of soluble and

dispersible metal(loid)s in both a liming-induced pH increase as well as in a drying event. My data

suggest that the composition of the soil solid phase, i.e. C

org

and sesquioxide content, may control

colloid release and the distribution of the respective metal between soluble and colloidal phases.

Future studies should investigate more soils across a wider range of composition in order to gain a

better understanding about the role of the composition of the soil sorbents not only in soluble but also

in dispersible metal(loid) mobilization. In addition, field experiments should clarify the relevance of the

identified mechanisms to colloidal metal leaching.

xviii

Zusammenfassung

Kolloidale Schwermetalle sind dafür bekannt, dass sie entlang präferentieller Fließwege im Boden

leicht verlagerbar sind. Daher stellen sie ein bedeutendes Risiko für die Verschmutzung des

Grundwassers dar. Für die Abschätzung möglicher Schwermetallauswaschungen an kontaminierten

Standorten - nicht nur in gelöster sondern auch in kolloidaler Form - ist ein Verständnis der

Mobilisierungsmechanismen löslicher sowie dispergierbarer Metallionen unabdingbar. Darüber hinaus

sind Kenntnisse über die Zusammensetzung der Kolloide wichtig. Gegenstand der vorgelegten

Doktorabeit ist die Untersuchung dieser physiko-chemischen Mobilisierungsmechanismen löslicher

und dispergierbarer Metalle anhand zweier Beispielfälle: (1) Einfluss einer Trocknung und

Wiederbefeuchtung von Böden auf die Mobilisierung von löslichem und dispergierbarem Cd, Cu, Pb

und Zn sowie (2) Einfluss einer Bodenkalkung auf die Mobilisierung von löslichem und

dispergierbarem Pb, As und Sb unter besonderer Berücksichtigung der Kolloidzusammensetzung.

Ich überprüfte die Frage, ob 1) Trocknung von Bodenproben nicht nur die Hydrophobizität der

Festphase, sondern auch die der dispergierten Kolloide erhöht und 2) ob Trocknung von

Bodenproben zu einer Erhöhung der gelösten und kolloidalen Metallkonzentrationen führt. Im

Zusammenhang einer Kalkung überprüfte ich die Frage, ob 3) bei steigendem pH-Wert die dadurch

induzierte Erhöhung der Partikelladung zu einer stärkeren Dispergierung von Kolloiden führt und es

deshalb zu einer Zunahme an kolloidalem Pb, As und Sb kommt. Darüber hinaus nahm ich an, dass

4) die Zugabe von Ca

die Wirkung des pH-Wertes durch Begünstigung von Kolloidflockung

verringert. Zusätzlich überprüfte ich die Annahme, dass 5) sowohl eine Änderung des pH-Wertes als

auch der Valenz der Gegenionen die Qualität und die Größenverteilung der Kolloide beeinflussen. Ich

untersuchte 6) den Einfluss eines pH-Anstieges auf die Mobilisierung von löslichem und

dispergierbarem Pb, As und Sb in einem kontaminierten Boden, der reich an organischer Substanz

ist. Dabei überprüfte ich die Frage, ob 7) zweiwertige Kationen die Wirkung des pH-Wertes

abschwächen durch a) ein „Neutralisieren“ der negativen Kolloidladung und durch b) Immobilisierung

gelösten As und Sb durch die Bildung anorganischer Präzipitate. Ein wesentlicher Aspekt der Arbeit

war dabei die Entwicklung von Methoden zur Kolloidcharakterisierung: Die C-18-Methode erlaubt die

Bestimmung der Hydrophobizität von Kolloiden und multi-stage tangentiale Ultrafiltration (MTUF)

gestattet die Charakterisierung aufgetrennter Kolloide und daran assoziierter (Halb-)Metalle in

Bodensuspensionen. Zusätzlich entwickelte ich einen Ansatz, der durch eine Kombination von

Ergebnissen aus MTUF und Zentrifugation die Abschätzung des Prozentsatzes an dispergierten

organisch- und mineralisch-dominierten Kolloiden ermöglicht. Alle Fragen wurden an gefilterten

wässrigen Extrakten (1.2 µm) von Schüttelversuchen überprüft.

Um die Fragen 1 und 2 zu überprüfen bestimmte ich Trübe, Zetapotential, Hydrophobizität und Größe

der dispergierten Kolloide feldfrischer und luftgetrockneter Böden eines Rieselfeldes. Ich maß gelöste

xix

und kolloidale Konzentrationen an Schwermetallen und organischem C in den Suspensionen und

bestimmte die Hydrophobizität der Bodenfestphase.

Während Trocknung zu einem Anstieg der Hydrophobizität der Festphase führte, erhöhte sich die

Hydrophobizität der dispergierten Kolloide nicht. Trocknung mobilisierte organische Kolloide, aber

immobilisierte (organo-)mineralische Kolloide. Die Konzentrationen an kolloidalen Schwermetallen

nahm bezüglich Cd und Zn in fast allen Böden und bezüglich Cu und Pb in einigen Böden ab,

wohingegen die Konzentrationen an den untersuchten gelösten Schwermetallen anstieg.

Die auf die Kalkung bezogenen Versuche führte ich mit Oberbodenproben von Schießplätzen durch

und erhöhte dabei den pH-Wert der Schüttelproben durch die Zugabe von verdünnten Lösungen

einer einwertigen (KOH) oder zweiwertigen (Ca(OH)

) Base. Die Fragen 3 und 4 untersuchte ich an

einem an organischer Substanz reichen Boden (C

org

: 21 %). Die Fragen 3 bis 5 wurden an einem

Boden mit einem niedrigen C/Fe-Verhältnis (C

org

: 9 %) überprüft, während auf Frage 6 und 7

bezogene Versuche an einem Boden mit einem hohen C/Fe-Verhältnis (C

org

: 8 %) durchgeführt

wurden. Die zur Überprüfung von Frage 5 erforderliche Fraktionierung der Bodenkolloide erfolgte

durch die entwickelte MTUF-Methode. In den Suspensionen bestimmte ich Zetapotential,

Partikelgröße, Trübung sowie gelöste und kolloidale Konzentrationen an Pb, As, Sb, Fe und

organischem C.

Die Ergebnisse des an organischer Substanz reichen Bodens zeigten, dass Kolloide aufgrund ihrer

Ladung als auch durch sterische Effekte stabilisiert werden. Die Kombination beider Effekte

stabilisiert dispergierte organische Kolloide bei pH 4 weitaus besser als Kolloide, die weniger C

enthalten. Während in der Anwesenheit von KOH die Konzentrationen an kolloidalem Pb

kontinuierlich anstiegen, erhöhten sie sich in Anwesenheit von Ca(OH)

erst bei pH-Werten > 5,8.

Letzteres ist auf eine Kompensation der negativen Kolloidladung durch das zweiwertige Ca-Kation

zurückzuführen. Die Charakterisierung der Kolloide mittels MTUF zeigte, dass die Valenz der

Gegenionen die Menge, die elementare Zusammensetzung als auch die Größe der Kolloide steuert.

Eine pH-Erhöhung durch KOH führte zur Dispergierung von Sesquioxiden und organischen Kolloiden

(9 nm – 220 nm). Kolloidales As und Pb sind dabei mit Sesquioxiden assoziiert, die teilweise durch

organischen C stabilisiert sein können. Eine pH-Erhöhung durch Ca(OH)

unterdrückte die

Dispergierung der Sesquioxide und induzierte die Bildung von größeren Kolloiden (0,22 – 1,2 µm) wie

Fällungsprodukte von Pb und „Brückenbindungsprodukte“ von Ca und DOM, die auch Pb enthalten

können. Während die Daten des Bodens mit dem hohen C/Fe-Verhältnisse die Bildung anorganischer

Präzipitate ausschließen, deuten sie darauf hin, dass die Valenz der Gegenionen die Mobilität der

Sorbenten (d. h. Kolloide und DOM) als auch die Sorptionskapazität der Sorbenten steuert. Ein

kalkungsinduzierter pH-Anstieg stellt kein Risiko für die Mobilisierung dispergierbaren As und Sb dar,

so lange zweiwertige Kationen die Lösung dominieren. Allerdings darf die Mobilisierung von löslichem

Sb nicht außer Acht gelassen werden.

Meine Ergebnisse zeigen, dass unterschiedliche Mechanismen für die Mobilisierung von löslichen

und dispergierbaren (Halb-)Metallen bei einem kalkungsinduzierten pH-Anstieg als auch bei einer

Trocknung verantwortlich sind. Sie weisen darauf hin, dass die Zusammensetzung der

Bodenfestphase, d. h. der C

org

- und Sesquioxidgehalt, die Kolloidfreisetzung als auch die Verteilung

des jeweiligen (Halb-)Metalls zwischen gelöster und kolloidaler Phase steuern kann. Zukünftige

Studien sollten mehr Böden über eine breitere Zusammensetzung hinweg untersuchen, um ein

besseres Verständnis über die Rolle der Zusammensetzung der Bodensorbenten - nicht nur bei der

Mobilisierung von löslichen sondern auch von dispergierbaren (Halb-)Metallen - zu erlangen.

Ergänzend sollten Feldversuche die Relevanz der identifizierten Mechanismen zur Auswaschung

kolloidaler Metalle klären.

1 General Introduction

Heavy metal contamination of soil through both anthropogenic and natural sources

has occurred throughout the world and has been recognized as a global problem. It

may pose a significant risk to human (Mani and Kumar, 2005) and environmental

health through e.g. groundwater pollution (Sorvari et al., 2006) and toxicosis of

wildlife (Lewis et al., 2001). Sewage farms (Blume et al., 1980) and shooting ranges

(Neite et al., 1999) are examples for sites severely contaminated by heavy metals.

According to Herms and Brümmer (1984) the ecological relevance of heavy metals

in soils is mainly controlled by their soluble and labile pools, representing those

fractions which are easily bioavailable and translocatable. Therefore, solubility and

bioavailability of heavy metals have become the main focus of many environmental

studies. While these studies commonly investigated dissolved metal fractions

(generally defined as < 0.45 µm) only (for instance Tack et al., 1999) recently

published results have shown that a significant fraction of conventionally defined

dissolved fractions is actually existent in colloidal form (Douglas et al., 1993;

Sanudo-Wilhelmy et al., 1996; Lyvén et al., 2003), which McCarthy and Zachara

(1989) describe – beside the solid and dissolved phase – as a third phase in the

subsurface environment. The association of heavy metals and metalloids with

mobilized colloids is controlled by the affinity of the respective metal to the colloids

(Pourret et al., 2007), the properties of colloids (Lyvén et al., 2003) and soil solution.

The distribution of a metal between these phases will determine its fate and

transport in the environment (Hassellöv et al., 1999).

Due to their small size (Brezesinski and Mögel 1993; Pourret et al., 2007) and their

large surface to volume ratio (Kretzschmar et al., 1999) colloids tend to be stable in

suspensions. As a result, colloids are hardly filtered by porous media (i. e. physically

and physico-chemically) if they travel through preferential flow paths (Seta and

Karathanasis, 1997). Unlike truly dissolved metals colloid-associated metals may

therefore easily be translocated into deeper parts of the soil and groundwater.

Several authors (Denaix et al., 2001; Jensen et al., 1999) even say that colloids

facilitate the subsurface transport and mobilization of heavy metals at contaminated

General Introduction

sites. Failure to consider this pathway may thus result in an underestimation of not

only distances which contaminants may possibly migrate but also of the amount of

contaminants which is being translocated with the water flow (McCarthy and

Zachara, 1989). The different behaviour of “truly” dissolved and colloid-associated

heavy metals in soil and aquatic environments requires a differentiation between

these two forms of metals in order to gain a realistic understanding about potential

metal release, spatial distribution, bioavailability and the ultimate impact on human

health.

To understand the relevance of colloidal metal mobilization to potential metal

leaching on contaminated sites the knowledge of the underlying mechanism is

crucial. With respect to the binding of metals to colloids the properties of colloids are

of vital importance. Prevailing environmental conditions, on the other hand, may

control the mechanisms which govern colloid release and stability.

1.1 Soil colloid characteristics, characterization and mobilization

Colloids are characterized by properties such as composition, surface charge,

surface area, size and shape (Ranville et al., 1999). These parameters govern

colloid interactions with pollutants as well as colloid mobilization and stability.

1.1.1 The role of colloid composition and size in metal binding to colloids

Colloid composition determines – together with physico-chemical parameters of the

soil solution such as pH and ionic strength – the surface charge. Brady and Weil

(2002) describe colloid surface charge to be responsible for the sorption of heavy

metals. Several studies underline the relevance of colloid composition in order to

estimate the fate and bioavailability of associated elements. Pourret et al. (2007)

differentiate between an organic and inorganic colloidal pool. The authors found that

individual metals are partitioned in these two colloidal pools. An important role was

attributed to colloidal organic matter since a large fraction of trace metals was found

to be associated with it (Tanizaki et al., 1992; Dupré et al., 1999). Similarly, Lyvén et

al. (2003) report Fe- and organic C-based colloids to be the most important potential

carriers for other elements. Not only differ both types of colloidal carriers in their size

General Introduction

distribution but also in their trace metal content. They conclude metals to have

varying affinities to these different types of colloidal pools. Colloid composition,

together with the type of bonding the metal is linked to the colloid, control trace

element geochemical cycles or their fate in the environment (Pourret et al., 2007).

Kretzschmar et al. (1999) describe the particle size as one of the most important

characteristics of mobile colloids since they determine the specific surface area of

colloidal particles, which controls the degree of contaminant sorption. The size

fractionation of soil colloids together with the characterization of their elemental

composition may thus provide important information for a better understanding and

prediction of their role as carriers of different pollutants (Buffle and Leppard, 1995).

Therefore, the use of powerful methods for suspended soil colloid fractionation is

crucial. Since the stability of suspended soil colloids is very much subject to changes

in ionic strength, separation techniques which cause little disturbance of sample

equilibrium are an indispensible prerequisite for a successful fractionation. This

prerequisite is met by common fractionation techniques such as ultrafiltration and

centrifugation (Nifant’eva et al., 2001), which are considered a powerful tool for the

investigation of colloidal particles in natural environments (Kretzschmar et al., 1999).

While the investigation of colloid composition with respect to the bindings of metals

has already gained some attention there are no studies focusing on colloid

composition in the context of colloid mobilization.

1.1.2 Colloid mobilization in soils

Colloid charge is known to be a key parameter for colloid stability (Séquaris and

Lewandowski, 2003). In addition, size and shape are other important factors

controlling colloid transport and deposition in soil (Ranville et al., 1999). Therefore,

colloid properties such as charge and size play an important role in colloid

mobilization. In order to estimate potential leaching of metals from contaminated

sites not only in dissolved but also in colloidal form the knowledge of underlying

mechanisms governing colloidal metal mobilization is crucial. Colloid mobilization is

controlled by physical factors such as shear forces (Kaplan et al., 1993; Rousseau et

General Introduction

al., 2004) as well as (physico)-chemical paramters of the soil solid (Kretzschmar et

al., 1993) and solution phase (Kaplan et al., 1996). However, the dominating

mechanisms may vary for different environmental conditions. The presented thesis

investigates these physico-chemical mechanisms of colloidal metal mobilization in

two exemplary scenarios, namely (1) drying and rewetting of soils and (2) liming of

soil in the context of soil stabilization and remediation. Both, drying as well as liming,

are known to be conducive to the potential mobilization of colloid-associated heavy

metals.

1.2 Mobilization of soluble and colloidal metals in soils

1.2.1 The impact of drying on the mobilization of soluble and colloidal metals

Under natural conditions almost any soil is subject to wetting and drying cycles.

Drying may alter the physico-chemical properties of the soil such as conformation of

soil organic matter (Hurraß and Schaumann, 2006) and hydrophobicity (Dörr et al.,

2000; Dekker et al., 2001). Upon rewetting, these changed physico-chemical

properties of the soil may affect the composition of the soil solution, which is

reflected – amongst others – in increasing dissolved concentrations of organic

carbon and dissolved heavy metals (Wang et al., 2002; Tom-Petersen et al., 2004).

However, at the moment there is only little knowledge about the impact of wetting

and drying cycles on the physico-chemical properties of colloids. Wan and Wilson

(1994a) suggest colloid hydrophobicity to be an important parameter controlling

colloid retention in soil. The authors observed an increasing retention of colloidal

latex particles and bacteria with increasing particle hydrophobicity. Furthermore,

Schaumann et al. (2002) reported that the state of moisture of soil samples

containing organic matter affects colloidal characteristics of aqueous soil extracts.

The authors found increasing colloid particle sizes and decreasing molecular sizes of

humic associates following drying. These drying-induced changes of physico-

chemical properties may alter transport parameters of both colloidal and dissolved

phase. Therefore, the understanding of possible differences in colloid mobilization

and associated heavy metals from dried and field-moist samples together with the

General Introduction

underlying mechanism is essential not only for the design of batch and column

experiments but also for the risk assessment of contaminated sites.

1.2.2 The impact of liming on the mobilization of soluble and colloidal

metal(loid)s

Liming is a common method suggested for soil remediation and stabilization of

shooting range soils (EPA, 2001). Whilst drying as a physical impact factor affects

soil solution chemistry indirectly, the application of lime onto the soil directly affects

the chemical properties of the soil solution as well as the solid phase. The induced

pH-increase of the soil solution is meant to immobilize or precipitate heavy metals in

soils (Lindsay, 1979). However, the increase in pH as well as the possibly resulting

mobilization of dissolved organic matter may promote the mobilization of colloids

(Kretzschmar et al., 1993; Kaplan et al., 1996). On the other hand, the addition of Ca

ions into the soil leads to a compression of the diffusive double layer (McBride,

1994) facilitating colloid flocculation. While both processes have been studied

individually in depth (Kretzschmar et al., 1993), their net effect on colloid mobilization

is yet unclear.

The metals of main concern on shooting range soils are Pb, As and Sb. While the

relevance of colloidal Pb (Denaix et al., 2001; Jensen et al., 1999) and the impact of

liming on the mobilization of soluble Pb is well documented (Turpeinen et al., 2000),

studies on the pH-induced mobilization of colloidal Pb are missing.

Since numerous publications report increasing dissolved As concentrations with

increasing pH (for instance Tyler and Olsson, 2001; Xu et al., 1988) significant

effects of lime application are to be expected. However, there is only little knowledge

about As in association with colloids. Slowey et al. (2007) found As(V) sorbed to

poorly crystalline Fe(III)-hydroxides and Tanizaki et al. (1992) demonstrated As not

only to be present as simple ion but also associated with small (10

- 500 Da)

inorganic colloids. Similarly, Slowey et al. (2007) reported the occurrence of As on

colloids < 20 nm and ~ 100 nm.

General Introduction

Similar to As, up to now, there is only little evidence of the occurrence of colloidal

Sb. Buddemeier and Hunt (1988) found

125

Sb to be associated with colloids of a size

range between 3 and 50 nm and Lyvén et al. (2003) report Sb linked with organic

carbon colloids. Since Sb is reported to be strongly sorbed to Fe-oxides at pH-values

< 6.5 (Tighe et al. 2005; Edwards et al., 1999) colloid-associated Sb might become

relevant in a liming induced-pH increase.

1. 2 Objectives of this thesis

This thesis aims to elucidate the mobilization of soluble and colloidal metal(loid)s

under the influence of two different environmental factors. It investigates the

underlying mechanisms and characterizes colloid properties and composition using

batch experiments by focusing on the following two aspects:

1) Impact of drying on the amount and properties of colloids dispersed as well as

the mobilization of soluble and colloid-associated heavy metals (Pb, Cd, Cu,

Zn) on a sewage farm at Berlin-Buch.

2) Impact of liming on the mobilization of soluble and colloid-associated Pb, As

and Sb in soils from former shooting range sites.

In order to better understand the role of hydrophobicity and the composition of the

mobilized colloids, the development of appropriate methods was essential. The

development of a method for the determination of hydrophobicity of suspended soil

colloids is described in chapter 2. The development of a multi-stage tangential

ultrafiltration (MTUF) method, which lives up to the requirements of the separation of

colloids in soil suspension, served the characterization of the separated colloids

together with associated metal(loid)s. The application of MTUF is meant to elucidate

the role of differently composed colloids in metal(loid) mobilization. Furthermore, it

demonstrates that a combination of results gained by MTUF and centrifugation may

be drawn on in order to provide an estimate for (a) colloids mainly composed of

mineral and (b) colloids predominantly consisting of organic matter. The results of

this study are shown in chapter 5.

General Introduction

Chapter 2 describes the development of a method using C18 column material which

serves the determination of hydrophobicity of suspended soil colloids. This method

provides the basis for the following study of the underlying mechanism of colloids

released after drying.

In chapter 3 I investigated whether (a) drying of soil samples increases not only the

hydrophobicity of the solid phase but also the hydrophobicity of the dispersed

colloids and whether (b) drying of soil samples increases the colloid-bound and

dissolved metal fractions.

Chapter 4 is dedicated to the study of the effect of pH and the role of counterion

valency on the mobilization of soluble and colloidal Pb. In addition, it elucidates the

mechanisms which control colloid release as a function of pH and counterion

valcency. In the context of liming, I tested the hypothesis that the pH-induced

increase in particle charge leads to a greater dispersion of colloids with increasing

pH and thus increases mobilization of colloidal Pb. Furthermore, I hypothesized that

the addition of Ca

counteracts the pH effect, thus resulting in lower colloidal Pb

concentrations.

Chapter 5 introduces the developed design of a multi-stage tangential ultrafiltration

system suitable for the fractionation of suspended soil colloids. It compares results

obtained by MTUF and centrifugation and demonstrates how a combination of both

results allows for an estimate of colloids dominated by organic and mineral material.

Furthermore, I hypothesized that increasing pH leads to an increased mobilization of

colloidal Pb, As and Sb and that the addition of Ca

reduces this pH-effect. In

addition, I tested the assumption that both a change in pH and counterion valency

impact on the quality and size distribution of colloids. These hypotheses were tested

on a soil with a low C/Fe-ratio.

Chapter 6 investigates the influence of a pH-increase on the mobilization of soluble

and colloidal Pb, As, and Sb in a polluted organic-rich soil (i. e. a soil with a high

C/Fe-ratio). I tested the hypothesis whether divalent cations attenuate the pH-effect

General Introduction

by a) “neutralizing” negative colloid charge and b) immobilizing dissolved As and Sb

by the formation of inorganic precipitates.

The mobilization of heavy metals and metalloids is the subject of the following

chapters. In order to distinguish between the different forms of mobilized metals they

are denoted “dissolved” and “colloidal”.

2 A Method for the Determination of Hydrophobicity

of suspended Soil Colloids

Sondra Klitzke and Friederike Lang

Published in: Colloids and Surfaces A: Physicochemical and Engineering Aspects

2.1 Abstract

Colloids play a crucial role in the translocation of trace elements in soils. Recent

studies provided hints that colloid hydrophobicity may be an important factor

controlling colloid (im)mobilization in soils. However, existing methods for the

determination of hydrophobicity are limited to the bulk soil. Therefore, we developed

a method to determine the hydrophobicity of suspended colloids in aqueous soil

suspensions, which was based on a distribution between a polar and a non-polar

phase. The proposed method uses 30 mg of an unpolar solid phase (C18-column

material) which are mixed with 10 mL of suspension for 2 hours. The turbidity of the

suspensions is measured before and after mixing. The ratio of the colloids in the

hydrophilic aqueous and the hydrophobic solid phase is calculated as a measure of

colloid hydrophobicity. This method was successfully tested on differently

hydrophobized goethite particles. At DOC concentrations exceeding 20 mg l

-1

organic molecules sorbed to C18-material limit the applicability of the method.

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

2.2 Introduction

Colloids play a crucial role in the translocation of heavy metals (Keller and

Domergue, 1996; Jensen et al., 1999; Denaix et al., 2001) and organic contaminants

(Brown and Peake, 2003; Bergendahl, 2005; Totsche et al. 2006) in soils. Their

mobilization and stability in soils is controlled by physico-chemical factors such as

pH, ionic strength and DOM concentration in the soil solution, as well as steric

effects (Dekker et al., 2001). Furthermore, there are hints in the literature that

hydrophobicity is another important factor controlling colloid retention, stability and

sorption capacity. Wan and Wilson (1994a) demonstrated an increased retention of

colloidal latex particles and bacteria with increasing particle hydrophobicity in an

unsaturated sand column experiment. Similarly, under saturated conditions

hydrophobic colloids showed a lower recovery than hydrophilic colloids, allowing the

authors to conclude hydrophilic colloids are more mobile than hydrophobic ones

(Wan and Wilson, 1994a) as they sorb less strongly to both the gas-water and solid-

water interfaces. Breiner et al. (2006) reported hydrophobic organic molecules affect

colloid stability in aqueous solutions by altering the surface properties of colloids.

They postulate an enhanced hydrophobic nature of the colloids would reduce their

transport in the environment. Another parameter hydrophobicity has an effect on is

the sorption capacity of colloids. According to Liu and Lee (2006) hydrophobicity is

thought to enhance the sorption capacity of colloids for hydrophobic organic

compounds. Breiner et al. (2006) suggest organic matter-coated inorganic colloids

facilitate the sorption of hydrophobic organic contaminants. Different conditions are

reported to induce hydrophobicity, for instance drying, this being explained by

changes in the molecular conformation of the organic matter (Liu and Lee, 2006;

Ma’shum and Farmer, 1985). In addition, McHale et al. (2005) found wax coatings

arising from vegetation to render the surface of small soil particles hydrophobic.

Hydrophobicity is inversely correlated with the wettability of water (Doerr et al.,

2000). For pure mineral colloids the surface charge is closely related to the

hydrophobicity of colloids. In contrast, organo-mineral colloids of similar surface

charge may vary strongly in hydrophobicity, depending on the accessibility of

functional groups to surrounding water. For these colloids, hydrophobicity might be a

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

more suitable indicator for the estimation of sorption capacity and mobility than

surface charge.

There are a number of methods for determining the hydrophobicity of the soil solid

phase such as the water droplet penetration time test (WDPT; Doerr et al., 2000;

Letey, 1969), the Wilhelmy plate method as described by Bachmann et al. (2003)

and the molarity of ethanol droplet (MED) test (Roy and McGill, 2002). All these

methods, however, can not be applied to dispersed colloids in soil suspensions.

Recently, Guiné et al. (2003) proposed a method to determine the hydrophobicity of

colloidal bacteria in aqueous suspensions using hexadecane as an extracting

solvent. We adopted this approach to colloidal soil suspensions (1:10 water extracts)

in a preliminary experiment (unpublished data) but results indicated that this method

can not be used for colloidal soil suspensions due to incomplete separation of the

aqueous phase and the organic solvent. This incomplete separation may be

attributed to the presence of surface-active substances in the soil suspension (such

as dissolved organic matter), which cause an emulsion, rendering a complete phase

separation impossible. Therefore, the aim of this paper is to discuss a newly

developed method for the determination of the hydrophobicity of suspended soil

colloids in soil extracts.

2.3 Material and Methods

The two tested methods described in the following are based on a distribution of

suspended hydrophobic goethite particles between (a) two liquid phases

(Dichlormethane-method) and (b) a hydrophobic solid phase (C18-method).

2.3.1 Approach for method testing

In view of the findings from van Oss and Giese (1995) that interfacial interactions

control hydrophobicity we based the suitability test of the methods described in the

following sections on the assumption that there exists a relationship between

hydrophobicity and the surface charge of the mineral particle. That is: the higher the

charge of the particle, the higher the tendency to remain in suspension. The closer

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

the particle charge to zero, the worse the hydration process and the higher the

tendency to “escape” from water, i. e. attraction by an unpolar phase increases.

Hence, if the methods tested proved successful, the resulting hydrophobicity values

should increase with decreasing amount of the zeta potential of the colloids.

We applied the methods to differently treated goethite particles in aqueous

suspensions. 50 mg goethite was suspended in water, using ultrasonic treatment for

30 min followed by the addition of appropriate quantities of stock solutions [NaH

(c(PO4) = 3800 mg L

-1

), Mg(NO

)

, Al(NO

)

c(Mg, Al) = 1000 mg L

-1

)]. The volume

was made up to 1 L using deionised water. Final concentrations of the individual

solutions were c(PO

) = c(Al) = 80 µM, c(Mg) = 120 µM). For the Na treatment,

goethite was “washed” in dilluted NaOH solution (c = 1 µM) before being suspended

in water. The mixtures were allowed to equilibrate for 16 hours whilst shaking.

Aliquots of all suspensions were adjusted to pH 2 – 8 using NaOH and HNO

(c =

0.1 M) and the zeta potential of the colloids was determined (Zetasizer 2000,

Malvern) after completed equilibration.

The goethite was synthesized by ageing of ferrihydrite, which precipitated after

mixing Fe(NO

)

O and KOH solutions at a molar Fe/OH ratio of 0.05

(Schwertmann and Cornell, 1991). For a detailed description see Mikutta et al.

(2006).

2.3.2 Hydrophobic phases tested for the separation of hydrophobic colloids

2.3.2.1 Dichlormethane-Method

Dichlormethane (DCM) has a higher specific weight and a higher polarity than

hexadecane (as used by Guiné et al., 2003), allowing for a better phase separation.

3 mL of DCM were added to 12 mL of suspension of each pH-level, in which the

turbidity had been determined. The mixture was vortexed for 1 min (level 6) and left

to stand for 30 min to allow for phase separation. The aqueous phase was

subsequently transferred into the cuvette for turbidity measurement (T2; 2100P ISO

Turbidimeter, Hach). Hydrophobcity was calculated according to equation (2.1) in

section 2.3.3. Samples of each pH-level were set up in triplicate.

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

2.3.2.2 C18-Method

This separation method used a hydrophobic solid phase (C18 microparticles: 40 µm

diameter, Octadecyl Prep LC packing, Bakerbond) instead of a non-polar organic

solvent to separate hydrophobic colloids.

The method was applied to goethite particles treated with NaH

and Al(NO

)

different pH values (triplicate samples). The suspensions were filtered over a 15 µm

Nylon mesh (Nybolt) prior to turbidity measurements (T1). Thirty mg of C18

microparticles and 10 mL of suspension were mixed in a glass tube allowing for

hydrophobic colloids to sorb onto their surface. The tubes were put on a horizontal

shaker (KS501digital, IKA Labortechnik) at 100 rpm for 2 hours. At the end of the

shaking period, C18 particles were separated using the 15 µm Nylon mesh and the

turbidity of the filtered suspensions was determined (T2). Hydrophobicity was

calculated according to equation (2.1) in section 2.3.3.

2.3.3 Calculation of hydrophobicity

The calculation of hydrophobicity (H) is based on the distribution coefficient of

colloids between a polar and an apolar phase as described by Guiné et al. (2003).

The turbidity of the colloidal suspension is measured prior (T1) and after (T2) the

extraction of hydrophobic colloids. Hydrophobicity is then calculated according to

equation (2.1).

−

(2.1)

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

2.3.4 Further optimization of the C18-method

In order to determine the optimal quantity of C18 microparticles required for a certain

turbidity of the suspension we conducted a capacity test. Microparticle aliquots from

10 to 120 mg were shaken with 10 mL of filtered (15 µm) suspension of P-treated

goethite particles. The sorbed quantity of goethite particles was determined using

turbidity measurements as described in section 2.3.2.2.

In order to determine the optimal shaking time we added 10 mL of goethite

suspension (previously treated with NaH

) to 30 mg of C18 microparticles. The

mixtures were put on a horizontal shaker (100 rpm) for 1, 2, 3, and 4 hours. Particle

separation and turbidity measurements were conducted as described in section

2.3.2.2.

To test any competitive displacement between dissolved organic carbon (DOC) in

soil solutions and hydrophobic colloids from the C18 microparticles we exposed the

microparticles to a DOC solution. This solution was obtained from the aqueous

extract of forest floor samples (100 g litter, 500 mL of deionised water), which was

left to stand for 24 hours and subsequently filtered (prefiltered: folded filter

(Sartorius, 132); 0.45 µm cellulose-acetate (Sartorius, Type 11106 – 47 – N)).

Dissolved organic C concentration of the extract was measured by a total organic

carbon analyser (TOC – 5050 A, Shimadzu). Solutions of 20 and 100 mg DOC L

-1

were chosen to represent average and high DOC concentration of aqueous soil

extracts. They were prepared by appropriate dilution of the forest floor extract using

deionised water. 10 mL of each solution were added to 30 mg C18 microparticles

(triplicates) and equilibrated for 2 hours on a horizontal shaker at 100 rpm. The

solution was filtered over a 15 µm nylon mesh that had been previously rinsed with

deionised water. The retained microparticles were then rinsed with deionised water

and air-dried on the mesh. 10 mL of a suspension of P-treated goethite particles

were added to 30 mg DOC-treated (DOC-20; DOC-100) as well as to untreated C18

microparticles. The suspensions were equilibrated, filtered and their turbidity

determined as described in section 2.3.2.2.

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

2.4 Results and Discussion

2.4.1 Dichlormethane-Method

The results show a clear relation between the zeta potential and the hydrophobicity

and are in line with our assumption (figure 2.1). The hydrophobicity determined by

the two phase extractions of Al-treated goethite was higher than the Na- and Mg-

treated versions at a similar absolute surface charge. These results may be

explained by the interaction of Al with the organic solvent. Aluminium forms transition

complexes with the chlorinated solvent as described in the Diels-Alder-Reaction

(Fichte, 1986). Therefore, this method has some limitations and may only be applied

to soils with little or no Al. The second, alternative method using C18-microparticles

is supposed to overcome not only the problem of phase separation of the adapted

method of Guiné et al. (2003) but also Al interactions (DCM-method).

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

-40 -30 -20 -10 0 10 20 30 40 50

Zeta potential [mV]

Hydropho

bicity

P-treated

Al-treated

Na-treated

Mg-treated

Figure 2.1: Hydrophobicity of differently treated goethite (as determined with dichlormethane) as a

function of the zeta potential

2.4.2 C18-Method

These results are similar to those found using the DCM-method, with a clear

relationship between the zeta potential and the hydrophobicity (figure 2.2). The data

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

“gap” between the points at 1.8 mV (P-treatment) and 27.8 mV (Al-treatment) is

explained by the high coarseness of the Al-coated goethite particles in that charge

range, preventing them from being able to pass through the 15 µm Nylon mesh. The

relative maximum of H is a function of the particle surface characteristics and

amounts to 54 % for the P- and Al-coated goethite particles.

The result of the capacity test is displayed in figure 2.3. This graph indicates that 60

mg microparticles per 12 FNUs is the optimal quantity (sorption maximum).

However, if this amount was extrapolated to higher turbidity values (50 FNU), the

C18 microparticles started to coagulate and resulted in lower uptake values than

smaller quantities of microparticles. Therefore, we derived an optimal quantity of 30

mg, which may be used for suspensions of a turbidity up to 35 FNU (which is

according to our experience a relatively high amount of dispersed colloids found in

extracts of sandy soils, and much higher than the FNU of the soil solution). The

different shaking times did not show any significant variation in the sorbed amounts

of colloids (results not shown) and we chose to conduct the main experiment with a

shaking time of 2 hours.

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

-40 -30 -20 -10 0 10 20 30 40

Zeta potential [mV]

Hydrophobicity

P-treated

Al-treated

Figure 2.2: Hydrophobicity of Al- and P-treated goethite (as determined with C18 microparticles) as a

function of the zeta potential

A Method for the Determination of Hydrophobicity of suspended Soil Colloids

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 20 40 60 80 100 120 140

Quantity of C18 microparticles [mg]

Sorbed percentage of P-coated Fe colloids

Figure 2.3: Capacity test of C18 microparticles (error bars depict one standard deviation)

The comparison of DOC-treated (DOC-20) with untreated C18 microparticles

showed no significant difference in the quantity of the sorbed goethite particles (table

2.1). However, at very high DOC concentrations (DOC-100) a slight difference could

be detected. This may be attributed to sorption of organic molecules to the C18-

material. Therefore, we recommend the application of this method to soil solutions of

no more than 20 mg DOC l

-1

Table 2.1: Result of the comparison between goethite particles sorbed on DOC-treated and untreated

C18 microparticles (VC: variation coefficient

)

DOC-100 DOC-20 Untreated

Average 62.3 % 53.4 % 50.7 %

VC 5.2 % 8.8 % 5.2 %

2.5 Conclusion

The C18 method as described can be used for the determination of hydrophobic

colloids in soil suspensions. The application of our method can help to better relate

the properties of soil colloids to their function as carrier of pollutants in soils.

3 Hydrophobicity of Soil Colloids and Heavy Metal

Mobilization – Effects of Drying

Sondra Klitzke and Friederike Lang

Published in: Journal of Environmental Quality

3.1 Abstract

Drying of soil may increase the hydrophobicity of soil and affect the mobilization of

colloids after re-wetting. Results of previous research suggest colloid hydrophobicity

to be an important parameter in controlling the retention of colloids and colloid-

associated substances in soils. We tested the hypothesis that air-drying of soil

samples increases the hydrophobicity of water-dispersible colloids and whether air-

drying affects the mobilization of colloid-associated heavy metals. We carried out

batch experiments with field-moist and air-dried (25°C) soils from a former sewage

farm (sandy loam), a municipal park (loamy sand), and a shooting range site (loamy

sand with 25 % C

org

). The filtered suspensions (< 1.2 µm) were analyzed for

concentrations of dissolved and colloidal organic C and heavy metals (Cu, Cd, Pb,

Zn), average colloid size, zeta potential, and turbidity. The hydrophobicity of colloids

was determined by their partitioning between a hydrophobic solid and a hydrophilic

aqueous phase. Whilst drying increased hydrophobicity of the solid phase it did not

affect the hydrophobicity of the dispersed colloids. Drying decreased the amount of

mobilized mineral and (organo-)mineral colloids in the sewage farm soils, but

increased the mobilization of organic colloids in the C-rich shooting range soil. Dried

samples released less colloid-bound Cd and Zn than field-moist samples. Drying-

induced mobilization of dissolved organic C caused a redistribution of Cu from the

colloidal to the dissolved phase. We conclude that drying-induced colloid

mobilization is not caused by a change in the physico-chemical properties of the

colloids. Therefore, it is likely that the mobilization of colloids in the field is caused by

increasing shear forces or the disintegration of aggregates.

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

3.2 Introduction

The influence of drying on the physico-chemical properties of the soil solution and

the solid phase has been well researched. Dried soil samples showed a drastic

increase in the concentration of dissolved organic matter (DOM) after rewetting

(Bartlett and James, 1980; Baskaran et al., 1994; Courchesne et al., 1995; Münch et

al., 2002; Kjærgaard et al., 2004b), leading to a decrease in the pH of the soil

solution (Courchesne et al., 1995). The observed increase in DOC concentration

may be attributed to the disruption of microbial biomass (Christ and David, 1996)

and the breakdown of aggregate bonds (Raveh and Avnimelech, 1978). As DOC

contributes to an enhanced mobilization of dissolved heavy metal species by forming

soluble metal complexes (Brümmer et al., 1986) drying and rewetting may also be

conducive to enhancing metal leaching. In addition, drying was also found to

increase water repellency of the solid phase (Dekker et al., 2001), due to increasing

hydrophobicity.

Numerous recent studies showed that colloid-bound heavy metal transport plays a

crucial role in soils (Keller and Domergue, 1996; Jensen et al., 1999; Denaix et al.,

2001) and is found to be of greater importance than transport as dissolved ions (Egli

et al. 1999; Jensen et al., 1999). Several authors describe the influence of drying

and rewetting as important factors controlling colloid release. El-Farhan et al. (2000)

observed the highest peak of particle mass recovery after the infiltration on an

initially dry field soil. Similarly, field studies of Jann et al. (2002) revealed increasing

mobilization of colloids following dry periods. The authors attributed this

phenomenon to micro-erosion and abrasion induced by shear forces. Likewise,

Denaix et al. (2001) found markedly increased concentrations of mobilizable Pb-

containing colloids in the soil water after dry periods. However, the drying-induced

mobilization of colloids seems to be limited to the initial phase of re-wetting: In the

frame of column studies using dried soil (soil-water potential: 15500 hPa) Kjærgaard

et al. (2004b) observed an initial increase in colloid release followed by a constant

decrease as the number of pore volume increases. This initial increase may be

explained by slaking of aggregates due to the compression by trapped air during the

wetting phase (Le Bissonnais, 1996). In the same study, Kjærgaard et al. (2004b)

investigated the effect of initial soil matrix potential on water-dispersible colloids,

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

revealing that as a result of enhanced interparticle bonding or cementation of

colloids drying leads to a decrease of dispersible colloids in the soil suspension.

As drying of soil may induce changes in the solid phase such as increasing

hydrophobicity it is postulated that this may impact on colloid dispersibility. Few

studies provide first hints on colloid hydrophobicity as an important parameter

influencing colloid retention in soils. Wan and Wilson (1994a) demonstrated an

increased retention of colloidal latex particles and bacteria with increasing particle

hydrophobicity in an unsaturated sand column experiment. Similarly, under saturated

conditions hydrophobic colloids showed a lower recovery than hydrophilic colloids.

Thus, the authors concluded that hydrophilic colloids are more mobile than

hydrophobic ones (Wan and Wilson, 1994a,b) since they sorb less strongly to the

gas-water and solid-water interfaces. These findings, together with the observation

that drying can increase the hydrophobicity of the bulk soil, are in conflict with the

concept of colloid mobilization after soil drying. In addition to having an impact on

colloid mobilization, colloid hydrophobicity also plays a key role in metal

bioavailability. Carvalho et al. (1999) reported that the relative hydrophobicity of

metal-colloid complexes may affect their bioavailability by enhancing their transport

across membrane lipid bilayers.

In addition to directly affecting colloid properties, drying and rewetting of soils may

affect the dispersibility of colloids by altering the physico-chemical properties of the

soil solution, for example increasing DOC concentration of the soil solution is known

to enhance colloid stability (Kretzschmar et. al., 1999).

Moreover, air-dried soils are commonly used for various soil analyses. Whereas it is

well documented that heavy metal as well as organic C concentrations in the

extracts of air-dried and moist soils differ from each other (Wang et al., 2002; Tom-

Petersen et al., 2004), studies of the effect on the colloid-bound fractions are absent.

Thus, the results of this study may provide important information for the design of

analytical procedures.

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

As the effect of drying-induced hydrophobization on the amount and properties of

dispersible colloids has not been investigated and the underlying mechanisms of

colloid release after drying are unknown the aim of this paper was to investigate

whether

(a) drying of soil samples does not only increase the hydrophobicity of the solid

phase but also the hydrophobicity of dispersible colloids.

(b) drying of soil samples increases the colloid-bound as well as dissolved metal

fractions.

3.3 Material and Methods

3.3.1 Soil Samples

The experiments were conducted with soil samples taken from the A

horizons of the

Berlin sewage farm (Buch soil) classified (according to World Reference Base for

Soil Resources, 1998) as Regosol with sandy clay loam texture and the Berlin

municipal park Tiergarten (TG) classified as Cambisol with sandy loam. They are

both developed on quartenary fluvial sands, however, the sediment structure of the

Tiergarten park has been changed at the beginning of the last century by adding

building rubble and dredged lake sediment. Both sites were chosen since they were

contaminated by heavy metals due to the impact of sewage water (Buch soil) and

building rubble (TG soil). The experimental site on the sewage farm has been

divided into 3 subplots of similar texture (hence the labeling B1, B2, B3). The entire

site is characterized by a high small-scale heterogeneity. Additionally, we

investigated the top soil (0 – 10 cm) of a former shooting range site near Gütersloh

(GS), Lower Saxony, Germany, where oak trees have been growing for the past 30

years. The soil is a podzolic Cambisol with a texture of loamy sand. It developed on

fluvial sandy stream deposits of the Pleistocene. This soil was included in the study

due to its high contamination by Pb, which is known to have a high affinity to colloids

(Egli et al., 1999; Jensen et al., 1999; Denaix et al., 2001) and because of its high

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

content of organic matter. We sampled the A

horizon, which was interrupted by the

forest floor O

-horizon. For the analyses we mixed both horizons.

The samples were homogenized and sieved (2 mm) after the removal of roots and

stones. The soils did not contain any significant amounts of aggregates before

homogenization and sieving. The field-moist soil samples were put in plastic bags

and stored in a refrigerator. An aliquot of the samples was dried at 25 °C until

constant weight and kept in air-tight plastic containers at room temperature until use.

To characterize the soil samples the pH was determined in de-ionised water and

0.01 M CaCl

solution, using 10 g of air-dried soil and 25 ml of the respective

solution. The suspension was left to stand over night before the pH was measured

using WTW inoLab pH-meter. Concentrations of organic C, N and S (C

org

, N

org

)

were measured by a C/N/S-analyzer (Elementar, vario EL III) with soil dried at

105°C. From these parameters, the C/N-ratio was derived. For the determination of

the water content, field-moist samples were dried at 25 °C and 105 °C, until constant

weight. Table 3.1 shows the analyzed properties of the sampled soil horizons. Total

metal concentrations of the soil were determined by nitric acid-assisted digestion in

closed vessels (180°C, 6 hours). After cooling, the digested samples were filtered

(Schleicher & Schuell, ∅ 150 mm, type 0790 ½, Ref. No. 10301645), transferred into

25 mL volumetric flasks, and subsequently filled to the mark with deionised water.

The solutions were analyzed for Pb, Cu, Cd and Zn concentrations. For

measurement details refer to section “analyses”.

The mobilization of colloidal and dissolved C, Cd, Cu, Pb and Zn was studied using

triplicate field-moist and air-dried (25 °C) sample s (< 2 mm).

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

Table 3.1: Properties of the soil samples used for the experiments (n. d.: not determined)

Soil parameter

H2O

5.7

4.7

5.5

4.1

3.6

CaCl2

5.1

4.2

5.0

3.9

2.9

Conductivity [dS m

-1

]

0.135

0.184

0.268

1.473

0.195

org

[g kg

-1

]

50.0

29.0

22.0

56.0

245.0

tot

[g kg

-1

]

5.0

3.0

2.0

4.0

9.0

tot

[g kg

-1

]

1.6

0.8

0.5

3.1

n. d.

C/N ratio [-]

9.7

9.3

11.4

12.5

27.0

Pb [mg kg

-1

]

132

248

141

11 000

Zn [mg kg

-1

]

528

559

1218

n. d.

Cd [mg kg

-1

]

n. d.

Cu [mg kg

-1

]

265

n. d.

Water content (25 °C) [wt.%]

9.7

10.8

13.5

18.8

46.4

Water content (105 °C) [wt.%]

9.9

10.9

16.2

19.4

47.4

3.3.2 Experimental setup

Based on the procedure of Curtin et al. (1994), who used a soil to water ratio of 1:10

to determine clay dispersibility, we chose a soil to water ratio of 15 g field-moist soil

to 150 ml total solution volume in combination with turbidity measurements (see

below). The weight of the dried soil samples was corrected for the loss of water.

Samples were shaken for 16 hours, using an end-over-end-shaker (18 rpm; GFL

3040).

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

At the end of the experiment pH and conductivity were measured in the suspensions

prior to filtration (1.2 µm cellulose-nitrate filter, Sartorius, Type 11303 – 047N). We

determined total organic carbon concentrations (denoted as TOC in the following) as

well as total concentrations of Cd, Cu, Pb, and Zn, turbidity, hydrophobicity, particle

size, and zeta potential of the filtrate. An aliquot of the suspension was

ultracentrifuged at 300 000 g for 1 h at 10 °C (Beckman Optima TL) in order to

separate colloids larger than 14 nm and a density > 1.2 g cm³ (based on Stoke’s law)

from the solution. The supernatant was transferred into plastic vessels, acidified with

nitric acid

and analyzed for Cd, Cu, Pb, Zn and organic C (DOC) concentrations

(considered as truly dissolved). The difference between concentrations in

ultracentrifuged and not ultracentrifuged samples accounts for colloidal fractions

(operationally defined) of the above mentioned elements.

3.3.3 Analyses

Cadmium, Cu, Pb and Zn concentrations of the colloidal suspensions were

determined after microwave-assisted nitric acid digestion (CEM, MARS Xpress)

according to the procedure described in EPA method 3015 (EPA 1994). The Zn

measurement was carried out using a flame atomic absorption spectrophotometer

(Perkin Elmer 1100 B) at a wavelength of 213.7 nm. Copper, Pb and Cd analyses

were determined using a graphite furnance atomic absorption spectrophotometer

(Varian, Spectra AA 880Z; wavelength: 327.4 nm, 283.3 nm and 228.8 nm,

respectively). The organic C concentration of the suspensions and solutions was

measured by a total organic carbon analyser (TOC – 5050 A, Shimadzu). The

turbidity was determined by a turbidimeter (Hach 2100P ISO). Hydrophobicity of the

colloids was measured as a partitioning coefficient between the aqueous suspension

and a solid hydrophobic phase (C

Prep LC Packing, diameter: 40 µm, Bakerbond,

Type 7025 – 00). Ten mL of filtered suspension were added to 30 mg C

spheres

and shaken for 2 hours at 100 rpm (IKA Labortechnik, KS501 digital) in order to

allow for hydrophobic colloids to sorb onto the solids. Afterwards, the suspensions

were filtered over a 15 µm Nylon gauze to remove the C

spheres. The turbidity (T2)

of the filtrate obtained was re-measured. Hydrophobicity (H) was calculated

according to the following formula (equation 2.1):

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

−

(2.1)

with T1 being the turbidity of the suspension prior to C

treatment. Further details

about the methed are described in Klitzke and Lang (2007).

The contact angle of the bulk soil samples as a measure for hydrophobicity of the

solid phase was determined by indirect measurements. The field-moist samples

were measured by the capillary rise method as described by Adamson (1990) and

air-dried samples by the Wilhelmy plate method (Bachmann et al., 2003). The use of

two methods was considered essential due to the different wettabilities of the

samples caused by the drying process. Figure 3.1 depicts the relation of the 3

parameters contact angle, hydrophobicity as determined by C

spheres and

wettability.

Figure 3.1: Relation between contact angle, hydrophobicity and wettability

The size distribution of the colloids was analysed by dynamic light scattering (HPPS

– High Performance Particle Size, Malvern Instruments) and from the calculated

particle size distribution curve (based on the volume) an average diameter was read

0° 90° 180° Contact angle

0 % 50 % 100 % Hydrophobicity (C18)

high low Wettability

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

off. The Zeta potential was determined by a Zetasizer 2000 photon correlation

spectrometer (Malvern Instruments) on the basis of electrophoretic mobility

measurement of the colloids.

For each parameter of the complete sample set a paired t-test was conducted to

determine significant differences between results from field-moist and air-dried

samples. Sine the properties of the Gütersloh soil differ greatly from the other soils

included in the study, a second t-test was carried out based on the Buch and

Tiergarten soils only. In addition, statistically significant differences between the

triplicates of field-moist and air-dried samples of individual soils were determined by

an unpaired t-test. We used a level of significance of 95 % (P < 0.05) for both tests.

3.4 Results and Discussion

Statistically significant differences for each parameter of the complete sample set

between field-moist and air-dried samples are displayed in table 3.2. In that context it

has to be mentioned that the results of the t-test are only a helpful tool if there are no

inverse effects, i. e. if all soils show the same tendency (either increasing or

decreasing concentrations of a parameter following drying). If this condition is not

met, opposing effects cancel each other out and the result of the t-test is distorted. It

has to be mentioned that the significance of effects for individual soils can be

indirectly determined by the error bars displayed in the graphs, representing twice

the standard deviation at a confidence level of 95 %.

Table 3.2: Significance of the effects of soil drying on the given soil characteristics, (+) indicates

statistically significant differences between field-moist and air-dried samples (P < 0.05). They relate to

assessed parameters across all examined soils.

diss

- Cu

diss

- Cd

diss

- Zn

diss

- DOC +

coll

- Cu

coll

- Cd

coll

+ Zn

coll

+ COC -

Contact

angle + Turbidity

- Hydrophobicity

- Zeta

potential

- pH

conduct

-ivity -

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

3.4.1. Influence of drying on the composition of the dissolved phase

For most of the soil samples drying did not lead to a major, significant change in pH

and conductivity (table 3.2 and 3.3). In the case of the Gütersloh soil the observed

increase in conductivity following drying may be attributed to the drastic increase in

TOC of the suspensions.

Table 3.3: pH and electrical conductivity in suspensions (s:w ratio: 1:10) of field-moist and air-dried

soil samples (± 1 standard deviation)

Parameter

TG GS

suspension field-moist samples

5.6 ± 0.0

4.6 ± 0.0

5.5 ± 0.1

4.1 ± 0.0

3.7 ± 0.0

suspension air-dried samples

6.0 ± 0.1

5.0 ± 0.1

5.7 ± 0.3

4.2 ± 0.1

3.6 ± 0.1

Conductivity [dS m

-1

]

Suspension of field-moist samples

0.051 ±

0.001

0.086 ±

0.002

0.096 ±

0.003

0.760 ±

0.008

0.155 ±

0.003

Conductivity [dS m

-1

]

Suspension of air-dried samples

0.056 ±

0.001

0.086 ±

0.002

0.094 ±

0.003

0.738 ±

0.031

0.177 ±

0.002

The DOC concentrations in the air-dried samples were significantly higher than

concentrations in the field-moist samples (table 3.2), the change being especially

pronounced for the Gütersloh soil (figure 3.2). These findings are in accordance with

observations reported by Baskaran et al. (1994), Courchesne et al. (1995), and

Kaiser et al. (2001).

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

100

200

300

400

500

600

700

800

900

B1 B2 B3 TG GS

org

-concentration [mg kg

-1

]

dissolved - field-moist

dissolved - air-dried

colloidal - field-moist

colloidal - air-dried

Figure 3.2: Dissolved and colloidal C

org

concentrations in suspensions of field-moist and air-dried

samples (error bars depict one standard deviation)

In most samples, drying did not lead to a significant increase in dissolved

concentrations of Cd and Zn (figure 3.3a and 3.3b). As both metals have a very low

affinity to organic C (McBride, 1994) their mobilization is not controlled by the

increasing concentration of DOC.

Dissolved Cu concentrations increased drastically following drying (fig 3.3c) and are

statistically significant for all soils except for the Gütersloh soil. These findings are

consistent with results reported by Tom-Petersen et al. (2004). The authors attribute

this phenomenon to increased concentrations of dissolved organic C which leads to

the formation of dissolved organic Cu complexes. The exception of the Gütersloh

soil may be explained by the extremely high Pb concentrations, exceeding dissolved

Cu concentrations by more than two orders of magnitude and therefore displace Cu

from being complexed by DOC.

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

100

150

200

250

300

350

400

B1 B2 B3 TG GS

Cd concentration [µg kg

-1

]

dissolved - field-moist

dissolved - air-dried

colloidal - field-moist

colloidal - air-dried

Figure 3.3a: Dissolved and colloidal Cd concentrations in suspensions of field-moist and air-dried

samples (error bars depict one standard deviation)

B1 B2 B3 TG

Zn concentration [mg kg

-1

]

dissolved - field-moist

dissolved - air-dried

colloidal - field-moist

colloidal - air-dried

Figure 3.3b:

Dissolved and colloidal Zn concentrations in suspensions of field-moist and air-dried

samples (values of Gütersloh soil are below quantification limit; error bars depict one standard

deviation)

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

400

800

1200

1600

2000

2400

2800

3200

B1 B2 B3 TG GS

Cu conentration [µg kg

-1

]

dissolved - field-moist

dissolved - air-dried

colloidal - field-moist

colloidal - air-dried

Figure 3.3c: Dissolved and colloidal Cu concentrations in suspensions of field-moist and air-dried

samples (error bars depict one standard deviation; air-dried colloidal concentrations of the B1 sample

not determined)

In contrast to Cu, the effect of drying on the mobilization of dissolved Pb varied for

the different samples. The Buch samples showed constant, the Tiergarten and the

Gütersloh soils (according to the t-test) significantly increasing dissolved Pb

concentrations after drying (figure 3.3d). In the Gütersloh soil the increase can be

ascribed to the extremely high concentration of DOC, leading to the mobilization of

dissolved organic Pb complexes (McBride, 1994). In the Buch soils we did not

observe any increase in dissolved Pb concentrations despite increasing DOC

concentrations. The higher affinity of Cu to soil organic matter (McBride, 1994) and

the higher stability constants of Cu for humic acids (Lubal et al., 1998) and Cu-EDTA

complexes (Lindsay, 1979) could explain why the drying-induced release of organic

C does not lead to a mobilization of Pb in the Buch soils but to a preferred

complexation of Cu. In the Gütersloh soil, however, the extremely high Pb

concentrations exceed the Cu concentrations, therefore, despite the higher affinity of

Cu for organic matter it is displaced by Pb from the organic complex.

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

100

1000

10000

100000

B1 B2 B3 TG GS

log Pb-concentration [µg kg

-1

]

dissolved - field-moist

dissolved - air-dried

colloidal - field-moist

colloidal - air-dried

Figure 3.3d: Dissolved and colloidal Pb concentrations in suspensions of field-moist and air-dried

samples (error bars depict one standard deviation)

3.4.2. Influence of drying on the composition and properties of dispersible

colloids

All analyzed soils had substantial amounts of colloids in the suspension except for

the Tiergarten soil, which hardly releases any colloids at all.

Based on turbidity measurements, the statistical analysis of all samples does not

show any significant effect of drying on the mobilization of colloids (table 3.2). This is

because drying shows inverse effects for different soil samples: The Gütersloh soil

shows increasing colloid mobilization after drying, while the Buch soils show a

decreasing and the Tiergarten soil constant number of mobilized colloids following

drying. Therefore, in the statistical analysis, two opposing effects “cancel” each other

out. The decrease and constancy in colloid mobilization are in contrast to results

obtained from field studies reported by Denaix et al. (2001) who conducted their

study on soil of similar texture and pH and Jann et al. (2002; calcareous gravel),

indicating colloid mobilization after re-wetting of dry soil. However, decreasing colloid

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

concentrations are supportive of findings in sandy soils of different clay content as

reported by Kjærgaard et al. (2004a,b).

The decreasing amount of mobilized colloids may be explained by three processes:

(I) Enhanced aggregation during the drying process as described by Thill and

Spalla (2003).

(II) Drying-induced disintegration of organomineral complexes as suggested

by Peltovuori and Soinne (2005). The break up of weak bonds between

organic matter and hydroxides (Haynes and Swift, 1989) may reduce the

stability of mineral colloids resulting in a decrease in colloid mobilization.

(III) Hydrophobization of colloids as will be explained in the following

paragraph.

The difference in changes of turbidity was most pronounced for the samples B1 and

Gütersloh. Whereas soil B1 demonstrated a decrease in turbidity following drying,

the Gütersloh soil showed the opposite (figure 3.4). The high content of organic

matter of the latter creates the potential for a 3-dimensional network forming under

field-moist conditions (Schaumann et al., 2000). Such a cross-linked gel structure of

organic material could possibly impede the release of colloids into the solution. In

previously dried soil, however, any such restrictive network that existed would have

likely been destroyed and would take time to re-develop. Thus, upon rewetting,

colloids would initially be able to move freely into solution, resulting in an enhanced

release of colloids. An additional factor that may have contributed to the increased

release of colloids from the organic-rich Gütersloh soil is thought to be the disruption

of biomass (Christ and David, 1996) and possible release of organic matter

associated with microbial death and cell lysing upon drying. The higher inital water

content of the Gütersloh soil may also have been a factor.

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

B1 B2 B3 TG GS

Turbidity [FNU]

field-moist samples

air-dried samples

Figure 3.4: Turbidity in the soil suspensions of field-moist and air-dried samples (error bars depict

one standard deviation)

With the exception of the Tiergarten samples our results demonstrate a significantly

increasing contact angle of the solid phase following drying (figure 3.5, table 3.3),

which is consistent with an increasing water repellency (Goebel et al., 2004). This

result implies that drying increases hydrophobicity and agrees with observations

reported by Dekker et al. (2001) and Hurraß and Schaumann (2006). Surprisingly,

the change in hydrophobicity of the dispersible colloids does not follow the change in

hydrophobicity of the solid sample. We observed smaller changes in the

hydrophobicity of the colloids, but in general they remained hydrophilic (figure 3.6).

We postulate drying renders the colloids of the solid phase hydrophobic but these

hydrophobic colloids are no longer suspended after drying. This assertion is

supported by the work of Wan and Wilson (1994 a,b) who observed that hydrophobic

colloids are less mobile. As a result only colloids which are not affected by the

hydrophobization process, i. e. hydrophilic ones, are found in the analyzed

suspensions. Another possible explanation for the occurrence of hydrophilic colloids

is the presence of amphiphilic organic molecules as they were found to be an

important factor controlling wettability (Hurraß and Schaumann, 2006). Their non-

polar groups are thought to sorb onto the hydrophobic surface whilst the polar

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

groups would point towards the aqueous phase rendering the surface hydrophilic

and thus allowing better mobilization of colloids of hydrophobic samples.

100

110

120

130

140

B1 B2 B3 TG GS

contact angle [°]

field-moist samples

air-dried sample

hydrophobic

hydrophilic

Figure 3.5: Contact angle of field-moist and air-dried samples (error bars depict one standard

deviation)

-2%

10%

12%

B1 B2 B3 TG GS

Hydrophobicity [%]

field-moist samples

air-dried samples

Figure 3.6: Hydrophobicity of colloidal suspensions of field-moist and air-dried soil samples (error

bars depict one standard deviation)

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

Our findings indicate that drying of the soil does not lead to a major change in the

physico-chemical properties of the suspensions: The Zeta potential as well as the

particle size of the suspended colloids remained constant (data not shown). The

extremely high conductivity values of the Tiergarten soil (table 3.2) account for a

high ionic strength of the soil solution and explain the lack of colloids in the

suspension.

Concentrations of colloidal organic C remained constant or showed an increase

(figure 3.2) for most soils with the differences being most pronounced and

statistically significant (t-test) for the Tiergarten and Gütersloh soils. Our results

support the findings of Kjærgaard et al. (2004a,b) who observed an increase in

colloidal organic C following drying. We explain this phenomenon by the

disintegration of organomineral complexes (Peltovuori and Soinne, 2005) and the

disruption of microbial biomass (Christ and David, 1996) in the C-rich Gütersloh soil.

Our results reveal no increase in the colloid-bound metal fraction (data not shown)

as would be expected from suggestions in the literature (Denaix et al., 2001). We

observed a clear decline in both concentrations and fractions of colloid-bound Cd

and Zn (table 3.3). Our data suggests a preferential cementation of inorganic

particles of the soil matrix induced by the drying process leading to a lower degree of

colloid dispersibility after rewetting.

Colloid-bound Cu concentrations decreased in two soils following drying despite

increasing colloidal organic C concentrations. This implies that there is no correlation

between these two parameters although Cu has a high affinity to organic matter

(McBride, 1994). These findings suggest: (I) Colloidal Cu in these soil suspensions

may be mainly bound to inorganic particles, which may possibly have an organic

surface coating. (II) Since the suspensions of air-dried samples showed elevated

concentrations of DOC one might also hypothesize an equilibrium between colloidal

Cu and dissolved organically complexed Cu, i. e. high DOC concentrations could

lead to a desorption of Cu from the colloids. This assumption would be in line with

increasing concentrations of dissolved Cu in the suspensions of the air-dried

samples.

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

Whereas the drying process leads to a significant decrease in colloidal Pb

concentrations for the Buch and the Tiergarten soils, concentrations remain constant

in the Gütersloh sample. If the Gütersloh soil is left out in the t-test (table 3.3),

changes for the Buch and Tiergarten soils are significant. This difference could be

ascribed to the lower pH

H2O

and the high organic C content of the Gütersloh soil

allowing for the formation of Pb-organic precipitates (Lang et al., 2005). However,

the fractions of colloidal Pb for all soils (data not shown) remain unchanged by the

drying process. For the Buch soils, colloidal Pb accounts for between 70 and 96 % of

total Pb in suspension, whereas in the Gütersloh soil dissolved Pb is dominating.

Similarly to Cu, the data set does not reveal any correlation between colloidal

organic C and colloidal Pb. These findings lead to the conclusion that colloidal Pb in

the Buch soils is predominantly bound to inorganic colloids.

3.5 Conclusion

Our batch experiments showed that drying of the studied soils does not lead to a

uniform increase in the concentration of water-dispersible colloids. We conclude that

the influence of drying on the dispersibility of pedogenic colloids is likely to depend

on their composition. In a C-rich matrix mainly organic colloids are mobilized by soil

drying, whereas in a more mineral matrix (organo)-mineral colloids are immobilized,

possibly due to enhanced cementation during the drying process. Drying-induced

(im)mobilization of colloids does not always go along with the change in

concentration of colloid-associated heavy metals. This indicates that drying does not

only affect the mobilization of colloids but also the equilibrium between colloidal and

dissolved species. Absolute concentrations of the investigated colloid-bound heavy

metals were found to decrease in almost all soil samples for Cd and Zn, and in some

soils, for Cu and Pb also. This decrease may either be attributed to the

immobilization of colloids (as observed for Zn, Cd and in some samples also for Pb)

or to the mobilization of heavy-metal sorbing DOM leading to a redistribution of the

metal between colloidal and dissolved phase (as observed for Cu). Since drying did

not influence the physico-chemical properties of the colloids our results suggest that

drying-induced colloid mobilization in the field is most likely due to shear forces (as

already explained by Jann et al., 2002) or the dispersion of macroaggregates

Hydrophobicity of Soil Colloids and Heavy Metal Mobilization – Effects of Drying

(Kjærgaard et al., 2004a) rather than by a major change in physico-chemical

properties of the colloids. Future studies should assess whether the presented

results are also applicable to intact cores.

4 Increasing pH releases colloidal Lead in a highly

contaminated Forest Soil

Sondra Klitzke, Friederike Lang, and Martin Kaupenjohann

Accepted for publication in: European Journal of Soil Science

4.1 Abstract

Colloids can play an important role in the leaching of lead (Pb) in soils, and liming to

increase pH may produce conditions conducive to colloid release. We studied the

effect of pH and the role of counterion valency on the mobilization of Pb in two

topsoil horizons of a former shooting range. In batch experiments, the release of

both dissolved and colloidal Pb was studied at a pH range between 3 and 7. The pH

was adjusted with solutions of nitric acid (pH 3) and KOH and Ca(OH)

(pH 4 to 7)

and the chemical composition, size and charge of the mobilized colloids were

determined. In the presence of the monovalent K

-ion concentrations of colloidal and

dissolved Pb increased markedly with increasing pH. Colloids were stabilized not

only by electrostatic but also by steric repulsion. Organic colloids seem to dominate

at low pH of the KOH-treatment, at pH > 4 mineral particles were also dispersed.

Even though the presence of the Ca

-ion reduced the concentrations of colloidal Pb

more than did the K

-ion, our results of the Ca(OH)

treatment show that the

relevance of both colloidal and dissolved Pb increases at a pH of about 5.8. Risk

assessment on limed sites should therefore take into account both dissolved, and

colloidal Pb in judging the likelihood of Pb leaching.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

4.2 Introduction

The risk of ground water pollution by lead (Pb) is of particular interest at highly

contaminated sites such as rifle ranges, where the soil can contain as much as 80.9

g Pb kg

-1

(Knechtenhofer et al., 2003). These sites are often in forests on acid soils.

Liming the soil has frequently been proposed as a means to reduce the risk of Pb

leaching from such sites (Illera et al., 2004; Grønflaten et al., 2005). Several

investigations have analysed the effect of pH on the mobility of Pb in soils. They

assessed the success of Pb immobilization by liming (Grønflaten et al., 2005) or the

risk of Pb mobility by acidification (Illera et al., 2004). Generally, the solubility of Pb

decreases with increasing pH up to its solubility minimum at pH 6 (Herms and

Brümmer,

1984). At pH > 6, an enhanced release of soil organic matter promotes the

formation of soluble organo-Pb complexes (Sauvé

et al., 1998).

Recent studies have emphasized the significance of colloids for the translocation of

strongly sorbing pollutants in soils. They have shown that colloids are transported

much faster than conservative tracers (substances which show no interaction with

the soil matrix and therefore percolate at the same speed as soil water), due to size

exclusion and electrostatic repulsion from the soil matrix (McKay et al., 1993).

Transport facilitated by colloids has also been suggested as a major pathway for Pb

leaching from soils (Egli et al., 1999; Denaix et al., 2001). Colloids might explain the

apparent contradiction between Pb leaching measured in soil and the strong affinity

of Pb to the soil solid phase. However, in the context of environmental risk

assessment of sites contaminated by Pb and soil remediation, attention is usually

paid to only the mobilization of truly dissolved species (see, for instance, Tack et al.,

1999). So far there have been only few studies on the role played by colloids in

forest topsoil, which contain large concentrations of organic matter. Wang and

Benoit (1996) suggested that colloid-bound Pb species might be of high relevance at

the boundary between the forest floor and Ah horizon.

Increasing soil pH causes significant mobilization of colloids (Kaplan et al., 1996).

This can be explained by the increasing negative surface charge of particles with

increasing pH (Kretzschmar

et al., 1993; McBride,

1994). This causes both stronger

repulsion by the negatively charged soil matrix (Kretzschmar

et al., 1993) and

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

decrease of interparticle attraction (Kaplan et al., 1997). As the surface charge

increases, larger electrolyte concentrations are needed to collapse the electric

double layer (Kaplan et al., 1996). Consequently, with increasing pH, colloidal

suspensions remain stable even at large concentrations of background ions.

The valency of the dominating counterion in suspensions plays a key role in colloid

stability: Whereas a monovalent cation tends to disperse colloids, multivalent cations

tend to suppress dispersion. The bivalence of the Ca

-ion leads to a compression or

even collapse of the diffuse double layer, facilitating particle aggregation. Besides,

Kretzschmar and Sticher (1997) noticed that the stabilizing effect of organic matter is

reduced by increasing concentrations of Ca

, which they thought might be due to a

bridging effect of the divalent Ca

. Therefore, the addition of the divalent Ca

to the

soil by application of lime is likely to destabilize the colloids. The resulting net effect

of the stabilizing pH-effect and the destabilizing Ca-effect, however, is unknown.

Furthermore, ‘bridging’ reactions between Ca

and dissolved organic matter as

reported by Dahlgren and Marrett (1991) might generate new colloidal particles.

In summary, the literature suggests that there is little colloid-facilitated Pb transport

in acid conditions, but with increasing pH the mobilization of colloidal Pb from soils

must be expected in addition to the mobilization of organically complexed Pb. Even

though the stability of colloids at high pH is well documented, we know of no studies

on the detachment of colloids from soil with increasing pH. We suggest that those

factors that increase colloidal stability should also mobilize the colloids in soil, an

idea that has yet to be proved experimentally. Besides, as far as we know, no one

has directly investigated the effect of increasing pH on the amount and type of Pb-

containing colloids that can be mobilized. Furthermore, there are no estimates of

colloidal Pb relative to dissolved species over a range of pH. Thus, our objectives

were

(1) to investigate the effect of pH on the mobilization of dissolved and colloidal Pb

from a heavily contaminated soil,

(2) to elucidate the mechanisms of colloid mobilization by determining particle

size, zeta potential and chemical composition of colloids dispersed at various

pHs, and

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

(3) to investigate the role of the divalent Ca

-ion as opposed to the monovalent

-ion in the process of pH-induced colloid mobilization.

4.3 Materials and Methods

4.3.1 Soil sampling and soil characterization

Samples were taken from a former shooting range in North Rhine-Westphalia,

Germany, which was wooded with 30-year-old oak trees. The soil’s parent material is

a fluvial, sandy stream deposit of the Pleistocene. The soil itself is a Podzolic

Cambisol on top of which is an Oh horizon of the forest floor. For one set of

experiments we mixed Ah and Oh material (0 – 10 cm; to give soil 1) taken from two

soil profiles, which we dug behind the former target. For the comparison of a pH

increase in the presence of a divalvent (Ca

) cation we used soil samples of the

same topsoil but not exactly at the same position (soil 2). Table 1 shows some basic

properties of the soil horizons.

Table 4.1: Some properties of the analysed soils

pH (CaCl

) pH (H

O) C

org

[g kg

-1

]

C/N Pb

tot

[g kg

-1

]

tot

[g kg

-1

]

Soil 1

2.9 3.6 245 27 11.0 4.4

Soil 2

2.8 3.3 209 27 13.0 3.1

We determined the pH in deionized water and 0.01

CaCl

solution, using 10 g

field-moist soil and 25 ml of the solution (pH Meter 761 Calimatic, Knick). The

suspensions were left to stand for 1 hour before the pH was measured. Organic C

and N (C

org

and N

tot

) were measured by a C and N analyser (Carlo Erba Instruments,

type C/N NA 1500 N) after we had dried the soil at 105 °C. For the determination of

the water content, field-moist samples were dried at 105 °C until they reached

constant weight. To measure total acid-soluble Fe and Pb concentrations the soil

was digested in closed vessels (two replicates), following the procedure described

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

by Ilg et al. (2004). Ten ml of concentrated nitric acid was added to 2 g of field-moist

soil. The vessels were heated to 180°C for 6 hours. After cooling, the digested

samples were filtered (Schleicher and Schuell, diameter 150 mm, type 0790 ½,

Reference No 10301645), transferred into 25 ml volumetric flasks, and subsequently

each was filled to the mark with deionized water. The solution was analysed for Pb

and Fe as described in the section ‘analyses’.

4.3.2 Dispersion experiments

There are currently no standard methods for the determination of the stability of soil

colloids. However, batch experiments are commonly used to determine the

dispersibility and stability of colloids. They vary with respect to soil:water ratio and

shaking time (Kaplan et al., 1996; Czigany et al., 2005). Batch studies by Miller and

Baharuddin (1986) revealed that the percentage of clay that was dispersed by this

method was strongly related to soil loss by surface run-off, indicating that this kind of

experimental set-up relates to field observations.

We studied the mobilization of colloidal and dissolved Pb in batch experiments using

field-moist samples (< 2 mm). Based on the results of preliminary experiments, we

chose a soil:water ratio of 5 g field-moist soil to 50 g solution. This ratio proved to be

wide enough to prevent an artificial generation of colloids by abrasion. At narrow

soil:water ratio we observed a decrease in mean colloid size. This might be due to

abrasion of mineral particles as has been suggested, for example by Curtin et al.

(1995) for concentrated suspensions. Kinetic studies showed that equilibrium is

reached after about 8 hours of shaking (figure 4.1). Therefore we used a shaking

time of 12 hours on an end-over-end-shaker for our dispersion experiments.

Colloids are defined as particles sufficiently small to remain in suspension (Brady

and Weil, 2002) and able to scatter light (Brezesinski and Mögel, 1993). Since

particles greater than 1 nm and smaller than 1 µm show these properties, colloids

have been defined as particles of that size range (Brady and Weil, 2002). However,

these size limits are not uniformly defined in the literature. Scheffer and

Schachtschabel (1998) characterize colloids to be of a maximum size of 2 µm. For

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

reasons of practicality, we chose a filter of a pore size of 1.2 µm to separate

dispersed colloids in the suspension of our batch system from coarser soil material.

The used ultra-centrifugation technique allowed us to separate colloids of a size

range between 4 and 10 nm from the suspensions (4 nm: organo-mineral colloids

(averaged density d = 2.65 g cm³); 10 nm: organic colloids (assumed density 1.2 g

cm³)).

Time [hours]

0 5 10 15 20 25

Total mobilized Pb concentration [mg kg

-1

]

Figure 4.1: Total mobilized Pb concentration as a function of shaking time (soil 1; pH of soil

suspension: 3.9)

4.3.3 Effect of pH increase in the presence of a monovalent counterion

To investigate the effect of increase in pH we adjusted the soil suspensions to pH 3,

4, 5, 6, (±0.2) and 7 (±0.5) with three replicates each. The pH was adjusted with

0.1

KOH and 0.1

HNO

. A monovalent base as opposed to a divalent was used

to avoid the interference of Ca-specific coagulation and bridging effects with pH

effects. To obtain pH 3 and 4 the required quantities of HNO

and KOH were added

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

at the beginning of the experiment. To obtain pH 5, KOH was added gradually during

the first hour, and to obtain pH 6 and 7 KOH was added during the first 2 hours as

the buffering of added hydroxide was slow. The ionic strength of the individual

treatments was fixed at 13.5 ± 2.5 mmol (pH 3 to 6), and 15.5 ± 3.0 mmol (pH 7) with

0.2

KNO

solution.

4.3.4 Effect of pH increase – comparison of counterion valency

For the comparison of a pH increase induced by a monovalent and divalent base we

set up the following two series by adding either KOH or Ca(OH)

. We added 0.005

Ca(OH)

and 0.05

Ca(NO

)

4 H

O solutions to adjust the pH to 3.9, 5,2, 5.8, and

6.9. Since the amount of dissolved Ca(OH)

added to the samples by the 0.005

Ca(OH)

solution was not enough to obtain pH 5.2 and 5.8 in the suspensions we

also added solid Ca(OH)

. We adjusted pH 6.9 by adding 0.005 M Ca(OH)

solution

and solid Ca(OH)

. To illustrate an exemplary comparison with a monovalent base 3

series were adjusted for pH values 3.3, 4.8 and 6.3 by adding appropriate quantities

of 0.1

KOH and 0.1

KNO

solutions. The ionic strength of each individual batch

of each treatment was kept constant by adding appropriate quantities of either 0.05

Ca(NO

)

4 H

O or 0.1

KNO

solution and amounted to 46 ± 16 m

At the ends of both experiments, the pH was checked in the suspensions prior to

filtration through a 1.2-µm cellulose-nitrate-membrane (Sartorius, Type 11303 –

047N). An aliquot of the filtrate was acidified (to pH < 1) with concentrated nitric acid

for the measurement of the total Pb concentration in solution. In addition, we

determined concentrations of total organic C. We investigated the effect of a pH-

increase and for soil 1 we also measured optical density, particle size and zeta

potential (see analytical methods below). An aliquot of the filtrate was

ultracentrifuged at 300 000 g for 1 hour at 10 °C (Beckman Optima TL) to separate

the colloids. The supernatant was transferred into Eppendorff plastic caps, acidified

with nitric acid and analysed for concentrations of dissolved Pb and organic C. The

difference of the Pb and C concentration between un-ultracentrifuged and

ultracentrifuged samples accounts for colloidal fractions of Pb and C. At pH 5, 6 and

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

7, the colloidal residues were digested with concentrated nitric acid

in closed vessels

at 180° for 6 hours and subsequently analysed for Fe.

4.3.5 Analyses

Lead and Fe were measured with a flame atomic absorption spectrophotometer

(Perkin Elmer 1100 B), at a wavelength of 217.0 nm for Pb and at 248.3 nm for Fe.

Preliminary experiments showed no difference between results gained from acid-

digested colloidal suspensions as described by the EPA method 3015 (EPA, 1994)

and undigested suspensions. The concentration of organic C in the solutions was

measured by a total organic carbon analyser (TOC – 5050 A, Shimadzu).

The optical density determined as light absorption at 525 nm was taken as a

measure for the relative amount of dispersible particles as spezified by Kretzschmar

et al. (1997) on a spectrophotometer (Beckmann Instruments, DU 60 Series). At pH

3, 4 and 5 direct absorbance measurements were possible. However, at pH 6 and 7,

samples were too concentrated, and dilution (with millipore water) was required. This

dilution reduces the ionic strength and might thus have led to a disaggregation of

larger particles. However, the smaller particles created still have an influence on the

measured absorbance. We observed that colloid-free dissolved organic matter

solution also absorbs light at 525 nm. To correct for the interfering absorbance of this

material, the absorbance of ultracentrifuged, colloid-free samples was measured.

This absorbance was subtracted from the absorbance of the un-centrifuged

samples.

Particle sizes (average diameter) were determined by dynamic light scattering

(HPPS – High Performance Particle Size, Malvern Instruments) at a wavelength of

633 nm and a scattering angle of 173° after previous calibration of the instrument

with a standard of known particle size. Zeta potential was calculated based on the

electrophoretic mobility of the colloids, which was analysed by a Zetasizer 2000

photon correlation spectrometer (Malvern Instruments).

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

At pH 5 and 7 we characterized the colloidal residue by scanning electron

microscopy (SEM; Hitachi S-2700) coupled with an energy-dispersive analysis of X-

rays (EDX; SAMx, software ‘IDFix’ - acquisition parameter of the beam: accelerating

voltage: 20 kV, beam current: 500 nA). The suspension was applied to a glass

support and the suspension of pH 7 additionally coated with C.

The data presented in the graphs is given as the average value of three replicates

(soil 1) and two replicates (soil 2) ± standard error.

4.4 Results and Discussion

4.4.1 Effect of a pH increase in the presence of a monovalent base

4.4.1.1 Effects of pH on Pb mobilization

At pH 3, only dissolved Pb was detectable. The concentrations of dissolved Pb

concentration reached a minimum at pH 4, followed by an increase with increasing

pH (figure 4.2a). Because the solubility of dissolved organic matter increases with

increasing pH (Jardine et al., 1989), the concentration of this matter increased

(figure 4.2b) and was strongly correlated to dissolved Pb concentrations from pH 4 to

pH 7. Various studies have shown DOM-induced Pb mobilization

(e.g. Sauvé et al.,

1998). However, the mobility minimum observed in our study was at lower pH than in

other studies. Herms and Brümmer (1984) for example found least Pb mobility in

topsoils in the pH range 5 to 6.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

2 3 4 5 6 7 8

Zeta potential [mV]

-40

-30

-20

-10

2 3 4 5 6 7 8

Mobilized C [mg kg

-1

]

5000

10000

15000

20000

25000

30000

COC

DOC

2 3 4 5 6 7 8

Mobilized Pb [mg kg

-1

]

400

800

1200

1600

coll

diss

Figure 4.2a-c: Mobilized Pb (a), organic C (b), and zeta potential (c) of soil suspensions extracted

from soil 1 at various pHs of the KOH treatment (after 1.2 µm filtration). Pb

coll

: colloidal Pb, Pb

diss

dissolved Pb, COC: colloidal organic carbon, DOC: dissolved organic carbon.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

The strong mobilization of Pb even at pH 5 from our soil samples might be due to the

large concentration of organic matter in the soil. At pH 3 and 4 concentrations of

extractable C were in the range given for the concentration of dissolved organic C in

soil solution of acid forest soils (10-90 mg C l

-1

, Kalbitz et al., 2000), and were about

10 times larger at pH 7. The very large concentration of dissolved organic C

especially at higher pH can be explained by the use of a monovalent base (KOH) as

opposed to the divalent Ca(OH)

to increase the pH in our experiment. Such large

concentrations will not be found under ‘natural’ conditions, but may occur after

application of alkaline fertilizers or after clear cutting.

The results demonstrated a continuous increase of the colloidal Pb fraction in the pH

range from almost zero at pH 3 to 1200 mg kg

-1

at pH 7 (figure 4.2a). At pH 4, 36 %

of extractable Pb and 34 % of extractable C were bound to colloids. Our results

accord with those from Wang and Benoit (1996), who found that 50 % of Pb in soil

solutions from OA horizons of the Hubbard Brook experimental site (pH 4 to 4.7)

were bound to colloids. Thus humic rich forest topsoils may pose exceptional risk of

colloid mobilization, and much larger than in arable soils at pH above 6 (e.g. Kaplan

et al., 1997), which have been used in previous studies.

At pH 7 colloidal Pb was the dominant species (approximately 70 % of total

extractable Pb). The increase of colloid mobilization with increasing pH is a well

known phenomenon. Increasing surface charge as observed in our study from zeta

potential measurements (figure 4.2c) enhances colloid mobilization and stability (e.g.

Kaplan et al., 1997).

Colloidal Pb concentrations increased with increasing optical density of the

suspensions (figure 4.3). These results indicate that the increase in colloidal Pb was

caused mainly by stronger particle dispersion at high pH and less likely by a greater

affinity of Pb for colloids.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

2345678

Optical density [-]

0.0

0.5

1.0

1.5

2.0

Figure 4.3 Optical density of soil suspensions extracted from soil 1 at various pHs of the KOH

treatment (after 1.2 µm filtration)

4.4.1.2 Effects of pH on colloid properties

Two EDX spectra of colloids at pH 7 and 4 are shown as examples in figure 4.4a

and b. They are representative for the colloids dispersed at the individual pH. At pH

7 we detected Al, Si, O, K and Pb as the main components of dispersible colloids

(figure 4.4a). Lead might be associated with aluminosilcates, which are part of the

colloidal fraction as proposed by Kaplan et al. (1997). In addition, the strong

correlation between concentrations of colloidal Fe and concentrations of colloidal Pb

(pH 5 to 7, r = 0.99; data not shown) indicates that Fe is a further component of the

colloids. At pH 4, the comparatively small intensity of the Al signal together with the

very high C and O peaks indicate the presence of organic Pb-containing colloids

(figure 4.5). The SEM/EDX analyses confirm the presence of different types of

colloids at pH 4 and 7. However, increasing concentrations of colloidal organic C

(figure 4.2b) were significantly correlated with colloidal Pb concentrations. This

observation allows for the assumption of organic C being a component of the

colloids across the entire pH range measured.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

Figure 4.4a+b: EDX spectra of colloids dispersed at (a) pH 7 and (b) pH 4 of soil 1 (KOH treatment)

K Ti Cu

keV

500

1000

1500

2000

0 9.000

Intensity [counts]

Energy [keV]

K Ti Cu

keV

500

1000

1500

2000

0 9.000

Intensity [counts]

Energy [keV]

keV

500

1000

1500

2000

2500

0 9.000

Intensity [counts]

Energy [keV]

keV

500

1000

1500

2000

2500

0 9.000

Intensity [counts]

Energy [keV]

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

Figure 4.5: SEM image of colloids dispersed at pH 4 (soil 1, KOH treatment)

Soil minerals mobilized by changes in charge and stabilized by organic coatings

might be the form of colloids mobilized at pH 6 and 7, whereas organic

macromolecules or metal-organic precipitates, which have been suggested to occur

in acid soils (e.g. Nätscher and Schwertmann, 1991; Fotovat et al., 1997) might be

the colloids mobilized in the low pH range. At pH 6, larger concentrations of colloidal

Fe suggest the presence of colloidal Fe oxides or colloidal Fe-organic-complexes

(Jones et al., 1993). Above a certain ratio of Pb to dissolved organic matter Pb might

no longer be soluble and may be present in colloidal form as shown by Lang et al.

(2005). Remarkably, the ratio of Pb to dissolved organic C (50 µg Pb g

-1

C ± 12) is

significantly less than the Pb loading of colloidal organic C (82 µg Pb g

-1

C ± 9). This

hypothesis that metal organic salts are the colloids in acid soils accords with the

observations from Hill and Aplin (2001) who compared the colloids of rivers draining

carbonate and silicate terrains. They concluded from their results that organic

macromolecules and associated metals, which were mobilized in soils by weathering

induced by acid and dissolved organic matter, form the prevailing colloids of the low-

pH acid silicate terrains. Similarly, our findings accord with the results of Saar and

Weber (1980) who observed the formation of insoluble Pb

-fulmic acid complexes at

a very small ratio of Pb

to fulvic acid.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

From pH 3 to 6 the mean particle size remained almost constant, whereas we found

larger particles at pH 7 (figure 4.6). This is accords with the hypothesis of different

colloid composition at different pH. Kaplan et al. (1993) found that large colloids (750

nm) consist mainly of quartz and clay minerals. Humic-rich colloids sampled from

surface bogwaters were mostly smaller than 220 nm (Huynh and Jenkins, 2001).

2 3 4 5 6 7 8

Particle size [nm]

200

400

600

800

1000

1200

Figure 4.6: Size of mobilized colloids of soil 1 at various pHs of the KOH treatment (after 1.2 µm

filtration)

The particle charge increased continuously with increasing pH (figure 4.2c),

obviously being one reason for colloid mobilization. Given the proposed differences

in colloid composition with increasing pH, we think that the pH-zeta potential relation

is similar for different types of colloids.

4.4.1.3 Mechanisms of colloid mobilization

Compared with data in the literature our results show a strong mobilization potential

of colloidal Pb, even at low pH. The increasing Pb mobilization with increasing pH

and surface charge of the colloids clearly showed that electrostatic stabilization is

one of the reasons for colloid mobilization in the soil. Organic matter might act as

coating around inorganic particles, and as suggested by Kaplan et al. (1996), might

mask the surface charge of the underlying mineral and lead to more negative values.

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

However, the zeta potential measured in our study was not extraordinarily large.

Kaplan et al. (1993) found values from -32 to -44 mV for mobile colloids (we found -

14 to -34 mV). This implies that surface charge might not be the only reason for

colloid mobilization in our case.

Kaplan et al. (1997) suggested that organic coatings prevent inorganic particles from

aggregating not only by electrostatic but also by steric repulsion. Similar

observations were reported by Heil and Sposito (1993) who postulated that

electrostatic and steric properties reduce the van-der-Waals-forces between

particles and thus lead to an enhanced colloidal stability. Kretzschmar et al. (1993)

reported the steric stabilization effect of clay colloids to be affected by the

macromolecular configuration of humic substances that changes with the proton

concentration. These steric effects might be important for our soil because of the

large concentration of organic matter and may mobilize colloids even at low pHs. In

agreement, Burba and van Bergh (2002) found surprisingly stable humic-rich colloids

in acid bogwater samples (pH 3.5 – 4.5). Thus, the classical model of colloid

mobilization and stability might underestimate the risk of colloid mobilization at the

boundary between forest floor and mineral horizons and might be extended for

organic horizons by inclusion of steric stabilizing effects.

4.4.2 Effect of pH increase – comparison of counterion valency

The comparison of an increase in pH on the dispersibility of colloidal Pb showed a

much stronger mobilization in the presence of the monovalent K

than in the

presence of the divalent Ca

(figure 4.7a). Whereas in the KOH treatment, the

concentration of colloidal Pb showed a strong and uniform increase, in the presence

of Ca

the pH-induced mobilization of colloidal Pb was much less distinct. The very

strong 7-fold increase occurred only at pH > 5.8. This prounounced increase in

concentration of colloidal Pb, which goes along with a similarly pronounced increase

in colloidal organic C (data not shown), accords with the following conceptual model

explaining the role of particle charge in colloid mobilization: at low pH (i.e. little

deprotonation) the bridging effect of the Ca

masks the increasing charge effect, as

described by Schäfer et al. (2000), and therefore leads to an immobilization of

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

colloids, as observed by Dahlgren and Marrett (1991), i. e. no colloidal Pb or only

very small concentrations. With increasing pH, and the resulting greater degree of

deprotonation of functional groups, the increasing forces of electrostatic repulsion

can no longer be fully compensated by Ca bridging. This is the onset of a strong pH-

induced mobilisation of colloids. In addition to this model assumption, the dispersion

of Pb(OH)

at pH 6.9 could occur. Even though concentrations of dissolved Pb

exceed the solubility product and therefore indicate complexation of Pb by dissolved

organic C, the dispersion of precipitated Pb(OH)

cannot be excluded.

2 3 4 5 6 7 8

Ca2+: mobilized colloidal Pb [mg kg-1]

K+: mobilized colloidal Pb [mg kg-1]

200

400

600

800

1000

1200

1400

Ca2+

K+

Figure 4.7a: Concentrations of colloidal Pb of soil suspensions extracted from soil 2 at various pHs

by KOH and Ca(OH)

(after 1.2 µm filtration)

Whereas concentrations of dissolved Pb increased with increasing pH in the KOH

treatment, they decreased in the Ca(OH)2 treatment to a minimum at pH 5.8, after

which concentrations slightly increased again (figure 4.7b). This observation may be

explained by a combination of both charge and ‘bridging’ effect, i.e. the interaction of

the cation with dissolved organic C. In the KOH treatment, the mobilization of

dissolved organic C increased with increasing pH as a result of the deprotonation of

functional groups. There are no ‘bridging’ interactions between the monovalent K+-

ion and dissolved organic C, and thus allows a high degree of complexation of Pb2+

with dissolved organic C. In the Ca(OH)2 treatment, however, the pH-induced

mobilization of dissolved organic C is reduced by the divalent Ca2+, which ‘bridges’

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

between negatively charged molecules of dissolved organic C (Dahlgren and

Marrett, 1991), leading to the formation of precipitates of Ca and organic C. These

precipitates may in part be present as colloids in our filtrate. Thus, less dissolved

organic C is available to complex Pb and therefore Pb precipitates, resulting in a

minimum of dissolved Pb concentrations at pH 5.8. However, at pH exceeding 5.8,

the increasing concentration of mobilized dissolved organic C is no longer masked

by Ca2+. Therefore, dissolved organic C molecules are then available to complex Pb

and lead to an increased mobilization of dissolved Pb. Concentrations of dissolved

Pb of the Ca(OH)2 treatment exceed those of the KOH treatment between pH 3 and

4. Considering the comparatively high Ca concentration in solution, we think that this

is an additional displacement of Pb from exchange sites, resulting in concentrations

of Pb greater than those in the KOH treatment.

3 4 5 6 7 8

: mobilized dissolved Pb [mg kg

-1

]

200

400

600

800

1000

: mobilized dissolved Pb [mg kg

-1

]

100

150

200

250

Figure 4.7b: Concentrations of dissolved Pb of soil suspensions extracted from soil 2 at various pHs

by KOH and Ca(OH)

(after 1.2 µm filtration)

Increasing pH releases colloidal Lead in a highly contaminated Forest Soil

4.5 Environmental Relevance

The batch experiments determine the dispersibility and quality of colloids under the

described physico-chemical conditions and give an indication on the effect of

increasing pH. However, these results might not be translated directly to the field

scale when the effects of liming are evaluated. They do not allow for a derivation of

absolute concentrations in the soil solution under field conditions. However, our

results indicate a high risk of mobilization of colloidal Pb at the site we studied. We

conclude that in acid forest soils with organic layers large amounts of colloids may

be mobilized, especially at the boundary between organic and mineral horizons. At

low pH organic colloids seem to be of particular relevance. These colloids may in

part consist of metal-organic salts, may be stabilized by steric effects and might be

much more stable than colloids in soils containing less organic carbon. Even though

the presence of Ca

strongly reduces the concentration of colloidal Pb in

comparison with K

, our findings show that there is a risk that both colloidal and

dissolved Pb increase at a pH value above 5.8. Therefore, the danger of Pb leaching

not only in dissolved, but also in colloidal form should be taken into account in the

risk assessment of limed sites. Further studies are needed to assess whether the

colloids remain mobile in the subsoil.

5 Lead, Antimony and Arsenic in dissolved and

colloidal Fractions from an amended Shooting

Range Soil as characterized by Multi-stage

Tangential Ultrafiltration and Centrifugation

Sondra Klitzke

, Jason Kirby

, Enzo Lombi

, Rebecca Hamon

, and Friederike Lang

Berlin University of Technology, Department of Soil Science

CSIRO Land and Water, Adelaide, Australia

5.1 Abstract

Size and composition of colloids determine their relevance as carriers of heavy

metals in soils. Liming may alter these characteristics of colloids. In batch studies,

we compared the influence of increasing pH and cation valency on the mobilization

of soluble and colloid-associated Pb, As and Sb by adding Ca(OH)

and KOH to soil

samples of a contaminated shooting range site. Multi-stage tangential ultrafiltration

and centrifugation were used for the size fractionation of colloids in suspensions.

Whereas the monovalent K-ion induced the dispersion of smaller (100 kDa – 220

nm) organo(-mineral) colloids, the divalent Ca-ion suppressed the dispersion of the

latter and led to the formation of larger colloids (220 – 1200 µm), presumably

“bridging”-products between Ca, Pb and DOM. Whilst both techniques inherit

shortcomings such as problems with build-up on membranes (filtration) and

incomplete size fractionation of colloids with varying density (centrifugation) the

combination of both allows for (i) an estimate of colloids mainly composed of mineral

material and colloids consisting mainly of organic matter as well as (ii) the

differentiation between “free” colloidal organic C and organic C associated with

mineral colloids of different size classes.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

5.2 Introduction

Colloids, which are particles of a diameter < 1 µm (Ranville et al., 2005) play a

crucial role in the movement of contaminants in the environment. These colloidal

particles can consist of clay minerals, iron (Fe)-, aluminium (Al)- and manganese

(Mn)-oxides as well as –hydroxides, organic macromolecules and microorganisms

such as viruses and bacteria (Brady and Weil, 2002). Each type may vary in their

response to physico-chemical changes in the environment. The large specific

surface to volume ratio enables colloids to sorb large quantities of organic and

inorganic pollutants (Kretzschmar et al., 1999). Their transport and deposition

behaviour is not only controlled by their physico-chemical properties such as surface

chemistry and composition, but also by the size and shape of the colloids (Ranville

et al., 1999). The size fractionation of soil colloids together with their characterization

may help to identify their composition. This information may thus allow for a better

understanding and prediction of their role as carriers of different pollutants (Buffle

and Leppard, 1995).

5.2.1 Separation methods for the size fractionation of colloids

Several techniques such as centrifugation, ultrafiltration, and field-flow fractionation

(FFF) have been applied to separate or fractionate colloidal particles in soil

suspensions (Buffle and Leppard, 1995). As opposed to FFF, ultrafiltration and

centrifugation are techniques which do not cause a change in sample equilibrium

(Nifant’eva et al., 2001). This is crucial in work on colloid chemistry. While

centrifugation techniques can introduce errors in particle size fractionations through

differences in particle densities (e. g. due to the presence of organic matter; Wu et

al., 2003) common ultrafiltration techniques also inherit a few shortcomings such as

(I) adsorption of macromolecules on the membranes and (II) the aggregation of

smaller colloids to larger colloids at the membrane surface due to the formation of a

thick diffusive layer induced by the filtration process (Buffle and Leppard, 1995). In

addition a drawback associated with every filtration process is the artefact which

may be caused by non-spherical particles larger than the pore size of membranes.

The concept of multi-stage tangential ultrafiltration (MTUF), however, addresses

these two mentioned limitations: Adsorption of macromolecules can be overcome by

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

the use of specially selected filter discs. Coagulation can be minimized by filtering

the sample over filters with successively decreasing pore size, by applying a

tangential flow and a very low flow rate of the cross flow to minimize the diffusive

layer (Buffle and Leppard, 1995). Multi-stage tangential ultrafiltration has been

applied with success in freshwater systems to understand the size distribution of

aquatic humic substances (Burba et al., 1998) as well as the association of heavy

metals (Buykx et al., 2000) and organic compounds (Gadel et al., 2000) with

particulate or colloidal matter. However, no work has yet been carried out on soil

suspensions. Depending on the texture and the physico-chemical properties of the

soil matrix, soil solution may contain higher concentrations of dissolved and colloidal

species than aquatic water samples rendering the filtration process more difficult.

The MTUF system can be modified for the separation of natural colloidal material in

soil solutions to prevent changes in sample ionic strength through the re-circulation

of the < 100kDa filtrate (defined as “dissolved”) as described by Buykx et al. (2000).

Considering the strong influence of ionic strength on soil colloid stability, this

technique of a closed system loop presents a powerful method to separate colloids

by means of multi-stage tangential ultrafiltration. The aim of our work is to apply the

modified MTUF system to the separation of soil extracts and to compare the results

produced by MTUF and centrifugation.

5.2.2 The impact of liming on shooting range sites

On shooting range sites, where the metals of main concern are lead (Pb), arsenic

(As) and antimony (Sb), the colloid-facilitated transport of contaminants may be

relevant. Whilst numerous publications show that colloids are a major pathway for

Pb in soils (Jensen et al., 1999, Egli et al., 1999; Denaix et al., 2001) detailed studies

for colloid-associated As and Sb are missing. However, considering the high affinity

of As (Kabata-Pendias & Pendias, 2001) and Sb (Tighe et al., 2005, Lintschinger et

al., 1998) to Al- and Fe-oxides one may derive colloid-associated mobilization of

these two elements to be a relevant process in soils. Despite the recognition of the

sorption capacity for metals and the translocation potential of colloids, risk

assessment studies often do not differentiate between colloid-associated and

dissolved metals.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

The application of lime is a common technique suggested for the remediation of

contaminants at shooting range sites (EPA, 2001). The increase in soil pH due to the

addition of lime has been shown to increase the retention of Pb in soils (McBride,

1994). Although hardly investigated, liming might have a significant effect on the

generation and stability of colloids. In fact, the release of colloids in soils is controlled

by the physico-chemical properties of the soil solution, with pH (Kaplan et al., 1996)

and dissolved organic matter (DOM) concentration (Kretzschmar et al., 1999)

playing a key role in governing colloid stability. Liming has a number of effects on

these parameters governing the mobilization of both soluble and colloid-associated

metals:

a) pH increase

b) addition of Ca

ions into the soil

Colloid stability is enhanced at elevated pH due to increasing negative particle

charge (Kretzschmar et al., 1999). In addition, the increase in pH induces a

mobilization of dissolved organic C which may lead to the formation of organic

surface coatings on minerals. These coatings are reported to enhance colloid

stability (Kaplan et al., 1997).

Beside the possible generation of colloids such as Ca-arsenate (Sadiq, 1997) and

“bridging” products between Ca and DOM (Dahlgreen and Marett, 1991) the addition

of the Ca

ion has a destabilizing effect on colloids. The higher valency of the Ca

ion leads to a compression or even collapse of the diffusive double layer (McBride,

1994), facilitating particle aggregation. In addition, Kretzschmar and Sticher (1997)

observed the stabilizing effect of organic coatings to be reduced by increasing Ca

concentrations, which may be due to a bridging effect of the bivalent Ca.

The resulting net effect of the stabilizing pH-effect and the destabilizing Ca-effect,

however, is unknown.

5.2.3 Aim and scope of the work

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

Results of Klitzke et al. (in press) showed that an increasing pH leads to the

mobilization of Pb-bearing colloids and to a change in composition of the mobilized

colloids. These observations are followed up by the size fractionation of the colloids

in the presented study. The aim of the paper is (I) to investigate the impact of liming

on soil solution and colloid chemistry and (II) to compare the results produced by

MTUF and centrifugation by testing the following hypothesis:

1. MTUF is a powerful tool for the fractionation of soil colloids and has a better

applicability than centrifugation.

2. Increasing pH enhances the mobilization of colloidal Pb, As and Sb

3. The addition of Ca

reduces this pH-effect

4. Both a change in pH and valency of the counterion impact on the quality and

size distribution of colloids.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

5.3 Material and Methods

5.3.1 Soil collection and characterization

The study soil was collected from the A

-horizon of a former shooting range location

in Lüerdissen, North Rhine-Westphalia, Germany, classified (according to World

Reference Base) as Cambisol with a texture of loamy silt. The soil was air-dried,

homogenized, sieved < 2 mm and stored in plastic-bags at room temperature until

further analysis. The soil pH and electrical conductivity were determined in 1:2.5

(mass to volume) soil suspensions using deionised water (Millipore; WTW inoLab).

The water content was determined on field-moist samples dried to a constant mass

at 105°C. Total organic C and N (C

org

, N

org

) were determined on soil dried at 105 °C

using a C/N-analyzer (Elementar, vario MAX CN Elemental Analyzer). The

determined soil-chemical characteristics are presented in table 5.1. Total trace metal

concentrations in the < 2 mm soil sample were determined in duplicate using a

microwave-assisted acid digestion procedure (0.25 air-dried soil, 5 mL aqua regia)

and inductively coupled plasma - optical emission spectroscopy (SpectroFlame,

Spectro Analytical Instruments) and hydride generation - atomic absorption

spectroscopy for As and Sb (GBC 906AA). Oxalate-extractable Fe and Al were

determined by adding 100 mL of ammoniumoxalate/oxalic acid solution to 5 g of air-

dried soil. The mixture was shaken for 2 hours and subsequently filtered over a

folded filter. The obtained filtrates were analyzed by flame atomic absorption

spectroscopy (Perkin Elmer 1100 B) at wavelengths of 309.3 nm (Al) and 248.3 nm

(Fe). Total metal and oxalate-extractable concentrations in the soil are presented in

table 5.1.

Table 5.1: Some physico-chemical parameters, total heavy metal and oxalate-extractable Fe (Fe

)

and Al (Al

) concentrations [mg kg

-1

] of the

used soil, determined as duplicate

H2O

Conductivity [µS cm

-1

] C/N ratio [-] C

org

[%] N

tot

[%]

5.0 53.1 17.7 8.9 0.5

Pb As Sb Al Al

Fe Fe

Ca Mg K P

8 844

70 122 12 118 1 956 10 642

3236 143

1 854

1 792

2 816

694

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

5.3.2 Soil batch extractions

The < 2 mm soil sample was incubated at 60% water holding capacity for at least 1

week (max. 3 weeks) prior to extractions. Soil suspensions were prepared at a soil to

solution ratio of 1:10 (5 g incubated soil: 50 mL solution). The sample treatments

were prepared in duplicate by adjusting soil suspension pH’s to 6.5 ± 0.1 using 0.01

M KOH or 0.009 M (saturated) Ca(OH)

solutions. Differences in ionic strengths

between KOH and Ca(OH)

treatments were compensated for by the addition of

0.01 M KNO

solution as a background electrolyte to KOH suspensions. A control

treatment was used to compare pH changes from the original pH values in soil

suspensions through the addition of 0.01 M KNO

background electrolyte without pH

adjustment. Soil suspensions were shaken on an end-over-end shaker for 16 h (10

rmp) and filtered to < 1.2 µm (cellulose-nitrate; Sartorius) prior to colloid

fractionation.

5.3.3 Colloid fractionation - membrane filtration

A modified multi-stage tangential ultrafiltration system was developed based on an

instrument previously outlined by Burba et al. (1995). The MTUF system consisted of

four chambers of 25 mm radius (~ 2 - 3 ml volume) made from a Perspex (high purity

acrylic polymer) material. This system has a smaller total solution volume and a

slightly larger filtration area/volume than the original one designed by Burba et al.

(1995). This translated in a shorter time required for equilibration (see below) and

the system could cope better with solution having higher concentrations of organic

and inorganic colloids (soil extracts) as compared to freshwater samples. The

individual chambers were inserted with individual membrane filters of the following

pore sizes: 0.45 µm (mixed-cellulose-ester; Millipore), 0.22 µm (mixed-cellulose-

ester; Millipore), 0.1 µm (polycarbonate; Whatman), and 100 kDa (polyethersulfone;

Omega, Pall Gellman). The individual MTUF chambers were staked with filters of

decreasing pore size to allow for the sequential filtration of the < 1.2 µm soil

suspensions. The soil suspensions were circulated in the MTUF system using a

liquid chromatography pump (Dionex) at a flow rate of 0.1 ml min

-1

. To prevent the

build-up of colloidal material at filter membrane surfaces, a tangential flow rate of 3

ml min

-1

was applied to each chamber using Tygon tubing (inner diameter 1.3 mm)

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

and a peristaltic pump. The < 1.2 µm soil suspensions were separated into the

following colloidal fractions: 1.2 - 0.45 µm, 0.45 - 0.22 µm, 0.22 - 0.1 µm, 0.1 µm -

100 kDa, and < 100 kDa, with the filtrate passing through the 100 kDa membrane

operationally defined in this study as the “dissolved”.

The MTUF system was cleaned before analysis by soaking in 1 % HNO

for 12 h

and thoroughly rinsed with deionised water (Millipore). MTUF system tubing and O-

rings were rinsed with 1 % HNO

and deionised water. Prior to the size fractionation

of the sample the liquid chromatography pump was cleaned to remove potential

trace metal and organic C contamination by flushing with 30 % isopropyl alcohol, 0.1

% HNO

and deionised water (Millipore). In order to prevent any release of organic C

into the suspensions, the membrane filters were prepared by soaking in 0.1 M NaOH

for 1.5 h, rinsing with deionised water and stored in 0.5 M HNO

until analysis. The

individual membranes were flushed with 50 ml deionised water (Millipore) prior to

MTUF separations. MTUF system blanks using deionised water found trace metal

and organic C concentrations below instrumental detection limits (Agilent 7500c

ICPMS, 0.2-0.5 µg l

-1

and Skalar C/N Analyser; 0.5 mg l

-1

After the MTUF system was filled with solution and air bubbles removed from

chambers approximately 5 to 10 mL of the “dissolved” < 100 kDa filtrate was re-

circulated to the sample chamber (1

chamber, 0.45 µm filter) at a flow rate of 0.1 ml

min

-1

. The re-circulation of the “dissolved” fraction is essential for the complete

separation of colloidal material through all chambers (Buykx et al., 2000). This

avoids the addition of external solutions as described by (Burba et al., 1998) that can

affect the ionic strength of samples. At the end of this equilibration period (minimum

9 h as determined by breakthrough experiments, data not shown) the individual

chambers were emptied and the solutions analysed for pH, electrical conductivity,

organic C, and trace metal (i.e. As, Sb, Pb, Fe, Mn and Al) concentrations. Dissolved

concentrations were subtracted from total concentrations in the individual chambers

in order to determine the colloidal trace metal and organic carbon concentrations for

each size fraction. In order to determine possible trace metal losses to filters at the

completion of MTUF separations the individual membranes were microwave -

assisted acid digested using 5 ml of concentrated HNO

and analyzed for total As,

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

Sb, Pb, Fe, and Mn by inductively coupled plasma-mass spectrometry (Agilent

7500c). Recovery rates were calculated by summing up the individual element

content of the suspensions of the individual size fractions and – if need be – their

corresponding filter membranes. The sum was compared with the quantity initially

introduced into the system (set to 100 %).

5.3.4 Colloid fractionation - centrifugation

A centrifugation technique was used to compare colloidal size fractions in soil

suspensions determined by MTUF. The centrifugation cut-offs were selected

according to MTUF membrane filter sizes: 0.45 µm, 0.22 µm, 0.1 µm, and 100 kDa

(~ 9 nm). A subsample (15 ml) of the < 1.2 µm soil suspensions was centrifuged at

times calculated based on an averaged particle density of 2.65 g cm

-3

using the

modified Stokes’ equation described by Tanner and Jackson (1947). At completion

of centrifugation the supernatant was removed into separate plastic vessels. An

aliquot was acidified with 10 µl of 7 M HNO

. The acidified samples were analyzed

for Pb (flame atomic absorption spectrophotometry, Perkin Elmer 1100 B), Fe

(graphite furnace atomic absorption spectrophotometry, Varian, Spectra AA 880Z)

and total organic C (TOC – 5050 A, Shimadzu). An aliquot of the suspension was

microwave - assisted digested using concentrated HNO

as described in the EPA

method 3015 and subsequently analyzed for As and Sb (ICPMS, Varian).

The turbidity of soil suspensions was determined using a turbidimeter (Hach 2100P

ISO). In order to check the size of the particles in the supernatant the averaged

volume-based size distribution of the colloids was analysed by dynamic light

scattering (DLS, Malvern Instruments). The Zeta potential was calculated based on

the electrophoretic mobility of the colloids, which was analysed by a Zetasizer 2000

photon correlation spectrometer (Malvern Instruments).

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

5.3.5 Estimation of the fraction of organic- and mineral-dominated colloids –

theoretical background

Both separation methods are not able to strictly fractionate colloids according to their

size. Other colloid properties control their assignment to different fractions. These

controlling properties differ depending on the fractionation method: The separation of

colloids in the filtration process is controlled by the size of the colloids as well as by

the shape of the colloids, which may enable colloids larger than the specified pore

size to penetrate the membrane anyway (and hence resulting in the passage of

colloids which are larger than the actual pore size). With centrifugation it is not only

the size but even more so the density of the individual colloids.

On one hand, these different fractionation principles of the applied techniques may

limit the comparability of results. On the other hand the combined analysis of

different results may help to quantify different groups of colloids as outlined in the

following: Colloids can be classified into two different categories: a) colloids whose

composition is dominated by organic material and b) colloids whose composition is

dominated by mineral material such as Fe-, Al-, Mn-oxides and clay minerals.

Provided that the filter permeability is equal for both colloid groups, MTUF is able to

separate colloids into size classes, irrespective of the composition of the colloids. In

contrast, centrifugation is only able to completely separate colloids according to their

size, which are of similar density (equal or greater than the averaged density, which

is 2.65 g cm³ in our case). In our centrifugation experiment, colloids dominated by

mineral material will be fractionated correctly, while colloids dominated by organic

material may be assigned to a smaller size fraction or even to the dissolved fraction.

The degree of fractionation of organo-mineral colloids by centrifugation and hence

the assignment to one of the groups depends greatly on the averaged density of the

entire colloid, i. e. an organo-mineral colloid which is dominated by organic matter

has a lower density than a colloid which is mainly composed of minerals. Based on

these facts the combination of the results obtained from both methods can provide

an estimate for the fraction of an element associated to i) colloids dominated by

mineral material (centrifugation) and to ii) colloids dominated by organic material

according to equations (5.1) and (5.2):

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

Percentage of element associated to mineral-dominated colloids = % of colloid-

bound element as determined by centrifugation (5.1)

Percentage of element associated to organo-dominated colloids = (% of colloid-

bound element as determined by MTUF) – (% of colloid-bound element as

determined by centrifugation) (5.2)

Equation 5.1 and 5.2 can be applied to each individual size fraction. The estimate

inherits some degree of uncertainty due to the unknown composition of organo-

mineral colloids to which the individual element is associated with and may therefore

only provide a ballpark figure. Positive results of equation (5.2) indicate the

percentage of the individual element associated with organo-dominated colloids,

negative results hint for the occurrence of the “filter effect”, i. e. centrifugation result

exceed MTUF results. This is the case when larger colloids pass through the filter

membrane. Since the percentage caused by this artefact proved very low it may be

neglected in the calculation.

The overall percentage of an element bound to organo-dominated colloids across all

size fractions can be calculated using equation (5.3):

Total

(Organo-dominated colloids)

= (100 % - MTUF(<100 kDa)) – (100 % - Centrifugation(<

100 kDa)).

5.3.6 Further characterization of the colloids

In addition, energy dispersive X-ray (EDX; SAMx, software “IDFix”) analysis of the

colloids was used to gain an understanding of the composition of the colloids in the <

1.2 µm filtrate. For the SEM/EDX analyses, the suspensions were applied to a glass

carrier and were not coated prior to analysis.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

5.4 Results and Discussion

5.4.1 Multi-stage tangential ultrafiltration

5.4.1.1 Instrumental

The averaged recovery rates for the individual elements (table 5.2) following

separation showed some loss for the metals. Iron recoveries were lower than that

found for other elements and variable between samples. The reproducibility of the

element distribution across the different size fractions amounted to 85 – 100%. The

analysis of digested filters proved that 10 to 18% of the overall Pb content (KOH

treatment) and 25 to 30% of the overall Pb content (Ca(OH)

treatment) were

attached to the membranes. This residual Pb was found mainly to be associated with

the larger membranes (0.22 and 0.45 µm). The organic C recovery rate was slightly

larger than 100%.

Table 5.2: MTUF recovery rates for the individual elements, averaged over the entire experiment

(VC: variation coefficient)

Fe Mn As Sb Pb C

org

Average

recovery rate 68 % 87 % 85 % 94 % 85 % 118 %

VC 29 % 8 % 8 % 5 % 8 % 7 %

5.4.1.2 Colloid size and element distribution as determined by MTUF

Results of the control sample (pH of suspension: 4.7) clearly showed that hardly any

metals were associated with colloids (table 5.3 and figure 5.1). The dissolved phase

(< 100 kDa) dominated for all elements. Approx. 20% of organic C was present in

colloidal form, being distributed in similar amounts over the different size fractions. A

small percentage of Fe (9 %) was found in the colloidal fractions and may be due to

its association with organic matter, possibly as Fe-DOM precipitates, which occur

according to Jansen et al. (2005) at a pH of 4.5.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Mn Fe As Sb Pb OC

Elemental Size Distribution (%)

< 100 kDa

0.1 µm - 100 kDa

0.22 - 0.1 µm

0.45 - 0.22 µm

1.2 - 0.45 µm

Figure 5.1: Elemental size distribution of the control sample (pH 4.7) following colloid separation by

MTUF (OC: organic carbon)

Table 5.3: Dissolved and colloidal concentrations of the individual elements in the different treatments

as obtained by MTUF (diss.: dissolved, coll: colloidal)

An increase in the soil suspension pH to 6.5 using KOH resulted in a complete

change in both concentrations and elemental distribution (table 5.3 and figure 5.2).

The concentration of most of the analytes increased in comparison to the control

Mn [µg l

-1

]

Diss. Coll.

Fe [µg l

-1

]

Diss. Coll.

Sb [µg l

-1

]

Diss. Coll.

Pb [µg l

-1

]

Diss. Coll.

As [µg l

-1

]

Diss. Coll.

org

[µg l

-1

]

Diss. Coll.

Control

714

126

130

2675

KOH a

1777

4234

267

6844

7550

453

KOH b

2004

3422

184

6365

10690

187

Ca(OH)

246

142

924

642

Ca(OH)

299

186

1312

670

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

sample. The large observed increase in dissolved Pb concentrations may be due to

the strong increase in DOC concentration leading to a complexation of Pb

(Turpeinen et al., 2000). Arsenic was found to be re-distributed between the

dissolved and colloidal phase: in comparison to the control treatment (figure 5.1)

dissolved concentrations decreased and colloidal concentrations increased, the

latter probably due to the dispersion of sesquioxides. Interestingly, dissolved

concentrations of As did not increase with increasing pH as commonly reported in

the literature (for instance Tyler and Olsson, 2001). This may be due to the

increased dispersion of sesquioxides, providing more sorption sites for As. The

colloidal size fractionation pattern found for Pb and As is similar to that observed for

the sesquioxide-forming elements (Fe and Al), which are predominantly found in the

smaller colloidal size fractions (100 kDa – 220 nm). The identification of Al, Fe, Pb

and oxygen in EDX-spectra suggests Pb to be associated with colloidal sesquioxides

and clay minerals (figure 5.3). These observations are consistent with previous

findings by Kabata-Pendias and Pendias (2001) who report the main sorbents of As

to be Fe- and Al-oxides.

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Mn Fe Al As Sb Pb OC

Elemental Size Distribution (%)

< 100 kDa

0.1 µm - 100 kDa

0.22 - 0.1 µm

0.45 - 0.22 µm

1.2 - 0.45 µm

Figure 5.2: Elemental size distribution of the KOH treatment (pH 6.5 ± 0.1)) following colloid

separation by MTUF (OC: organic carbon)

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

Figure 5.3: EDX spectra of the KOH treatment

Inspite of high concentrations, DOM did not displace As from the mineral colloid

surfaces. This displacement might be expected according to results published by

Redman et al. (2002) who found indications for the sorption competition between

natural organic matter (NOM) and As.

Similar to the control treatment, organic C was mainly present in the dissolved

fraction of the KOH treatment. However, the concentration of colloidal C is still larger

in the KOH treatment than in the control treatment (table 5.3). We speculate, from

results reported in the literature (e.g. Tipping and Higgins, 1982; Kretzschmar et. al,

1997) that in part organic C is associated with inorganic colloids in the form of

coatings enhancing colloid stability.

Although the state-of-the-art describes Sb to have similar adsorption-desorption

properties as As (for instance Lintschinger et al., 1998) we did not find any parallels

between these two elements in our experiment. With increasing dispersion of

sesquioxides we only detected a small percentage of Sb in the colloidal fractions.

keV

500

1000

1500

2000

2500

3000

3500

4000

0 10.00

Na Cl

Energy [keV]

Intensity [counts]

keV

500

1000

1500

2000

2500

3000

3500

4000

0 10.00

Na Cl

Energy [keV]

Intensity [counts]

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

Contrarily to the findings of Johnson et al. (2005) who reported Sb to be completely

sorbed at pH < 7, we still observed significant dissolved Sb concentrations following

a pH increase to 6.5. In contrast to results published by Flynn et al. (2003), who

observed only a minor influence of pH on the mobilisation of total Sb in soil water,

concentrations of both dissolved and colloidal Sb were found to increase following a

pH increase using KOH. A possible explanation for Sb to be present in the dissolved

fraction is its association with DOM as previously described by Lintschinger et al.,

(1998). This assumption is supported by results of Mergenthaler and Richner (2002),

who found a substantial fraction of Sb (17 to 44 %) bound to organic matter.

When the pH of the soil suspension was increased to 6.5 using Ca(OH)

, the

distribution between dissolved and colloidal fractions changed strongly in

comparison to both the KOH treatment and the control (figure 5.4). Ca(OH)

addition

resulted in significantly lower dissolved concentrations of the individual elements

than KOH addition (table 5.3). Arsenic, Fe, Pb, and DOM may have been retained by

the soil matrix and/or were associated with particles larger than the filter cut-off (1.2

µm). This assumption may be attributed to the formation of “bridging” products

between Ca and dissolved (Dahlgreen and Marett, 1991) or particulate organic

matter, which are also thought to contain As and Fe (for example Fe-DOC-Ca-As).

Similarly, from results reported by Celi et al. (2001), who found ortho-phosphate

sorption on goethite to be increased in the presence of Ca

, we conclude enhanced

As sorption due to bridging reactions on Fe-oxides. These oxides were retained by

the soil matrix and contribute thus to a decrease in dissolved As concentration in the

soil suspension. Besides, the formation of Ca-arsenate cannot be excluded. Our

results are in contrast to those obtained by Jones et al. (1997) who reported

increasing As concentrations following liming. The authors concluded As

concentrations to correlate positively with pH, however, not with precipitation-

dissolution reactions. Iron may have precipitated as metal hydroxide (Jansen et al.,

2002). Due to the low concentrations of protons and DOM, Pb is not displaced from

exchange sites. As a consequence of the described reactions, concentrations of

dissolved As, Fe, Pb and DOM in the suspensions decreased.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Mn Fe As Sb Pb OC

Elemtental Size Distribution (%)

< 100 kDa

0.1 µm - 100 kDa

0.22 - 0.1 µm

0.45 - 0.22 µm

1.2 - 0.45 µm

Figure 5.4: Elemental size distribution of the Ca(OH)

treatment (pH 6.5 ± 0.1)) following colloid

separation by MTUF (OC: organic carbon)

The EDX-spectra of the Ca(OH)

treatment (figure 5.5) identified C, O, and Ca as the

main elements of the dispersed colloids and supports our assumption of the

formation of “bridging” products between Ca and DOM. The absence of Pb from the

colloidal fraction may be explained by two facts: (a) the competition between Ca and

Pb for binding sites may lead to a displacement of Pb from the colloid (Harter, 1979)

(b) the concentrations of Pb were below quantification limit. Therefore, the

identification of Pb-bearing colloids remains an assumption which could finally not be

confirmed by EDX measurements. As and Sb were below quantification limit in this

treatment.

Elemental concentrations of the colloidal fractions may have decreased in

comparison to the KOH treatment due to the suppression of colloid mobilization by

the presence of the divalent Ca

cation as outlined by Kretzschmar et al. (1999). We

observed a shift in the distribution pattern of colloid-associated Pb from smaller

fractions (100 kDa – 220 nm as in the KOH treatment) to larger fractions (220 – 450

nm). Since Pb concentrations in solutions are below the solubility product of Pb(OH)

the presence of larger colloidal size fractions may be due to the formation of

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

“bridging” products between Ca-DOC-Pb as pointed out by Dahlgreen and Marett

(1991).

Figure 5.5: EDX spectra of the Ca(OH)

treatment

5.4.2 Centrifugation

5.4.2.1 Colloid size and element distribution as determined by centrifugation

In general, the distribution of colloid-associated elements among the different size

fractions as determined by centrifugation shows similar trends as the filtration-based

distribution. However, the size of colloids determined by laser scattering strongly

differs from the calculated size. Reasons for this apparent discrepancy and for

differences between centrifugation and filtration results are discussed in the section

below.

5.4.2.2 Turbidity and zeta potential

The turbidity of the individual supernatants continuously decreased with decreasing

cut-off (results not shown) and is in agreement with the assumption of colloidal

keV

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

0 10.00

Energy [keV]

Intensity [counts]

keV

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

0 10.00

Energy [keV]

Intensity [counts]

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

particles being removed during the filtration process. In parallel, the zeta potential of

colloids was found to become more positive with decreasing particle size (table 5.4).

This suggests colloidal particles with negative charges were being removed from

suspensions with decreasing particle size cut-off. The zeta potential of colloids in the

Ca(OH)

treatment was more positive than that of the KOH treatment. This

observation may be explained by the formation of complexes between Ca and

anionic groups of NOM (natural organic matter) “neutralising” some of the negative

charge resulting in enhanced particle aggregation (Tiller and O’Melia, 1993).

Table 5.4: Zeta potential, Fe and DOC concentration of the supernatants of the different treatments

(centrifugation)

Control KOH Ca(OH)

Zeta

potential

[mV]

Zeta

potential

[mV]

[µg l

-1

]

DOC

[mg l

-1

]

Zeta

potential

[mV]

[µg l

-1

]

DOC

[mg l

-1

]

Gesamt (<1.2µm) -13.8 -21.4 3700 131.6 -14.3 788 85.3

< 0.45 µm -14.0 -22.9 3650 126.9 -13.9 751 82.6

< 0.22 µm -13.6 -22.1 3550 127.0 -14.2 723 82.5

< 0.1 µm -11.4 -23.4 3450 128.2 -14.4 665 82.3

< 100 kDa -11.1 -14.8 2450 123.9 -8.5 557 82.1

5.4.3 Colloid characterization by comparison and combination of results

obtained by centrifugation and MTUF

5.4.3.1 Estimation of the fraction of organic- and mineral-dominated colloids

In most cases, centrifugation showed a similar colloidal distribution pattern as MTUF.

In general our results indicate that the distortion of the results is more affected by

varying densities of the colloids rather than by the shape of colloids allowing for the

penetration of particles larger than the actual pore size. The density effect gets most

obvious for organic colloids. Calculating the centrifugation time required for colloids

of a certain size to settle requires the assumption of an average density of 2.65 g

-3

. However, the density of organic colloids may be much below that average (1.1

g cm

-3

(Citeau et al., 2001),

< 1.6 g cm

-3

(Bethwell 2004)). Thus, we might have

underestimated the fraction of COC by using centrifugation as a separation

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

technique as can be illustrated in the control treatment (figure 5.6) where

centrifugation revealed a much lower percentage of COC than MTUF. This

assumption is in agreement with findings by Wu et al. (2003) who found particles

larger than the cut-off, which was calculated based on Stoke’s law assuming an

effective colloid density of 2.65 g cm

-3

. They attribute this observation to the

presence of organo-clay complexes with a density < 2.65 g cm

-3

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Fe Fe* As As* Sb Sb* Pb Pb* OC OC*

Elemental Size Distribution (%)

< 100 kDa

0.1 µm - 100 kDa

0.22 - 0.1 µm

0.45 - 0.22 µm

1.2 - 0.45 µm

Figure 5.6: Comparison of the elemental size distribution of the control treatment between results

gained from MTUF and centrifugation (*; OC: organic carbon)

Results of the KOH and the Ca(OH)

treatment obtained by equation (5.2) and (5.3),

which differ significantly from zero (t-test with p < 0.1) are displayed in table 5.5.

Deviations between the sum of each individual size fraction and the total amount of

organo-domintated colloids (as obtained by equation 5.3) are ascribed to the fact

that organo-dominated colloids still occur in size fractions for which the t-test does

not show any significant difference between the data obtained by centrifugation and

MTUF.

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

In the KOH treatment, Pb and Fe were only found in the smallest colloidal fraction

(100 kDa – 100 nm) with sesquioxides most likely being coated by organic C (figure

5.7). This assumption is confirmed by the sharp change in zeta potential between

Table 5.5: Percentage of the individual element associated with mineral-dominated and organo-

dominated colloids in the respective size fractions of MTUF and centrifugation as obtained by

duplicate samples of the KOH- and Ca(OH)

treatment. The difference between MTUF and

centrifugation in the individual size fractions (equation 5.2) and the total sum of colloids of all size

fractions (equation 5.3) are listed for data points with p < 0.1.

Element Size range Percentage of

the individual

element

associated with

colloids in the

respective size

fraction [%]

(MTUF)

Percentage of the

individual element

associated with

mineral-dominated

colloids [%]

(Centrifugation )

Percentage

of the

individual

element

associated

with organo-

dominated

colloids [%]

KOH

Pb 0.22 – 0.1 µm 34 0 34

Pb 0.1 µm – 100 kDa 17 9 8

Pb Total (equ. 5.3) 48

Fe 0.22 – 0.1 µm 41 3 38

Fe Total (equ. 5.3) 33

As 0.22 – 0.1 µm 36 5 31

As Total (equ. 5.3) 33

Ca(OH)

Pb 0.45 - 0.22 µm 14 2 12

Pb 0.22 – 0.1 µm 2 4 -2

Pb Total (equ. 5.3) 25

Fe 1.2 – 0.45 µm 1 5 -4

the 100 kDa and 0.1 µm supernatants, suggesting the separation of particles of

negative charge. In the fraction of 100 – 220 nm centrifugation results differ from

those obtained by MTUF: no colloidal Pb, Fe and organic C could be detected. We

presume individual ions of Pb and Fe in the 220 – 100 nm fraction to be associated

with organic matter (OM), which is not separated due to its lower density and

therefore simulates the absence of colloid-associated Pb and Fe. This assumption is

supported by the estimate provided by equation 3 (table 5.5). Similarly to Fe and Pb,

As was found in the smaller colloidal size fraction (100 kDa – 100 nm), where it may

be associated with sesquioxides (Kabata-Pendia, 2001). In the larger size fraction

(100 – 220 nm) As was only detected to a small degree. This observation

substantiates the assumption of As being associated with OM, either by “bridging” of

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

Fe between As and OM or directly with OM as described by Saada et al. (2003), who

found As to be associated with amine groups of DOM. Since OM of this size fraction

cannot be separated by means of centrifugation it is thought not to occur as coatings

of sesquioxides, but as a “free” colloid. The assumption of As being associated with

organic colloids is supported by results obtained by equation 5.3 (table 5.5).

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Fe Fe* As As* Sb Sb* Pb Pb* OC OC*

Elemental Size Distribution (%)

< 100 kDa

0.1 µm - 100 kDa

0.22 - 0.1 µm

0.45 - 0.22 µm

1.2 - 0.45 µm

Figure 5.7: Comparison of the elemental size distribution of the KOH treatment between results

gained from MTUF and centrifugation (*; OC: organic carbon)

In the Ca(OH)

treatment the centrifugation results for Fe and As differ from those

found using the MTUF system (figure 5.8). The lower percentage of these elements

in the colloidal fractions of the MTUF results, which was also observed for Fe in the

control treatment, may be attributed to an artefact associated with the filtration

process. For Fe in the Ca(OH)

treatment, this assumption could be confirmed by the

result obtained by equation 5.3 (table 5.5). The pore size of the filter membranes is

based on the assumption of spherical particles. However, colloidal material is often

non-spherical. Therefore, longish particles of a rod- or needle-like shape such as

these of Fe oxides would pass the membrane of smaller pore sizes anyway and thus

decrease the concentration in the size fraction to which they would actually belong to

and increase the concentration in the fraction below. In the Ca(OH)

treatment, the

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

colloid distribution of As follows the pattern found for Fe. In contrast to results

obtained with MTUF, colloid-associated Pb was only detected to a minor degree.

This observation supports the previously mentioned assumption of Pb being

associated with DOM in the form of Ca-DOC-Pb-complexes. Due to its low density

this compound is thought not be separable by means of centrifugation. The presence

of an organo-Pb-colloid is substantiated by the result of equation 5.3 (table 5.5).

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Fe Fe* As As* Sb Sb* Pb Pb* OC OC*

Elemtental Size Distribution (%)

< 100 kDa

0.1 µm - 100 kDa

0.22 - 0.1 µm

0.45 - 0.22 µm

1.2 - 0.45 µm

Figure 5.8: Comparison of the elemental size distribution of the Ca(OH)

treatment between results

gained from MTUF and centrifugation (*; OC: organic carbon)

5.4.3.2 Particle size measurements by dynamic light scattering (DLS)

Particle size measurements were used in order to confirm the cut-offs calculated for

the centrifugation experiment and to check the presence of any particles larger than

the cut-off, which may occur due to a density lower than the one assumed in the

calculation. The measurement technique was found to be subject to strong

fluctuations especially as cut-offs were getting smaller and thus fewer particles were

dispersed in suspension. With increasing centrifugation time (i. e. decreasing cut-

off), the particle radius decreases (results not shown) in both treatments (Ca(OH)

and KOH). However, with decreasing cut-off, larger particles become increasingly

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

relevant again. Dynamic light scattering measurements did not show any larger

particles in the Ca(OH)

-treatment as indicated by MTUF and as suggested in the

literature (Celi et al., 2001).

Iron concentrations of all treatments decreased with decreasing particle size,

whereas organic C concentrations hardly decreased (table 5.4). This observation

was consistent with the assumption of inorganic particles (for instance oxides)

getting separated during centrifugation (i. e. the number of large inorganic particles

decreased). Meanwhile, the number of larger particles of smaller density (such as

organics and clays) could remain constant and thus contribute to an increased

averaged particle radius. Similar observations are reported by Wu et al. (2003) who

attribute this phenomenon to the average hydrodynamic diameter of polydisperse

samples to be biased towards the larger particles. Besides, they attribute this

discrepancy to the non-spherical shape of particles.

The increasing presence of particles of larger size with calculated decreasing cut-off

may also be explained by the low density of organic matter (table 5.6): the lower the

density the larger the particle which remains in solution. In addition, it has to be

mentioned that with decreasing particle size and with decreasing optical density of

the sample the impact of contaminating dust particles may be increasingly relevant.

Table 5.6: Cut-off for different densities based on the same centrifugation parameters

Averaged density:

2.65 g cm

Density of organics:

1.5 g cm

Density of organics

1.2 g cm

Cut-off [nm] Cut-off [nm] Cut-off [nm]

450 818 1293

220 400 632

100 182 287

9 16 26

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

5.5 Conclusion

The combination of results gained from both MTUF and centrifugation allows for an

estimate of the percentage of mineral-dominated and organo-dominated colloids.

Moreover, the combination of both results allows for the conclusion whether colloidal

organic C is associated with oxides or available as “free” colloidal organic C.

However, each method inherits its own shortcomings and restrictions. In the filtration

process, the spherical pore size of the filtration membranes allows larger particles of

different shape (for instance rod- or needle-like) to pass the filter anyway and may

thus distort the concentration pattern. On the other hand, centrifugation eliminates

problems caused by membranes (such as build-up and adsorption). However, the

assumption of an averaged particle density inevitably brings about an unclear cut-off.

The major shortcoming of the centrifugation technique in studies dealing with

colloids is the inability to completely separate colloidal organic C and associated

elements from the suspension. Since colloidal organic C is thought to play a major

role in environmental chemistry of colloids and associated contaminants we suggest

the use of MTUF over centrifugation for colloid fractionation studies if a combined

approach of both methods is not feasable.

Our study showed that the valency of the counterion controls quantity, elemental

composition and size of the colloids. Increasing the pH using KOH led to the

dispersion of sesquioxides and organic colloids. Colloidal Pb was found to be

associated with sesquioxides which may be stabilized by organic C. This treatment

also induced the mobilization of smaller colloids (100 kDa – 220 nm).

Increasing the pH using Ca(OH)

suppressed the dispersion of sesquioxides and

induced the formation of larger colloids such as precipitation products of Pb and

“bridging”-products of Ca and DOM, which may also include some Pb (for instance

DOC-Ca-DOC-Pb). However, the precise composition of Pb-containing colloids in

the Ca(OH)

treatment could finally not be confirmed by EDX measurements.

An understanding of colloid size and composition is essential in understanding the

fate and behaviour of associated contaminants in terrestrial environments in order to

Dissolved and colloidal Pb, Sb and As as characterized by MTUF and Centrifugation

determine the risk to human and environmental health from the mobilization,

stabilization and transport of colloidal fractions in potential stabilization procedures.

Further research needs to be undertaken to characterize colloidal fractions in soils

and determine the potential lability of associated elements in environments.

6 Mobilization of soluble and dispersible Pb, As,

and Sb in a polluted, organic-rich Soil – Effects

of pH Increase and Counterion Valency

Sondra Klitzke and Friederike Lang

Submitted for publication to the Journal of Environmental Quality

6.1 Abstract

Liming is a common technique suggested for the stabilization of shooting range

sites. We investigated the effect of a pH increase on the mobilization of soluble and

dispersible (colloidal) Pb, As, and Sb. Our hypothesis was that the addition of

divalent cations counteracts the pH-induced mobilisation of soluble and colloidal

metal(loid)s. We determined soluble and dispersible As, Pb, Sb, Fe and C

org

concentrations in the filtered (1.2 µm) suspensions of batch extracts of topsoil

samples (C

org

: 8 %) from a former shooting range site following a pH increase to

values between 3.5 and 7 by adding a monovalent (KOH) or a divalent (Ca(OH)

)

base. In the Ca(OH)

-treated samples, soluble metal(loid) concentrations were 62 %

to 98 % lower than in those treated with KOH. Similarly, Ca reduced dispersible Pb

by 95 %, but had little or no impact on dispersible As and Sb. Our results show that

Ca attenuates the mobilisation of both soluble and dispersible Pb and As in the soil.

We conclude that the counterion valency controls the mobility of metal(loid)s by

affecting the mobility and sorption capacity of the sorbents (e.g. colloids, organic

matter).

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

6.2 Introduction

The contamination of shooting ranges by Pb, As and Sb may pose a significant

environmental concern (USEPA, 2005). Amongst these three metals, Pb is of

principle concern due to the high content in bullets (95 %; Neite et al., 1999) which

often leads to severe contamination of the soil. The management of shooting ranges

suggests liming for the remediation of such sites in order to reduce Pb migration

(USEPA, 2005). Lime application impacts on the soil chemistry in two ways: (a) it

leads to an increase of the soil pH and (b) and increases Ca concentration in the soil

solution. With increasing pH, Pb concentrations in the soil solution decrease

(Lindsay, 1979) due to enhanced retention by soil minerals (McBride, 1994).

However, at higher pH, an enhanced release of soil organic matter promotes the

formation of soluble organo-Pb complexes (Sauvé et al., 1998). In addition, the

displacement of Pb from exchange sites by the addition of Ca ions (Harter, 1979)

may contribute to increased dissolved Pb concentrations after lime application. While

the impact of liming on dissolved Pb is well documented (for instance Turpeinen et

al., 2000), detailed studies on the response of dissolved As and Sb to lime

application are scarce if not absent. Whereas Pb in soils predominantly occurs as

heavy metal cation, the metalloids As and Sb exist as negatively charged oxyanions.

This explains relevant differences in sorption and retention of the metals.

Arsenic retention in soil is controlled by two types of sorbents, namely sesquioxides

and organic matter. Sesquioxides are widely recognized as the main sorbents for As

(for instance Kabata-Pendias and Pendias, 2001). Dissolved As concentrations in

soil are reported to increase with increasing pH due to desorption from sesquioxides

(Hsia et al., 1994; Smith et al., 1998). Recent studies also attribute an important role

to soil organic matter in As retention (Warwick et al., 2005; Buschmann et al., 2006)

with humic acids being found to play a major role in As retention in soil when the Fe-

oxide content is low (Warwick et al., 2005). Suggested binding mechanisms for As

with soil organic matter are the formation of inner- and outer-sphere complexes with

phenolic and carboxylic groups (Mukhopadhyay and Sanyal, 2004) and H-bridges

(Buschmann et al., 2006). Mobilization of dissolved organic C (DOC) induced by pH

increase might therefore also mobilize associated As.

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

In addition to (de)sorption reactions, precipitation is a further mechanism which

controls As solubility. Adding Ca

cations into the soil could also facilitate the

formation of Ca-As-precipitates (Sadiq, 1997), resulting in a decrease in dissolved

As. Similarly, Swash and Monhemius (1995) reported that the addition of lime

reduced dissolved As concentrations.

Fe (hydr)oxides (Spuller et al., 2007), Al oxides (Tighe et al., 2005, Lintschinger et

al., 1998, Brannon et al., 1985) and humic acid (Tighe et al., 2005) are reported to

be - as with As - important sorbents for Sb in soil. Sb was found to be strongly

sorbed to amorphous Fe-oxide at pH 3 to 6.5 (Tighe et al., 2005) and to hematite at

pH < 7 (Edwards et al., 1999). In contrast to these mineral sorbents, the sorption of

Sb to humic acids strongly decreases with increasing pH (Tighe et al., 2005). The

authors attribute an important role to humic acid retention of Sb in soils with high

organic matter content. Based on these findings, one may presume an increase in

dissolved Sb concentrations following a pH increase if organic matter is the main

binding partner of Sb in the soil. Furthermore, the addition of Ca may lead to the

formation of Ca[Sb(OH)

]

, which is suggested to control Sb(V) mobility at high

concentrations (Johnson et al., 2005).

In order to assess the effect of liming on the mobilisation of pollutants colloidal forms

have to be considered in addition to dissolved ones. Several studies emphasize the

role of colloids for heavy metal transport in soils (Jensen et al., 1999, Egli et al.,

1999; Denaix et al., 2001). A pH increase enhances the release and stability of

colloids (Kretzschmar et al., 1999) liming may induce the mobilization of colloid-

associated heavy metals. However, Ca

has a destabilizing effect on colloids

leading to a compression or even collapse of the diffusive double layer (McBride,

1994) which facilitates particle aggregation. In addition, Kretzschmar and Sticher

(1997) observed that the stabilizing effect of organic coatings was reduced by

increasing Ca

concentrations due to the bridging effect of the bivalent Ca

Numerous studies indicate colloid-associated Pb mobilization and translocation to be

a relevant pathway for Pb movement in soils (Jensen et al., 1999). However, there

are very few studies on the occurrence of colloidal As (Buschmann et al., 2006) and

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

Sb (Buddemeier and Hunt, 1988) and detailed information on mobilization

mechanisms and binding of the metalloids onto the colloids are absent. Considering

the high affinity of As to sesquioxides (Kabata-Pendias and Pendias, 2001) and

complexation with dissolved organic matter (Buschmann et al., 2006) we assume

colloidal As mobilization to be important in soils. The following processes and

binding mechanisms may be relevant for the association of As and Sb with colloids:

a) An increase in pH might produce new sorption sites, e. g. by the dispersion of As-

and Sb-sorbing colloids. b) Complexation of As (Mukhopadhyay and Sanyal, 2004;

Buschmann et al., 2006) and Sb (Tighe et al., 2005) with humic acids may enhance

their sorption to colloidal organic matter. c) A pH increase might induce the formation

of cation bridges in the presence of polyvalent cations (for instance bridging products

between Ca, As and dissolved organic matter (DOM) as described by Redman et al.,

2002).

Whilst the role of sesquioxides in the context of As and Sb retention in soils is well

documented, there are only few studies investigating the role of organic matter.

Furthermore, there are no studies focusing on colloid mobilization in organic-rich

soils. Therefore, the aim of our work is to investigate the influence of a pH increase

on the mobilization of soluble and dispersible Pb, As, and Sb in a soil with mainly

organic sorbents. We tested the hypothesis that divalent cations attenuate the

mobilising effect of increasing pH by a) “neutralizing” negative colloid charge and

therefore contributing to enhanced colloid flocculation and by b) immobilizing

dissolved As and Sb by the formation of inorganic precipitates such as Ca

(AsO

)

and Ca[Sb(OH)

]

6.3 Materials and Methods

6.3.1 Soil sampling and soil characterization

We selected a soil with a comparatively high C

org

and low sesquioxide concentration.

Samples were taken from the top soil (Ah-horizon) of a former shooting range in

North Rhine-Westphalia, Germany, wooded with 30-year-old oak trees. The soil’s

parent material is a fluvial, sandy stream deposit of the Pleistocene. The soil was

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

classified (according to World Reference Base) as Podzolic Cambisol with a texture

of loamy sand.

We determined the pH in deionized water and 0.01 M CaCl

solution, using 10 g air-

dried soil and 25 ml of the respective solution (pH Meter 761 Calimatic, Knick). The

suspensions were left to stand for 1 hour before the pH was measured. Organic C

and N (C

org

and N

tot

) were measured by a C and N-analyser (Carlo Erba Instruments,

type C/N NA 1500 N) after drying the soil at 105 °C. For the determination of the

water content, field-moist samples were dried at 105 °C until they reached constant

weight. To measure total acid-soluble As, Al, Fe, Mn and Pb the soil was digested

with concentrated HNO

in closed vessels (two replicates), following the procedure

described by Ilg et al. (2004). The digested samples were filtered, transferred into

volumetric flasks and filled to the mark with deionized water. This solution was

analyzed as described in the analysis section. Total Sb of the soil was determined by

X-ray fluorescence analysis (Philips PW2400 X-ray spectrometer DY 818) in the

form of a powder tablet. Oxalate-extractable Fe and Al were determined by adding

100 mL of ammoniumoxalate/oxalic acid solution to 5 g of air-dried soil. The mixture

was shaken for 2 hours and subsequently filtered over a folded filter. The obtained

filtrates were analysed as described in the analysis section. The determined soil

physico-chemical characteristics are presented in table 6.1.

Table 6.1: Soil properties

Paramter Value

H2O

4.3

CaCl2

3.7

C/N 36.05

C [g kg

-1

] 80.0

N [g kg

-1

] 2.2

Pb [mg kg

-1

] 495

As [mg kg

-1

] 4

Sb [mg kg

-1

] 34

Al [mg kg

-1

] 2534

Fe [mg kg

-1

] 933

Mn [mg kg

-1

] 27

[mg kg

-1

] 700

[mg kg

-1

] 318

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

6.3.2 Dispersion experiments

As there are currently no standard methods for the determination of the stability of

soil colloids batch experiments are commonly used to determine the dispersibility of

colloids. We studied the mobilization of colloidal and soluble Pb, As and Sb in batch

experiments using field-moist samples (< 2 mm) and a soil to water ratio of 1:10. The

samples were shaken for 16 hours on an end-over-end-shaker (10 rpm; GFL 3040).

For the comparison of a pH increase induced by a monovalent and a divalent base

we set up the following series: K-treatment: We added appropriate quantities of 0.1

M KOH and 0.1 M KNO

solution to adjust the pH to 3.5, 4.3, 6.2 and 7.1. Ca-

treatment: We added 0.005 M Ca(OH)

and 0.05 M Ca(NO

)

4 H

O solutions to

adjust the pH to 3.7, 4.4, 5.3, and 6.1. To obtain pH 5.3 and 6.1 we also added solid

Ca(OH)

. We adjusted the pH to 6.7 by adding 0.005 M Ca(OH)

solution and solid

Ca(OH)

. The ionic strength of each individual batch of both treatments was kept

constant by adding required amounts of KNO

and Ca(NO

)

solution, respectively,

amounting to 15 mM.

At the end of the experiment pH was checked in the suspensions prior to filtration

through a 1.2-µm cellulose-nitrate-membrane (Sartorius, Type 11303 – 047N). An

aliquot of the filtrate was acidified (pH < 1) with concentrated nitric acid for the

measurement of the total Pb, As, Sb, and Fe concentration in solution. In addition,

we determined concentrations of total organic C (TOC), turbidity, particle size and

zeta potential (analytical methods see below). An aliquot of the filtrate was

ultracentrifuged at 156 000 x g for 3 hours at 10 °C (Beckman Optima TL) in order to

separate colloids larger than 10 nm and a density > 1.2 g cm³ (based on Stoke’s law)

from the solution. The supernatant was transferred into plastic containers, acidified

with nitric acid and analysed for concentrations of dissolved Pb, As, Sb, Fe and

organic C (DOC). The difference of the metal and C concentration between non-

ultracentrifuged and ultracentrifuged samples accounts for colloidal metal and C

fractions (defined operationally).

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

6.3.3 Analyses

Arsenic was measured by ICP-MS (Varian), Pb and Fe by graphite-furnance atomic

absorption spectroscopy (Spectra AA 880Z, Varian) at a wavelength of 217.0 nm

(Pb) and 248.3 nm (Fe), respectively. Oxalate-extractable Al and Fe concentrations

were determined by flame atomic absorption spectroscopy (Perkin Elmer 1100 B) at

wavelengths of 309.3 nm (Al) and 248.3 nm (Fe), The organic C concentration of the

solutions was measured by a total organic carbon analyser (TOC – 5050 A,

Shimadzu).

The turbidity of soil suspensions was determined using a turbidimeter (Hach 2100P

ISO). In order to check the size of the particles in the supernatant the averaged

volume-based size distribution of the colloids was analysed by dynamic light

scattering (DLS, Malvern Instruments). The Zeta potential was calculated based on

the electrophoretic mobility of the colloids, which was analysed by a Zetasizer 2000

photon correlation spectrometer (Malvern Instruments).

The error bars depicted in the figures represent the experimental error of triplicate

samples.

6.4 Results and Discussion

6.4.1 Colloid mobilization

With increasing pH the amount of dispersed colloids increased in both treatments

(figure 6.1). In the presence of the Ca

cation, however, this increase is offset to

higher pH values. In addition, the amount of dispersed colloids is less than in the

presence of the monovalent K

cation. With increasing pH zeta potential values of

the colloids become increasingly negative (figure 6.2). Again, this pH-induced charge

effect is less pronounced in the Ca treatment than in the K treatment. While K

adsorbs non-specifically, specific adsorption of Ca

to anionic groups of natural

organic matter leads to the compensation of the negative colloid charge as explained

by Tipping and Cooke (1982). Therefore, Ca

results partly in enhanced colloid

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

flocculation (Tiller and O’Melia, 1993) which is reflected in decreased turbidity values

of the Ca treatment (figure 6.2). The particle size ranged from approximately 300 to

1000 nm. The pH and the valency of the counterion had no significant effect on

colloid size (data not shown).

3 4 5 6 7 8

Turbidity [FNU]

K Ca

Figure 6.1: Turbidity as a function of pH for the KOH and Ca(OH)

treatment (bars depict

one standard error)

-50

-40

-30

-20

-10

3 4 5 6 7 8

Zeta potential [mV]

K Ca

Figure 6.2: Zeta potential as a function of pH for the KOH and Ca(OH)

treatment (bars depict one

standard error)

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

6.4.2 Mobilization of soluble and dispersible Pb, As, and Sb

The comparatively low concentration of sesquioxides and high concentration of

organic C of our study soil suggests organic matter to be the dominant binding

partner of Pb, As and Sb. An increase in pH results in the enhanced release of both

soluble and dispersible organic C with this increase in the presence of the

monovalent cation being much more pronounced than in the presence of the divalent

cation (data not shown).

6.4.2.1 Lead

The concentrations of both colloidal and dissolved Pb increase following the pH

increase induced by addition of KOH (figure 6.3a and 6.3b). In the presence of the

divalent Ca cation however, dissolved and colloidal concentrations differ from the

KOH-treated samples. The decrease of dissolved Pb concentrations reached a

minimum at pH 5.3 after which they rose with increasing pH. Since Ca decreased

the pool of mobilized DOC available for Pb-complexation, at pH values near 5.3 Pb

is to a great part removed from solution by either precipitation as Pb(OH)

or by

adsorption onto the solid phase. At pH values greater than 5.3, however, the

negative charges of deprotonated carboxyl groups exceed the amount of positive

charges of the Ca cation. As a consequence, concentrations of DOM and thus also

concentrations of dissolved organically bound Pb increase.

3 4 5 6 7

Ca: Mobilized Pb

diss

[mg kg

-1

]

K: Mobilized Pb

diss

[mg kg

-1

]

Ca K

Figure 6.3a: Mobilized dissolved Pb as a function of pH for the KOH (K) and Ca(OH)

(Ca) treatment

(bars depict one standard error)

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

Colloidal Pb in the Ca(OH)

-treated samples could only be detected at pH > 4.4

where the increasingly negative particle charge induced by a pH increase is no

longer compensated by Ca cations. However, Ca (in comparison to K) did not

attenuate the concentration of colloids (i. e. the turbidity) to the same degree as

concentrations of colloidal Pb. Therefore, Pb-displacement from the colloid by Ca

due to competition for binding sites as described by Harter (1979) might additionally

cause a decrease in colloidal Pb concentrations.

3 4 5 6 7

K: Mobilized Pb

coll

[mg kg

-1

]

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Ca: Mobilized Pb

coll

[mg kg

-1

]

K Ca

Figure 6.3b: Mobilized colloidal Pb as a function of pH for the KOH (K) and Ca(OH)

(Ca) treatment

(bars depict one standard error)

6.4.2.2 Arsenic

Dissolved As increases with increasing pH in the presence of the monovalent cation

(R² = 0.94, figure 6.4) in accordance with results reported by Smith et al. (1998),

suggesting As-desorption from sesquioxides. In addition, the pH-induced dissolution

of humic acids to which (according to Warwick et al., 2006) As is adsorbed to, may

explain the increase in dissolved As. Colloidal As concentrations (figure 6.4) in this

treatment are much lower than dissolved concentrations suggesting As has only a

minor association with colloids. Even though colloidal Fe and colloidal organic C

concentrations increase continuously with increasing pH (data not shown), colloidal

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

As concentrations remain constant and do not show any pH-dependence. This

implies limited availability of colloidal binding sites for As. The high ratio between

organic C and Fe and organic C and Al suggests possible As-sorption sites provided

by sesquioxides to be greatly reduced by the presence of organic coatings. Possible

As-binding sites on organic matter, such as amino groups (Thanabalasingam and

Pickering, 1986), are more abundant near the mineral surface where they are being

adsorbed (as described in the model proposed by Kleber et al., 2007). This reduced

availability caused by direct blockage of sorption sites provided by the mineral

surface and by a reduction in the accessibility of these sites allows for a very limited

As sorption. A further explanation for the occurrence of only little colloidal As is the

presence of Fe-organic clusters. The accessibility of organic functional groups

suitable for the binding of As in these clusters would be very low.

200

400

600

3 4 5 6 7 8

Mobilized As [µg kg

-1

]

dissolved As - K

dissolved As - Ca

colloidal As - K

colloidal As - Ca

Figure 6.4: Mobilized dissolved (blank symbols) and colloidal (filled symbols) As as a function of pH

for the KOH (K) and Ca(OH)

(Ca) treatment (bars depict one standard error)

The addition of the Ca cation led to a continuous decrease in dissolved As with

increasing pH. These findings are in line with Voigt et al. (1996), who report the

addition of Ca(OH)

to reduce exchangeable As. Since the formation of Ca

(AsO

)

in our suspensions

may be ruled out based on calculations of the solubility product,

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

the observed decline in dissolved As may be attributed to increased “bridging” with

increasing pH as suggested by the mechanism described by Redman et al. (2002):

negatively charged As anions and DOM are linked by positively charged Ca cations.

Similarly, Thanabalasingam and Pickering (1986) found polyvalent cations enhance

As sorption to humic acids providing evidence for our assumption that Ca crosslinks

between several arsenate and DOM ions to form larger compounds. Since we did

not detect any colloidal As in the Ca(OH)

treatment we assume the formed bridging

products to be either retained by the soil matrix or to be dispersed in suspension in a

size exceeding our filter cut-off (1.2 µm).

6.4.2.3 Antimony

Similarly to As, organic matter is postulated to be the main binding partner of Sb in

this soil due to the low concentrations of Fe and Al. Dissolved Sb concentrations

increase following a pH increase in both treatments, with the increase being more

pronounced after the addition of KOH (figure 6.5). The increase in dissolved Sb in

the KOH treatment is in agreement with Tighe et al. (2005), who reported sorption of

Sb(V) by humic acid decreasing as the pH increases. In the KOH treatment,

dissolved Sb shows the same response to a pH increase as dissolved As.

While interactions between As, Ca and dissolved organic matter have already been

reported (Redman et al., 2002) studies describing the role of cation bridging between

organic matter and Sb are absent. In the presence of Ca cations, our results show

an increase in dissolved Sb concentrations with increasing pH suggesting that Sb

does not bridge to non-mobile or aggregated organic matter to the same extent as

As. We ascribe this observation to the different accessibility of the negative charge

of these two anions caused by differences in their chemical structure: In the arsenate

anion, As forms covalent bonds with oxygen- and hydroxyl-groups in a tetrahedric

structure with the negative charge being located at the deprotonated oxygen atom at

the edge of the ion, where it is freely accessibly by other ions in solution. In the case

of Ca

this allows for bridging reactions between As and DOM leading to the

flocculation of DOM-Ca-As-consisting compounds. In the Sb[(OH

)]

-complex,

however, the negative charge is located at the Sb atom which is coordinated by 6

octahedral OH-groups (King et al., 1995) with these OH-groups shielding the

negative charge. The shielded negative charge of the Sb anion would render

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

bridging between increasingly negatively charged DOM and Sb by the Ca cation

more difficult. The assumption that Sb does not bridge with Ca ions is supported by

the fact that the ratio of Sb/DOC in both treatments (data not shown) does not differ

significantly from each other. Since Sb may be associated with DOM (Tighe et al.,

2005) and therefore co-dissolution with DOC may be possible, the equal Sb/DOC-

ratio indicates that the lower mobilization of soluble Sb in the Ca(OH)

treatment is

due to less DOM mobilization and not due to bridging of Ca.

Antimony was not found to be associated with colloids in any treatment (figure 6.5).

The precipitation of Ca-antimony could be ruled out since dissolved concentrations

of Sb were below the solubility product.

100

200

300

2 3 4 5 6 7 8

Mobilized Sb [µg kg-1]

dissolved Sb - K

dissolved Sb - Ca

colloidal Sb - Ca

colloidal Sb - K

Figure 6.5: Mobilized dissolved (blank symbols) and colloidal (filled symbols) Sb as a function of pH

for the KOH (K) and Ca(OH)

(Ca) treatment (bars depict one standard error)

Mobilization of soluble and dispersible Pb, As, and Sb – pH and Counterion Valency

6.5 Conclusion

Our experiments demonstrate that Ca counteracts the pH effect. Whereas the

formation of Ca

(AsO

)

precipitates may be ruled out Ca was found to impact on the

mobilization of sorbents. The presented results suggest that the counterion valency

controls the mobility and sorption capacity of the sorbents (i.e. colloids, DOM):

(a) It governs (i) the mobilization of colloids by impacting on colloid charge and (ii)

the mobilization of DOM by flocculation.

(b) It governs the interactions of the metal(loid)s with the soil sorbents and therefore

impacts on dissolved (i.e. Pb-DOC-complexation, formation of insoluble “bridging

products” between As, Ca, and organic matter) and colloidal (competition for binding

sites between Ca and Pb) metal(loid) concentrations.

Future research should address the role of the Ca

cation in the interactions

between As/Sb, DOM and Ca

in defined systems.

Our data indicates that while liming-induced pH increase does not pose any major

risk of colloidal As and Sb mobilization as long as divalent cations dominate in

solution the mobilization of soluble Sb needs to be considered. Risk assessment of

limed sites should not only address dissolved, but also colloidal Pb concentrations to

avoid potential Pb leaching. Further studies are needed to address the role of

varying sorbent composition in colloidal metal(loid) mobilization.

7 Synthesis and general Conclusions

7.1 The governing role of soil solid phase composition in soluble and

dispersible metal(loid) mobilization in the context of drying and liming

The overall data suggest that the composition of the soil solid phase, i. e. C

org

and

sesquioxide concentration, may turn the balance in favour or against the mobilization

of the individual dissolved or colloidal metal(loid) phase. In the following sections, I

will discuss this assertion in more detail and elucidate the role of the composition of

the soil sorbents and dispersible colloids in the context of both soluble and

dispersible metal(loid) mobilization.

7.1.1 Drying

My results of chapter 3 demonstrate that the effect of drying on colloid mobilization

depends on the concentration of soil organic C: Whilst in “low C

org

-soils” (C

org

concentration between 2.0 and 5.6 %) drying led to either an immobilization of

colloids or to no significant change in colloid mobilization, it led to a pronounced

increase in the amount of mobilized colloids in a soil with a C

org

concentration of 24.5

% (“high C

org

-soil”). In the latter, organic colloids were mobilized while in the “low

org

-soils” mineral colloids were immobilized. In most cases of the “low C

org

-soils”

theses changes went along with a decrease in mobilized colloidal heavy metal

concentrations, whereas in the soil containing 24.5 % C

org

no significant increase in

colloidal heavy metal mobilization could be found. While drying-induced mobilization

of DOC had no effect on dissolved concentrations of Cd, Pb, and Zn it caused a

redistribution of Cu from the colloidal to the dissolved phase. These findings indicate

that the concentration of organic matter plays a key role in both the (im)mobilization

of colloids as well as the mobilization of soluble and colloidal heavy metals.

Synthesis and general Conclusions

100

7.1.2 Liming

The soils used for the investigations of the effects of a pH increase and counterion

valency (chapter 4 and 6) differ in their C

org

concentration and C

org

/Fe

tot

-ratio (table

7.1).

Table 7.1:

org

concentration and C

org

/Fe

tot

-ratio of the soils used in chapter 4 and 6

org

[g kg

-1

] Fe

tot

[g kg

-1

] Ratio C

org

/Fe

tot

(based on moles)

Soil 1 (chapter 4) 209 3.1 221

Soil 2 (chapter 6) 80 0.9 314

The ratio between C

org

concentration and added Ca(OH)

(as required to adjust pH

7, data not shown) is similar in both soils. This indicates that organic matter makes

up the major part of the base neutralization capacity and that the amounts of base

neutralizing functional groups (i.e. carboxyl groups) per mass unit of organic matter

are the same in both soils.

The soils show an onset in colloid mobilization at different pH values in the presence

of the Ca

ion (figure 7.1), which is in line with colloidal Pb mobilization: In the soil

with a C

org

concentration of 80 g kg

-1

I could detect a pronounced colloid mobilization

at pH > 4.4. This increase in mobilized colloids following Ca(OH)

-addition takes

place at lower pH values than in the soil with a C

org

concentration of 209 g kg

-1

where I observed a strong increase in mobilized colloids only at pH values > 5.8. The

onset in colloid mobilization goes along with a very pronounced increase in the

negative colloid charge (figure 7.2). At pH > 4.4, however, colloid charge of the soil

with a concentration of 80 g kg

-1

org

was much more negative than in the soil with a

concentration of 209 g kg

-1

org,

explaining

the stronger dispersion of colloids.

However, these results appear contradictory to findings by numerous authors who

report the mobilization and stabilization of colloids to be enhanced by DOM since it

increases the negative colloid charge (Kretzschmar, 1999; Kaplan et al., 1996; Heil

and Sposito, 1993).

Since the base neutralization capacity, i.e. the amount of base neutralizing functional

groups (i.e. carboxyl groups) per mass unit organic matter of the soils is similar this

Synthesis and general Conclusions

101

parameter may be ruled out to explain the different pH at which colloid mobilization

starts, but there are several possible explanations for this apparent contradiction.

Ionic strength. The ionic strength of the soil with a concentration of 209 g kg

-1

org

approximately 3 times greater than the ionic strength of the soil with a concentration

of 80 g kg

-1

org

. Since for same electrolytes, an increasing ionic stength results in a

decreasing zeta potential (Müller, 1996) this could explain the lower zeta potential of

the soil with a C

org

concentration of 209 g kg

-1

Different ratios between the C

org

and Fe

tot

concentrations of the soils. The zeta

potential of colloids is governed by the sum of negatively and positively charged

sites (i. e. negatively charged functional groups of organic matter and positively

charged surfaces of inorganic minerals such as Fe- or Al-(hydr)oxides). Different

ratios between the Fe

tot

and C

org

concentrations could therefore explain why the soil

with a concentration of 80 g kg

-1

org

shows greater negative charges at pH > 4.4

than the soil with a concentration of 209 g kg

-1

org

(figure 7.2): The relative amount

of dispersible Fe-(hydr)oxides in the soil with a C

org

/Fe

tot

ratio of 314 is lower than in

the soil with a C

org

/Fe

tot

ratio of 221. This means that the relative contribution of

positively charged Fe-(hydr)oxides to the averaged zeta potential is less, thus

resulting in lower overall negative charges than in the soil with a C

org

/Fe

tot

ratio of

221.

In addition, the larger Fe

tot

concentration of soil 1 may provide a further explanation.

Supposing the colloidal Fe concentrations in the suspensions are greater in soil 1

than in soil 2, the lower zeta potential in soil 1 could be a result of additional

crosslinking reactions between Fe and organic matter (Schaumann, 2006), resulting

in a compensation of negatively charged functional groups.

Synthesis and general Conclusions

102

3 4 5 6 7

Turbidity [FNU]

Corg: 8.0 %

Corg: 20.9 %

Figure 7.1:

Turbidity of soil suspensions of soil 1 and 2 with a high and a low concentration of organic

matter as a function of pH (Ca(OH)

treatment; bars depict one standard error)

-20

-18

-16

-14

-12

-10

-8

-6

-4

-2

3 4 5 6 7

Zeta potential [mV]

Corg 20.9 %

Corg 8.0 %

Figure 7.2: Zeta potential of soil 1 and soil 2 with a high and a low concentration of organic matter as

a function of pH (Ca(OH)

treatment; bars depict one standard error)

Synthesis and general Conclusions

103

These observations indicate that the composition of soil colloids

may play an

important role in colloid mobilization since it may control the onset pH of colloid

mobilization.

The experiments discussed in chapter 5 and 6 were conducted on soils with similar

org

concentrations, but greatly different concentrations in oxalate-extractable Al and

Fe. In the following, they are therefore referred to as high and low C/Fe-soil,

respectively. These two types of soils provide differently composed (a) sorbents for

dissolved metal(loid)s and (b) pools of potentially mobilizable colloids. My results

show that in these two soils a pH increase in the presence of a monovalent and a

divalent counterion leads to differences in the mobilization of both soluble and

colloidal Pb, As and Sb. In addition, different types of colloids are mobilized

providing various types of sorption sites for different metal(loid)s (Lyvén et al., 2003).

The provision of dispersed sorption sites does not only affect colloidal metal

concentrations but also dissolved ones. In the following, I will compare relevant

differences and outline similarities in soluble and dispersible metal(loid) mobilization

between the two types of soils and discuss responsible mechanisms.

7.1.2.1 Lead

I did not detect any difference in the influence of the KOH treatment on the

mobilization of colloidal and soluble Pb between the two types of soils. However, in

the Ca(OH)

treatment of the low C/Fe-soil, a pH increase from the soil original pH

5.0 to pH 6.5 resulted in decreasing dissolved Pb concentrations, whereas they

increased in the high C/Fe-soil following the pH minimum at pH 5.3. This difference

might be ascribed to the smaller ratio between dissolved Pb and DOC of the high

C/Fe-soil, implying enough DOC to allow for Pb complexation. In the low C/Fe-soil,

the higher concentrations of mobilized Fe diminished the DOC-pool available for Pb-

complexation, resulting in a decrease in dissolved Pb concentrations. However,

since in the low C/Fe-soil, there are only two data points, pH 4.7 and 6.5, it is

impossible to define the minimum of the curve and therefore I cannot rule out

increasing concentrations at higher pH.

Synthesis and general Conclusions

104

7.1.2.2 Arsenic

The response of As to a pH increase in the presence of a monovalent base differed

strongly between both soils. Whilst dissolved As concentrations in the high C/Fe-soil

increased, presumably due to the dissolution of organic matter to which according to

Warwick et al. (2005

)

As is associated with, concentrations dropped in the low C/Fe-

soil. I ascribe this difference in As distribution to the presence of different sorption

sites in both soils. In the low C/Fe-soil, micro-aggregate dispersion of sesquioxide-

bearing colloids may provide a high amount of high-affinity sorption sites for As,

leading to an increasing colloidal As and consequently decreasing dissolved As

concentration. In that context it is interesting to note that even though the total

concentration of colloidal As increased, the amount of As sorbed per g Fe-containing

colloids decreased from pH 4.7 (17.9 µg As/g Fe) to pH 6.5 (5.2 µg As/g Fe). This

indicates that with increasing pH, As desorbs from colloidal sesquioxides. However,

due to the great release of As-sorbing colloids, net concentrations of colloidal As still

increase.

A pH increase in the presence of a divalent cation led to decreasing concentrations

in both dissolved and colloidal As.

7.1.2.3 Antimony

Results of the low C/Fe-soil show a similar response to increasing pH as the high

C/Fe-soil: dissolved Sb concentrations increase following a pH-increase to 6.5, with

the increase being more pronounced in the KOH than in the Ca(OH)

treatment. The

observed pH-dependence of the low C/Fe-soil is in conflict with findings reported by

Tighe et al. (2005), who report > 99 % to be absorbed by amorphous Fe(OH)

regardless of pH. This mechanism would suggest very low dissolved Sb

concentrations. However, if sesquioxides or –hydroxides are partly coated with

organic matter – which is, considering the elevated DOC concentrations, very likely

the case in the low C/Fe-soil – organic matter would act as a main binding partner.

As the pH increases and organic matter begins to dissolve, associated Sb would

also dissolve. This mechanism could explain the increase in dissolved Sb with

increasing pH. It therefore points out a governing role of organic matter in Sb

mobilization. While there is no colloidal Sb in the high C/Fe-soil, there is some in the

Synthesis and general Conclusions

105

low C/Fe-soil. It increases slightly following the addition of KOH and drops after

Ca(OH)

-addition, suggesting a possible association of Sb with colloidal

sesquioxides, whose dispersion is enhanced in the presence of the monovalent

cation and decreased in the presence of the divalent cation.

7.1.3 Is there a critical zeta potential for colloid mobilization?

The zeta potential as a measure for the colloid surface charge is a governing

parameter for colloid stability (Séquaris and Lewandowski, 2003). Considering the

zeta potential across all measured samples my data set suggests colloid

mobilization to show a pronounced increase at a zeta potential < - 14 mV (figure

7.3). This is similar to results by Séquaris and Lewandowski (2003), who report

colloid release in batch experiments with agricultural top soil (C

org

: 2 to 5 %) to start

at zeta potential values < - 16 mV. However, both values are less negative than the

critical zeta potential reported in the study by Johnson (1999). At similar ionic

strength, the author found zeta potentials on quartz sand of – 34 mV. This difference

in critical zeta potential between colloids mobilized from top soil samples and quartz

sand could be ascribed to steric effects as described by Heil and Sposito (1993),

which come additionally into play on organic colloids or organic-coated mineral

colloids. Considering the organic-rich matrices (C

org

concentration ≥ 8 %) of all my

soil samples I presume steric stabilization might be an effect in addition to the zeta

potential, enhancing colloid stability as suggested by Kretzschmar et al. (1997).

Synthesis and general Conclusions

106

-40-35-30-25-20-15-10-50

Zeta potential [mV]

Turbidity [FNU]

Figure 7.3: Turbidity as a function of the zeta potential of all soils

7.2 General conclusions and derived implications for the management of

contaminated sites

In this thesis I could demonstrate that different mechanisms are responsible for the

mobilization of soluble and colloidal metal(loid)s in both a liming-induced pH

increase as well as in a drying event. My results also provide evidence that the

composition of the soil solid phase (i. e. its individual components) may control the

release of soluble and colloidal metal(oid)s (figure 7.4). It provides an indicator for

(a) the composition of sorbents (i.e. of possible associations of the individual

components of the soil solid phase with each other and the hereby resulting sites

available for the sorption of individual ions present in the soil solution) and (b) the

composition of the pool of potentially dispersible colloids. Actual colloid mobilization

is controlled by physico-chemical parameters of the soil solution such as pH, DOC

concentration, ionic strength and the valency of available cations. The composition

of the mobilized colloids governs the distribution of the respective metal(loid)

between the dissolved and colloidal phase. In addition, differently composed colloids

Synthesis and general Conclusions

107

showed different responses to environmental changes. While organic colloids were

mobilized in a drying event, inorganic colloids were immobilized.

Figure 7.4: Relation between the soil solid phase, soil sorbents and colloidal and dissolved

metal(loid)s

The data suggest that in the context of a liming-induced pH increase colloid

composition may determine the onset-pH of colloid mobilisation. Furthermore, colloid

composition governs whether a pH increase leads to an increase or decrease in the

respective concentrations of dissolved or colloidal metal(loid)s. Dissolved and

colloidal Pb may be increased following a liming-induced pH increase if a certain

minimum pH is exceeded. This pH minimum is determined by the ratio between

organic and inorganic colloidal components. Whilst the mobilization of colloidal As

and Sb does not seem to be of any relevance in a liming-induced pH increase, I

observed decreasing dissolved As concentrations, but increasing dissolved Sb

concentrations. It became obvious that the three metals of main concern on shooting

range sites may respond differently to a pH increase. According to the US CERCLA

(Comprehensive Environmental Response, Compensation, and Liability Act) List As

and Pb are ranked as the metals with the highest environmental concern since they

are “determined to pose the most significant potential threat to human health due to

their known or suspected toxicity and potential for human exposure” at seriously

contaminated sites (ATSDR, 2005). Since Pb makes out the highest percentage of

Composition of soil solid phase

Composition of pool of

potentially dispersible

colloids

Mobilized colloids

Composition of sorbents

Dissolved metal(oid)s

Colloidal metal(oid)s

Composition of soil solid phase

Composition of pool of

potentially dispersible

colloids

Mobilized colloids

Composition of sorbents

Dissolved metal(oid)s

Colloidal metal(oid)s

Synthesis and general Conclusions

108

the overall contamination I suggest increasing the pH as far as to the pH minimum of

soluble and colloidal Pb mobilization (i. e. in my studied acidic forest soil of chapter 6

to pH 4.4 and 5.3, respectively). This pH would also assure decreasing dissolved As

concentrations and risks only a minor increase in dissolved Sb concentrations.

These results suggest that on sites, which are contaminated by metal(loid)s differing

in their responses to lime application, a differentiated approach is necessary. The

optimum pH, i. e the pH at which the overall risk of both colloidal and dissolved metal

mobilization is minimized has to be worked out experimentally in these cases. This

should also consider the option that metals might get mobilized in the topsoil but

could be immobilized in deeper mineral horizons due to changing soil pH and soil

composition.

7.3 Outlook

With this study, I could demonstrate relevant processes of soluble and dispersible

metal(loid) mobilization in the context of a liming-induced pH increase and in a

drying event. However, the question which mechanism controls the mobilization in a

scenario where both, liming and drying, occur simultaneously will remain a matter of

debate: While my data in general suggests Ca to suppress colloid mobilization, Flury

(2007) found, that the presence of a liquid-gas interface, which may form during the

drying process, governs colloid stability. The author showed that if the colloid is

attached to the liquid-gas interface, a high ionic strength solution containing Ca ions

would not lead to the (according to the DLVO-theory) expected flocculation of

colloids. However, if the colloid is attached to the solid-liquid interface, the same high

ionic strength solution would induce the flocculation of colloids. These findings

indicate that effects caused by drying could override effects caused by liming.

Studies looking into the combination of both effects should therefore be the subject

of future research.

My results suggest that the composition of the soil solid phase, i. e. the

concentration of C

org

and sesquioxide governs the mobilization of colloids as well as

soluble and dispersible metal(loids) in the context of a drying and liming event.

Synthesis and general Conclusions

109

However, since the number of samples is too small for the derivation of any general

relation between the composition of the soil solid phase and mobilized colloidal

metal(loid)s future studies should investigate more soils across a wider range of

composition in order to gain a better understanding about the role of the composition

of the soil solid phase not only in soluble but also in dispersible metal(loid)

mobilization. In order to elucidate (im)mobilization mechanisms of As and Sb, future

studies should address the role of the Ca cation in the interactions between As/Sb,

DOM and Ca in defined systems. Up to now, numerous authors have raised

speculations about how Sb could possibly be associated to different sorbents of the

soil matrix, for instance by specific adsorption (Tighe et al, 2005) or organic

complexation (Lyvén et al., 2003). However, only very few studies have clearly

identified the binding mechanism (i. e. inner-sphere complexes on Fe-oxide

surfaces; Scheinost et al., 2006). Since the knowledge about the type of bonding is

crucial for the estimation of the mobilization potential more research should be

directed towards the identification of Sb binding types to different sorbents of the

soil.

To provide an estimate how much of the colloidal metal pool could potentially

become available in the dissolved phase once the colloid has been translocated to a

different environment, future research should investigate labile fractions of the

metal(loid)s associated to the colloid.

A next step following the batch studies of this thesis should include experiments,

which are conducted under more natural conditions in order to clarify the

environmental relevance of the identified processes. This should eventually lead to

the derivation of key parameters, which allow for the quantification of colloidal

metal(loid)s that could potentially be leached from soil to groundwater.

111

8 References

Adamson, A. W. 1990. Physical chemistry of surfaces. Wiley, New York.

Bachmann, J., Woche, S. K., Goebel, M.-O., Kirkham, M. B. and Horton, R., 2003. Extended

methodology for determining wetting properties of porous media. Water Resources Research

39 (12), Art. No. 1353.

Bartlett, R., James, B., 1980. Studying dried, stored soil samples - some pitfalls. Soil Science Society

of America Journal 44, 721–724.

Baskaran, S., Bolan, N. S., Rahman, A., Tillman, R. W., Macgregor, A. N., 1994. Effect of drying of

soils on the adsorption and leaching of phosphate and 2,4-dichlorophenoxyacetic acid.

Australian Journal of Soil Research 32, 447–64.

Bergendahl, J., 2005. Batch leaching tests: Colloid release and PAH leachability. Soil and Sediment

Contamination 14, 527–543.

Bethwell, C., 2004. Organische Substanz und ihr partikulären und adsorbierten Anteil im

Feinsediment eines bergbaubeeinflussten Fließgewässers. Aktuelle Reihe der BTU Cottbus

3, 109–118. http://www.tu-

cottbus.de/BTU/Fak4/Gewschu/downloads/aktuelle_reihe/ar_3_2004/ar_03_2004_10_bethwe

ll.pdf (verified 17th October 2007).

Brady, N.C., Weil, R.R., 2002. The Nature and Properties of Soils. 13th ed. Prentice-Hall. Upper

Saddle River, NJ.

Brannon, J. M., Patrick, Jr., W. H., 1985. Fixation and mobilization of antimony in sediments.

Environmental Pollution (Series B) 9, 107–126.

Breiner, J. M., Andersson, M. A., Tom, H. W. K., Graham, R. C., 2006. Properties of surface-modified

colloidal particles. Clay and Clay Minerals 54, 12–24.

Brezesinski, G., Mögel, H.-J., 1993. Grenzflächen und Kolloide. Spektrum Akademischer Verlag,

Heidelberg.

Brown, J. N., Peake, B. M., 2003. Determination of colloidally-associated polycyclic aromatic

hydrocarbons (PAHs) in fresh water using C18 solid phase extraction disks. Analytica

Chimica Acta 486, 159–169.

Brümmer, G. W., Gerth, J., Herms, U., 1986. Heavy metal species, mobility and availability in soils.

Journal of Plant Nutrition and Soil Science 149, 382–398.

Buddemeier, R. W., Hunt, J. R., 1988. Transport of colloidal contaminants in groundwater:

radionuclide migration at the Nevada Test Site. Applied Geochemistry 3, 535–548.

References

112

Buffle, J., Leppard, G. G., 1995. Characterization of aquatic colloids and macromolecules. 2. Key

Role of physical structures on analytical results. Environmental Science and Technology 29,

2176–2184.

Burba, P., Aster, B., Nifant’eva, T., Shkinev, V., Spivakov, B.Y., 1998. Membrane filtration studies of

aquatic humic substance and their metal species: a concise overview. Part 1. Analytical

fractionation by means of sequential-stage ultrafiltration. Talanta 45, 977–988.

Burba, P., Van den Bergh, J., 2002. On-site classification of humic-rich hydrocolloids and their metal

species by means of online multistage ultrafiltration. Environmental Science and Technology

36, 4114–4120.

Buschmann J., Kappeler A., Lindauer U., Kistler D., Berg M., Sigg L., 2006. Arsenite and arsenate

binding to dissolved humic acids: Influence of pH, type of humic acid, and aluminium.

Environmental Science and Technology 40, 6015–6020.

Buykx, S. E. J., van den Hoop, A. G. T., Cleven, R. F. M. J., Buffle, J., Wilkinson, K. J., 2000.

Particles in natural surface waters: chemical composition and size distribution. International

Journal of Environmental Analytical Chemistry 77, 75–96.

Carvalho, R. A., Benfield, M. C., Santschi, P. H., 1999. Comparative bioaccumulation studies of

colloidally complexed and free-ionic heavy metals in juvenile brown shrimp Penaeus aztecus

(Crustacea: Decapoda: Penaeidae). Limnology and Oceanography 44 (2), 403–414.

Celi, L., Presta, M., Ajmore-Marsan, F., Barberis, E. (2001). Effects of pH and electrolytes on inositol

hexaphosphate interaction with goethite. Soil Science Society of America Journal 65, 753–

760.

Christ, M. J., David, M. B., 1996. Temperature and moisture effects on the production of dissolved

organic carbon in a Spodsol. Soil Biology and Biochemistry 28 (9), 1191–1199.

Citeau, L., Lamy, I., van Oort, F., Elsass, F., 2001. Nature des sols et nature des colloïdes circulant

dans les eaux gravitaires: une etude in situ. Earth and Planetary Sciences 332, 657–663.

Courchesne, F., Savoie, S., Dufresne, A., 1995. Effects of air-drying on the measurement of soil pH in

acidic forest soils of Quebec, Canada. Soil Science 160, 56–68.

Curtin, D., Campbell, C. A., Zentner, R. P., Lafond, G. P., 1994. Long-term management and clay

dispersibility in two Haploborolls in Saskatchewan. Soil Science Society of America Journal

58, 962–967.

Curtin, D., Steppuhn, H., Selles, F., 1995. Determination of clay dispersion and flocculation in sodic

soil as influenced by soil:solution ratio. Communications in Soil Science and Plant Analysis

26, 209–221.

Czigany, S., Flury, M., Harsh, J.B., 2005. Colloid stability in vadose zone aggregate. Hanford

sediments. Environmental Science and Technology 39, 1506–1512.

Dahlgren, R. A., Marrett, D. J., 1991. Organic carbon sorption in arctic and subalpine spodosol B

horizons. Soil Science Society of America Journal 55, 1382–1390.

References

113

Dekker, L. W., Doerr, S. H., Oostindie, K., Ziogas, A. K., Ritsema, C. J., 2001. Water repellency and

critical soil water content in a dune sand. Soil Science Society of America Journal 65, 1667–

1674.

Denaix, L., R., Semlali, M., Douay, F., 2001. Dissolved and colloidal transport of Cd, Pb, and Zn in a

silt loam soil affected by atmospheric industrial deposition. Environmental Pollution 113, 29–

38.

Doerr, S. H., Shakesby, R.A, Walsh, R.P.D., 2000. Soil water repellency: its causes, characteristics

and hydro-geomorphological significance. Earth-Science Reviews 51, 33–65.

Doerr, S. H., Thomas, A. D., 2000. The role of soil moisture in controlling water repellency: new

evidence from forest soils in Portugal. Journal of Hydrology 231, 134–147.

Edwards, R., Lepp, N. W., Jones, K. C., 1999. Antimon. p. 333–337. In: Alloway, B. J. (ed.),

Schwermetalle in Böden. Springer Verlag, Berlin.

Egli, M., Fitze, P., Oswald, M., 1999. Changes in heavy metal contents in an acidic forest soil affected

by depletion of soil organic matter within the time span 1969–93. Environmental Pollution 105,

367–379.

El-Farhan, Y. H., Denovio, N. M., Herman, J. S., Hornberger, G. M., 2000. Mobilization and transport

of soil particles during infiltration experiments in an agricultural field, Shenandoah Valley,

Virginia. Environmental Science and Technology 24, 3555–3559.

Environmental Protection Agency (EPA) 1994. Method 3015: Microwave-assisted digestion of

aqueous samples and extracts. http://www.epa.gov/sw-

846/pdfs/3015.pdf#search=%22EPA%20method%203015%22 (verified 17

October 2007).

Environmental Protection Agency 2001. Best Management Practices for Pb at outdoor shooting

ranges. http://www.epa.gov/region2/waste/leadshot/download.htm (verified 17th October

2007).

Fichte, B. (1986, ed.), Organikum, VEB Berlin.

Flynn, H. C., Meharg, A. A., Bowyer, P. K., Paton, G. I., 2003. Antimony bioavailability in mine soils.

Environmental Pollution 124, 93–100.

Fotovat A., Naidu, R., Sumner, M. E. 1997. Water:soil ratio influences aqueous phase chemistry of

indigenous copper and zinc in soils. Australian Journal of Soil Research 35, 687–709.

Gadel, F., Serve L., Benedetti, M., Da Cunha, L. C., Blazi, J. L., 2000. Biogeochemical characteristics

of organic matter in the particulate and colloidal fractions downstream of the rio Negro and

Solimoes rivers confluence. Agronomie 20, 477–490.

Goebel, M.-O., Bachmann, J., Woche, S. K., Fischer, W. R., Horton, R., 2004. Water potential and

aggregate size effects on contact angle and surface energy. Soil Science Society of America

Journal 68, 383–393.

References

114

Grønflaten, L., Amundsen, L. Frank, J., Steinnes, E., 2005. Influence of liming and vitality fertilization

on trace element concentrations in Scots pine forest soil and plants. Forest Ecology and

Management 213, 261–272.

Guiné, V., Martins, J., Gaudet, J.P., 2003. Facilitated transport of heavy metals by bacterial colloids in

sand columns. Journal de Physique IV France 107, 593–596.

Harter, R. D. 1979. Adsorption of copper and lead by Ap and B2 horizons of several northeastern

United States soils. Soil Science Society of America Journal 43, 679–683.

Haynes, R. J., Swift, R. S., 1989. The effects of pH and drying on adsorption of phosphate by

aluminium-organic matter associations. J. Soil Sci. 40, 773–781.

Heil, D., Sposito, G., 1993. Organic matter role in illitic soil colloids flocculation: II. Surface charge.

Soil Science Society of America Journal 57, 1246–1253.

Herms, U., Brümmer, G., 1984. Einflussgrössen der Schwermetallöslichkeit und -bindung in Böden.

Zeitschrift für Pflanzenernährung und Bodenkunde 147, 400–424.

Hill, D.M., Aplin, A.C., 2001. Role of colloids and fine particles in the transport of metals in rivers

draining carbonate and silicate terrains. Limnology and Oceanography 46, 331–344.

Hsia, T-H., Lo, S.-H., Lin, C.-F., Lee, D.-Y. 1994. Characterization of arsenate adsorption on hydrous

iron oxide using chemical and physical methods. Colloids and Surfaces A: Physico-chemical

and engineering aspects 85, 1–7.

Hurraß, J., Schaumann, G. E., 2006. Properties of soil organic matter and aqueous extracts of

actually water repellent and wettable soil samples. Geoderma 132, 222–239 .

Huynh, L., Jenkins, P., 2001. A rheological and electrokinetic investigation of the interactions between

pigment particles dispersed in aqueous solutions of short-chain phosphates. Colloids and

Surfaces A – Physicochemical and Engineering Aspects 190, 35–45.

Ilg, K., Wilcke, W., Safronov, G., Lang, F., Fokin, A., Kaupenjohann, M., 2004. Heavy metal

distribution in soil aggregates: a comparison of recent and archived aggregates from Russia.

Geoderma 123, 153–162.

Illera, V., Garrido, F., Serrano, S., García-González, M.T., 2004. Immobilization of the heavy metals

Cd, Cu and Pb in an acid soil amended with gypsum- and lime-rich industrial by-products.

European Journal of Soil Science 55, 135–145.

Jann, S., Totsche, K. U., Jaesche, P., Kögel-Knabner, I., 2002. Release and transport of colloids from

the unsaturated zone under natural conditions: a lysimeter study. In Proceedings of colloids

and colloid-facilitated transport of contaminants in soils and sediments. International

Workshop. Foloum, Denmark. September 19–20, 2002.

Jansen, B., Nierop, K. G. J., Verstraten, J. M., 2002. Influence of pH and metal/carbon ratios on

soluble organic complexation of Fe(II), Fe(III) and Al(III) in soil solutions determined by

diffusive gradients in thin films. Analytica Chimica Acta 454, 259–270.

References

115

Jansen, B., Nierop, K. G. J., Verstraten, J. M., 2005. Mechanisms controlling the mobility of dissolved

organic matter, aluminium and iron in podzol P horizons. European Journal of Soil Science

56, 537–550.

Jardine, P. M., Weber, N. L., McCarthy, J. F., 1989. Mechanisms of dissolved organic carbon

adsorption on soil. Soil Science Society of America Journal, 53, 1378–1385.

Jensen, D. L., Ledin, A., Christensen, T. H., 1999. Speciation of heavy metals in landfill-leachate

polluted groundwater. Water Research 33, 2642–2650.

Johnson, C. A., H. Moench, Wersin, P., Kugler, P., Wenger, C., 2005. Solubitlity of antimony and

other elements in samples taken from shooting ranges. Journal of Environmental Quality 32,

248–254.

Jones, C. A., Inskeep, W. P., Neumann, D. R., 1997. Arsenic transport in contaminated mine tailings

following liming. Journal of Environmental Quality 26, 433–439.

Jones, R. I., Shaw, P. J, De Haan, H., 1993. Effects of dissolved humic substances on the speciation

o firon and phosphate at different pH and ionic strength. Environmental Science and

Technology, 27, 1052–1059.

Kabata-Pendias, A., Pendias, H., 2001. Trace Elements in Soils and Plants. CRC Press, Boca Raton.

Kaiser, K., Kaupenjohann, M., Zech, W., 2001. Sorption of dissolved organic carbon in soils: effects of

soil sample storage, soil-to-solution ratio, and temperature. Geoderma. 99, 317–328.

Kalbitz, K., Solinger, S., Park, J.H., Michalzik, B., Matzner, E., 2000. Controls on the dynamics of

dissolved organic matter in soils: A review. Soil Science 165, 277–304.

Kaplan, D. I., Bertsch, P. M., Adriano, D. C., 1997. Mineralogical and physicochemical differences

between mobile and nonmobile colloidal phases in reconstructed pedons. Soil Science

Society of America Journal 61, 641–469.

Kaplan, D. I., Sumner, M. E., Bertsch, P. M., Adriano, D. C., 1996. Chemical conditions conducive to

the release of mobile colloids from Ulitsol profiles. Soil Science Society of America Journal

60, 269–274.

Kaplan, D.I., Bertsch, P.M. Adriano, D.C., Miller, W.P., 1993. Soil-borne mobile colloids as influenced

by water flow and organic carbon. Environmental Science and Technology 27, 1193–1200.

Keller, C., Domergue, F.-L., 1996. Soluble and particulate transfers of Cu, Cd, Al, Fe and some major

elements in gravitational waters of a Podzol. Geoderma 71, 263–274.

King, R. B. 1995. Antimony. p. 174–176. In: King, R. B. (ed.), Inorganic Chemistry. Encyclopedia of

inorganic chemistry, Volume 1, Wiley, Chichester.

Kjærgaard, C., de Jonge, L. W., Moldrup, P., Schjønning, P., 2004a. Water-dispersible colloids:

Effects of measurement method, clay content, initial soil matric potential, and wetting rate.

Vadose Zone Journal 3, 403–412.

References

116

Kjærgaard, C., Moldrup, P., de Jonge, L. W., Jacobsen, O. H., 2004b. Colloid mobilization and

transport in undisturbed soil columns. II. The role of colloid dispersibility and preferential flow.

Vadose Zone Journal 3, 423–433.

Kleber, M., Sollins, P., Sutton, R., 2007. A conceptual model of organo-mineral interactions in soils:

self-assembly of organic molecular fragments into zonal structures on mineral surfaces,

Biogeochemistry 85, 9–24.

Klitzke, S., Kirby, J., Lombi, E., Hamon, R., Lang, F. Lead, antimony and arsenic in dissolved and

colloidal fractions from an amended shooting range soil as characterized by multi-stage

tangential ultrafiltration and centrifugation. In prep.

Klitzke, S., Lang, F. Mobilization of soluble and dispersible Pb, As, and Sb in a polluted, organic-rich

soil – Effects of pH increase and counterion valency. Journal for Environmental Quality,

submitted.

Klitzke, S., Lang, F., 2007. A method for the determination of hydrophobicity of suspended soil

colloids. Colloids and Surfaces A: Physicochemical and Engineering Aspects 303, 249–252.

Klitzke, S., Lang, F., 2007. Hydrophobicity of Soil Colloids and Heavy Metal Mobilization - Effects of

Drying. Journal of Environmental Quality. 36, 118–1193.

Klitzke, S., Lang, F., Kaupenjohann, M. Increasing pH releases colloidal lead in a highly contaminated

forest soil. European Journal of Soil Science, accepted.

Knechtenhofer, L. A., Xifra. I. O., Scheinost, A. C., Flühler, H., Kretzschmar, R., 2003. Fate of heavy

metals in a strongly acidic shooting-range soil: small-scale metal distribution and its relation to

preferential water flow. Journal of Plant Nutrition and Soil Science 166, 84–92.

Kretzschmar, R., Borkovec, M., Grolimund, D., Elimelech, M., 1999. Mobile subsurface colloids and

their role in contaminant transport, Advances in Agronomy 66, 121–181.

Kretzschmar, R., Hesterberg, D., Sticher, H., 1997. Effects of adsorbed humic acid on surface charge

and flocculation of kaolinite. Soil Science Society of America Journal 61, 101–108.

Kretzschmar, R., Robarge, W. P., Weed, S. B., 1993. Flocculation of kaolinitic soil clays: Effects of

humic substances and iron oxides. Soil Science Society of America Journal 57, 1277–1283.

Kretzschmar, R., Sticher, H., 1997. Transport of humic-coated iron oxide colloids in a sandy soil:

Influence of Ca

and trace metals. Environmental Science and Technology 31, 3497–3504.

Lang, F., Egger, H., Kaupenjohann, M., 2005. Size and shape of lead-organic associations. Colloids

and Surfaces A: Physicochemical and Engineering Aspects. 265, 95–103.

Le Bissonnais, Y. 1996. Aggregate stability and assessment of soil crustability and erodbility: I.

Theory and methodology. European Journal of Soil Science 47, 425–437.

Letey, J., 1969. Measurement of contact angle, water drop penetration time, and critical surface

tension. Proceedings of the Symposium on Water-Repellent Soils, University of California,

43–47.

Lindsay, W. L. 1979. Chemical equilibria in soils. John Wiley & Sons, New York.

References

117

Lintschinger, J., Michalke, B., Schulte-Hostede, S., Schramel, P., 1998. Studies on speciation of

antimony in soil contaminated by industrial activity. International Journal of Environmental

Analytical Chemistry 72, 11–25.

Liu, Z., Lee, C., 2006 Drying effects on sorption capacity of coastal sediment: The importance of

architecture and polarity of organic matter, Geochimica and Cosmochimica Acta 70, 3313–

3324.

Lubal, P., Siroky, D., Fetsch, D., Havel, J., 1998. The acidobasic and complexation properties of

humic acids Study of complexation of Czech humic acids with metal ions. Talanta 47, 401–

412.

Ma’shum, M., Farmer, V. C., 1985. Origin and assessment of water repellency of a sandy South

Australian soil, Australian Journal of Soil Research 23, 623–626.

McBride, M. B., 1994. Environmental Chemistry of Soils. Oxford University Press, New York.

McHale, G., Newton, M. I., Shirtcliffe, N. J., 2005. Water-repellent soil and its relationship to

granularity, surface roughness and hydrophobicity: a materials science view, European

Journal of Soil Science 56, 445–452.

McKay, D. L., Gillham, R. W., Cherry, J. A. 1993. Field experiments in a fractured clay till. 2. Solute

and colloid transport. Water Resources Research 29, 3879–3890.

Mergenthaler, B., Richner, T. 2002. Mobilität und geochemisches Verhalten von Antimon im Boden

von Schießanlagen. http://e-collection.ethbib.ethz.ch/show?type=dipl&nr=47 (verfied 17th

October 2007).

Mikutta, C., Lang, F., Kaupenjohann, M., 2006. Kinetics of phosphate sorption to polygalacturonate-

coated goethite, Soil Science Society of America Journal 70, 541–549.

Miller, W. P., Baharuddin, M. K., 1986. Relationship of soil dispersibliity to infiltration and erosion of

southeastern soils. Soil Science 142, 235–240.

Mukhopadhyay D., Sanyal S. K., 2004. Complexation and release isotherm of arsenic in arsenic-

humic/fulvic equilibrium study. Australian Journal of Soil Research 42, 815–824.

Münch, J.-M. , Totsche, K. U., Kaiser, K., 2002. Physicochemical factors controlling the release of

dissolved organic carbon from columns of forest subsoils. European Journal of Soil Science

53, 311–320.

Nätscher, L., Schwertmann, U., 1991. Proton buffering in organic horizons of acid forest soils.

Geoderma 48, 93–106.

Neite, H., Bannick, C. G, Bertram, H.-U., Eckert, P., Friedmann, L., Hülsenbusch, M., König, W.,

Nöltner, T., 1999. Bodenbelastung auf Wurfscheiben-Schießständen. Bodenschutz 2, 53–56.

Nifant’eva, T., Burba, P, Fedorova, O., Shkinev, V. M., Spivakov, B.Y., 2001. Ultrafiltration and

determination of Zn- and Cu-humic substances complexes stability constants. Talanta 53,

1127–1131.

References

118

Peltovuori, T., Soinne, H., 2005. Phosphorous solubility and sorption in frozen, air-dried and field-

moist soil. European Journal of Soil Science 56, 821–826.

Ranville, J. F., Chittleborough, D. J., Beckett, R., 2005. Particle-size and element distribution of soil

colloids: implications for colloid transport. Soil Science Society of America Journal 69, 1176–

184.

Ranville, J. F., Chittleborough, D. J., Shanks, F., Morrison, R. J. S, Harris, T., Doss, F., Beckett, R.,

1999. Development of sedimentation field-flow fractionation-inductively coupled plasma mass-

spectrometry for the characterization of environmental colloids. Analytica Chimca Acta 381,

315–329.

Raveh, A., Avnimelech Y., 1978. The effect of drying on the colloidal properties and stability of humic

compounds. Plant and Soil 50, 545–552.

Redman, A. D., Macaldy, D. L., Ahmann, D., 2002. Natural organic matter affects arsenic speciation

and sorption onto hematite. Environmental Science and Technology 36, 2889–2896.

Roy, J. L., McGill, W. B., 2002. Assessing soil water repellency using the molarity of ethanol droplet

(MED) test, Soil Science 167, 83–97.

Saada, A., Breeze, D., Crouzet, C., Cornu, S., Baranger, P., 2003. Adsorption of arsenic(V) on

kaolinite and on kaolinite-humic acid complexes. Role of humic acid nitrogen groups.

Chemosphere 51, 757–763.

Saar, R., Weber, J., 1980. Lead(II) complextaion by fulvic acid: how it differs from fulvic acid

complexation of copper(II) and cadmium(II). Geochimica et Cosmochimica Acta, 44, 1381–

1384.

Sadiq, M., 1997. Arsenic chemistry in soils: An overview of thermodynamic predictions and field

observations. Water, Air, and Soil Pollution 93, 117–136.

Sauvé, S., McBride, M., Hendershot, W., 1998. Soil solution speciation of lead(II): Effects of organic

matter and pH. Soil Science Society of America Journal 62, 618–621.

Schäfer, A. I., Schwicker, U., Fischer, M. M., Fane, A. G., Waite, T. D., 2000. Microfiltration of colloids

and natural organic matter. Journal of Membrane Science 171, 151–172.

Schaumann, G. E., Siewert, C., Marschner, B., 2000. Kinetics of the release of dissolved organic

matter (DOM) from air-dried and pre-moistend soil material. Journal of Plant Nutrition and Soil

Science 163, 1–5.

Scheffer, F., Schachtschabel, P., 1998. Lehrbuch der Bodenkunde. 14th ed. Ferdinand Enke Verlag,

Stuttgart.

Schwertmann, U., Cornell, R. M., 1991 Iron oxides in the laboratory, VCH Weinheim.

Seta, A. K., Karathanasis, A. D., 1997. Stability and transportability of water-dispersible soil colloids.

Soil Science Society of America Journal 61, 604–611.

Smith, E., Naidu, R., Alston, A. M., 1998. Arsenic in the soil environment: A review. Advances in

Agronomy 64, 149-195.

References

119

Spuller, C., Weigand, H., Marb, C., 2007. Trace metal stabilisation in a shooting range soil: Mobility

and phytotoxicity. Journal of Hazardous Materials, 141, 378–387.

Swash P.M., Monhemius A.J., 1985. The disposal of arsenical wastes: technologies and

environmental considerations. International Minerals and Metals Technology, 121–125.

Tack, F. M., Singh, S. P., Verloo, M. G., 1999. Leaching behaviour of Cd, Cu, Pb and Zn in surface

soils derived from dredged sediments. Environmental Pollution, 106 107–114.

Tanner, C. B., Jackson, M. L., 1947. Nomographs of sedimentation times for soil particles under

gravity or centrifugal acceleration. Soil Science Society Proceedings, 60–65.

Thanabalasingam P., Pickering, W.F., 1986. Arsenic Sorption by Humic Acids. Environmental

Pollution (Series B) 12, 233–246.

Thill, A., Spalla, O., 2003. Aggregation due to capillary forces during drying of particle submonolayers.

Colloids and Surfaces A: Physicochemical and Engineering Aspects 217, 143–151.

Tighe M., Lockwood, P., Wilson, S., 2005. Adsorption of antimony(v) by floodplain soils, amorphous

iron(III) hydroxide and humic acid. Journal of Environmental Monitoring 7, 1177–1185.

Tiller, C. L., O’Melia, C. R., 1993. Natural organic matter and colloidal stability: models and

measurements. Colloids Surfaces A: Physicochemical and Engineering Aspects 73, 89–102.

Tipping, E., Cooke, D., 1982. The effects of adsorbed humic substances on the surface charge of

goethite (α-FeOOH) in freshwaters. Geochimica et Cosmochimica Acta 46, 75–80.

Tipping, E., Higgins, D. C., 1982. The effect of adsorbed humic substances on the colloid stability of

haematite particles. Colloids and Surfaces 5, 85–92.

Tom-Petersen, A., Hansen, H. C. B., Nybroe. O., 2004. Time and moisture effects on total

bioavailable copper in soil water extracts. Journal of Environmental Quality 33, 505–512.

Totsche, K. U., Jann, S., Kögel-Knabner, I., 2006. Release of polycyclic aromatic hydrocarbons,

dissolved organic carbon, and suspended matter from disturbed NAPL-contaminated gravelly

soil material, Vadose Zone Journal 5, 469–479.

Turpeinen R., Salminen J., Kairesalo T., 2000. Mobility and bioavailability of lead in contaminated

boreal forest soil. Environmental Science and Technology, 34, 5152–5156.

Tyler, G., Olsson, T., 2001. Concentrations of 60 elements in the soil solution as related to the soil

acidity. European Journal of Soil Science 52, 151–165.

USEPA. 2005. Best Management Practices for Lead at Outdoor Shooting Ranges. USEPA.

Washington DC. Available at http://www.epa.gov/region2/waste/leadshot/epa_bmp.pdf

(verified 8 May 2007).

van Oss, C. J., Giese, R. F., 1995. The hydrophilicity and hydrophobicity of clay minerals. Clay and

Clay Minerals 43, 474–477.

Voigt, D. E., Brantley, S.L., Hennet, R.J.C., 1996. Chemical fixation of arsenic in contaminated soils.

Applied Geochemistry 11, 633–643.

References

120

Wan, J., Wilson, J. L., 1994a. Visualization of the role of the gas-water interface on the fate and

transport of colloids in porous media. Water Resources Research 30, 11–23.

Wan, J., Wilson J. L., 1994b. Colloid transport in unsaturated porous media. Water Resources

Research 30, 587–864.

Wang, E.X., Benoit, G., 1996. Mechanisms controlling the mobility of lead in the Spodosols of a

Northern Hardwood Forest Ecosystem. Environmental Science and Technology 30, 2211–

2219.

Wang, Z., Shan, X-Q., Zhang, S., 2002. Comparison between fractionation and bioavailability of trace

elements in rhizosphere and bulk soils. Chemosphere. 46, 1163–1171.

Warwick, P., Inam, E., Evans, N. 2005. Arsenic's interaction with humic acid, Environmental

Chemistry 2, 119–124.

World Reference Base for Soil Resources, 1998. Food and Agriculture Organization of the United

Nations. Rome.

Wu, Q., Riise, G., Kretzschmar, R., 2003. Size distribution of organic matter and associated

propiconazole in agricultural runoff material. Journal of Environmental Quality 32, 2200–2206.

121

Acknowledgements

During the time I was working on this dissertation I was supported, directly or

indirectly, by several people and institutions. I am thankful to

Prof. Dr. Martin Kaupenjohann for his confidence in my work, guidance,

continuous encouragement and freedom throughout my time in the department,

Dr. Friederike Lang for supervision, support, confidence and for many valuable

and inspiring discussions,

Prof. Dr. Rainer Schulin for being the co-examiner,

Nadine Kurowski for valuable insights into complex chemistry, Prof. Dr. Gabriele

Schaumann and Prof. Dr. Wolfgang Wilcke for helpful discussions,

Andrea Herre for indispensable advice on running the ICP-MS and for fruitful

discussions,

Claudia Kuntz for her help with analyses whilst I was working on the

INTERURBAN project and supportive advice throughout my time in the

department. Her motivation and enthusiasm have tided me over many ups and

down during numerous setbacks of the C18-method development.

Sabine Rautenberg for C/N-analyses, Kotan Yildiz for help and support with flame

AA and graphite-furnance AA measurements, Geerd Smidt for his help with

analyses, Ulrich Gernert from ZELMI for SEM-measurements, and Petra Marsiske

for X-ray fluorescence analyses,

Dr. Reimo Kindler, Dr. Katrin Ilg and Dr. Christian Mikutta for the friendly working

atmosphere in our common office and/or for valuable discussions and many

chances to bounce off ideas,

The staff of the soil chemistry and soil physics group at Berlin University of

Technology for the friendly, open and supportive atmosphere in the department,

Acknowledgements

122

Prof. Dr. Mike McLaughlin for giving me the chance to do part of my thesis within

his group at CSIRO Land and Water, Adelaide, Australia, and Dr. Enzo Lombi for

supervision,

Dr. Jason Kirby for all his help, advice, encouragement, patience and support on

setting up the at times troublesome and very temperamental MTUF instrument,

John Gouzos, Sharyn Zrna, Caroline Johnston for help with analyses in the

analytical lab of CSIRO,

all my former colleges from CSIRO Land and Water for the warm welcome and

the friendly atmosphere in the lab. Their positive, optimistic and laid-back outlook

on the ups and down in the life of a scientist have certainly contributed to keeping

up my motivation after repeated breakdowns of the MTUF instrument.

my parents and my sister Maren for support and encouragement throughout my

entire studies in environmental science and my thesis time,

David Meredith for his friendship, support and for proof-reading several

manuscripts,

my friends for their good company, the sometimes necessary distraction from

work and for providing a shoulder to lean on,

the university rowing group “TU-Rudertreff” for many enjoyable and relaxing

rowing trips on the Berlin waterways which provided a great opportunity to

“recharge my empty batteries” after occasional periods of exhausting lab

experiments.

The funding provided by the German Research Foundation (DFG) for the

INTERURBAN project (La 1398/2) and the shooting range site project (La 1398/3)

as well as by the German Academic Exchange Service (DAAD) is acknowledged.

123

Curriculum vitae

1974 born on February 27

in Schorndorf (Germany)

1993 Graduation (Abitur) at Maria-Merian-Schule, Waiblingen (Germany)

1993-1994 Au-Pair in Great Britain

1994-1996 Technical College “Chemisches Institut Dr. Flad”, Stuttgart

1996: “Staatsexamen”, Chemical-technical assistant

1996-1997 Internships in France (CEPHAC, Poitiers) and Great Britain

(DeMontfort University, Leicester)

1997-2002 Study of Environmental Sciences at the Universities of Lüneburg, Trier

and Tübingen

2002 Master of Science in Applied Environmental Geoscience, University of

Tübingen

2003-2007 PhD student at the Department of Soil Science, Berlin University of

Technology

2005 Research stay (7 months), CSIRO Land and Water, Adelaide, Australia

- Size-fractionation of colloids by means of multi-stage tangential ultrafiltration

- Bioavailability of heavy metals as indicated by diffusive gradients in thin films

A-1

Appendix

-2

Chapter 2: A method for the determination of hydrophobicity of suspended

soil colloids

Appendix 2.1: Hydrophobicity of differently treated goethite - DCM-method

P-treated goethite

Zetapotential

[mV]

1.7 8.0 96.6% 96.2% 96.2%

4.0 0.4 96.3% 96.2% 96.6%

4.2 -2.8 95.5% 93.6% 95.1%

4.4 -11.0 96.4% 96.6% 97.2%

4.6 -4.9 71.6% 88.5% n.a.

4.8 -4.0 81.8% 75.5% n.a.

5.0 -21.1 27.9% 32.6% 32.1%

5.6 -26.3 6.7% 3.1% 16.1%

6.5 -29.6 9.0% 9.1% 9.1%

7.0 -33.9 6.9% 6.9% 6.9%

7.5 -34.5 10.8% 10.8% 10.8%

8.0 -33.5 29.3% 29.3% 29.4%

Al-treated goethite

2.0 37.5 52.0% 51.4% 55.0%

3.0 33.1 43.5% 43.9% 52.3%

4.0 35.3 42.1% 42.0% 41.8%

5.0 43.0 9.2% 10.0% 10.1%

6.1 34.0 26.7% 26.7% 26.3%

7.0 29.4 34.4% 34.3% 34.2%

7.5 25.7 67.1% 67.2% 67.2%

8.0 24.6 79.3% 79.3% 79.1%

Mg-treated goethite

2.0 43.8 16.0% 17.1% 16.9%

3.0 37.1 4.8% 5.9% 5.5%

4.0 32.3 16.3% 16.6% 16.5%

5.0 30.3 29.9% 30.2% 29.8%

6.0 26.3 31.7% 31.6% 31.1%

7.0 23.6 85.4% 85.0% 85.1%

8.0 8.0 38.0% 37.8% 37.4%

Na-treated goethite

2.2 37.0 17.4% 17.6% 17.7%

3.2 35.7 13.0% 13.8% 14.8%

4.1 34.5 4.3% 5.1% 5.1%

5.1 38.1 12.6% 12.6% 12.6%

6.1 26.9 17.2% 17.1% 17.3%

6.9 25.3 28.1% 28.2% 28.4%

8.1 5.1 58.8% 58.6% 58.8%

n. a.: data not available

Hydrophobicity

Appendix

-3

Appendix 2.2: Hydrophobicity of Al- and P-treated goethite - C18-method

P-treated goethite

Zetapotential

[mV]

3.8 1.8 55.2% 51.8% 55.5%

4.0 -7.5 40.8% 36.9% 39.3%

4.4 -8.2 31.3% 35.1% 38.3%

4.6 -13.3 2.3% 0.0% 2.0%

5.1 -19.7 4.9% 0.0% 0.0%

5.2 -23 4.7% 5.8% 5.0%

7.1 -33.5 0.9% 0.0% 2.0%

Al-treated goethite

5.0 41.8 7.1% 11.1% 7.1%

5.8 37.6 6.6% 7.5% 11.0%

6.4 35.7 4.0% 6.2% 10.7%

7.0 30.2 15.2% 13.4% 13.0%

7.0 30.8 29.1% n.a n.a

7.2 27.8 51.5% 51.9% 51.9%

Appendix 2.3: Capacity test of C18 microparticles

C18 [mg]

10 61.1% 62.7% 60.7%

20 69.5% 61.0% 68.3%

30 71.3% 70.0% 63.8%

40 51.3% 71.0% 86.8%

50 77.3% 88.9% 85.0%

60 84.4% 85.4% 89.1%

70 83.4% 82.0% 82.4%

80 82.7% 88.6% 86.1%

100 90.3% 89.8% 92.9%

120 82.2% 68.4% 79.1%

Appendix 2.4: Result of the comparison between goethite particles sorbed

on DOC-treated and untreated C18 microparticles

Treatment

Untreated 52.6% 48.9% n.a.

DOC-20 50.4% 49.3% 59.7%

DOC-100 59.3% 65.7% 61.9%

n. a.: data not available

Hydrophobicity

Appendix

-4

Chapter 3: Hydrophobicity of soil colloids and heavy metal mobilization:

Effects of drying

Appendix 3.1: Properties of soil suspensions and soil solid phase

Soil Replicate pH EC Turbidity T1 Turbidity T2

Zeta potential

Particle size Contact angle

[µS m

-1

][FNU] [FNU] [mV] [nm] [°]

a 5.6 51.1 22.5 21.8 -18.0 550 84.4

B1, moist b 5.6 51.9 23.3 21.8 -17.9 300 85.4

c 5.6 51.1 21.6 20.1 -17.8 380 85.9

a 6.1 54.7 18.1 15.4 -17.0 290 116.4

B1, dried b 6.0 57.1 14.6 14.5 -15.9 300 116.8

c 5.9 55.5 9.64 9.70 -16.9 240 120.5

a 4.6 85.1 5.76 5.58 -16.3 280 83.7

B2, moist b 4.6 88.0 6.89 5.97 -15.9 600 83.9

c 4.6 83.7 6.34 6.13 -15.9 300 83.6

a 5.1 86.5 18.1 16.6 -16.2 900 120.4

B2, dried b 4.9 87.0 7.96 7.73 -15.8 300 122.6

c 5.0 84.1 7.50 n.a -16.4 300 120.7

a 5.5 93.4 10.5 10.2 -16.3 600 85.9

B3, moist b 5.4 97.6 9.49 8.81 -17.2 400 84.2

c 5.5 118.7 7.71 7.78 -14.6 700 85.5

a 6.0 93.1 8.23 7.20 -15.9 n.a. 121.2

B3, dried b 5.6 97.8 8.01 n.a. -16.5 290 121.7

c 5.6 91.4 6.44 n.a. -15.8 n.a. 127.5

a 4.2 768 0.98 n.a. -5.9 n.a. 88.8

TG, moist b 4.0 760 1.36 n.a. -1.6 n.a. 89.3

c 4.0 753 1.21 n.a. -2.9 n.a. 89.2

a 4.1 773 n.a. n.a. -5.9 n.a. 89.9

TG, dried b 4.2 722 0.47 n.a. -4.4 550 90.0

c 4.2 718 1.96 1.78 -7.2 n.a. 90.0

a 3.7 151.9 12.6 8.78 -17.4 n.a. 87.9

GS, moist b 3.7 156.1 4.80 4.80 -20.4 300 88.2

c 3.7 157.3 2.68 2.51 -18.1 320 87.7

a 3.6 177.9 14.5 14.7 -16.4 800 110.8

GS, dried b 3.6 178.4 13.2 13.1 -17.5 350 111.2

c 3.5 175.2 15.5 15.0 -17.8 300 107.5

n. a.: data not available

FNU: Formazin nephlometric unit

Appendix

-5

Appendix 3.2a: Dissolved and colloidal concentrations in soil suspensions

of field-moist and air-dried samples

Soil Replicate DOC COC Cd

diss

coll

diss

coll

[mg l

-1

] [mg l

-1

] [µg l

-1

] [µg l

-1

] [mg l

-1

] [mg l

-1

]

a 6.4 4.0 3.4 3.6 0.62 0.17

B1, moist b 5.8 3.8 3.9 4.2 0.61 0.11

c 6.5 3.5 4.1 5.8 0.61 0.17

a 11.1 3.4 6.0 0.8 0.80 0.06

B1, dried b 11.5 3.0 6.2 0.8 0.74 0.07

c 11.6 3.3 6.2 0.5 0.76 0.04

a 9.7 2.6 23.0 10.3 1.29 0.25

B2, moist b 8.5 2.9 22.8 10.3 1.18 0.24

c 8.0 3.1 24.0 6.8 1.25 0.18

a 16.0 4.8 28.8 6.3 1.33 0.13

B2, dried b 15.5 4.1 29.1 5.2 1.29 0.12

c 15.2 9.5 29.1 n.a. 1.31 n.a

a 11.9 1.5 13.8 6.4 0.93 0.12

B3, moist b 9.9 3.5 15.0 8.3 0.99 0.14

c 9.1 2.9 14.5 7.0 0.98 n.a.

a 15.2 4.1 17.4 1.9 0.95 0.04

B3, dried b 16.7 3.1 17.3 3.1 0.98 0.04

c 16.2 2.8 17.5 2.1 0.9 0.0

a 8.7 2.2 17.3 5.8 0.73 0.06

TG, moist b 10.3 0.0 13.8 5.1 0.64 0.05

c 7.8 2.1 13.0 4.7 0.58 0.05

a 17.1 6.1 14.8 1.5 0.55 0.04

TG, dried b 17.6 4.8 16.5 1.1 0.61 0.02

c 17.3 4.9 15.6 2.4 0.61 0.03

a 19.6 5.8 1.3 1.4 below QL below QL

GS, moist b 19.1 5.0 3.1 0.9 below QL below QL

c 20.4 3.9 2.0 0.1 below QL below QL

a 38.2 7.2 0.7 0.5 below QL below QL

GS, dried b 37.3 8.6 0.7 0.4 below QL below QL

c 39.7 7.8 0.7 0.3 below QL below QL

DOC: dissolved organic carbon

COC: colloidal organic carbon

n. a.: data not available

QL: quantification limit

DL: detection limit

Appendix

-6

Appendix 3.2b: Dissolved and colloidal concentrations in soil suspensions

of field-moist and air-dried samples

Soil Replicate Cu

diss

coll

diss

coll

[µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

]

a 70.9 28.6 2.1 42.3

B1, moist b 72.6 17.9 2.0 40.7

c 73.9 62.6 n.a. n.a.

a 208.3 n.a. 2.1 18.4

B1, dried b 186.1 n.a. 1.7 17.5

c 179.2 n.a. 8.8 18.7

a 162.4 18.2 50.3 63.6

B2, moist b 173.4 0.0 18.5 48.3

c 185.1 0.0 7.6 179.5

a 240.7 n.a. n.a. 25.6

B2, dried b 239.4 39.4 9.6 20.5

c 245.0 65.1 12.2 18.1

a 35.4 125.8 1.5 24.1

B3, moist b 70.4 266.3 0.0 29.3

c 131.8 45.5 0.2 22.3

a n.a. n.a. 0.0 18.0

B3, dried b 226.6 45.0 1.8 12.0

c 219.9 33.4 0.7 11.1

a n.a. n.a 54.7 26.4

TG, moist b 52.0 100.3 58.5 26.4

c 49.0 93.1 60.9 20.5

a 167.7 16.0 n.a. n.a.

TG, dried b 166.2 12.6 84.7 16.5

c n.a. n.a. 73.3 14.9

a 18.4 19.0 1348.8 351.5

GS, moist b 30.2 10.6 1098.0 426.2

c 22.7 n.a. 1421.5 561.6

a 17.0 6.4 2564.6 662.6

GS, dried b 14.3 9.6 2675.6 537.2

c 14.0 4.5 2756.4 471.9

n. a.: data not available

Appendix

-7

Chapter 4: Increasing pH releases colloidal lead in a highly contaminated

forest soil

Appendix 4.1: Total mobilized Pb concentration as a function of shaking

time (soil 1; pH of soil suspension: 3.9)

Time [h] a b c

2 0.8 1.1 0.8

4 1.3 1.5 1.3

8 1.4 1.6 1.7

24 1.5 1.5 1.6

Appendix 4.2: Concentrations and properties of the soil supsensions (soil 1)

pH Replicate Pb

diss

coll

DOC COC

[mg l

-1

] [mg l

-1

] [mg l

-1

] [mg l

-1

]

a 23.2 0.0 14.8 1.9

pH 3 b 22.4 0.0 14.7 0.9

c 22.6 0.8 13.0 2.6

a 2.0 1.7 32.2 24.0

pH 4 b 2.5 1.4 31.9 24.8

c 2.0 1.6 31.8 25.1

a 11.8 5.5 223.8 n.a.

pH 5 b 11.4 5.8 226.7 72.3

c 11.2 5.7 224.6 69.6

a 23.4 15.7 551.1 236.2

pH 6 b 24.5 16.7 545.7 296.8

c 29.6 15.5 640.5 305.2

a 51.0 58.5 1278.8 743.0

pH 7 b 41.2 61.3 1212.8 650.0

c 50.5 53.5 1208.4 700.9

pH Replicate Optical Density Zeta potential Particle size Fe

coll

[-] [mV] [nm] [mg l

-1

]

a 0.000 -16.6 400 -

pH 3 b 0.000 -12.5 500 -

c 0.000 -14.0 n.a. -

a 0.062 -17.8 260 -

pH 4 b 0.065 -18.2 280 -

c 0.050 -20.3 250 -

a 0.342 -23.9 1025 0.23

pH 5 b 0.342 -23.0 600 0.52

c 0.314 -23.5 350 0.15

a 0.754 -28.5 n.a. 0.36

pH 6 b 0.778 -28.9 300 n.a.

c 0.885 -26.6 300 0.42

a 1.826 n.a. 1000 n.a.

pH 7 b 1.798 -36.9 1050 n.a.

c 1.932 -32.7 600 3.1

* concentration of colloidal residues, dissolved in 20 mL concentrated nitric acid

DOC: dissolved organic carbon; COC: colloidal organic carbon

n. a.: data not available

Determination of the soil water content (105°C; soil 1):

Water loss

54.70%7.7224

Mobilized Pb concentrations [mg l

-1

]

Weight soil moist [g] Weight soil dried [g]

4.226

Appendix

-8

Appendix 4.3: Concentrations and properties of the soil supsensions (soil 2)

pH Replicate Pb

diss

coll

DOC COC

KOH [mg l

-1

] [mg l

-1

] [mg l

-1

] [mg l

-1

]

a 9.0 0.5 311.2 5.8

pH 3.3 b n.a. n.a. n.a. n.a.

c n.a. n.a. n.a. n.a.

a 11.5 20.5 504.0 236.4

pH 4.8 b 11.5 18.0 494.8 229.0

c n.a. n.a. n.a. n.a.

a 52.0 45.4 2422.0 337.0

pH 6.3 b 51.8 35.3 2539.0 311.0

c n.a. n.a. n.a. n.a.

Ca(OH)

a 11.9 0.1 20.0 1.8

pH 3.9 b 11.8 0.0 19.3 2.1

c 11.7 0.0 20.0 1.2

a 1.3 0.0 49.9 2.2

pH 5.2 b 1.2 0.1 48.9 0.0

c 1.2 0.1 54.3 0.0

a 0.9 0.0 71.4 0.0

pH 5.8 b 0.7 0.1 72.9 0.0

c 0.7 0.2 74.1 0.0

a 1.7 0.9 173.9 7.4

pH 6.9 b 1.7 0.9 162.8 42.5

c 1.7 0.4 166.1 36.1

KOH Replicate Turbidity Zeta potential Particle size

[FNU] [mV] [nm]

a 29.5 -11.3 963

pH 3.3 b n.a. n.a. n.a.

c n.a. n.a. n.a.

a 331.3 n.a. 464

pH 4.8 b 73.0 -16.1 n.a.

c n.a. n.a. n.a.

a n.a. n.a. n.a.

pH 6.3 b n.a. n.a. n.a.

c n.a. n.a. n.a.

Ca(OH)

Replicate Turbidity Zeta potential Particle size

[FNU] [mV] [nm]

a 0.44 -3.7 1459

pH 3.9 b 0.63 -3.4 n.a.

c 0.30 (-5.4) n.a.

a 0.20 -3.5 1879

pH 5.2 b 0.43 -3.3 1132

c 0.68 -3.2 1458

a 0.81 -1.6 820

pH 5.8 b 0.20 (1.1) 1132

c n.a. -5.0 879

a 8.47 -15.4 687

pH 6.9 b n.a. -13.3 1879

c 8.30 -13.3 318

Determination of the soil water content (105°C; soil 2):

Water loss

39.5%

Weight soil moist [g] Weight soil dried [g]

1.90 1.15

Appendix

-9

Chapter 5: Lead, Antimony and Arsenic in dissolved and colloidal fractions

from an amended shooting range soil as characterized by multi-stage

tangential ultrafiltration and centrifugation

Appendix 5.1:

Sample number time volume EC

Data breakthroughcurve

[min] [ml] [mS cm

-1

]

1 0 5.193 64.83

Initial Solution: 2 30 4.636 56.38

0.5 M CaCl

; EC: 66.03 mS cm

-1

3 60 3.277 43.94

Rinsing solution: 4 90 3.280 35.44

0.01 M CaCl

; EC: 2.15 mS cm

-1

5 120 3.267 27.42

6 150 3.211 23.340

7 180 3.203 18.340

flow rate: 0.1 mL min

-1

8 210 3.187 15.260

tangential flow: 20 mL min

-1

9 240 3.216 12.510

10 270 3.101 10.230

11 300 3.148 7.880

12 330 3.167 5.578

13 360 3.178 4.483

14 390 3.185 3.488

15 420 3.110 2.744

16 450 3.184 2.451

17 480 3.154 2.233

18 510 3.240 2.083

19 540 3.271 2.130

20 570 3.208 1.858

21 600 3.073 2.135

22 630 3.133 2.101

23 660 3.115 2.035

24 690 3.134 2.122

25 720 3.195 2.112

26 750 3.168 2.171

27 780 3.071 2.075

28 810 3.075 2.081

29 840 2.890 2.100

30 870 3.060 2.026

31 900 3.060 2.101

32 930 2.979 2.093

33 960 3.143 2.081

34 990 3.098 2.095

35 1020 3.096 2.085

36 1050 3.189 2.070

37 1080 3.192 2.067

38 1110 3.193 2.073

39 1140 3.335 2.051

40 1170 3.284 2.060

41 1200 3.300 1.963

42 1230 3.324 1.963

43 1260 3.336 1.980

44 1290 2.908 1.953

45 1320 3.018 1.988

46 1350 3.106 1.978

47 1380 3.083 1.974

48 1410 3.156 1.965

Appendix

-10

Appendix 5.2: Concentrations in the inidivual MTUF chambers

Blank Volume pH EC Mn Fe Al As Sb Pb Pb

corr

[ml] [µS cm

-1

] mg l

-1

Initial solution 24.08 4.7 1430 744 159 n.d. 33 161 3172 n.d. 69.1

Chambers:

1.2 µm - 0.45 µm 2.48 4.2 1555 766 103 n.d. 30 160 2811 n.d. 106.5

0.45 µm - 0.22 µm 2.42 4.1 1476 727 182 n.d. 30 151 2636 n.d. 100.5

0.22 µm - 0.1 µm 2.56 4.1 1466 769 174 n.d. 32 157 2625 n.d. 111.3

0.1 µm - 100 kDa 2.48 4.1 1477 780 135 n.d. 33 160 2708 n.d. 112.1

< 100 kDa 14.14 4.1 1473 714 126 n.d. 27 130 2675 n.d. 71.0

KOH - A Volume pH EC Mn Fe Al As Sb Pb Pb

corr

[ml] [µS cm

-1

] mg l

-1

Initial solution 22.29 6.5 1350 169 6788 n.d. 59 329 18470 18470 191.5

Chambers:

1.2 µm - 0.45 µm 2.27 6.4 1455 106 2128 n.d. 22 301 6556 9648 164.5

0.45 µm - 0.22 µm 2.58 6.4 1413 96 1848 n.d. 20 271 6503 19421 180.4

0.22 µm - 0.1 µm 2.74 6.3 1439 313 23288 n.d. 158 429 44873 56433 377.5

0.1 µm - 100 kDa 2.36 6.4 1402 233 16396 n.d. 106 352 31000 31794 367.5

< 100 kDa 12.34 6.5 1446 94 1777 n.d. 20 267 6365 6365 187.2

KOH - B Volume pH EC Mn Fe Al As Sb Pb Pb

corr

[ml] [µS cm

-1

] mg l

-1

Initial solution 23.51 6.6 1311 142 6813 3448 47 208 15646 15646 445.4

Chambers:

1.2 µm - 0.45 µm 2.322 6.6 1252 91 2106 1356 17 176 6894 8408 463.48

0.45 µm - 0.22 µm 2.568 6.7 1240 92 2119 1369 17 174 6931 9452 460.92

0.22 µm - 0.1 µm 2.684 6.6 1387 290 20469 7900 138 283 37563 46817 704.6

0.1 µm - 100 kDa 2.630 6.6 1410 210 13544 5381 87 231 29194 29612 637.28

< 100 kDa 13.306 6.7 1331 94 2004 1346 17 184 6844 6844 453.0

corr

: Pb associated with the filter membrane

OC: organic carbon

n.d.: not determined

Pb concentrations of digested filter membranes

Filter membrane

0.45 µm 703 352

0.22 µm 3173 635

0.1 µm 3018 2484

100 kDa 170 110

The filter membranes were digested in 10 ml concentrated HNO

µg l

-1

µg l

-1

µg l

-1

µg l

-1

Appendix

-11

Ca(OH)

- A Volume pH EC Mn Fe As Sb Pb Pb

corr

[ml] [µS cm

-1

] mg l

-1

Initial solution 21.30 6.3 102 61 502 11 143 2137 2137 43.9

Chambers:

1.2 µm - 0.45 µm 2.28 6.2 104 54 256 10 150 917 3747 45.3

0.45 µm - 0.22 µm 2.65 6.2 100 53 255 9 146 892 2857 37.8

0.22 µm - 0.1 µm 2.21 6.3 101 56 311 10 145 1010 1608 57.5

0.1 µm - 100 kDa 2.30 6.4 97 52 235 9 144 895 1190 59.4

< 100 kDa 11.86 6.4 94 52 246 9 142 924 924 52.7

Ca(OH)

- B Volume pH EC Mn Fe As Sb Pb Pb

corr

[ml] [µS cm

-1

] mg l

-1

Initial solution 23.21 6.4 83 83 559 14 210 2503 2503 49.3

Chambers:

1.2 µm - 0.45 µm 2.45 6.0 92 70 319 11 186 1268 3585 52.5

0.45 µm - 0.22 µm 2.68 6.0 96 64 284 10 179 1277 3510 49.0

0.22 µm - 0.1 µm 2.44 6.0 95 72 624 12 191 1734 2712 60.9

0.1 µm - 100 kDa 2.26 6.0 95 90 521 14 229 1441 1598 66.6

< 100 kDa 13.37 6.1 90 65 299 11 186 1312 1312 46.9

corr

: Pb associated with the filter membrane

OC: organic carbon

Pb concentrations of digested filter membranes

Filter membrane

0.45 µm 64.5 56.83

0.22 µm 49.6 59.94

0.1 µm 13.2 23.89

100 kDa 6.8 3.515

The filter membranes were digested in 10 ml concentrated HNO

Determination of the soil water content (105°C):

Water loss

27.8%

26.6%

Weight soil dried [g]

0.751

0.536

µg l

-1

µg l

-1

µg l

-1

Weight soil moist [g]

0.742

1.023

Appendix

-12

Appendix 5.3: Concentrations in the centrifugates

Blank

a b a b a b a b a b

Total 308 370 54 53 285 291

3.6 3.8

37.5 36.8

< 0.45 µm 299 299 52 53 282 282

3.6 3.8

37.1 37.2

< 0.22 µm 282 410 52 54 285 286

3.7 3.8

37.2 36.3

< 0.1 µm 288 385 53 52 273 278

3.6 3.5

37.8 36.4

< 0.009 µm (100 kDa) 260 283 52 54 274 279

3.5 3.9

36.8 36.4

KOH

a b a b a b a b a b

Total 4000 3400 50 49 706 659 8.4 7.6 138.8 124.4

< 0.45 µm 3900 3400 50 48 654 651 8.4 7.5 130.8 123.0

< 0.22 µm 3700 3400 47 44 647 670 8.4 7.5 130.9 123.1

< 0.1 µm 3600 3300 43 44 700 642 8.4 7.8 130.6 125.9

< 0.009 µm (100 kDa) 2600 2300 40 37 660 629 7.7 7.1 126.6 121.2

Ca(OH)

a b a b a b a b a b

Total 778 799 21 18 396 395 2.7 2.7 86.2 84.5

< 0.45 µm 743 759 18 19 394 405 2.5 2.6 82.3 82.9

< 0.22 µm 714 733 19 18 405 407 2.6 2.5 82.2 82.7

< 0.1 µm 662 669 18 17 400 399 2.5 2.5 81.9 82.8

< 0.009 µm (100 kDa) 590 524 14 16 379 387 2.4 2.4 83.0 81.2

Determination of the soil water content (105°C):

Water loss

KNO

treatment

KOH/Ca(OH)

treatment

Weight soil moist [g] Weight soil dried [g]

0.490 0.376

0.85 0.65

28.5%

23.5%

µg l

-1

µg l

-1

As Sb Pb organic C

mg l-1

µg l-1

Fe As Sb Pb organic C

Fe As Sb Pb

mg l

-1

mg l

-1

organic C

Appendix

-13

Appendix 5.4: Physico-chemical properties of the total solutions and centrifugates

Total

a b a b

Blank

4.6 4.6 1291 1294

KOH

6.6 6.6 658 656

Ca(OH)

6.7 6.7 92.8 94.7

Blank

a b a b a b

Total 0.79 1.18 -14.5 -11.5 303 287

< 0.45 µm 0.59 0.83 -9.4 -8.1 334 319

< 0.22 µm 0.45 0.52 -2.2 -4.2 334 303

< 0.1 µm 0.26 0.34 2.0 1.1 458 319

< 0.009 µm (100 kDa) 0.16 0.15 2.9 0.7 1132 1879

KOH a b a b a b

Total 17.20 8.72 -24.9 -28.9 682 687

< 0.45 µm 8.64 6.82 -25.0 -29.5 272 303

< 0.22 µm 5.64 4.06 -23.2 -23.5 319 708

< 0.1 µm 1.57 1.50 -10.9 -10.5 339 274

< 0.009 µm (100 kDa) 0.33 0.46 3.0 1.2 879 1132

Ca(OH)

a b a b a b

Total 7.17 6.06 -15.9 -15.9 319 319

< 0.45 µm 4.34 5.21 -17.4 -17.1 192 654

< 0.22 µm 2.77 2.53 -13.7 -11.3 192 303

< 0.1 µm 0.88 0.99 -6.9 -2.4 319 319

< 0.009 µm (100 kDa) 0.26 0.26 -2.6 2.2 n.a. 879

n.a.: not available

Zeta potential [mV] Particle size [nm]Turbidity [FNU]

Particle size [nm]Turbidity [FNU]

Turbidity [FNU] Zeta potential [mV] Particle size [nm]

Zeta potential [mV]

pH EC [µS cm

-1

]

Appendix

-14

Chapter 6: Mobilization of colloidal and dissolved Pb, As and Sb in a polluted,

organic-rich soil – effects of pH increase and counterion valency

Appendix 6.1: Properties of soil suspensions

KOH Replicate pH EC Turbidity Zeta potential Particle size

[µS cm

-1

][FNU] [mV] [nm]

a 3.5 3680 1.01 -13.0 980

pH 3.5 b 3.6 3470 2.37 -14.3 879

c 3.5 3600 1.82 -13.3 682

a 4.3 1324 5.82 -20.5 319

pH 4.3 b 4.3 3210 9.65 -19.5 524

c 4.3 3180 5.25 -21.2 319

a 6.2 2190 13.3 -20.0 1083

pH 6.2 b 6.1 2170 9.9 -18.2 625

c 6.3 2170 18.2 -18.4 536

a 7.0 1457 30.16 -34.0 682

pH 7.1 b 7.1 1472 40.00 -34.5 2137

c 7.2 1467 (9.3) -35.2 319

Ca(OH)

Replicate pH EC Turbidity Zeta potential Particle size

[µS cm

-1

][FNU] [mV] [nm]

a 3.6 1277 0.99 -3.1 100

pH 3.7 b 3.6 1264 1.41 -6.2 879

c 3.8 1236 1.85 -3.7 n.a.

a 4.4 842 0.78 (-10.4) 1523

pH 4.4 b 4.4 832 0.93 -3.4 842

c 4.4 829 0.63 -2.2 682

a 5.3 333 3.79 -12.8 446

pH 5.3 b 5.3 327 3.56 -12.1 682

c 5.3 325 (1.75) -10.3 682

a 6.1 53.2 16.7 -18.7 853

pH 6.1 b 6.1 54.9 12.5 -18.0 651

c 6.2 56.8 13.6 -18.9 808

a 6.7 84.5 11.9 -15.7 334

pH 6.7 b 6.7 79.5 16.1 -15.9 319

c 6.6 76.1 11.5 -16.7 319

n. a.: data not available

FNU: Formazin nephlometric unit

Appendix

-15

KOH Replicate Pb

diss

coll

diss

coll

diss

coll

[µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

]

a 600 0.0 29.9 0.0 8.7 0.1

pH 3.5 b 550 0.0 18.0 0.0 4.1 0.0

c 550 200 17.1 0.0 3.0 0.3

a 500 350 19.2 2.1 7.4 0.0

pH 4.3 b 450 300 17.7 3.9 5.3 0.0

c (1100) 0 (32.3) 0.0 (21.3) 0.0

a 4250 650 36.7 2.9 21.9 0.0

pH 6.2 b 2950 400 30.8 1.9 20.9 0.0

c 3700 500 28.2 1.6 12.4 0.0

a 5250 1050 48.9 7.3 25.0 0.0

pH 7.1 b 4800 700 39.9 4.6 17.5 0.3

c 4300 500 38.4 2.6 20.4 0.0

Ca(OH)

Replicate Pb

diss

coll

diss

coll

diss

coll

[µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

] [µg l

-1

]

a 254.7 (59.7) 13.0 0.3 3.0 0.0

pH 3.7 b 297.0 0.0 15.5 0.5 3.4 0.0

c 238.2 1.5 15.2 0.0 2.0 0.0

a 73.5 0.0 16.3 0.0 3.3 0.0

pH 4.4 b 74.5 0.0 13.8 0.0 1.7 0.3

c 73.1 0.0 15.9 0.0 2.0 0.2

a 43.0 10.0 9.8 0.8 2.6 0.0

pH 5.3 b 43.8 6.9 10.0 0.4 2.7 0.0

c 39.2 6.9 10.7 1.0 3.2 0.0

a 95.8 27.6 7.9 0.1 6.4 0.0

pH 6.1 b 111.9 35.3 7.4 0.2 5.3 0.0

c 102.2 (160.7) 7.7 0.3 5.3 0.1

a 95.8 27.6 7.9 0.1 6.4 0.0

pH 6.7 b 111.9 35.3 7.4 0.2 5.3 0.0

c 102.2 (160.6) 7.7 0.3 5.3 0.1

Appendix

-16

KOH Replicate DOC COC Fe

diss

coll

[mg l

-1

] [mg l

-1

] [µg l

-1

] [µg l

-1

]

a 19.4 1.0 106.6 0.0

pH 3.5 b 18.4 1.6 102.6 33.2

c 19.0 1.9 90.4 (759.1)

a 52.6 16.4 417.5 108.5

pH 4.3 b 50.5 21.7 477.6 166.2

c 51.3 10.9 (197.1) 242.6

a 605.5 104.6 6535.5 665.7

pH 6.2 b 441.1 n.a. 4219.0 (366.7)

c 462.6 n.a. 4163.5 545.6

a 1034.2 136.8 9024.0 762.3

pH 7.1 b 834.2 170.4 6605.0 (2188.9)

c 703.6 289.0 5414.9 991.4

Ca(OH)

Replicate DOC COC Fe

diss

coll

[mg l

-1

] [mg l

-1

] [µg l

-1

] [µg l

-1

]

a 17.9 1.0 (54.0) 6.2

pH 3.7 b 16.2 1.4 19.6 (25.1)

c 15.4 1.1 22.4 10.0

a 20.2 1.9 7.9 30.9

pH 4.4 b 21.6 1.8 8.2 38.1

c 20.3 2.1 8.8 27.4

a 33.1 (3.4) 70.3 10.3

pH 5.3 b 34.2 1.2 74.1 3.3

c 34.2 1.6 71.4 0.0

a 62.4 5.4 251.5 32.4

pH 6.1 b 62.4 2.7 252.3 36.0

c 66.0 0.8 241.7 49.0

a 62.4 5.4 0.0 0.0

pH 6.7 b 62.4 2.7 0.0 0.0

c 66.0 (0.75) 0.0 4.3

Determination of the soil water content (105°C):

Water loss

26.70% KOH treatment

30.40% Ca(OH)

treatment

0.92 0.64

Weight soil moist [g]

1.20

Weight soil dried [g]

0.88